Wynne - Cicero On The Philosophy of Religion PDF

CICERO ON THE PHILOSOPHY OF RELIGION
During the months before and after he saw Julius Caesar assassinated
on the Ides of March,  BC, Cicero wrote two philosophical
dialogues about religion and theology: On the nature of the gods and
On divination. This book brings to life his portraits of Stoic and
Epicurean theology, as well as the skepticism of the new Academy,
his own school. We meet the Epicurean gods who live a life of
pleasure and care nothing for us, the determinism and beauty of
the Stoic universe, itself our benevolent creator, and the reply to both
that traditional religion is better served by a lack of dogma. Cicero
hoped that these reflections would renew the traditional religion at
Rome, with its prayers and sacrifices, temples and statues, myths and
poets, and all forms of divination. This volume is the first fully to
investigate Cicero’s dialogues as the work of a careful philosophical
author.
. . .  is Associate Professor of Classics in the Department

of World Languages and Cultures at the University of Utah. He is a
scholar of later ancient Greek and Roman philosophy and religion,
and a specialist on Cicero, Stoic and skeptical philosophy, and the
philosophy of religion in the ancient world.
Downloaded from https://www.cambridge.org/core. Johns Hopkins University, on 05 Jan 2020 at 04:37:24, subject to the Cambridge Core terms
of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781107707429
CICERO ON THE
PHILOSOPHY OF RELIGION
On the Nature of the Gods and On Divination
J. P. F. WYNNE
University of Utah
University Printing House, Cambridge  , United Kingdom
One Liberty Plaza, th Floor, New York,  , USA
 Williamstown Road, Port Melbourne,  , Australia
–, rd Floor, Plot , Splendor Forum, Jasola District Centre, New Delhi – , India
 Anson Road, #–/, Singapore 
Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of education,
learning, and research at the highest international levels of excellence.
www.cambridge.org
Information on this title: www.cambridge.org/
: ./
© J. P. F. Wynne 
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 
Printed in the United Kingdom TJ International Ltd., Padstow, Cornwall
A catalogue record for this publication is available from the British Library.
 ---- Hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
parentibus
meis
Contents
Acknowledgments page viii

List of Abbreviations x
Introduction: Cicero and the Translation of Philosophy from

Greece to Rome 
Chapter : Cicero’s Project in On the Nature of the Gods and
On Divination 
Chapter : Velleius the Epicurean 
Chapter : Balbus the Stoic and Cotta the Skeptic 
Chapter : Quintus’ Stoic Case for Divination 
Chapter : Marcus’ Arguments against Divination 
Chapter : Marcus’ Stance on the Central Question 
Appendix : Terminology in DND and Div. for Religious Virtues

and Vices, and Greek Equivalents 
Appendix : Velleius’ Strategies against his Opponents 
Appendix : Balbus’ Classification of the Gods 
Bibliography 
General index 
Index locorum antiquorum 
vii
Acknowledgments
This book began as a  dissertation at Cornell. I would like to thank

my doctoral committee members, Terence Irwin and Hayden Pellicia. To
my committee chair Charles Brittain, I owe a particular and ongoing debt
of thanks. Other Cornellians, present or former, to whom I owe thanks
from some part of the genesis of the book include Tobias Torgerson,
Aaron Kelsh, Erik Kenyon, Brent Hannah, Anthony Hunter, Scott
MacDonald, Michael Fontaine, Tad Brennan, and Gail Fine.
I am very lucky in my colleagues and students (again present or former)
at Northwestern, a wonderful place to study and to teach the classics and
philosophy. My sincerest thanks for all their various help, kindness, advice,
undeserved patience, and intellectual company to Sara Monoson, Richard
Kraut, Ann Gunter, Will West, Robert Wallace, Marianne Hopman,
Reginald Gibbons, Francesca Tataranni, David Ebrey, John Schafer,
Baron Reed, and Kenneth Seeskin.
Malcolm Schofield was kind enough to read a draft of this book and
discuss it with me at length. I am very grateful for this experience, and for
all the ways in which he improved my work.
The ‘Second Saturdays’ work-in-progress group of the Chicago Area
Consortium in Ancient Greek and Roman Philosophy read and com-
mented on a number of pieces relevant to this book. My thanks to the
ancient philosophy community in Chicago, and in the Midwest at large,
among whom are (or were) Elizabeth Asmis, Agnes Callard, Gabriel
Richardson Lear, Constance Meinwald, Emily Fletcher, Jason Rheins,
and Dhananjay Jagannathan. An audience at the  Northwestern–
Pisa conference at the University of Pisa heard and helped to improve an
early version of Chapter .
During my final editing of the manuscript while on leave from North-
western in –, the Tanner Humanities Center at the University of
Utah was kind enough to arrange facilities for me as a visiting scholar.
viii
Acknowledgments ix
The anonymous readers from the Press have improved the book
immeasurably, as has my editor Michael Sharp. He has also shown the
patience of Job.
Despite the many sources of advice and improvement I have mentioned,
I am a stubborn individual, and the errors or infirmities in the book are
firmly and entirely my own.
Above all, to my family ShawnaKim, Jack, and Elizabeth: thanks.
Abbreviations
DL Diogenes Laërtius, Lives of the eminent philosophers

DRN Lucretius, De rerum natura (On the nature of things)
LS Long and Sedley ()
LSJ Liddell, Scott, Jones et al. ()
OLD Oxford Latin dictionary (nd ed.) = Glare ()
RE Paulys Realencyclopädie der classischen Altertumswissenschaft =
Wissowa et al. (–)
SB Shackleton Bailey (–) or ()
SVF Stoicorum veterum fragmenta = von Arnim (–)
TLL Thesaurus linguae latinae (–)
When I cite certain of Cicero’s works central to understanding his philo-

sophical writing, I abbreviate or translate their conventional Latin titles as
shown in the table below. I cite all other works by a commonly used title,
whether in English or in the original language.
My label Latin title Other common titles
Consolation Consolatio
Div. De divinatione On divination
DND De natura deorum (On) the nature of the gods
Hortensius Hortensius
Laws De legibus On the laws
Letters to Atticus Epistolae ad Atticum
Letters to his friends Epistolae ad familiares
Letters to Quintus Epistolae ad Quintum
fratrem
On augury De auguriis
On duties De officiis On obligations
On ends De finibus (bonorum et On the ends of goods and evils, On
malorum) moral ends
List of Abbreviations xi
(cont.)
My label Latin title Other common titles
On fate De fato
On friendship (Laelius) de amicitia Laelius
On glory De gloria
On invention De inventione
On old age (Cato) de senectute Cato
On the orator De oratore On the ideal orator
On the parts of De partitione oratoria
rhetoric
Republic De re publica
Timaeus Timaeus
Tusculans Tusculanae disputationes Tusculan disputations
What I cite, following convention, as the Academica, is a portmanteau text.

Its books are the surviving parts of two of the drafts Cicero made of his
dialogue about the skeptical Academy:
Academica Part of the first book of the second of the two drafts. Cicero
book  called this draft the Academic books (Academici libri). Scholars
sometimes call it the Academica posteriora.
Academica What Cicero called the Lucullus, one of the books of the first
book  draft. Scholars sometimes call this draft the Academica priora.
Introduction: Cicero and the Translation
of Philosophy from Greece to Rome
The Romans did not understand their own religion. They were the heirs to
immemorial practices in honor of their gods. But when they paid the gods
cult, they did not know the meaning of what they did, nor the nature of
the gods they worshipped. The result was that they moved like strangers
through their own city, looking for a way to feel at home.
Or that is how Cicero paints it (p.  below). The problem was one for
intellectuals, perhaps a small class, who wanted not only to practice their
religion, but also to understand it in a rigorous way. One remedy was the
antiquarianism of Varro, who aimed by historical study to recover the
intentions of the religion’s founders. But Cicero’s suggestion was to apply a
new resource to the problem: Greek philosophy.
In the event, Cicero carried through this project in two dialogues, On
the nature of the gods and On divination (De natura deorum, thus DND,
and De divinatione, thus Div.). In this book, I aim to interpret the whole
and some parts of these texts themselves, not the Hellenistic philosophy for
which they are excellent sources, nor what Cicero’s own sources might
have been. Now I do not argue that in these dialogues Cicero wrote, or
aimed to write, as an “original philosopher.” But I do think that he shaped
the drama, the characters, their speeches, and his own authorial com-
ments, as parts of literary and philosophical wholes. I think that he thus
suggested a unique and interesting answer to the dialogues’ questions.

A third dialogue, On fate was planned to complete this sequence. Cicero indeed wrote an On fate,
but not as he had planned it. I argue below (p. –) that we may read DND and Div., without On
fate, as a completed project.

Was Cicero an “original” philosopher? If the question is whether he brought forth ideas or arguments
that he did not get from anybody else, then it seems probable that he did so in his writings on
political philosophy, especially in the Republic, Laws, and On friendship. But it is rare to find
evidence that he does so in what I call the Late Sequence dialogues. Nor was it his objective to do
so. But why must the philosopher who always produces original ideas be valued higher than
somebody who presented or used the ideas and arguments of others in an original way? Both
promote wisdom, which is why universities employ historians of philosophy today.
of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781107707429.001
 Cicero and the Translation of Philosophy from Greece to Rome
My goals are first, to interpret that answer, and second, to show by
example that Cicero’s philosophical dialogues may fruitfully be read in
this way.
A century ago, when philological science held Cicero’s philosophical
writing in low esteem, it would have looked silly to write this book. The
best scholars, and what seemed the best evidence, suggested that Cicero’s
philosophical dialogues would not reward a reader who looked for unity of
literary form and a fertile intellectual project. It was even said that Cicero
was unable or unwilling fully to understand the philosophy he wrote
about. My book is one attempt to make good on a slow rise in scholarly
opinion of Cicero since then. In this introduction I shall collect some
reasons to allow the assumption that Cicero’s dialogues might sustain my
sort of treatment.
Cicero’s philosophical dialogues fall into two kinds. One kind includes
dialogues in which Cicero’s skepticism, and the Hellenistic philosophers
whom he treats, are rarely drawn to the reader’s attention (Republic,
Laws, On old age, On friendship). The second kind includes dialogues
in which Cicero points to his skeptical agenda and to his Hellenistic
material. These are the sequence of dialogues marked out by Cicero at
Div. .: Hortensius, Academica, On ends, Tusculans, DND, Div., and On
fate. I shall call the latter set of dialogues the “Late Sequence” because
these works were written in sequence towards the end of Cicero’s life.
I shall draw on evidence from and about the other dialogues when
I explore Cicero’s approach to the dialogue form in general. But the
arguments of this introduction are intended to apply principally to the
Late Sequence.
I note one further convention here. Cicero puts “himself” into his
dialogues as a character, but I think we should not simply assume that
these characters reflect accurately the historical Marcus Tullius Cicero.
Rather than labor these distinctions, I shall from now on call Cicero the
historical person, or his authorial voice, “Cicero,” but his avatar in each
respective dialogue “Marcus.” His brother’s avatar, who appears in Div., is
“Quintus.”
. An Older View of the Late Sequence

[Cicero] sive diffitetur sive confitetur, omni tempore secutus Ennianum illud
philosophari est mihi necesse, at paucis, nam omnino haut placet
et degustandum non ingurgitandum sibi ratus, cum senex civibus philosophia
explicanda utilis esse statuisset, tanta facilitate et celeritate libros scripsit edi-
ditque, ut Graecorum exemplis vix legendis vacaret. miranti vel Attico respondit
. An Older View of the Late Sequence 
apographa sunt, minore labore fiunt: verba tantum adfero, quibus abundo.
bene profecto actum nobiscum esset, si optimorum librorum vel Panaetii ac
Posidonii apographa nobis reliquisset. at nego Ciceronem eum fuisse qui philo-
sophum Graecum veritatem spinosa arte exputantem et in viscera penetrantem
sequi aut vellet aut posset. foro natum erat hoc ingenium, non scholae.
[Cicero], whether he admits or denies it, followed Ennius’ line every time,
I must philosophise, but only a little, for to philosophise entirely displeases,
and reckoned that he should taste, but not gorge himself. When, as an old
man, he had decided to make himself useful to his fellow citizens by
explaining philosophy, he wrote and published books so quickly and glibly
that there was scarcely time left to read his Greek models. He replied to
Atticus, who was perhaps wondering about this, “they are copies, they come
out with relatively little work. I add only the words, with which
I overflow.” Certainly he would have done well by us, had he left us
‘copies’ of the best books of, say, Panaetius and Posidonius. But I deny that
Cicero was a man who was either able, or wished, to follow a Greek
philosopher thinking out the truth with his thorny art, and penetrating to
the guts of it. Cicero’s genius was born for the forum, not for the
lecture hall.
If these words were only a little less kind, you might guess they came from
Theodor Mommsen, Cicero’s chief deprecator (see p.  below). In fact
Hermann Usener wrote them, in the preface to his Epicurea (p. lxv).
The attitude to Cicero that Usener reports is important to the story of
Cicero’s philosophical writings, because it represents well the attitude that
prevailed during a crucial period in the history of classical scholarship,
during the nineteenth and early twentieth centuries. Epicurea was, and still
is, a landmark collection of texts for the study of Epicurus. Usener
influenced Hermann Diels, author and editor of Doxographi graeci and
Die Fragmente der Vorsokratiker, and Hans von Arnim, editor of Stoicorum
veterum fragmenta (SVF). Doxographi graeci was a foundational contribu-
tion on our sources for philosophy of the centuries before and after Plato
and Aristotle. Die Fragmente der Vorsokratiker is still the standard collec-
tion of texts for the Presocratics, SVF for the older Stoics. In his preface to
SVF, von Arnim signs on to the picture of Cicero “whose authority in

Usener here flatly denies Cicero’s own claim at Tusculans .. Usener supplements Cicero’s version
of Ennius’ line from Gellius NA ...

Usener quotes Letters to Atticus .. =  SB. See below pp. –.

For example, while David Konstan’s Stanford Encyclopedia of Philosophy bibliography (Konstan
) lists Arighetti’s as the “best edition,” Usener is the “most complete collection of fragments.”

For Usener’s influence on Diels, and on the project of Doxographi Graeci in particular, see Mansfeld
and Runia () vol. pp. –. For his influence on von Arnim, see SVF p. iii.

On Diels and Doxographi Graeci see Mansfeld and Runia (–) vol. Chapter .
 Cicero and the Translation of Philosophy from Greece to Rome
philosophical matters Usener has weighed in Epicurea.” Diels, meanwhile,
seems to compliment Cicero when he quotes him for the epigraph to
Doxographi Graeci: “it is one of slow wit who follows little streams and
does not see their sources.” But the compliment delivers an insult, since
for Diels Cicero is a little stream whom, at best, we can follow to his
sources. “Excessive ignorance of ancient philosophy” hindered Cicero
and he followed the Greek “tentative and anxious like a blind man.”
When we confront a philosophical speech in Cicero with what Diels
thought was its Greek source, we expose “either Cicero’s mendacity or
his stupidity.” For those of us who rate Cicero’s philosophical writing
higher than this, it is confounding to see first-rate scholars attack him in
such terms. It helps to recall how different from our own was the intellec-
tual climate of nineteenth-century Germany in which they trained.
Mommsen had a theory that civilizations rise and fall and that the
ancient Mediterranean had a long story with just such an arc, of which
in the History of Rome he wrote the last chapter. The final pages are a sort
of cabinet of freaks in which Mommsen collects proofs that by the end of
the Roman Republic, ancient civilization was decadent. Here we encoun-
ter Cicero’s Late Sequence, written “with equal peevishness and precipita-
tion.” Mommsen thought the dialogues were in “rude imitation of the
popular writings of Aristotle,” “stitched together” by Cicero from which-
ever writings of Hellenistic philosophers “came or were given into his
hand, into a so-called dialogue.” Cicero exhibited
. . . that sort of bungling, which a man of letters, who has not attained to
philosophic thinking or even to philosophic knowledge and who works
rapidly and boldly, shows in the reproduction of dialectic trains of thought.
In this way no doubt a multitude of thick tomes might very quickly come
into existence—“They are copies,” wrote the author himself to a friend who
wondered at his fertility; “they give me little trouble, for I supply only the
words and these I have in abundance.” Against this nothing further could
be said; but any one who seeks classical productions in works so written can
only be advised to study in literary matters a becoming silence.

cuius in philosophicis auctoritatem in Epicureis examinavit Usener, p. xix.

tardi ingeni est rivulos consectari, fontes rerum non videre, from Cicero, On the orator, ..

Diels probably meant to call to mind Academica ., where Varro’s character tells Marcus that he
did not write philosophy in Latin because he instead tells friends who are interested to go directly to
the Greeks, and “to drink from the sources rather than follow little streams,” ut. . . a fontibus potius
hauriant quam rivulos consectentur.

nimia vetustae philosophiae ignorantia; vacillans et anxius ut caecus. Diels () , describing
Marcus’ summary of Greek opinions at Academica ..

vel fraudem vel socordiam, Diels () .

Dickson’s translation, sanctioned by Mommsen, vol.  pt.  () p. .
. A More Positive View of Cicero’s Dialogues 
Although Epicurea and SVF appeared half a century after Mommsen’s
History, the sort of views about history that Mommsen represented are
visible in Usener and von Arnim’s attitudes, as they were from Diels
when he published Doxographi graeci in . Now, these were brilliant,
and most responsible, scholars. As we shall see, there was what seemed to
be good evidence for their view of the Late Sequence. Thus, although
I shall reject their view as others have now done, we should remember
one good reason for these scholars’ extravagance against Cicero was that
it was in keeping with some big ideas, and the exemplary scholarship, of
their times.
. A More Positive View of Cicero’s Dialogues

Since the last quarter of the twentieth century there has been a revival in
the understanding of Hellenistic philosophy. This revival owes a large debt
to the philologists I have mentioned, and to others like them, who
flourished a century earlier. For it has shared one part of their approach,
the thought that, to quote Long and Sedley (LS p. ), “Hellenistic
philosophy is a jigsaw.” So it is, if we define Hellenistic philosophy more
strictly as the philosophy of the Greek schools after the death of Aristotle
and before Cicero began to write. For, with the exception of some texts
from Epicurus, no complete piece of Greek philosophical writing from
that time survives. Yet the thought that was contained in the lost texts is
recoverable. If we cut up Cicero, and many other sources, into jigsaw
pieces, we can rearrange them by school, thinker, and topic, so as to make
pictures of the Hellenistic period. “So far as I could,” wrote von Arnim, “I
projected in [SVF] a full and accurate image of Chrysippus’ philosophy.”
When we open volumes II and III of SVF, we see that this “image” is made
of small jigsaw pieces cut from many texts. Painstaking collection by
scholars like von Arnim ultimately bore fruit in today’s Hellenistic revival,
because it turned out to be true that the “images” they made could yield
coherent and fertile reconstructions of Hellenistic thought. We should
keep cutting up Cicero for such purposes. I hope that work like mine can
make the pieces more useful still.

It is not my intention here to trace in detail the history of influence among these scholars. For
introductions to this history see the relevant articles in Briggs and Calder ().

Chrysippeae philosophiae accuratam atque plenam imaginem, quatenus potui, hoc opere adumbravi,
SVF p. iii.
 Cicero and the Translation of Philosophy from Greece to Rome
But we have seen that the nineteenth- and early twentieth-century
scholars matched their care for the philosophy to which Cicero is a witness
with a negative view of Cicero’s own philosophical acumen. By and large,
historians of philosophy do not share that negative view today. When
scholars have built reconstructions of Hellenistic philosophy in large part
on Cicero’s reports, those reconstructions have seemed to be well
founded. It would be implausible now to think that Cicero was stupid
or mendacious in writing about philosophy. That has suggested to many
that Cicero’s texts might be worth a look not only as sources, but also as
objects of study in themselves. We may see Hellenistic philosophy, defined
more loosely as the whole life of the Hellenistic schools in antiquity, not in
fragments, but available to us in rich and continuous texts. Sure enough,
for decades there have been rallying-cries to the study of Cicero’s dialogues
on their own merits. A number of scholars have taken up the challenge,
explicitly or implicitly. The jigsaw approach to Cicero has remained
dominant, but in interpreting DND and Div. as I do I am not doing
anything unheralded or unprecedented – despite Mommsen’s warning.
One positive view of the Late Sequence I shall call the encyclopedia view.
The encyclopedia view takes its cue from Cicero’s success as a philosoph-
ical source. It supposes Cicero’s dominant goal in the Late Sequence to be
the opening up of philosophy to a new Roman audience by the compos-
ition of a philosophical “encyclopedia.” On this reading, the topics covered
in the dialogues form a curriculum that covers a representative selection of
Hellenistic philosophical topics. A given speech represents Cicero’s version
of the Stoic view on the dialogue’s topic, or the Epicurean view, or the
skeptic’s reply, while a given speaker is cast to represent some particular, or
an ideal, member of a school.

Some examples of studies where Cicero has been vital in reconstructing even relatively technical
aspects of Hellenistic thought: Bobzien () who draws extensively on On fate and Div., Brittain
() who draws on Academica, Graver () who draws on Tusculans, and many of the essays in
Sedley ().

Some rallying-cries: Boyancé () and preface to Boyancé (), Douglas () and (),
Striker (), Powell’s introduction to his (a), Smith (), Schofield (). In one way or
another these works also make a start on reading the dialogues as they recommend we should.

Schultz’s () commentary on Div.  similarly takes Div. seriously as a unified whole (see
pp. –). Coverage of all the philosophical dialogues is in Süss (), although he regards the
quality of the dialogues as uneven, and Woolf (). Woolf reads many of the dialogues carefully
as Cicero’s own projects. Atkins () is not on Late Sequence dialogues, but reads the Republic
and Laws in a similar spirit, as does Zarecki ().

In addition to readings more focused on the philosophical purposes of the dialogues, there is recent
work exploring Cicero’s artful use of the dialogues as part of his rhetorical and philosophical life. See
Steel () Chapter , Gildenhard (), Baraz ().

A sophisticated defense of the Late Sequence using the encyclopedia view appears in Striker ().
. A More Positive View of Cicero’s Dialogues 
There are good reasons to think that to make his dialogues useful as an
encyclopedia was one goal to which Cicero gave thought. For one sort of
audience whom the dialogues seem to envisage is unlearned Romans who
could not get their philosophy from the Greeks. These Romans, before
Cicero, could read only works of Epicurean authors in Latin whom Cicero
thought incompetent. (See pp. – below.) Of Cicero’s Late Sequence
writings, the work most obviously intended to address this need was the
Hortensius, for it was meant to turn new readers to philosophy.
To further strengthen the conclusion that one purpose of the Late
Sequence was as an encylopedia, we find that the Sequence appears to
have a curricular order and completeness. Cicero lists the Sequence at Div.
.–. When we read this list, we see that it begins with the call to
philosophy in Hortensius, before turning to Cicero’s favored school
(Academica) and the ethical foundation that philosophy offers (On ends).
Next, the Tusculans offer doses of philosophical medicine to soothe
common sources of distress – the fear of death, for example – in a way
that should appeal to the beginner at philosophy at least as much as to the
savant. Further, the Sequence overall is probably meant to cover some of
each of the three conventional parts of Hellenistic thought: logic
(Academica), ethics (Tusculans, On ends), and physics (DND, Div., On
fate). Indeed, it was Cicero’s declared hope in Div. . that he would
have time to treat every topic of philosophy in Latin. In sum, the Late

For evidence of attention to unlearned readers, see: Marcus’ reply to Varro at Academica .;
Tusculans .– which paints the Romans as generally unversed in philosophy; Div. ., which
envisages transmission of philosophy to Cicero’s fellow citizens. Tusculans . speaks of “opening
up” the sources of philosophy and Div. . describes Tusculans as “opening up” matters necessary
for living well. Tusculans .– imagines readers with a liberal education, whose refined literary taste
means that they might read Plato and other Socratics, but who would never read ugly texts by
Epicurus or the Latin Epicurean authors.

Arguably, the end of the Lucullus (Academica .) signals a plan whereby the characters of the
sequence Hortensius, Catulus, and Lucullus were to keep going with a series of discussions of
the ethical and physical topics raised in the latter part of Lucullus. So perhaps there was a plan
that the Academica was to cover logic. See Griffin () – with her references, and note her
counter-arguments.

But it is also possible that this impression is illusory. Although the epistemological content of
Academica makes it look like a “logical” work to us, Cicero never describes it thus. He describes it as
his advocacy of the New Academy, which might mean that he thinks it gave an Academic treatment
of philosophy in general (see Griffin ). Neither On ends nor the Tusculans is anything close to a
comprehensive treatment of ethics. The Sequence gives little coverage to politics, a philosophical
topic in which Cicero certainly took an interest – did Cicero think the Republic and Laws could fill
that gap, even though they are not in the Late Sequence style? DND, Div., and On fate cover only a
limited part of physics, and not all their material is obviously physical rather than logical, or ethical.
The sketched preface in Timaeus suggests Cicero toyed with some more general treatment of physics
from a skeptical point of view.
 Cicero and the Translation of Philosophy from Greece to Rome
Sequence seems designed with one eye to new philosophical readers who
wanted to use it as something like a comprehensive curriculum. Thus the
encyclopedia view contains truth to the extent that making a philosophical
encyclopedia was one of Cicero’s goals for the Late Sequence.
But Cicero had other goals, which were often more important to him
than the encyclopedia view alone would suggest. There is another part of
the audience that the dialogues envisage: the learned. These are Romans
who already have some, or much, philosophy from the Greeks. To see who
these Romans might be, it helps to remember how Cicero’s texts were
disseminated. We know from his letters that Cicero sent off manuscripts
for the copyists whom Atticus retained. Thus there was a system of
production and distribution not entirely within Cicero’s control. But this
was little like modern publication and distribution of books. The immedi-
ate audience for the dialogues would be people who were part of Cicero
and Atticus’ intellectual and social circle. Such people very often had some
degree of Greek education and philosophical learning. Among these were
readers like Atticus himself, Varro, or Brutus, whose knowledge of

Cicero envisages these readers who use his dialogues to get acquainted with philosophy, perhaps for
the first time. But were there any such readers? We know that pirate copies were made of some of
the Late Sequence, notably On ends, from drafts in Atticus’ possession. (Letters to Atticus
.a = SB, . = SB) The respective owners of the copies are named as Cornelius
Balbus and Caerellia. The former appears elsewhere in the letters as a politician and not an
intellectual, so perhaps he was someone who wanted to learn about philosophy for the first time
from On ends. Then again, perhaps he made the copies out of politeness to Atticus, or perhaps he
was more learned in philosophy than he appears. Cicero says of Caerellia, an acquaintance to whom
he had owed money (Letters to Atticus .=SB, Letters to Atticus .=SB), that she was
mirifice. . . videlicet studio philosophiae flagrans, “no doubt aflame with a wonderous zeal for
philosophy.” (Letters to Atticus .a=SB) This strikes the eye as sarcastic, and Cicero
certainly resents that Caerellia has read his private draft. But it is also possible that Cicero
thought either that Caerellia was a learned and enthusiastic reader, or that she was unlearned in
philosophy but earnestly hoped to learn.

The obvious evidence of this readership comes from the small circle of knowledgeable
contemporaries who serve as dedicatees and speakers in the dialogues, for example Brutus, Varro,
Atticus, Quintus, Torquatus, or the recently deceased Cato. The sort of “study abroad” in Athens
depicted at the start of On ends Book  was real. Many professional philosophers were living in Italy
under Roman patrons, so that a philosophical education was available to interested aristocrats.
Rawson () surveys much of the evidence for these intellectual circles. Cicero implies that he
was attacked for taking this interest beyond a gentlemanly minimum (On ends .), but he was not
the only one. In his depiction of them Cicero may well exaggerate the degree to which some of these
men had all the arguments at their fingertips, but it is plain from Cicero’s letters and from their own
literary endeavors that they and others at Rome were capable philosophical hobbyists. For the
intellectual culture visible in Cicero’s letters, see McConnell ().

On the distribution of books, see Kenney () and Starr (). Murphy () collects
evidence for the early readership of Cicero’s dialogues.
. A More Positive View of Cicero’s Dialogues 
philosophy was probably comparable to Cicero’s own. In other words,
Cicero writes in the dialogues in part for people who already know about
the material he is presenting, or who at least could look it up, or ask a
philosopher, in Greek. For these readers, a Latin encyclopedia of philoso-
phy would be of little use.
This philosophically learned readership is important to bear in mind
when we try to understand the Late Sequence. For it explains a common
phenomenon faced by philosophical interpreters. Often Cicero, or a
character, will merely mention or allude to some argument or philosoph-
ical position, or give a very spare account of it, so that modern interpreters
need to turn to other Hellenistic sources to understand what Cicero, or the
character, is talking about. I often follow that procedure in this book. The
trick makes dramatic sense: inside the fictional world of the drama, both
the character and his interlocutors are supposed to know their philosophy.
But as to why Cicero does this for his readers, if you took only the
encyclopedia view of the dialogues, you would face a puzzle. Why would
Cicero refer to ideas in a way that will confuse his readers who would be, ex
hypothesi, ignorant of those ideas? But once we have in mind the learned
readers, we may suppose that Cicero sometimes merely mentions argu-
ments or views because there is a part of his readership, on whom he is
often focused, who will get the reference, or who can easily follow it up.
But then we might wonder, if Cicero wrote in part for readers to whom
it is not his purpose to teach the material, what were they supposed to get
out of the Late Sequence? And, if there is more to the Sequence than a
philosophical encyclopedia, are even unlearned readers given more than
just an introduction to the material? Let us look at some of Cicero’s
answers to these questions.
In his preface to On ends, Cicero distinguishes four types of his critics
(.): (a) people who dislike philosophy entirely, (b) people who think it
should not be done too much, (c) people who are educated in Greek and
who would prefer to read philosophy in that language, and (d) people who
think he should write about something else. It is to people of type (c) that
he gives by far his most detailed reply. (On ends .–) But these are not
people who are hostile to, nor necessarily ignorant of, the philosophical
content of the Late Sequence. Against them Cicero argues:

For the wealth of philosophical culture in Cicero’s letters, see McConnell (). For more specific
evidence that some recipients of Cicero’s letters could be expected to get philosophical jokes, see
Griffin (). For his Epicurean friend Matius, see Griffin ().
 Cicero and the Translation of Philosophy from Greece to Rome
nam si dicent [sc. his opponents] ab illis [sc. the Greeks] has res esse tractatas,
ne ipsos quidem Graecos est cur tam multos legant quam legendi sunt. quid
enim est a Chrysippo praetermissum in Stoicis? legimus tamen Diogenem,
Antipatrum, Mnesarchum, Panaetium, multos alios, in primisque familarem
nostrum Posidonium. quid? Theophrastus mediocriterne delectat cum tractat
locos ab Aristotele ante tractatos?
For if they will say that the Greeks have already covered these topics,
then there is no reason to read even so many of the Greek authors
themselves as one is supposed to read. For what, in the case of the Stoics,
did Chrysippus leave out? Yet we read Diogenes, Antipater, Mnesarchus,
Panaetius, and many others, not least our friend Posidonius. Does
Theophrastus give us only moderate pleasure, when he deals with topics
already covered by Aristotle? (On ends ., translation from Annas and
Woolf ())
This group of critics, then, have quite a reading list, and Cicero implies
(“we read”) that some of them have read it. They do not need to be filled
in on the material in the Late Sequence. Instead, Cicero needs to convince
them of what he argues in On ends .: that his own project is worth
writing (see also p.  below).
Turning to the philosophically unlearned readers, we find that even the
most hostile among them, the critics of type (a), are not wholly unlearned
(non admodum indoctis, On ends .). In the Tusculans Cicero applies
another word to his target audience: “educated” (eruditi).
sed eos, si possumus, excitemus, qui liberaliter eruditi adhibita etiam disserendi
elegantia ratione et via philosophentur.
But let us rouse up, if we can, those with a liberal education, both to
philosophise methodically, and also to employ a polished style of discourse.
(Tusculans ., translation from Douglas (), modified)
Since Cicero thinks that Roman educated culture in his day extended to
oratory and poetry but not to philosophy, it is clear that he means to

Perhaps this is the “royal” we. But the argument seems to depend on his opponents’ granting at least
that one should read all the Greeks on the list, so it is more likely that Cicero graciously suggests that
his opponents do read them.

I read philosophentur, an emendation by Sauppe with some weak manuscript support, adopted, for
example, by Kühner () and Dougan (), rather than philosophantur, which has the support
of by far the stronger part of the manuscript tradition. The context shows that Cicero’s purpose is to
stimulate those who have an education in eloquence to use it in a new way, to write well about
philosophy, rather than to excite to greater efforts some people who already write well on the
subject.
. A More Positive View of Cicero’s Dialogues 
inspire some people educated in oratory and poetry to bring their existing
skills to bear in a new area.
If we bear in mind all these readers whom Cicero envisages for the Late
Sequence, we gain the following picture of his goals in writing. He hopes,
first, to produce a comprehensive example of how philosophy could be
translated. That will inspire his Greek-reading, philosophically learned
readers. Next, he hopes to give an example of how philosophy can be
brought to life in letters and in its application to Roman culture. That will
inspire his educated readers in general. Neither of these goals is incompat-
ible with the encyclopedic element in the Sequence. Indeed, addressing
both goals over a full range of philosophical material will help to convince
and to inspire. But these goals yield reasons for Cicero to do more than be
encyclopedic. Let us see what more Cicero might try to do.
First, we shall look at Cicero’s translation project. By “translation”
I mean translation from Greek into Latin, but also from a Greek literary
and cultural context to a Roman one.
It is plain on all views of the Late Sequence that Cicero took some
existing material – Hellenistic philosophy in Greek – and then did some-
thing to it. But Cicero’s own metaphor for what he did to it is not
transcription, translation, or even teaching. It is that he “illuminates”
philosophy (illustrare). Specifically, he will illuminate it, for the first
time, with “the lamp of Latin letters.” (lumen litterarum Latinarum,
Tusculans .) This metaphor bears two meanings, both of which
I think Cicero intends. The first is that by illuminating philosophy with
Latin he makes it visible for the first time to those who can only “see”
Latin: monoglot Romans. Thus he will “open up the sources” of philoso-
phy. (fontes aperiemus, Tusculans .) But the second meaning is that he
will benefit philosophy by, so to speak, showing it in a new light. What do
I mean by this latter point?
Cicero is optimistic about the prospects for translation into Latin.
In replying to Romans who oppose translation, he says in the preface to
On ends:
. . . ita sentio et saepe disserui, Latinam linguam non modo <non> inopem, ut
vulgo putarent, sed locupletiorem etiam esse quam Graecam. quando enim

Here I agree with Smith () and Schofield () that Cicero’s emphasis in the Tusculans on
rhetorical sophistication in philosophical writing is important to how we should understand his
project.

Academica .; Tusculans ., .; Div. .. Lucretius uses the same metaphor, DRN ..
 Cicero and the Translation of Philosophy from Greece to Rome
nobis, vel dicam aut oratoribus bonis aut poetis, postea quidem quam fuit quem
imitarentur, ullus orationis vel copiosae vel elegantis ornatus defuit?
. . . my view is, as I have often argued, that, far from lacking in resources,
the Latin language is even richer than the Greek. When, after all, have we,
or rather our good orators and poets, lacked the wherewithal to create either
a full or a spare style in their work, at least since they have had models to
imitate? (On ends ., translation from Annas and Woolf ())
Now you might take Cicero to say something silly here: that Latin as he
found it was already better equipped for philosophy than was Greek. That
was obviously not true. Admittedly, Cicero found that sometimes pre-
philosophical Latin could throw up happier terminology than Greek, as
one might expect from any language. But most of the time he had had to
work hard to find or to coin Latin equivalents for Greek terms of art. The
Greeks themselves, he points out, had needed to coin technical terms, so
he could hardly have expected to find such words ready-made in Latin.
(On ends .) Thus I think that Cicero’s point is rather that given someone
to imitate, good Latin authors will in the future do well at philosophical
writing. He plans to give them someone to imitate. Thus he must think
not that Latin is already better equipped for philosophy than is Greek, but
rather that it starts with more raw materials out of which make philosoph-
ical terms of art (cf. Tusculans .–).
But Cicero thinks that the Latin language is not the only Roman
advantage he can bring to philosophy. The prefaces to the five books of
the Tusculans amount, in no particular order, to a theoretical history of
philosophy, and a picture of a place for Rome in that history. Let us survey
this story. The human soul, says Cicero, is sick and in need of healing.
(Tusculans .) Things would have been all right if only we were born able
to grasp the nature of things, and thereby to lead our lives according to
nature. (Tusculans .) But, as it is, our natural endowment consists only

Cicero’s fellow pioneer, Lucretius, is clear on Latin’s inadequacies, DRN .–. Some of the
technical vocabulary that Latin did already have was borrowed from Greek, as Cicero of course
admits. (On ends .)

In the Tusculans Marcus distinguishes labor (toil) from dolor (pain). He points out that both words
would be rendered in Greek by the ambiguous πόνος. “This time you’re poor in words, o Greece,
who always think you overflow with them,” he exclaims. (o verborum inops interdum, quibus
abundare te semper putas, Graecia!, Tusculans ..) In Div. Cicero is pleased that Latin gives
divinatio an august connection with the divi, the gods, while Greek μαντική is derived from the less
salubrious μανία, meaning “frenzy” or “insanity.” (Div. .)

On Cicero’s ambitions as a translator, and late Republican Latin’s ability to carry the project, see
Poncelet (), Powell (b), Moatti () –, and White ().
. A More Positive View of Cicero’s Dialogues 
in “tiny little sparks” (parvulos. . . igniculos) of understanding, the “seeds of
the virtues” (semina. . . virtutum). (Tusculans .) Furthermore, these
starting-points are degraded under the social influence of those who have
not done philosophy, and end up deeper in confusion and vice. (Tusculans
.–) But there is a medicine for the soul: philosophy. (Tusculans .) It
is the method by which we may heal ourselves, working (presumably) from
the first glimmers of understanding with which we were endowed and
which have survived social distortion. (Tusculans .–) The idea of
wisdom, Cicero will go on to say, is immemorial, and there were always
people regarded as wise, whether in the age of myth, like Ulysses and
Nestor, or in history, like the Seven Sages. (Tusculans .) Since we are not
naturally wise (as he told us in Tusculans .–), the implication is that
there must have been people doing something like philosophy in prehis-
toric times, though they did not call it that. (Tuculans .) Philosophy so
called began with Pythagoras and its tradition continued through
Anaxagoras, Archelaus, Socrates, Plato, the Academy, and hence to Cicero.
(Tusculans .–) Cicero will now introduce it fully and openly into
Roman letters for the first time. (Tusculans .–) If he succeeds, then the
Romans who take up his challenge will start the process of looking to their
natural glimmers of understanding, undoing vicious social influence, and
working from there towards wisdom. But Cicero hints that Romans start
with two sorts of advantage. These advantages cannot come from natural
endowment, since that must be the same in all humans. They must lie in
the social influences that Romans suffer.
The first advantage is that many of Rome’s institutions are younger than
the recent tradition of philosophy. Cicero – with due caution – conjectures
that since Pythagoras was influential in Italy at the time when Rome threw
out the kings, perhaps some of the Republic’s institutions were influenced
by him or his followers. (Tusculans .–) As was customary among
Pythagoreans (he says), these influences were not publicized, but expressed
indirectly. For example, the Pythagoreans used to contemplate song and
harp music, and early Romans sang to the flute – for the Twelve Tables
even had to ban singing that led to “another’s injury.” (alterius iniuriam,
Tusculans .) Later generations with some philosophical interest, includ-
ing notable historical figures like Laelius and Scipio in the second century
BC, showed similar interests in their lives, but did not write about them.
(Tusculans .) So the social influence to which centuries of Romans had
been subject, distorting their natural glimmers of understanding, was
perhaps less damaging than some, since it may itself have been shaped,
quietly, by philosophy.
 Cicero and the Translation of Philosophy from Greece to Rome
The second advantage for Romans is more straightforward: “I have
always held the view that in their own discoveries our Roman forebears
showed greater wisdom than the Greek, while what they received from the
Greeks they made better.” (Tusculans .) The evidence for this assertion
is that Roman “customs and rules of living” (mores et instituta vitae) are
better, as are Rome’s constitution, laws, and warfare. (Tusculans .) The
Romans are better at gravitas, at consistency (constantia), at magnanimity
(magnitudo animi), at ethical rectitude (probitas), at good faith (fides), and
at virtue in general. (Tusculans .) These are all virtues one might expect
to gain from philosophy. But, before any doctrina, any teaching (Tusculans
.), “What they had from nature, not acquired from books, is beyond
comparison with Greece or any other nation.” (Tusculans .) As I say,
in this philosophical context it is unlikely that Cicero means that Romans
are better at birth than are Greeks. It is rather that Romans, being relatively
unschooled, are closer to nature as they grow up. In consequence, they
have better retained their natural seeds of the virtues. Thus they have
pursued what they pursue better than have the culturally advanced Greeks.
Then on two counts, the rude Romans are well placed to start on
philosophy. Just as the Latin language has resources waiting to be
developed, so the Romans are ready to become a philosophical nation.
Did Cicero really think that Roman culture had these advantages? They
look like a patchwork of jingoism, run up for exactly the kind of educated
but parochial Roman Cicero hoped to inspire. He admits that the “his-
tory” of philosophy at Rome in the preface to Tusculans Book  is
guesswork from circumstantial evidence. Further, as we shall see (pp.
–), a radical skeptic like Cicero has reason to regard the picture of
philosophy he paints with suspicion: if he were sure we could “heal”
ourselves philosophically using some helpful endowment of natural con-
cepts, he might be less inclined to skepticism. Nevertheless, that he offers
his Roman readers this place in philosophical history tells us something
about how he intends them to be inspired by the Late Sequence. There
was also a bedrock of plausibility (which is not to say truth) under Cicero’s
salesmanship. Many Hellenistic philosophers would share the view that
philosophy helps to develop the seeds of virtue we naturally share, in order
to cure the foolishness from which we suffer and which cultural influences

. . . meum semper iudicium fuit omnia nostros aut invenisse per se sapientius quam Graecos aut accepta
ab illis fecisse meliora . . . Translation from Douglas (), modified.

iam illa, quae natura, non litteris adsecuti sunt, neque cum Graecia neque ulla cum gente conferenda.
Translation from Douglas ().
. A More Positive View of Cicero’s Dialogues 
tend to reinforce (see pp. – and – below). The story of a
philosophical tradition stretching back at least to Socrates is correct.
Finally, Cicero of course had real respect for the traditions of Rome,
despite the troubles of his own day, so that what he says may have seemed
more plausible to him than it does to us.
But Cicero’s hopes for a Roman philosophical tradition do not seem to
be invested in Romans discovering new ideas or arguments. He never
envisages a class of “original” Roman philosophers. Instead, he looks
forward to taking Greek ideas and to making a Latin philosophical
literature:
quod si haec studia traducta erunt ad nostros, ne bibliothecis quidem Graecis
egebimus, in quibus multitudo infinita librorum propter eorum est multitudi-
nem qui scripserunt; eadem enim dicuntur a multis, ex quo libris omnia
referserunt: quod accidet etiam nostris, si ad haec studia plures confluxerint.
But if these studies are brought over to the Romans, we shall not even need
the Greek libraries in which there is an endless throng of books, because of
the throng of those who have written them. Many writers say the same
things, and as a result they have stuffed the world with books. This will
happen to the Romans too, if many of us flood into these studies.
(Tusculans ., translation from Douglas (), modified)
Cicero does not regard innovation in content as a universal feature of the
tradition of philosophical literature. But, as the last sentence here indicates,
he does not disapprove of such repetition. The mockery of Greek grapho-
mania is affectionate. If he accepts as plausible something like the picture
of philosophy we have just seen, this reaction is explicable: the philosoph-
ical tradition communicates a method for healing oneself by moving from
the same natural glimmers of understanding to the same kind of wisdom.
So, while the Late Sequence is full of the attribution of particular argu-
ments to particular philosophers, it is not surprising that the general course
of philosophy should be similar over time, and that ideas and arguments
should be repeated. If that is how he thinks philosophy goes, sheer
originality of idea and argument is not something by which Cicero will
set much store. But where the Romans have room to do well is in the
literary style of philosophy.
Cicero regards his predecessors in Latin philosophical writing – who
were Epicureans – as bad writers for the educated reader. (Tusculans .,
.–) They do not express what they think using the tools of oratory.
fieri autem potest ut recte quis sentiat et id, quod sentit, polite eloqui non possit.
sed mandare quemquam litteris cogitationes suas, qui eas nec disponere nec
 Cicero and the Translation of Philosophy from Greece to Rome
illustrare possit nec delectatione aliqua adlicere lectorem, hominis est intemper-
anter abutentis et otio et litteris.
Now, it can happen that somebody who has a correct view cannot express it
with polish. But that anybody should commit to writing his thoughts,
when he can neither arrange nor illuminate them nor attract the reader
with any charm—that is the behavior of a man who abuses intemperately
both leisure and literature. (Tusculans ., translation from Douglas (),
modified)
Cicero’s criticism here has nothing to do with content. It would apply to
an author with whose view he agreed as well as to an Epicurean. The
educated reader, equipped with the critical tools of the rhetor, has to be
addressed by somebody like Cicero, who can use those same tools well. If
not, then “I do not see why [the Epicureans] need to be read except to each
other, by those who hold the same views.” (Tusculans .) When Cicero
said he would be the first to bring the “lamp” of Latin letters to bear on
philosophy (Tusculans .), it was not that he was overlooking the Epicur-
eans. It was that he did not think they had a lamp.
Exactly which tools of rhetoric does Cicero think will illuminate phil-
osophy? No doubt all could help in some way. But “illumination” carries
with it certain connotations that point to certain features of the Late
Sequence more than to others. One such connotation emerges from
Cicero’s On the parts of rhetoric. Speaking of the choice and arrangement
of words, he explains “illuminated speech” (oratio illustris) to his son:
est enim haec pars orationis, quae rem constituat paene ante oculos; is enim
maxime sensus attingitur, sed et ceteri tamen et maxime mens ipsa moveri
potest. sed quae dicta sunt de oratione dilucida, cadunt in hanc illustrem
omnia. est enim plus aliquanto illustre quam illud dilucidum. altero fit ut
intellegamus, altero ut videre videamur.

For more on this passage, see Smith (), Gildenhard () –.

. . . cur legendi nisi ipsi inter se, qui idem sentiunt, non intellego. Translation from Douglas ().

The Latin Epicureans Cicero has in mind are, if not others, Amafinius and Rabirius, Academica .,
Tusculans .– cf. On ends .–, Tusculans .. Cicero says that there were sane multi, “fairly
many” of their books around, Tusculans .. Here we come to a mystery: what about Lucretius?
About Lucretius’ “poems” (poemata), Cicero wrote in  BC to his brother Quintus, that they
shone with multis luminibus ingenii, multae tamen artis, literally, “many lamps of talent, but with
great technique.” (Letters to Quintus ..) So why does Cicero not acknowledge in the Late
Sequence that he has at least one predecessor who wrote illuminatingly about philosophy in Latin?
One guess is that Cicero read something by Lucretius not about philosophy, and that he did not
know our DRN. But the context in Cicero’s letter inclines me against this: to judge from the title of
Sallustius’ lost poem, DRN is an apt point of comparison with the Empedoclea. The other,
distasteful possibility is that Cicero deliberately ignored Lucretius in the Late Sequence, because
such an impressive Epicurean poet was inconvenient.
. A More Positive View of Cicero’s Dialogues 
For it is this department of oratory which almost sets the fact before the eyes
—for it is the sense of sight that is most appealed to, although it is
nevertheless possible for the rest of the senses and also most of all the mind
itself to be affected. But the things that were said about clear speech all
apply to the illuminated kind. For what is illuminated is worth considerably
more than that clarity above mentioned. By the one it comes about that we
understand what is said, but by the other that we seem to see it. (On the
parts of rhetoric , translation from Rackham ())
Illumination in this narrow sense – illuminating by choice and arrange-
ment of words – is not broad enough to cover all the effects that Cicero
aims for in the Late Sequence. But the wider idea of “illumination” it
exemplifies is apt: to illuminate is to make the subject seem to stand before
the reader’s eye. Cicero takes material from philosophers who write clearly,
so that they may be understood, but will illuminate it, so that the reader
seems to see. How the dialogues achieve this is obvious. They bring
philosophical speakers in front of our imagination. But, writing for his
rhetorically educated peers, Cicero will have to do a good job of this.
The speakers will have to be convincing characters who give coherent
speeches that fit their dramatic context. Since some of his readers will be
learned about philosophy, the back-and-forth will also need to make
dialectical sense to somebody with philosophical acumen. This, then, is
something more than an encyclopedia that we should expect from the Late
Sequence.
A second connotation for “illumination” comes from Cicero’s subse-
quent description of the Tusculans itself, at Div. .. The Tusculans at large
“opened up the subjects most necessary for happy living.” Cicero then
says that Book  of the Tusculans,
eum locum complexus est, qui totam philosophiam maxime illustrat; docet enim
ad beate vivendum virtutem se ipsa esse contentam.
contained that commonplace that most of all illuminates the whole of
philosophy, for it teaches that virtue is self-sufficient for happy living.
The “commonplace” is not an argument to which Marcus in Tusculans
Book  is committed, because he gives it in his capacity as a skeptical
interlocutor. Neither is Cicero the author of Div. committed to it. Thus
Cicero would not say that this topic illuminates by teaching his reader that
virtue is sufficient for happiness, since it is not Cicero’s intention that it

res ad beate vivendum maxime necessarias aperuerunt.
 Cicero and the Translation of Philosophy from Greece to Rome
should teach the reader this. The way this argument “illuminates the whole
of philosophy” must, for Cicero, be less direct. It shows how philosophy
can help us to deal with no less than the spectre of human fragility, of
which Cicero was sometimes terrified. (Tusculans .)
For the purpose of interpreting DND and Div., I can sum up what
I want to add to the encyclopedia view in two principles. The other
chapters of my book aim to show these principles’ worth by putting them
into action. They are:
The learned reader principle: to get the most from Cicero’s dialogues, we
should read them as though we were learned readers, who are already
aware of many of the philosophers, views, and arguments we read
about, or could find out about them for ourselves fairly easily. For us,
this means reading the dialogues with our other sources for Hellenis-
tic philosophy at our elbow.
The literary unity principle: we should read the dialogues as though
Cicero has an overall goal in writing each of them, which influences
his choice of schools for presentation and criticism, his characters,
their arguments, and so on, and we should read each speech as written
for a coherent character who chooses a set of arguments in response
to a given dramatic setting.
In the remaining sections of this introduction, I shall proceed as follows.
In section . I shall re-examine an old challenge to a position like mine,
the single-source hypothesis. In sections . and . I look at the evidence
for two approaches that Cicero took to his literary project: the dialogue
form at large, and his application of it to skeptical writing in particular. In
section . I look ahead to the two dialogues with which I shall be
concerned in the rest of the book.
. The Single Source Hypothesis

Source criticism, or Quellenforschung, is an attempt critically to determine
the sources of a given text and, perhaps, to reconstruct those sources if they
are lost. In the nineteenth and the earlier part of the twentieth century,
much of the scholarly attention devoted to Cicero’s dialogues was in

It is not my view that Cicero never made mistakes in pursuit of his goals. For some exceptions to the
literary unity principle in Div., see Chapter  section ., but cf. Chapter  section ..
. The Single Source Hypothesis 
pursuit of his sources. A particular version of this source criticism
depended on the single source hypothesis: that in most or all of his dialogues,
for long stretches of prose and perhaps in whole speeches or even books,
Cicero followed and translated a single source closely, making changes only
where needed in order to retain a Roman setting. On this hypothesis, it
might be possible to reconstruct much of each source that Cicero used.
The practice of single source criticism for the Late Sequence has failed to
achieve much consensus. In fact, the frustrations of single source critics
helped to summon the animus we saw directed at Cicero at the start of this
introduction. Frustration resulted in part from disappointment: if, as was
reasonable at the time, Cicero was read with the single source hypothesis in
mind, and yet attempts to identify and reconstruct single originals failed to
achieve accepted results, then it must be that Cicero could not or would
not transcribe his sources faithfully. But once we abandon the single source
hypothesis, we have another explanation for its failure: in writing speeches
for his dialogues, Cicero was his own man. He might “transcribe” a Greek
source for a long stretch if so minded, but he could also assemble words,
sentences, arguments, and topics as he wished, from many sources, from
memory, or by creation de novo.
Now it is not my intention to revisit the arguments against single source
criticism as it was and is practiced. Rather, I want to address the hypothesis
itself, for two reasons. First, at its simplest, it is at loggerheads with my
view of the dialogues, because it denies Cicero much agency in shaping
them. Second, the evidence for, and attractions of, the hypothesis were
once very strong, strong enough that its summary dismissal today repre-
sents a failure of imagination. But other data, and advances in scholarship,
are now sufficient to reject the hypothesis decisively. This decisive rejection
has not yet been set out in one place. So I shall aim to set out the evidence
for and against the hypothesis here.
So far as I can see, there were three main pieces of evidence in favor of
the hypothesis, as follows:

See Pease () vol.  pp. – n.  for source-critical bibliography on DND from the period
when single-source criticism was in vogue. Pease refers only to “the more important suggestions,”
but he gives sixty citations spread between  and . Not all of these are straightforwardly
source-critical. For example, he cites Boyancé (), who was hostile to source-criticism. For Div.,
see Pease (–) n.

Some decisive criticisms were Boyancé (), Douglas () –. An illustration of the
curiously low esteem in which it is held is that a late twentieth century editor of On duties, a text
which explicitly claims to stay close to its sources, and where our evidence for the identity of those
sources is good, felt the need to give a special explanation for investigating them: Dyck ()
–.
 Cicero and the Translation of Philosophy from Greece to Rome
. Letters to Atticus .. (= SB): “They are copies, they come out
with relatively little work. I add only the words, with which
I overflow.” This sentence is often quoted (as in the passages from
Usener and Mommsen with which I began). It was interpreted as
Cicero’s confession that the dialogues were made up of something
close to mere translations of a source.
. Haste. Cicero composed the Late Sequence in less than two years, and
all of it except the Hortensius in little more than a year. It seemed
unreasonable to expect anything more than transcripts at this rate of
production. Sure enough, there are marks of haste in the writing, for
example the traces in DND of a draft in which the drama was spread
over more than the single sitting narrated in our text.
. The Epicurean text we call Philodemus, On piety, written during the
time when Cicero was involved in philosophy in Italy, and found on
papyrus at Herculaneum. This was published in  with the
observation, never seriously challenged, that it seemed to be closely
related to part of Velleius’ Epicurean speech in Book  of DND. It
was thus plausible that excavators had in fact found one of Cicero’s
single sources. This must have seemed a spectacular confirmation of
what ) and ) suggested.
With the facts in )–) accepted widely, and interpreted as described, the
evidence in the nineteenth century for the single source hypothesis must
have seemed formidable. In the face of the arid style of the single source
critics, we should add that their hypothesis is very exciting: it offers to
conjure texts of Hellenistic philosophy from the grave. There was nothing
either silly or boring about taking the hypothesis very seriously.
The respective answers to the three pieces of evidence for the single-
source hypothesis are these:
. Letters to Atticus .. (= SB): There are two parts to this
answer. The first is that it is most obscure what Cicero meant by the often-
quoted sentence. For one thing, the context is disputable. The sentence
comes in a short paragraph, to which the letter turns abruptly from other
matters. The style of the paragraph is compressed. Part of the text is
garbled in our manuscripts. The paragraph runs:
De lingua Latina securi es animi. dices †qui alia quae scribis†. ἀπόγραφα
sunt, minore labore fiunt; verba tantum adfero, quibus abundo.
 
DND ., .. See Dyck () –. Drummond and Walpole ().
. The Single Source Hypothesis 
About the Latin language, your mind is at rest. You will say, ‘How about
the other things you’re writing?’ They are copies (apographa), they come
about with relatively little work. I just contribute the words, which I have in
plenty. (Trans. from Shackleton Bailey (), substantially adapted for
neutrality between interpretations)
The tone of the opening sentence of this paragraph traditionally is
thought to be something like, “set your mind at rest,” as if Atticus has
voiced some worry. If so, it is not clear what the worry was. Perhaps
Atticus was worried that Cicero’s use of language in his philosophical
writing was poor. That was apparently Usener and Mommsen’s reading
(pp. – above). Another interpretation would be that Atticus objected
to Cicero’s coining of a Latin philosophical vocabulary. But neither of
these suggestions seems plausible when we note how Cicero reasserts, and
then answers, Atticus’ worry. For if Atticus was worried about the quality
of Cicero’s writing, then Cicero would not have reassured Atticus by
playing down the amount of effort involved. Meanwhile, Shackleton
Bailey supposes that “the Latin language” refers not to the tongue, but
to Varro’s work with that title, or perhaps to some such work Cicero was
contemplating. (Shackleton Bailey  ad loc.) If so, then perhaps
Atticus’ concern is beyond recovery today. In short, if it is right that
Atticus needed to be pacified, we do not know what he was
worried about.
Cicero then seems to reassert the worry on Atticus’ behalf. What new
challenge Cicero posed is lost, since the words between the daggers in my
quotation (translated “How about the other things you’re writing?”)
represent nonsense in all our manuscripts. No satisfactory restoration is
yet available. Cicero appears to have imagined that Atticus would refer to
something else that he, Cicero, was writing. These other writings are
usually interpreted as the philosophical dialogues, but there is no direct
evidence for that interpretation. Cicero then replies that these writings are
apographa. This Greek word is mysterious. Its cognate verb, literally to
write from, means to copy, to enter in a list, or to bring a formal charge
against (LSJ sv.). apographa is either the plural of a neuter noun, or an
adjective agreeing with such. In either case it is extremely rare. What does
it mean?
The Greek rhetorician Dionysius of Halicarnassus, a younger contem-
porary of Cicero, uses the adjective to contrast the orator Lysias’ verbal
portrait of a client, which he calls an “archetype,” with Isaeus’, which he
calls an apographos. Dionysius intends the term pejoratively, saying that the
apographos portrayal is “obviously a fiction of the rhetorical art,” in contrast
 Cicero and the Translation of Philosophy from Greece to Rome
to the vividly portrayed archetype. (οὐ λανθάνοντα ὅτι πέπλασται ῥητορ-
ικῇ τέχνῃ, Isaeus ) So the word seems to mean a one-dimensional
character, a transparent copy of a fleshed-out original. Meanwhile, Pliny
the Elder applies the singular noun to a marketable copy of a portrait
painting of a girl, explaining the term as “a copy, which they call an
apographon.” (exemplar, quod apographon vocant, Natural history .)
So it is likely that with the word apographa Cicero means to call
his writings “copies.” But it is by no means clear that he means
“mere transcripts” in the sense of uncreative translations. Another possi-
bility, for example, is that Atticus had raised a concern about Cicero’s
choice or portrayal of the Roman characters in the dialogues. Letters to
Atticus . was written a week or so after Cicero had reported to
Atticus the completion of the first version of the Academica (Letters to
Atticus .. = SB) and about a month before Cicero decided
to rework the dialogue with Varro among its characters (Letters to
Atticus ..– = SB). Perhaps Atticus’ concern helped to
prompt Cicero’s rethink. If so, then in our passage Cicero replies that
the characters will not give offense, because they are obviously literary
fictions. If the metaphor were of visual portraits, then it would be easy to
make sense of his next remark, “I add only the words” – a curiously
redundant point if it were about translation – as meaning, “I just give
them their lines.” That reading would be as consistent with my view as
with the single source hypothesis, since Cicero could say this about
carefully written characters in carefully shaped dialogues. However that
may be, given these many uncertainties about what it means, proponents
of the single-source view were unwise to put much weight on Letters to
Atticus ...
The second part of the answer to ) is that even if we assumed that in
Letters to Atticus .. Cicero does admit that he made mere transcripts
with little effort, this would not prove much, since a great deal of evidence
countervails. Against this background, Cicero’s “admission” would look
like playful self-deprecation.
First, for other purposes I examine below ample evidence in the letters
themselves that Cicero effortfully crafted, and re-drafted, his dialogues
(pp. –). This seems inconsistent with mere transcription.
Second, I add here that a month after Letters to Atticus ., Cicero’s
opinions of his finished Academica project seem at odds with a single
source view: “I finished off the whole subject of Academic philosophy . . .
how well I cannot say, but so carefully that nothing further could be
done,” “unless my share of amour propre deceives me, it has turned out
. The Single Source Hypothesis 
better than anything in its genre now existing, even in Greek.” The first
quotation suggests that Cicero put more than mere word-for-word trans-
lation into the project in its various drafts. The second suggests that he
thought the result was in some way original by comparison with any
similar Greek text.
But the main font of evidence contrary to the single-source interpret-
ation of Letters to Atticus .. is in Cicero’s descriptions of his writing
in the dialogues themselves, which we have already seen in section . of
this introduction. In the preface to On ends Book , Cicero upbraids
learned Romans who turn their nose up at Latin:
quid? si nos non interpretum fungimur munere, sed tuemur ea quae dicta sunt
ab iis quos probamus eisque nostrum iudicium et nostrum scribendi ordinem
adiungimus, quid habent cur Graeca anteponant iis quae et splendide dicta sint
neque sint conversa de Graecis?
What of it, if I do not perform the task of translator, but preserve the views
of those whom I consider sound while contributing my own judgment and
order of composition? What reason does anyone have for preferring Greek
to that which is written with brilliance and is not a translation from
the Greek? (On ends ., translation from Annas and Woolf ())
The isolated and ambiguous evidence of Letters to Atticus .., then, is
not persuasive when set against the plentiful and unambiguous evidence
the other way.
. Haste. How fast did Cicero write the Late Sequence? We can make a
rough but helpful estimate. As a unit of length familiar to readers of Cicero
we can use the shorter sections into which his corpus (except the letters)
was divided by Alexander Scot. These are not exactly of uniform length,
since their divisions follow the syntax of the Latin. But they are roughly so.
We might say that their usual length is about that of a paragraph in today’s
academic English. Now the first work of the sequence, Hortensius, is lost
and the last, On fate, is mutilated. But Cicero wrote those two before and
after the other members of the sequence, respectively, so we can drop their
composition from our calculations. Let us estimate how long it took him

absolvi nescio quam bene, sed ita accurate ut nihil posset supra, Academicam omnem quaestionem libris
quattor, Letters to Atticus ..=SB. nisi forte me communis φιλαυτία decipit, ut in tali genere ne
apud Graecos quidem simile quicquam, Letters to Atticus ..=SB. Translations from
Shackleton Bailey (), modified.

Glucker () –. These are the shorter “sections” that appear in modern editions, usually
indicated with Arabic numerals.
 Cicero and the Translation of Philosophy from Greece to Rome
to write the five dialogues he completed in between: Academica, On ends,
Tusculans, DND, and Div. On ends seems to be mentioned in Cicero’s
letters to Atticus in mid-March  BC. Let us say that Cicero started the
Academica–Div. sequence then, although he could well have begun it
earlier. When Cicero finished Div. is hard to say, but he certainly did
some work on it after Caesar’s death on the Ides of March  BC. So let
us say that it was complete at the end of that March, although this date
might be early. This conservative estimate, from mid-March  BC to the
end of March  BC, gives Cicero about  days for our five works.
How much did Cicero write in this time? The length of the Academica
presents some difficulties, since Cicero completed two versions of it, of one
of which we have half, and of the other less than a quarter. He also worked
out some part of an intermediate draft. The writing of the two versions
overlaps with the writing of our other texts so that, unlike Hortensius and
On fate, we cannot drop the Academica from our calculations. We must
therefore guess how much of it Cicero wrote. The first version consisted of
two book-length dialogues, the Catulus and the Lucullus. We have the
whole Lucullus. To guess the length of the whole first version, let us say
that the Catulus was the same length as the Lucullus. The final version
divided the same material into four Academic books, with different speakers
and setting. Just how much new writing this required we do not know.
The new writing may have been extensive, especially as Cicero considered
yet a third version for a few days. Let us then guess that the Academic books
required as much writing as the first version again, so that we should
quadruple the length of Lucullus for our guess at the total number of
section-lengths Cicero finished in the course of the Academica project.
Our other difficulty in calculating how much Cicero wrote is DND Book
, from which a significant amount of text is lost. Since Cotta’s speech in
Book  is a point-by-point reply to Balbus’ speech in Book , let us guess
that Book  was originally the same length as Book .
If we add the existing lengths to the guessed lengths I have suggested, we
calculate that Cicero wrote about  sections of the sequence Academica

Letters to Atticus .=SB. de Epicuro suggests that Cicero is planning On ends Book .

For the date of composition of Div., see the judicious discussion in Wardle () –, and
Durand (), Falconer (), and Giomini ().

For the complex evidence for Cicero’s composition of Academica, see Griffin ().

DND .–, –.

This may well be an overestimate, since Balbus’ is the longest speech in the Late Sequence and Book
 of DND is its longest book.
. The Single Source Hypothesis 
to Div. in his  days. That is about six sections a day, on average. The
“on average” matters. On some days, for example, Cicero was traveling and
presumably wrote little. Moreover, in the same period he wrote the eighty-
five sections of On old age, the lost Consolation (Div. .–), and his
translation of part of Plato’s Timaeus, to push his average output of all
philosophical prose over six sections per day. He often had non-
philosophical writing on hand, too. Thus an average of six paragraphs a
day of philosophical prose finished for the Late Sequence is a most
impressive pace. But, especially if Cicero’s task was not, for the most part,
original argument, but rather the shaping and presentation of the argu-
ments of others, six sections is hardly so much that it is incredible that he
should have worked with a degree of care. We might compare On duties, a
later work that Cicero himself says was based on a single source, but not
slavishly so. On one recent estimate Cicero wrote On duties at a rate of
thirty-one sections per day. Next to this, six or seven sections a day of
something more complex than single source “transcription” looks
plausible.
Cicero’s pace of writing becomes easier to understand when we look at
his circumstances in those  days. Admittedly, they were very varied.
We have much the best evidence for the earlier part of the period, from
March through August  BC, at which point daily letters to Atticus break
off. At times Cicero was busy, likely more so in the period later in  and
in  BC when he was often at Rome or when, in December  BC, he
played host to Caesar (Letters to Atticus .=SB). But we know from
his letters that he also had plenty of leisure, at least during the time of daily
letters. His time then was distributed between his villas at Astura, a
relatively isolated seaside area, Tusculum, where his main library was,
and Arpinum, his ancestral estate. He was bereaved at the death of his
daughter, Tullia. This bereavement affected him severely for some
months. Early on in our period he seems to have spent most of his waking
hours in literary pursuits. On th March  he wrote to Atticus from
Astura:

Academica:  x  =  sections; Tusculans:  +  +  +  +  = ; On ends:  +  +
 +  +  = ; DND  +  x  = ; Div  +  = ; total =  sections. 
sections /  days = about . sections per day.

Cicero refers to On old age as a recent work at Div. .. Cf. Letters to Atticus .. = SB from
 May  BC.

Including On old age (and the Timaeus (fifty-two sections) yields / = about . sections/
day. The Consolation means that the real figure was higher. For the sake of argument, if the
Consolation was one-hundred sections long, Cicero’s average was about . sections/day.
 
Dyck () –. See Treggiari () –.
 Cicero and the Translation of Philosophy from Greece to Rome
in hac solitudine careo omnium colloquio, cumque mane me in silvam abstrusi
densam et asperam, non exeo inde ante vesperum. secundum te nihil est mihi
amicius solitudine. in ea mihi omnis sermo est cum litteris. eum tamen inter-
pellat flatus; cui repugno quoad possum, sed adhuc pares non sumus.
In this lonely place I do not talk to a soul. Early in the day I hide myself in a
thick, thorny wood, and don’t emerge till evening. Next to yourself solitude
is my best friend. When I am alone all my conversation is with literature,
but it is interrupted by fits of weeping, against which I struggle as best I can.
But so far it is an unequal fight. (Letters to Atticus .=SB, translation
from Shackleton Bailey ())
The next day, he reports that, “Reading and writing bring me, not solace
indeed, but distraction;” on  March, he rejects Atticus’ advice to return
to his “old ways” and says that he is satisfied with what he has achieved in
his consolation “through literature” (per litteras); on th May, he writes
through the whole of each day (dies totos) since even though it does not
lighten the load, it distracts him.
While the acute stage of Cicero’s grief does not seem to have lasted
through the summer of  BC and his preoccupations became more
varied, writing remains visible as the largest object of his effort. While at
Arpinum, he gave great attention to redrafting the Academica and putting
the final touches both to it and to On ends, along with preparing both as
gifts, to Varro and Brutus respectively. In July , by which time he seems
to have become less obsessive in his literary pursuits, Cicero nevertheless
spent some time “before dawn” working for a few hours (ante lucem, Letters
to Atticus .=SB). The letters suggest, then, that during our
 days Cicero had some, and sometimes many, hours in the day to give
to reading and writing – and that he used them. That being so, I suggest
that the Late Sequence dialogues were written at a most impressive speed,
but not with such haste that we must discount their quality ahead of
reading them.
. On piety. That this papyrus text from Herculaneum, which was
written during Cicero’s life is closely related to Velleius’ Epicurean speech
in DND Book , either as a source or as a descendant of the same source,
has been accepted very widely since its first publication. Thus it can be
surprising to put even an older edition of On piety like Gomperz’ next to

Translation from Shackleton Bailey (), slightly altered.

me scriptio et litterae non leniunt sed obturbant, .=SB; consuetudinem, ..=SB;
.=SB. Translations from Shackleton Bailey ().
. The Single Source Hypothesis 
Velleius’ speech and see that the correspondences, although persuasive in
broad strokes, are not as thoroughgoing as one might expect.
The most important parallel is that Velleius, in his catalog of philoso-
phers’ views about the gods, mostly mentions the same philosophers as
does On piety and mostly in the same order, as is emphasized by Diels’
printing of the two texts in parallel columns. Although the eminence of
most of the authors in the catalog and the broadly chronological order of
their listing could provide an alternative explanation for this coincidence,
on balance it convinces that there is some close relationship of sources
between the catalogs in the two texts. But in other respects even the
catalogs do not look alike. There is only one close parallel of wording,
otherwise Velleius seems to identify and phrase the philosophers’ views
differently than does the author of On piety. So far as one can tell from the
fragmentary papyrus text, Velleius gives more space to Epicurean counter-
argument against the other views listed than On piety does. We should
accept that Cicero wrote Velleius’ catalog by following On piety or a text
very like it, but what we learn is not that he translated the model word-for-
word, but rather that he adapted it to his purposes.
A second parallel between the two texts is that while Velleius’ speech
features first the catalog of philosophers, second a criticism of the poets,
and third a statement of Epicurean theology, the papyrus text features a
catalog of philosophers, a criticism of poets, arguments for the existence of
the gods, and a set of evidence of the Epicureans’ piety. So we might think
that aside from the differences in the last stated sections on Epicurean
theology, and the much greater length of the papyrus section on poets,
Cicero has given Velleius’ speech the same structure he found in On piety.
This must have seemed the more compelling to readers of a text like
Gomperz’, since in Gomperz’ edition the three parts of On piety appear
in the same order as do their apparent counterparts in Velleius’ speech.
But the picture is different today, when Dirk Obbink has published the
first part of his edition of On piety (). Obbink examined the confusing
evidence, created by the complex history of the treatment of the papyrus
and its copies, far more closely than did most other editors. Convin-
cingly, Obbink reverses the order of On piety. The exposition of Epicur-
ean piety comes first, followed by the sections on philosophers and poets.

Gomperz’ edition () has been influential in that Diels relied on it in Doxographi Graeci ().
 
Diels () –. The passage on Anaximenes, DND . = On piety p.  Gomperz.

For a sophisticated treatment of the parallels, see Obbink () –, with his references.
 
See Obbink (), Introduction, especially pp. –. Obbink () , –.
 Cicero and the Translation of Philosophy from Greece to Rome
When we match this with the radically smaller space Velleius devotes to his
remarks on poets and the largely distinct content of his positive, Epicurean
arguments, Velleius’ speech looks very different from the papyrus text. If,
again, we envisage Cicero writing the speech with On piety before him, we
now have to conclude that he adapted parts of it very substantially, left out
other parts, and introduced much that was new, whether with other texts
as his guides or from his own knowledge. Thus, far from supporting the
single source hypothesis, the comparison of On piety with Velleius’ speech
now supports a view of Cicero as the creative user of a source to compose a
speech of his own devising.
To sum up: where the evidence for the single source hypothesis used to
appear strong, and the hypothesis was well worth exploring, closer exam-
ination and related scholarly advances leave the evidence either uncertain,
outweighed by opposing data, or so changed that it points in the opposite
direction. We should not think of Cicero in his dialogues merely as a
transcriber of single sources for Hellenistic philosophy.
. Cicero and the Revival of the Classical Dialogue

Later in the Hellenistic period, continuous prose tended to replace dia-
logues as the form of choice for technical exposition. The authors that
Cicero names as the important models for his and Varro’s dialogues –
Plato, Aristotle, and Heraclides of Pontus – had written centuries before.
So when Romans like Varro and Cicero wrote dialogues they were
engaged in a literary revival of a form that had flourished hundreds of

Obbink himself thinks that, if Cicero had On piety in hand, he “rolled very quickly through” the
part on Epicurean piety, “reduced some  columns [of the criticism of poets] . . . to a single
paragraph” and put the criticism of philosophers up front, affording himself a survey of
philosophical views at the outset of the dialogue and an apt portrayal of the Epicureans as he
understood them, as “cultural renegades bent on undermining the whole of traditional learning”
(Obbink , pp. –).

For the exceptions, see Hirzel () –. An example of a (relatively early) Hellenistic
dialogue to which Cicero certainly gave attention was a work on the soul in three books by
Dicaearchus (Tusculans ., cf. Cicero’s reading list in Letters to Atticus ..=SB). My
point is not that nobody wrote technical dialogues in the Hellenistic period – they did – but rather
that the important technical books of the period usually were not dialogues, and that Cicero does
not look to Hellenistic dialogists as his models, or left no evidence that he did so.

There were some predecessors in writing Latin dialogues. A Marcus Junius Brutus had written them
in the previous century, On the orator .–, cf. Pro Cluentio . This cannot be “M. Brutus
(the father of the conspirator)” as Powell (b)  suggests. Rather, he is Iunius no.  in RE.
Gaius Scribonius Curio was another Latin dialogist contemporary with Cicero, Brutus ,
Suetonius, Lives of the Caesars ., Hirzel () –. Given Cicero’s casual references to
Brutus and Curio as dialogists, it seems likely to me that there were other technical dialogues in
Latin prior to Varro and Cicero. There were, obviously, other sorts of dialogues in drama and satire
. Cicero and the Revival of the Classical Dialogue 
years before and was not the vital genre of technical literature in their own
day, not even in Greek. When we read Cicero’s dialogues, we should
remember we are reading not someone striving toward a contemporary
Greek model, but Cicero’s attempt to do better at the illumination of
philosophy than Greek authors of his own day, by going back to classical
examples of the fourth century BC.
In this section, I will present some of the evidence for this view of
Cicero’s approach to dialogues. It shows in particular that Cicero and his
circle had models to whom they looked when they discussed the dialogue
form. Naturally enough, Cicero seems to have referred to these models
most often when he was beginning to write dialogues, so that much of the
evidence here is directly concerned with dialogues he wrote earlier than the
Late Sequence. But we see arise here many aspects of the Late Sequence
dialogues, which I shall turn to in the next section.
We shall look in a moment at Cicero’s use of his models Plato, Aristotle,
and Heraclides, but let us note first his acquaintace with classical dialogists
was broader than just these three. Whether directly or from a quotation, in
On invention .– the young Cicero translated a comic little scene “in
Aeschines the Socratic” between Aspasia, Xenophon, and Xenophon’s
wife. In a stiff footnote to a May  letter to Atticus, Cicero returns
his verdict on the writings of Antisthenes, “Cyrus  pleased me the same
way as the rest of Antisthenes’ stuff—the man is sharper than he is
cultured.” Most importantly, Cicero was more than familiar with Xeno-
phon’s dialogues, and especially with the Oeconomicus. In Cicero’s On old
(see Hirzel () –). In April  Cicero thanks Atticus for his mirificos cum Publio dialogos,
“wonderful dialogues with Publius” (Letters to Atticus .. = SB). When Cicero uses the
transliterated Greek dialogus he usually seems to mean a piece in the dialogue genre – had
Atticus been playfully reporting conversations in the form of a literary dialogue, much as Cicero
himself does in his miniature dialogue with Quintus filius (Letters to Atticus . = SB)?

Of course, this is not to say that there were no recent and interesting Greek dialogues in Cicero’s
day. It is possible that Antiochus’ Sosus (Academica .) was a dialogue. A proper name for a title is
suggestive. Some of the more elaborate arguments that Sosus was a dialogue are made in the hope
that it is the source for a part of Academica: Hirzel () – and cf. Hirzel () n,
Glucker () –.

Cf. Hirzel’s chapter on Varro and Cicero, entitled ‘Wiederbelebung des Dialogs’ (, –).
Zoll () proposes a “Platonrenaissance” in the second and first centuries at Rome, especially
marked with Cicero ().

This is presumably from Aeschines’ Aspasia, a work that DL rates as “stamped with the Socratic
character” (τὸ Σωκρατικὸν ἦθος), ..

I am not sure what the “” signifies here. DL lists four different works of Antisthenes named Cyrus
(DL ., ). In any case, the β´ is Shackleton Bailey’s correction of the vulgate reading. He takes
this “” to refer to the second listed Cyrus dialogue, on kingship.

Κῦρος β´ mihi sic placuit ut cetera Antisthenis, hominis acuti magis quam eruditi, Letters to Atticus
.a. = SB, and cf. Tusculans ..
 Cicero and the Translation of Philosophy from Greece to Rome
age Cato, explicitly his mouthpiece (On old age ) says that “Xenophon’s
books are very useful for many matters,” and follows this with a
summary and quotations from Oeconomicus .. Cicero did not cull
these details from some intermediate source, for he tells his son Marcus
in On duties that, in his youth, he translated the whole Oeconomicus into
Latin (On duties .), a boast corroborated by Quintillian (Institutes
..).
Cicero, then, was well grounded in classical Greek dialogues. But he and
his peers take three authors specifically as models: Plato, Aristotle, and
Heraclides of Pontus. Although he is by far the least well known of the
three, the literary achievement of Heraclides’ dialogues was much admired
in antiquity. DL introduces the list of Heraclides’ works with, “There are
in circulation writings of his of the greatest beauty and highest quality—
dialogues . . .” (.) The impressive quality of these dialogues evidently
inspired some technical literary criticism. DL says that Heraclides com-
posed dialogues in both of ‘comic’ and ‘tragic’ diction. (.)
Furthermore:
ἔστι δ’ αὐτῷ καὶ μεσότης τις ὁμιλητικὴ φιλοσόφων τε καὶ στρατηγικῶν καὶ
πολιτικῶν ἀνδρῶν πρὸς ἀλλήλους διαλεγομένων. . . . ἄλλως τ’ ἐν ἅπασι
ποικίλος τε καὶ διηρμένος τὴν λέξιν ἐστὶ καὶ ψυχαγωγεῖν ἱκανῶς δυνάμενος.
He has as well a certain middle style of conversation, representing philoso-
phers and generals and statesmen in discussion with each other. . . . And
otherwise in all his writings he is varied and lofty in his style and sufficiently
able to capture the mind. (DL ., translation from Schütrumpf, Stork,
van Ophuijsen, and Prince ())
As DL tells it, Heraclides held considerable interest specifically as an
author of dialogues that (to use Cicero’s metaphor) illuminated their
subject matter.
If we assume that Heraclides’ reputation in Cicero’s day was not unlike
what DL has to say of him, it is not surprising that Cicero took him as a
model dialogist. Indeed, Cicero took Heraclides as his first model. Con-
cerning his revisions of Academica, and Atticus’ suggestion that Varro and

multas ad res perutiles Xenophontis libri sunt.

On other Ciceronian references to Xenophon, see Zoll () n.

φέρεται δ’ αὐτοῦ συγγράμματα κάλλιστά τε καὶ ἄριστα· διάλογοι . . . Translation from
Schütrumpf, Stork, van Ophuijsen, and Prince (), modified.

On Heraclides’ dialogues, see Gottschalk () –, Fox (). For Heraclides in general see
Gottschalk (), Schütrumpf et al. (), Fortenbaugh and Pender ().
. Cicero and the Revival of the Classical Dialogue 
Gaius Aurelius Cotta (narrator of On the orator, who will speak for the
Academy in DND) be the speakers, he writes:
si Cottam et Varronem fecissem inter se disputantis, ut a te proximis litteris
admoneor, meum κωφὸν πρόσωπον esset. hoc in antiquis personis suaviter fit,
ut et Heraclides in multis et nos in sex de re publica libris fecimus. sunt etiam de
oratore nostri tres mihi vehementer probati.
If I had made Cotta and Varro discuss it between them, as you suggest in
your last letter, I should have been a ‘silent character.’ This is quite
agreeable if the characters belong to history. Heraclides did it in
many works, and I myself in my six books on the Republic. And there
are my three On the orator, of which I entertain a very good opinion. (Letters
to Atticus ..– = SB , translation from Shackleton Bailey (),
modified)
Here we see that Heraclides served, both for Cicero and for Atticus, as a
model. Cicero proposes to Atticus Heraclides’ historical dialogues as a
reason to think that such dialogues can be done well. Indeed, Atticus
evidently enjoyed the Heraclidean style because in his attempts to broker
a literary exchange between Cicero and Varro he tried to elicit from each
“something Heraclidean.”
On the other hand, Heraclides’ limitations as a model were felt early, at
least by Cicero’s audience. In  BC while his Republic was in progress
Cicero wrote to his brother Quintus that he had read out the first two
books to Sallustius and received a worrying response:
admonitus sum ab illo multo maiore auctoritate illis de rebus dici posse si ipse
loquerer de re publica, praesertim cum essem non Heraclides Ponticus sed
consularis et is qui in maximis versatus in re publica rebus essem; quae tam
antiquis hominibus attribuerem, ea visum iri ficta esse; oratorum sermonem in
illis nostris libris qui essent de ratione dicendi, belle a me removisse, ad eos
tamen rettulisse quos ipse vidissem; Aristotelem denique quae de re publica et
praestanti viro scribat ipsum loqui.
[Sallustius] pointed out that these matters could be treated with much more
authority if I spoke of the commonwealth in my own person. After all, he
said, I was no Heraclides of Pontus but a consular, one who had been
involved in most important state affairs. Speeches attributed to persons so
remote in time would appear fictitious. In my earlier work on the theory of

From Varro to Cicero, Letters to Atticus .. = SB, . = SB; from Cicero .. =
SB, .. = SB, .. = SB.

For an examination of Cicero’s reasons to use great men of the past as characters, see Steel ()
–.
 Cicero and the Translation of Philosophy from Greece to Rome
oratory, he said, I had tactfully separated the conversation of the orators
from myself, but I had put it into the mouths of men whom I had
personally seen. Finally, that Aristotle’s writings on the state and the
preeminent individual are in his own person. (Letters to Quintus .. =
SB, translation from Shackleton Bailey ())
Cicero goes on to say that he was shaken by this advice, and that he
entertained the possibility of abandoning the Heraclidean form of the
books. In the end, he seems to have added prefaces in his own voice rather
than give up the historical setting. Note that all these observations on the
advantages and disadvantages of the two dialogue forms – Heraclidean and
Aristotelian – came from Sallustius. Even in  BC Cicero was not alone
among his friends in the ability to give some thought to different models
for the dialogue. Of course, Cicero himself was not unreflective about
Heraclides as a model, and he quickly departed from it when it suited him.
But as we have just seen in Letters to Atticus . (p.  above), when he
departed, it was in favor of another model, Aristotle.
Let us turn, then, to Aristotle, and look at two letters in which we see
Cicero and Atticus acknowledge him as a model. As we have seen, in
earlier dialogues like On the orator or his Republic, Cicero used historical
characters in the Heraclidean manner. This meant that in those earlier
works, Cicero could not pay Varro the compliment of writing him a part
in a dialogue, as Atticus consistently urged (cf. Letters to Atticus ..,
p.  above). In  BC, while writing the Republic, Cicero defended this
decision, reassuring Atticus that, “I am making a suitable occasion to
address him [i.e. Varro] in one of the prefaces which I am writing to each
book, as Aristotle did in what he calls his ‘exoteric’ pieces [i.e. Aristotle’s
dialogues].” (Letters to Atticus .. = SB) Meanwhile, in Letter
.., Cicero comments on his switch, in the Late Sequence, away from
the Heraclidean model: “. . . But my recent compositions follow the
Aristotelian pattern, in which the other roles in the dialogue are subordin-
ate to the author’s own.” Cicero’s appearance in his own dialogues, of
course, required a contemporary setting and characters. Between these two
letters, we see Cicero justify two of the most strikingly un-Platonic (and so,
to us, apparently original) aspects of his dialogues as borrowings from
Aristotle. These are, first, the appearance of his own character and, second,

itaque cogitabam, quoniam in singulis libris utor prohoemiis ut Aristoteles in iis quos ἐξωτερικοὺς vocat,
aliquid efficere ut non sine causa istum appellarem. Translation from Shackleton Bailey ().

. . . quae autem his temporibus scripsi Ἀριστοτέλειον morem habent, in quo ita sermo inducitur
ceterorum ut penes ipsum sit principatus.
. Cicero and the Revival of the Classical Dialogue 
the prefaces to each work, and sometimes to individual books, that he
wrote in his own voice. It is likely that if we could read Aristotle’s
dialogues, their literary influence on Cicero would prove wider still.
Cicero’s third model, then, was Plato. The breadth of Cicero’s know-
ledge of and enthusiasm for Plato’s texts is well documented. We should
be careful with this evidence. We cannot always be sure that Cicero makes
a reference to Plato from his own knowledge rather than from what he
found in another text. But the sheer number of citations (and translations)
in Cicero’s corpus, and his tendency to effusive praise for deus ille noster
Plato, “that god of ours, Plato” (Letters to Atticus ..=SB), makes an
eloquent case that Cicero was very well read in the master’s corpus and
thought highly of his philosophical and literary monuments. It is not my
purpose here to make that case over again. Rather, I want to look at some
indications that Cicero reflected on Plato’s dialogues when he was deciding
how best to write his own, so that Lactantius echoed Cicero’s verdict on
Plato in his own verdict on Cicero, summus ille noster Platonis imitator,
“our greatest imitator of Plato.”
Of Plato’s works, Cicero translated at least the Protagoras and the part
of the Timaeus of which his version survives. When Cicero came to write
his first dialogue, On the orator (completed in  BC), although Heraclides
was the model for Cicero’s use of characters from history, Plato is also an
important influence. The narrator in Cotta’s scene-setting for the

In On the orator ., Crassus says the perfect orator would be one who “could speak on either side
of the issue in all matters, in the Aristotelian manner,” Aristotelio more de omnibus rebus in utramque
partem possit dicere (cf. On ends ., Orator ). This sets up Aristotle and the Peripatos in
competition with the Academy as proponents of the rhetorical practice of speaking on either side of
an issue (cf. Tusculans .). (See Long () –.) Zoll () – thinks that Aristotle’s
dialogues, rather than just his rhetorical theory and Peripatetic teaching methods, involved speaking
in utramque partem rather as we find in Cicero’s dialogues.

See especially Degraff (), Long () –. The evidence also emerges in studies of Cicero as
translator like Poncelet () and Powell (c). For Cicero as a creative adapter of Plato see
Douglas (), Boyancé () –. For two recent interpretations of Cicero’s philosophical
responses to Plato, rather than his use of Plato as a model for the dialogue form (although the two
subjects are of course not entirely distinct), see Annas () and Gildenhard ().

Lactantius, Divine institutes .., cf. ... For more on antiquity’s judgment of Cicero as Plato
Latinus see Quintillian, Institutes .. (and also perhaps ..), [Longinus], De sublimitate ,
Zoll () –.

Four fragments survive, Priscian .., .., .., Donatus On Phormio ...

It is possible that Cicero made both translations after On ends (cf. On ends .), which would make
the last two years of his life still more crowded than we already know them to have been. But if the
Protagoras translation was a youthful work like the Oeconomicus then perhaps he could overlook it in
On ends. On the other hand, his Timaeus is probably later than On ends, since it is certainly later
than Academica (Cicero, Timaeus ) and likely contemporary with the Tusculans (see the seeming
reference to a separate discussion of Pythagoreanism at Tusculans ., and the discussion of dating
in Sedley ()).
 Cicero and the Translation of Philosophy from Greece to Rome
conversation of Book  makes a fuss of recalling Plato’s Phaedrus. Cotta
tells Marcus that Scaevola explicitly set out to recall the Phaedrus, saying
(On the orator .), “Crassus, why don’t we imitate the Socrates of Plato’s
Phaedrus?” He has spotted a shady plane tree, just as Socrates and Phaedrus
do, and he and Crassus decide to sit and talk under it. What may be a
pun on his own name by Plato in the Phaedrus (the plane tree is a
πλάτανον, platanon, a) becomes clearly playful in Cicero’s use of
platanus in close proximity to Platonis. Cicero used Plato in thinking about
another detail of the dialogue, too. Atticus had complained that Cicero
removes Scaevola from the dialogue too early. Cicero replies
non eam temere dimovi sed feci idem quod in πολιτείᾳ deus ille noster Plato.
cum in Piraeum Socrates venisset ad Cephalum, locupletem et festivum senem,
quoad primus ille sermo habetur, adest in disputando senex, deinde cum ipse
quoque commodissime locutus esset, ad rem divinam dicit se velle discedere
neque postea revertitur. credo Platonem vix putasse satis consonum fore si
hominem id aetatis in tam longo sermone diutius retinuisset.
I did not drop him [i.e. Scaevola] casually, but followed the example of our
divine Plato in his Republic. Socrates calls on Cephalus, a rich, genial old
gentleman, in the Piraeus. During the opening talk the old fellow is present
at the discussion, but then, after speaking himself and very nicely too, he
says he has to go and attend to a sacrifice, and does not reappear. I imagine
that Plato thought it would not be convenable to keep a man of Cephalus’
age too long in so protracted a conversation. (Letters to Atticus ..=SB,
translation from Shackleton Bailey ())
Notice first that, as with Heraclides and Aristotle, Cicero presents Plato
as an authoritative model. Atticus is meant to back down from his com-
plaints once he agrees that Plato did what Cicero has done. But the appeal
to Plato is much more involved than the invocation of precedent, and it is
not just an appeal to general features of Plato’s style. Instead, Cicero
reconstructs Plato’s specific creative process in presenting and then remov-
ing the elderly Cephalus, and compares his own policy on Scaevola in the
sentences after my quotation. The parallel between Scaevola and Cephalus
is a subtle one (unlike that between Cato and Cephalus in On old age)
and without the comment in Cicero’s letter we might miss it as easily as

DL . records, but does not endorse, the story that the Phaedrus was the first dialogue Plato
wrote. The Phaedrus, with its material on rhetoric, is in any case a natural point of reference for On
the orator, but if Cicero was familiar with this tale, perhaps that provided another link.

Where Socrates likes the idea of sitting on the grass (Phaedrus c), Cicero makes the two Romans
call for cushions (On the orator .). Perhaps this is a humorous contrast, or perhaps it is an
allowance for their relatively advanced age.
. Cicero and the Skeptical Dialogue 
Atticus did (cf. Schütrumpf () ). This suggests that Cicero’s use
of Plato as a model was not limited to the moments where he makes
an obvious allusion. Rather, he could use Plato’s monuments as a
sounding-board when devising elements of his own work that strike us
as original.
. Cicero and the Skeptical Dialogue

So far I have argued that Cicero and his peers set out to revive the dialogue
form they found in Classical Greek authors. That revival is clearest in
evidence I have presented to do with dialogues Cicero wrote before the
Late Sequence. With the Late Sequence, Cicero was caused to think about
the dialogue form afresh. For he wrote the Late Sequence so as to call to
the reader’s attention his Academic skepticism. The Hortensius was meant to
turn the reader to philosophy, but the next in the series, the Academica,
indicated which school was Cicero’s own. After that, the rest of the Late
Sequence dialogues are given prefaces which point out their skeptical
authorship. This led Cicero to bend the form of his dialogues to the new
purpose. But before I detail that new form, I should say what his
skepticism was.
Just as Cicero and his peers had brought new life to an old literary form,
in the Late Sequence he writes from the point of view of an outmoded
philosophical school, the New Academy. As suited his temperament,
Cicero thought that he remained a friend to the patrimony of Socrates
and the Academy, skepticism, when “stupidity” (tarditas) had left it
almost deserted by others, and even by the Greeks. (DND .) Skepti-
cism admits of many variations, so I should say first out of what sort
of skepticism I think Cicero writes in the Late Sequence. This is a
difficult and complex question, so I can give here only my controversial
opinion.
When the Academy turned to skepticism, they had an opponent in the
confidence of early Stoics that our minds can have a firm grasp on the

For many more and stronger claims about Platonic parallels in On the orator see Schütrumpf ()
and Zoll () –. Annas () explores deeper parallels between Cicero’s and Plato’s
respective Laws.

Here I set aside the possibility that Cicero went through a non-skeptical phase. See Görler ()
vs. Glucker (). In any case, at the start of DND Marcus seems to accept that he is a pupil of
Philo (.), from which it would follow that he is at any rate not an Antiochean in the
conversation of DND. Other treatments of Cicero’s overall standpoint on philosophy and
skepticism are Taran (), Steinmetz (), Glucker ().
 Cicero and the Translation of Philosophy from Greece to Rome
world. In Cicero’s telling, the dispute with the Stoics that arose over this
issue defined the Academy’s history into the earlier part of Cicero’s own
lifetime. The Stoics held that much of the mental life of rational animals
like ourselves is made up of rational impressions in the mind, whether
impressions from the senses or impressions we imagine in thought. That
these impressions are rational means that they have content, they say or
mean something. When an impression says or means some assertion, that
assertion is either true or false, and in virtue of the assertions they make,
our impressions can thus be true or false. For example, I at present have a
visual impression that has the rational content, “there is a blue book on my
desk.” If it is true that there is a blue book on my desk, my impression is
thereby true. Similarly, when I say to myself mentally, “two plus two
equals four,” I have a true impression that two plus two equals four.
Further, I am able to assent to these impressions (or their content), that
is to say, to take them to be true, or to withhold my assent. For me to
assent to an impression is thus for me to form an ordinary, but confident
or “dogmatic,” sort of belief. The goal, of course, is to assent to the true
impressions and withhold assent from the false ones. Now the Stoics
thought that of our true impressions, there are some which are such that
they cannot be false. In fact, my impression that there is a blue book on the
table – formed in good light, my eyes working well – might well be an
impression of this kind. I cannot have a false impression of this sort, with
just this causal history and thus this clarity and distinctness. These are
called “cataleptic” (i.e. grasping) or “cognitive” impressions, because if and
only if I assent to one of them, then I have succeeded in grasping the world.
Not only have I taken a true impression to be true, I have taken to be true
an impression which cannot be playing me false, which describes the world
with guaranteed reliability. The Stoic wise person, the sage, is able always
to distinguish these cataleptic impressions from false impressions and from
those that are true but not cataleptic. The sage will withhold assent from
any impression that might be false (whether false impressions or those true
but not cataleptic), and assent only to cataleptic impressions. Thus the
Stoic sage is infallible: she forms only true beliefs.
Against the Stoics, the Academics argued that there could be no impres-
sion of such a sort that no impression just like it could be false. To take a

This account of the Stoic position is based on Cicero’s description, Academica .–, .,
.–, ., ..
. Cicero and the Skeptical Dialogue 
visual example, suppose you cannot tell apart two twins, Castor and
Pollux. You see Castor nearby, in good light. You form the visual impres-
sion that “that is Castor.” This impression, it would appear, is true, and
clear, and distinct. Thus it seems at first to be a cataleptic impression for
the Stoics. But if you had seen Pollux in just the same way, you would have
an indistinguishable impression that “that is Castor” – and this impression
would have been false. Thus your actual impression, of Castor, cannot be
cataleptic, because there could be an impression just like it that was false. If
you assent to it, then, for all you can tell, you might be assenting to a false
impression, so you have hardly gained a secure grasp on the world as it is.
But the Academics offered arguments like this across the whole range of
impressions you might have, from sensory impressions to philosophical
thoughts. Indeed, they not only disputed Stoic epistemology, they argued
against every philosophical school’s thesis on any question, so as to show
that there was not evidence to guarantee to us the truth of any proposition.
We shall see the Academics Cotta and Marcus engaged in this practice in
DND and Div., respectively, arguing against Stoic and Epicurean the-
ology. With these arguments in hand, the Academics claimed to have
shown the Stoics that there are no cataleptic impressions.
The Academics pointed out that conclusion that there are no cataleptic
impressions would have a dire consequence for the Stoics. For the Stoics
think that one should assent only to cataleptic impressions. If there are
none, then one should never assent, never take anything to be true. But the
Stoic could ask, do you Academics draw this conclusion? Do you really
think you should never take anything to be true? Here the Academics seem
to have parted ways in the generation before Cicero. Some, Cicero’s
teacher Philo of Larissa among them, seem to have answered that the
Academy merely used the Stoic premise that one should assent only to
cataleptic impressions in argument against the Stoics. Philo himself
thought one might take impressions to be true, provided that he acknow-
ledged that any impression might be false. Let us call Philo’s view Miti-
gated Academic skepticism. But Clitomachus of Carthage took the Stoic
premise to heart, and concluded that the wise person will never take
anything to be true. When challenged as to how one could live that way,
he answered that one can live by following what seemes plausible (proba-
bile in Cicero’s Latin) or like the truth (veri simile), as Philo might, but

This account of the Academic position is based on Cicero’s description, Academica .–,
.–, ..
 Cicero and the Translation of Philosophy from Greece to Rome
without taking any impression to be true. Let us call Clitomachus’ view
Radical Academic skepticism.
Does Cicero write the Late Sequence as a Radical or a Mitigated skeptic?
It seems to me that he writes consistently in these dialogues as a Radical
skeptic. That is to say, he presents himself and “Marcus” as aiming to take
nothing to be true, but rather to follow what seems plausible or like the
truth. (In Academica . Marcus admits that he does not always live up to
this principle, but that very admission shows that it is the principle against
which he measures himself.)
There is not space here to establish Cicero’s Radical skepticism in the
Late Sequence at large. But I shall offer some evidence for DND and Div..
My reading of this evidence will seem stronger in conjunction with
another of Marcus’ remarks about himself in the Academica: that he yearns
for the truth, “But just as I judge that this is most beautiful, to see the
truth, just so I judge it most vicious to approve falsehoods in place of
truths.” (.) To Marcus’ mind, a Mitigated skeptic risks doing what is
most vicious, in the pursuit of what is most beautiful. A Radical skeptic
turns away from this risk, keeping chaste his ardor for the unreachable
truth. This, I think, is the reason why Marcus in the Academica explicitly
prefers Clitomachus’ Radical interpretation of Academic skepticism.
(.) This principle – that one should not approve what is false in place
of the truth – also seems to me to point to, although not to prove, Cicero’s
Radical skepticism when read in conjunction with the two examples of
skepticism that I shall now give from DND and Div..
In my first example, Cicero gives a summary of Academic skepticism in
the preface to DND. He says:
non enim sumus ii quibus nihil verum esse videatur, sed ii qui omnibus veris
falsa quaedam adiuncta esse dicamus tanta similitudine ut iis nulla insit certa
iudicandi et adsentiendi nota. Ex quo exsistit illud, multa esse probabilia, quae

Cicero characterizes Philo and Clitomachus’ positions as interpretations of Carneades: did he give his
epistemological “views” just for the sake of argument against the Stoics, or was he committed to a view
like Clitomachus’? On this question see Allen (), Schofield (). For Clitomachus’ view see
Academica .–, ., .–, .–, for Philo’s Mitigated view see Academica ., .,
.. Of these disputes I accept the reconstruction of Brittain () Introduction, Chapter  and
Chapter ), summarized also at Brittain () x–xi. For clarity I omit the “Roman Books” view that
Brittain finds in Philo’s late career. While controversial (as any interpretation will be, given the difficult
subject and small evidence) it has precedent, especially in its distinction between radical and mitigated
skepticism, see e.g. Striker (c)  and n, Frede () –. An expansive response to
Brittain is Glucker (). For two other recent views see Thorsrud () –, n, Thorsrud
() –, n, and Lévy () –.

Görler () reads this passage differently, supposing that Cicero is proud to be a great opiner.

sed ut hoc pulcherrimum esse iudico, vera videre, sic pro veris probare falsa turpissimum est.
. Cicero and the Skeptical Dialogue 
quamquam non perciperentur, tamen quia visum quendam haberent insignem
et inlustrem iis sapientis vita regeretur.
For we are not those to whom it seems that nothing is true. Rather, we are
those who say that to all truths falsehoods are linked by a likeness so great
that there is no certain mark in them by which to judge and assent. Whence
emerged the following, that many things are plausible and that, although
they were not cognitively grasped, the life of the wise man is nevertheless
ruled by them, because they have a certain vivid and illuminated appear-
ance. (DND .)
This description suggests Cicero’s specifically Radical skepticism for two
reasons. First, it says that truths have no mark by which one can distinguish
them in order to assent. This suggests Cicero thinks one ought not to
assent to them, because we cannot distinguish them from falsehoods.
Second, Cicero says that the wise man is ruled by plausible things (prob-
abilia) because of a vividness in how they appear. Together these two
points suggest strongly that the wise man will be ruled by what seems
vividly plausible rather than by assent to impressions as true. Now, a
Mitigated reading of this passage is admittedly possible. First, when Cicero
says there is no certain mark by which to assent, he leaves it open that
perhaps the wise man nevertheless assents, as a Mitigated skeptic might do.
Second, while he says that the wise man is ruled by what seems plausible,
he does not say that he is ruled by it exclusively, nor does he say that the
wise man is ruled by what seems plausible without assenting to it. Neverthe-
less, on the balance of probabilities I take DND . to give a Radical
account of the Academic skepticism to which Cicero is committed. The
last sentence of DND gives some confirmation to this, when Marcus favors
a view because “it seemed to tilt closer to a likeness of the truth” (ad
veritatis similitudinem videretur esse propensior), in contrast to his Epicurean
interlocutor who prefers a view because it “seemed more true” (verior . . .
videretur) (DND .). Marcus seems to govern his views by what is like
the truth rather than by what is true. Especially when taken in conjunction
with the intuition that Marcus reports in the Academica, that one should
not put what is false in place of the truth, it seems likely that Cicero and
Marcus will not take to be true what seems like the truth, but rather will
follow it for its vividness.
To take up my second example, at the end of Div. Cicero gives Marcus
this summary of Academic method:
cum autem proprium sit Academiae iudicium suum nullum interponere, ea
probare quae simillima veri videantur, conferre causas, et quod in quamque
 Cicero and the Translation of Philosophy from Greece to Rome
sententiam dici possit expromere, nulla adhibita sua auctoritate iudicium
audientium relinquere integrum ac liberum, tenebimus hanc consuetudinem,
a Socrate traditam, eaque inter nos, si tibi, Quintus frater, placebit, quam
saepisse utemur.
But since it is the peculiar procedure of the Academy to introduce no
judgment of its own, to approve what seems most like the truth, to compare
the cases <on every side> and to bring out whatever can be said for each
view, to leave the judgment of listeners free and unimpaired by making no
use of its own authority, if you please, my brother Quintus, we will hold
onto this habit that was passed down from Socrates, and use it as often as
we can between ourselves. (Div. .)
This characterization of the Academy once again looks Radical. Cicero says
that the Academy makes its decisions on the basis of what seems most like
the truth. Again, Marcus does not specify that the Academy approves what
is most like the truth and ought not take it to be true. But that is, again, the
most natural way to read the passage. The evidence, then, points to Cicero
and the Marcuses of DND and Div. writing and speaking as Radical
skeptics.
In the rest of the book, I shall express myself as though Cicero writes
DND and Div. from this Radical perspective (although, outside the
context of the historical debate in the late Academy, I shall drop the
capital R and simply call him a “radical” skeptic): that one ought not to
assent, ought to aim to take nothing to be true. I do this for the reasons
I have given, but also for reasons outside the scope of this book. Much of
the time, what I say may easily be adapted to an understanding of Cicero as
a Mitigated skeptic, who takes his views to be true, but in such a way that
he thinks they might be false.
Let us now see how Cicero shapes his dialogues in accordance with his
skepticism. One attraction of the Academy for Cicero, as we see in Div.
., was that it dealt with the problem of authority. This problem
stemmed from the organization of Hellenistic philosophy into schools.
Would-be students of philosophy, as Cicero depicts it, were expected to
shop around until they found a teacher whom they liked. Suppose this
teacher was a Stoic, for example. Now the students were expected to
become faithful followers of Stoicism. If there was some Stoic doctrine
they did not understand, their responsibility was not to question or to

For a different view, that in these dialogues Cicero himself was a mitigated skeptic of a sort, and
indeed that he was the author of mitigated Academic skepticism, see Thorsrud () Chapter .
For other views of Cicero the Mitigated skeptic, see the references in n.  above.

Academica ., On ends .–.
. Cicero and the Skeptical Dialogue 
correct it, but rather to find out how to accept it. In this respect,
Hellenistic philosophical education tended to be alien to our practices
today. Cicero was disturbed, too. How could a choice made quickly and
by an ill-equipped beginner be the right way to choose the subsequent
raft of philosophical, and potentially life-altering, views? In DND,
Cotta the Academic makes fun of Velleius the Epicurean for parroting
views that “Epicurus meandered on about while he was nodding off.”
(quae Epicurus oscitans halucinatus est, .) Cicero tells us in DND’s
preface that Academics, by contrast, reject such methods, like the infamous
Pythagorean proof by Pythagoras’ authority: “he said it himself,” ipse dixit
(.).
Many aristocratic Roman students of philosophy were especially vulner-
able to these problems of authority. Preparing for a political career, they
would not have a lifetime to devote to study, and could not hope to
become as accomplished in philosophy as their teachers were. If they chose
the wrong lecturer early in their studies, they might never be able to correct
the errors that flowed from that choice. For these reasons, as well as for the
theoretical epistemological reasons of the Academica, Cicero recommends
the skeptical Academy to Romans: it should be by “rational argument”
(ratio) that a view is made to seem like the truth, not by “authority”
(auctoritas). (DND .)
How, then, does Cicero turn the dialogue form to skeptical use?
Consider the position of an Academic teacher as described in Div.
.. On the one hand, he wants to expound the various arguments on
all sides of a question, so that each listener is left with a free judgment
about it. On the other, he is aware one or another answer may seem
plausible to each student at a given time, and happy that the student will
live by this view. Teaching the arguments for any position will therefore be
very risky, since it might give the appearance that the teacher puts his
authority behind the arguments’ conclusions. If a student gets this impres-
sion, he might find the conclusion plausible because he suspects it seems
plausible to the teacher. But that is exactly what the teacher tries to avoid.
If the teacher is careful to remind students about the properly skeptical way
to use his teaching, it is still easy to imagine students discussing what
positions seem to excite the teacher more, why he chose some conclusion
and not another to present, and so on. Even Carneades’ own students seem
to have engaged in such guessing games, although their disagreement among
themselves suggests that he was skillfully inscrutable (see p.  n. ).

Academica ., Tusculans ..
 Cicero and the Translation of Philosophy from Greece to Rome
Cicero describes a complex history of responses by Academic teachers to this
problem, but he tends to pick out two forms of Academic teaching in
particular. We need not suppose that he reports two rigid procedures that
historical Academic teachers always followed. It seems more likely that these
are two theoretical approaches to what happened in the classroom, which in
practice was no doubt more various. But it will be helpful to see the clear
distinction that Cicero makes, and to compare this with a distinction among
his Late Sequence dialogues.
First, Cicero tells us that the Academics would sometimes talk in
utramque partem, “for either side” of an issue. Cicero’s presentation of
the exercise may well owe something to Carneades’ infamous visit to
Rome, on which Carneades gave first a speech arguing that justice is
natural, and then a speech arguing that it is not natural. The second
form of Academic exposition that Cicero describes did not involve two
speeches. Instead, a member of the audience would set the topic by giving
a “proposition” (propositum), of the form “It seems to me that . . .” The
Academic teacher would then argue against the propositum. In On ends,
Marcus complains that in the old days the member of the audience would
honestly report some view he held, presumably hoping for a salutary dose
of counter-argument, but that the practice had changed so that people
would give a propositum they wished to hear refuted. (On ends .)
Both these forms of exposition have the features that Academics wanted.
Both allow the listeners to form their own judgments. But also, as a
remedy for the problem of authority, in both the Academic teacher always
argues for a conclusion that is dictated to him by the form of exposition,
arguing either in succession for and then against some proposition, or
against a propositum raised by someone else. The listeners can always tell
 
On the orator ., Tusculans .. Lactantius, Divine institutes, ..–, cf. On ends ..

On ends .–, Tusculans ., On the orator .. Cicero gives the propositum method a complex
history. He associates it with Socrates (On ends .–, Tusculans .), with Arcesilaus (On the
orator ., On ends .), with Carneades (Tusculans . and perhaps Tusculans .), and,
mysteriously, with some version of the Academy contemporary with the drama of On ends ..
Rather than simply give his own continuous speech, the philosopher replying to the propositum
might, like Arcesilaus, have invited the poser of the propositum to defend it in conversation. This
would seem to make the distinction between the propositum method and speaking in utramque
partem a matter of degree rather than kind. For the further complexity of the association with
Socrates and his methods, see Schofield () .

Of course, in the argument-on-either-side form, it seems the teacher could choose the proposition
to argue first for, then against. Thus pupils might try to divine the teacher’s leanings from his
choice of topic. But if the teacher kept his arguments well balanced, and tended to choose obvious,
often-discussed, topics in the way that Cicero does, it seems likely that he could keep his own
reactions private.
. Cicero and the Skeptical Dialogue 
that the Academic teacher does not put his authority behind anything he
says, because the teacher is not free to choose to argue for any particular
conclusion.
Cicero took these two forms of Academic spoken exposition and turned
them into two distinct forms of dialogue. This is explicit in his preface to
On fate. Cicero writes,
quod autem in aliis libris feci, qui sunt de natura deorum, itemque in iis, quos
de divinatione edidi, ut in utramque partem perpetua explicaretur oratio, quo
facilius id a quoque probaretur quod cuique maxime probabile videretur, id in
hac disputatione de fato casus quidam ne facerem impedivit.
But a chance occurrence prevented me from doing, in this discussion on
fate, what I have done in my other books which are concerned with the
nature of the gods and likewise in those which I produced on divination,
that a continuous speech should be set out for either side, so that each
person might more easily approve what seemed most plausible to each. (On
fate , translation from Sharples (), modified)
Cicero explains that, following Caesar’s death, the consul-designate Hirtius

visited. Having some leisure, Hirtius asks to hear Marcus declaim. Marcus
offers him the choice of a rhetorical display or philosophical argument.
Hirtius chooses the latter:
. . . quoniam. . . hanc Academicorum contra propositum disputandi consuetu-
dinem indicant te suscepisse Tusculanae disputationes, ponere aliquid ad quod
audiam, si tibi non est molestum, volo.
. . . since your discussions at Tusculum show that you have adopted this
Academic habit of arguing against something proposed, I would like, if you
don’t mind, to propose something on which I might hear you. (On fate ,
translation from Sharples ())
Marcus agrees and they sit to talk. The text that follows is lost, but since
when it resumes Marcus is giving a long speech on fate, presumably
Hirtius gave some propositum on the subject. Hirtius’ reference in the
Tusculans is clear: the conversation of each book begins with an interlocu-
tor stating what seems to be the case to him. “If anyone wishes, let him say
what he wants to hear argument about,” says Marcus (Tusculans .), and
the proposita then give the target of all Marcus’ arguments: “it seems to me

Since Marcus seems mostly to argue against Chrysippus’ position on fate, I suspect Hirtius’
propositum was mihi videtur omnia fato fieri, “it seems to me that everything happens by fate.”
In Div. Quintus promised a discussion of that thesis “in another place” – a curious piece of
prescience from a character (.).
 Cicero and the Translation of Philosophy from Greece to Rome
that death is an evil”, “that pain is the greatest of all evils,” “that the wise
man suffers distress,” “that the wise man can lack any trouble in his soul,”
“that virtue cannot be sufficient for living happily.” Each of these
proposita is a thesis that seems extremely plausible before philosophical
argument and each, taken seriously, conduces to a distressing life. Cicero
regards it as his job as Academic speaker to “cover up my own opinion and
relieve others of error, so that we may inquire into what is most like the
truth.” (Tusculans .) By getting his interlocutors to withhold assent
from the distressing beliefs he hopes to relieve them of especially troubling
error, and he allows them to form a suitably Radical view of what seems
like the truth on each issue, at least for now.
We can, then, divide the surviving parts of the Late Sequence between
these two types of dialogue. Academica, On ends, DND, and Div. feature
continuous speeches for either side of an issue. Tusculans and On fate
feature arguments against a propositum. Cicero arranges the machinery of
the dialogues so as to dramatize the outcomes of an Academic expos-
ition. He, as the author, is the authority figure. He arrogates some of
this authority to himself by attaching prefaces to the dialogues in which he
frames their project and importance and by narrating, in the same author-
ial voice, all the events of the conversation. When we “hear” a character
speaking, we hear Cicero narrate the speech. But Cicero then refuses to put
this authority behind any philosophical argument, since those arguments
are distributed among characters in the narrative. While we know that
Cicero is an Academic, and he often appears as an Academic character, we
see that even the Academic characters only deliver arguments demanded by
their role in the conversation. Thus Cicero never interposes his own
authoritative view, leaving the reader’s judgment free. But he also
dramatizes the other part of the procedure, that at a particular time one
side of the issue may seem to each person more like the truth. This he
achieves by his own presence as a character in the dialogue. Thus he

Tusculans ., ., ., ., ..

I thus disagree with Griffin () , who thinks that the Tusculans are dogmatic. In Div. .,
aperuerunt means “to open up,” i.e. to make available, and it is the locus, i.e. the set argument
contained in Tusculans , which docet, “teaches,” that virtue is sufficient for happiness. Tusculans
 as a whole does not teach this.

The exception is On ends Book , where Piso receives no answering speech.

In what follows, I am in substantial agreement with Schofield ().

Some of his prefaces are more intimately connected with the conversation that follows than others.
See p.  below.

There is, of course, some bleeding from the identity of the “Marcus” characters to Cicero the
author. In On ends ., for example, the authorial voice claims to add his own judgments, by which
he seems to mean the views of the character.
. Cicero and the Skeptical Dialogue 
can either, as in On ends, show us his skeptical self wrestling with and
changing his mind about which side of a conundrum seems most like the
truth or, as in DND and Div., construct a stable view for himself (see
Chapter ).
These two forms of Academic writing required some changes to Cicero’s
earlier thinking about dialogues. As we saw in Letters to Atticus .
(pp.  and  above), Cicero changed his habit from the Heraclidean
model of characters from history to the Aristotelian model of taking a role
himself and using his contemporaries as interlocutors. He gives more of his
reasons for this change, as Letter . continues:
ita confeci quinque libros περὶ Τελῶν ut Epicurea L. Torquato, Stoica
M. Catoni, περιπατητικὰ M. Pisoni darem. ἀζηλοτύπητον id fore putaram
quod omnes illi decesserant. haec Αcademica, ut scis, cum Catulo, Lucullo,
Hortensio contuleram. sane in personas non cadebant; erant enim λογικώτερα
quam ut illi de iis somniasse umquam viderentur. itaque ut legi tuas de
Varrone, tamquam ἕρμαιον adripui. aptius esse nihil potuit ad id philosophiae
genus, quo ille maxime mihi delectari videtur . . .
In the five books which I composed On ends I gave the Epicurean case to
L. Torquatus, the Stoic to M. Cato, and the Peripatetic to M. Piso.
I thought it would excite no jealousy, since none of them was still living.
This treatise on the Academy I had given, as you know, to Catulus,
Lucullus, and Hortensius. It must be confessed that the matter did not fit
the persons, who could not be supposed ever to have dreamed of such
abstrusities. So when I read your letter about Varro I seized upon it as a
godsend. No name could have been better suited to that brand of philoso-
phy [=Antiocheanism] in which he seems to me to have taken a particular
pleasure. (Letters to Atticus ..– = SB , translation from
Shackleton Bailey (), modified)
For one thing, Cicero was under social pressure (“jealousy”) to give a part to
friends and acquaintances or to an important literary colleague like Varro
(cf. Letters to Atticus .=SB). But for another, he knows that even his
older contemporaries like Hortensius, Catulus, and Lucullus did not really
know about the details of philosophy. It was therefore implausible to make
them speak such details. But his objective in the Late Sequence was to take
the matter of technical Greek philosophy and “illuminate” it, to bring it to
life. He therefore needed speakers whom his educated readers – many of
whom would know many of the characters personally – could believe knew

For a more detailed study of Cicero’s skepticial use of the for and against structure of On ends, see
Brittain (), while Schofield () analyzes the skeptical propositum structure in Tusculans.
 Cicero and the Translation of Philosophy from Greece to Rome
enough. This, then, is a reason to switch from the Heraclidean model to
the Aristotelian: he can have his own character conduct the Academic parts
of the argument with people who plausibly knew and were enthusiastic
about their arguments. Of course, the stock of such people even among
Cicero’s age group was probably small. So we see him take steps to make
knowledge plausible. He finds Cato in a heap of Stoic books, primed for his
part in On ends (.). In DND he uses Stoic and Epicurean speakers,
Velleius and Balbus, who do not seem to have been known for much
besides their respective philosophies (DND . cf. On the orator .). In
Div. his brother Quintus has been reading DND and wants to talk about
it – perhaps we are to think he has been swotting up.
This concludes my survey of the evidence for the thought that Cicero
put into his dialogues, in relation to his classical models and for the overtly
skeptical writing of the Late Sequence. Now from Cicero’s classical
models, and from his scorn of his Epicurean predecessors who could not
illuminate their material, we might extrapolate the following conjecture.
He complained that, while the Epicureans may present their views accur-
ately, they could not make them vivid or attractive. So why, he asked,
would anybody read them who does not already agree with them (see p. 
above)? Now, we do not know enough about the full variety of Hellenistic
philosophical writing in Greek to say what it was all like. But it seems that
much of it might have been vulnerable to the criticisms which Cicero
makes against the Latin Epicureans – even if its philosophical content was
of the highest quality. Epicurus’ own writings surely would not meet
Cicero’s demands, nor would the Herculaneum texts. As to the Stoics,
Cicero certainly regarded their greatest thinker, Chrysippus, as a poor
stylist. (On the orator .) Thus perhaps Cicero thought not that the
Latin Epicureans were uniquely bungling, but that a lot of Greek philo-
sophical writing since Plato, Aristotle, and Heraclides had gone the same
way. Roman philosophical writing, with Cicero as its exemplar, would
bring written philosophy back to life.
. Cicero’s Theological Trilogy

In the catalog of his works in Div. .–, Cicero describes DND, Div. and
On fate as follows:
quibus rebus editis tres libri perfecti sunt de natura deorum, in quibus omnis
eius loci quaestio continetur. quae ut plane esset cumulateque perfecta, de
divinatione ingressi sumus his libris scribere; quibus, ut est in animo, de fato
si adiunxerimus, erit abunde satis factum toti huic quaestioni.
. Cicero’s Theological Trilogy 
Once [the Tusculans] were produced, I completed three books On the
nature of the gods, in which the whole inquiry about that topic is contained.
So that I might complete this inquiry clearly and copiously, I have begun to
write these books On divination. If, as I have in mind to do, I join to them
On fate, then a generously sufficient amount will have been done for this
whole inquiry. (Div. .)
This description shows that one quaestio, “inquiry”, runs through all three
of DND, Div., and On fate. In Chapter  I shall show what that “inquiry”
is, and in my remaining chapters I shall show how carefully DND and Div.
conduct it. But, while I use its material, I do not offer an extended
interpretation of On fate. Why not?
On fate is partly lost. Among other parts, we have lost the propositum
that Hirtius gives (see pp. – above), and we have lost the end of the
dialogue. We do not know what Marcus argues against, how he goes about
the argument as a whole, or whether he, or Cicero, said anything which
helped to illuminate Cicero’s own stance on the inquiry of DND and
Div.. Thus while On fate provides rich material for the jigsaw approach
to Hellenistic views of fate, and therefore sheds much light on DND and
Div., it is not amenable to my methods in this book.

Given what Marcus seems to be arguing in On fate (cf. n.  above) it is unlikely that he agrees
with the Stoics. But there is some evidence that, for some reason, he advocated a view of fate
related to, but distinct from, the Stoic one. On fate would thus have shown us an aspect of theology
where Cicero was closer to the Stoic view than in Div. but further than he seems in DND. One of
the short fragments of On fate (fr.  Yon, Sharples) is the following passage of Servius’ comment on
Aeneid . volvitque vices: definitio fati secundum Tullium, qui ait, fatum est conexio rerum per
aeternitatem se invicem tenens, quae sub ordine et lege variatur, ita tamen ut ipsa varietas habeat
aeternitatem, “A definition of fate according to Cicero, who says, ‘fate is a reciprocal
interconnection of events through all eternity, which varies according to order and a law but so
that the very variation itself has eternity.’” We might compare this with Quintus’ definition of fate
in Div. .: ordinem seriemque causarum, cum causa causae nexa rem ex se gignat. Ea est ex omni
aeternitate fluens veritas sempiterna. “An order and series of causes, where cause bound to cause
produces an event out of itself. It is eternal truth flowing from all eternity.” Servius’ “Tullian”
definition is similar to Quintus’ definition on the surface, especially in the emphasis on eternal
order and the interconnection of things. It might, indeed, be compatible with Stoicism, although it
is not the same as other well-known Stoic definitions, SVF .–. But it is technically quite
different from Quintus’ definition. Quintus says that fate is about the connection of causes and gets
the Stoic account of causation just right (a cause is a body producing an incorporeal effect “at” or
“for” another body), and he says that an important consequence of this is that propositions about
the future or the past can be made true now. The “Tullian” definition is about the connection of
res, not causes. The “eternity” involved in it is the eternity of some natural law whereby the world
proceeds. This natural law seems to imply physical determinism but not necessarily the
involvement of a divine, rational planner. There is no mention of eternal truth. Could it be that
this definition really is Cicero’s own rather than a Stoic one and that it was found in On fate?
 Cicero and the Translation of Philosophy from Greece to Rome
Is this a problem? Given my emphasis on the literary unity of Cicero’s
writings in the Late Sequence, you might think that missing a part of the
trilogy described in Div . leaves too large a gap. But there is reason to
think that DND and Div. on their own give us a complete picture of
Cicero’s project. This is thanks to a very plausible history we can recon-
struct, of Cicero’s writing of DND, Div., and On fate.
Div. seems to have been begun before the death of Caesar, when Cicero
still felt pushed out of public life (Div. .). But some of it was
certainly finished after Caesar’s assassination (e.g. Div. .). Cicero says
that, returned to political influence, he might no longer have time for
philosophy. (Div. .) Thus in Div. . he is not sure whether he will write
On fate. In the event, he did write On fate, but in the first section of that
work (p.  above) he tells us that he has changed his plan for its structure.
The circumstances he cites – Hirtius asking for a display – are a literary
artifice. Cicero’s dialogues are not transcripts of real conversations.
But the propositum format is shorter than speeches on either side,
since only one side of the case is argued. It is natural to think that this
was one of Cicero’s reasons to write On fate as he did, as he seems to imply
in On fate .
So what Cicero suggests happened is this. He was writing Div. when
Caesar was killed. Foreseeing that his leisure would be curtailed, he wrote
up a bibliography and coda for all his dialogues. He prepended this to a
book he had in hand, book  of Div., as its preface. Finding some time, he
finished On fate in a shorter format than originally he had planned. Now
you might argue that all this is merely a narrative gambit by Cicero,
designed to convey the disruption and activity into which he was thrown
rather than to record how he wrote the dialogues. You might be right. But
it seems to me that what Cicero suggests is very plausible. I am inclined to
accept it as the truth.
If the picture Cicero gives us is correct, then when he put the finishing
touches to Div. he was not sure that he would ever write On fate, nor of the
form the work would take if he found time to write it. So when we see that
the end of Div. seems to be an ending appropriate to the project set out in
DND (as I shall argue in Chapter ), we are entitled to think that Cicero
wrote that ending of Div. on the assumption that it might stand as the end

On the date of Div., see n.  above.

But Hirtius really was with Cicero around the time On fate was written and he did ask for a
declamation, Letters to Atticus .. = SB on  April  BC.
. Cicero’s Theological Trilogy 
of the inquiry he began in DND. Thus I think it legitimate to draw the
conclusions I do about DND and Div. as a unity, even though vital parts of
On fate are lost. For this reason, in the chapters that follow I use evidence
from On fate where appropriate, but I interpret DND and Div. as two
works that form one whole.

I anticipate here another objection to what follows. Some readers interested in Cicero’s project in
writing his dialogues will be disappointed that I rarely mention the political context in which
Cicero wrote DND and Div.. That is, I have suggested that Cicero claims to help Roman society
by giving it new intellectual resources, but I do not ask whether or how DND and Div. comment
on their contemporary political realities, like Caesar’s dictatorship and death. Those realities were,
of course, extraordinarily urgent. They must have intruded on Cicero’s mind. When we look for
Cicero’s comment on these circumstances, one place to look is in these two texts. It is all the more
natural to do so when Cicero says that the philosophical writings constituted a continuation by
other means of his political life. (Div. .–) Nevertheless, I find that we do not need to look to
political realities to explain anything about the large-scale features of DND and Div.. The
intellectual project of DND and Div., which I outline in this introduction and in Chapter 
below, is sufficient on its own to explain the dialogues on a large scale. That Cicero might also have
other, political reasons for giving the dialogues the form and arguments that he does is not
excluded by this conclusion. But if we try to reconstruct such further reasons, we should take
account of the full and properly understood philosophical and literary sophistication of
these works.
 
Cicero’s Project in On the Nature of the

Gods and On Divination
I shall argue that Cicero tells us the point of DND in the dialogue’s first
sentence:
cum multae res in philosophia nequaquam satis adhuc explicatae sint, tum
perdifficilis, Brute, quod tu minime ignoras, et perobscura quaestio est de natura
deorum, quae et ad cognitionem animi pulcherrima est et ad moderandam
religionem necessaria.
While many matters in philosophy have not at all had sufficient treatment
yet, inquiry into the nature of the gods—as hardly escapes you, Brutus—is
particularly difficult and thoroughly opaque. This inquiry is both most
beautiful for the mind to grasp, and necessary for the moderation of
religion. (DND .)
Cicero tells Brutus and the reader that inquiry into the nature of the gods
is attractive in two ways. First comes beauty. Beauty will matter in the end
but it is not what Cicero takes up in the rest of his preface. Instead he
elaborates the second point: moderation of religion. Staging a philosophical
inquiry into the nature of the gods in the hope of moderating religion is,
I shall argue, Cicero’s project in DND and Div..

Whether with cognitionem or the alternative reading agnitionem, in recent centuries this phrase has
most often been interpreted with animi as an objective genitive: “best for the grasp of the soul,” that
is, best in order to understand our own souls (see Davies (), Mayor (), Pease ()). But
that does not seem to be the use that Cicero makes of the inquiry in DND and Div., or of the beauty
we encounter in it. I prefer to take animi as a subjective genitive, in which I have the support of
Walsh (): “the noblest of studies for the human mind to grasp.”

Cicero seems to suggest that Brutus had some marked acquaintance with the question of the nature
of the gods. Perhaps this is flattery or refers to Brutus’ general philosophical learning (for which see
Sedley ). Perhaps the Antiochean Brutus’ treatise On the blessed life included some material on
contemplation as imitation of the divine (cf. Tsouni () –). Or perhaps he had ruminated
on the sort of Antiochean theology suggested at Boys-Stones () – or Blank ()
–.

For the crucial role of beauty in Cicero’s understanding of DND and Div., see Chapter  section
.. and Chapter  section ..

Cicero’s Project in On the Nature of the Gods and On Divination 
That a preface should tell you the point of the work to follow might
sound unsurprising. But with Cicero’s philosophical works it can be
doubted. The doubts stem from a letter to Atticus, who oversaw copying
of Cicero’s works. In the letter Cicero confessed that he attached to a
manuscript of his On glory the preface he had already used for Book  of
the Academica. “This happened because I have a volume of prefaces from
which I am in the habit of selecting when I have put a work in hand.”
(Letters to Atticus .. = SB ) The letter suggests that Cicero is
careless about his prefaces and that any of them could easily be cut from
one work and pasted into another. If so then perhaps the prefaces are
rhetorical exercises, standing free from the work to which they are glued.
But there can be no such doubts about the preface to DND. It is plain
that a large part, if not all, of it was written specifically for a philosophical
work on the nature of the gods. For half of it is directly and explicitly
concerned with the significance of the question of the gods’ nature and the
import of philosophers’ views in the matter. (DND .–, .–)
The remainder of the preface takes up the defense of philosophical
writing and of Cicero’s Academic skepticism in particular (DND
.–). Now Cicero could have taken this passage from a prewritten
preface and fitted it into DND. But it melds seamlessly with its surround-
ings. For, Cicero’s opening points with which we began (DND .), that
the question of the nature of the gods is difficult and important, are also
useful to him in this defence of skepticism. In the nature of the gods, he
suggests, we find a particularly important question, on which philosophers
are particularly prone to disagreement, because it is particularly difficult.
This should help us to see why skepticism, the withholding of assent to any
dogmatic answer about a question, can be due caution: on the nature of
the gods, no one can (yet) be confident of her answer, and a wrong answer
would be a disaster. Thus the conceit of Cicero’s quaestio, his “inquiry.”
In the preface to DND Cicero presents this inquiry not only as the
philosophical investigation about which his characters are to dispute, but
also as a quaestio in the sense of a session of a court: “on this topic (quo
quidem loco) it seems I should summon all people to judge what of these
<philosophical views> is true.” (DND .) We the audience are to hear
the speakers Cicero brings before the inquiry. The preface concludes that
we should see

id evenit ob eam rem quod habeo volumen prohoemiorum. ex eo eligere soleo cum aliquod σύγγραμμα
institui. Translation from Shackleton Bailey ().

quo quidem loco convocandi omnes videntur, qui quae sit earum vera iudicent.
 Cicero’s Project in On the Nature of the Gods and On Divination
. . . quid de religione pietate sanctitate caerimoniis fide iure iurando, quid de
templis delubris sacrificiisque sollemnibus, quid de ipsis auspiciis, quibus nos
praesumus, existimandum sit (haec enim omnia ad hanc de dis inmortalibus
quaestionem referenda sunt): profecto eos ipsos, qui se aliquid certi habere
arbitrantur, addubitare coget doctissimorum hominum de maxuma re tanta
dissensio.
. . . what we should reckon about religion, piety, holiness, rites, good faith,
and oaths, about temples, shrines, and solemn sacrifices, about the auspices
that I myself oversee (for we must relate all these things to our inquiry about
the immortal gods); certainly such great disagreement among the most
learned, about a matter of the greatest importance, will compel those who
judge that they themselves have some certainty, to hesitate. (DND .)
Thus even if part of the preface to DND were drawn from a roll of
prefabricated paragraphs, all of it is fitted to the dialogue it introduces,
either directly, or indirectly by recommending that skepticism is an
appropriate response to this inquiry in particular. The preface tells us the
point of the dialogue and also, as I shall argue, the point of Div.
If I am right about Cicero’s project, it poses a puzzle. For in Cicero’s
Rome religio, what I have innocently called “religion,” was in important
respects unlike what we tend to call “religion” today. For example, theo-
logical beliefs do not seem to have played much part in how a Roman and
an augur like Cicero, who “oversaw the auspices,” regulated his religio.
But on the face of it beliefs are all that a philosophical inquiry could
change, or at least all that it could change directly. So what could Cicero
mean when he says that his quaestio will moderate religio? We turn first to
that puzzle.
. Action, Belief, and Roman religio

Many students of “Roman religion” stress a difference between what they
study and the religions we in the west tend to think of today. In the
western world the religions that come quickest to mind, like Christianity,
Judaism, or Islam, tend to require certain beliefs. For example, if you do
not believe in God, many Christian authorities would say, you are not, or

When Cicero wrote Div., he had been a member of the college of augurs for about eight years. For
the circumstances of his election, see pp. – of Linderski (). Marcus in the Laws claims to
be very proud of his position as an augur (.). In Div., Cicero has Quintus remind Marcus, and
the reader, of this aspect of his own interest in the subject of the dialogue (.), and Marcus alludes
to debates inside the college of augurs in his own speech (., cf. p.  n.  below). Letters to his
friends .. gives us a sense of how these debates played out in writing. For the tradition that Cicero
wrote his own On augury, see p.  n.  below.
. Action, Belief, and Roman religio 
not fully, a Christian. So we tend to think of religions as requiring at least a
certain orthodoxy. But what we put under the heading “Roman religion”
was a set of institutions that do not seem to have required any particular
beliefs from their adherents. Instead they required only particular actions.
So we think of “Roman religion” as requiring not orthodoxy but rather
what is sometimes called orthopraxy.
In this book I use this widely accepted model of “Roman religion” as
orthopractic. I do so because, as we shall see, I think it is also the model
Cicero gives to his characters in DND and Div.
For my purposes, what do I mean by “Roman religion”? We can get an
approximate answer from an early part of Cotta’s skeptical speech in Book
 of DND, where he declares his determination to defend something like
“Roman religion.” Cotta is a traditionally minded member of the chief
priestly college, the pontifices, just as Cicero himself was a member of the
chief college of diviners, the augurs. Thus we should expect from Cotta an
attempt to sum up what Roman religio amounts to, from the point of view
of someone in traditional religious authority:
cumque omnis populi Romani religio in sacra et in auspicia divisa sit, tertium
adiunctum sit si quid praedictionis causa ex portentis et monstris Sibyllae
interpretes haruspicesve monuerunt, harum ego religionum nullam umquam
contemnendam putavi mihique ita persuasi, Romulum auspiciis Numam sacris
constitutis fundamenta iecisse nostrae civitatis, quae numquam profecto sine
summa placatione deorum inmortalium tanta esse potuisset.
Although the whole religion (religio) of the Roman people is divided into
rites and auspices (and a third part is added when the haruspices or the
interpreters of the Sibylline books have given some predictive warning
derived from portents or prodigies), I myself think that none of these
religious duties (religionum) is ever to be despised, and I have persuaded

Some textbook examples of this description of Roman religion: Beard, North, and Price () vol.
pp. x, –, Scheid () –, –. Recently some scholars have asked whether this picture
captures all there is to say about Roman religious life. Examples are Ando () –, Scheid
(). Rüpke () offers a story of “rationalization” of religion in the late Republic, but by this he
does not mean in general the sort of philosophical rationalization I attribute to Cicero, but rather
“the ordering and systematization of concepts, practices or instruments used to reach particular ends”
(p. ). Boys-Stones () – similarly sketches the significance of ancient orthopraxy for ancient
philosophical approaches to religion.

Cicero himself was able to entertain other models. Gildenhard (, pp. –) points out that
in De domo sua  Cicero implies that if the pontifices side with Clodius’ impiety, then they are not the
religious authorities after all. Gildenhard says that, “on this premise, it is the philosopher who tells
the pontiffs what is and is not holy, and not the pontiffs who decide whether an act of religio has been
performed.” (p. , emphasis added) This exordium thus goes further than does any voice in DND
or Div..
 Cicero’s Project in On the Nature of the Gods and On Divination
myself that Romulus, with the institution of the auspices, and Numa, with
the rites, laid the foundations of our state which would certainly never
have been so great without the greatest propitiation of the immortal gods.
(DND .)
Let us say that the “Roman religion” of Cicero’s day was roughly what
Cotta calls here religio (the singular) or religiones (the plural). These are the
sorts of institutions (like priesthoods, temples, or temple buildings and
property) or required actions (like sacrifices, prayers, the taking of the
auspices, the expiation of prodigies, or the swearing of oaths) that a
pontifex like Cotta, or an augur like Cicero, might be called on to oversee.
If we open a book on “Roman religion” today we are likely to find more or
less this agglomeration of subjects. For Cicero’s idealized but recognizable
version of Cotta’s traditional institutions and requirements, we may read
his Laws. (Laws .–) This loose set of institutions and practices, then,
is what I shall generally call “Roman religion” in this book. It is this
religion that Cicero’s characters assume to demand actions, but not beliefs,
from its practitioners.
Now when philosophers (ancient or modern) think about the idea of an
action they tend to conclude that it is complex, and that it requires an
agent to have some beliefs. “Action” here means something more than a
movement of the body. It refers to the actions of an agent, the sort of
actions for which she is responsible. When my leg twitches thanks to some
reflex, or when somebody else forces my arm to move, those movements of
my body are not my actions. Further, when philosophers try to decide
what action has been done, the agents’ mental attitudes, like her beliefs and
desires, often come into play. When somebody utters a falsehood, whether
or not she lied depends on whether she meant to deceive. If somebody
decides not to pay the taxes he owes, it might be tax evasion, or an honest
mistake, or a political protest, and our assessment of the morality of his
decision will differ accordingly. When we look at actions from this
philosophical perspective, the claim that Roman religion is orthopractic
and not orthodox might look naïve. For if actions are in part to be

On priesthoods under the republic, see Beard (). North () weighs some of the difficulties
in treating Roman traditional religio as “religion,” and considers the role of Romans like Varro and
Cicero in forming the modern notion of “religion.”

To take the three philosophical schools represented in DND as examples: for the Stoics on action,
see p.  n.  below. In Academica Marcus’ Academic engagement with the Stoic theory of action
is at .–. For the Epicureans, see Lucretius .–, .–. Discussions, which
focus on issues of freedom, responsibility, and the swerve, include Furley (), Purinton (),
Bobzien (), and O’Keefe ().
. Action, Belief, and Roman religio 
distinguished by what the agent believes, then perhaps doing certain
actions requires certain beliefs.
But in fact when we say orthopraxy requires only actions, we use
“actions” in a qualified sense: Roman religion required public or outward
actions. That is, it required the right bodily movements and utterances, at
the right time, in the right place, in the right dress, and so on. These are
aspects of actions that (in principle) other people can observe. So far as
religious requirements went, in making such movements a Roman could
think what she liked about them. If she dropped a pinch of incense on an
ember at the right time and place, she would have succeeded in her
relevant religious duty whether she thought she was doing it for a god
who was identical with, or who simply looked like, the statue in front of
her, or for some force of nature whom the statue personified, or for no god
at all – and so on. For clarity I shall distinguish two terms from now on.
On the one hand, I shall call the outward, observable aspect of an action a
“performance.” On the other, I shall call the fully described action, with its
agent’s relevant beliefs and desires taken into account, an “action.” Using
the terms that way, we may say that the orthopractic Roman religion
demanded certain performances, but did not mind what actions those
performances were part of.
My general answer to our puzzle about the moderation of religion will
rely on the distinction I have just made between action and performance.
In very large part, the participants in Cicero’s inquiry do not hope to
change which performances were required at Rome. But just because the
traditional religion that he accepted demanded only performances, it does
not follow that a private individual like Cicero, or any Roman, was unable
to think that such religiously correct performances were made into pious or
impious actions by what the agent believed she was doing. If two men
burn some incense exactly as a pontifex would recommend, one to do
honor to a god of whom he has an accurate view, the other with the
thought that no god knows what he is doing or would care that he does so,
a Roman could think that both men made a religiously correct perform-
ance, but that the first man did a pious action while the second did an
impious one. Hellenistic philosophy offered a choice of theoretical justifi-
cations for such views. While philosophers at large could agree that they all

Of course, among the motivations of somebody making a religiously correct performance is likely to
be precisely the desire to make a religiously correct performance as such. Thus in my terms, a
religiously correct performance is likely to be part of an action aiming (perhaps among other things)
at orthopraxy.
 Cicero’s Project in On the Nature of the Gods and On Divination
performed rituals as demanded by a priest like Cotta, an adherent of each
dogmatic school would add that, because of their differences in theology,
she herself satisfied the ritual prescription well while representatives of the
other schools did not. If we ask by what set of values “pious,” “impious,”
or “well” are measured, the answer is not the criteria of religious correct-
ness, which in this example we suppose are satisfied by all, but rather the
criteria of the ethics and theology recommended respectively by each
philosophical school.
We should reflect that in this way DND and Div. are further evidence
for the orthopractic view of Roman religion. The characters do not
propose to change any points of “outward” Roman religion. It is not
that they are unaware that the religious tradition has changed over time
and could well change again. The debate between Balbus and Cotta
I examine in Chapter  is in part premised on past changes. But no
character in DND or Div. proposes further changes of his own. The
changes the characters propose are in what to believe about the religion.
Futhermore, when confronted with one another’s contradictory views
about what to believe about religious orthopraxy, the characters do not
accuse one another of failing to adhere to its demands. This is despite the
clear implication of the speeches of the characters in DND, that each
thinks that the others are impious, or superstitious, or both. This behavior
itself shows that Cicero imagined such debates would proceed on the
assumption that the requirements of the “outward” religion as such did
not extend to what participants should believe about the nature of
the gods.

Beard, North, and Price () vol. pp. x, – point out that for Romans through much of the
period of Republic, there was little opportunity for an individual Roman to find fulfillment in a
private religious group, as we might find fulfillment by converting to Christianity or Buddhism.
DND shows us how educated Romans of the Late Republic could adopt philosophical schools (or
other schemes of thought, perhaps) and privately interpret their public religious actions in a way that
these too might give them a similar sort of fulfillment.

Moatti ()  similarly says that in Div. Cicero will not question “Roman religion as an
ensemble of practices and rites. He simply criticizes a mode of belief . . .” But I cannot agree with her
further thought that, for Cicero, the mode of belief in question is one that is “founded on passivity
and a philosophy (Stoicism) that is fundamentally incapable of thinking through religio.” Rather,
the mode of belief in question is rash assent to answers to the Central Question.

Balbus is suspicious of the worship of what he sees as vices (see pp. –). But he does not
explicitly propose any changes to such cults at Rome.

Some students of anthropology might think that even to accept the model of orthopraxy and belief
that I give here, where beliefs about orthopractic religion may render religious actions pious or
impious, gives too much weight to the notion of “belief.” For it is sometimes argued that “belief”
itself is a culturally contingent notion, conditioned by Abrahamic faiths, not necessarily found
elsewhere. Thus you might say that even my model “Christianizes” if not Roman religion, then at
least the Romans’ experience of their religion. (Feeney  pp. – gives an introductory
. Theological Facts and Conventional Piety 
. Theological Facts and Conventional Piety

The last section was a start in solving our puzzle about philosophy and
religion. But it remains that Cicero said that his inquiry is ad moderandam
religionem necessaria, “necessary for the moderation of religion.” (.) Yet
I should not think that in this sentence he means by “the moderation of
religion” reform of the performances that Roman religion required. For
I have said that no character in the dialogues calls for significant change to
those forms. Then what does Cicero mean? I shall argue that Cicero thinks
philosophical inquiry into the nature of the gods can moderate religious
performances by changing the agent’s beliefs about those performances, so
that she at least has no false beliefs such as would, if she had them, turn the
performances into impious or superstitious actions.
In order to get at the importance of this inquiry, in the preface to DND
Cicero assumes a certain common-sense view of piety and religion, and
points out that the fact of the matter about the nature of the gods makes
this view either true or false. Thus by holding fixed some theological beliefs
and considering alternative sets of theological facts, he furnishes the reader
with an example of the importance of the truth or falsity of theological
beliefs in deciding the right or wrong view of religion. I shall examine that
argument from his preface in this section. In the next section I will ask
how, if we did not hold our beliefs fixed, Cicero thinks philosophical
inquiry can change them for the better.
Why would Cicero need to tell us that theological facts matter for
religion? Is that not obvious? In fact, for the reasons we have seen, it
might not have been obvious to a Roman. For an orthopractic religion
might be conceived and justified in many ways. Perhaps, for example, it is
a set of ceremonies useful for holding society together, regardless of the
nature of the gods. Cicero sets out to suggest how, for one plausible
understanding of religion and its related virtues, the theological facts do
matter.
I turn first to the long passage of DND’s preface where Cicero presents
the importance of the question of the nature of the gods. I shall quote this
crucial text in full. Cicero has just said that everybody but Protagoras (who
hesitated), Diagoras of Melos, and Theodorus of Cyrene (who both
discussion of such questions in the context of Roman religion.) I answer that in Cicero’s corpus in
general, and in the Academica in particular, there is a great deal of painstaking discussion of what
various philosophers had understood by what we would call “beliefs” (of which Chapter  section
. gives a sample). It is these ideas I intend when I speak of “beliefs” about religion, and not
necessarily to invoke any modern notion of specifically “religious belief” or “faith.”
 Cicero’s Project in On the Nature of the Gods and On Divination
declared that there were no gods) has agreed that there are gods. Further,
those who believe in the gods have arrived at this most plausible view duce
natura, “with nature as a guide.” (.) This is supposed to explain why
Cicero will write about the nature of the gods: their existence may be
assumed, so that their nature is what is at issue. He continues (Roman
numerals are mine):
[i] (.) . . . qui vero deos esse dixerunt tanta sunt in varietate et dissensione,
ut eorum infinitum sit enumerare sententias. (.) nam et de figuris deorum et
de locis atque sedibus et de actione vitae multa dicuntur, deque is summa
philosophorum dissensione certatur; quod vero maxime rem causamque con-
tinet, utrum nihil agant nihil moliantur omni curatione et administratione
rerum vacent, an contra ab iis et a principio omnia facta et constituta sint et ad
infinitum tempus regantur atque moveantur, in primis magna dissensio est,
eaque nisi diiudicatur in summo errore necesse est homines atque in max-
imarum rerum ignoratione versari.
[ii] sunt enim philosophi et fuerunt qui omnino nullam habere censerent
rerum humanarum procurationem deos. quorum si vera sententia est, quae
potest esse pietas quae sanctitas quae religio? haec enim omnia pure atque caste
tribuenda deorum numini ita sunt, si animadvertuntur ab is et si est aliquid a
deis inmortalibus hominum generi tributum; sin autem dei neque possunt nos
iuvare nec volunt nec omnino curant nec quid agamus animadvertunt nec est
quod ab is ad hominum vitam permanare possit, quid est quod ullos deis
inmortalibus cultus honores preces adhibeamus? in specie autem fictae simula-
tionis sicut reliquae virtutes item pietas inesse non potest; cum qua simul
sanctitatem et religionem tolli necesse est, quibus sublatis perturbatio vitae
sequitur et magna confusio; (.) atque haut scio an pietate adversus deos
sublata fides etiam et societas generis humani et una excellentissuma virtus
iustitia tollatur.
[iii] sunt autem alii philosophi, et hi quidem magni atque nobiles, qui
deorum mente atque ratione omnem mundum administrari et regi censeant,
neque vero id solum, sed etiam ab isdem hominum vitae consuli et provideri;
nam et fruges et reliqua quae terra pariat et tempestates ac temporum varietates
caelique mutationes, quibus omnia quae terra gignat maturata pubescant, a dis
inmortalibus tribui generi humano putant, multaque quae dicentur in his
libris colligunt, quae talia sunt ut ea ipsa dei inmortales ad usum hominum
fabricati paene videantur.

By duce natura, Cicero may mean that most philosophers have arrived at theism as a result of their
study of natural science. But it seems more likely that Cicero means to agree with the speakers in
DND, that human nature – our psychology, we might say – tends to lead to belief in the gods. Of
course, this view is not so pausible today, when we know so many atheists or agnostics. An
argument against the view, represented by Cicero here, that ancient intellectual history was
overwhelmingly theist, is Whitmarsh ().
. Theological Facts and Conventional Piety 
[i] (.) . . . But there is such great disagreement and diversity among
those who have said that there are gods that it would be an endless task to
list their views. (.) For they make many claims about the gods’ forms,
about where they are and their seats and about the life they lead, and these
issues are contested as a great controversy among the philosophers. But
there is especially great disagreement about the issue which most of all
comprises the case and the matter at hand: whether the gods do nothing, work
at nothing, and take no care of, nor govern, affairs, or whether, on the contrary,
everything was made and set up by them from the beginning and is ruled and
set in motion by them into infinite time. Unless this issue is decided, it is
necessary that humanity live in the highest error and in ignorance of the
greatest matters.
[ii] For there were and are philosophers who hold that the gods take no
care whatsoever of human affairs. If their view is true, what piety (pietas)
can there be, what holiness (sanctitas), what religion (religio)? For all these
things are to be rendered (tribuenda) to the gods’ persons with purity and
without pollution (pure atque caste), on this condition: if the gods notice
them and if something is rendered by the immortal gods to the race of
humans. But if they neither can help us nor wish to, neither care at all what
we do nor notice, if there is nothing which can flow from them through to
human life, what reason is there (quid est quod) for us to apply to them cult,
honors, and prayers? Just as with the other virtues (virtutes), there can be no
piety in the appearance of invented pretense. Of necessity, holiness and
religion are taken away when piety is – and when those are taken away, a
troubled life follows, and great disorder. (.) Indeed it might be that when
piety towards the gods is taken away, then even good faith (fides) is taken
away, and the community (societas) of the human race, and with them the
most excellent virtue, justice.
[iii] But there are other philosophers, these being great and noble
philosophers, who hold that the whole cosmos is governed and ruled by
the mind and reason of the gods, and not only that, but even that the gods
care about human life and are provident. For they think that the immortal
gods render to the human race crops and the rest of what the earth bears,
weather and the variety of the seasons and the changes of the heavens by
which ripens and matures everything that the earth brings forth. They
talk about many things (which are collected in these volumes) that are
such that the immortal gods almost seem to have made those very things
for human use. (DND .–)
Cicero does not name any philosophers in this passage. But the obvious
representatives of the two positions he describes are in [ii] the Epicureans,
who deny the gods’ involvement in our lives, and in [iii] the Stoics, who
endorse it.
My paragraph [i] is satisfying reading for modern students of ancient
Rome (or Greece). For many of us face puzzles about how the Romans
 Cicero’s Project in On the Nature of the Gods and On Divination
understood the gods they represented in literature, art, or religion, and
whom they spoke of as playing a part in history. Did they really think the
gods looked like humans? Are the gods on the Capitoline, or Olympus, in
heaven, or everywhere? Cicero shows us in DND that we are not wrong to
be puzzled, since in antiquity, too, people (philosophers if nobody else)
asked similar questions. He also shows us some possible answers to these.
But Cicero wants to focus on one question in particular: do the gods
govern the world and care about human life, or do they not? This, he says,
maxime rem causamque continet, “most of all comprises the case and the
matter at hand.” He means, I think, that this is the key question before the
inquiry he is about to stage, because it is the key question for humanity’s
relationship with the divine, and thus, for the moderation of religion.
I shall call it:
The Central Question: Do the gods care for us?
By this phrasing I mean to suggest both (a) can they and do they act in
our world and our lives, and (b) do they care about us, so that they exercise
their abilities on our behalf?
In my paragraphs [ii] and [iii] Cicero spells out what is staked on the
Central Question. Let us take his explanation step by step. First, in
paragraph [ii], we are to consider the Epicurean view that the gods do
not notice what we do and would not care if they did, that they can do
nothing for us and would do nothing even if they could. If this is true, says
Cicero, “what piety can there be, what holiness, what religion?” This is a
rhetorical question to which the answer turns out to be “none.” Piety,
holiness, and religion would be “taken away.” We have Cicero’s first move:
if the gods do not care at all about us, there can be no piety, holiness, or
religion. Now that is at first sight a baffling claim. Cicero does not say (as a
Stoic might) that if the gods ceased to care, our familiar world would
dissolve altogether. His point is rather that if it turns out that the gods have
never cared, we would be here, but piety would not. But Roman religion in
the sense of outward performance could no doubt go on in such an
absence of divine care. Indeed if the Epicureans were right, Roman religion
in that sense had already gone on for many centuries in just such an
absence. So it must be that Cicero here does not use “piety,” “holiness,” or
even “religion” in the purely outward sense. He uses them to mean
something that there can be only when the gods care. What is this?
We see the answer from a careful reading of the following sentences of
[ii]. Cicero’s next move is to give us another conditional: if the gods notice
what we do and if they render us something in return, then piety, holiness,
. Theological Facts and Conventional Piety 
and religion are to be rendered to the gods. Here the language of ethics
enters the argument. We have a reason and perhaps an obligation to give
pious religion to the gods if the gods notice the gift and themselves give in
return. But if, Cicero goes on to say, they do not notice or care or help us,
“what reason is there (quid est quod) for us to apply to them cult, honors,
and prayers?” Again, this is a rhetorical question to which the answer is,
“none.” Now these conditionals (if – and only if – the gods care about us
then we should render them pious religion) are supposed to explain the
original claim that if the gods do not care at all about us, there can be no piety,
holiness, or religion. So piety, holiness, and religion must be such that we
can say of them that there can be (e.g.) piety only when there is a reason
why there should be piety.
Cicero reveals what sort of thing he is talking about: “Just as with the
other virtues, there can be no piety in the mere appearance of invented
pretense.” Now a virtue must have a purpose, and a virtuous action must
have a purpose. Suppose I say that my courage led me to do something to
no end. If I am right that there was no end, then what led me to do it was
not, in fact, courage. Similarly, if there is no benefit in having, or obliga-
tion to have, a certain trait of character – no purpose to it, let us say – that
trait is not a virtue. So virtues and virtuous actions have the logical features
Cicero wants for his argument. There can only be a virtue where there is
something to be had or to be done.
There is more evidence for this reading in that piety (pietas) and holiness
(sanctitas) are used as terms for virtues throughout DND. Specifically, they
are used as for these terms understood roughly as the Stoics do, in
opposition to the way that the Epicureans understand them. In Appendix
 I present the standard Stoic Greek definitions of the virtues of εὐσέβεια,
“piety” (defined as “knowledge of the service of the gods”) and ὁσιότης,
“holiness” (defined as “justice towards the gods”). I then quote Cicero’s
Stoic speaker Balbus, or his Academic Cotta when arguing against the
Epicureans, as they give pietas and sanctitas respectively the same defin-
itions as in the Stoic Greek sources. I think Cicero also has these meanings
in mind for pietas and sanctitas in my paragraphs [i]–[iii] of DND .–.
Now knowledge of the service of the gods, and justice towards the gods,
have clear purposes. Cicero’s term for the “service” in question is cultus.
This noun can mean the cultivation of a field, a friend, or a friendship, as
well as the cultivation or worship of the gods. It implies a real interaction
between the cultivated and the cultivator, where one benefits the other.

For another survey of these terms in DND, see Rüpke () –.
 Cicero’s Project in On the Nature of the Gods and On Divination
Knowledge of this sort of service to the gods has as its purpose getting the
human side of such interactions right. So if gods and humans cannot
interact or benefit one another, then there can be no such purpose to such
knowledge. Thus what we thought was the virtue of piety, knowledge of
the service of the gods, is robbed of its assumed purpose and is not a virtue,
at least not in the way that we thought. Similarly, justice towards the gods
has as its purpose acting justly towards the gods. But that requires that
there is some way in which humans could act justly towards the gods. We
must be related to them in such a way that we have duties of justice
towards them, and we must be able to fulfill those duties. But if we are not
in any sort of community with the gods, so that we have no duties of
justice to them, and in any case we cannot interact with them, then we
cannot act justly towards them. What we thought was the virtue of
holiness, justice towards the gods, is robbed of its assumed purpose.
Therefore it is not a virtue, at least not in the way that we thought.
If the “piety” and “sanctity” that the Epicureans would take away refer
to virtues, what about the trickier case of “religion”? What does that mean
and how is it taken away? One possibility is that in this case Cicero uses
religio, too, for a virtue. We find this use elsewhere in his corpus. In his
youthful On invention he called religio a kind of justice not opposite to but
neighboring the vice of superstition. (On invention .–) In the On
the parts of rhetoric he says that superstition imitates religio as rashness
(a vice) imitates courage (a virtue) or severity (another vice) imitates justice
(another virtue) (Parts ). Thus perhaps Cicero just means that religio, in
the sense of another virtue by which outward religious practice is thought
to be done well, would be taken away.
But there is another possibility. Perhaps here Cicero uses religio in the
way that Cotta will use it against Velleius: religio quae pio cultu deorum
continetur, “religio which is comprised by pious service to the gods.” (DND
.) Religio in this use means not simply religious performance but
rather right religion, religious performance made with the virtue of piety.
What Cicero goes on to say in [ii] meshes with this reading. When the
gods care and respond to what we do, Cicero says, there is a reason to give
“cult, honors, and prayers” to them. “Cult, honors, and prayers” sound
like the sort of performances called for by traditional religion. He does not
say that if the Epicureans are right we should or would stop making those
performances. Rather, he says we would have lost our reason to make them
and, along with that, the piety we thought we had with which to make

Of course, in those circumstances there can also be no such knowledge.
. Theological Facts and Conventional Piety 
them well. For piety, being a virtue, is not to be found in “the appearance
of invented pretense” (specie. . . fictae simulationis). This phrase strikes
me as carefully written. First, a cult performance is a species, an appearance,
something observable. On a Stoic view of the world, this could be the
appearance of an action pious because it was done for gods who care. But
according to the Epicureans nobody has that sort of piety because the gods
do not care. So at most the outward action is the appearance of a pretense
to piety towards gods who care. Further, it is not that somebody who
earnestly tries to act piously is pretending, but failing to achieve a possible
piety towards such gods. Tradition, and philosophers like the Stoics, have
invented such a notion of piety in the first place. Thus religio in the sense of
outward religious actions done virtuously would be taken away along with
the virtues of piety and holiness.
An Epicurean might well complain that in this passage Cicero has his
thumb on the scales. As we shall see in the next chapter, Epicureans
defended their own conception of piety and, like the Stoics, advocated
the maintenance of traditional religious practice. They interpreted piety as
the virtue whereby one venerates gods who, happily, do not care about
humans. So the Epicureans would reject the claim of [ii] that, if the gods
do not care about us, then there can be no piety. Thus Cicero does not use
a philosophically neutral notion of piety in his preface. This is certainly of
a piece with Cicero’s general attitude towards Epicurus and his school.
Although in principle Academic skepticism should have made him more
neutral, Cicero never leaves much doubt that he thought Epicurus was a
bungler. So it is not surprising that rhetorically speaking the Epicureans’
role in the preface is to have only the troubling consequences of their view
put on show.
But there is more to be said in defense of Cicero’s use in this preface of
ideas of piety, holiness, and religion slanted against the Epicureans. First, it
would be reasonable for him to think that piety as understood by common
sense had to do with behaving correctly in one’s relationship with others
who paid attention to what one did, and who gave in return. Aeneas would
not be pious in an obvious sense if he were mistaken in thinking that those
whom he served cared what he did. So it seems fair of Cicero to suggest
that the Epicureans would do away with the received notion of religious
piety. Indeed, as we shall see in Chapter , Epicurus would readily agree to

Cicero and his characters often use this sort of language when attacking Epicurean ethics. See
Academica ., On friendship , On ends ..
 Cicero’s Project in On the Nature of the Gods and On Divination
this. This is what I mean when I say that Cicero assumes a certain
traditional or common-sense idea of piety in the preface.
A second point to notice in Cicero’s defence is that he does not say that
the Epicureans are wrong. Admittedly, in paragraph [iii] he makes the
consequences of the Stoic position sound more attractive. But immediately
after my quoted text he will tell us that the skeptic Carneades came up
with many provocative arguments against the Stoics. The preface, with its
skeptical set-up, is not meant to leave us convinced of Stoic theology. On
the contrary: about the facts of the gods’ nature it is meant to leave us
worried, but open-minded. Thus in this preface, Cicero unblinkingly faces
the possibility that the Epicureans are right. If they were, then the received
purpose of outward religious practice, and the virtues needed to pursue
that purpose, would have either to be abandoned, or radically to be
reconceived.
But Cicero does not think that the consequences of the Epicurean view
end there. He tells us in [ii] that in the first place, troubled life and great
disorder are bound to follow. Then perhaps along with piety, holiness, and
religion will be taken away good faith, the community of the human race,
and justice itself. So at first sight, Cicero seems to think that if Epicureans
are right then Roman society – or human society in general – might
collapse. Such a thought would be implausible, especially today when
wider experience of influential deists and atheists suggests that they are
unlikely to cause society’s downfall. So you might think that Cicero is up
to a rhetorical trick familiar from his speeches in court, where, regardless of
the real importance of the case, he tells the jury that the very survival of the
Republic is in their hands. But I think Cicero’s point is not, in fact, that
society will collapse in the obvious sense. He is not predicting riots.
First, notice what Cicero does not say here. He does not say that the
consequences he points to are a result of what people believe. He does not
say that if the Romans came to be Epicureans in large numbers then,

Graver () – says that Cicero takes the Epicurean position as “target” in the preface to
DND. This is a reasonable position, since (as Graver points out) the consequences of Epicureanism
are described in such dire terms, and since Marcus sides with Balbus in DND ., when Balbus has
disagreed with the Epicurean answer to the Central Question. However, I do not agree. First,
refuting Epicureanism is certainly not the “stated aim of the treatise.” (p. , emphasis added)
Second, affording the audience free and informed judgment on the Question is the stated aim.
(DND .–) Third, Cotta’s arguments against Balbus are very compelling, and were probably
all the more compelling before some of his speech was lost from the MS tradition, so that Marcus’
preference in DND . can come as a surprise. This would be strange if the clear goal of the
dialogue were to discourage the Epicurean answer to the Central Question.

E.g. Pro Roscio Amerina , Pro Murena –.

Of course, at Rome in the s BC, further social collapse would have been a reasonable prediction.
. Theological Facts and Conventional Piety 
because they no longer believed that gods care about us, they would cease
to act in the just and socially cohesive ways that traditional piety motivates.
Rather, he refers to the fact of the matter about the gods. If, as it happens,
the Epicureans are right and the gods do not care about us, then it follows
already that piety and perhaps justice are taken away. As Cicero states it, in
an Epicurean world there would be no piety, and there might be no
society, even if all Romans were Stoics who labored under the delusion
that the gods care and who tried to act in Stoically pious, just, and cohesive
ways. So I think that Cicero’s point here is not about how the Romans’
beliefs will lead them to behave if Epicurus is right. Instead, it is about how
a theoretician should analyze the ethics of Roman behavior. It helps to
notice that the verb I have translated “take away” (tollere) is ambiguous
between literal removal and refutation by argument. If the Epicureans are
right, piety as traditionally understood is refuted, or metaphorically
removed from the world, in the sense that it is shown never to have
been there.
In examining the wider social consequences of Epicurean theology,
Cicero first says that “good faith” (fides) might be threatened if Epicurus
is correct. This “faith” is of course not the faith of modern religions, a kind
of belief. Rather it is the good faith with which one acts bona fide. At Rome
fides could plausibly be said to be backed by religion. It was worshipped as
a goddess, Fides, with a temple on the Capitoline (cf. DND ., .,
.). Good faith was what one showed in fulfilling an oath and what one
looked for in an oath-taker. Cicero gives us a discussion of the ethics of
taking an oath in On Duties (.–). If I take an oath by Jupiter,
should I keep it because I am afraid that the god will otherwise get angry?
Cicero reminds the reader that good philosophers argue that a god will not
be prone to anger:
est enim ius iurandum affirmatio religiosa; quod autem affirmate, quasi deo
teste promiseris, id tenendum est. iam enim non ad iram deorum, quae nulla
est, sed ad iustitiam et ad fidem pertinet. nam praeclare Ennius:
O Fides alma apta pinnis et ius iurandum Iovis.
qui ius igitur iurandum violat, is fidem violat, quam in Capitolio vicinam Iovis
optimi maximi . . . maiores nostri esse voluerunt.
. . . for a sworn oath is a religious (religiosa) affirmation: and if you have
promised something by affirmation with a god as a witness you must hold
to it. What is relevant here is not the anger of the gods, which does not
exist, but justice and good faith (fidem). For what about Ennius’ splendid
words,
‘Oh winged and nurturing Faith, and oath sworn in Jupiter’s name!’
 Cicero’s Project in On the Nature of the Gods and On Divination
Therefore anyone who violates a sworn oath violates Fides, whom our
ancestors wished to dwell on the Capitol, as ‘neighbor to Jupiter Optimus
Maximus’ . . . (On duties ., translation from Griffin and Atkins (),
modified)
Such a view of good faith is what Cicero must have in mind in [ii]. When
I take an oath in good faith, even an oath in favor of another human,
I understand myself to enter into an arrangement with a god whom I have
taken as my witness. My good faith leads me to honor the promise in part
because a god has witnessed it. But if no god witnesses what we do, then
good faith understood in this way is, like piety, a fiction. Perhaps there are
other theories about why, if that were so, it would still be desirable to keep
our word and to be trustworthy. An Epicurean could say that we keep our
word so as to ensure our own peace of mind, and that reflection on divine
tranquility is helpful to us in doing so. But by the truth of such Epicurean
theology good faith understood as a part of justice in fact overseen by a god
would be taken away, because there is no divine overseer.
Still more dramatic is Cicero’s next claim, that “the community of the
human race” (societas generis humani) might also be taken away. Perhaps in
part this is just a consequence of the loss of good faith: if the desire to act in
good faith is not really part of a relationship between ourselves and the
gods, then our community is not on the footing that Roman tradition
suggested. But perhaps Cicero is driving at a bigger point. We can see
this point against some background in his thoughts about societas we find
in his dialogue the Laws.
In the first book of the Laws, there is much discussion of what Marcus
calls societas, “community.” Marcus argues that the community of the

Compare Cicero’s argument at Laws . that the opinion that rational gods look after us or
threaten punishment is useful: it strengthens oaths or treaties, discourages crime and leaves “citizens’
community with one another holy,” sancta. . . societas civium inter ipsos. This argument, that what
people think about the gods matters, is distinct from the discussion of the consequences of what
might be true of the gods in the preface to DND.

I must be cautious in how I use the Laws to interpret DND. Laws is incomplete as we have it. When
Cicero wrote it and whether he completed it is uncertain. How far a skeptic like Cicero endorses
what his character says in any dialogue or would wish it to be read into other works is never clear. In
that regard the Laws is a particular challenge because it contains no obvious acknowledgment of
Cicero’s skepticism. Indeed it can seem rather dogmatic (see especially Laws .), although I agree
with the consensus that it is not so (see Görler , Atkins  pp. –). But the Laws
contains (in Book ) an idealized version of Roman religious law and (in Book ) a discussion of the
fundamental nature of law. Both are given by Cicero’s own character. I think it is fair to use these
books as background for the sort of issues Cicero might have in mind in writing the general remarks
in the preface to DND.

Laws . , , , , , , , , . societas is most literally “companionship” or “society.”
But in translating Stoic political ideas, Cicero seems to use societas for the Greek κοινωνία (see
. Theological Facts and Conventional Piety 
human race is based on our shared law. For him this law is simply our
reason that (if we all used it rightly) would lead us all in the same way.
(Laws .) But Marcus thinks that the gods, too, have reason. So he
thinks that they, too, are part of our community of the rational. (Laws
.–) Now members of this community, he argues, should have a
natural desire to be just to the other members of the community. By this
he means that we do not desire justice for some further goal, like assuring a
safe society for our own needs. Rather, we naturally want to be just to
others for their own sake. If we were to think that this was not a natural
desire, but rather that we desire to be just with (e.g.) our own Epicurean
advantage in mind as a further end, Marcus argues this would undermine
not only real justice to other humans, but also justice from us to the gods.
(Laws .) Now one attraction that rational virtue in this community
holds for us is that we can see it perfected in divine nature – we want to
show ourselves worthy of what we have in common with the divine. But
suppose, as Cicero contemplates in the preface to DND, there are no gods
in our community, because no gods care about us. That might make it less
attractive to be in a community with our fellow human beings where we
desire to help others for their own sake and not for our own. For perhaps
thinking of gods who are not active parts of the community would lead us
to imitate the Epicureans and leave the community, as he suggests the
Epicureans in fact did. (DND .) Such is the climactic criticism that
Cotta will bring against Velleius at the end of DND Book . (DND
.–, see pp. – below.) Note that this need not mean that
the Epicureans literally wall themselves up in their gardens. They might
carry on participating in what looks like the old community, but now for
selfish reasons, so that their society is not the web of mutual concern that it
used to be.
The threat from Epicurean theology that Cicero paints, then, is not that
life in Italy might descend into chaos in an obvious, outward sense. Instead
it is that society, religion, and their attendant virtues (as those things were
conventionally understood) would be hollow fictions if the Epicureans
were right. This means that, for a thinking Roman who realized that this
was the case, confusion about why such apparent goods are good at all
would follow. In order to be true, beliefs about such values would have to
be refounded, perhaps on Epicurean or similar principles.
Wynne () n). Given Cicero’s emphasis on the source of societas in shared reason,
“community” seems apt.
 Cicero’s Project in On the Nature of the Gods and On Divination
. How Philosophical Inquiry Can Moderate Religion

We have seen that from Cicero’s philosophical point of view, it is import-
ant that the facts about the gods can vitiate or underwrite accepted beliefs
about religion. But it is our beliefs, not how the gods are, that our
philosophizing can change. So it ought to be that philosophical inquiry
moderates religion by changing what we think. How should it change
what we think, and why should this amount to moderation for the better?
One way for Cicero’s inquiry to moderate religion for the better would
be for it to change its audience’s views to fit the facts. Cicero does not
discuss this sort of moderation of religion in his preface. But, as we shall
see in my next two chapters, in the rest of DND, his Epicurean and Stoic
speakers do so. Velleius the Epicurean holds that the Stoics overwork their
god, and “impose an eternal slave-master on our necks, whom we must
fear through all the nights and days.” (DND .) By contrast, he says,
Epicurus showed how the gods could lead a tranquil life, without work in
keeping with their happiness, so that we should have no fear of them. His
view of the gods brings it about ut deos pie coleremus et ut superstitione
liberaremur, “that we give cult to the gods piously,” acknowledging their
true happiness, “and that we are liberated from superstition,” because we
do not fear their anger (DND .). Meanwhile the Stoic Balbus (who, we
may assume, regards Velleius’ position as impious) says that Stoic theology
leads one to rationalize and reinterpret the traditional, poetic understand-
ing of the gods, which is in some respects a false and distorted understand-
ing. This is, says Balbus, a distinction between religion (religio), practiced
by those who have done such reinterpretation, and superstition (super-
stitio), practiced by those who have not (DND .). He says that the
etymology of superstitio is that Roman ancestors applied it to people who
constantly sacrificed and prayed so that their children might be their
survivors (superstites). (DND .).
Both Epicureans and Stoics, then, are concerned to avoid both impiety
and superstition. For Stoics, right religion is, so to speak, between impiety
and superstition. Impious religious practice is done in the belief that the
gods care for us less than they do, while superstitious practice is done in the
belief that the gods care more than they do. For the Epicureans, supersti-
tion and impiety come together. They say that a view of the gods like the
Stoic one will lead to superstitious practice, since a Stoic believes that the
gods notice and care about what she does, and impious practice, since for

imposuistis in cervicibus nostris sempiternum dominum, quem dies et noctes timeremus.
. How Philosophical Inquiry Can Moderate Religion 
the gods to care about and provide for us is inconsistent with their divine
happiness.
Although he does not address it in the preface to DND, in the preface to
the first book of Div. Cicero himself will encapsulate the consequences of
reaching the wrong sort of view about the nature of the gods:
nam cum omnibus in rebus temeritas in adsentiendo errorque turpis est, tum in
eo loco maxime, in quo iudicandum est, quantum auspiciis rebusque divinis
religionique tribuamus; est enim periculum, ne aut neglectis iis impia fraude
aut susceptis anili superstitione obligemur.
For error and rashness in assent is vicious in any matter, but it is especially
so on that question where we must judge how much credit to give to
auspices, to divinity, and to religion. For there is a danger that we shall be
involved either in an impious fraud (if we neglect these matters) or in the
superstition of an old woman (if we accept them). (Div. .)
The dangers here are an impious fraud (inpia fraus) or an old woman’s
superstition (anilis superstitio). I suggest that an impious fraud is what a
Roman would be involved in when she practiced religion if the Stoics’
answer to the Central Question is true but she had adopted the Epicurean
answer. For then she would do the outward actions of a pious person, but
fraudulently because she would in fact do them in vicious error about the
gods who had done so much for her and who saw what she was doing.
Meanwhile, an old woman’s superstition is what a Roman would be
involved in when he practiced religion if the Epicurean answer to the
Central Question is true but he had adopted the Stoic answer. For he
would suppose that the gods saw and cared about what he was doing, and
that they could affect his life, when they cannot.
In DND we have a similar picture of the relationship between piety and
superstition from the skeptic Cotta in his reply to Velleius. Velleius has
advertised Epicureanism as a liberation from superstition, but Cotta shoots
back:
nam superstitione, quod gloriari soletis, facile est liberare, cum sustuleris omnem
vim deorum. nisi forte Diagoram aut Theodorum, qui omnino deos esse
negabant, censes superstitiosos esse potuisse; ego ne Protagoram quidem, cui
neutrum licuerit, nec esse deos nec non esse. horum enim sententiae omnium
non modo superstitionem tollunt, in qua inest timor inanis deorum, sed etiam
religionem, quae deorum cultu pio continetur.
For liberation from superstition (in which you Epicureans are accustomed
to glory) is easy when you have taken away all the power of the gods. Unless
perhaps you think that Diagoras or Theodorus, who deny that there are
gods at all, could have been superstitious. I don’t think even Protagoras
 Cicero’s Project in On the Nature of the Gods and On Divination
could have been superstitious, who allowed neither claim, neither that there
are gods, nor that there are not. For all these mens’ views destroy not only
superstition, in which there is empty fear of the gods, but also religion,
which is comprised by pious cult of the gods. (DND .)
Cotta’s criticism uses the conceptual map of pious religion between

superstition and impiety. Velleius boasts of freeing people from supersti-
tion. But Cotta asks if this is to be counted a success when Velleius has
done it at the cost of moving people from superstition right through pious
religion and out into impiety. By robbing the gods of their power,
Epicurus has left a view of them equivalent in its implications for religion
to those of thinkers who deny or doubt the gods’ existence.
We now see one way in which philosophical inquiry could moderate
religion in the sense of outward practice: discovery of the truth. It could
give the inquirer and his audience true beliefs (or, hopefully, even know-
ledge) about the nature of the gods. For example, if the Epicureans were
right, then the Romans until Cicero’s day had been making religiously
correct performances as parts of susperstitious and impious actions,
because they believed that the gods cared. If philosophical inquiry then
discovered the truth, the Romans might drop that superstitious and
impious belief and start to do pious religious actions.
But Cicero’s radical skepticism means that this cannot be the only kind
of moderation he invites us to consider. For he expects that his inquiry will
not lead to discovery of truths about the gods. He hopes that it will lead to
hesitation and suspension of judgment on the matter. If so, how will the
inquiry moderate religion?
Before I answer that question, I should address a problem for my
skeptical reading of Cicero’s preface. For in paragraph [i] Cicero said that
unless the issue before his inquiry is decided (diiudicatur) humanity will
live in error or ignorance (error, ignoratio). This might suggest that Cicero
would in fact be dissatisfied with a skeptical outcome of the inquiry. For it
suggests that he thinks ignorance a bad outcome. A skeptic claims to live
without knowledge and thus, in a sense, to live in ignorance.
But we find elsewhere in Cicero’s philosophical dialogues another sense
of ignorance: ignorance not simply as lack of knowledge, but also as false
opinion. For example, consider this passage from On ends where another
Epicurean, Torquatus, says in support of the use of natural science:
omnium autem rerum natura cognita levamur superstitione, liberamur
mortis metu, non conturbamur ignoratione rerum, e qua ipsa horribiles
existunt saepe formidines.
. How Philosophical Inquiry Can Moderate Religion 
By knowing the nature of all things we are freed from superstition and
liberated from the fear of death. We are not thrown into confusion by
ignorance (ignoratio) and by the chilling fear that often results from ignor-
ance alone. (On ends ., translation from Annas and Woolf ())
Again, terrors do not emerge from sheer lack of knowledge. Rather they
emerge from the false beliefs that the gods might harm us and that we
might suffer after death. So both of these Epicurean speakers use “ignor-
ance” to mean, so to speak, positive ignorance: false beliefs. So I suggest
that Cicero, too, uses ignoratio in [i] in the sense of positive ignorance.
False belief is something that a skeptic would seek to avoid. So Cicero as
a skeptic could agree with the dogmatic philosophers that ignorance in this
sense is undesirable.
I think Cicero does not, in the end, endorse the conditional that, unless
the question of the gods’ care for us is finally decided, we must live in error
and positive ignorance. For he proposes another option: if the case appears
undecided, we could withhold assent from any belief, and thereby live
without false beliefs. This is the sort of outcome he expects from the
inquiry in DND . (quoted p.  above). In Div. . he suggests that
rash assent leading to error is what leads both to impiety and to supersti-
tion. A skeptical version of the moderation of religion, then, would be
suspension of judgment on the question of the nature of the gods. If
impiety and superstition are false beliefs, then by suspending judgment
Cicero, and others led to suspension by the arguments exhibited in DND
and Div., can avoid both.
This skeptical moderation of religion might look second best. Would it
not be preferable to achieve dogmatic but true beliefs, or even knowledge,
about the nature of the gods? Cicero would probably accept that in
principle such discoveries would be preferable, since in his Academica even
the most radical skeptic wishes to discover the truth. (Academica .–)
But Cicero would argue that in the absence of any clear evidence one way
or another, suspension of judgment is better than rash assent. It was not
just mistaken assent about religion that he called an especially foul error in
Div. . (p.  above), but rather rash assent, which might be assent to the
truth. Further, even after suspension of judgment one might find oneself
with views about the gods, views that do not amount to beliefs about what

It is possible that with error and ignoratio Cicero gives respectively Stoic and Epicurean descriptions
of false opinion in the matter of the gods. If so, his point in my paragraph [i] of DND .– (p. 
above) might be that the dogmatic parties to the dispute about the gods’ involvement would say
between them that until the dispute is settled people will live either in error or in ignorance. They
would omit another option, suspension of judgment.
 Cicero’s Project in On the Nature of the Gods and On Divination
is true. For as a radical skeptic Cicero would not accept the truth of any
claim, but instead he may follow what seems to him for the present
plausible (probabile) or like the truth (veri simile) (see pp. – above).
Thus Cicero himself may moderate his religious actions not only by
suspending judgment about the truth of the nature of the gods but also
by engaging in carefully balanced inquiry so that their nature seems this
way or that to him on some basis that seems responsible. He may encour-
age his readers to form their own free judgments of this sort. In Chapter 
I shall investigate what, if any, of his “own” judgments he indicates in our
dialogues. But despite these weak and ephemeral judgments of what seems
like the truth, the crucial aspect of skeptical moderation is that Cicero will
not accept as true any beliefs about an action that, if he accepted them as
true, would leave the action impious or superstitious. Perhaps his resulting
action is not pious, or is less pious than it could be, in principle, if he knew
the truth. But if knowledge or warranted beliefs about the gods’ true
nature are beyond him, he thinks it is better to avoid than to risk impiety
and superstition.
I can now sum up the project that, I have argued, Cicero has in mind
when he looks for moderation of religion from philosophical inquiry into
the nature of the gods. modero, to “moderate,” means to impose modi on
something. A modus is a measure or a bound. Philosophical inquiry may
impose two sorts of bound on Roman religion, that is, on Roman religious
practice: that it be neither superstitious nor impious. It will not do this by
bringing about any substantial reform of outward religious performance.
Rather, it will make each religiously required performance an action
neither superstitious nor impious by leading the religious actor to lack
the false beliefs about the gods which would make the performance super-
stitious or impious. The inquiry will achieve this either by discovering the
truth about the nature of the gods or, as Cicero expects, by leading us to
suspend judgment about the truth of their nature. The Central Question
about the gods’ nature for this purpose is whether they care about and
intervene in human life. For if a religious agent believes falsely that they do
care and intervene then she acts superstitiously, but if she believes falsely
that they do not care or intervene then she acts impiously.

See Lewis and Short (), OLD, TLL s.v.

Moatti () – explores in DND and Div. how Cicero thinks philosophy can save us from
superstition, but puts less emphasis on how it can save us from impiety. Santangelo ()  and
– concentrates on Cicero’s ambition in Div. to relieve religio of superstitio. He gives helpful
detail on the meanings of the term superstitio.
. Cicero’s Project in its Intellectual Context 
. Cicero’s Project in its Intellectual Context

You might ask: why would Cicero choose precisely that project for these
dialogues? One answer, it seems to me, is that the various theological
teachings of Hellenistic philosophers did indeed suggest to an attentive
student that the question of the gods’ involvement in our lives was a crucial
issue. But as I shall now point out, Cicero’s project also responds to a
further set of questions we can trace in Cicero’s own corpus and in the
(often scanty) monuments of the intellectual society of his day.
“We were strangers and lost in our own city until your books played the
role of hosts, leading us home so we could at last recognize ourselves and
where we were.” (Academica .) So says Marcus to his interlocutor
Varro in the first book of Cicero’s final draft of the Academica. He seems to
refer mainly to Varro’s Antiquities human and divine and to include that
huge work’s antiquarian investigation of Roman religion: “you have
opened up for us . . . the laws governing our rites and priesthoods . . .
the titles, classes, duties, and origins of everything human and divine.”
(.) Marcus’ reason for admiring these books is important background for
the project of DND and Div. Let us see why this is so.

Rawson () surveys the evidence for the lively intellectual efforts of the first century BC at
Rome. Moatti () argues from the whole of this evidence for a general conclusion about the
motivations for these efforts in some ways similar to the one I have attributed to Cicero for DND
and Div., namely that, in the face of the long political chaos and series of civil wars through which
they lived, people had a sense of bewilderment at, and loss of understanding of, their own
institutions, of the sort that Cicero tries to repair by importing Greek philosophy. “To defend
that tradition [i.e. that from which Rome had gradually emerged], the ‘last of the Republicans’ had
sought, not to refound the state [as Augustus would], but to enable it to endure, to safeguard its
stability and conservation, even if that involved finding new bases for it” (p. , emphasis added).

Scholarship on Varro has tended to agree with this characterization of the purpose of his antiquarian
work. “Religious antiquities, Varro stresses, must be preserved, since they are by common consent
threatened by the negligence of the citizenry . . . If the religious edifice on which the res publica
depended were weakened, would not the whole structure of society begin to collapse? Indeed, much
of the work of the Roman antiquarians has been seen rightly, as a literature of crisis trying to place
against what the antiquarian sees as a chaotic and dangerous world . . . an idealized picture of the
way things were and should be again in order to correct that process of deleterious change.” (Tarver
() ).

nam nos in nostra urbe peregrinantis errantisque tamquam hospites tui libri quasi domum deduxerunt,
ut possemus aliquando qui et ubi essemus agnoscere. tu . . . sacrorum iura tu sacerdotum, tu sedum
regionum locorum tu omnium divinarum humanarumque rerum nomina genera officia causas aperuisti.
Translations from Brittain ().

As I suggested above (p. ), although Marcus describes Varro’s remedy as though it helped Romans
at large, it is likely that only a small class would read his writing. In practice, Cicero’s philosophical
efforts were probably similarly for an elite audience, even if in principle they are not intended to be
esoteric (cf. p.  below). Brunt () concludes that, “It seems probable that the theological
doubts and contradictions of the philosophic schools had little effect on Roman religious practices,
or so far as concerns the mentality of most Romans, on the beliefs associated with them.” (p. )
 Cicero’s Project in On the Nature of the Gods and On Divination
The Marcus and Varro of the Academic books are, of course, fictionalized
characters speaking fictional lines. But the content of the real Antiquities
was much as Marcus describes it. Most of the work’s great length is lost,
but hefty reports and quotations survive, primarily in Christian sources.
To Varro, Christian sources are hostile sources, at least for the history of
religion. But with due care some part of Varro’s project and findings seems
to be recoverable. Augustine, for example, offers what he takes to be a
charitable interpretation of Varro’s procedure. The Antiquities human and
divine, says Augustine, came in two distinct parts, about human and divine
matters respectively. This was because Varro intended a work not about
the nature of gods and men at large, but about Rome. “Just as the painter
is prior to the painting, the builder to the building, so cities are prior to
what cities institute.” For Varro thought Rome, and not the gods, had
instituted Rome’s religion. Further, he seems to have reconstructed early
Roman religion as different in some respects from that of his own day. For
their city’s first  years, he said, Romans worshipped without names or
images for the gods. Jupiter (or the god later called such) he seems to
have regarded as akin to a monotheistic deity, set over the others, “no
different from the god of the Jews.” Later there were changes and
additions in which, for example, the temple of Jupiter was built on the
Capitoline. Varro expressed some dissatisfaction with what the founders
and reformers made: if he were founding a religion from scratch, he would
make it more in harmony with the principles of nature.
Varro’s researches were, at least in part, intended to recall his contem-
poraries to forgotten religious lore. Some gods, he said, were falling into
disuse, so that he would rescue them, like Aeneas who carried the Penates
out of burning Troy. To this end he cataloged all the gods he could find
On the other hand, it seems possible that Marcus’ diagnosis of alienation and bewilderment in the
face of the traditional religion was true for people even beyond those with the leisure and education
for intellectual pursuits, even if the philosophical help that he designs for them in DND or Book
 of his Laws would not, in practice, reach them. But were Romans at large picking up and
considering philosophical ideas in other ways? So far as I know, the evidence does not allow us to
answer this question.

sicut prior est . . . pictor quam tabula picta, prior faber quam aedificium, ita priores sunt civitates quam
ea, quae a civitatibus instituta sunt . . . . Fr.  Cardauns = Augustine, City of God ..

Fr.  Cardauns = Augustine, City of God ., ..

Fr.  Cardauns = Augustine, De consensu evangelistarum .., cf. fr  Cardauns = John Lydus,
De mensibus ..

Fr.  Cardauns = Augustine, City of God ..

Fr. a Cardauns = Augustine City of God ..
. Cicero’s Project in its Intellectual Context 
and suggested to which god the Romans should pay cult for which
purpose. He also tried to revive some neglected interpretations of Roman
practice. For example, when the temple of Jupiter was planned, he said, all
the gods who had altars on the site agreed by way of augury to have their
altars moved, except three. These were Mars (god of war), Juventas (the
goddess Youth), and Terminus (god of boundaries), whose primitive altars
were still to be found inside the precincts of Jupiter’s temple. This
signified, said Varro, that Rome would always be martial and young, and
that her borders would not yield.
Such is the treatise that Marcus in the Academica congratulates for
making Romans at home in their own city. Marcus seems to think, then,
that the Romans had lost something: the full understanding of their own
institutions and religion. This resulted, he suggests, in the alienation of
Romans from their own institutions, a problem that Marcus thinks the
Antiquities can help significantly to address by informing Romans about
their religion’s full meaning. The rules of the religion were not (for the
most part) revealed by the gods, but were rather the creation of states-
men. They could not be justified just by appeal to divine command, but
they could be understood through the history of their design and develop-
ment. It is that history which the antiquarian could try to recover. The goal
of Varro’s treatise thus understood has much in common with Cicero’s
own project in DND and Div. Both works equip, or re-equip, Romans
with a helpful understanding of their own religion, an understanding that
had been lacking or lost.
Nor, seemingly, were Cicero and Varro alone in this sort of project. In
her summary of intellectual theology at Rome in the Late Republic,
Elizabeth Rawson counts (for example) in addition to Cicero five other
Roman authors on augury from the first century BC, or two Roman works
on aspects of the pontifical college from the same period. In some cases
little of these authors’ work survives. But their number suggests a general
concern with historical or theoretical treatment of the religion. In Div.,

Fr.  Cardauns = Augustine, City of God, ..

Fr.  Cardauns = Augustine, City of God ..

The exception in the pages of DND and Div. is the Etruscan art of haruspicy, whose original
handbook was divulged by the supernatural man-boy Tages – or so says a rather satirical Marcus at
Div. ..

A recent co-ordination of the real Varro and Cicero’s projects is Rüpke () Chapters –.

Rawson () –. It is commonly accepted that Cicero wrote a work, now lost, On augury.
So far as I know, the positive evidence for the existence of this work are citations by the fourth
century AD grammarian Charisius in his Ars grammatica (pp. ,  Barwick). Cicero may allude
to a plan for such a work at Letters to his friends .., and perhaps Div. ..
 Cicero’s Project in On the Nature of the Gods and On Divination
Marcus mentions a controversy between the augurs Appius Claudius
Pulcher and Gaius Marcellus over whether augury was a real art, or was
invented for civic reasons. Several others contributed books on the topic,
the contents of which seem to have been “strongly antiquarian.” (Rawson
() )
We can glean a little more about another of Cicero’s intellectual
contemporaries, the “Pythagorean” speculator Nigidius Figulus, who fea-
tures in Cicero’s sketched Timaeus (section ). He wrote works On the gods
(in at least nineteen books), On entrails, On private augury, On dreams, and
he translated a book on lightning strikes, the latter four works being to do
with divination. So it is clear that Nigidius had a close interest in the
investigation of religion. He also wrote a work on language, the Philological
notebooks (Commentarii grammatici), in a passage of which he argued for
the pejorative connotation of the adjectival suffix -osus, quoting an old
poem, saying that to be religiosus was to have an immoderate and excessive
religiosity and thus to be superstitious, a character flaw which might lead
one into excessive religious performance. It is plausible to think that, as for
Cicero, to avoid such excesses (or perhaps their contrary defects) was part
of Nigidius’ motivation for his studies.
Thus in DND and Div. Cicero responded to an intellectual concern of
his times at Rome: to rediscover, or simply to find, the right way to
understand the performances of the traditional religion. We can see
Cicero’s approach more clearly if we look at the Laws and Academica as
background. As we have seen (p.  n.  above), we should not press
too hard to reconcile the Laws with other dialogues. Rather, we should use

Div. ., cf. Laws .–.

“Pythagorean” and physical speculation, Cicero, Timaeus ; On the gods, Macrobius, Saturnalia
..; On entrails, Gellius, Noctes Atticae ..; On private augury, Gellius, Noctes Atticae ..;
On dreams, brontoscopy, John Lydus, De ostentis, , .

We can find evidence of the sort of unease about religion that may have motivated the intellectuals
in Cicero’s works for wider audiences. Even in speeches Cicero can argue against some (notional?)
atheists in favor of providence both from Rome’s imperial success and from cosmological arguments
(Pro Cluentio ); that Sicily had fallen into superstition because of the false belief that the gods
were angry with the island and that the Romans should “heal” the Sicilian religion by removing the
ground of this belief (In Verrem ..–); (sarcastically) to the pontifices that Clodius, who
infiltrated the women-only rites of the Bona Dea, was motivated by anilis superstitio, “an old
woman’s superstition” (cf. Div. ., p.  above) that the gods wanted him to infiltrate the rites and
that he should be told by the pontifices to impose some modus, limit, on his religion (cf. DND .) –
presumably, by removing the belief which led to the offending performance (De domo sua ). So
Cicero thought that his contemporaries were alive to the idea that one’s beliefs about religion might
lead one astray, either into wrong performances or into wrong beliefs about those performances, and
to be open to intellectual argument about those beliefs. It would follow that intellectual argument
could lead one right or astray. See also Gildenhard () part III and the religious material
collected from speeches by R. J. Goar ().
. Cicero’s Project in its Intellectual Context 
it as background. But it is very useful background here, since in it Cicero
gives us an extended attempt to reconceive (mostly) traditional Roman
religion in line with a philosophical theory of politics. We saw that in Laws
Book , Marcus argues that we humans share reason with one another and
with the gods, and that this is the basis of our justice towards one another
(p.  above). A creator god, he says, gave us this reason. (Laws .–)
Our shared human nature is also reflected in our shared capacity for
superstition, which is cross-cultural even if its expression, like the worship
of animals in Egypt, sometimes is not. (Laws .) When Cicero comes to
his formulation of idealized Roman religious law in the second book of the
Laws, he aims to give a traditional rather than a new code. Changes to the
required performances are few. As it happens, they are often in line with
the project of promoting in DND virtuous religious actions, since what are
rejected are performances which are unusually incompatible with this
project: the worship of vices is abandoned (Laws .) and collections of
money for religious use are banned in part because they lead to superstition
(Laws .). Meanwhile, the traditional religious performances retained
are interpreted in the context of the theory of Laws Book . For example,
Marcus considers whether to retain temple buildings or, like the Persians,
to torch such shrines lest the gods be enclosed. He decides to keep the
buildings, but on the grounds that they encourage Romans to think of the
gods as neighbors, who have houses alongside their own, and therefore as
members of their own society. (Laws .–) A preamble to the
religious laws encourages the Romans to see the religion in the light of
true gods, whose reality can be apparent through the order of the cosmos
and arguments much like Balbus’ Stoic position in Book  of DND. (Laws
.–) So in the terms of Book  of the Laws, Marcus in Book  of the
Laws wants to use the religion to promote Romans’ natural, rational
capacity for justice towards the gods, and to inhibit their natural capacity
for superstition. The Laws, then, shows us how a philosophical theology
could give meaning to traditional Roman religion, and virtue to its
practitioners.
Now Marcus’ exposition in the Laws seems very confident. But of
course, given his skepticism, if Cicero had a view about the right way to
interpret traditional religion, it was probably not so confident. Sure

Through such interpretations, Marcus in the Laws is sympathetic to the visual metaphors of
traditional religious buildings and cult images (cf. Laws .). It is tempting to see in this an
answer to Varro’s apparent admiration of early Roman religion, which he thought was aniconic.
(Varro’s Antiquities frs. , , and  Cardauns.)
 Cicero’s Project in On the Nature of the Gods and On Divination
enough, in Academica, as Marcus talks his way through a catalog of
disagreements among dogmatists, we come to the example of natural
theology. The Stoics may argue, he says, that the cosmos is a god, that it
has a mind, and that it ordered and moves itself. But they will run into
Aristotle, who will argue that the order of the cosmos was never created.
They may argue that god created everything (besides himself, we assume)
and that he did so for our sakes, but they will run into the Peripatetic
physicist Strato of Lampsacus, who will argue that nature and not god
made everything, freeing god from such labors. “But I don’t assent to
Strato, or to you either,” Marcus sums up, “now one view seems more
plausible, now the other.” (Academica .–) So in Academica,
Marcus claims to suspend judgment on some of the theological claims
which underwrote the religious views of the Laws: are we really given
reason by a creator god, and do the gods really care about us in such a way
that we could meaningfully think of them as members of our society?
These are puzzles closely related to the Central Question of DND.
This background in Cicero’s corpus, then, gives us sharper focus on his
project in DND and Div. Varro, the author of the Antiquities as portrayed
in Academica , who chose not to write philosophy and instead to gather
new data, looked on the state religion as founded by statesmen and
changed over a long tradition. He offered his data as a way to recall the
Romans to the fullness of their orthopractic tradition, by reminding them
of its development and complexity, regrettable though some aspects of the
tradition might be from the point of view of natural philosophy. Cicero,
by contrast, is much less interested in the history of traditional rules for
performance. His approach is to accept the performances as given by
tradition, and then to supply from philosophical investigation intellec-
tually rigorous ways to interpret those performances, so that one may
render them pious, or at least so that one may avoid the false beliefs about
them which would make them impious or superstitious. These interpret-
ations may be entirely new. In Varro’s metaphor, Cicero’s goal is not
necessarily to reconstruct the meaning that the “painter” intended in the
“painting” of Roman religion, but rather to arrive at a philosophically
grounded interpretation of the picture. In this way Cicero has little room
for the distinction between philosophical and civic theology made in
Varro’s work. He suggests that the right way for the Romans to return
from alienation is that the legacy of those who devised the religion, namely
its required performances, should indeed be retained, but that it should be

nec Stratoni tamen adsentior nec vero tibi; modo hoc modo illud probabilius videtur.
. How the Project Shapes the Dialogues 
moderated from without by the philosophical views, or skeptical integrity,
of religious agents. Philosophical theology itself yields a civic benefit.
. How the Project Shapes the Dialogues

The last plank in my argument that Cicero pursues the project I have
suggested is a brief examination of the dialogues themselves. For it is not
just that the preface to DND piques the reader’s interest with the questions
I have examined. Rather, Cicero chooses and shapes the arguments por-
trayed throughout DND and Div. in order to pursue the debate on the
Central Question.
It is hard to know which Cicero thought were the “main” philosophical
views of the gods at Rome in the s BC. To us, perhaps, the Epicureans
and the Stoics seem the salient contenders. But that might be an accident,
the result of their portrayal in DND and the survival of Epicurean texts at
Herculaneum. Cicero knew a number of Antiocheans, including Varro the
author of the Antiquities and Brutus the addressee of DND, so their
theological views, where distinct from the Stoic view, could also have been
options for inclusion in DND (see p.  n. ). Indeed Cicero reminds the
reader of this when Marcus observes that a representative of Antiochus is
the only significant omission from the scene. (DND .) Meanwhile, in
Div. Quintus will reveal that he is most sympathetic to the Peripatetic view
of divination. (.) The Peripatetic Strato’s view was the foil for Stoic
theology in the Academica, and from the sketch of a dialogue on physics
which we call the Timaeus it appears that the Peripatetic Cratippus was
intended to contribute. In the Timaeus Nigidius was also to appear,

Feeney () – draws a similar contrast between Varro and Cicero’s respective projects.
Notice that by characterizing Varro’s project as antiquarian rather than philosophical, I do not
mean that it was merely antiquarian, or that the conclusions Varro reached about early Roman
religion were not informed by philosophical knowledge. I am inclined to agree with van Nuffelen
() against Moatti () or Momigliano () that Varro hoped to recover some truths about
philosophical subjects that informed early Roman religion. In this way his project has something in
common with, say, Balbus’ treatment of traditional religion in DND Book . Cicero, by contrast,
seeks to bring new philosophical insight to bear from the outside, so that it would be an additional
bit of luck, rather than the goal of his project, if it turned out that some of this insight were already
encoded in parts of the tradition. Rüpke () contemplates the historicizing nature of Varro’s
project. Wallace-Hadrill () – puts these questions into the wider context of his
“cultural revolution” in late Republican Rome.

Cf. Furley (). But Balbus’ point has to do with Antiochus’ claim, explored at length in On ends
Books –, that the Stoic and Peripatetic theories of value differed only verbally. This is not directly
relevant to the subject matter of DND. So Balbus’ remark is dramatically appropriate, but does not,
I think, give us Cicero’s reason for omitting an Antiochean character from DND.

Cicero seems to find Aristotle’s own theology elusive, blurred perhaps by Aristotle’s dialogues. See
DND ., ., but cf. the more familiarly Aristotelian view of Academica ..
 Cicero’s Project in On the Nature of the Gods and On Divination
described as a Pythagorean, and probably to deliver the translated excerpt
from Plato’s Timaeus that makes up the rest of that text. That excerpt
contains plenty of theology. So even if there had not been an obvious
alternative to Stoic and Epicurean theology in the views of Antiochus,
nevertheless Peripatetic, Platonic, and “Pythagorean” theologies were also
current for Cicero. Of the dogmatic options available to him, why did
Cicero choose to examine the Epicurean and Stoic views, and only those
views, in detail?
If you took the “encyclopedia” view of Cicero’s later philosophica
(pp. –), you might insist that, since Cicero’s dominant aim was to
provide a philosophical encyclopedia, Cicero presumably did choose the
Stoic and Epicurean views just because they struck him as the main views
that readers needed to know about. But my position is that although
Cicero had some thought for an encyclopedia, he was more interested in
shaping his dialogues as works each with unity and an aim. My position
can make good sense of his choice of the Epicureans and the Stoics. For
consider that the Central Question of Cicero’s project, as I interpret it, is
whether the gods care about us and act in our lives. But the Stoics and
Epicureans are polar and rich opposites on this question.
The two schools are polar opposites in that the Stoics hold that god fated
every last detail about our world, to include everything about us, and did
so for our benefit, while the Epicureans hold that the gods have never had
anything to do with us or our world, cannot do anything to or for us, and
would not if they could.
Cicero could have got a similar polar contrast from other schools. In
particular, we already have seen that in the Academica he drew the same
contrast but with Aristotle or Strato representing the view that no divine
mind created or acts in the world. But, by opposing the Epicureans in
particular to the Stoics, Cicero also achieves a particularly rich contrast.
Both Stoics and Epicureans see the gods as like us in important respects.
The Stoics see gods and humans in a community of the rational with
duties to one another. The Epicureans see the gods’ perfect happiness as

See Furley () for another discussion of Cicero’s omission of Peripatetic views.

For example, Brunt ()  concludes from Cicero’s choices that Stoic, Epicurean, and
Academic theologies “had the most currency in contemporary Rome.” Brunt nevertheless sees
some of the polar contrast to be drawn between the Epicureans and Stoics, but characterizes Balbus’
Stoic gods as more rationalized, and less personally caring, than I do (pp. –).

Bénatouïl () shows how thoroughly, and controversially, the Stoics made their god fate every
last detail of the world.

Academica .. In DND . Velleius interprets Strato’s position as attributing a sort of divinity
to nature, but that does not seem to be part of Marcus’ interpretation in Academica Book .
. How the Project Shapes the Dialogues 
implying precisely that the gods are not part of any community with us,
and rather that they are above any involvement in the business of the
cosmos or of our lives, so that their life of pleasure is assured. Further, both
schools think we should imitate the gods in these respects so far as we can. For
Cicero’s Epicureans, our community with one another is regrettable,
merely a remedy for our weakness. Thus these two answers to Cicero’s
question lead to very different visions not merely of right religion, but also
of happy and virtuous human life in general. By way of comparison, the
Aristotelian god also does not care for us. But Aristotelians do not draw
the conclusion that we should all seek to imitate god in that respect. The
choice of exactly the two schools he chooses, then, makes most clear what
is at stake in Cicero’s Central Question.
Moving to Div., Cicero there includes speeches only for a specific Stoic
view of divination (Quintus’ speech in Book ) and Academic arguments
targeted at that view of divination in particular (Marcus’ speech in Book ).
The Epicureans rejected divination entirely, and thus could have spoken in
opposition to the Stoics, or could have been given their own speech and
counter-argument. But in Div., unlike in DND, Epicurus is summarily
dismissed, “babbling about the nature of the gods.” (balbutientem de natura
deorum, Div. .) Why? Cicero does not seem to have rated Epicurean
arguments against the Stoics very high: unless Carneades had come along,
Marcus says at the end of Div., the Stoics might have been judged “the only
philosophers” by Marcus’ day. (soli. . . philosophi, Div. .) But I suspect
there is more to it than that. The Epicurean view of divination was purely
negative. The positive Epicurean theology and physics on which the nega-
tive view rested was covered by DND (and would be expanded on in On
fate, if Cicero got around to it). Now if Cicero’s main concern were to be
encyclopedic he might still have included the Epicurean arguments and a
reply. But if I am right, he left them out because there was no positive and
fertile Epicurean view to engage with. A Stoic argument for divination is
therefore picked out as the richest representative of the view that the gods

The only other school which gets some airing in Div. – the Peripatetics, who held that “natural”
divination is real but “artificial” divination is not, a view to which Quintus himself in fact
subscribes – could have had its own treatment (see Div. ., ., .). But as Quintus and
Marcus present it, the Peripatetic view is assimilable for dialectical purposes to that part of the Stoic
view which argues that natural divination is real. The Peripatetic view is therefore covered efficiently
as part of the general positive and rich view that divination is real. Meanwhile, according to Cicero,
Panaetius had distinguished himself among Stoics by his doubts about divination at large, and
rejection of some aspects of it, but Cicero gives his arguments little space in Div. (See Academica
., Div. ., ., ., .. DL . says that Panaetius simply denied the reality of
divination, but Cicero seems likely to be more accurate in this regard.)
 Cicero’s Project in On the Nature of the Gods and On Divination
give us helpful information, a view which implies a positive answer to the
Central Question of DND.
The choice of schools and speakers is not the only element of the
structure of the dialogues shaped by the Central Question and its vital
role in the moderation of religion. In the chapters ahead, I shall argue that
Cicero the author often reminds the reader of the Question at significant
junctures. Now, so far as the speakers of DND are aware, their topic of
discussion is the nature of the gods in general, and neither the Central
Question nor religion specifically. (DND .) Yet Cicero shapes their
conversation so that the Central Question and religion receive emphasis.
Both Velleius’ speech and Cotta’s response to Velleius culminate in
drawing the consequences for religion from the Epicurean answer to the
Central Question. (DND .–, .–, see Chapter  section .)
Cicero has Cotta specifically request that Balbus give us the Stoic argu-
ments that the gods run the world and care about us, that is to say, his
answer to the Central Question. (DND .) Balbus does so at enormous
length. (DND .–) But by a dramatic trick Cicero also emphasizes
that part of Balbus’ speech that deals with religion, and makes this a key
bone of contention in Cotta’s response (see below, pp. –). Mean-
while, Div. is set in train by a conversation about just these disagreements
in DND, and features a conclusion that revisits the Central Question.
(Div. .–, cf. Chapter  section .) In any case, Cicero’s choice of
divination as a topic for further elaboration is well explained by the Central
inquiry. As Quintus points out in opening the conversation of Div., if the
gods give us divination, then it follows that they do indeed care for us.
(Div. .)
 
Velleius the Epicurean
In this chapter, I examine Cicero’s first step in the project I outlined in the
previous chapter. In sections . and . I interpret the Epicurean answer
to the question of the nature of the gods, given by Gaius Velleius in his
speech in the first book of DND. In section ., I follow the consequences
for Roman religion that Cicero draws from Velleius’ theology, through
both Velleius’ speech and that of his skeptical opponent, Cotta.
For an Epicurean like Velleius, good theology stems from the premise
that the gods are happy and eternal. But a grave obstacle to good theology is
the human psyche: we are naturally acquainted with the happy and eternal
gods, yet we are prone to draw the wrong, and troubling, conclusions from
this acquaintance. We are fortunate, then, to live after Epicurus, the man
who grasped the true implications of divine happiness. The centrality of
happiness in Velleius’ speech is such that Cotta, his opponent, coins two
words for it: beatitas or beatitudo (DND .). Epicurean beatitude is the
absence of pain and anxiety, a passive and self-sufficient quietness. With
such happiness in mind, through Velleius and Cotta, Cicero offers to
careful readers a vision of the traditional religion performed by happy
Romans in honor of happy gods. A key property of these gods is that,
being passive and self-sufficient, they do not care for us at all. Velleius’
answer to the Central Question is “no” (see p. ). For that very reason, he
thinks the Romans should worship the gods as an ideal.
I must first mention a related controversy. Epicureans like Velleius
believed that the gods exist. So when an Epicurean made a religious
performance, she made it for an ideal, but for one who, she thought,

At one point, Cotta agrees with Posidonius in doubting the sincerity of Epicurus’ theism. But this
comes in the rhetorical climax of his speech, when he thinks he has shown that Epicurus’ doctrines
are laughable and reprehensible. (DND .) The direct textual evidence of Epicurus’ theism is
strong and univocal, see especially Letter to Menoeceus . Indeed in a calmer moment Cotta himself
admits that the much more plentiful textual evidence available in his day was similarly unambiguous.
(DND .) On accusations like Cotta’s see Obbink ().

 Velleius the Epicurean
nevertheless existed. Today there is controversy over how Epicurus’ gods
both exist and yet act in our lives only as our ideals. Perhaps Epicurus
thought that the gods exist only as our ideals, as the conglomerations of
atoms that impact our minds when we imagine the gods. Or perhaps he
thought that the gods have eternal bodies of human shape but are so
distant, or are made of such evanescent matter, that they could have no
direct effect on us even if they wished to. Partisans on both sides of this
modern debate find some of their evidence in DND Book . But in order
to understand Velleius on the Central Question and religion, it is not
necessary for us to settle his place in the debate over how Epicurus’ gods
exist. What matters is the view he attributes to Epicurus, that the gods
exist but act for us only as our ideals. That is so on either modern view of the
existence of Epicurus’ gods. Thus I do not choose between those views in
this chapter.
I should also mention another important question that is not my focus
in this chapter. Thanks to the resemblances of Velleius’ speech to the
Herculaneum On piety, Book  of DND is often studied with its sources in
mind (see pp. –) Even scholars who give Cicero agency in this book’s
composition tend to see it as their task to detect, and then to explain,
Cicero’s use of and supplements to his sources. But my intention here is
to study on its own terms the product of whatever process of writing
Cicero went through.
When we come to Velleius’ speech, we have just read Cicero’s preface. If
as learned readers we recognized in the preface that the party who
answered that the gods do not care for us were Epicureans, we know
now that we have one of them before us. If we are unlearned and did
not recognize this, then Cicero has Velleius enlighten us straightaway.
Thus, guided by Cicero’s preface, what we expect is an intellectual bomb-
thrower whose views would do away with traditional piety, or with society,
and even with justice itself. Velleius is indeed a combative and confident
speaker. But in other ways Cicero allows Velleius to confound our negative
expectations. Velleius tries to keep his speech short and, although he does
not think so (DND .), he succeeds. Cotta praises the speech as an
unusually clear and stylish exposition of a difficult subject. (DND .)
I think Cicero means us to agree with Cotta. Most of Velleius’ speech is a
single, well-organized, and concise argument for the Epicurean position on

For this debate, see Konstan () and Sedley () with their references. Some older treatments
of Epicurean theology are Festugière () and Lemke ().

See for example Obbink (), Essler ().
. Velleius’ Theology 
the Central Question. It depends on a central argument lasting just two
sections. (DND .–) This argument is followed by an exposition of
some of its satisfying consequences (DND .–). It is preceded by a
catalog of its deadly implications for rival theologies. (DND .–) In
my interpretation I shall switch the order so that I treat the main argument
and its theological consequences first (section .) and the catalog second
(section .).
Although Velleius is unaware that he is part of a dialogue about the
Central Question and its implications of Roman religion, Cicero shapes
his speech with that project in mind. Further, Cotta’s reply to Velleius
helps to bring out those religious implications in full. Thus in section .
I put together Velleius’ and Cotta’s speeches to see how Cicero suggests
that the Epicureans would moderate Roman religion.
. Velleius’ Theology

Velleius’ short main argument proceeds from the nature of the human
psyche to the existence, eternity, and happiness of the gods. It draws on the
department of Epicurean thought called “canonics” or, roughly speaking,
what we call epistemology. (DND .) The “canon” in question is the set
of true criteria by which we ought to judge whether further claims are true.
Two parts of this canon rely on the Epicurean dogma that all sensations are
true. To take the sense of sight as our example, Epicurus thought that we
see thanks to skeins of atoms that are cast off by visible bodies, and that are
images of the bodies that cast them off. When an image impacts in our
eyes, we see the body from which it came. A stream of images from one
body allows us to see it over time. Furthermore, said Epicurus, when
I visualize, then too my imagination or “mind’s eye” is impacted by
images. For example, when I call to mind an absent friend, I open up
my imagination to skeins of atoms that have flown from his body to my
mine. This view allowed Epicurus to take an optimistic view of our
cognitive abilities both in sensation and in imagination. If I see a straight
stick that looks bent in water, then that sensation is true, in that it results
from the impact on my eye of a real image of the stick, an image that has
indeed been “bent” by passing through water. Meanwhile, if I hallucinate a
griffin, that hallucination is true, in that an image of a griffin is indeed
hitting my imagination. (LS )
Of course, Epicurus did not think we are free from error. We introduce
error in our beliefs about our sensations or imaginings. If, in the examples
just given, I judge that “that stick is bent” or that “there is a griffin over
 Velleius the Epicurean
there,” then I introduce error where before there was none. If I were
skillful in my judgments, I would be able to judge from the true images
that “that stick is straight, but appears bent due to refraction,” or that,
“this image of a griffin has struck my imagination, and not my eye.” But
sometimes I will fail at these judgments, and draw false conclusions from
the true sensations: that the stick is bent, or that I see a griffin over there.
(LS ) Sensations, then, are always true, and function as criteria for
judging further truths, if only we would use them well.
Criteria of a second sort are preconceptions (πρόληψις, prolēpsis). After
many sensations of some kind of thing – say, of trees – I naturally form a
preconception of trees. The preconception is an accurate idea of a tree that
comes to mind when I come across the word “tree,” giving the word
meaning, so that I can talk about trees and, if I want to, investigate them
scientifically. Although the preconception is accurate, we are fallible, so the
accuracy of our preconceptions does not guarantee that our investigations
will go well. In the case of the gods our preconception evidently has at least
a “visual” element: Greeks and Romans imagined the gods with human
bodies. But this “visual” element does not mean that we ever see gods with
our eyes. For of course, unlike trees, we never see gods around us. Our
preconception of the gods is thus formed only by the images of the gods in
the imagination.
It is the doctrine that natural preconceptions are accurate on which
Velleius relies for his main theological argument. The first part of this
argument establishes that there are gods. Now this part of the argument is
important for understanding the basis of Velleius’ theology – his inference
from our preconceptions about the gods to the gods’ existence. This part of
the argument is thus very important for my interpretation of Velleius’
speech. But it has also aroused a good deal of puzzlement and confusion.
Therefore, since I hope to show that Velleius speaks and argues well,
I must lay these confusions to rest. Thus I shall comment on this part of
his argument in some detail. For ease of reference I give it here in full. For
ease of reference, I have made my own divisions of the argument,
numbered in bold with round brackets.
(i) ea qui consideret quam inconsulte ac temere dicantur, venerari Epi-
curum et in eorum ipsorum numero de quibus haec quaestio est habere debeat.

On Epicurean preconceptions in general, see DL ., Epicurus Letter to Herodotus –, LS
Chapter , Manuwald (), and Striker (b) –.

See pp. – below.
. Velleius’ Theology 
(ii) solus enim vidit primum esse deos, quod in omnium animis eorum
notionem inpressisset ipsa natura. quae est enim gens aut quod genus hominum
quod non habeat sine doctrina anticipationem quandam deorum, quam appel-
lat πρόληψιν Epicurus id est anteceptam animo rei quandam informationem,
sine qua nec intellegi quicquam nec quaeri nec disputari potest.
(iii) quoius rationis vim atque utilitatem ex illo caelesti Epicuri de regula
et iudicio volumine accepimus. quod igitur fundamentum huius quaestionis est,
id praeclare iactum videtis.
(iv) (a) cum enim non instituto aliquo aut more aut lege sit opinio
constituta maneatque ad unum omnium firma consensio, (b) intellegi necesse
est esse deos, quoniam insitas eorum vel potius innatas cognitiones habemus;
(v) de quo autem omnium natura consentit, id verum esse necesse est; esse
igitur deos confitendum est.
(i) Someone who considers how rashly and carelessly [other theologies]
are uttered, must revere Epicurus and count him among those about whom
we inquire [i.e. among the gods].
(ii) For he alone saw first that there are gods because nature itself has
impressed the idea of them in everybody’s soul. For what nation or kind of
human beings is there that does not have some ‘preconception’ of the gods,
without being taught it? Epicurus calls this a prolēpsis, that is, some
representation of an object preconceived by the soul, without which
nothing can be understood, or inquired into, or dicussed.
(iii) We were given the power and utility of this argument in that
Olympian book of Epicurus’, On the canon and criterion. Thus you see that
this foundation of our inquiry was admirably laid down.
(iv) For (a) since the opinion was brought about not by any insti-
tution, or custom, or law, and persisted unanimously as a firm consensus of
all people, (b) it is necessarily understood that there are gods (intellegi
necesse est esse deos), because we have implanted, or rather inborn, concep-
tions of them.
(v) But whatever everyone’s nature agrees on, that is necessarily true.
Thus we must acknowledge that there are gods. (DND .–)
Now I shall offer my interpretation of this controversial argument.
Velleius appeals to an alleged ethnographic fact, that people of all nations
have both a preconception of the gods, and a belief that the gods exist.
From this he argues that there are, indeed, gods. Even if we grant Velleius’
ethnographic premise, this sort of argument can look very suspicious. “If
we all believe it, then it is true” might once have proved that the Earth does
not move. But Velleius’ argument is not so simple. His point is that the
preconception of the gods is natural in humans. Nature has pressed the

Obbink () – questions Velleius’ use of consensus arguments and suggests that Cicero
himself rendered the Epicurean position this way, reproducing Stoic consensus arguments.
 Velleius the Epicurean
idea into our minds (ii), so that our common nature agrees about it (v). It
is natural concepts Velleius thinks are always right. Thus that we all have a
concept, or that that concept appears in all societies, is not supposed to be
direct evidence that that concept is accurate. Instead, it is direct evidence
that the concept is natural. After all, if there is no culture whose people
lack some concept, then that is evidence that the concept is cross-cultural
and the product not of nurture, but of nature (ii). Now, one might reply,
perhaps natural preconceptions are universal, but it does not follow that all
universal concepts are natural. But in fact Velleius does not seem to think
that universality on its own proves naturalness. Instead, in (iv)(a), he
suggests some further criteria. First, no cultural institution or custom
produced the belief that there are gods. Second, once it had arisen, this
belief persisted as a stable consensus, which suggests that it did not encoun-
ter resistance from some other, perhaps natural, concepts. So if we discover
a universal belief, Velleius allows that further historical and sociological
tests may be applied before we may declare the belief’s associated concepts
natural: it should not have arisen from mere custom, and it should have
remained a stable consensus. Finally, if we decide that a concept is indeed
natural, then we invoke further reasons to think it is right, namely, Epi-
curus’ optimism about our natural ability to perceive the world accurately,
represented in his theory of “canonics” mentioned in (ii).
But Velleius’ argument has offered further puzzles. A first puzzle has
been that the argument seems oddly repetitive. Velleius seems to say the
same thing three times in (ii), (iv), and (v): since there is a natural
consensus that the gods exist, the gods exist. That would be a dilatory
way for Velleius to speak when he wants to fit so much into his short
speech. Further, (iii) reads as though, far from repeating himself, Velleius
thinks he is developing an argument by stages. He says that (ii) and (iii)
are supposed to show us that Epicurus laid some foundation for what
follows. Thus it would be helpful to find some way in which (ii) gives us a
premise for (iv) and (v). A second puzzle is the following. The relation-
ship between the preconception of the gods and the belief that there are gods

Note that I do not here take a position on whether our preconception of the gods is “innate” (see
Konstan , Sedley ). Velleius’ claim that nature impresses it on our minds could mean
either that we have an innate disposition to form the preconception, or that nature brings it about
in all of us after birth.

Here I agree with Sedley () – against Obbink () –.

For these puzzles, see Sedley () –.

For complaints that the structure of the argument is informal, see Sedley () , or repetitive,
see Konstan () .
. Velleius’ Theology 
seems obscure. In (iv) and (v) Velleius seems to treat the two, and their
correctness, as interchangeable. Yet these are importantly different things,
and to substitute them for one another would seem to confuse Velleius’
position. For example, I might have an accurate preconception of the
dodo, but no belief that dodos exist. So might I not be similarly disposed
towards the gods? Since I think this argument is at the heart of Velleius’
speech, I will offer a detailed attempt to solve these puzzles.
I begin by taking the hint in (iii) and identifying a first premise in (ii).
This premise, I think, is that if I am to understand anything about (for
example) dogs, I must have a preconception of dogs that I call to mind
when I study or talk about them. Now I suppose in (ii) “understood”
(intellegi) could mean successful understanding, in the way that Einstein
understood General Relativity and I do not. If so, then (ii) already implies
that there are gods, and Velleius will indeed repeat himself. But it is better
to take “understand” to mean any understanding, in the sense that I must
have some understanding of what sort of thing General Relativity is before
I can talk about it at all. Thus in (ii) Velleius seems to be saying we need
a preconception to think or talk about something at all, not just to
think about it correctly. Taking “understand” in this way, let us draw
from (ii) this Velleius’ first premise:
V. If we understand something about x, we have a preconception of x.
Next, (iv)(a) says that, without any cultural influence and stably over
time, we share an opinion. This opinion is spelled out in (iv)(b): that there
are gods. Now in the underlined part of (iv)(b), “understood” (intellegi)
sometimes is taken simply to draw from (iv)(a) a conclusion we under-
stand, successfully, that there are gods. If this were so, then indeed
Velleius would repeat his conclusion in (iv) and (v). But it is better to
take “understood” in the same way as I took it in (ii): if we all believe that
there are gods, then we must understand that there are gods, in the sense
that we must have some idea what we are thinking about. If so, then in (iv)
(b) Velleius still has not concluded that there are gods. Rather, the purpose
of (iv) is to infer from our natural opinion that there are gods, to the
conclusion that we have a natural preconception of the gods.
Specifically, the next premise conveyed in (iv) relies on the universality
and non-cultural origin of the belief. If human beings come to the belief
that there are gods, everywhere and without cultural prompting, then we
can see that,

Cf. Striker (b) –.
 Velleius the Epicurean
V. We naturally understand that there are gods.
But if this is so, we must have had the preconception before the belief,
everywhere and without cultural prompting. That is to say, given premises
V and V, it follows that, as Velleius says in (iv)(b),
V. We naturally have a preconception of the gods.
We still have not concluded that there are gods.
In (v) Velleius proceeds to his conclusion. Velleius helps himself to a
given of Epicurean canonics, that,
V. Whatever we all naturally agree about is true.
Then he concludes that, “we must acknowledge that there are gods,” i.e.,
V. It is true that there are gods.
Now it is clear that we need to give to Velleius a further unstated premise,
since V does not follow from the conjunction of V with any previous
premise(s). This further premise must be, as is implicit in (iv),
V. Our natural preconception of the gods is such that we all agree that
there are gods.
There is a parallel for V in DND . (quoted on p. ), where Velleius
says that we have a preconception of the gods with the result that (ut) we
think the gods are happy and immortal. This seems to me to furnish an
interpretation of the implicit premise V: our preconception of the gods,
which is a necessary condition for us to be able to think about the gods,
and therefore in turn for the belief that there are gods, is of such a kind that
it also results in this belief.
Thus, if we accept that Velleius’ argument is as I have interpreted it in
V–V, he does not repeat himself, but rather proceeds from an initial
premise given in (ii), through a series of distinct inferences valid given
Epicurean canonics, to a unique conclusion in (v). Further, we do not have
to attribute to him a confusion of the preconception of the gods with the
belief that they exist.
Despite all that I have said for it, this argument for the existence of the
gods is hardly satisfactory on its own. For one thing, the premises V and
V, that Velleius draws from Epicurean canonics, are controversial, even
implausible, for somebody who is not an Epicurean. Velleius does nothing
to justify them. Thus you might ask, is it not strange that Velleius would
not help his audience, and Cicero would not help his readers, to
. Velleius’ Theology 
understand the argument, for example by providing the sort of commen-
tary I have just offered? I answer that Velleius’ behavior fits the drama. He
is talking to people who are probably aware of at least the basic infor-
mation about Epicureanism with which I prefaced my discussion of his
argument. So he simply draws on the relevant premises in canonics
without explaining them further. He might not convince his friends who
are not Epicureans, but they can see where he is coming from. But then,
what about Cicero’s readers? This passage will be much more meaningful
to philosophically learned readers than to their unlearned counterparts.
This is intelligible given my “learned reader” principle (see p. ). But
what about the unlearned reader, also on Cicero’s mind, who may want
further information? All this reader is given is a reference to where to find
help: Epicurus’ theory of the preconception. Is this a weakness in Cicero’s
writing of the scene?
I think this difficulty is less acute when we see Velleius’ argument in
its context in DND. For in the preface to the dialogue, Cicero has
already said that nearly all of us come to think that there are gods, “with
nature as a guide.” (duce natura, DND .) We might say that the
reader has been asked to grant this claim about human psychology as a
basis for what follows. Further, Balbus will later agree with Velleius that
universal and stable consensus is evidence that this belief is inborn.
(DND .–) Even Cotta will say that belief in the gods “cannot be
burnt out” of his mind and that everybody else, except the particularly
impious, agrees. (ex animo exuri non potest, DND .) Thus Balbus,
Cotta, and Cicero concede the plausibility of Velleius’ conclusion, and
some of its premises. Thus the unlearned reader, when he sees Velleius’
speech in the context of DND, may at least grant for the sake of
argument what Velleius has said. From the point of view of my
literary unity thesis (see p. ) it is reasonable that Velleius merely
mentions this basis for his position.
A more serious problem for the argument lies in its appeal to universal
consensus. The problem is not that Velleius draws an unwarranted con-
clusion from the evidence he adduces: in principle, the use of a carefully
examined consensus to establish that a belief is natural does not seem
unwarranted. The problem is that he simply asserts as evidence, from his
armchair as it were, that there has been such a consensus, everywhere and
at all times. This alleged fact could only be determined (if it could now be
determined at all) by exploration and very long empirical study. Now,
although a Stoic like Posidonius had done some ethnography, Velleius is
not unusual among Hellenistic philosophers in asserting merely from his
 Velleius the Epicurean
armchair that belief in the gods is found universally. Greek and Roman
philosophers, after all, had yet to hear of a people who did not, in the
estimation of Greek and Roman philosophers, exhibit belief in some sort
of god. If they had, Cotta or Sextus Empiricus would tell us about it
precisely to counter the argument from universal consensus. So Velleius
could be excused on the grounds that the universality of belief in the gods
seemed plausible on all sides. But Velleius is uniquely at fault for his further
assertions, about the non-cultural origins of, and persistent consensus
about, the belief. Those are subtle and specific assertions about how things
were from deep in prehistory. How could Velleius, or Epicurus, possibly
know that these assertions are true? This seems to me to be a weakness at
the heart of Velleius’ position, both about the existence of the gods and
about the further facts of their nature. But it is to these further facts that
we now turn.
Velleius informs us that our preconception of the gods is such that, in
addition to believing that there are gods, we form a view of their nature.
This view is the basis for his answer to the Central Question, and for his
account of religion:
Quod quoniam fere constat inter omnis non philosophos solum sed etiam
indoctos, fatemur constare illud etiam, hanc nos habere sive anticipationem,
ut ante dixi, sive praenotionem deorum (sunt enim rebus novis nova ponenda
nomina, ut Epicurus ipse πρόληψιν appellavit, quam antea nemo eo verbo
nominarat)—hanc igitur habemus, ut deos beatos et inmortales putemus. quae
enim nobis natura informationem ipsorum deorum dedit, eadem insculpsit in
mentibus ut eos aeternos et beatos haberemus. Quod si ita est, vere exposita illa
sententia est ab Epicuro, quod beatum aeternumque sit id nec habere ipsum
negotii quicquam nec exhibere alteri, itaque neque ira neque gratia teneri, quod
quae talia essent inbecilla essent omnia.
Seeing that almost everybody, not only among philosophers, but also
among the unlearned, agrees that there are gods, we confess that the
following is established, too: that we have a ‘pre-conception,’ as I put it
before, or a ‘pre-notion’ (for we must give new terms to new material, just
as Epicurus himself called it a prolēpsis, though previously nobody had
referred to it by that name)—anyhow, we have this preconception with
the result that we think the gods are happy and immortal. For the same

Some examples of Posidonius’ ethnography: Edelstein and Kidd (–) T, FC, FA,
Fab, F, F–.

Velleius moves without comment from the conclusion that the gods are immortal to the thought
that they are eternal. If the latter means that they are not only immortal but also ungenerated, this is
a substantial step. Even the sort of mechanism given by Lucretius in DRN .- could yield
a belief that the divine figures are immortal but that they have not existed from all eternity.
. Velleius’ Theology 
nature that gave us an impression of the gods themselves, has carved into
our minds that we should hold them eternal and happy. This being so,
Epicurus correctly promulgated his view that, “Whatever is happy and
eternal neither has any trouble of its own nor gives any trouble to another.
And thus it is in the grip neither of anger nor of grace, for everything of
that sort is weak.” (DND .–)
The quotation from Epicurus is Cicero’s translation of the first of a
collection of Epicurus’ sayings we call the Principal doctrines. Epicurus
encouraged his followers to memorize such collections. Along with the
second saying, that “death is nothing to us,” the first aims to free those
who accept it from what Epicurus thought were the fundamental fears of
unenlightened humanity: death and the gods.
The mechanism by which our natural preconception of the gods gives
us the belief that the gods are happy and eternal, Velleius does not specify.
We can fill in a mechanism he may have in mind by appeal to Cicero’s
contemporary, the Epicurean poet Lucretius. Lucretius tells us that the
humans of old used to experience, asleep and awake, images of huge,
strong figures in human form. (DRN .–) Velleius would say
these images were the basis for these people’s preconception of the gods.
Lucretius goes on:
aeternamque dabant vitam, quia semper eorum
suppeditabatur facies et forma manebat,
et tamen omnino quod tantis viribus auctos
non temere ulla vi convinci posse putabant.
fortunisque ideo longe praestare putabant,
quod mortis timor haut quemquam vexaret eorum,
et simul in somnis quia multa et mira videbant
efficere et nullum capere ipsos inde laborem.
They endowed [the gods] with everlasting life, because their appearance was
in perpetual supply and the form remained unchanged, and more generally
because they supposed that beings with such strength could not be over-
come by any force. And hence they supposed them to be supremely blessed,
because none of them seemed oppressed by fear of death, and also because
in their dreams they saw them perform many marvelous acts with no
trouble to themselves. (DRN .–, translation from LS A())

Here and below, when I translate gratia as “grace,” I do not mean to impute a Christian view of
grace to Velleius or to Cotta. Specifically, they do not use the word to mean a gift given to the
undeserving. They seem to mean that attitude whereby one helps others for those others’ own sake,
and thus a gift freely given in that sense.
 
Cf. DND ., On ends ., DL .–. Letter to Herodotus , On ends ..

Principal doctrines , Philodemus’ “tetrapharmakos” at LS J, On ends ..
 Velleius the Epicurean
The first four lines here show us why the images of the gods led early
humans to attribute immortality to the gods. The latter four lines show
why the images suggested that the gods were, in Epicurean terms, happy. It
is some such mechanism, no doubt, that Velleius thinks naturally led to
the belief that the gods are happy and eternal. That belief will form the
main premise for Velleius’ view of religion, as we shall see in section ..
But Velleius’ opponents could ask: if we naturally have this preconcep-
tion of the gods, and its attendant beliefs, why do most of us not agree
with Epicurus about theology? Velleius explains that, on its own, the
opinion that the gods are happy and eternal would be all we need for
piety. “But,” he continues, “to make sure of this opinion, the soul looks for
a form for god, and life, and action or activity of mind.” (DND .)
That Velleius makes the soul the subject of this sentence suggests that, just
as he has been speaking of natural processes that sculpt a preconception in
the soul, here he gives us another fact of human psychology independent
of human agency. It is not as though we could choose not to desire
confirmation of the opinion that the gods are happy and immortal. We
might conjecture why this is so: once we have decided that there are
eternal, happy gods, we are naturally curious about what they are like,
and whether they might have anything to do with us. We cannot help but
wonder whether, just as our own happiness depends on our form, and way
of life, and activities, the gods might not need to be active, too. Indeed, the
passage of Lucretius I have just quoted continues seamlessly into some
further judgments that humans of old made about their gods. Faced with
natural phenomena they could not understand, or feared, these people
concluded that the gods were responsible for those phenomena: the
heavens, the seasons, weather, lights in the sky. (DRN .–) In
the circumstances, this seems a reasonable, albeit false, conclusion: the
ancients had the belief (and according to Velleius, the true belief ) that
powerful-looking gods existed, so when they could not find another way to
explain things that inspired wonder or terror, they put two and two
together, and saw not only human form in the gods, but also activity,
the production of what impressed and frightened them. But in Epicurus’
eyes, this conclusion was a disaster for humanity. For, it gave the wrong
answer to the Central Question, and made us fear the gods.
What does a follower of Epicurus, like Velleius, think we should do in
this predicament? Luckily for us, we live after Epicurus. Epicurus is the

sed ad hanc confirmandam opinionem anquirit animus et formam et vitam et actionem mentis atque
agitationem in deo.
. Velleius’ Theology 
heroic figure whose clarity of mind and dexterous manipulation of his
preconceptions allowed him to reach the right conclusions about the gods.
This sort of insight, and the happiness it brought, means that Velleius is
probably serious in attributing something close to divinity to Epicurus.
And luckily for us, Epicurus passed his reasoning to his followers so that it
is now available to all. Velleius gives us some of this reasoning in the
sections after the main argument.
Let us look at an example of this help that Epicurus left us. He thought
that the gods have human form: arms, legs, heads, and so on, of the same
shape as our own. Velleius says that “in part nature suggests, in part reason
teaches” this. (DND .) Nature suggests it in that, according to Velleius,
our preconception of the gods has human shape, since that is how we
imagine the gods, “for what other form ever comes to anyone awake or
asleep?” (DND .) The verb “suggests” (admonet) is appropriate for
the way a non-rational preconception will lead us to a belief. Now Velleius’
claim might be true for Greeks and Romans but as Cotta will point out, it
would appear to be false for humanity at large: witness the Egyptians,
whose god Apis took the form of a cow. (DND .) But reason, Velleius
says, whose appropriate verb is “teaches” (docet), supplies arguments that
the Greeks and Romans have hewed closer to their natural preconception.
One argument is that the gods must be most beautiful, but no shape is
more beautiful than the human form – a violent disagreement with the
view of thinkers like Plato and the Stoics that the simple sphere is most
beautiful. (DND .– cf. DND .–) Another argument Velleius
gives will be a useful point of reference in the next section of this chapter.
This is that:
quoniamque deos beatissimos esse constat, beatus autem esse sine virtute nemo
potest nec virtus sine ratione constare nec ratio usquam inesse nisi in hominis
figura, hominis esse specie deos confitendum est.
Since it is agreed that the gods are most happy, but that nobody can be
happy without virtue, and that virtue cannot be consistent without reason,
and that reason cannot be present anywhere but in a human figure, we must
acknowledge that the gods have human appearance. (DND .)
This argument relies on premises drawn from elsewhere in Epicurean
theory: that virtue is necessary for happiness (On ends .–), that it

See Erler ().

ac de forma quidem partim natura nos admonet partim ratio docet . . . quae enim forma alia occurrit
umquam aut vigilanti cuiquam aut dormienti?
 Velleius the Epicurean
requires reason, and so on. Perhaps the strangest-looking idea (for an
ancient philosopher) is that only a human shape can support reason. This
follows from Epicurus’ materialism, much as a materialist about the mind
today might argue that there can only be reason in a human brain. For
Epicurus, the soul and its reasoning part are made of fine atoms, but they
need the structure of a body to hold them together (see LS ). We see the
reasoning part held together in the right way by the human anatomy, but
no reasoning part in any other animal. Thus we cannot expect to find
reason in anything that lacks the form of the anatomy of the relevant part
of the human body.
Note two further implications of this passage that will be of use in my
next section. These are, first, that having reason in a body of human form,
the gods must have souls. Second, having a soul in human form, the gods
must have sensation (sensus). For an Epicurean, this would also follow
immediately from the gods’ happiness: since happiness consists in a life
of pleasure, the gods must be able to feel pleasure.
Let us now leave Velleius’ positive arguments for Epicurean theology,
and move on to his treatment of Epicurus’ opponents.
. Velleius’ Treatment of his Opponents

Now that we have established the heart of Velleius’ speech, I would like to
go back to the earlier part of the speech, in which Velleius attacks his
opponents. For I think that the argument we have just explored in section
. helps us to understand Velleius’ attack.
Before the catalog of opponents itself, Velleius launches a broadside at
the theologies of the Stoics and of Plato’s Timaeus (DND .–). This
beginning looks puzzling. Instead of explaining a clear structure for what is
to come, Velleius bursts into the middle of specific controversy. Further,
Plato and several Stoics will receive a second treatment during the catalog
to follow, and the Stoics will come under fire yet again at the end of the
speech. Thus not only is this opening section poorly introduced, it seems
redundant. What is its purpose?
Both from the point of view of Velleius’ dramatic situation, and from
that of Cicero’s project, we can make some sense of Velleius’ opening. We
are told that Velleius begins with a reprise of what he had already begun
before Marcus came in. (DND .) Perhaps this explains Velleius’ abrupt
start: Balbus and Cotta know what he is up to, and he expects Marcus to
pick up the thread. Indeed, Velleius’ own reason for starting with an attack
on Plato and the Stoa is obvious enough. As has been drawn to the reader’s
. Velleius’ Treatment of his Opponents 
attention (DND .–), his interlocutors represent respectively the Stoa
and the Academy. An attack on the Stoics and on Plato, the founder of the
Academy, is thus a natural choice: “your Plato” he says to Cotta and
Marcus (Plato vester, DND .), “your [Stoic] Providence” to Balbus
(Pronoea. . . vestra, DND .). Cotta and Marcus, of course, are New
Academics. Thus, while they have some allegiance to Plato, they have no
allegiance to Timaeus’ views. The attack on Timaeus is therefore at best an
indirect and rhetorical attack on Cotta and Marcus. Velleius may be
excused on this point to the partial extent that he takes not Plato’s own
views to task, but rather “the god from Plato’s Timaeus” (Platonis de
Timaeo deum, DND .). The tenor of his criticism is mostly to ask
how Plato could come up with anything so daft, rather than to accuse
Plato of believing it.
But if Velleius’ own reason for starting his speech as he does is given by
his conversational setting, Cicero has other reasons. The speech follows
swiftly on the preface we examined in Chapter . In the preface, Cicero
presented two opposing views on the Central Question, but he did not
identify the parties who hold these views. Some readers might not be sure.
So Cicero fixes this quickly. We are told before his speech that Velleius is
an earnest Epicurean. (DND .) Immediately, he directs his attack on
the providential and creationist aspects of Timaeus’ and the Stoics’ theolo-
gies, that is to say, precisely against their positive answer to the Central
Question. Velleius’ fusillade of arguments, then, functions not so much as
the dialogue’s serious exposition of these arguments, but rather to plunge
us into the disagreement among philosophers that Cicero has just adver-
tised in the preface. We are shown, not told, that an Epicurean represents
the view that gods did not make the world and do not care for us. We have
just read that such a person would like to take away traditional piety and
perhaps “the community of the human race” (DND .). But now we are
confronted with Velleius, passionate and overflowing with arguments for
this radical view. It is suddenly not so easy to see whose side we are on.
Velleius’ speech settles down to its systematic structure at DND .,
where he moves from his first attack on Timaeus’ and Stoic theology to
his catalog of his opponents. He responds to each opponent in turn. He
treats a long list of philosophers and then, briefly, the poets and the
opinions of the vulgar. As we saw in my introduction (pp. –), it is
plausible that Cicero wrote this catalog by reordering the list of

The only clear exception is the attribution to Plato of the denial that any figure is more beautiful
than the sphere, DND . .
 Velleius the Epicurean
theological views given in an Epicurean text at least closely related to the
Herculaneum On piety. But Cicero appears to supplement this text’s
Epicurean replies to each opponent, and his supplements seem to have
certain consistent features that call for explanation. What sense can we
make of Velleius’ tirade?
Velleius does not tell us what approach he means to take to his catalog,
or how he grounds any of his criticisms of his opponents. But it seems to
me that we can detect a pattern in his criticisms. For many of them draw
on premises which an Epicurean would think follow simply from the true
and natural preconception uncovered in Velleius’ main argument. That
preconception led naturally to the belief:
N: There are gods, and gods are happy and eternal.
When I say the premises follow simply, I mean first that they follow either
from N alone or from N and some other natural preconceptions. Second,
they seem to follow by inferences of the sort we can make easily, without
philosophical training. From N follow straightforwardly:
G: There are gods.
H: The gods are happy.
E: The gods are are eternal.
We have seen that both directly from the preconception and indirectly
(DND ., quoted p. ) from H follows:
F: The gods have human bodily form.
The steps in the argument from H to F also showed us that Velleius must
think that:
A: The gods have souls.
S: The gods have sensation.
Let us now see how Velleius puts these premises to use. He has two ways
to deal with his opponents. The first way is to call attention to contradic-
tions or other incoherences among their views. The second way, some-
times combined with the first, is to point out that the opponent’s view
contradicts one or more of the seven lettered Epicurean claims we have just

Velleius tends to use the contradiction or incoherence criticism more as he moves later in his
broadly chronological list. This tendency is understandable since Velleius attributes more complex
positions to more recent and better documented philosophers. Simple views are not so vulnerable to
accusations of self-contradiction.
. Velleius’ Treatment of his Opponents 
derived, directly or via some other basic Epicurean notions, from Velleius’
preconception of the gods. It is striking that these are all the resources that
Velleius needs to refute the twenty-eight opponents in his catalog. In the
table in Appendix  I have indicated where Velleius uses each strategy.
This suggests that Velleius has a fairly restricted set of theological claims in
mind throughout his speech, and that Cicero intends to make him look a
coherent and elegant speaker within the limits of his material – as Cotta
says. (DND .)
Velleius sometimes seems to abuse his opponents immoderately. The
views of Timaeus and the Stoics are “worthless and fabricated” (futtilis
commenticiasque, DND .). The views of those in the catalog are pretty
much “the dreams of the deranged” (delirantium somnia, .).
Empedocles “slips most disgracefully” (turpissime labitur, .),
Heraclides of Pontus gives us “puerile myths” (puerilibus fabulis, .),
Cleanthes writes “as though deranged” (quasi delirans, .), and so on.
Cotta will point out that this sort of tone is characteristic of Epicureans
(.–). But there is a serious point to the terms Velleius uses. For the
belief that there exist eternal and happy gods is supposed to be natural:
everybody naturally has it, or at least had it at some point in a natural
development. The other lettered premises are supposed to follow simply
from that natural belief, requiring at most some other natural preconcep-
tion and our natural ability to reason well. So to arrive at a view that
contradicts one of these premises does not require merely the mistake of a
responsible philosopher. It means that the thinker has, at some point, flown
in the face of her accurate and natural conceptual endowment. I see the
gods in my mind’s eye, happy and eternal, and yet come to a view of them
that contradicts either this vision itself, or its conjunction with other,
equally natural and accurate preconceptions. This sort of error might rightly
be labeled a malfunction of the imagination, like madness or the non-
sequitur delusion of a dream. Velleius’ list of opponents in the catalog may
well have been determined by Cicero’s source, but the nature of the list fits
Velleius’ approach since he enumerates and replies to individual philoso-
phers, not to whole schools. Where schools could not rightly be said to
dream or go mad, he takes on the individuals whose minds “slipped,” the

Despite Cicero’s usual hostility to Epicurus, I do not think he tries to undermine Epicurean
theology by making Velleius’ catalog look silly, as e.g.McKirahan (), and Obbink ()
 suggest. That said, Cotta’s surprise that Velleius had all the facts for the catalog at his fingertips
may well be Cicero’s admission that such knowledge was unlikely in a Roman Epicurean. (DND
.)
 Velleius the Epicurean
subjects of psychological failure. In sum, just as Velleius’ positive theo-
logical arguments are based on human psychology, so his treatment of his
opponents is to put his finger on where their psyches went astray into a
sort of delusion, beckoned on by the compulsion we saw in DND .
(pp. –).
The catalog of opponents’ failings prepares us for the contrast with
Epicurus. For Velleius he was a unique figure who taught theological views
but who was himself “divine” enough to handle his preconceptions suc-
cessfully. (DND ., .) Now, the heroic status which Velleius gives
Epicurus helps us with an uncomfortable feature of his strategy. For we
could ask him, if we all have these natural preconceptions, why are
Epicurus’ views so controversial? That they are controversial, after all, is
the point of the catalog. But that Epicurus is a hero shows that Velleius
thinks that, although the cardinal points of Epicurean theology follow
simply from our preconceptions, we are compelled to know more and it is
hard to avoid psychological failures and bad philosophy of a sort that push
us away from the simple and correct chains of inference. Most of us need
Epicurus to save us.
We can see now the place of his catalog of opponents in the argumen-
tative structure of Velleius’ speech. The premises that, along with accus-
ations of self-contradiction, are used to refute all opponents, follow simply
from the preconception uncovered in the main argument, so that by
contrast with Epicurus the opponents are failures. Velleius’ coherent
strategy throughout is to work from his “facts” of human psychology.
Those facts give him, on the one hand, the preconception from which he
derives the existence, happiness, and eternity of the gods and, on the other,
the tendency of humans to draw further but false, and indeed deranged,
theological conclusions in the face of their preconceptions, so long as they
lack either Epicurus’ abilities or his “divine” aid. The catalog of opponents
is of a piece with this picture of our psyche, and what might seem the
immoderate tone of its criticism of the opponents as suffering from
madness is exactly motivated by Velleius’ ideas about psychology. Now

It seems harsh for Velleius to cast this blanket aspersion over the whole list of philosophers, since
some of them he accuses only of self-contradiction, not of contradicting their natural
preconceptions. Perhaps we are to make exceptions, or perhaps Velleius thinks that espousing
contradictory views requires a similar sort of psychological failure.

In addition to the problems encountered by early humans we saw in Lucretius (pp. –) Velleius
reminds us, in passing, of one other factor since then: people are subject to social influences towards
false views, like the blandishments of poetic depictions of the gods or the opinions of the vulgar
(DND .).
. Velleius’ Treatment of his Opponents 
this coherence would have been clearer if Velleius had not put the cart
before the horse when he gives the catalog of his opponents before the main
argument that tells us the basis for his rebuttal of those opponents. Why
does he do so and why did Cicero? For, as I have argued above (pp.
–), we can no longer suppose that Cicero wrote the speech in this
order simply because he followed sources like the Herculaneum On piety,
since we now know that On piety ended with the catalog. Cicero chose
this order.
Velleius must assume that his listeners Cotta, Balbus, and Marcus are
aware already of at least the basics of Epicurean theology. He is therefore
free to put the catalog first for rhetorical purposes. The effect is to put
Epicurus in a heroic light, by contrast with the failure of all his rivals.
Cicero, meanwhile, has an obvious reason of his own to put the catalog
first. He told us in the preface that there was huge diversity among the
philosophers in their view of the nature of the gods, but he has not told us
all those views. He limited himself to the two sides of the Central
Question. Velleius’ catalog, coming soon after the preface, thus fills this
gap for readers. But this is hardly sufficient reason for Cicero to put the
cart before the horse. The catalog would fill the gap even if it appeared later
in the speech. I suspect Cicero had a further motivation, to do with his
larger goals for DND. We have just learnt that Velleius earnestly represents
a philosophy that Cicero told us in the preface would overthrow traditional
piety, and perhaps even justice itself. We know, then, that Velleius is a
radical reformer of our ideas. By starting with the initial attack on the
Stoics and Plato, and then by having him catalog the failings of all rival
philosophers, of the poets, and of the vulgar as instances of delusion,
Cicero shows us the thoroughness and supposed rigor of this radical
attitude and prepares us for Velleius’ positive portrayal of Epicurus and
his revolutionary views.
Velleius’ theology for our purposes, then, is this. Human beings natur-
ally form a correct preconception of eternal, happy gods, with bodies of
human form. This shows us that those gods exist. But the human psyche
does not rest easy with this preconception. Instead, it casts about for
further explanations of the gods’ character or way of life. In this subsequent
reasoning we are sadly prone to mistakes, especially under cultural

Obbink puts a similar point: “Given his purpose in writing, it seems understandable that Cicero,
who consistently portrays the Epicureans as cultural renegades bent on undermining the whole of
traditional learning, should have given pride of place to the Epicureans’ criticism of the
philosophical and mythical tradition.” (, p. )
 Velleius the Epicurean
influence but also (I have presumed, relying on Lucretius) when we
confront features in the natural world we do not understand. We tend
to think that the gods are active, and that they show anger or grace towards
us. Thus philosophers may end up with theologies that fly in the face of
their original preconceptions of the gods and of happiness. These theolo-
gies are a kind of madness. Epicurus was able without help to resist these
mistakes and to come to the proper conclusion: the eternal and happy gods
will not care for us one way or another. We ought to take up his help and
reach the same conclusion. Such is Velleius’ answer to the Central
Question.
. Religion and Epicureanism

We saw that Velleius brought his main argument as far as the first of
Epicurus’ Principal sayings, in which the happiness and eternity of the gods
was said to show that they have neither anger nor grace (DND .,
pp. –). Velleius then concludes his main argument as follows:
Si nihil aliud quaereremus nisi ut deos pie coleremus et ut superstitione
liberaremur, satis erat dictum; nam et praestans deorum natura hominum
pietate coleretur, cum et aeterna esset et beatissima (habet enim venerationem
iustam quicquid excellit), et metus omnis a vi atque ira deorum pulsus esset;
intellegitur enim a beata inmortalique natura et iram et gratiam segregari;
quibus remotis nullos a superis inpendere metus.
If we were seeking no more than piously to worship the gods and to be freed
from superstition, then what has been said would be enough. For then, on
the one hand, we humans would worship with piety the gods’ preeminent
nature, since that nature is both eternal and blessed—for whatever excels is
justly worshipped—and, on the other, we would have been purged of any
fear of violence or anger from the gods. For it is understood that anger and
grace do not mix with a happy and immortal nature. When we have taken
away anger and grace, no fear of the heavenly gods hangs over us. (DND
.)
Cicero makes Velleius draw this consequence of the main argument and of
Epicurus’ Principal doctrines  first because this principle doctrine gives a
clear answer to Cicero’s Central Question: no, the gods do not care for us.
Second, Velleius shows how this answer to the Central Question moder-
ates religion, one of Cicero’s purposes in DND and Div.. He says that
Epicureans will give cult to the gods. In Cicero’s usage in the preface to
DND, this seemed likely to refer to the performances required by
orthopractic religion. But, adds Velleius, Epicureans shall do so piously
. Religion and Epicureanism 
and without superstition. They will be pious because they will believe
correctly that the gods are eternal and happy and that such excellence
makes the gods worthy of worship. They will be without superstition
because they will not fear the gods in the way that believing they might
care about us would allow us to fear them. Velleius, then, shows us here
that in the preface Cicero was using a traditional notion of piety, but not
the only one. The Epicurean version, indeed, is intended as a different way
to do what Cicero wants: to impose limits on what one may piously believe
about religion. In Cicero’s version in the premise, piety was between two
limits: impiety on one hand (believing that the gods care for us less than
they do) and superstition on the other (believing that they care more than
they do). In Velleius’ scheme, one could not be both pious and super-
stitious. For Velleius, superstition and impiety are the same thing, namely
the belief that the gods care for us. This for him involves both the false
belief, which he would call superstition, that we should fear the gods, and
the false belief, which he would call impiety, that the gods are not happy in
the Epicurean sense of the word.
Now Principal doctrines ’s crucial point in relation to religion, that the
gods lack anger and grace, is to be inferred in particular from divine
happiness. The Stoics, for whom the virtuous are not subject to everyday
passions, will agree with the Epicureans that a god’s happiness implies that
he is not subject to anger. But they will not agree that it implies that he
lacks grace in the Epicurean sense. For Velleius wants grace to imply that a
gracious being will want to help others when it does not need to for the
sake of its own happiness. But the Stoic god helps human beings when he
does not need to for the sake of his own happiness. Here we have a
fundamental point of disagreement between Velleius and the Stoic view,
and it turns on the ethical question of the nature of happiness. The final
section of Velleius’ speech sets out this disagreement. (DND .–)
Epicurean happiness is the life of pleasure. Not long before he wrote
DND, Cicero wrote in On ends a discussion of this view of happiness for
the Epicurean Torquatus. Torquatus says, “there can be no doubt that
pleasure is the highest and ultimate of all goods, and that to live happily is
nothing other than to live with pleasure.” (On ends .) But the
extreme of pleasure is not to add more and more individual and exciting
pleasures, like food or sex, but rather to achieve the absence of physical pain
and psychic distress, tranquility. (On ends .–, ) Mental pleasure

non potest esse dubium quin id [i.e. voluptas] sit summum atque extremum bonorum omnium, beateque
vivere nihil aliud sit nisi cum voluptate vivere.
 Velleius the Epicurean
can overcome physical pain, so that the sage may endure physical illness
and injury by recourse to her mental delight. (On ends .) Her happy life
of pleasure is thus achieved not by hedonism in our everyday sense, but by
a disciplined, modest life, run by rational virtues that allow the virtuous
person to feel that her pleasure is secure. (On ends .–) Further,
Epicureans thought that the basic ground on which this view of happiness
rests – that ultimate value is found in pleasure and pain – should be
obvious to all, since it is in pleasure and pain that we directly feel value,
good or bad. (On ends .–) This direct sensation of value, and its
consequent preconceptions of happiness or the good, plays a role in the
view of happiness similar to the preconception of the gods in theology:
from it the rest should be inferred.
Velleius’ gods must be happy (beati), then, in this Epicurean sense.
Since this sense of happiness is naturally available to us just as is the
preconception of the gods, we ought all to infer from these two precon-
ceptions that the gods have this kind of happiness. But many of us, for
example the Stoics, do not infer this. Velleius says:
et quaerere a nobis Balbe soletis quae vita deorum sit quaeque ab is degatur
aetas. ea videlicet qua nihil beatius nihil omnibus bonis affluentius cogitari
potest. nihil enim agit, nullis occupationibus est implicatus, nulla opera moli-
tur, sua sapientia et virtute gaudet, habet exploratum fore se semper cum in
maximis tum in aeternis voluptatibus. hunc deum rite beatum dixerimus,
vestrum vero laboriosissimum.
You [Stoics], Balbus, are in the habit of asking us what life the gods have,
and how they pass their time. The answer is that it is a life happier than
which nothing can be imagined, nothing more abundant with every good.
For a god does nothing, is entangled in no business, strives at no tasks,
rejoices in his own wisdom and virtue, and has it for certain that he will
always have pleasures that not only are of the greatest magnitude, but also
are eternal. We shall say rightly that this god is happy, but that your god
struggles extremely. (DND .–)
Velleius gives the gods each of the aspects of happiness Torquatus defines
in On ends Book . They are wise and virtuous and take mental joy in this.
Their security in their pleasure is eternal and they do not need to under-
take even those labors which the Epicurean human sage will undertake out
of necessity. There is no possibility of distress, either present or future, to
disturb them. Velleius contrasts this god with the Stoic version, who must
constantly labor to keep every aspect of the world going. For Velleius the
Stoic world has an unhappy, self-torturing god that intrudes itself into our
. Religion and Epicureanism 
lives because of its unnecessary concern for us (DND .), fating events
and providing us with divinatory information (DND .). From such a
nightmarish thought, Velleius finds that Epicurean cosmology, where the
cosmos is made by the chance collision of atoms, is a welcome release.
(DND .)
his terroribus ab Epicuro soluti et in libertatem vindicati nec metuimus eos quos
intellegimus nec sibi fingere ullam molestiam nec alteri quaerere, et pie sancte-
que colimus naturam excellentem atque praestantem.
Loosed by Epicurus from such terrors and emancipated into freedom, we
do not fear those whom we understand never make any trouble for
themselves and do not seek any trouble for another, and we give
cult piously and with holiness to that excellent and outstanding nature.
(DND .)
Religious performances, then, are made into pious actions when done in
favor of a god possessed of Epicurean happiness, who therefore takes no
interest in us and should not scare us. This passage marks the ends of
Velleius’ arguments. Cicero has given him a peroration on happiness that
establishes an Epicurean answer to the Central Question and points to the
upshot of this answer for religion.
Velleius, then, makes some illuminating but bare statements of the
Epicurean attitude to religion in general. For Cicero’s project, this leaves
a crucial question unanswered: how would the Epicurean attitude apply to
Roman religion in particular? Velleius’ attitude might suggest that an
Epicurean would wish to alter the demands of the Roman orthopraxy.
Indeed, it seems likely that if Epicurus had designed a society and religion
from scratch, he would not have prescribed all the practices of the Roman
state. But in reality Epicureans found themselves living in cities with
established orthopraxies. The Herculaneum On piety – whose subject we
reconstruct as piety, not theology at large – proposes a litany of evidence
both that Epicurus himself made at least all the religious performances
legally required where he lived, and that he recommended to other
Epicureans that they do the same (lines – in Obbink’s edition).
The papyrus sums up the evidence it will present:
[ὁ δ’] Ἐπίκουρος φανή[σεται] καὶ τετηρηκὼς [ἅπαν]τα καὶ τοῖς φί[λοις
τ]ηρεῖν παρεγ[γυηκ]ώς, οὐ μόνον [διὰ τ]οὺς νόμους ἀλλὰ διὰ φυσικὰς
[αἰτιάς·] προσεύχεσθαι γὰρ ἐν τῶι Περὶ [βίων] οἰκεῖον εἶναι [ἡμεῖ]ν ̣ φησιν,
οὐχ ὡς [δυσ]μενῶν τῶν [θεῶν] εἰ μὴ ποιή[σομεν,] ἀλλὰ κατὰ τὴν ἐπίνοιαν
τῶν [ὑπερβ]αλλουσῶν [δυνά]μει καὶ σπου[δαιότ]ητι φύσεων [ἵνα τά]ς τε
τελ{λ}ε̣[ιότητας] γ̣ιγν ̣ώσκω[μεν καὶ] τοῖς νόμοις [συμπερι]φόρας·. . .
 Velleius the Epicurean
Furthermore, it will appear that Epicurus loyally observed all the forms of
worship and enjoined upon his friends to observe them, not only on
account of the laws but for physical reasons as well. For in On lifecourses
he says that to pray is natural for us, not because the gods would be hostile
if we did not pray, but in order that, according to the understanding of
beings surpassing in power and excellence, we may realize our fulfillments
and social conformity with the laws. (On piety ll. – Obbink, text and
translation from Obbink ())
Epicurus, then, did not wish to alter the performances of traditional

religion. Instead, he wanted Epicureans to make their traditional perform-
ances into pious actions by believing Epicurus’ answer to the Central
Question.
In DND Book , it is not Velleius but rather Cotta who tells us about
this part of the Epicurean position. In the last ten sections of his speech,
Cotta turns from the Epicurean views on the nature of the gods to an
attack on the Epicureans’ notion of piety (DND .–). He, too,
makes it clear that Epicurus claimed to adhere to the local orthopraxy in
Athens where he lived:
”at etiam de sanctitate de pietate adversus deos libros scripsit Epicurus.”at quo
modo in his loquitur: ut <Ti.> Coruncanium aut P. Scaevolam pontifices
maximos te audire dicas, non eum qui sustulerit omnem funditus religionem nec
manibus ut Xerxes sed rationibus deorum immortalium templa et aras everterit.
“But[,” say the Epicureans, “] Epicurus wrote books about piety towards
the gods.” But how does he talk in these books? So that you would say you
were listening to Tiberius Coruncanius or Publius Scaevola, pontifices
maximi, not to the man who demolished all religion down to its founda-
tions, who overturned the temples and the altars, not with his hands like
Xerxes, but with arguments. (DND .)
Cotta’s incredulity indicates that Epicurus claimed not to overthrow

Athenian religion. Cotta is himself a pontifex. Two of the pontifices whom
he says he follows as particular authorities are Tiberius Coruncanius and
Publius Scaevola. (DND .) So when he says that Epicurus talks like
those two, he means that Epicurus strove to sound like somebody with a
legitimate commitment to traditional religion – in Epicurus’ case, to
Athenian rather than to Roman religion, of course. That he puts Epicurus’
point in Roman terms shows that Cotta thinks Epicurus’ attitude can be
translated from the Athenian context to the Roman one. By way of
confirmation, Cotta later says, in response to the Epicurean claim that
the gods do not care about us, that if it is so, then
. Religion and Epicureanism 
. . . quid veneramur quid precamur deos, cur sacris pontifices cur auspiciis
augures praesunt, quid optamus a deis inmortalibus quid vovemus? “At etiam
liber est Epicuri de sanctitate.” ludimur ab homine non tam faceto quam ad
scribendi licentiam libero.
. . . why should we revere the gods, why should we pray to them? Why do
the pontifices preside over the rites, why do the augurs preside over the
auspices? What do we hope for from the immortal gods? What do we
promise to them? “But there’s Epicurus’ book On holiness.” We are made
sport of by a man who was not so much witty, as unrestrained in the licence
he took when writing. (DND .–)
Cotta’s imaginary Epicurean says that Epicurus’ books could underwrite
Roman practices like augury or pontifically sanctioned rites. Through
Cotta’s incredulous eyes, then, we see how an Epicurean would want to
import Epicurus’ piety to Rome: the required performances are to be left
intact, but they are to be understood in a new way. Velleius does not
demur.
Roman religion, like Athenian religion, began to be practiced long
before Epicurus. So Velleius, who thinks that Epicurus alone drew the
right conclusions about the gods, must conclude that Roman religion
was founded by people who had a very distorted view of those gods. No
doubt these founders attributed anger and grace to their objects of
worship. Therefore, unlike Varro or the Stoics, Velleius cannot hope
to recover any forgotten theological insights that lay behind early Roman
religion. If he wants to keep it, he has to give the traditional orthopraxy
some entirely new meaning. Perhaps Cotta refers to this radical re-
interpretation of traditional religion when he compares Epicurus to
Xerxes. This comparison may not be wholly pejorative. In Cicero’s
Laws, Marcus attributes Xerxes’ destruction of Athens’ temple buildings
to his belief that the gods should not be cooped up when the cosmos at
large is their temple. (Laws .) In that telling of the story, Xerxes
did not take himself to lack piety, or to have destructive intentions,
towards the Athenian gods. On the contrary, he was motivated by his
own, un-Athenian views about the pious way to treat those same gods.
Epicurus, similarly, wants to be rid of traditional beliefs about the gods

Understood this way, Cicero’s picture of Epicurean attitudes to traditional religion accords with the
views scholars have attributed to Epicurus himself. See Obbink () –, Algra () –,
and Penwill (). A dated but still rich account is Festugière () –.

Lucretius is explicit that it was early man’s mistake in attributing control of the cosmos to the gods
that led to religious performances, DRN .–.
 Velleius the Epicurean
in order to make his own idea of piety, rather than to leave behind no
sort of piety at all.
There is more to be said about Cotta himself, the pontifex who opposes
dogmatic theology. But I leave that to my next chapter because Cotta’s role
comes most into its own in his exchange with Balbus. I shall close this
chapter with Cotta’s opposition to Velleius’ view of religion in particular.
This section is the climax of Cotta’s speech in Book , and in it Cotta
levels three criticisms. The first two argue that Velleius has not given us, as
he claimed he would, a basis for piety without superstition. The third argues
much more forcefully, that Epicurus’ reimagining of religion tears love
itself out of human life. Let us take each point in turn.
On the count of piety, Cotta says that while Epicurus may have
protested that he was pious because he made the right religious perform-
ances, there can be no piety towards the Epicurean gods. If we, consist-
ently with the Epicurean view, conceive of pious religion as a
demonstration of admiration at the gods’ excellence, then Cotta declines
to find the Epicurean gods, who are passive hedonists, excellent. Thus
Cotta, in effect, rejects the Epicurean re-interpretation of piety by rejecting
Epicurean ethics. For him, Epicurus’ admiration for uncaring hedonists is
a vice, not a virtue, and therefore is not piety. But if, Cotta says, we instead
conceive piety more traditionally, as the virtue by which we interact well
with beings with whom we are bound by duties of justice, then of course
Epicurean piety simply is not that. (DND .–)
On the count of superstition, Cotta attacks the Epicurean strategy of
minimizing entirely the degree to which the gods care for us. (DND
.–) Velleius may be able to free us from superstition this way,
since we do not end up believing that the gods care about us where they do
not. But, as just argued, he does so at the cost of throwing us into impiety.
The third count on which Cotta attacks Velleius’ view of religion is
more surprising. He takes on Velleius for denying grace to the gods: when

Cotta will say that Epicurean piety rips religio out of the human soul, so the comparison to Xerxes is
at least partly pejorative. Xerxes’ overturning of altars, after all, is less easy to excuse than his
demolition of temple walls. Lucretius offers, with approval, a comparable image of a victorious
Epicurus trampling religio underfoot. (DRN .–) Perhaps this suggests not radical reconceiving
of a religion that was begun from the wrong motives, but rather destruction of that religion, even
though Lucretius is not formally opposed to religious gestures undertaken with Epicurean piety.
(DRN .–) But Lucretius uses the term religio not, as Cicero does, for the performances of an
orthopractic religion or for a pious version of those performances, but rather for a traditional
religion of fear and placation of the gods. (DRN .–, ; .; .; .) So he may well
mean simply that Epicurus defeated religio in the latter sense. That would be compatible with
Cicero’s portrayal of Epicurean attitudes.
. Religion and Epicureanism 
Epicurus did so, he “pulled up at the roots religion from the human soul.”
(ex animis hominum extraxit radicitus religionem, DND ..) First, Cotta
points out that not only will the Epicurean gods not care for us, they will
not care for each other, either. Unlike Stoic gods they will not be social
beings, concerned for others, at all. Then, he draws a consequence for the
Epicurean view of human ethics:
ut enim omittam vim et naturam deorum, ne homines quidem censetis, nisi
inbecilli essent, futuros beneficos et benignos fuisse? nulla est caritas naturalis
inter bonos? carum ipsum verbum est amoris, ex quo amicitiae nomen est
ductum; quam si ad fructum nostrum referemus, non ad illius commoda quem
diligemus, non erit ista amicitia sed mercatura quaedam utilitatum suarum.
prata et arva et pecudum greges diliguntur isto modo, quod fructus ex is
capiuntur, hominum caritas et amicitia gratuita est; quanto igitur magis
deorum, qui nulla re egentes et inter se diligunt et hominibus consulunt. quod
ni ita sit . . .
For, to set aside the power and nature of the gods, do you not think that
humans, too, would have been beneficent and benevolent, were they not
weak? Is there no natural dearness between good people? The very word
‘love’ (amor) is dear, from which ‘friendship’ (amicitia) is derived. If we refer
friendship to our own enjoyment, not to what is advantageous for the
person we love, that will not be friendship, but rather a transaction in what
each needs for his own use. We love pastures and fields and flocks of cattle
in the latter way, because we take produce from them, but dearness and
friendship between human beings is a matter of grace. Thus how much
more gracious is the friendship of the gods, who, although they lack
nothing, both love one another, and care for humans. If that is not so,
[there follows the quotation on p. ]. (DND .)
Cotta’s criticism assumes that the Epicurean gods are an ideal for human
ethics. If the gods are not friends with one another through love of the
friend for his own sake, then ideal human beings will not be so either.
Thus good human beings will have no “natural dearness.” Instead, because
unlike the gods they cannot be secure in their life of pleasure without other
people, they will make deals to mutual advantage with other humans. Yet
Cotta says this is not how we are naturally. Our intuitions tell us that to
love another for her own sake is good. Thus, contrary to Epicurean claims,
it is Velleius whose position flies in the face of our natural instincts.
One way to put the point is this. Epicurus and his followers liked to
protest the immense value they found in friendship, in justice, and in

This position bears a close resemblance to the one which Laelius takes up in Cicero, On friendship,
especially in sections – of that work.
 Velleius the Epicurean
social life in general. But for Cotta, the case of the gods reveals the
hollowness of those protests. For the gods, having perfect security of
pleasure without anybody else’s help or forbearance, do not trouble
themselves about anybody else. This shows that a human Epicurean sage
will be friendly, just, and social only because he needs to in his human
weakness. If only he were as strong as a god, he would not care about
others at all.
This criticism of the Epicurean religion must be what Cicero had in
mind in his preface when he said that the community of the human race
would be taken away were Epicurean theology true (see p. ). It seems to
me to be a deep criticism. For you might think that a great attraction of
Velleius’ Epicurean religion was as follows. When Cotta denies that there
can be piety for beings who do not give in return (DND .), this might
put us in mind of a stereotype of Roman religion: do ut des, “I give so that
you give.” Roman religion often looks like a self-interested business,
seeking gain from the gods. A cynic could read Cicero’s own preface in
this vein: if the gods do not give anything back, there is no virtue in
worshipping them. In fact, Cotta will be frank that the city’s gain is his
own reason for maintaining Rome’s cults. (DND .) Of course, Epicur-
ean ethics, too, refers all virtuous actions ultimately to one’s own happi-
ness. But an Epicurean worshipper does not expect any crude, pragmatic
returns, like riches or victory in battle. Her attention is directed at
admiration of the beautiful and perfectly happy gods. The benefits that
will accrue to her from this admiration are, so to speak, spiritual: they
improve her state of mind. But Cotta’s criticism takes away this seemingly
impressive feature of Epicurean religion. For the picture of piety which
Cotta opposes to the Epicurean version is not do ut des, but rather a desire
on either side to give simply out of love. On this opposing view, the gods
give benefits to humans because they wish to benefit us for our own sake,
and if we are virtuous we, similarly, worship them out of a desire to do
right by them. It is the same as when good people help one another not out
of self-interest, but because they care for one another. Even if the Epicur-
ean gods were not too lazy and selfish to deserve worship, Cotta suggests,
Epicurean piety is itself a selfish affair.

On ends .–; LS .

Something like this seems to be the freight of the charge that Cicero likes to lodge against those who
make friendship or justice stem from need rather than from nature (presumably the Epicureans): On
friendship –, Laws .–.
 
Balbus the Stoic and Cotta the Skeptic
Nature, says Balbus, is not only craftsmanlike, but is actually a craftsman.

(natura non artificiosa solum sed plane artifex, DND .–) The world
not only has the appearance of rational design, it was indeed designed, and
was made, by a rational god. Balbus argues at length that the cosmos is
“governed” by the gods (administrari, DND .–) who “care about
us” (consulere nobis, DND .–) and even about individual humans
(DND .–). It was made for the sake of gods and of humans, who
share it as our common home. (DND ., .) Against Velleius, here
we have a stout defender of the contrary answer to Cicero’s Central
Question (see p. ): the gods do care for us.
But Balbus also describes one category of the gods thus: “From . . .
physical theory has flowed a great mass of gods, who, dressed in human
appearance, have supplied the poets with myths, but have crammed
human life with every supersitition.” (DND .) Who are these “con-
fected and fictional gods” (.) who feature in “impious myths” (DND
.)? To name a few, they include Jupiter, Juno, Minerva, Vesta, and
Janus (DND .–) – many of the leading gods at Rome. Stoicism
might lend piety to some ideal, rational religion. But will it do for Roman
religion? Balbus thinks so. Cotta, both a skeptic and a pontifex, is not
convinced.
This is the side of the Central Question I examine in this chapter, and
the one to which Cicero turns our attention in Books  and  of DND.
Balbus defends the Stoic view (Book ), then Cotta delivers a skeptical
response (Book ). Thus although the exchange on Stoicism is formally of
equal importance in the structure of the DND’s debate, Cicero allows it
much greater length than he does to Epicurus. He also gives it a narrative
exclamation mark. For the dialogue ends with his contrast of Velleius’

ex ratione . . . physica magna fluxit multitudo deorum, qui induti specie humana fabulas poetis
suppeditaverunt, hominum autem vitam superstitione omni referserunt.

 Balbus the Stoic and Cotta the Skeptic
preference for Cotta’s case with Marcus’ preference for Balbus’. (.) We
last see Velleius, so to speak, sitting on the sidelines with Marcus, a
spectator at the main event. The reason for this is not far to seek. Cicero
is careful to stage a debate where the reader is indeed free to make up his
own mind on the Central Question. As we saw in Chapter , the
Epicurean view receives fairer treatment in DND  than Cicero’s preface
might have led one to expect. In the end, Cicero tells us that Marcus liked
Balbus’ speech more than Cotta’s – but not whether Marcus liked Balbus’
speech more than Velleius’. But in my final chapter I shall argue that
Marcus in Div. does express a preference for Balbus’ view of religion.
Cicero himself also thinks that Stoicism is worth taking more seriously
than is Epicureanism. Thus he stages more grandly the debate on the
Stoa’s prospects for moderating religion.
In this chapter I shall start by examining the dramatic aspects of the
exchange between Balbus and Cotta (section .). For the drama around
their respective speeches helps us to situate the speeches in Cicero’s
project. Then I shall examine Balbus’ speech in section . and Cotta’s
in section ..
. Balbus versus Cotta on Religion

In Chapter , we saw that the implications of Epicurean theology for
religion form the climax to Velleius’ and Cotta’s speeches in DND Book .
In Books  and  this connection is not emphasized in the same way. It is,
nevertheless, emphasized, and it is my main purpose in this section to
point out that emphasis. I shall argue that in this part of the dialogue
Cicero keeps his project before the reader’s mind by dramatic means,
engineering a personal clash between Balbus and Cotta that puts the
spotlight on the adequacy of Stoic theology to moderate Roman religion.
Another purpose for this section is that I wish to remove a false assumption
that has been common in scholarly interpretations of the latter two books
of DND. This is the assumption that Balbus represents theism and support
for traditional religion, while Cotta represents opposition to both these
causes. On the contrary, both Balbus and Cotta are theists and supporters

This assumption is visible in, for example, Bringmann () , and Momigliano () . Of
course, there are many other interpretations of DND . which do not make the objectionable
assumption, though they may be mistaken in other ways. E.g. Levine, who offers the literary analysis
that Cicero was afraid of Cotta’s view being misunderstood as negative dogmatism (Levine ()
–); a biographical explanation by Glucker ()  n. , but cf. successful refutations by
Görler (), DeFilippo ()  n. ; Mora () –, who accepts that . is corrupt,
. Balbus versus Cotta on Religion 
of the traditional religion, but Cotta is the arch-traditionalist, and Balbus is
the would-be reformer of related beliefs.
First, recall again that while we readers know of Cicero the author’s
project and that we are to consider the Central Question, the characters in
DND have taken as their topic simply the nature of the gods. (DND .,
., .) In this way the characters naturally give their answers to the
Central Question in the context of their whole theologies, where if only the
Question had been put before them, their speeches would have been more
circumscribed. But Cicero must therefore write the drama in such a way
that the reader keeps the Question in mind and is steered towards the
relevant theology. Let us see how this is achieved in Books  and  of DND.
Cicero bookends Balbus’ speech with two short conversations in which
Balbus and Cotta trade remarks on Cotta’s dual status as a pontifex and an
Academic skeptic. (DND .–, .–.) Balbus attacks first, immedi-
ately before and after his speech:
eundem equidem mallem audire Cottam, dum qua eloquentia falsos deos
sustulit eadem veros inducat. est enim et philosophi et pontificis et Cottae de
dis inmortalibus habere non errantem et vagam ut Academici sed ut nostri
stabilem certamque sententiam.
I would prefer to hear this same Cotta introduce true gods with that same
eloquence with which he refuted the false [sc. Epicurean] ones. For it
belongs to a philosopher and a pontifex and a Cotta to have about the
immortal gods not a wandering and roving view like the Academics, but
rather a stable and sure one like our [Stoics]. (DND .)
haec mihi fere in mentem veniebant quae dicenda putarem de natura deorum.
tu autem Cotta si me audias eandem causam agas teque et principem civem et
pontificem esse cogites et, quoniam in utramque partem vobis licet disputare,
hanc potius sumas . . . potius huc conferas. mala enim et impia consuetudo est
contra deos disputandi, sive ex animo id fit sive simulate.
These things, more or less, came to mind as what I thought I should say
about the nature of the gods. But you, Cotta, if you heed me, would make
the same case and consider that you as both a leading citizen and a pontifex
should take up this case rather than [the opposing one], since you [Aca-
demics] are allowed to argue on either side . . . For it is a bad and an
impious habit to argue against the gods, whether it comes from the heart or
is pretended. (DND .)
which seems unlikely in view of the testimony of Augustine City of God .. I give my own
interpretation of DND . in Chapter . Pease () and in his commentary (–)
surveys some older interpretations.
 Balbus the Stoic and Cotta the Skeptic
Balbus takes full account of the subtleties of Cotta’s roles in these two
criticisms, as follows.
Balbus’ criticisms are not of Cotta as merely an adherent of the religion
the pontifices prescribe. Balbus might think an Academic is capable of
making those performances in a religiously correct way, albeit impiously
from the point of view of Stoic ethics. The epistemological standard he
proposes is for philosophers and for those in charge – leading citizens,
aristocrats like men of Cotta’s family, and the pontifices. It is the pontifices,
after all, who not only must perform what is prescribed but also must give
the prescriptions. In the former passage just quoted (DND .), Balbus
poses his epistemological problem: there is an instability in the views of a
skeptic. Presumably he means by this that a radical skeptic will hold her
views weakly and according to how things seem to her at the time, not with
the conviction that she has in hand sufficient reason to take them to be true.
Thus she may find it easier to revise her views than does a dogmatist like
Balbus. Such potential instability of views about the gods might on its own
keep a pontifex from doing a good job. For Cotta might change his mind
too often about what is required orthopractically in difficult cases. Worse
still, who is to say that his magpie mind will not light on the view that the
gods wish him to use his position to overthrow the traditional orthopraxy
altogether? Furthermore, even if instability of views does not prevent Cotta
from doing his job as required by religious constraints, it might well lead
him into episodes of impiety or superstition. For example, he might swing
to the view that he is not performing his duties for a god who cares when, in
fact, a god cares. That would be impious.
In my second passage (DND .), Balbus does accuse Cotta of risking
impiety (as opposed to religious incorrectness), but in doing so he still
honors the rationale behind Cotta’s skeptical practices. He does not, for
example, infer from Cotta’s anticipated arguments against Stoic theism
that Cotta himself denies the existence of the gods. Balbus is aware that
such arguments might be “pretended” (simulate) only in order to answer
his own. Rather, he says that the practice of even pretended arguments
against the gods is impious. The thought seems to be that it is impious to
devise or to give such arguments even if one does not endorse them. So
skeptical practice is impious because it leads one habitually to give such
arguments. Balbus implies that that is especially bad for somebody in
religious authority.
Before we see Cotta’s answers to these two charges, note what they
imply for Balbus’ understanding of his own Stoicism and for the structure
of his speech. For although Cicero’s characters are not always convinced of
. Balbus versus Cotta on Religion 
the arguments they are assigned, we are told that Balbus is an expert Stoic.
Thus we may presume the Stoicism he expounds reflects his own opin-
ions. Now Balbus says that as a pontifex Cotta should have stable views
and should not have to speak against the gods, not even for the sake of
argument. To meet these demands, it seems, Cotta could pick any dog-
matic theism he liked. But Balbus wants Cotta to introduce the true gods
(as in DND .). For Balbus, the true gods are the Stoic gods. Thus,
Balbus must imagine that Stoicism could be a philosophy for a pious
pontifex. He must think the apparent tensions between Stoicism and
Roman religious orthopraxy to which I pointed at the start of this chapter
can and should be resolved. We shall see soon how Balbus hopes to do this
(section . of this chapter).
Next, see the implication for the makeup of Balbus’ speech. Shortly
after DND ., with the superiority of his theology for pontifical purposes
fresh in his mind, Balbus proposes the following structure for his speech:
omnino dividunt nostri totam istam de dis inmortalibus quaestionem in partis
quattor.
[] primum docent esse deos,
[] deinde quales sint,
[] tum mundum ab his administrari,
[] postremo consulere eos rebus humanis.
nos autem hoc sermone quae priora duo sunt sumamus; tertium et quartum,
quia maiora sunt, puto esse in aliud tempus differenda.
Our Stoics divide that whole inquiry about the immortal gods into four
parts.
[] First, they teach that there are gods;
[] then, what the gods are like;
[] next, that the cosmos is governed by the gods;
[] finally, that the gods care about human affairs.
But in this conversation, let’s take up the first two. The third and fourth,
because they are longer, I think should be carried over to another time.
(DND .)

DND .. Quintus Lucilius Balbus was a senator and is probably one of the two Balbi Cicero
mentions as Stoics at On the orator .. Otherwise the historical Balbus is obscure to us. No doubt
Cicero wanted to choose a Roman, even a relatively obscure one, who could plausibly be assigned an
in-depth knowledge of Stoic theology. Perhaps it is playful that this long, minutely argued, but less
eloquent speech is delivered by a man called Balbus (meaning “the stutterer”). See also Goulet
(–) vol.  number A.. Balbus makes a contrast with the historical Cotta, who was an
orator whom Cicero admired (see Malcovati () vol.  –). “Cotta” also narrates On the
orator. see also Goulet (–) s.v. “Cotta.”
 Balbus the Stoic and Cotta the Skeptic
Balbus at first intends his speech to end with the Stoic view of what the
gods are like, in part no doubt because that is the question the characters
believe is before them. But note that Balbus speaks these words with DND
. fresh in his mind. We may conclude that he plans to finish with the
climax of section [] as we have it, the declaration that a Stoic theology
would help one to right religion and to avoid superstition (DND .–).
Just as Velleius’ speech ended with a conclusion from the nature of his
gods to the moderation of religion, so would have Balbus’.
To Balbus’ intention to omit parts [] and [] of his speech, Cotta
replies, “Not at all, for we are at leisure, and our debate is on issues which
should come even before business.” (DND .) Thus Balbus gives parts
[] and [] of his speech after all. Here Cicero’s hand is clear. For the
Stoic-sounding answer to the Central Question in the preface to DND was
phrased thus: “that the whole cosmos is governed and ruled by the mind
and reason of the gods” and “even that the gods care about human life and
are provident.” (DND .) Thus parts [] and [] of Balbus’ speech argue
with precision for the result Cicero represented in the preface as one
important answer to the Central Question. Thus we see that while Balbus’
initial plan for the speech and Cotta’s reply might seem gratuitous dra-
matic details, by introducing them Cicero contrives to have his cake and
eat it too. For [] and [] must follow [] and [] in the argument, and in
any case are the parts of the speech directly relevant to the Central
Question, so that Balbus’ reflections on religion in section [] might seem
less prominent than did Velleius’ equivalent peroration. Cicero thus has
Balbus address the full Stoic answer to the Central Question. But now we
also see that Balbus has his own eye on section [], and its reflections on
traditional gods, as something of a climax in his message to Cotta. Thus we
are told that Balbus attaches importance to his own use of his view of the
nature of the gods as a way to moderate religion – which was one of
Cicero’s stated purposes in the first sentence of the preface.

minime vero . . . nam et otiosi sumus et his de rebus agimus, quae sunt etiam negotiis anteponenda. Mayor
() ad loc. notes that Cotta’s words allude to Plato, Phaedrus b. This elegant turn suggests
that Cicero had some careful purpose in mind for them.

Auvray-Assayas () and () shows that a huge transposition of the text as found in the MSS is
made in every modern edition of DND Book . She demonstrates that originally the transposition
was made on the basis merely of a conjecture due to Politian. She then argues (a) that there is no
philological reason to accept this conjecture and (b) that the MSS text makes for interesting
philosophical reading. I would resist the argument at (a), since it seems to me that . (qui locus
est proximus) answers the structure adumbrated at . (tertius est locus) and must be connected to
. as it is in modern editors’ texts.

deorum mente atque ratione omnem mundum administrari et regi; etiam ab isdem hominum vitae
consuli et provideri.
. Balbus versus Cotta on Religion 
Cotta is rankled enough by Balbus’ challenges to answer them. This is
not part of his skeptical counter-argument, but a set of personal remarks:
“but before the matter at hand, a few things about me.” (sed ante quam de
re, pauca de me, .)
non enim mediocriter moveor auctoritate tua Balbe orationeque ea quae me in
perorando cohortabatur ut meminissem me et Cottam esse et pontificem; quod
eo credo valebat, ut opiniones, quas a maioribus accepimus de dis immortalibus,
sacra caerimonias religionesque defenderem. ego vero eas defendam semper
semperque defendi, nec me ex ea opinione, quam a maioribus accepi de cultu
deorum immortalium, ullius umquam oratio aut docti aut indocti movebit. sed
cum de religione agitur, Ti. Coruncanium P. Scipionem P. Scaevolam pontifi-
ces maximos, non Zenonem aut Cleanthen aut Chrysippum sequor, habeoque
C. Laelium augurem eundemque sapientem quem potius audiam dicentem de
religione in illa oratione nobili quam quemquam principem Stoicorum. . . .
habes Balbe quid Cotta quid pontifex sentiat; fac nunc ego intellegam tu quid
sentias; a te enim philosopho rationem accipere debeo religionis, maioribus
autem nostris etiam nulla ratione reddita credere.
For I am no little moved by your authority, Balbus, and by your speech
that, in its conclusion, urged me to recall that I am a Cotta and a pontifex.
[Cf. DND ..] I believe that meant that I should defend the rites, the
ceremonies, the religious duties (religiones), and the opinions about the
immortal gods that we have received from our ancestors. Assuredly I shall
always defend them and have defended them always. Nor will the speech of
any man, neither a learned man nor unlearned, ever move me from that
opinion that I have received from the ancestors about the cult of the gods.
But when religion is at issue, I follow Tiberius Coruncanius, Publius Scipio,
and Publius Scaevola the pontifices maximi, not Zeno or Cleanthes or
Chrysippus, and I have Gaius Laelius, an augur and at the same time a
wise man, whom I would rather hear speak about religion in his noble
oration than any Stoic chief. [The text quoted on pp. – appears here.]
There you have the view of a Cotta and pontifex, Balbus. Now make me
understand what your view is: for I must receive a theory of religion from a
philosopher like you, but I must trust our ancestors even with no
theory given. (DND .–)
Cotta’s rhetorical tactic is to shift the burden Balbus has placed on him
back onto Balbus. Now he, Cotta, is the unshrinking defender of trad-
itional religion as such. Balbus, with his rationalizing theories, is the threat
to tradition. Let us examine briefly how Cotta argues that this is so.

For more detail on Cotta’s skepticism and its relation to his pontifical duties, see Wynne (). For
another view, see DeFilippo ().
 Balbus the Stoic and Cotta the Skeptic
Although in DND .– he refers to . in particular, I think Cotta
also answers the instability charge of .. He reports how his own behavior
and disposition seem to him: “Assuredly I shall always defend them and have
defended them always.” He does not claim to believe the ancestral opinions
about the immortal gods in a dogmatic way but rather that he stably makes a
defence of them. In this way, lack of theological views or unstable theological
views might be no threat to a pontifex’s performance of his duties, provided
that the pontifex decided to act on ancestral authority. Meanwhile, against
the charge of impiety Balbus brought against even insincere skeptical
arguments against the gods (DND .), Cotta turns the tables. As one
who trusts the ancestors, Cotta has ample basis to accept what seem to be
the traditional views about theology and religion, for example, that there are
gods. But Balbus, as a Stoic, makes acting on such a belief dependent on
having sufficiently good evidence in hand that one is warranted to assent to
it dogmatically. Cotta, of course, will go on to argue that Balbus offers no
such evidence. “You bring forward all these proofs as to why there are gods,”
Cotta will say later, “and you make doubtful a matter that is, in my view,
least doubtful.” (DND .) Thus Cotta, who trusts his ancestors in this
matter, feels secure in supporting the religion they have handed down, but
fears that Balbus, who demands a “theory of religion,” will be left bereft
when Cotta shows that the Stoic theory is ill-founded.
Cotta is also a theist, or as much a theist as an Academic skeptic can be.
He says so both to Velleius and to Balbus. To Velleius he says:
itaque ego ipse pontifex, qui caerimonias religionesque publicas sanctissime
tuendas arbitror, is hoc quod primum est, esse deos, persuaderi mihi non
opinione solum sed etiam ad veritatem plane velim. multa enim occurrunt
quae conturbent, ut interdum nulli esse videantur. Sed vide quam tecum agam
liberaliter: quae communia sunt vobis cum ceteris philosophis non attingam, ut
hoc ipsum; placet enim omnibus fere mihique ipsi in primis deos esse.
So I myself, a pontifex, who think that the rites and public religious duties
are to be defended as most sacred, I would want entirely to be convinced
about that which is the first issue, that there are gods, not only as a matter of
opinion but even as regards truth. For many points rush in to confuse, so
that sometimes there seem to be no gods. But see how generously I will deal
with you: I won’t touch what is common to you and other philosophers,
like this position itself—for nearly everybody holds, and I myself am among
the first to do so, that there are gods. (DND .–)

adfers haec omnia argumenta cur dii sint, remque mea sententia minime dubiam argumentando dubiam
facis.
. Balbus versus Cotta on Religion 
And to Balbus:
. . . quod inter omnis nisi admodum impios convenit, mihi quidem ex animo
exuri non potest, esse deos, id tamen ipsum, quod mihi persuasum est auctoritate
maiorum, cur ita sit nihil tu me doces.
. . . what is agreed among all people, except the utterly impious, cannot be
burnt out of my mind: that there are gods. But that very point, of which
I am persuaded by the authority of our ancestors—you teach me nothing of
why it is the case. (DND .)
Cotta seems to agree with Cicero’s observation in the preface that almost
everyone believes in the gods “with nature as a guide.” (DND .) He does
not put the point in the form of a dogmatic-sounding claim about human
nature. Rather he concedes how two sorts of data appear to him, psycho-
logical and sociological. In his own psychology he can report that the thesis
“there are gods” cannot be “burnt out” of his mind. Similarly, in society,
he sees that nearly everybody agrees with him. That includes the Roman
ancestors, by whom he is further persuaded. Now he does not hold this
view as a “matter of truth.” (DND .) For unlike a dogmatist he does
not take himself to have discovered the truth of the matter. All a skeptic
can do is report how things seem to him, and to Cotta it seems that there
are gods. To this he enters one exception. When he encounters arguments
for the existence of the gods, his skeptical training calls counter-arguments
immediately to mind. In the immediate grip of these (“many points rush
in to confuse”) it can seem to Cotta that there are no gods. But this is a
phenomenon that arrives in intellectual discussion. In the public assembly,
it would be difficult to deny the gods’ existence. (DND .) Cotta is a
thorough theist, then, and it is only when dogmatists try to help him out
that he calls this view into question.
Thus I draw two conclusions about the staging of Books  and . First,
we are shown where to look for the material that interests us as readers of
Cicero’s project, examining the Central Question. But, second, we have
also been shown two interlocutors neither of whom is hostile either to
theism or to the traditional religion. Each thinks that his set of views (or
suspensions of judgment) keeps orthopraxy pious, and that the other’s

Cotta says that he would nevertheless like to be convinced as a matter of truth that there are gods.
This might sound unskeptical, or at any rate not compatible with the most radical sort of Academic
skepticism in which the skeptic does not take his views to be true. Should such a skeptic even wish
to discover the truth? But Cicero’s answer to this question is “yes,” as is clear from Academica ..
Cotta’s desire here is thus compatible even with the most radical, Clitomachean Academic
skepticism.
 Balbus the Stoic and Cotta the Skeptic
dogmatism or skepticism threatens to lead its adherents into impiety or
superstition. In sections . and . I shall now devote more space to
Balbus’ side of the dispute. That is because his Stoic position, as one of the
two answers to the Central Question, is of great importance to my reading
of DND. Moreover, it is the view that Marcus will prefer in the end. In
section . I shall then turn back to Cotta.
. Balbus and the Central Question

Balbus’ speech is very long and detailed. A full exposition of his theology
would be beyond the scope of this book. In this section, then, I shall
concentrate on two areas that are of direct relevance to my interpretation
of Cicero’s project. In section .. I shall first exhibit Balbus’ extremely
emphatic answer to the Central Question, that the gods do care for us.
Then, in section .., I shall expand on one aspect of this answer, its
stress on beauty. I take up this latter point because I think it is the bridge
that links the emphasis on beauty in the opening sentence of DND (“most
beautiful for the mind to grasp,” see p. ) to its reappearance in the
closing sentences of Div. (., “the beauty and the order of the cosmos
compel one to admit. . .,” see pp. –).
.. That God Governs the World and Cares about Us

In this section I shall often speak of “Balbus’ god” or “the Stoic god” and
by this I mean the cosmic active principle (on which more in a moment).
Like many ancient polytheist theologians, Balbus will often use “god” or
“gods” interchangeably to mean roughly what we call “god,” or the divine
in general. Thus, for example, Balbus’ four theses in his speech are phrased
to respond to polytheist questions – “that the gods govern the world” etc. –
but often his arguments for these theses focus without embarrassment on
the cosmic god. Sometimes, though, when Balbus says “gods” he intends
specifically the gods plural, as when he means to include the sun, moon,
and stars, or the souls of heroes. I call the Stoic god “god” without the
article like “God,” since this usage helps to communicate that the Stoic
god is a good and very powerful creator and is in those ways like the God
of Abrahamic faiths. But I give him a small g, to remind us that he is not
supposed to be that sort of transcendent, omnipotent God. I call him

For a somewhat different reading of the point of the staging in Books  and , see DeFilippo ()
–.
. Balbus and the Central Question 
“him” because that is his grammatical gender in Greek and Latin when the
Stoics refer to him as a “god” (deus, θεός). It reminds us that he is in many
ways a person. But it would also be helpful to think of him as an “it,” a
sexless and non-human animal.
I begin with a summary of some foundations in Stoic thought necessary
to follow the rest of the chapter. The Stoics were materialists or, more
accurately, corporealists. They held that there are only bodies, which is to
say objects in three dimensions that can act, be acted upon, or (as in almost
all cases) both. The most basic bodies, two spatially continuous prin-
ciples, exist eternally, inseparably, and in the same volume. One of the two
would lack all qualities if it were not acted upon and is called “matter.”
(DL .) The other acts on the matter, giving it qualities and forming
from it the elements, their compounds, and bodies of the familiar sort we
see around us, to include ourselves. Balbus speaks of the active principle
as a kind of heat or fire, moving of its own volition. (DND .) Balbus
and other Stoics think that there are many gods. But it is the active
principle that is most fundamentally the Stoic god. Not only is it the
mover of all nature, Natura itself, it is also rational as we are. Indeed Stoics
sometimes call the active principle λόγος, logos, “reason” in Greek (DL
.), the word Cicero translates with ratio (DND .). Balbus argues
that like us this active principle of nature must have a rational mind, a
“leading part” in Stoic terminology, that rules the rest of us lesser natures.
(DND .–) It follows, he says, that this Nature must be the best and
most worthy of power. Thus not only it is rational like us, but, better than
nearly all of us, it has perfect reason and is “wise.” (sapiens, DND .,
.–)
What does a Stoic like Balbus mean by “wisdom” or “perfect reason”?
This is easiest to understand in the human case. A human child is born
without reason, a brute. But her senses supply “impressions” to her mind
and she remembers these. As she ages her mind sorts her memories and
builds concepts. At some point in late childhood or early adolescence she
has a sufficient assembly of concepts to become rational. Now when
she has a sensory impression, or entertains a thought or utters a sentence,

Of course, speaking of a goddess, with feminine pronouns, would have achieved this goal just as
well, but that is not what the ancient Stoics did when they spoke specifically about their deus. They
did argue that some other, feminine nouns also referred to god, like natura, “nature” or εἱμαρμενή,
“fate.”

SVF .–, .–.

For the active principle as god, and its role in making everything else besides the material principle,
see Sedley (). Cf. Cooper () and his references, Lapidge ().
 Balbus the Stoic and Cotta the Skeptic
she does so in light of her relevant concepts and thus she sees or thinks
or speaks something meaningful. In Stoic terms she says something, a
“sayable” (λεκτόν), and only the rational can say anything. Now one sort
of thing we might say, an “assertible,” might be true or false. Another
consequence of her acquisition of reason is that our growing child becomes
interested in truth. She can assent to a thought or a sensory impression if
she thinks that what it says is true. But for the moment she can get this
wrong: her stock of concepts is imperfect enough that she may mistake a
true impression for a false one. This is where doing philosophy comes in.
The Stoics propose that we should sort through our concepts, our beliefs,
and the further evidence and arguments they supply, to acquire ever more
accurate concepts and well-founded true beliefs. Eventually, in the ideal
case, this assembly would become knowledge and we would thus have
become wise. Our true, harmonious, self-consistent knowledge would be
such that we would never again be fooled by a false impression. From then
on we would form only true beliefs.
The full importance to the Stoics of becoming wise emerges when we
see how they think we act responsibly and what things they think are
valuable. For the Stoics, to act is to form a certain sort of belief, of a form
something like, “it is appropriate now for me to do x.” They think that
when a person assents to belief of this sort, his mind releases an “impulse”
that makes his body move so that he does x. But recall that to form a
belief was to take one’s impression to say something true. If I make a
mistake about this then I might take to be true an impression that is false.
I might think, “although I am not hungry and it will damage my health, it
is appropriate for me now to take one more slice of this delicious cake,
even though Thomas has not yet had a slice.” By assenting to this false
belief, I have done something bad. It is intemperate (I damage myself for
superfluous pleasure) and unjust (I rob Thomas of his desert). But now
recall that the wise person, the “sage,” never assents to a false belief. Thus
she never does anything bad, but only what is in truth appropriate for her
to do. She would never be fooled by the cake. Therfore to be wise is also to
be virtuous: one will move through life never doing anything bad. But what
about those of us who are not yet wise? We are vulnerable to forming false
beliefs because there are inconsistencies and falsehoods among our existing

For the Stoics on reason and sayables, see LS –, , with Frede (), on wisdom, LS , ,
, .

For the Stoic theory of action see LS – with Inwood () –, Brennan ()
–, Brennan () Chapters  and . Cicero’s familiarity with this theory is apparent in
On fate, especially sections –, and Academica, especially .–, .–.
. Balbus and the Central Question 
concepts and beliefs. Thus we might do things that are not appropriate.
The Stoics are hard on this failing. There is no middle ground: they call all
of us who are not virtuous vicious. It is a moot point whether they thought
any human had ever perfected his reason so as to become virtuous. Either
all or almost all of us are vicious. But god is wise.
What, then, is divine wisdom like? For you might guess that it is quite
different from the human variety. But Balbus attributes wise psychology to
his god most explicitly in the following passage:
ipsius vero mundi, qui omnia conplexu suo coercet et continet, natura non
artificiosa solum sed plane artifex ab eodem Zenone dicitur, consultrix et
provida utilitatum oportunitatumque omnium. atque ut ceterae naturae suis
seminibus quaeque gignuntur augescunt continentur, sic natura mundi omnis
motus habet voluntarios, conatusque et adpetitiones, quas ὁρμὰς Graeci vocant,
et is consentaneas actiones sic adhibet ut nosmet ipsi qui animis movemur et
sensibus. Talis igitur mens mundi cum sit ob eamque causam vel prudentia vel
providentia appellari recte possit (Graece enim πρόνοια dicitur). . .
The nature of the cosmos itself, which both contains and holds together
everything in its embrace, is called by the same Zeno [the founder of
Stoicism] not only craftsmanlike, but actually a craftsman, caring and
provident for all that is useful and advantageous. And as other natures are
severally engendered, and grow, and are held together by their own seeds,
thus the nature of all the cosmos has voluntary motions, efforts, and
impulses, which the Greeks call hormai [impulses], and it directs consistent
actions at these in the way that we do who are moved by senses and by
souls. Therefore, since the mind of the cosmos is such [i.e. rational and
impulsive], for that reason it can rightly be called either ‘practical wisdom’
or ‘providence’, which is said pronoia [= providence] in Greek;. . . (DND
.–)
God, then, acts like an animal with a soul and sensation. Its body (that is
to say, the cosmos) moves as a result of impulses that it controls by
“consistent actions”, that is to say, by assents to impressions consistent
with its other rational endowment, its beliefs and actions. That it is wise
then allows Balbus to call it “practical wisdom” and “providence.” These
are terms for virtues the human sage might have. Thus the Stoic god has
the same sort of reason, action, and virtue as does the human sage.

Each of consentaneus and actio might have several meanings. But I take consentaneus in the sense of
“rationally consistent” (see On ends ., On duties .) and actio in the philosophical sense of a
Stoic “action,” see On fate  where Stoic actiones are distinguished from impulses and are grouped
with assents.

For prudentia as a rendering of φρόνησις see On duties .. For πρόνοια the Stoic virtue see SVF
..
 Balbus the Stoic and Cotta the Skeptic
As the final piece of the background, let us survey what things the Stoics
think are valuable. They maintain that only one thing is choiceworthy:
virtue. For it is this perfection of our reason that always leads us to do the
good thing, to be good ourselves, and to be happy. Conversely, it is vice,
the lack of virtue, that leads us to be bad and unhappy. Thus the Stoics
limit their list of good things to “virtue and what shares in virtue.” The
sage’s virtue is good, by having virtue the sage is good, by resulting from
her virtue her actions and feelings are good, and so on. Equivalently, fools
like ourselves are bad because our vice is bad, our actions are bad, our
feelings are bad, and so on. Thus the list of evils is “vice and what
participates in vice.” Everything else is neither good nor bad but rather
“indifferent.” This includes things that most people find to be of far from
indifferent value: life, death, family, social honor, health, wealth, and so
on. But when the sage acts she pursues or avoids such things, preferring
one to another. So the virtuous person chooses well among indifferents.
The criterion for this choice is what is natural, since the Stoics define the
happy life as “life in agreement with nature.” In most circumstances,
obviously, it will be most natural to pursue life over death. But sometimes,
in exceptional circumstances, we should pursue death over life, for example
if called on to sacrifice ourselves in just defence of others. For we are social
animals, say the Stoics, so it is natural for us to take other humans’ needs
into account. There is one further outwork of this view of value to
consider. The Stoics are most famous for their view that ordinary human
emotions, like fear, anger, or grief, are the result of vice and are to be
extirpated. For to be vulnerable to these emotions involves beliefs about
value that only a vicious person could have. If I fear loss, pain, or death
then I must believe that those things are bad, when they are not. They are
indifferent. A virtuous person could deal with them as such, secure in her
own goodness. The sage will not suffer anger, fear, grief and the rest, but
rather will feel untroubling feelings of her own, like joy.
With all this background in hand, let us turn to Balbus on the Central
Question. The Stoics’ cosmic god is not an omnipotent Creator in the
manner of the Abrahamic God. He does not make the physical world from
nothing. Rather, he finds himself joined to a material principle that is
eternal, just as he is. He gives all this matter all of its properties. It is in this
large but finite and inescapable physical realm that he can do what he likes.

Key texts for Stoic value theory are conveniently collected in LS Chapters  and .

For the Stoics on the emotions, see Cicero Tusculans  and , Frede (), Brennan () and
(), Graver ().
. Balbus and the Central Question 
We know one thing he has chosen to do in this circumstance: he made the
cosmos as we know it. Now the Stoics, notoriously, do not think that this
state of affairs is permanent. They think that the processes of circulation
and combustion by which our world and its stars are kept running lead the
universe, periodically, to burn up entirely. We call this the “conflagration.”
Once the world has become a pure, undifferentiated “flash” in this way,
god is left alone with matter. He begins, again, the process of making
elements, compounds and bodies so as to produce a world like ours afresh.
Since god always makes the best choices at the first go, he never experi-
ments with this process. He makes the world, from scratch, the same way
each time, so that it is like the one we know.
But why? Since we share her human nature, when we assess actions of
the Stoic sage we can see what things might suggest themselves to her as
natural to pursue in a given set of cirumstances. But god’s nature and
circumstances are quite different, most of all during the conflagration
when he must decide whether and how to organize the cosmos anew.
Why did he do as he did this time, and as he always does?
Balbus answers this question, in terms so straightforward that it is easy
to miss their significance. The answer comes in section [] of his speech,
“that the gods care about us,” where Balbus argues that the natural items
we use were in fact made for our use. Now in what follows it is important
that the Stoics used the word “cosmos” (κόσμος, mundus) with three
distinct meanings. One of these meanings is god the rational animal, the
eternal creator. A second meaning is the order of things the creator brings
about between conflagrations, the “cosmos” as the familiar structure of the
world today. “Cosmos” with the first of these meanings refers to the maker
of what it refers to with the second. A third meaning is as the “world” in
the sense of the community of all people, humans and gods.
Now let us turn to Balbus’ answer to why god made the world:
principio ipse mundus deorum hominumque causa factus est, quaeque in eo
sunt ea parata ad fructum hominum et inventa sunt. est enim mundus quasi
communis deorum atque hominum domus aut urbs utrorumque; soli enim
ratione utentes iure ac lege vivunt. ut igitur Athenas et Lacedaemonem Athe-
niensium Lacedaemoniorumque causa putandum est conditas esse, omniaque
quae sint in his urbibus eorum populorum recte esse dicuntur, sic quaecumque
sunt in omni mundo deorum atque hominum putanda sunt.

Balbus’ description of the conflagration is at DND .. My account of the physical process of the
conflagration here draws on Cooper (). For different Stoic views of the conflagration and its
place in god’s providential plan, see Mansfeld (), Salles () and (b).

For the various meanings of cosmos, see SVF .–.
 Balbus the Stoic and Cotta the Skeptic
In the first place, the cosmos itself was made for the sake of gods and
humans, and everything that is in it was prepared and devised for the
enjoyment of humans. For the cosmos is as it were a common home or
settlement of both gods and humans, for only they live by law and statute
using reason. Therefore just as one must think Athens or Sparta was
founded for the sake of the Athenians or the Spartans, and that everything
that is in these settlements rightly belongs to their peoples, so one must
think that whatever things are in all the cosmos belong to gods and to
humans. (DND .)
The “cosmos” here is made. It is distinguished from the cosmic commu-
nity of gods and humans who inhabit it. So it must be the cosmos in the
sense of the current order of things. God made and ordered these things,
quite simply, for our sake. By “us” here I mean rational beings, including
god himself, the other gods, and humanity. Furthermore, everything in it
was made for human use (as well as for the gods). Now I think that Balbus
means this assertion quite literally. For he immediately launches into a list
of things made for our use, which proceeds methodically from the heavens
inwards through air, sea, and land to the benefits stored up for us under
the ground. (DND .–). The heavens, for example, are a spectacle
and give us timekeeping and the seasons (DND .); the soul of the pig
has the same function in the cosmic order as salt can have, to keep pork
fresh until we want to eat it (DND .); only humans mine the
commodities under the ground (DND .). Just as musical instruments
must be made for those who can play them, all the bounty of agriculture is
made for those who can use it. (DND .–)
What does Balbus mean when he says that the cosmos is “as it were a
common home or settlement”? Applying my learned reader principle (see
p. ), that Balbus is talking to people who already know about Stoic
philosophy, and that Cicero wrote with some such readers in mind, it
helps to simulate that knowledge by looking at a parallel from a passage of
Stoic doctrines attributed by Eusebius to an Epitome of philosophical views
prepared by the first century BC philosopher Arius Didymus:

We might wonder whether bed-bugs, scorpions, and other natural irritants or threats were made for
our use. But Chrysippus said that lice are useful for waking us up and mice for making us tidy.
(Plutarch, Stoic Self-contradictions d = SVF .) We learn from Lactantius that “the
Stoics . . . say . . . that there are many plants and animals whose utility is so far hidden; but that
they will be discovered with the passing of time, just as today need and use have discovered many
things unknown in earlier ages.” (On the anger of God  = SVF .) So Balbus, too, would
probably say that if we think something is not useful, that is because we have not yet discovered
its use.
. Balbus and the Central Question 
ὃν γὰρ τρόπον πόλις λέγεται διχῶς τό τε οἰκητήριον καὶ τὸ ἐκ τῶν
ἐνοικούντων σὺν τοῖς πολίταις σύστημα, οὕτω καὶ ὁ κόσμος οἱονεὶ πόλις
ἐστὶν ἐκ θεῶν καὶ ἀνθρώπων συνεστῶσα, τῶν μὲν θεῶν τὴν ἡγεμονίαν
ἐχόντων, τῶν δ’ ἀνθρώπων ὑποτεταγμένων. κοινωνίαν δ’ ὑπάρχειν πρὸς
ἀλλήλους διὰ τὸ λόγου μετέχειν, ὅς ἐστι φύσει νόμος· τὰ δ’ ἄλλα πάντα
γεγονέναι τούτων ἕνεκα. οἷς ἀκολούθως νομιστέον προνοεῖν τῶν ἀνθρώ-
πων τὸν τὰ ὅλα διοικοῦντα θεόν, εὐεργετικὸν ὄντα καὶ χρηστὸν καὶ
φιλάνθρωπον δίκαιόν τε καὶ πάσας ἔχοντα τὰς ἀρετάς. διὸ δὴ καὶ Ζεὺς
λέγεται ὁ κόσμος, ἐπειδὴ τοῦ ζῆν αἴτιος ἡμῖν ἐστι.
For in the same way as ‘city’ has two meanings, the dwelling-place, and the
system resulting from the combination of residents and citizens, so also the
cosmos is, as it were, a city composed of gods and men, in which the gods
have the rule, and the men are subject. There is, however, a community
between them, because they partake of reason, which is nature’s law: and
for their sakes all other things have been made. From which things it
follows that we must believe that the god who administers the whole is
provident, being beneficent, and helpful, and friendly to man, and just, and
possessed of all virtues. For that reason, the cosmos is also called ‘Zeus’,
since it is the cause of our life (zēn). (Eusebius Preparation for the Gospel
.. = SVF ., translation from Gifford (), adapted)
Our author here argues that the “cosmos” which is the cause of our life can
be seen to be as a god that administers an appropriate habitation for the
“cosmos” in the sense of the “city” of gods and humans. Now in Stoic
political theory, all and only the wise are citizens of the cosmic city-state.
For only they understand and obey the laws, that is to say the demands of
reason, well enough to be fully “naturalized” members of the state. So only
the gods and the few, if any, wise human souls are in fact citizens. But all
of the rest of us who live under the demands of the reason, that is to say
under the same law, form a “community” along with the wise. Thus the
world as a whole contains a society like the United States, a community of
both citizens (the wise) and of resident and socially integrated aliens (the
rest of us). In this sense the cosmic community of all gods and humans is as
it were (οἱονεί, quasi) a city-state but is not strictly a city-state. It fails to be
a city-state not because it is not literally a community of inhabitants living
in the same place and under the same law – it is literally such a commu-
nity – but because many of the inhabitants are not citizens. This is why
Balbus, too, says that the cosmos is “as it were a common home or

DL . = SVF ., Stobaeus ..i = SVF ..

There is also a city-state in the strict sense, the city of the gods and any wise humans. See my ‘God’s
Indifferents’ (). For other views, see Schofield (), Obbink (), Vogt () Chapter .
 Balbus the Stoic and Cotta the Skeptic
settlement of both gods and humans, for only they live by law and statute
using reason.” (DND ., quoted pp. – above.) On this basis he
argues by analogy that the other bits of the cosmos must have been made
for us, just as we would say that the architecture and infrastructure of
Chicago was made for its people and not vice-versa. Thus: everything else
was made for us, the rational.
Now we might still wonder: why did god make the world? We have seen
his purpose in making the cosmos as he made it. It was for us. But now we
recall that as part of this same process of creation he made us, too, and all
the other gods. What is the point of us? Balbus has already told us this. It
comes in the course of an argument that the cosmos is perfect in that it
lacks nothing.
Scite enim Chrysippus, ut clipei causa involucrum vaginam autem gladii, sic
praeter mundum cetera omnia aliorum causa esse generata, ut eas fruges atque
fructus quos terra gignit animantium causa, animantes autem hominum, ut
ecum vehendi causa arandi bovem venandi et custodiendi canem; ipse autem
homo ortus est ad mundum contemplandum et imitandum—nullo modo
perfectus sed est quaedam particula perfecti.
For Chrysippus said shrewdly that, just as the cover is made for the sake of
the shield, but the sheath for the sake of the sword, in the same way
everything else (other than the cosmos) was generated for the sake of some
further thing, like the crops and the fruit that the earth produces for the
sake of animals, while animals are generated for the sake of humans, like the
horse for transport, the ox for ploughing, and the dog for guarding and for
hunting. But the human being himself arose to contemplate and to imitate
the cosmos. He is not at all complete, but he is a tiny part of the complete
cosmos. (DND .)
The important principle here is that “everything else besides the cosmos
was generated for the sake of some further thing.” It is clear that “cosmos”
in this sense is god himself, the creator, because we have seen in DND
. (pp. – above) that the cosmos in the sense of the created
order was made for the sake of other things, for gods and humans. Thus
now we learn that the creator did not come into being for the sake of
something else, but everything else did. This implies that humans were
created for the sake of some further thing. For the sake of what? “To
contemplate and to imitate the cosmos.” But “the cosmos” in this passage
has meant god himself, the creator. So god made us for his own sake, to
contemplate and to imitate him.
It is helpful to compare another statement from Chrysippus on the same
matter, this one transmitted by Cato the Stoic in Cicero’s On ends:
. Balbus and the Central Question 
praeclare enim Chrysippus, cetera nata esse hominum causa et deorum, eos
autem communitatis et societatis suae, ut bestiis homines uti ad utilitatem suam
possint sine iniuria.
Chrysippus made the famous remark that all other things were created for
the sake of humans and gods, but that humans and gods were created for
the sake of their own community and society; and so humans can use
animals for their own benefit with impunity. (On ends ., translation
from Annas and Wolfe ())
At first sight this looks like a different answer to the question of our
purpose: we were made for the sake of our community and society, not for
god’s own sake. But I think it is a different way of putting the same
point. For god is a member of our community and society, the society of
the rational. A way to play our part into this society is to use our reason,
or to try to use our reason, to imitate god. Many people, of course,
succeed in this project only marginally, or misconceive it in (for example)
the manner of the Epicureans. But by using reason they are, in fact,
trying to get things rationally correct and so, even if they do not know it,
in the Stoics’ eyes they strive to imitate god. Thus when Balbus says that
our purpose is to contemplate the cosmos, he refers to a part we have to
play in the cosmic society. For if we contemplate it successfully then we
shall realize that the cosmos was ordered by its own rational mind, and
we shall realize that we have social ties to this mind and to our fellow
rational creatures in general. I think that Cato has the same social role in
mind for us when he says that we were made for the society we share with
the gods.
The Stoic god knew what he was getting into when he planned to make
us. Very, very few, if any, humans become wise. Thus very few of us
succeed fully at the task of contemplating and imitating god. Thus very
few of us become citizens of the city of the wise. So god must be not only a
political animal, who values his citizenship in the city of the wise. He must
also be a social animal, in that he must value the mere society he shares
with us rational but vicious beings who have not achieved citizenship in
the cosmic city. In this sense, god cares about us. In the sense that he has
provided a world for us, he cares for us. This is Balbus’ general answer to
the Central Question.
Very well, god aims to shape the cosmos as our common home. But
what steps has he taken to this end? Balbus tells us in a passage immedi-
ately following his characterization of Zeno’s view of Nature as a crafts-
man (DND .–, quoted on p.  above). In acting as a virtuous
craftsman,
 Balbus the Stoic and Cotta the Skeptic
haec potissimum providet et in is maxime est occupata,
(a) primum ut mundus quam aptissimus sit ad permanendum,
(b) deinde ut nulla re egeat,
(c) maxume autem ut in eo eximia pulchritudo sit atque omnis ornatus.
[Nature] above all provides for the following things and is most taken up
with them
(a) first, that the cosmos should be as fit as possible to endure,
(b) then, that it should lack nothing,
(c) but most of all, that there should be in it exceeding beauty and every
adornment. (DND .)
Here we have an insight into god’s virtuous purposes in running the
world as it is: (a)–(c). But we have seen Balbus say also that the cosmos and
everything in it is made for the sake of gods and humans. Are these
statements in conflict? Not necessarily, and one way they might go
together is that god might have some particular set of amenities he wants
to make for our sake. In fact (a)–(c) are, I suggest, a list of the features god
is occupied with in providing the fabric he has built for his cosmic society
to live in.
In (a), the cosmos that god tries to make stable cannot be the cosmos in
the sense of the two principles that survive eternally. God, the active
principle, can change neither his own eternity, nor that of the passive
principle. (a) must refer to the cosmos in the sense of the current system of
the world which god has designed. God can only make it as fit as possible to
endure because eventually the conflagration will overtake it. (b) might
sound like a sort of metaphysical principle of plenitude: god sees to it that
everything which could come into being does so. But this does not seem to
be what god does. We could imagine many things that god could have
made, even with the limited resources available to the Stoic god, but which
he has not. So I think (b) is to be read in light of DND . (quoted above,
p. ). There we saw the web of purposes and dependences that god has
organized. God sees to it that every niche in this cosmic economy is filled
so that no other part of it suffers a lack. Herbivores have plants, god has

If god cares so much about the cosmos’ survival, why build in the conflagration? This is a vexed
question. Mansfeld () argues that Chrysippus thought the conflagration preferable to the
cosmic order between conflagrations. Far from the conflagration being a regrettable interruption
of our world order, our world order would thus be an instrument to bring about the conflagration.
But even if Mansfeld were right about Chrysippus, it is clear that Balbus, perhaps following
Cleanthes, thinks that our current world order is preferable to the conflagration (DND ., cf.
Salles , b).
. Balbus and the Central Question 
contemplators, and so on. Consideration of god’s purpose in (c) I leave to
sub-section .. below.
Why does god want to give us exactly these three advantages in our
cosmic home? Balbus does not go on to elaborate the point. But if we put
ourselves in the position of god the architectus of the cosmos as Balbus calls
him (DND .), god the architect or city planner, we can get some idea.
Suppose you were given fairly plentiful resources to design and to build a
house. What would you aim at in your design? You would want it not to
fall down, which equates to Balbus’ point (a). You would want it to have
all the rooms and amenities that its likely inhabitants would require,
together with the physical plant to sustain these – point (b). You would
want it to be attractive and meaningful, as in purpose (c).
If those aims seem important but arbitrarily chosen, it is instructive to
turn to Cicero’s younger contemporary, the architect Vitruvius. Vitruvius
lists exactly three chief concerns of an architect in building:
haec autem ita fieri debent, ut habeatur ratio firmitatis, utilitatis, venustatis.
(i) firmitatis erit habita ratio, cum fuerit fundamentorum ad solidum
depressio, quaque e materia, copiarum sine avaritia diligens electio;
(ii) utilitatis autem, <cum fuerit> emendata et sine inpeditione usus
locorum dispositio et ad regiones sui cuiusque generis apta et commoda
distributio;
(iii) venustatis vero, cum fuerit operis species grata et elegans membrorumque
commensus iustas habeat symmetriarum ratiocinationes.
And all these buildings must be executed in such a way as to take account of
durability (firmitas), utility (utilitas), and loveliness (venustas):
(i) Durability will be catered for when the foundations have been sunk
down to solid ground and the building materials carefully selected
from the available resources without cutting corners;
(ii) the requirements of utility will be satisfied when the organization of
the spaces is correct, with no obstacles to their use, and they are
suitably and conveniently orientated as each type requires.
(iii) Loveliness will be achieved when the appearance of a building is
pleasing and elegant and the commensurability (symmetriarum) of
its components is correctly related to the system of modules. (Vitru-
vius, On architecture .., translation from Schofield ())
Vitruvius’ three concerns of the human architect correspond to Balbus’

three divine purposes, and in the same order. As to (a) and (i), Balbus’ god
makes the cosmos as fit as possible to survive, Vitruvius gives his building
the foundation and materials it needs to stand firm. As to (b) and (ii),
 Balbus the Stoic and Cotta the Skeptic
Balbus’ god ensures the world lacks nothing it needs, Vitruvius arranges
and divides the space available on the site of the building for its use. As to
(c) and (iii), Balbus’ god aims at beauty, Vitruvius aims at loveliness.
Indeed Vitruvius’ theory of proportions (symmetriai, ..) is very close
to the Stoic definition of beauty, “the proportion of the parts to one
another and to the whole.” The coincidence between Vitruvius’ prin-
ciples of building and the goals of Balbus’ god is very suggestive. It does
not have to mean that there was any direct influence from Hellenistic
architectural theory into this aspect of Stoic theology or vice-versa,
although perhaps there was. But it shows that when Balbus gives priorities
to his creator for the housing of the cosmic community, he is thinking in
the way that an architect of his day might have thought about how to
accommodate his clients.
To conclude this sub-section, notice that Balbus’ answer to the Central
Question is starkly opposed to Velleius’ in the sentiments it implies are
pious for the worshipper. For Velleius, it was a consequence of the gods’
invulnerability and Epicurean virtue that they are passive and in no way
social. So of course they did not make, neither do they act in, our world.
That they would never wish to intervene in our lives is part of what makes
Velleius admire and want to worship them. But for Balbus, the virtuous
creator and his divine helpers have in mind first of all to make and to join a
society of the rational, and then to form everything else in the world for
the sake of those rational creatures. This virtuous creation and help is part
of what makes Balbus want to worship it and them. Velleius admires the
gods in part because they do not care for us, Balbus admires the gods in
part because they do.
.. Beauty and the Central Question

We saw in DND . (p.  above) that in his capacity as cosmic architect
god most of all pursues “beauty and every adornment.” In Vitruvius’

Cicero, Tusculans .; Galen, On the doctrines of Plato and Hippocrates, ..; Stobaeus ...

Some further perspectives on Stoic creationism and providence are the essays collected in Salles
(a), and Dragona-Monachou (), Frede (), and Sedley () –.

Another interpretation of this word, pulchritudo, would be that it is equivalent to the Greek Stoic
term τὸ καλόν, “the fine.” Τὸ καλόν is formed from the Greek adjective καλός which can mean
“beautiful” as well as “fine.” For the Stoics, “the fine” is another way to describe virtue or the good.
Thus you might think that what Balbus means in DND . is that god aims most of all to bring
about virtue and goodness. But I have argued elsewhere that this cannot be right (Wynne ()
–). In Stoic contexts Cicero translates τὸ καλόν as honestum (e.g. On ends . cf. Plutarch
On Stoic self-contradictions c; On ends .). It is in Stoic contexts where he renders the related
. Balbus and the Central Question 
terms, what god wants to provide for our home is “loveliness” above all else.
At first sight this is puzzling. The Stoics can seem dour – why would their
god put beauty before usefulness? Yet once your eye is open to it, you see
“beauty” (pulchritudo) everywhere in Balbus’ speech. It is one of the
signatures his creator gives to the world. To help us to understand why,
I shall look at two stages where beauty is key to Balbus’ arguments.
The first stage is early on in section [], “that there are gods.” Balbus has
argued that there is a universal consensus that there are gods (DND
.–) and proposes to conclude that we have a natural preconception
that this is so: “it is born into all people, and as it were carved into their
souls, that there are gods.” (omnibus. . . innatum est et in animo quasi
insculptum esse deos, DND .) For the Stoics, as for the Epicureans, a
natural preconception must be an accurate preconception, so if Balbus is
right then he would have proved to his own satisfaction there are gods.
But this poses a question: how do we naturally acquire this preconception?
We do not, of course, see gods around in the way we see dogs. We might
reply that, according to Balbus, we see nothing but the cosmos or its parts,
and it is a god. But we can imagine that Stoics would not wish to rest on
this point, for the divinity of the world around us is not obvious to every
casual observer, so that it would be implausible that simply from seeing the
world around us we get our concept of the divine. Further, unlike Velleius,
Balbus cannot appeal to an influx to the mind of true images of the gods
(cf. p.  above). Rather, he gives four causes of our preconception owed
to the Stoic Cleanthes. The last runs:
quartam causam esse eamque vel maximam aequabilitatem motus conversionem
caeli, solis lunae siderumque omnium distinctionem utilitatem pulchritudinem
ordinem, quarum rerum aspectus ipse satis indicaret non esse ea fortuita.
[Cleanthes says that] the fourth cause [of our concept of god], and that
perhaps the greatest, is the evenness, the motions and the rotation of the
heaven, the contrast of the sun and moon and all the stars, the utility,
beauty, order, the very sight of which (he said) shows sufficiently that these
things are not by chance. (DND .)
Greek noun τὸ κάλλος that he uses pulchritudo (Tusculans . cf. Stobaeus .. = SVF .;
DND . cf. Aëtius . = SVF .). Τὸ κάλλος is a Stoic indifferent, most obviously beauty as
we find it in the body (DL ., Plutarch, On common conceptions c, Sextus Empiricus,
Adversus mathematicos .). This latter is the sort of beauty, then, that Balbus’ god aims to
provide. For Stoic beauty and its relation with the fine, see also Bett (), and Graver ().

On Stoic preconceptions, see LS , Schofield (), Brittain ().
 Balbus the Stoic and Cotta the Skeptic
There follows a comparison with seeing a house or a gymnasium: you
would not see such a building and conclude that its rationally motivated
features could have happened without some designing intelligence. Simi-
larly, the visible qualities of the heavens, beauty among them, tell us that
they have some designer. But this is supposed to give us our concept of
god. Thus we are supposed to be able to see a rational “hand” behind the
sky even when we have not yet conceived of god. According to Balbus,
Cleanthes thought beauty was a clue to rational design available to those
who might not even have expected it.
The relation in which this beauty puts us to god is further explained a
few lines later. Balbus has moved from Cleanthes’ four causes of our
concept of god to Chrysippus’ related description of learning his first idea
of god from nature:
an vero, si domum magnam pulchramque videris, non possis adduci ut, etiam si
dominum non videas, muribus illam et mustelis aedificatam putes tantum ergo
ornatum mundi, tantam varietatem pulchritudinemque rerum caelestium,
tantam vim et magnitudinem maris atque terrarum si tuum ac non deorum
inmortalium domicilium putes, nonne plane desipere videare?
Or indeed, should you have seen a large and beautiful house, you could not
be induced to think it was built for the mice and the weasels—even if you
should not see its master. Therefore such great decoration of the cosmos,
such great differentiation and beauty of the heavenly realm, such great size
and energy of the sea and the land—if you should think that these are your
home and not the home of the immortal gods, surely you would be seen
simply to have lost your mind? (DND .)
The conclusion of this argument is potentially misleading. For it could
suggest that the cosmos is the home of the gods and not your home. But
this, as we have seen, is not Balbus’ view. So he must mean that the cosmos
is not only your home, but is also the home of the gods. This we are
supposed to conclude from the sheer scale of the beauty of the heavens,
and so on. Now this point, unlike Cleanthes’, relates beauty not to its artist
but to its appreciator. Do you think that humans like yourself can be the
only intended audience for beauty on the scale of the heavens? No, there
must be some other and greater beings who, like you, appreciate beauty:
those other rational animals, the gods.

Both Rackham () and Walsh () translate “by” the mice and weasels. But this destroys the
analogy. Balbus’ point is that the house is not built for mice and weasels, since it is built for humans.
Thus the cosmos is not built only for human beings, since it is also built for the gods.
. Balbus and the Central Question 
Between the two last quoted passages, we see that when beauty is made
specifically to be appreciated, it is made by the rational for the rational.
When found in nature it is a clue that somebody like us made our world,
and did so with observers like himself in mind.
We have already seen why god might have done this and why he might
want to give us the concept of his own nature. For he made us to
contemplate and to imitate him, so we had better be able to spot his work
and think about its maker. In ., as part of his argument that the gods
care about us, Balbus has just given us DND . (pp. – above).
Everything was made for the use of gods and humans. He begins his
catalog of human uses thus:
iam vero circumitus solis et lunae reliquorumque siderum, quamquam etiam ad
mundi cohaerentiam pertinent, tamen et spectaculum hominibus praebent;
nulla est enim insatiabilior species, nulla pulchrior et ad rationem sollertiamque
praestantior; eorum enim cursus dimetati maturitates temporum et varietates
mutationesque cognovimus. quae si hominibus solis nota sunt, hominum facta
esse causa iudicandum est.
Now in fact the circlings of the sun and moon and of the rest of the stars,
although they have to do with the cohesion of the cosmos, nevertheless they
also provide a spectacle for humans. For there is no sight that it is harder to
have enough of, none more beautiful or that stands out better to reason and
intelligence. For it is when we have measured the stars’ courses that we
know the ripenings, the differentiation and changes of the seasons. If these
things are observed (nota) only by humans, we must judge that they were
made for humans’ sake. (DND .)
Presumably the heavenly movements are not known only to humans. The
gods must know about them too. But what is unique to us is that we are
rational creatures with the sense of sight, who look from the perspective of
the Earth, around which the visible heavenly movements are made. So it is
the sight of the heavens, from our perspective at their centre, that is known
only to us. This, with its usefulness and above all beauty, must be intended
for us. Notice that, since Balbus has said that the heavenly bodies are self-
moving gods (DND .–), those lesser, created gods cooperate with
their creator in bestowing this vision on us.
Now I shall move to a second instance where beauty does special work
in Balbus’ argument. This comes in part [] of his speech, “that the cosmos
is governed by the gods.” At the start of part [], he advertises a subdivision
into three kinds of argument, of which “the second is one which teaches
that all things are controlled by a sentient nature and that everything is run
 Balbus the Stoic and Cotta the Skeptic
most beautifully by it.” (DND .) The arguments of this subdivision
duly appear at DND .–. Now these arguments, which Balbus calls
arguments from the “most beautiful” running of the world to the thesis
that the cosmos is governed by gods, look very like what we would call
arguments from design. Imagine you found a watch in a forest, a modern
creationist might say – surely you could not look at it and say that it came
about by chance? The intricacy of its mechanism implies design. But living
things are similarly intricate, she goes on to say, and therefore they are
designed. Balbus’ parallel examples include some clocks, but also the
course of a ship or a magnificent orrery built by Posidonius to model the
motions of the sun, the moon, and the five planets. (DND .–) But
Balbus’ general label for the property that cannot have come about by
chance (cf. DND . quoted p.  above) in these examples is not
intricacy or complexity, but beauty.
Balbus captures this outlook in a passage that contrasts his answer to the
Central Question with Velleius’:
hic ego non mirer esse quemquam qui sibi persuadeat corpora quaedam solida
atque individua vi et gravitate ferri mundumque effici ornatissimum et pul-
cherrimum ex eorum corporum concursione fortuita? hoc qui existimat fieri
potuisse, non intellego cur non idem putet, si innumerabiles unius et viginti
formae litterarum vel aureae vel qualeslibet aliquo coiciantur, posse ex is in
terram excussis annales Enni ut deinceps legi possint effici; quod nescio an ne in
uno quidem versu possit tantum valere fortuna.
Should I not now be amazed that there is anybody who can convince
himself that some solid and indivisible bodies are moved by force and
weight and that the cosmos, most decorated and most beautiful, is made
by chance collisions of these bodies? I don’t understand why somebody who
judges that that could have happened should not also think that, if
uncountably many images of the  letters, made of gold, or of whatever,
are gathered into some vessel, then one could make the Annals of Ennius,
such that the poem can be read in order, by pouring these letters out on the
ground. I doubt that luck could manage so much as one line. (DND .)
Balbus’ understanding of probability is unsatisfactory. He thinks that it
is not even possible that the Annals could appear if you poured out
innumerable Scrabble tiles. In fact, it is vanishingly unlikely, but it could
happen by chance. But all that is needed for the arguments of the
Epicureans with their infinitely large and old universe is that something
could happen by chance, because then somewhere out in infinity it will

secunda est autem quae docet omnes res subiectas esse naturae sentienti ab eaque omnia pulcherrime geri.
. Balbus and the Central Question 
have happened. But the probabilities are likely not where the force of
Balbus’ point is really supposed to lie. When you encounter a great work of
art like the Annals, the Aeneid of Cicero’s day, you would find it very hard
to imagine that it came about by chance. It has all the hallmarks of a
rational artist behind it. This is to say, for Balbus, that it is beautiful.
We can understand better the connection between beauty and reason in
Balbus’ views if we look at part of his argument against the Epicureans,
that the gods can be spherical, like the cosmos or the sun, because the
sphere is the most beautiful shape.
globus (sic enim σφαῖραν interpretari placet), ex planis autem circulus aut
orbis, qui κύκλος Graece dicitur, his duabus formis contingit solis ut omnes
earum partes sint inter se simillumae a medioque tantum absit extremum, quo
nihil fieri potest aptius—sed si haec non videtis, quia numquam eruditum
illum pulverem attigistis, ne hoc quidem physici intellegere potuistis, hanc
aequabilitatem motus constantiamque ordinum in alia figura non potuisse
servari?
But whereas two shapes are most outstanding, among solids the sphere (let
us translate sphaera thus), but among plane figures the circle or ring, which
is said kuklos in Greek, it occurs in only these two shapes that all their parts
are most similar to one another, and that the outer surface is a distance from
the centre that could not be made more fitting. But if you [Epicureans]
cannot see this, because you have never touched that educated dust, could
you not understand even this much physics, that this evenness of motion
and constancy of order could not have been preserved in any other form?
(DND .–)
The “educated dust” is the geometer’s sandbox. So what Balbus has just
said is supposed to be a point about geometry as understood by the
educated, albeit not a very complicated one. The parts of a circle, or of
the surface of a sphere, are all very similar to one another. These internal
relations can only be understand by someone with basic geometrical
training and therefore with reason. But that suggests that beautiful prop-
erties like sphericity can only be brought about by a craftsman with the
same faculty. Balbus thus describes the sphere and the circle as satisfying
what seems to be in other sources the Stoic definition of beauty, “the
proportion of the parts to one another and to the whole” (see p.  nn. 
and ). Now beauty for Balbus, as for the Stoics in general, does not
always seem to be a matter of quantitative or geometrical proportion.

Balbus’ general views on beauty, as well as of its role in our concept of the divine, find a very close
parallel in Aëtius . = SVF ..
 Balbus the Stoic and Cotta the Skeptic
Much of the beauty he describes in the cosmos is more a matter of fitting
together its web of purposes and dependences in a way that is more
qualititative and to do with elegant fitting of means and ends. But these
too, Balbus thinks, require a rational craftsman and, when we perceive
them in an object, they allow us to infer that the craftsman made it.
I shall return to Cicero’s experience of beauty and its importance for the
Central Question in my last chapter. For now, imagine the world through
Balbus’ eyes. A social being, he sees his fellow humans, and he sees gods
with whom he shares reason processing magnificently across the sky.
Everything else he sees was made to help him, or the rest of the human
race, or the gods. But besides what is useful, the world was made to be
beautiful. To be awed by the heaven at night is not to see beauty of a
different kind than we find in an art gallery. It is the same kind of beauty,
and is made by a good person (among other purposes) for the same
purpose as a painting, for us to see. The third of Balbus’ proofs that the
cosmos is run by gods is simply an expression of admiratio, “wonder,” as he
contemplates the “beauty” of what divine providence has made. (DND
., .) He runs on at this for fifty-six rapturous sections. (DND
.–) Faced by such overwhelming evidence of the hand of a caring
god, Balbus might wish to express his awe and gratitude in a meaningful
way. That is where religion comes in.
. Balbus’ Stoic View of Traditional Religion

Now that we are acquainted with Balbus’ theology, we can see that when it
comes to Roman religion he faces a puzzle that, put in the broadest way, is
not so different from what Velleius confronts: both of them can honor
some seemingly vital phenomena of the religion, but cannot honor others.
Velleius can do justice to the iconography and popular imagining of the
gods as blessed beings of human shape. But he cannot allow that religion is
done for gods who care. Balbus, meanwhile, can supply the latter, but his
cosmic god, while a blessed and virtuous person, is a very long way from
the Jupiter portrayed in cult statues or in myth. Even when Balbus begins
to elaborate his polytheism and introduces the heavenly bodies as gods, the
geometrically uplifting sun, moon, and stars seem rationalized gods
indeed. If Balbus is to use belief in gods like these to moderate Roman
religion, he must show how the religion’s traditional performances –
statues and temples, sacrifices and prayers – could be made piously by
people who know and embrace Stoic theology. It is naturally in part [] of
his speech (see p.  above), “what the gods are like,” that Balbus comes
. Balbus’ Stoic View of Traditional Religion 
most directly to grips with this question. In this section I shall examine
that second part of Balbus’ speech in detail.
Balbus’ discussion of “what the gods are like” amounts to a classification
of the gods. I shall say that he gives us gods of five types (which he
explicitly distinguishes from one another) on two levels (which he distin-
guishes only implicitly). I illustrate this classification in the table in
Appendix . There are gods of two types in the higher level, which
I shall call the paradigmatic gods. I call these paradigmatic because in the
theological arguments in the other parts of his speech, these are the gods
Balbus tends to consider. In the table I number Balbus’ types of god for
convenience. The only god of type  (the cosmos) and the gods of type 
(the heavenly bodies) are found in the higher, paradigmatic level. The
lower level I shall call the cultural gods because Balbus is most interested in
the origin of their place in human culture. Among these we find type ,
some gods named metonymically for their beneficial powers or the benefits
they bestow, type , deified heroes, and type , aspects of the natural
world. Balbus unreservedly argues for the view that the paradigmatic gods
are gods. When he comes to the cultural gods, while he accepts in one way
or another that they are divine, to varying degrees Balbus seeks to reform
existing beliefs about them. In particular, he thinks that existing beliefs
about the gods of type  have gone badly awry.
From the table we can get a first idea of Balbus’ approach to the
traditional gods. First note that – very loosely speaking – as we move
down the table, Balbus’ endorsement of the gods’ divinity becomes more
qualified, but the gods become more prominent in traditional religion.
This is an indication of the depth of the problem that Balbus faces.
Second, note that of the paradigmatic gods as Balbus describes them, only
two were worshipped at Rome (Sol and Luna, the sun and moon). By
contrast, all the gods at the cultural level were worshipped at Rome. Balbus
brings them up exactly for that reason. His task with the cultural gods is to
explain how they can be the proper objects of cult when they are not, or
seem not to be, his paradigmatic gods.
Velleius gave us a history in which everybody’s natural preconception of
the gods had suffered serious distortion, until Epicurus single-handedly

Brennan () - observes “two kinds of polytheism” in Stoicism, which describe
respectively “nominalist” deities (roughly equivalent to my cultural gods) and “planetary” ones
(equivalent to my type  gods). He asks whether this reflects different approaches in different Stoic
sources, or two approaches to polytheism that the same Stoic could take (p. ). Balbus
accommodates gods of both Brennan’s kinds, which shows that at least as Cicero understood it, a
Stoic could support both of these accounts of polytheism at once.
 Balbus the Stoic and Cotta the Skeptic
stepped forth to save us. For Velleius, traditional religion reflects this
serious, ancestral distortion. Velleius’ reconsideration of religion then
needed to be most radical: although he can accept features like the
human form and reputed blessedness of the traditional gods, to moderate
the tradition he must apply Epicurus’ entirely independent discoveries (see
Chapter  section .). As the table shows, Balbus also submits a cultural
history of religion. But note that in each of the three cultural levels, Balbus
says that the origin of the Roman gods in human culture was acceptable or
even good. This means that he thinks that there was truth in the ideas
behind the founding of Rome’s cults, and at least a kernel of truth implicit
in the orthopraxy or attendant myths that have been handed down. Balbus
thus can hope to show not only that Stoicism can bring to bear some
helpful new ideas from outside, but also that many Stoic truths are already
latent in the tradition. He is more a supporter of traditional religion as
traditional than is Velleius.
Let us then take Balbus’ classification of the gods type by type. At the
paradigmatic level, Balbus acknowledges that he faces something like the
reverse of Velleius’ problem:
restat ut qualis eorum natura sit consideremus; in quo nihil est difficilius quam
a consuetudine oculorum aciem mentis abducere. ea difficultas induxit et vulgo
imperitos et similes philosophos imperitorum, ut nisi figuris hominum constitutis
nihil possent de dis immortalibus cogitare; cuius opinionis levitas confutata a
Cotta non desiderat orationem meam.
It remains for us to consider what [the gods’] nature is like. On this topic,
nothing is harder than to draw the mind’s gaze away from what our eyes are
accustomed to. This difficulty has taken in both unsophisticated people
generally and philosophers who are like the unsophisticated, so that they are
unable to think at all about the immortal gods except in the form of human
figures. The ridiculousness of this opinion was refuted by Cotta, and it does
not call for a speech from me. (DND .)
As we have seen, it is important to Balbus that the gods are in some ways
like us, especially in their rational psychology. But his paradigmatic gods
are nothing like us in the shape or constitution of their bodies. Thus his
paradigmatic gods meet resistance in minds whose habits have been shaped
by life at Rome.
Now Balbus’ difficulty here is more pointed than he makes it sound. For
Balbus, like Velleius but on different grounds, accepts the view that
whatever natural preconceptions we might have are true. One sort of
evidence for a natural preconception is a universal consensus among
. Balbus’ Stoic View of Traditional Religion 
peoples, since such a consensus is more easily explained as the product of
human nature than of culture. Thus there follows Balbus’ problem. For
there was of course a widespread, albeit not universal, consensus in the
ancient Mediterranean that the gods could rightly be thought of in human
form. This is certainly the case for the gods of myth and poetry. With the
gods of cult, we cannot know how literally adherents in general took
anthropomorphic symbols. We recall Augustine’s report that Varro made
his religiously learned speaker, one of the Scaevolae, say that the gods of
myth were inimical to civic religion. (City of God .) So those in charge
of Rome’s religion probably did not think mythical representations of the
gods should be taken literally. Still, it is easy to imagine that performing
traditional religion and hearing its related stories fostered a tendency in
many to think of the gods in human shape – precisely the anthropomorph-
ism that Velleius welcomes as evidence for the truth. Thus Balbus must
convince his listeners that this particular tendency does not reflect a
natural, and therefore true, preconception of the gods. He must “draw
the mind’s gaze away from what our eyes are accustomed to.” His
approach is to argue that the gods of the paradigmatic level, which are
spherical and not in the least human-looking, are indeed gods.
Let us then look in more detail at Balbus’ arguments that the gods of
types  and  are gods. The only god of type  is the cosmos. Balbus
spends most of the relevant paragraphs (DND .–) mocking
Epicurus’ contention that a sphere, the shape of the Stoic cosmos, is not
the most beautiful solid (see .–, quoted on p.  above).
The mockery is not idle. For he considers not only the cosmos but also
the heavenly bodies spherical (., pulcherrima forma, “most beautiful
shape”). To establish the superior and therefore divine beauty of the sphere
over that of the human body is thus an important part of Balbus’ task in
drawing our minds away from accustomed images. Further, as Balbus
remarks, too much special argument here for the divinity of the cosmos
would be idle, since much of the rest of his speech serves that function.
Type  gods are the heavenly bodies, both the fixed stars and the seven
planets, the stars that we see “wander” against the background of their
fixed counterparts. The planets are the sun, the moon, and what we would
call today the five visible planets, Mercury, Venus, Mars, Jupiter, and
Saturn. The five planets, which we call by the names of the gods, are
known to Balbus as the “star of Mercury” (stella Mercurii) and so on.

On arguments from consensus, see Schofield (), Brittain ().
 Balbus the Stoic and Cotta the Skeptic
(DND .–) No identification of the god and the heavenly object is
implied merely by the name.
Balbus’ first step in showing the divinity of all these bodies is to describe
the amazing regularity and precision of their motions. (DND .–) In
the case of the fixed stars, this does not take much arguing. Of course, they
go around once every sidereal day. For the various planets there is more
work to do. The sun moves around (as the Stoics saw it) once per solar day,
while moving against the fixed stars along the ecliptic over the course of
 / days (.), producing the year, the course of the ecliptic rising
and dropping in the zodiac to produce the seasons. The moon travels a
similar path against the fixed stars monthly, and has its phases (DND
.). Both of these sets of movements are not only marvelously regular,
but have direct benefits for us – timekeeping for one thing, but also the
terrestrial manifestations of the seasons, the growth of crops, and so forth.
With the other five planets, showing regularity is harder again. Their
movements against the fixed stars appear much less regular than those of
the sun or moon. But Balbus reminds us that this is merely an appearance.
Greek mathematical astronomy had long ago shown that although the five
planets’ motions were more complex, they were as regular and predictable
as the moon’s and the sun’s. Balbus says that the five planets have “two
continual revolutions.” (continuas conversiones duas, DND .) He means,
I suppose, that we can model their apparently irregular motions as the sum
of two sorts of regular motion. Which two such motions he has in mind is
not clear. Perhaps he means that the apparent motion of each planet can be
explained as the sum of, first, the motion of the heavens as a whole and,
second, the planet’s own movements around the ecliptic. Ancient astron-
omers could model each planet’s motion mathematically, so that moving
according to the required geometry was a feat that suggested that the
source of the planet’s motion, too, understood mathematics. Further-
more, Balbus then reminds us that even these complex movements are
cyclical and return to where they began to start again. To underline the
point, he gives us approximate observational data for the periods of
the planets’ movements, Saturn the longest at thirty solar years, Mercury
the shortest at a single solar year. (DND .–)

The talk of two motions must call to mind the motions of the Same and the Different in Plato’s
Timaeus, see e–b, translated by Cicero in what we call his Timaeus, sections –. But
perhaps Balbus has in mind some sort of further mathematical model of the motions in terms of the
sum of the revolutions of more than one sphere, like that going back to Eudoxus and Calippus. For
the tradition of celestial mechanics, and the Stoics’ apparent lack of interest in it, see Jones ()
–.
. Balbus’ Stoic View of Traditional Religion 
All this regularity is very well. For Balbus it indicates that there is reason
behind the motions of the heavenly bodies. But it is not enough to get
Balbus his conclusion that the heavenly bodies themselves are gods. Why
should we not think that the reason behind these motions is just that of the
cosmos itself? That an orrery must be designed does not show that its
model planets themselves have reason. Balbus’ extra premise here is
revealed in his description of the fixed stars. (DND .–) He says that
the stars are not moved in virtue of being fixed to a rotating sphere of
aether or the like, but rather that they move under their own power. The
same presumably goes for the sun, moon, and other planets. This would
make the celestial bodies, including the fixed stars, self-moving parts of the
cosmos much as we are. Thus the reason and intelligence evident in their
motions must, in the first instance, be their own. In order to move
according to the regular rules we can discern only with our reason, they
themselves must understand these rules. Thus Balbus triumphantly
reaches his conclusion in DND .. He is satisfied that he has given us
a multitude of gods (as many as there are fixed stars in the sky + seven
planets + the cosmos) that fit our preconceptions of the gods, but are not
of human bodily shape. These are what I call his paradigmatic gods.
Where Balbus’ task at the paradigmatic level was to turn our mind’s eye
away from conventional representations of the gods to his own, celestial
deities, at the cultural level his task is to account for the worship of the
traditional gods. His exchange with Cotta further shows that not just any
explanation will do. Balbus must show how these gods and their cults are
such that Stoic theology will support rather than undermine them.
To begin with the gods of type , Balbus introduces them as follows:
multae autem aliae naturae deorum ex magnis beneficiis eorum non sine causa
et a Graeciae sapientissimis et a maioribus nostris constitutae nominataeque
sunt. quicquid enim magnam utilitatem generi adferret humano, id non sine
divina bonitate erga homines fieri arbitrabantur.
But the wisest men of Greece, and our own ancestors, established and
named many other divine natures, and not without cause. For the wise
thought that whatever brought great utility to the human race did not come
about without divine goodness towards humans. (DND .)
He sums them up thus:

utilitatum igitur magnitudine constituti sunt ei di qui utilitates quasque
gignebant, atque is quidem nominibus quae paulo ante dicta sunt quae vis sit
in quoque declaratur deo.
 Balbus the Stoic and Cotta the Skeptic
Thus the gods who brought about each thing of use were established as
gods for the greatness of these useful gifts. Indeed, what power there was in
each god is made plain by the names we have just spoken about. (DND
.)
These gods, then, are classified not by their intrinsic properties, but by
their origin in human culture. They were wisely or traditionally established
as gods at some time in Greek and Roman history, and they were named.
They were picked out by the benefit they bestow, and named after their
power to bestow it. Nor is this history mere conjecture. In the case of his
Roman examples, Balbus says that some of the relevant cults were started
by known founders at various dates in the historical period. (DND .)
Balbus gives those who established these gods a lot of credit. Various
Roman forebears set them up, which for Cicero is always a good thing, but
for Balbus the Stoic the bigger praise is probably that the “wisest men of
Greece” did the same. It is likely that he does not mean “wisest” in the
technical Stoic sense, since the Stoics do not think there have been many
sages. But the term is still a compliment. Further, Balbus gives some degree
of approval to these founders’ rationale: it was “not without cause.” The
rationale was that great benefits did not come about without divine
goodness towards us. We know why Balbus approves of this sentiment.
It is his view that a good god made everything else in the cosmos for us,
gods and humans. So he thinks that to set up cults to this divine goodness
is an appropriate response. At the same time, and whatever he may think
about the wisest men of Greece, Balbus almost certainly does not think
that eminent Roman ancestors were sages. It is their foundations he will
discuss directly. So he will not expect perfect consistency or correctness in
the rationale behind these foundations.
With these type- gods Balbus is interested in how they came to be
named. This is natural, since their names are transparently meaningful:
“Good Faith,” “Abundance,” and so on. Thus it is obvious that there must
have been some rationale in their naming. Balbus distinguishes three items
in his analysis of this rationale: the benefit bestowed (utilitas) or “thing
itself’” (res ipsa), the god (deus), and the power by which the benefit is
bestowed or controlled (vis). He describes three sorts of naming that
happened:

For the supporting evidence – much of it from Livy – see Pease (–) ad loc. Whether or not
this history is right, it is plausible that Cicero thought it was.
. Balbus’ Stoic View of Traditional Religion 
[i] itaque tum illud quod erat a deo natum nomine ipsius dei nuncupabant,
ut cum fruges Cererem appellamus vinum autem Liberum. . .
[ii] tum autem res ipsa, in qua vis inest maior aliqua, sic appellatur ut ea ipsa
vis nominetur deus, ut Fides ut Mens . . .
[iii] quid Opis quid Salutis quid Concordiae Libertatis Victoriae; quarum
omnium rerum quia vis erat tanta ut sine deo regi non posset, ipsa res
deorum nomen optinuit.
[i] Thus sometimes they used to dub that which was born from the god
with the name of the god himself, as when we call grain ‘Ceres’ but
wine ‘Liber’ . . .
[ii] but sometimes the thing itself, in which there is some some greater
power, is so called that the power itself is named a god, like Fides
(“Good Faith”) or like Mens (“Mind”). . .
[iii] What about Ops (“Abundance”), what about Salus (“Security”). . .?
Because the power of these things was so great that they could not
be controlled without a god, the thing itself supplied the name of
the gods. (DND .–)
(The text continues with a problematic list of gods to which I shall return
to on p. .) [i] and [iii] are a neat pair: in [i] the benefit is named after the
god, in [iii] the god is named after the benefit. [ii] is more complex: the
god is named after the power in the benefit. DND . (as quoted pp.
–) clarifies that this power is also in the god.
As I read DND .–, [i] does not describe a cultural origin for a kind
of god. Ceres and Liber must already have been thought gods, and so
named, for Balbus’ example to work. They will appear again in the list of
type  gods and seem at home there. I am not aware of any cult that
worshipped grain or wine simply as such. So I think that [i] is an illustra-
tion of the sort of metonymical naming that Balbus says happened in [ii]
and [iii]. It is meant to make the metonymies of [ii] and [iii] plausible by
appeal to a similar idiom in regular use.
Why does Balbus complicate matters by giving distinct naming histories
in [ii] and [iii]? I suggest this is because all the examples in [ii] can by
construed as referring to divine rational virtue: Fides and Mens as above,
but also Virtus (“Virtue”) and Honor (“Honor”). But all the examples in
[iii] can be construed, for the Stoics, as mere indifferents: Ops and Salus
but also Concordia (“Concord”), Libertas (“Liberty”), and Victoria (“Vic-
tory”). If that is right, it explains why Balbus’ emphasis in [iii] is on the
need for divine control over indifferents (which could help or harm) while

One could construe “concord” as a reference to the inner concord of a virtuous soul, “liberty” to the
true freedom of the sage, “security” to the sage’s inability to be harmed, or “abundance” to the sage’s
 Balbus the Stoic and Cotta the Skeptic
in [ii] he emphasizes the divine power of the virtues, which can only help.
But in any case, various Romans named the god they “established” or
“consecrated” metonymically, after the benefit the god bestowed or the
virtue with the god bestowed it. Note that as Balbus sees it, these ancestors
were not so far from expressing his own answer to the Central Question:
the gods’ virtue leads them to bestow benefits on the human race. It is
striking that Balbus construes the motivation of the consummate Roman
leaders who founded these cults, the likes of Marcus Aemilius Scaurus and
Marcus Claudius Marcellus, as gratitude on behalf of the human race at
large, not of the city of Rome in particular. (DND .)
There is an ambiguity in what Balbus says the ancestors were up to. On
one reading, he says that “Fides,” “Ops,” and so on were not really names
for gods. Rather, they were names for benefits or beneficial powers for
which divine goodness was thought a necessary condition. DND .
(quoted on p. ) seems to point in this direction. The cult of, say,
Fides, would then have functioned indirectly as worship of the gods who
have or bestow the “named” benefit, good faith. But on another reading,
Balbus does think that these names referred to gods, but that they do so via
some metonymical figure. When you say “Fides” you figuratively refer to
the cosmic god, perhaps, or to all the gods, who have or bestow the virtue
of good faith. DND . (quoted on pp. –) seems to point in this
direction. Perhaps, given the flexibility with which Balbus applies his
models of naming in this passage, the ancestors sometimes intended one
option and sometimes the other, or did not see the distinction. After all,
Balbus is not committed to the systematicity or clarity of the historical
decisions he describes here, since the Roman ancestors were not sages. But
Balbus, seeing the large measure of laudable insight in the cults handed
down, can calibrate his own beliefs about what each performance means
quite precisely. Let us look at some evidence for how he does so.
Later in his speech, Balbus will confirm that he understands these type-
“gods” as properties or benefits of the gods, rather than independent gods
in themselves:
sequitur ut eadem sit in is quae humano in genere ratio, eadem veritas
utrobique sit eademque lex, quae est recti praeceptio pravique depulsio. ex
quo intellegitur prudentiam quoque et mentem a deis ad homines pervenisse
(ob eamque causam maiorum institutis Mens Fides Virtus Concordia consecra-
tae et publice dedicatae sunt; quae qui convenit penes deos esse negare, cum
true wealth, namely virtue. But Balbus’ description of these things, as needing a god to rule them,
points in the other direction.
. Balbus’ Stoic View of Traditional Religion 
eorum augusta et sancta simulacra veneremur: quod si inest in hominum genere
mens fides virtus concordia, unde haec in terram nisi ab superis defluere
potuerunt?). . .
It follows that there is the same reason in [the gods] as there is in the human
race, there is the same truth in both, and the same law, which is the
injunction of what is right and the expulsion of what is immoral, whence
we understand that wisdom and intelligence came to people from the gods.
And it was for that reason, through the institutions of our ancestors, that
Mens (‘Mind’), Fides (‘Good Faith’), Virtus (‘Virtue’), and Concordia
(‘Concord’) were enshrined and dedicated by the state. How is it appropri-
ate to deny that these things are in the possession of gods, when we worship
their solemn and holy images? Because if there is intelligence, good faith,
virtue, and concord in the human race, from where did these flow to earth if
not from those above? (DND .)
In this passage, where Balbus has returned to more abstractly physical

arguments, he reserves the word “gods” for the beings who have, and give
us, good faith, virtue, and so on. The institution of the cults of Mens,
Fides, and the rest, he offers as evidence that the ancestors thought the
gods have the same reason and law as we do. This argument requires that
the “mind” and “good faith” we refer to when we worship are not
independent gods, but human-style virtues of the gods. Presumably,
Balbus has it in mind that the gods “above” who possess these goods,
and give them to us, are the paradigmatic gods of types  and , the
cosmos and the heavenly bodies (as well, perhaps, as any gods of type  as
there may be).
Presumably, then, the following is what Balbus thinks a Roman sage
would tell himself as he approached a ceremony in honor of (for example)
Fides. “There is no god called Fides distinct from other gods, like the
cosmos or the heavenly bodies. This cult was set up by my ancestors in
honor of the virtue of good faith, as evoked by the name of its ‘god.’ But
good faith is a virtue of the gods, and one on which my fellow Romans and
I depend. My own good faith is a gift from the gods, who gave me my
reason, and conditions that allowed me to perfect it. Thus understood, my
performance will be a pious action in favor of the gods.”
Before we leave the type- gods, I would like to return to the problem-
atic list of gods I set aside, which appears at DND .: Cupido (“Eros” or
“Lust”), Voluptas (“Pleasure”), and Venus Lubentina (“Pleasing Venus”).
This list needs careful handling. For Balbus seems to imply criticism of
these gods’ divine status. If so, and even if only implicit, this would be the
only criticism of an aspect of traditional orthopraxy anywhere in DND or
 Balbus the Stoic and Cotta the Skeptic
Div. For Balbus presents these as established gods, with “consecrated”
names, but then seems to imply that they might not really be gods, virtues,
or benefits of the gods. If so, they should not have been consecrated in the
way that they were. Now Balbus’ hostility to these particular gods is
puzzling. What seems to unite the list is a connection with pleasure,
especially sexual pleasure, and the desire for it. But the Stoics have no
problem with pleasure, or with sex enjoyed in the right circumstances, and
they think that there can be right circumstances. The sage is notoriously
“erotic,” that is to say, interested in romantic or sexual relationships. Not
only this, but Zeno, founder of the Stoa, made Cupido’s Greek equivalent
Eros a, or the, patron deity of his ideal republic. So it is surprising that
these gods are a source of concern for Balbus. He could easily have put
them in the same category as “Abundance”: preferred indifferents to be
controlled by virtue. But instead he says that their names “are consecrated,
the terms for vicious and unnatural things – although Velleius reckons
otherwise, nevertheless these very vices often assault nature quite
fiercely.” (DND .) So Balbus refers Cupido and Venus Lubentina
not to erotic desire in general, but to vicious lust in particular. Voluptas he
must think evokes a vicious, passionate pleasure.
That Balbus intends some criticism of the cult of these gods is hard to
avoid. For how could a Stoic accept the worship of vices, or of the
consequences of vice? Even Marcus in Cicero’s Laws, not a Stoic, expels
the cults of vices from his idealized Rome. (Laws .) Nevertheless,
Balbus does not make any criticism explicit. On the contrary, he classifies
these gods so that it is possible to see how they might fit into his scheme.
He says they are of the kind described in my division [iii] of DND .
(p.  above). We might interpret as follows. Urges contrary to virtue try
to overthrow the natural order of things: they “assault nature quite
fiercely.” But divinely ordained nature can fight back in the form of our
reason. Thus the virtuously rational divine powers are again to be thanked
for our power, in principle, to control these lusts. In this way these vices or
vicious phenomena are like “Abundance” and so forth: without divine
virtue they could not be controlled. Thus understood, the performance of

See Schofield (), Nussbaum ().

See Schofield () Chapter , Boys-Stones ().

Reading naturam. The whole quotation is, vocabula consecrata sunt, vitiosarum rerum neque
naturalium – quamquam Velleius aliter existimat, sed tamen ea ipsa vitia naturam vehementius saepe
pulsant.

He will soon use the word cupiditates, related to Cupido, to refer to vicious lusts which virtuous gods
do not have. (.)
. Balbus’ Stoic View of Traditional Religion 
their cults might be useful to Stoics in promoting self-control. If Balbus
thinks that Stoic ethics can accommodate these gods it must be along some
such lines. But it is tempting to think that he would eliminate them if he
could. That he does not say so is evidence that he is doing his best to
accommodate everything that is traditional in Roman religion.
Let us turn now to the gods of type . These we might call deified
heroes: divinities who were once mortals, but who have achieved some sort
of apotheosis.
suscepit autem vita hominum consuetudoque communis ut beneficiis excellentis
viros in caelum fama ac voluntate tollerent. hinc Hercules hinc Castor et Pollux
hinc Aesculapius hinc Liber etiam . . . hinc etiam Romulum, quem quidam
eundem esse Quirinum putant. quorum cum remanerent animi atque aeterni-
tate fruerentur, rite di sunt habiti, cum et optimi essent et aeterni.
But [the conventions of] human life and common habit have allowed
people to raise to the heaven men who have excelled in the benefits [they
have given], thanks to their reputation and to people’s gratitude. Hence
Hercules, hence Castor and Pollux, hence Aesculapius, hence Liber too . . .
[Balbus explains that this Liber, the son of Semele, is not the same as the
Liber in type ] . . . hence even Romulus, whom some consider to be the
same as Quirinus. Since their souls would persist and enjoy eternity, they
were duly taken to be gods, since they were the best and eternal. (DND
.)
Balbus once again approves of the cultural process he describes. He says
that the heroes were taken to be gods rite (‘duly’ in my translation), a word
that is ambiguous between religious or secular procedural correctness. He
says that the heroes’ souls enjoy persistence after death and are “the best,”
two divine qualities. Balbus endorses the divinity and continued existence
of the type  gods.
Just as the type- gods had to do with benefits to humanity that require
divine goodness, so type- gods have excelled in benefits. Hercules’ best
known contribution today is that he rid the world of some monsters in his
labors. But he also loomed large as a prototypical beneficent hero, victori-
ously facing down danger for the general good. Such a figure appealed to
the Stoics. In On ends Cicero has the Stoic Cato say that Hercules, along
with Liber, had an impulse to “save the human race,” and was someone

When we consider these type- gods, we should remember that when his daughter Tullia died, only
a few months before he wrote DND, Cicero made an earnest effort to build a shrine, a fanum, for
her, in order to achieve an “apotheosis.” (Letters to Atticus ..=SB)

For some philosophical uses of Hercules, see On ends .–; On duties .; Seneca, De
constantia ..; Epictetus, Dissertations .–.
 Balbus the Stoic and Cotta the Skeptic
with the resources and strength to do so. (On ends .) Now Hercules
was involved in the prehistory of Rome. His story, which was to become
prominent in the Augustan authors, was as victor over the bandit or
monster Cacus, who lived on or near the future site of Rome. As our
sources (all later than Cicero) tell it, the area was then inhabited by, among
others, some Arcadian Greeks, whose king, Evander, helped to set up the
altar and cult of Hercules that had persisted down to classical times at the
Ara Maxima, the Great Altar to Hercules. This particular cult of Hercu-
les at Rome, then, was according to the myth a commemoration of a
specific benefit done to people who would welcome Aeneas to Italy. But
Balbus elects to emphasize not Hercules’ contribution to Rome’s forebears,
but rather the response of humanity in general to such generous heroes.
Aesculapius, Greek Asclepius, was a great healer. Castor and Pollux help
soldiers and sailors. Liber, son of Semele, is Dionysus as depicted, for
example, in Euripides’ Bacchae. He gave humanity wine. Romulus, of
course, founded Rome – although Balbus may well not accept that he
should be identified with the god Quirinus. If not, he agrees with Scipio in
Cicero’s Republic that Romulus, although a remarkable enough man to
prompt stories of apotheosis, died a natural death. (Republic .–)
Although he accepts them, Balbus handles the type- gods gingerly. In
their legends, more unites them than he lets on. They were all thought, in
some version of their respective myths, to be the sons of mortal mothers by
divine fathers. Sometimes, bad divine behavior ensued. Most notably it
was the anger of the jealous Hera that led to Hercules’ labors. Balbus can
admit none of these stories literally understood. As we shall see, Balbus’
type- gods, who include the gods of myth like Juno, the Roman equiva-
lent of Hera, do not walk among us in the shape of humans or other
animals, do not have genitals, do not seduce or rape women, and do not
suffer jealousy or anger. Not only this, but the deeds of the heroes
themselves were questionable at times. For example, Hercules murdered
his own children in place of Eurystheus’ in Euripides’ Hercules furens. This
was a traditional philosophical example of unreason, indeed of insanity.
(Academica ., Sextus Empiricus, Adversus mathematicos .–.)
Again, Romulus’ homicide of his own twin, Remus, was an action we
might call, at best, open to various interpretations. So how is Balbus to
deal with the worship of heroes to whom such stories are attached? One
option is what Balbus will do for gods of type . He could interpret, for

Dionysius of Halicarnassus, Roman antiquities .–, Virgil, Aeneid .–, Ovid, Fasti
.–, Livy ., Propertius ..
. Balbus’ Stoic View of Traditional Religion 
example, Hercules’ story as an allegory for the victorious triumph of divine
virtue over laborious challenges. Thus the humanity of Hercules and with
it all the troubling aspects of the myth could be explained away. This is
what Cornutus, another Stoic interpreter of the traditional gods, would
later do. (Cornutus, Theology – Lang) But Balbus chooses to accept
that type- gods were indeed mortals who became gods. He must therefore
winnow out the chaff of scandalous myths, but accept literally the kernel of
these stories, that these men once lived, “excelled in benefits,” and are now
“the best and eternal.”
Balbus’ acceptance of the gods of type  reflects an accepted but now
rather obscure Stoic doctrine. The most helpful report is in DL: “And
[the Stoics] say that there also exist some daimones who have a sympathy
with human beings, and are overseers of human affairs; and the surviving
souls of virtuous men are heroes.” (.) Daimones – whence
“demons” – had often appeared in the philosophical tradition as beings
intermediate between gods and humans. So it is not strange that the
Stoics had their own version of this view. Meanwhile the souls of sages,
presumably because they are physically as well as epistemically more
coherent than foolish souls, survive for a longer time the death of the rest
of the body. They live on as “heroes,” former mortals who have graduated
to a semi-divine life. Even before death, the difference between the
virtuous reason of a sage and the virtuous reason of a full-blown god is
one of size, not of kind, so that the difference in kind between a human
sage and a god is one only of body and mortality. (Posidonius fr.  EK,
Seneca Letter ) Souls with that kind of divine status that also “enjoy
eternity,” as Balbus puts it, might rightly be called gods.
Thus Balbus makes no in-principle objection to mortals taking the best
of their own deceased to be gods. Yet his endorsement is lukewarm. “The

The evidence is collected at SVF .–. The reports besides DL’s lump the Stoics in with
Plato, Pythagoras, Thales, and so on. Nevertheless, they establish that Chrysippus held this sort of
view. See also Brennan () –, with his sources.

Φασὶ δ’ εἶναι καί τινας δαίμονας ἀνθρώπων συμπάθειαν ἔχοντας, ἐπόπτας τῶν ἀνθρωπείων
πραγμάτων· καὶ ἥρωας τὰς ὑπολελειμμένας τῶν σπουδαίων ψυχάς. Translation from Inwood
and Gerson (), modified.

E.g. Plato, Apology d, Symposium –; Epinomis –; Aristotle, Divination in dreams
b–.

When Balbus calls these heroes “eternal” he exaggerates: like the other gods besides the cosmos
itself, they will burn in the conflagration. So “eternal” must be used loosely, to mean “very long-
lived,” perhaps with some implication of the close relationship between the hero’s soul and the truly
eternal, divine principle. The latter might be what Balbus has in mind when he says heroic human
souls “enjoy eternity” – they enjoy the eternity of the truly immortal divine principle with which
they are in touch.
 Balbus the Stoic and Cotta the Skeptic
wisest men of Greece” introduced gods of type , but type- is a product of
convention. That suggests that a Stoic would want to tread more carefully
and see if these cults may involve one in vicious ideas. Balbus in fact
proposes conditions on which the endorsement of a type- god would be
proper. The person in question must be the best (i.e. a sage) and eternal in
soul. Since the Stoics have very high standards for sagehood, Balbus quietly
leaves it open to a Stoic to deny on further analysis that the heroes in
question were sages.
Nevertheless, it is best to read Balbus’ own endorsement of the trad-
itional gods he puts in type  as genuine. After all, he says that the
examples he mentioned are the best and eternal. Thus the Roman sage
who accepts one of these cults will think to herself as she approaches a
ritual performance for Hercules, “This cult is in honor of a man who was a
sage, who thus had a divine soul which now persists, who bestowed and
bestows benefits. Thus understood, my performance will be a pious action
in favor of this god.”
Last in Balbus’ classification, we turn to type- gods. In this category we
find what are clearly the leading gods of myth (or at least of the Greek
myth so influential in Roman literary culture), and are in many respects
the leading gods of Roman cult, especially of public cult as overseen by
Cotta and the pontifices – Jupiter, Juno, Minerva, Vesta, the Penates,
Janus, and so on.
Balbus gives easily his most complex cultural history to these gods. He
proposes two influences on how they came to be. First, they were charac-
terized by a physical argument or method. (ex ratione physica, DND .)
The physics on which the gods were based he considers to have been
sound. He calls its discoveries “well and usefully discovered.” (bene atque
utiliter inventis, DND .) Rather than with proper names the discoverers
of these physical ideas, or religious innovators informed by them, desig-
nated the type- gods with descriptive terms or phrases. These terms and
phrases, as we shall see, either described aspects of the divine as rightly
understood in physics, or aspects of a desirable sort of worship.
But these welcome developments were overtaken by a second influence.
The entities the physicists had described were “dressed in human appear-
ance.” (induti specie humana, DND .) This must mean that they were
represented in human shape. Thus they became fodder for the poets, who
distorted the gods’ nature still further in stories where the gods live and

For further consideration of the question of in what sense the Stoics thought human sages had
become heroes or divine, see Brouwer (), pp. –, –.
. Balbus’ Stoic View of Traditional Religion 
behave like ordinary humans. (DND ., .). Meanwhile, as is evident
in classical Latin, the meaning of the descriptive terms or phrases for the
gods was obscured or lost altogether. They came to seem to be arbitrary
designations for each god, in other words, proper names of the sort that
humans are given. The objects of worship identified and described in the
light of respectable physics had become the gods of myth, familiar from
Homer and the other poets. (DND .). Some of these myths became
impious. (DND .)
Balbus says that the consequences of this shift were dire:
videtisne igitur ut a physicis rebus bene atque utiliter inventis tracta ratio sit ad
commenticios et fictos deos. quae res genuit falsas opiniones erroresque turbu-
lentos et superstitiones paene aniles. et formae enim nobis deorum et aetates et
vestitus ornatusque noti sunt, genera praeterea coniugia cognationes, omniaque
traducta ad similitudinem inbecillitatis humanae. nam et perturbatis animis
inducuntur: accepimus enim deorum cupiditates aegritudines iracundias; nec
vero, ut fabulae ferunt, bellis proeliisque caruerunt, nec solum ut apud Home-
rum cum duo exercitus contrarios alii dei ex alia parte defenderent, sed etiam ut
cum Titanis ut cum Gigantibus sua propria bella gesserunt. haec et dicuntur et
creduntur stultissime et plena sunt futtilitatis summaeque levitatis.
You see how, from physical facts whose discovery was good and useful, was
wrenched a plan for confected and fictional gods. This state of affairs gave
rise to false opinions, distressing errors, and superstitions almost like those
of an old woman. Even the gods’ figures are known to us, their ages and
costumes and accessories, not to mention their family trees, their couplings
and siblings, and everything travestied into a likeness of human weakness.
For they are even brought before us with their souls in passion: we are
presented with the lusts, the distress, the rages of the gods. Nor indeed, so
the myths allege, did the gods lack for wars and battles, and not merely as in
Homer where different gods pick a different side among two opposing
armies to defend, but they even waged their very own wars with Titans and
with Giants! These things are said, and they are trusted, with great stupid-
ity. They are full of fatuity and the greatest inanity. (DND .)
We can see why Balbus might think that all these myths are impious.
They present the gods as less divine than they are, in human shape and
prey to the passions that (according to the Stoics) result only from vice.
Not only do the gods help defensively in human wars – which at a stretch
Balbus might be able to accept as a sign of their care for us – they fight as a
body against their creatures, the Titans and Giants. These are not Balbus’
just gods in community with all the rational. We can further see why
Balbus says such stories give rise to superstition. These would be gods for
us to fear, prone to anger, lust, and weakness. The sort of excessive
 Balbus the Stoic and Cotta the Skeptic
placation he will call superstition in DND .– (see pp. –
below) would seem justified if we thought the gods were like that.
In some ways, these gods of myth should not trouble a Stoic adherent of
Roman religion. Many of their faults are “known” from poetry, much of it
not Roman. The performer of religious actions is no more required to
believe in this than in any other theology. Strictly speaking, then, ortho-
practic Roman religiosity has little to do with the myths. But for Balbus,
the problem is that the myths have not remained culturally idle. They have
filled the gap left as an older, physical understanding of the gods was
forgotten. In some ways, as in cult statues or the use of what have become
meaningless proper names, religion even tends to suggest the myths over
other interpretations of its own meaning. Thus, thanks to the cultural
importance of the myths, for somebody in first-century BC Rome, or
indeed for readers of DND today, it might not be obvious what a Stoic
who says “Jupiter” might mean – if Jupiter is not a bearded figure throwing
thunderbolts, then what is he?
Balbus shows how Stoics might undo some of the distortions in the
myths by a program of allegoresis and etymology. In allegoresis he treats a
scandalous myth as allegory for a respectable bit of physics. In etymology
he gives a history of the name of each type- god. Now allegoresis and
etymology were common Stoic activities. There survives a compendious
parallel for Balbus’ program in the Theology of Lucius Annaeus Cornutus, a
Stoic of the century after Cicero. But Stoics in general, including the
older Stoics, tried to re-read names and myths in general, and poetry in
particular, in light of their own theories. Scholars have reconstructed their
reasons for doing so as follows. The Stoics thought that early people, in
the state of nature, lacked the ill social influences that we are under. They
were therefore able to get a clearer grasp of nature than are we. This
“wisdom” came to be encoded in myths, either put into allegories by early
people themselves, or through distortion as Balbus describes. Old, clear
concepts were preserved in the origins of words. Thus the Stoics, while
deploring the confusions, went looking to recover this deposit of primitive
insight. Now I think that this picture of Stoic allegoresis and etymology in
general is correct. But I am not sure that it is what Balbus in particular is

For Stoic allegoresis see Most (), and Boys-Stones (a), (b) Chapters –.

For Cornutus, see Most (), Boys-Stones (b) –, Ramelli ().

For Stoic debates on human origins see Boys-Stones (b) Chapter . For Stoic etymology see
Allen () and Long ().The key texts for the Stoic theory of etymology are Origen, Contra
Celsum . = SVF .; Augustine, De dialectica .
. Balbus’ Stoic View of Traditional Religion 
up to. For I am not sure that the chronology works, and I am sure that if
Balbus is trying to recover lost wisdom, he is bad at it.
To take chronology first, we ask, when does Balbus think the two
influences on the type  gods – correct physics and unwelcome distortion –
were exerted? He thinks that the two processes overlapped. This is visible
in his etymology of Neptune, as follows. Jupiter and Juno he thinks stand
for the sky, the joint realm of (respectively) the aetherial heaven and the
air. (DND .–) “There remained water and earth, with the result that
the myths divided up three kingdoms.” (.) So Balbus envisages that,
first, Jupiter and Juno were described according to physics, and second,
myths arrived, perhaps where Jupiter had two brother kings. In reaction to
the myths, two more portions of the divine were given descriptive names:
Neptune (connected to “swimming”) and Dis (connected to the “riches”
the earth offers). Thus myths and scientific naming were able to interact.
So when did these processes overlap? Balbus does not give a date. One
has the sense that he regards the specific events he reconstructs as imme-
morial. But he does not seem to set them in an utterly remote past. In his
etymologies, he assumes that the gods were dubbed by people who spoke
some comprehensible predecessor of Latin. But he thinks they were in
touch with speakers of Greek who had similar gods. (DND .–) For,
first, he thinks the names can be explained from familiar Latin words that
sound only a bit different: e.g. Diana from dies (“day”) and nox (“night”),
Janus from eundo (“going”), Jupiter from iuvans pater (“helpful father”) or
Jove from iuvare (“to help”). Second, he thinks that some names were
derived or just borrowed from Greek: Proserpina derived from Greek
Persephone, Vesta from Hestia, Apollo just borrowed. A period when
speakers of early Latin were in touch with similar speakers of Greek sounds
more like the legendary time of Latinus, Evander, or Aeneas familiar from
sources written soon after Cicero, Virgil’s Aeneid or the beginning of Livy’s
history, than it does the time of the first emergence of humanity. Again,
Balbus says that the physical gods, dressed in human shape, provided
material to the poets. Which poets? There were poets before Homer, but
Homer is the poet he names. (DND .) In Cicero’s writings people date
Homer to about a century and half before the foundation of Rome,
meaning they think Homer flourished about  BC. (Cicero, Republic
.–; Tusculans .) So the process of mythical distortion was still
ongoing not many centuries before Rome and its religion began. Finally,
we might ask, does Balbus credit the tradition that Romulus and his

aqua restabat et terra, ut essent ex fabulis tria regna divisa.
 Balbus the Stoic and Cotta the Skeptic
successor as king, Numa Pompilius, had an important part to play in
formulating Rome’s practices? (cf. DND .) He never says so, but if so,
his treatment of the type- gods implies that some physics reached the first
kings relatively undistorted. On the whole, the formative period of the
type- gods sounds like the centuries from the age of the Homeric heroes
down to the first decades of Rome.
If I am right about Balbus’ chronology, it is still possible that he looks
for the most primitive “wisdom” in his allegories and etymologies. Perhaps
he thinks the world began, or that humanity was in a state of nature, not so
long before the age of the Homeric heroes. Perhaps he thinks primitive
“wisdom” was passed down clearly for a long time and distorted only his
last millennium. But it seems more likely that he thinks the (admittedly
mysterious) old physicists who existed in this period were more like the
Stoics, making discoveries by philosophy or something like it, and not that
they were ingenuous primitives. In Div., Quintus will say that the
Babylonians have collected astrological observations for , years.
(Div. .) Cicero knows that this is ridiculous (cf. Div. ., p. 
below) but it suggests that his Stoics trace humanity’s history back much
further than the period Balbus appears to describe. If so, then Balbus does
not seem to look for the most primitive “wisdom,” but rather to show us
that Roman religion from the start was designed with more or less correct
physics in mind.
After chronology, my second point is that if Balbus were aiming to
excavate ancient truths for our benefit, he is not very good at it. His
allegoresis, which he attributes to Zeno, Cleanthes, and Chrysippus, runs
as follows: the myth is that Saturn, Greek Kronos, castrated his father
Caelum (“Heaven,” Greek Uranus), and was in turn bound by his son
Jupiter, Greek Zeus. Balbus thinks that the castration of Uranus is an
allegorical way of saying that divine aether produces without genitals. He
says that Saturn or Kronos represents time, which contains the movements
of the aether, and that Jupiter represents limitation of the courses of
heavenly, aetherial bodies into regular motions. (DND .) Now it seems
to me that Balbus’ reading of the myth goes by very quickly and that it is
obscure in some respects. In particular, it would take much more explan-
ation to say what correct physical theory Balbus wants us to think is
enclosed in the allegory as a whole. His point about the castration,

Perhaps Cicero imagines that these philosophers gave rise to the myths of Atlas, Prometheus, or
Cepheus, whose myths he interprets at Tusculans . as distorted descriptions of prowess at
cosmology.
. Balbus’ Stoic View of Traditional Religion 
I suppose, is clear enough. The rest has something to do with the
productive relationship between aether, time, and the structured heavens,
all three of which are accorded some divine status. But what the nature of
the productive relationship or the nature of the divine status might be is
opaque. So I think Balbus’ aim here is not to excavate some ancient physics
for our edification. Rather, his aim is merely to tell us that allegoresis of the
myth in terms of physical claims acceptable to a Stoic is possible. If learned
readers want to learn more, let them look in Zeno, Cleanthes, and
Chrysippus.
If Balbus’ aim were the recovery of the encoded physics, his etymologies
would also pose some puzzles. Often they reveal nothing more than that
Ceres, for example, goddess of fertility and agriculture, “bears crops” (from
gerere, to bear, DND .). These descriptions of the gods are consistent
with Balbus’ position, but they are so vague that they would be consistent
with many positions. Nor do they reveal much news of their own. With
Apollo and Proserpina, Balbus points out that a god’s name is borrowed or
derived from Greek, but stops there, without etymologizing the Greek
names in turn. To trace every god’s name all the way back to its physical
meaning does not seem to be the goal. At times, Balbus is more interested
in arguing that the names of the type- Roman gods were derived from earlier
terms than he is in the specific theories that originally produced those
terms. He certainly wants his audience to think that all the names could be
traced to the ideas of the old physicists, but in some cases he limits himself
to showing the first step or two of the trace.
In sum, Balbus applies the techniques of etymology and allegoresis not
primarily to reveal the physical principles that stand behind the type 
gods. Rather, he demonstrates that techniques are available that, when
applied more thoroughly than Balbus has time for, could show that the
myths and the names were derived in a complex process from some earlier
set of ideas. Not only that, but this set of ideas is such that a Stoic can

Is this particular Stoic use of etymology new with Balbus? I suspect not. Take this report about
Antipater’s etymology of the Syrian goddess Atargatis: “And yet Antipater of Tarsus, the Stoic, in
the fourth book of [his] On superstition, says that it is said by some that Gatis, the queen of the
Syrians, was such a gastronome that she ordered that ‘without Gatis (ater Gatidos, ἄτερ Γάτιδος) no
fish was to be eaten;’ and he says that through ignorance the many named her Atargatis, but stayed
away from fish.” καίτοι γε Ἀντίπατρος ὁ Ταρσεὺς ὁ ἀπὸ τῆς στοᾶς ἐν τετάρτῳ περὶ δεισιδαιμονίας
λέγεσθαί φησι πρός τινων ὅτι Γάτις ἡ τῶν Σύρων βασίλισσα οὕτως ἦν ὀψοφάγος ὥστε κηρῦξαι
ἄτερ Γάτιδος μηδένα ἰχθὺν ἐσθίειν· ὑπ’ ἀγνοίας δὲ τοὺς πολλοὺς αὐτὴν μὲν Ἀταργάτιν ὀνομάζειν,
ἰχθύων δὲ ἀπέχεσθαι. (Athenaeus .a = Antipater fr.  in SVF vol. .) Notice two things. First,
“the many” mistake a meaningful phrase for a meaningless proper name. Second, the context for
Antipater’s story was his On superstition. So perhaps for Antipater this story illustrated the sort of
naming mistake people made about Balbus’ type- gods.
 Balbus the Stoic and Cotta the Skeptic
plausibly argue that they were the same as the Stoic view of the world. So
Balbus does not have to reject the type- gods, or their traditional worship,
at all. Far from it: they were his gods and his worship all along. Stoics are
the very people to moderate Roman religion, because people like them
advised on its design in the first place.
So then, how does Balbus think the designers of Roman religion made it
a system to worship Stoic-style gods? In the first place, he understands
many of the type- gods as, in origin, references to his paradigmatic gods.
He says
sed tamen is fabulis spretis ac repudiatis deus pertinens per naturam cuiusque
rei, per terras Ceres per maria Neptunus alii per alia, poterunt intellegi qui
qualesque sint quoque eos nomine consuetudo nuncupaverit.
But now, once these myths are condemned and rejected, we will be able
understand who the god is who permeates the nature of each thing, Ceres
through the earth, Neptune through the sea, other gods through other
things, what they are like, and by what name custom has called them.
(DND .)
Compare Ceres and Neptune thus understood with what Balbus said
about the cosmic nature, rational and equipped with sensation, at DND
.–. He argued that the four “parts” of the cosmos, earth, water, air,
and the ether, are hot, and thus pervaded by the fiery cosmic nature.
(DND .–) “But we see that in the parts of the cosmos (for there is
nothing in the cosmos that is not part of the whole) are sense and
reason” (DND .). In ., then, Balbus has detected this same theory
of the cosmic nature behind some of the type- gods. Neptune is the part
of the cosmic nature in the sea, Ceres that part in the earth when
considered as the source of fertile harvests. To extrapolate: Juno is the
part in the air (DND .–); Dis, Greek Hades, is also the part in the
earth, considered as that from which everything terrestrial is made and into
which it decays (DND .); Jupiter is that part in the aetherial heaven
(DND .–). Since Balbus thinks that the heaven is where the mind of
cosmic nature as a whole resides (DND .–), Jupiter would seem to
be, as is consistent with his commanding role in myth, either the type-
 god himself, or the “leading part” (the Stoic ἡγεμονικόν) of the type-
god, its mind. Perhaps Balbus thinks that each of the other parts of the
cosmos is equipped with its own reason, an agent like the heavenly bodies.

videmus autem in partibus mundi (nihil enim est in omni mundo quod non pars universi sit) inesse
sensum atque rationem.
. Balbus’ Stoic View of Traditional Religion 
But perhaps he thinks each is a way to refer to that part of cosmic god that
is in and controls each respective part.
In the second place, we move to traditional gods that Balbus identifies
with type- gods, the heavenly bodies. By etymologizing Diana from “day”
and “night,” Balbus concludes that she was was originally the moon, which
makes the day as though night. (DND .) Her epithet Omnivaga (“All-
wandering”) conveys that she is one of the seven vagantes, the “wanderers”
or planets. Apollo, he asserts, is the sun. (DND .) Thus these two gods
are revealed to be type- gods after all, and cult doublets of Sol and Luna.
Thus understood, Roman religion allows Balbus to worship his
paradigmatic gods.
Balbus concludes part [] of his speech, “what the gods are like,” as
follows.
quos deos et venerari et colere debemus. cultus autem deorum est optimus
idemque castissimus atque sanctissimus plenissimusque pietatis, ut eos semper
pura integra incorrupta et mente et voce veneremur. non enim philosophi solum
verum etiam maiores nostri superstitionem a religione separaverunt. nam qui
totos dies precabantur et immolabant, ut sibi sui liberi superstites essent, super-
stitiosi sunt appellati, quod nomen patuit postea latius; qui autem omnia quae
ad cultum deorum pertinerent diligenter retractarent et tamquam relegerent,
sunt dicti religiosi ex relegendo, <tamquam> elegantes ex eligendo, <ex>
diligendo diligentes, ex intellegendo intellegentes; his enim in verbis omnibus
inest vis legendi eadem quae in religioso. ita factum est in superstitioso et
religioso alterum vitii nomen alterum laudis. Ac mihi videor satis et esse deos
et quales essent ostendisse.
These gods [i.e. Ceres, Neptune, and other type  gods] we must both
worship and pay cult to. But cult of the gods is best and least polluted, and
again most holy and most full of piety, when we worship always with both
mind and voice pure, faultless, and untainted. For not only philosophers
but even our own ancestors distinguished superstition from religion. For
people who spent whole days praying and sacrificing so that their children
would survive them (sibi superstites essent) were called ‘superstitious,’ a term
that later got a wider application. But people who carefully went back over
everything to do with the cult of the gods and as it were ‘re-read’ (re-
legerent) it, they were called ‘religious,’ as elegans from e-ligere (‘well-chosen’
from ‘to pick out’), diligens from di-ligere (‘careful’ from ‘to love’), intelle-
gens from intel-legere (‘intelligent’ from ‘to understand’). In all of these

Balbus’ intention here is to make it plausible that religio and its cognates are derived from a
compound of legere by mentioning other, more obvious compounds of legere. In addition to the
phonetic comparison, he probably intends us to see some of the senses of legere like “to gather” or
“to pick” as part of the original meanings of verbs like intel-legere, so that part of the original
 Balbus the Stoic and Cotta the Skeptic
words there is the same force of legere (‘to gather’, ‘pick’, or ‘read’) as there is
in religiosus (‘religious’). Thus it came about that between ‘superstitous’ and
‘religious,’ one is a term for vice, the other a term of praise. And it seems to
me that I have shown sufficiently that there are gods and what they are like.
(DND .–)
This passage is full of Cicero’s terminology from his preface (see Chap-
ter  and Appendix ). In DND .– (quoted on pp. – above) Cicero
wondered how to worship the gods “with purity and without pollution”
(pure atque caste), how piety and holiness could be secured. Now he has
Balbus tell us how the cult of the gods is “least polluted” (castissimus), done
with a pure (pura) mind, most holy and full of piety. The answer is that
such virtuous performance follows an intellectual activity, the “re-reading”
or “re-gathering” of traditional cult. This is the sort of interpretation that
Balbus has been engaged in. By it, the Stoic can come to believe correctly
that traditional practices were designed to be performed for the paradig-
matic gods themselves, in honor of their virtues and their gifts, or for the
divine souls of human sages. Thus her rational and linguistic faculty is
purified, with more accurate beliefs and concepts, and she moves closer to a
perfect virtue of piety. She will not fear the angry, lustful gods of myth, thus
she will have avoided superstition. She will accept that the gods she
worships care for her, her fellow citizens, and people in general, so she will
have avoided impiety. In Cicero’s terms, this is moderate religion.
But how should she perform in their honor? Some traditional perform-
ances make easy sense. For example, speaking in Jupiter’s honor seems
quite rational, since the cosmic god is supposed to have sensation in all his
parts and must, then, notice what we say to him. But Balbus’ paradigmatic
gods do not benefit in any obvious way from an elaborate animal sacrifice.
Might the cosmic god (for example) not think the Romans had mistaken
his nature when they sacrificed?
But Balbus thinks that sacrifice was written into the religion from the
start. Speaking of some of its original authors, he says:
cumque in omnibus rebus vim haberent maximam prima et extrema, principem
in sacrificando Ianum esse voluerunt, quod ab eundo nomen est ductum, ex quo
transitiones perviae iani foresque in liminibus profanarum aedium ianuae
nominantur. nam Vestae nomen a Graecis (ea est enim quae ab illis Ἑστία
dicitur); vis autem eius ad aras et focos pertinet, itaque in ea dea, quod est
rerum custos intimarum, omnis et precatio et sacrificatio extrema est. nec longe
meaning of “to understand” is “to choose between” or “discriminate”, or di-ligere so that part of the
original “to love” is “to pick out.”
. Balbus’ Stoic View of Traditional Religion 
absunt ab hac vi di Penates, sive a penu ducto nomine (est enim omne quo
vescuntur homines penus) sive ab eo quod penitus insident; ex quo etiam
penetrales a poetis vocantur.
And since in all matters the first and the last things have the greatest force,
they wished the leader in sacrifice to be Janus, because his name is taken
from eundo [‘going’], from which name shortcut passage-ways are named
‘Januses,’ and the doors in the entrance to ordinary houses are called ianuae.
For Vesta’s name is from the Greeks, since she is the one whom they call
Hestia [‘hearth’]. But the force of the name applies to altars and hearths,
thus in this goddess is the end of prayer and sacrifice, because she is the
guardian of matters internal. Nor are the ‘Penates’ far from this same force,
whether their name is from penus [‘larder’] since the ‘larder’ is everything
which people eat, or whether from that which resides ‘deep inside’ [penitus],
for which reason the poets call them penetrales (‘the ones deep inside’).
(DND .–)
To catch Balbus’ meaning, it helps to compare his treatment of Vesta
with Cornutus’ treatment of her Greek analogue Hestia. (Cornutus –
Lang) One meaning Cornutus gives to Hestia is that of the divine fire that
is the sustaining principle mixed into all parts of everything, into which
everything will (roughly speaking) burn up, and by which everything will
be made again at the conflagration. This is why the hearth is the centre of
the home, he says, because, like the divine principle in the cosmos, the
hearth houses the fire inside that sustains. Greek sacrifices begin and end
with Hestia, Cornutus continues, in acknowledgment of the divine fire’s
role at the beginning and end of the cosmos. Thus we see that Balbus, too,
wants Vesta to be the fire inside, the ultimate end of prayer and sacrifice, or
the Penates to be the inward part which feeds us all. He too gestures at
Vesta and the Penates as symbols of the all-pervading divine principle, and
thus proper objects of cult.
But it is over-subtle to think that such symbolism is Balbus’ only
objective in this passage. His more obvious purpose is to refer the names
Janus, Vesta, and Penates to a model for traditional Roman sacrifice. As
Balbus’ flexible model acknowledges, it is hard to say what a “typical”
Roman sacrifice was like even in one historical period, since sacrifices
ranged in scale and kind from big public processions to private occasions,
in honor of many idiosyncratic gods, with many items sacrificed. Let us
take one reconstruction of larger and more public animal sacrifices.

Here I follow Scheid in Rüpke () –. Scheid’s view of sacrifice emerged from his work
on the Arval Brethren, see Scheid () –. For questions about Scheid’s approach see
 Balbus the Stoic and Cotta the Skeptic
A procession entered the templum, the plot of land consecrated to the god.
On this plot would typically be built the god’s “house” with an altar in
front of it. There would be an initial libation and burning of incense.
Other gifts to the god might be given. The animal would be “immolated,”
which is to say covered with sacred meal, and then killed. Its innards
would be inspected to see if the god had accepted it. If so, the god’s
portion was symbolically handed over, for example by burning on the
altar. The remainder of the animal was then declared ready for mortal
consumption and eaten in a feast. The god was considered to be a fellow
participant at this feast. Balbus’ model captures the movement of this
process. Janus was given first place because he symbolizes movement, the
procession over the boundary into the templum. Sure enough, there is
evidence that Janus was given the first honors even at sacrifices or prayers
to other gods. Vesta is the last because she is the hearth or altar, the place
where the god’s portion is given over, at what you might think is the
culmination of the sacrifice. Or, as in a private house or the exceptional
temple of Vesta where the altar was not outside, Vesta is “inside,” away
from the worshippers in the god’s aedes at a temple or in the private space
reached by crossing into an ordinary house. The Penates are either the
place inside the house or aedes, or the “larder” from which everyone feeds,
the feast. By these double meanings – divine fire and the terms of the
ritual – Balbus assimilates sacrifice as a symbol of thanks to the divine fire
within, which gives us plenty. The designers of the religion made divine
patrons for the parts of the ritual to indicate its divine acceptance. To
speculate, it suits Balbus that some sacrifices culminated in a feast along-
side the god – it symbolized the bountiful gods’ place with us in the society
of the rational. In sum, in this analysis of the role of Janus, Vesta, and the
Penates in sacrifice, we see Balbus gesture at how the rituals of the religion,
and not just the names and depictions of its gods, might be re-read to
express views the Stoics would call pious.
My last point on Balbus’ etymologies is that they are plausibly Cicero’s
own work. It is possible that some of the many Stoics who had contact
with Romans had done for the Roman gods what their predecessors had
done for the Greek. Meanwhile, Varro had done non-Stoic etymology of
the same names. But Cicero is the pioneer of Stoic thought in Latin, so it is
plausible that Balbus’ etymologies are Cicero’s own. By contrasting them
Ando’s  review of Scheid (). For a collection of evidence on sacrifice see Wissowa ()
–.

Ovid, Fasti .–, Livy .; Wissowa () –.
. Balbus’ Stoic View of Traditional Religion 
with Cornutus’ equivalents, Stoic but Greek, and Varro’s, Latin but not
Stoic, we can pick out a theme. To begin from Jupiter, Balbus could have
looked for an equivalent to the Greek Zeus, “the father of gods and men”
(πατὴρ. . . θεῶν καὶ ἀνθρώπων, Cornutus  Lang) from “to live” (zēn,
ζῆν), as in Eusebius Preparation for the Gospel .. (see p.  above)
and Cornutus  Lang, because he is alive everywhere and causes the other
living things. Perhaps Varro’s Dies-piter, “day” or “sky father” could have
helped, linking Balbus’ heavenly Jupiter with the creative role of the
cosmic god (On the Latin language .–). But Balbus goes another
way: iuvans pater, “helpful father.” (DND .–) “Helpful father”
exactly captures Balbus’ theology. The cosmic god made us and the
type- gods to be like himself in respect of reason: he is our “father.”
But then he went on to help us: he makes and sustains the rest of the
cosmos for our sake. In the same way, Balbus goes on to emphasize divine
beneficence in many of his gods. He says that the Greek Stoics identify
Juno with the air. (ΗΡΑ from ΑΗΡ, i.e. (H)ēra from aēr, Cornutus 
Lang.) But for his part, Balbus derives her Latin name, like Jupiter, from
iuvare, to help (this time in agreement with Varro, On the Latin language
.). Dis, alternatively named Pluto, may in both languages be derived
from “wealth” (dives, πλοῦτος). Accordingly Cornutus says that Hades,
lord of the underworld or “earth” in Balbus’ estimation, is called Pluto,
“wealthy,” because all perishable things die and thus become his possession
in the end. ( Lang) But that is perhaps too depressing for Balbus. He
explains the etymology, “because everything both dies into the earth and
arises from the earth.” (quia recidunt omnia in terras et oriuntur e terris,
DND .) Thus Balbus makes Dis generous with his wealth rather than
simply host to the dead. Varro agrees with the cyclical interpretation, but
does not mention wealth. (On the Latin language .) These examples
show that by choosing among etymologies like those we find in Varro or
Cornutus, or by pioneering his own, Cicero not only makes Balbus’ type-
gods Stoic. He also hones them to his own purpose by making them
indicative of gods who care about us. This beneficence is the same quality
we saw Balbus emphasize in the gods of types  and .
This, then, is Balbus’ view of the traditional religion. Or, to put it in his
own terms in DND .– (pp. – above), it is his view of how the
“cult of the gods,” the cultus deorum, is to be pursued piously by religio,
“re-reading.” Perhaps with an unspoken exception or three, the cult of
each Roman god was originally, and is still for a discerning Stoic, a system
of gesture understood to honor in the right way the paradigmatic gods,
virtuous and helpful members of our rational society. Balbus thus
 Balbus the Stoic and Cotta the Skeptic
welcomes the traditional religion as traditional. He does not think it has
ever been a static tradition. Its early development was complex and
additions kept coming. But all these innovative foundations turn out to
have used a deep, or at least a respectable, understanding of how the gods
really are, and how we should treat them, even if in the case of the type-
gods their contributions have suffered distortion.
Now notice one more feature of DND .–. Balbus thinks that
some Roman ancestors devised the term religio to distinguish religious
people from the superstitious. At the time, the “superstitious” engaged in
too many ceremonies in the hope that their children would survive them
(would be superstites). These people must already have been confused by
threatening “gods” like those in myth. Thus it was already part of the early
tradition, preserved in the word religio, that the traditional cult of the gods
invites constant reflection and reinterpretation. Done right, this does not
result in radically new meanings for the cult. Although it may also bring
clarification of earlier ideas, as with the gods of types  and , it is chiefly
how the cult’s pious performance can be preserved from the distortion.
When Balbus imports the thought of the Stoics and looks for the evidence
that the same ideas prompted the design of the cult, he does what tradition
calls for. Perhaps this is just good luck. If Balbus found that Roman cult
was not compatible with Stoic ethics he would reject it. If he found that it
was not originally so, but could be reformed radically rather as Velleius
wants to do, he would do the latter. But as it is, he claims to be the one
who is truly at home with the tradition. In this sense he is a reformer: he
wishes to reform what was deformed.

On this account, you might think Balbus looks too friendly to traditional cult to be a good Stoic.
For there is evidence that Zeno was unfriendly to it. We read that in his Republic he recommended
that no temples be built (texts collected at SVF .). We should not mistake this for an edict to
destroy existing temples, but it does imply that Zeno thought having temples was not ideal. The
hostile Plutarch said that Zeno’s recommendation committed the whole school to hypocrisy if ever
they took part in traditional religion. (On Stoic self-contradictions c) Further, there are passages
where Seneca emphasizes that sages have no need of conventional religious performance: vis deos
propitiare? bonus esto. satis illos coluit quisquis imitatus est. “You want to win over the gods? Be good.
He has given them cult enough who has imitated them.” (Letters .) But if Zeno was against
traditional performances, he was not followed by the rest of the Stoa. Seneca himself, while he has
his doubts about gilding the horns of a sacrificial cow, approves of sacrificial performance when
done with pious intention, just as Balbus would lead us to expect. (De beneficiis ..) In the
doxographers, not only is the Stoic sage said to be pious thanks to the unity of the virtues (Stobaeus
..= SVF .), he also makes a study of the conventions of worship (DL .=SVF .).
The picture that emerges is that while the sage can exercise piety without performing her city’s
religion, in general she will choose to make the traditional gestures, understanding those gestures
piously. How the Stoics reconciled their determinist theology with religious actions like (say) prayer
is another question. For excellent treatments of all the matters in this note see Algra (), (),
and ().
. Cotta’s Response 
. Cotta’s Response

It is time to turn back to Balbus’ opponent. Shortly I shall examine some
key points in Cotta’s arguments against Balbus. As with Balbus’ speech, a
full treatment of Book  would be beyond the scope of this book. In
addition, Cotta’s speech has gaps in its transmitted text, so that it is hard to
pick up the thread of his general strategy at particular points. Instead,
I shall concentrate on arguments with which Cotta matches the parts of
Balbus’ speech I examined above. In sub-section .. we shall see an
objection Cotta submits to Balbus’ positive outlook on the Central Ques-
tion. In sub-section .. we shall look at Cotta’s approach to Balbus’
thesis that Stoicism and Roman religion are compatible. But first let us
examine the figure of Cotta more closely.
In section . of this chapter it emerged that Cotta defends the trad-
itional religion as handed down by his pontifical predecessors. We also saw
that it seems to him that there are gods. Here I shall examine the other
features of Cotta’s outlook on the relationship between skepticism and
religion. For although both are Academic skeptics, Cicero seems to have
designed Cotta to approach the matter differently than does his own
character, Marcus.
First, Cotta finds it most plausible that the traditional requirements of
Roman religion should be performed. That is to say, he is not just going
through the motions when he does his pontifical job, or when he defends
it. For as he says of himself, “I have persuaded myself that Romulus, with
the institution of the auspices, and Numa, with the rites, laid the founda-
tions of our state which would certainly never have been so great without
the greatest propitiation of the gods.” (DND ., see pp. – above.)
Cotta has persuaded himself using some historical evidence – Rome’s rise to
greatness – and a further argument, that if Romulus and Numa had not so
successfully established Roman propitiation of the gods, Rome’s rise would
not have occurred. For the sort of historical facts and arguments he has in
mind, we can compare Livy’s stories of Romulus and Numa’s role in
founding the religion (Livy .–, –) and the great speech Livy
gives to Camillus, linking piety and impiety with Rome’s successes and

These lacunae were evidence for A. C. Clark in The Descent of Manuscripts (). He argues that
they are a result of the loss of whole pages, quires, and so on. If Clark was right, perhaps these parts
of books were lost by accident. If so, it is in a way good news for modern readers, since the losses are
arbitrarily related to their content, and thus might not distort our interpretation in a systematic way.
But crude, censorious editing is possible too. For example, perhaps Christian readers tore out pages
or quires to which they objected. (My thanks to an anonymous reviewer for the latter point.)
 Balbus the Stoic and Cotta the Skeptic
failures (Livy .–). Is this self-persuasion unskeptical on Cotta’s part?
No, because this persuasion is only a matter of historical argument. For
Cicero, and thus probably for Cotta too, historical argument is about
rhetorical plausibility and not about rigorous proof of the truth. (On the
orator .–) So Cotta can have persuaded himself of this view in the
way he describes without taking it be true.
Notice that Cotta’s reason to support traditional religious performance
has nothing whatsoever to do with the nature of the gods. So far as he has
claimed they might not care at all. “Propitiating” them might benefit
Rome just by strengthening the “propitiators’” moral fibre. This lack of
theology is not an accident. For Cotta consistently disclaims any view at all
about the nature of the gods, not even a skeptical claim about what seems
plausible for the moment. The matter always seems “thoroughly opaque”
to him. (perobscura, DND ., cf. ., .–) In summary, then: it
seems plausible to Cotta that there are gods and that it benefits Rome to
keep up the traditional religion. But he never forms a view on whether the
gods care about this, or are in a position to do anything in return.
Thus in contrast to Balbus and Velleius, Cotta takes Cicero’s skeptical
option in the moderation of religion. He avoids positive impiety or
superstition by refusing assent to the truth of either answer to the Central
Question. In this way he represents beside them a third possible outcome
to Cicero’s project in DND and Div. But there is also something unsettling
for Cicero about Cotta’s comfort in his total inability to form even a
provisional view on what is plausible about the nature of the gods. This
is that even in this predicament Cotta does not seem to feel alienated from
his religion. Perhaps he would have more time for an antiquarian approach
like Varro’s than he does for Velleius’ and Balbus’ philosophies, but at any
rate, far from appreciating Balbus’ attempt to lead the Romans’ home to a
theoretical understanding of cult, he thinks that the theories if anything
weaken his own grasp on what is pious.
One recent interpretation of Cotta is that his long consideration of
philosophical accounts of religion have led him to suspend judgment
about religious matters in such a way that he merely follows religious
views, taking them neither to be true nor even persuasive. On this
account, Cotta strives to keep religion out of the realm of philosophical
evaluation altogether. This seems to me to go too far. Of course, as an
Academic, Cotta’s philosophical reflections lead him to suspend judgment
about truth. But Cotta takes his views, for example that there are gods, to

DeFilippo () .
. Cotta’s Response 
be persuasive. As I have just argued, he has no view about the nature of the
gods not because he has no views about what is persuasive, but because it
happens to be a particularly difficult question. Further, Cotta has not
stopped philosophizing about religion. He is the host of the conversation
in DND and seemingly its instigator. Like Marcus even at his most
skeptical in the Academica (.–), Cotta would like to be persuaded
even as to the truth, at least of the existence of the gods. (DND .)
Cotta’s conjunction of skepticism and traditional religious practices are
thus something more like what Sextus Empiricus says of the Pyrrhonist
skeptic’s outlook:
τάχα γὰρ ἀσφαλέστερος παρὰ τοὺς ὡς ἑτέρως φιλοσοφοῦντας εὑρεθήσεται
ὁ σκεπτικός, κατὰ μὲν τὰ πάτρια ἔθη καὶ τοὺς νόμους λέγων εἶναι θεοὺς καὶ
πᾶν τὸ εἰς τὴν τούτων θρῃσκείαν καὶ εὐσέβειαν συντεῖνον ποιῶν, τὸ δ’
ὅσον ἐπὶ τῇ φιλοσόφῳ ζητήσει μηδὲν προπετευόμενος.
For perhaps the sceptic will be found to be safer than those who do
philosophy in another way; in line with his ancestral customs and laws,
he says that there are gods and does everything that tends to worship of and
reverence towards them, but as far as philosophical investigation is con-
cerned, he makes no rash moves. (Adversus mathematicos ., translation
from Bett ())
This sort of attitude to religion, where radical skepticism leads to sturdy
traditionalism, has had great influence on some modern thinkers. We can
see it as an inspiration for fideism, the view that skepticism about rational
inquiry leads us to fall back on faith or tradition. (Mutatis mutandis –
orthopractic Cotta does not have the idea of religious faith in the modern
sense.) Perhaps, in this way, Cotta, and not the other characters or even
Cicero, has turned out to be the most influential voice in DND.
.. Cotta and the Problem of Us

In this sub-section, I examine a key objection Cotta makes to Balbus’
answer to the Central Question. Now, as to his general strategy, Cotta
found that Balbus said a great many things that were “fitting and coherent
with one another.” (apta inter se et cohaerentia) Therefore Cotta will “not
so much refute [Balbus’] speech as inquire after the things I did not

Popkin () documents this history amply. It seems likely that many of the authors whom
Popkin treats, being conversant with Latin, may have taken more interest in Cotta and Cicero’s
Academic skepticism than Popkin mentions, since he is interested specifically in the influence of
ancient Pyrrhonism. Penelhum () reflects on this same fideist history.
 Balbus the Stoic and Cotta the Skeptic
understand.” (DND .) So, unlike what he did to Velleius, Cotta will
not try to convict Balbus of self-contradiction. Thus he intends to put
problems for Balbus’ position on some other bases. He does so from the
point of view of a philosophically well-informed listener who can adduce
his own reasons why Balbus’ arguments are puzzling to someone who is
not a Stoic. This approach is apparent in the argument I interpret here.
In sub-section .. of this chapter I argued that Balbus’ resounding
“yes” to the Central Question was based on our reason. God gave it to us so
that we could be part of a cosmic society with him and the other gods.
Reason is the basis of the relationship with us that leads god to make
everything else for our sake. It is on this point that Cotta mounts his attack
on Balbus’ theology.
The attack comes at DND .–. The beginning of the attack is lost,
but it is part of Cotta’s reply to part [] of Balbus’ speech, “that the gods
care about us” (DND ., cf. .). Cotta gives the final inference of the
argument thus:
nam si
C. stultitia consensu omnium philosophorum maius est malum quam si
omnia mala et fortunae et corporis ex altera parte ponantur,
C. sapientiam autem nemo adsequitur,
C. in summis malis omnes sumus, quibus vos optume consultum a dis
inmortalibus dicitis.
nam ut nihil interest utrum nemo valeat an nemo possit valere, sic non
intellego quid intersit utrum nemo sit sapiens an nemo esse possit.
For if,
C. by universal consensus of philosophers, foolishness is a greater evil
than all the evils of fortune or the body that could be opposed to
it, but
C. nobody attains wisdom, then
C. we, who you say are cared for very well by the immortal gods, are in
the most evil circumstances.
For just as it makes no difference whether nobody is healthy or nobody
can be healthy, thus I do not understand what difference it makes whether
nobody is wise or nobody can be wise. (DND .)
As Cotta promised, this argument does not accuse Balbus of a self-
contradiction. Balbus could even accept all of C–C provided that, first,

non tam refellere eius orationem quam ea quae minus intellexi requirere.
. Cotta’s Response 
he could suppose that foolishness is a greater evil in the sense that it is
evil and other challenges are not, and that, second, he was allowed to
understand our evil circumstances as all our own fault. For (if we make the
relevant allowances) premises C and C are welcome to him. Following
them, he can agree that we are all evil and as unhappy as can be in the sense
that we have not achieved wisdom. What he will not accept is the informal
implication of C, that our unhappy case looks puzzling next to his claim
that the gods themselves care for us very well. Where Cotta might interpret
our “evil circumstances” as an evil world, Balbus thinks our unhappiness is
our own fault, our misunderstanding of a good world. Thus it is not at
odds with the benevolence of the gods. Cotta is aware that Balbus keeps
premise  consistent with itself by holding that each of us could make
ourselves good and happy, and thus that we are responsible for our own
evils. But Cotta wants to know how this excuses god when, as a matter of
fact, we make ourselves bad and unhappy. It is as though god planned a
world where we could, in principle, be healthy, but also planned that in
the event every one of us was sick.
Notice what a fundamental part of Balbus’ position Cotta attacks. To be
good or bad is a property only of the rational, since the only things good or
bad are virtue, vice, or those which share in them. Thus by giving us reason
god gave us the opportunity not only to contemplate, to imitate him, and
to strive to be good – he also gave us the capacity to be bad, and ingeni-
ously so. Cotta says, “if the gods gave humans reason, they gave us
malice.” (DND .)
Perhaps some Stoics would reply that Cotta exaggerates when he says
that all humans are bad. Perhaps there have been some very few good
sages. But the score is billions of bad and unhappy mortals against at most
a few of the good and happy. Balbus’ god may have made us a universe
that is full of useful and beautiful things. But, Cotta points out, for a Stoic

I use “evil” simply as the noun corresponding to “bad,” since that is traditional in discussions of the
theogical “problem of evil.” I do not mean “evil” to connote extremely bad, as it sometimes
does today.

Cotta’s argument has something in common with contemporary debate on the problem of evil. In a
classic formulation of the problem, J. L. Mackie () pointed out today’s theists are generally
committed to the propositions () “there is an omnipotent, omniscient and perfectly good God,”
and () “there is evil.” But he conceded that there is nothing formally contradictory about these two
propositions. So he then made the case for () “if () is true, then () is false.” If someone could
succeed in proving (), then contemporary theism would be shown to be irrational. Debate
therefore centers on (). Similarly, Balbus accepts  and  but also asserts that we are cared for
very well by the gods. What Cotta does not understand is his basis for holding all these thoughts in
his head at once.

si enim rationem hominibus di dederunt, malitiam dederunt.
 Balbus the Stoic and Cotta the Skeptic
that is of no account against our foolishness. Cotta’s objection, then, is a
version of the problem of evil. Balbus has a quick way with some other
varieties of this problem. If you appeal to him with natural disasters and
unfair distribution of indifferents, he can say that those are things we
should rise above. If you confront him with anybody’s ethical evil, he can
say that it is the fault of that person alone and that it need make only her
unhappy, a precisely condign punishment. But Cotta’s question is about
god’s decision to make humanity in general to be the way we are: would a
good god have condemned all of us fools to evil and unhappiness, just so
that we could have the reason that is the very cause of our unhappiness?
This is a deep challenge to Balbus’ theology.
What is there to be said in Balbus’ defence? He did not address this issue
in his speech in Book . It is possible that he replied, or that Cotta put
some reply in his mouth, in the lost section of Cotta’s argument (immedi-
ately before the lacuna in DND .). But Cotta gives us some indications
of what Balbus’ defence would rest on. One point is that, “But again and
again you argue that this is the fault of humans, not of the gods . . .” (sed
urgetis identidem hominum esse istam culpam non deorum, .) Although
everything in the Stoics’ world was fated by god, each of our actions
depends on our assent. Thus my character is a crucial factor in determining
each thing that I think or do. To make this vision of ethical responsibility
in a deterministic world rigorous, Chrysippus devised for the Stoics a
theory of necessity and possibility that allowed that, although each action
is fated, it is up to us. Thus whether we decide to do the work required to
make ourselves virtuous or not is up to us. God is therefore excused. He
may have made the world so that each of us was fated to be bad and
unhappy, but we and not he are responsible for our evil and unhappiness.
Since our evil is the only evil, the Stoic god is responsible for no evil and
his benevolence is not contradicted. This, then, is the consistent position
that we can attribute to Balbus. Cotta anticipates a remark from Balbus
that when a child uses his inheritance badly, we do not blame the parent
who left the estate (DND .). God gave us the means to be happy, it is
up to us to use it well.
What does Cotta find hard to understand about this? As he said in DND
. (p.  above), it is that he is unsure that god can be excused his
design decisions just by the subtle conclusion that each of us, in the Stoic
sense of “could,” could do otherwise than we do. God is like a professional

See On fate –, Bobzien () Chapter . That god is responsible for everything except evil is
already clear in Cleanthes Hymn –.
. Cotta’s Response 
philosopher who knows that some difficult teaching is likely to be misun-
derstood and lead students astray, but gives it anyway, on the grounds that
it is the students’ fault if they misunderstand. (DND .) In .–
Cotta has a battery of examples to illustrate the point. We abuse the very
capacities that make reason seem a gift, so that we can be all the more
creative in evil. Medea, when she killed their children to undo her
husband, was using her reason. (DND .–) The characters of comedy
give reasons and arguments for their silly or vicious goals. (DND .–)
Roman history and the law-courts of Cotta’s own day furnish endless
examples of inventive crimes. Balbus will agree that all these actions are
evil, that they are evil precisely because they are done using reason, and
that god gave us the reason we use. Cotta redescribes this situation so that
it seems grotesque.
For the reasons which Balbus would insist on, this argument of Cotta’s
would not shake a convinced Stoic. It does not try to attack the Stoic using
only premises she would accept. But as a skeptic Cotta does not need
always to disarm his opponent in that way. All he needs to do is give
arguments of equal weight to his opponent’s. He gives a thoughtful listener
a weighty reason to wonder if Balbus’ sunny theology holds up: if god
made us evil and unhappy so that he could have our company, does he
really care for us well? It seems that the world whose beauty enraptures
Balbus is in fact a sort of hell. Cicero gives us something to ponder about
Marcus’ (and, as I shall argue in Chapter ., his own) preferred answer to
the Central Question.
.. Cotta’s Reply to Balbus on Religion

Now that we have seen what a hefty weapon Cotta wields against Balbus’
answer to the Central Question, we turn to his reply to Balbus’ arguments
that Stoic theology can render Roman religion piously done. The target is
thus part [] of Balbus’ speech, “what the gods are like,” the part of the
speech summarized in the table in Appendix . Religion officially becomes
the topic in DND ., and Cotta sticks with this topic all the way to ..
It is a significant focus of his speech.
Cotta’s tactic, again, is not to aim for formal refutation, but rather to ask
about what he does not understand. What he does not understand in this
case is Balbus’ various attempts to accommodate the gods of types – (or
even of type ) in a theology focused on the cosmic god of type . He says:
 Balbus the Stoic and Cotta the Skeptic
quando enim me in hunc locum deduxit oratio, docebo meliora me didicisse de
colendis diis inmortalibus iure pontificio et more maiorum capedunculis his,
quas Numa nobis reliquit, de quibus in illa aureola oratiuncula dicit Laelius,
quam rationibus Stoicorum.
Now that my speech has led me to this topic, I shall tell you that I have
learnt better things about how the gods are to be worshipped according to
pontifical law and ancestral custom from those little pots which Numa left
to us, about which Laelius speaks in his tiny little golden speech, than from
the arguments of the Stoics. (DND .)
Laelius’ aureola oratiuncula, “tiny little golden speech,” probably men-
tioned the little pots as evidence that Numa, reputed founder of many
aspects of the religion, was frugal in his religious practices. Now if we
had not examined Balbus’ ideas about religion closely, Cotta’s answer here
might look misplaced. Balbus did not intend to change anything about
pontifical law or ancestral custom in the sense of changing any prescrip-
tions for performance. So why does Cotta object that Balbus has taught
him nothing about legal or customary worship? He does so because Balbus
wants to help Cotta understand why the traditional prescriptions were
instituted, and what they mean. Cotta replies with an apt, if satirically
small, point of comparison. From Laelius’ citation of the pots he learnt
something about what a founder figure intended by his prescriptions. As a
priest, he can report that the pots were more use than all the Stoic
theorizing, but in just the way that the Stoic theorizing was supposed to
be useful.
What does Cotta not understand about Balbus’ theories? Cotta does not
object to the rationalization of traditional beliefs about the gods per se.
Indeed he alleges that he wants more of it. The problem is rather that
Balbus’ methods do not work to Cotta’s satisfaction. For example, in the
case of some Greek myths, he says
atque haec quidem <et alia> eius modi ex vetere Graeciae fama collecta sunt.
quibus intellegis resistendum esse, ne perturbentur religiones; vestri autem non
modo haec non refellunt verum etiam confirmant interpretando quorsum
quidque pertineat.
And these particular stories, and others of this kind, have been assembled
thanks to their ancient acclaim in Greece. You understand that they must
be resisted lest religion be distressed. But your Stoics not only do not refute
the stories, rather they even strengthen them by interpreting what each one
refers to. (DND .)

See the texts cited by Mayor () ad loc.
. Cotta’s Response 
As he presents it here, Cotta would not mind if Balbus wanted simply to
reject these myths as prone to produce superstition. It is Balbus’ methods
for adopting the myths that Cotta worries about.
When Cotta refers to problematic Greek myths, he does not mean that
any myth is problematic. Rather, what he has just done, as he says, is to go
through assemblies of myths, all of good repute. (DND .–) Each
group of myths is about some one god, but each myth in the group
attributes to the god different parentage. But if two people do not share
both parents, they cannot be the same person. Thus if we adopt all five of
Cotta’s distinct myths about Mercury, for example, it follows that there are
five different Mercuries. So does Balbus, who wants to accommodate all
the myths by interpretation, intend to multiply Mercury? But then the
number of type- gods that religion should acknowledge will have grown
in a worrying way. Now Balbus could have an answer to this charge. He
might say, for example, that each myth of Mercury’s origin discloses a
distinct piece of physical information about the god to which it refers.
These bits of physical information need not be contradictory. So Balbus
does not have to concede that the myths require five Mercuries. But
Cotta’s argument here is indicative of a trend. He tends to object that
Balbus’ methods yield too many gods.
Cotta sets this agenda early on in his reaction to part [] of Balbus’
speech, “what the gods are like:”
nec vero volgi atque imperitorum inscitiam despicere possum, cum ea considero
quae dicuntur a Stoicis. sunt enim illa imperitorum: piscem Syri venerantur,
omne fere genus bestiarum Aegyptii consecraverunt; iam vero in Graecia multos
habent ex hominibus deos . . . Herculem Aesculapium Tyndaridas Romulum
nostri aliosque compluris, quos quasi novos et adscripticios cives in caelum
receptos putant. haec igitur indocti; quid vos philosophi, qui meliora? omitto
illa, sunt enim praeclara: sit sane deus ipse mundus. . . . quare igitur pluris
adiungimus deos? quanta autem est eorum multitudo!
I cannot look down on the ignorance of the mob and the uneducated, when
I reflect on what the Stoics say. From the uneducated you get these sorts of
things: the Syrians revere a fish, the Egyptians have consecrated almost
every kind of beast. Now, even in Greece they have many gods made out of

I agree with Mayor () against e.g. Ax () and Pease () in reading nostri over nostrum. It
is, of course, possible that Cotta gives a list of the heroes of Greece, marking out Romulum nostrum
as the only Roman on the list. But that makes the sequence of thought strange: why would Cotta
restrict his list of the uneducated to Greeks and barbarians? If we read nostri then he moves from the
Greek heroes to a list of the heroes accepted by the Romans. Cf. DND ., p.  above.

This claim probably stems from the iconography of either Atargatis or Dagon.
 Balbus the Stoic and Cotta the Skeptic
humans . . . our own people have Hercules, Aesculapius, Castor and Pollux,
Romulus, and many others, who they think were admitted to heaven as
new and naturalised citizens. So that’s what the unlearned say—what do
you philosophers say, how is it better? I pass over your well-known
arguments, let the cosmos itself be a god at least . . . So why do we add
many gods? But how great is the crowd of them! (DND .–)
Cotta proceeds to argue that the Stoics match the stupidity of the crowd
point for point. They deify stars which they name after beasts, matching
the mistake of the Syrians and Egyptians. They make gods of dead
humans, matching the mistake of the Greeks and Romans. (DND
.–) But although these are particular arguments, “why do we add
many gods?” could stand as the slogan for all of Cotta’s engagement with
Balbus on religion. He will grant the existence of the singular, type- god
(“let the cosmos itself be a god”) in order to raise difficulties about the
introduction of further gods. Thus what he wants to question is Balbus’
methods: if Balbus wants to get from one god to many in such-and-such a
way, then he will find himself led to yet more, and too many, gods.
Cotta takes on Balbus’ methods directly, as we saw in DND .
(p.  above) or in an attack on allegoresis (explicatio fabularum) and
etymology (enodatio nominum) (DND .–). But recent scholarship
has given pride of place among Cotta’s argument to a volley of arguments
of the type we call the sorites, or “heap.” (DND .–) Cotta explicitly
attributes these arguments to Carneades. (DND .) We have a close
enough parallel for some of them in Sextus Empiricus to say with Sextus
that they were written down by Carneades’ pupil, Clitomachus. (Adversus
mathematicos .) Here we find Cicero, directly or indirectly, and from a
source or from memory, using a set of arguments whose provenance we
know. We can see what he does with them. Let us see first, then, what
these arguments given by Carneades were like.
Arguments in the sorites or “heap” type can be put in various forms, but
one such would be this. The question is raised, how many grains of sand
make a heap? One grain on its own is not a heap. A million grains is a
heap. Furthermore, one grain of sand is negligible. The addition of just
that one grain could never turn a collection that is not yet a heap into a
heap. Thus we can say, if one grain of sand is not a heap, then neither are
two. If two grains of sand are not a heap, then neither are three. On we go,

In this summary interpretation of Carneades’ theological sorites arguments I am indebted to
Couissin () and especially to Barnes () and Burnyeat (). I agree with Burnyeat in
particular. For Cicero’s knowledge of the sorites as such, see Academica ., .–.
. Cotta’s Response 
until we reach: if , grains of sand are not a heap, then neither are a
million. Thus we seem to reach a false conclusion from true premises by
valid reasoning. Such arguments would appeal to skeptics since they
confront dogmatists with apparently unattractive options. Either the dog-
matist can give up one of the plausible sounding claims about what a heap
is or whether a single grain of sand makes a difference, or he can give up on
if-then reasoning.
It seems that Chrysippus’ response for the Stoics to such arguments was
to reject the thesis that the addition of one more grain cannot turn a little
collection into a heap. For him, the source of the apparent trouble is not
that there is no definite point at which a heap appears, but rather that we
do not know exactly where it is. Thus Chrysippus could hold on to the
soundness of if-then reasoning where all of its conditionals are true. But
he can say that somewhere in the sequence of conditionals between “if one
is not a heap, then neither are two” and “if , are not a heap, then
neither are a million” is a false conditional, perhaps “if , are not a
heap, then neither are ,.”
The sort of sorites argument that Carneades used against Chrysippan
theology, and which Cotta exploits against Balbus, was meant to capitalize
on Chrysippus’ response to the sorites in general. Suppose Carneades
could construct a sequence of conditionals that ran from “if Zeus is a
god, then so is Poseidon” to “if a brook is a god, then so is a puddle.”
Suppose that a Stoic would agree that Zeus is a god, but a puddle is not.
Then, according to Chrysippus’ own response, the Stoic would have to
conclude that one of the if-then claims in between “Zeus is a god” and “a
puddle is a god” was false. Thus if Carneades could only show that each of
his if-then claims is true, or at any rate that each is compelling to a Stoic,
then the Stoic’s rational approach to theology would totter. She would
have to accept either that Zeus is not a god, or that a puddle is a god, or
that there is something fishy about if-then reasoning after all. Moving from
one god to many would seem a task that risks much.
Since each premise in such a sorites must be examined, Carneades’
theological sorites are rather short. We do not go a million gods deep. We
go just far enough to reach some candidate “god” who all would agree is
not a god. Thus unlike with the sand argument I gave above, Carneades

Strictly speaking Chrysippus seems to have formulated the sorites not with the strongest form of
Stoic conditional, which was true if and only if the truth of its antecedent was incompatible with the
falsity of its consequent, but with a truth functional or material conditional, which is true if and
only if it is not the case that both its consequent is false and its antecedent true. See Burnyeat ()
–.
 Balbus the Stoic and Cotta the Skeptic
does not need to skip a long sequence of conditionals and say “on we go,
until we reach . . .” If one conditional in the chain seems weak or contro-
versial, he can provide it with the support it needs. Here is a sample from
Sextus:
καὶ μὴν εἰ ὁ ἥλιος θεός ἐστιν, καὶ ἡμέρα ἂν εἴη θεός· οὐ γὰρ ἄλλο τι ἦν ἡμέρα
ἢ ἥλιος ὑπὲρ γῆς. εἰ δ’ ἡμέρα ἐστὶ θεός, καὶ ὁ μὴν ἔσται θεός· σύστημα γάρ
ἐστιν ἐξ ἡμερῶν. εἰ δὲ ὁ μὴν θεός ἐστι, καὶ ὁ ἐνιαυτὸς ἂν εἴη θεός· σύστημα
γάρ ἐστιν ἐκ μηνῶν ὁ ἐνιαυτός. οὐχὶ δέ γε τοῦτο· τοίνυν οὐδὲ τὸ ἐξ ἀρχῆς.
Then again, if the sun is a god, day too would be a god; for day is none
other than the sun above the earth. But if day is a god, the month too will
be a god; for it is a composite of days. And if the month is a god, the year
too would be a god; for the year is a composite of months. But this is not so;
therefore neither is the original point . . . (Sextus, Adversus mathematicos
., translation from Bett ())
The Stoics think the sun is a god. Thus Carneades constructs a sequence of
conditionals that move them from that claim to the claim that the year is a
god. But he thinks the Stoics would reject that latter. Thus the Stoics have
Carneades’ intended three unwelcome options: they must reject the divin-
ity of the sun, endorse the divinity of the year, or reject the reasoning of
the argument. Each conditional is supposed to rely on some thought
analogous to the addition of a single grain of sand. The distinction
between, for example, the sun and the sun above the earth is supposed
to be too small to allow us to attribute divinity to the one but not the
other. But instead of leaving each of these moves to a mechanical principle,
as we did in the sand example, Carneades specifies in which respect each
pair are too similar for the Stoics to make only one of them a god.
So then, what distinctive use does Cotta make of these arguments? The
most obvious distinction is that Sextus deploys them for the conclusion
that there are no gods (Adversus mathematicos .) but Cotta does not.
For Cotta says, “Carneades used to say these things, not to refute the gods
(for what is less fitting for a philosopher?) but so as to make the case that
the Stoics explain nothing about the gods.” (DND .) Cotta certainly

Many of Carneades’ grounds for similarity seem unconvincing to me. If the sun is a god then of
course the sun, when above the earth, is a god. But the state of affairs, the sun being above the earth,
need not be. Day seems to be the state of affairs, not the sun. So why, if the day is a god, need it
follow that the month is one too? Perhaps Carneades’ arguments seem most attractive if you already
have the sense that the Stoics’ accounts of some polytheist gods look like special pleading. If they
went to these lengths to make such an arbitrary range of entities divine, why these rather than those?

haec Carneades aiebat, non ut deos tolleret (quid enim philosopho minus conveniens), sed ut Stoicos nihil
de dis explicare convinceret.
. Cotta’s Response 
sees the big picture. For Carneades the skeptic cannot have meant to
prove, once and for all, that there are no gods. Instead, like Sextus, his
larger goal must have been to match the Stoic arguments that there are
gods. But of the various conclusions available from these arguments, it is
possible that Sextus is right that Carneades stated the one given, that none
of the “gods” mentioned from the start of the argument are gods. Indeed in
general Carneades may have stated these sorites arguments with Sextus’
conclusion that there are no gods. For he may have thought that this was
the sting in the tail for the Stoics: they gave a theology that argued for
gods, he uses their own methods to argue that there are not. If so, then
Cotta describes Carneades’ purpose accurately but adapts the arguments
for his own purposes by leaving out the atheist conclusions in which Sextus
is interested.
A second difference is that Sextus’ Carneades uses his sorites against
Stoic gods in general. For example, in the argument quoted above (p. )
he argues against one of Balbus’ type- gods. But Cotta only uses his
(surviving) sorites against cultural gods. Where he attacks the divinity of
the sun, moon, or heaven, it is in arguments whose steps depend on myths
and the identification of, for example, the sun with Apollo. (DND .,
.) Thus those arguments deal with the gods of type . In Cotta’s hands
these arguments are not against Stoic theology in general, but rather
against Balbus’ methods in accommodating cultural gods.
A third difference is that, in keeping with his general strategy, Cotta
does not claim to have convicted Balbus of incoherence or irrationality. In
Sextus’ handling, at any rate, Carneades’ formal treatment amounts to an
argument that the Stoics are either deceived about the consistency of their
views, or are irrational by their own standards in holding them. Cotta
instead introduces these arguments as something Carneades “used to say”
(DND .), asking “if I should follow you, tell me, what I should say to
somebody who questions me as follows? . . .” (DND .). Cotta does not
assert the arguments or that Balbus has no way to respond to them. It is
just that he, Cotta, does not understand what Balbus can say. Thus the
reader is left suspending judgment on the issue and Cotta will not “follow”
Balbus.
A fourth difference is that Cotta is often interested not only in the
existence of the gods he mentions, but also in whether they deserve
worship. For example, Cotta argues that if the Roman gods are to be
worshipped, then so are the Egyptian gods Isis and Osiris. But if Isis and
Osiris are to be worshipped, then so are the other barbarian gods. Cotta
says that the latter include cows and horses, ibises, hawks, crocodiles, fish,
 Balbus the Stoic and Cotta the Skeptic
and so on. (DND .) Balbus is thus left to choose between unwelcome
options, that cows should be worshipped or that the Roman gods should
not. This argument shows not only that among Cotta’s targets in the
sorites arguments is Balbus’ defence of traditional religion, but also that
Cotta looks at the question from a specifically Roman perspective. He asks
not only “are these gods worthy of worship?” but also “are these gods
whom the Romans should accept as such or worship?”
This emphasis in Cotta’s sorites is of not only theoretical import. For
the treatment of new claimants to divinity was a real issue for the
authorities at Rome, especially as the empire expanded. At one extreme
might be introduction of the cult of a new god at Rome itself. But even
acknowledging the divinity of gods overseas could have conseqences. Cotta
alludes to one such real case, one for which the historical Cicero sat on the
investigating committee of senators. Here is the relevant sorites:
an Amphiaraus erit deus et Trophonius? nostri quidem publicani, cum essent
agri in Boeotia deorum inmortalium excepti lege censoria, negabant immortalis
esse ullos qui aliquando homines fuissent. sed si sunt i di, est certe Erectheus,
cuius Athenis et delubrum vidimus et sacerdotem. quem si deum facimus, quid
aut de Codro dubitare possumus aut de ceteris qui pugnantes pro patriae
libertate ceciderunt? quod si probabile non est, ne illa quidem superiora unde
haec manant probanda sunt.
Or will Amphiaraus be a god and Trophonius? Our taxmen, at any rate,
used to deny that there are any immortals who were once humans, when
there were fields in Boeotia belonging to the immortal gods exempted from
tax by the censors’ code. But if they are gods, Erechtheus certainly is, who,
as I saw myself, has a shrine and a priest at Athens. If we make him a god,
how can we hesitate either over Codrus or over others who fell fighting for
their country? But if the latter is not plausible, neither are those earlier
points from which the later ones flow. (DND .)
Amphiaraus and Trophonius were two heroes – gods of Balbus’ type  –
who were worshipped in Boeotia, a region of Greece north of Athens.
Land consecrated to the gods was exempt from tax under the law Cotta
cites. Thus the tax farmers of Boeotia argued that the two heroes were not
gods, in order to secure the tax from their land. The Senate’s decision was
that Amphiaraus and Trophonius were gods. So Cotta here reminds his
listeners that for the Roman authorities, the divinity of the two heroes was

For the case and Cicero’s involvement, see Ando () –. See also his Chapter  on
interpretatio Romana. North ()  points out that the Amphiaraus case happened at
roughly the dramatic date of DND.
. Cotta’s Response 
a question with practical implications. Then his sorites proceeds.
Erechtheus, legendary Athenian king, similarly had a cult, so by the same
criteria as Romans accepted Amphiaraus as a god they should accept him,
too. But then Codrus, a less prominent mythical king who died to save
Athens and was also accorded cult, should be accepted as divine, and so
on. But if it is not plausible that Codrus and other such heroes are gods
(and exempt from Roman tax!), then, by the logic of the sorites, the
decision about Amphiaraus was wrong. In this way, Cotta finds that
Balbus’ openness to type- gods threatens to undermine an established
decision.
In summary, Cotta’s sorites arguments are in large part intended to
show that Balbus has not succeeded to Cotta’s satisfaction in accounting
for the reality of the traditional gods, of types –, in such a way that the
Romans may worship them without superstition. For Balbus’ methods
seem to apply too widely, so that piety threatens to commit the Romans to
the worship, or acknowledgment, of a crowded, indeterminate mass of
gods. Further, we see Cicero shaping his material, for he uses Carneades’
arguments to Cotta’s purpose in a way that Sextus, with his different
purpose, did not use the same material.
Is Cotta right to be puzzled in this way? There seems to be an obvious
answer available to Balbus. This is that he has not set out to prescribe a
religion from scratch. Rather he offers support for Roman religion as it
actually is. His method is to reconstruct the historical circumstances of
each innovation. But then of course the choice of the Roman gods, and not
others, seems arbitrary. The gods of Rome were indeed selected by a long,
complex, and as a whole rather arbitrary set of decisions. Meanwhile,
among the many religious innovations of the past, there were of course
an indeterminate number of ways that the gods of religion might relate to
the paradigmatic gods. Again, this is not the fault of Balbus’ methods.
Rather his methods rightly reveal the complexities of history. In short, it is
not Balbus’ job to cut off Cotta’s sorites at the appropriate point. It is his
job only to interpret the data of Roman tradition, messy as they are. The
boundary of those data is given to him as his cut-off.
So is Cotta putting Carneades’ arguments to unfair use? No, he could
find sufficient grounds to press his question fairly. First, although Balbus is

For data on Erechtheus and Codrus, see entries in Kearns (), Appendix .

Of course, Sextus may have adapted the arguments, too. It is possible that Cotta’s arguments are in
fact closer to what Carneades gave. Cotta’s version could work against a lot of Stoic allegoresis and
divine etymology, notably what we find in Cornutus.
 Balbus the Stoic and Cotta the Skeptic
not prescribing a religion from scratch, his views on the type- gods
suggest that people with a Stoic-style world-view prescribed many early
and surviving aspects of the Roman tradition. Not only that, but such
people also gave rise to the Greek myths. If Balbus were an Egyptian or a
Syrian, no doubt he could come up with a way to see the hand of proto-
Stoics in the origin of the worship of Serapis or of the fish. Thus if the
old physicists were right to prescribe the worship of Jupiter but also right in
the views that gave rise to the myth of his many brothers, why does Balbus
think he should worship Jupiter but not adopt the brothers as gods, too?
I expect Balbus could say that Romans worship the paradigmatic gods by
their traditional conventions, the Syrians do so by theirs, and so on. Each
tradition is sufficient for its own citizens but not required of others. But,
second, as Balbus testifies, the nature of the Roman tradition was not to
guard some static set of gods over time. Balbus’ history of the religion is a
history of innovation and of new gods, all of which he accepts as trad-
itional. Thus if Balbus commits himself to the equal propriety of worship-
ping Jupiter and the Syrians’ fish, and if Cotta the pontifex were to follow
him in this, consistency could require Cotta to push for the introduction
of the fish to Rome as a god. It is not clear how Balbus proposes to block
this consequence.
These last two paragraphs of dialectic are only my informed speculation,
and one could go through many more rounds to and fro. It is into exactly
this sort of beard-stroking, I think, that Cotta means to send us with his
sorites, and with the other arguments that Balbus commits himself to too
many gods. For Cotta’s point is not that Balbus is definitely wrong, but
rather that Balbus has not yet told us enough to allow us to understand
how Stoic theology and traditional religion are compatible. If Cotta were
to use Balbus, rather than his pontifical predecessors, as his authority on
the nature of the religious tradition, religion at Rome might be confused
and distressed. These particular arguments are given in Cotta’s skeptical
role. But as it happens they fit neatly with what he tells us about his
personal outlook. Cotta will leave Stoic theorizing well alone, remain in
complete ignorance of the nature of the gods, and carry out the duties of a
pontifex simply as tradition requires.
We should come away from Books  and  of DND with respect for the
courage of Cicero’s skeptical practice. The manner in which Cicero faces

Here we might compare Balbus with Chaeremon who, in the century after Cicero, was an Egyptian
grammarian and allegorist of religious matters in the Stoic style. See Van der Horst (), Frede
().
. Cotta’s Response 
both what is most attractive and what is most troubling in Balbus’ ideas is
courageous. To a Roman searching for a philosophical religion, Balbus’
world is enticing indeed, a cosmos of beauty and beneficence to which the
bewildering old cults suddenly turn out to have been tailored. But Cicero
gives Cotta strong reasons to doubt this world, even to recoil from it. The
“benificent” god gave us not only beauty, but also Thyestean feasts and
pandemic evil. The accommodation of traditional religion to philosophical
theology, it turns out, may be so much irresponsible special pleading.
Cotta is such a forceful character that, after these objections, Marcus’
preference for Balbus’ speech at the end of Book  comes as a surprise.
Whichever way Cicero’s own ideas tended to run, we should see these two
books as what they are, a remarkable exercise in intellectual integrity: the
avoidance of rash assent, even in matters of moment. In the next chapter
we turn to Div., where, I shall argue, scrutiny of Cotta and Balbus’ debate
on the Central Question is narrowed and resumed.
 
Quintus’ Stoic Case for Divination
Consider again the following passage (cf. p.  above):

nam cum omnibus in rebus temeritas in adsentiendo errorque turpis est, tum in
eo loco maxime, in quo iudicandum est, quantum auspiciis rebusque divinis
religionique tribuamus; est enim periculum, ne aut neglectis iis impia fraude
aut susceptis anili superstitione obligemur.
For error and rashness in assent is vicious in any matter, but it is especially
so on that question where we must judge how much credit to give to
auspices, to divinity, and to religion. For there is a danger that we shall be
involved either in an impious fraud (if we neglect these matters) or in the
superstition of an old woman (if we accept them). (Div. .)
With this envoi Cicero’s preface to Div. gives way to the dialogue itself. It
puts back squarely before the reader the Central Question of DND (see
p.  above): do the gods care for us? For of the motivations that Cicero
gave in the preface to DND for the study of the nature of the gods, the
more salient was that it is ad moderandam religionem necessaria, “necessary
for the moderation of religion.” I argued that Cicero means that natural
theology puts modi, limits, on religion, becuse it keeps religious actions
between, on the one hand, impiety and, on the other, superstition (see
Chapter  section ., and Appendix ). Now, as we enter the drama of
Div., we are reminded that exactly this is the central project of the two
dialogues. For Cicero says that we run the risk of superstition if we
“accept” (susceptis) augury and religion. suscipio was the verb to describe
a Roman father’s ritual lifting up of a baby at the hearth, to accept it into
the family. If an augur “accepts” his auspices, he takes them to heart – he
thinks that they really obtain divine sanction so that, if they do not really
do so, he is superstitious. But if he “neglects” them (neglectis) he thinks that
they do not really obtain divine sanction so that, if they do obtain divine
sanction, his action is an impious fraud.

Quintus’ Stoic Case for Divination 
Div. involves more detailed and more technical philosophy than does
DND. My exposition in this and the next chapter must reflect that. We
can see why Div. has this texture. So to speak, Cicero has zoomed in on
one piece of Balbus’ and Cotta’s architecture of arguments. The implica-
tion of the dense and technical character of Div. is that Cicero thought
divination was worthy of very serious philosophical consideration. We
should first reflect for a moment on why Cicero was inclined to take the
subject seriously.
As is documented very amply in Div., divination was everywhere in
Roman life. There were private practices, like astrology (a recent arrival at
Rome in Cicero’s day), or attention to the premonitory power of one’s
dreams and reveries. But there were also the grand consultations of the
traditional religion. We saw that in DND Cotta described two of his three
parts of Roman religion as divinatory: augury and haruspicy (see p. 
above). Augury involved, primarily, the observation of the behavior of
birds, to determine whether the gods gave their assent to this or that
action of the state. Cicero himself was an augur, a member of the college of
officials who oversaw augury (see p.  n. ). Thus, just as in DND Cotta
the pontifex takes on the subject of the nature of the gods, so in Div.
Marcus the augur takes on the subject of divination. Meanwhile the
haruspices, whose office was seen as Etruscan in origin, consulted such
evidence as flaws in the entrails of sacrificial victims, or lightning strikes,
and thus were able to prescribe remedies for disturbances in Rome’s
relationship with the gods, notice of which the gods portended to the city
by the arrival of “prodigies,” strikingly strange events. In summary a
Roman, and especially a Roman of the ruling class, lived surrounded by
phenomena that might, according to tradition or to private belief, bring
news from the gods. Thus this Roman might be expected to expend
energy, and perhaps a good deal of energy, to mark and to interpret these
potential messages.
In Cicero’s world, what did intellectuals make of this ever-present aspect
of Roman life? It is hard to put ourselves into their minds in this regard.
Obviously, they had not considered the Christian condemnation of divin-
ation that has held sway for most of the time between Cicero and our day,
namely, that we should recoil from it, since if it works, it works by
demonic intervention. Meanwhile, I expect that to many of my twenty-
first-century academic readers, Roman divinatory practices simply seem
bizarre, or pseudo-scientific, or at any rate a waste of time. Cicero was
certainly aware that one could arrive at this latter sort of attitude, and may
have done so himself, as we see in the last sections of Div. Ancient
 Quintus’ Stoic Case for Divination
Epicureans, especially, were dismissive. But if ancient Roman (or Greek)
intellectuals found a dismissive view of divination obvious, they did not say
so. Their general experience seems to be best summed up by the opening
of Aristotle’s little treatise On divination in dreams (b–):
τὸ μὲν γὰρ πάντας ἢ πολλοὺς ὑπολαμβάνειν ἔχειν τι σημειῶδες τὰ ἐνύπνια
παρέχεται πίστιν ὡς ἐξ ἐμπειρίας λεγόμενον, καὶ τὸ περὶ ἐνίων εἶναι τὴν
μαντικὴν ἐν τοῖς ἐνυπνίοις οὐκ ἄπιστον· ἔχει γάρ τινα λόγον· διὸ καὶ περὶ
τῶν ἄλλων ἐνυπνίων ὁμοίως ἄν τις οἰηθείη. τὸ δὲ μηδεμίαν αἰτίαν εὔλογον
ὁρᾶν καθ’ ἣν ἂν γίνοιτο, τοῦτο δὴ ἀπιστεῖν ποιεῖ·. . .
The fact that all persons, or many, suppose dreams to possess a special
significance, tends to inspire us with belief in it, as founded on the
testimony of experience; and indeed that divination in dreams should, as
regards some subjects, be genuine, is not incredible, for it has a show of
reason; from which one might form a like opinion also respecting all other
dreams. Yet the fact of our seeing no reasonable cause to account for such
divination tends to inspire us with distrust. (Translation by J. I. Beare in
Barnes () vol. )
Aristotle’s attitude is that divination (in dreams, anyhow) has a certain
anecdotal plausibility, but that it seems very hard, or impossible, rationally
to explain its causes. Aristotle, of course, was writing in Greece three
centuries before Cicero, but his attitude captures the sort of puzzle faced
by the Hellenistic philosophers who inform Cicero, and by thoughtful
Romans, too. It was not only that the kinds of divination were variously
entrenched and recommended in law or in custom. There was also the
sense from experience and anecdote that, inexplicably, divination some-
times delivered results that were hard to dismiss. To illustrate this experi-
ence, Cicero uses a setting for Div. reminiscent of other dialogues he wrote
on personal themes, On friendship and On old age. That is to say, we are
flies on the wall at a private conversation between Marcus and his own
brother, Quintus. In this intimate setting, the brothers can call on their
familiarity with one another’s inner lives for moments where the hope, or
the threat, of divination, could grip. We learn, for example, that the two
had a pair of strange but strikingly premonitory dreams, both of which
predicted Cicero’s return from political exile. (Div. .–) Whether
these particular stories are fiction does not matter. Cicero shows us why, in
a society or a family stocked with such anecdotes, even the most critical
philosopher might keep an open mind, and allow that divinatory phenom-
ena might be an object of serious intellectual concern.

Schultz () – also emphasizes the intimacy of Div.’s setting.
Quintus’ Stoic Case for Divination 
Notice that once it is your concern, divination raises questions at the
heart of philosophy. Is the future fated now? Can anyone or anything
know it? What distinguishes a science from a pseudo-science? What,
indeed, is truth, what is meaning, and where do we find them? With such
questions at stake, it is no surprise that philosophers like the Stoics
subjected divination to careful scrutiny. Then there is Cicero’s Central
Question from DND: are there gods who care for us? The reality of
divination would prove at a stroke Balbus’ answer to the Central Question.
For if the gods send us information, and especially if they surround us with
signs as the Stoics claim, then they can and do care for us. In sum, Div. is
more than a pendant to DND. It is the technical scrutiny of a weighty and
potentially decisive aspect of DND’s project.
To frustrate “vicious” “rashness of assent” (Div. .) in these matters,
then, Cicero balances the two books of Div. between a long exposition of a
Stoic defence of divination in the mouth of Quintus (Book ), and a long
speech against this Stoic position in the mouth of his own character,
Marcus (Book ). In this and the next chapter I shall offer philosophical
interpretation of the main bone of contention in the two speeches respect-
ively. In this chapter I shall examine the Stoic case for divination that
Quintus gives. Now by his own account, Quintus’ Stoic case for divination
rests entirely on what he calls eventa or “outcomes.” (Div. .–, see
pp. –.) That is to say, he seeks to prove the reality of divination
simply by pointing out many cases where divinatory predictions came true.
By this he avoids basing his proof on what he calls a ratio or “theory” to
explain why they come true. The quotation from Aristotle we have just
seen gives us a sense of why this would be a promising strategy. Now in
fact in the latter part of his speech Quintus gives in addition some causal
explanations of divination. (Div. .–) Those explanations are worth
lengthy attention in their own right, but since Quintus puts no weight on
them in arguing for his conclusion, they are not part of the trajectory of
this book. Thus in this chapter we shall examine only the case from
outcomes. In the next chapter we shall see Marcus’ response to it.
The long, former part of Quintus’ speech (Div. .–), the part on
outcomes, presents some difficulties. For example, Quintus maintains and
emphasizes throughout what seems an unnecessarily strong distinction
between two kinds of divination, “artificial” and “natural.” But then he

Some recent studies of the causal explanations in Div. .– are Dragona-Monachou (),
Glucker (), Tarrant (), Guillaumont (), Brittain (), and the relevant parts of
Schäublin (), Yon and Guillaumont (), Wardle (), and Schultz ().
 Quintus’ Stoic Case for Divination
includes what seems to be a previously unremarked third kind, “from
conjecture.” Another difficulty is that Quintus often seems to move from
one topic to another at random, and to digress or to repeat himself. Most
of all, the thing reads as a welter of bizarre anecdotes. To quote Malcolm
Schofield:
This deluge of examples permits Cicero to indulge his skills as a story-teller
(not to mention his ambitions as a poet) on a much grander scale [than in
DND]. The reader may be forgiven for feeling sometimes that the real point
of Div.  is simply that it gives him the opportunity to do so. The
underlying philosophical thought is presumably that it is precisely an
authentically messy welter of allegedly divinatory experiences which gives
the best chance of persuading someone of the case for divination. . . . Of
course, there are alternative justifications conceivable: pile up the evidence;
if there is a lot of it, the reader may begin to think there is something in it.
But why the chaotic disorder of Quintus’ examples? (Schofield ())
As I have already suggested, I agree with Schofield that Quintus’ strategy is
to invoke the “authentically messy welter of allegedly divinatory experi-
ences.” But I hope to discover more about why Quintus takes this
approach, and how “the underlying philosophical thought” is supposed
to work.
My interpretation of Quintus’ speech benefits from my two principles
of interpretation in this book, the literary unity thesis and the learned reader
principle (see p.  above). For I think Quintus does not simply recite
“the” Stoic view of divination, and that he does not attempt to introduce
the topic from scratch to a layman. Rather, his speech has the following
complex goals. In his reply to Balbus in DND, Cotta challenged a Stoic
view that divination is an art. When he did this, Cotta attacked an account
of divination associated in our sources with many Stoics, but most closely
with Chrysippus. Quintus wishes to answer Cotta’s questions. Yet he does
so not by defending Chrysippus’ account, but rather by advancing a
second Stoic account that inherits many ideas from, but is not compatible
with, Chrysippus’. According to this new account, divination is not an art,
but rather a kind of prediction, which sometimes uses arts, and sometimes
does not. If I am right about these complexities in Quintus’ speech, they
can solve some of its difficulties. For one thing, some of Quintus’ apparent
muddles are not muddles, but rather are complexities where two different
Stoic accounts of divination are under discussion. For another, Quintus

For attempts to make sense of the complexities of Div. at the level of rhetoric and the formation of
discourse, see Beard (), Krostenko ().
. The Occasion for Quintus’ Speech 
observes his distinction between natural and artificial divination so rigidly
because that is the distinctive feature of his position, and where he sees the
strength of his answer to Cotta.
I shall offer this interpretation as follows. In section ., I shall show
that a response to DND, and to Cotta in particular, is Quintus’ goal.
In section ., I shall set out what I have called Chrysippus’ position
on divination, the foil to Quintus’ own version of the Stoic account. In
section ., I give Quintus’ account. I then show how Quintus applies his
new Stoic account of divination, first to what he calls artificial divination
(section .), and then to what he calls natural divination (section .).
. The Occasion for Quintus’ Speech

Cicero begins the dialogue of Div. with an unusual literary scheme. His
characters Quintus and Marcus discuss his (the real Cicero’s) recent opus
DND. This discussion sets in motion the rest of Div.. By this trick, Cicero
signals not only that Div. is a continuation of the project of DND, but also
that Div.’s characters might intend to respond to the earlier text which they
have read, just as we have. Sure enough, Quintus is explicit that his speech
will be a reaction to an inadequacy in the treatment of divination in Books
 and  of DND. He says,
eius rationi non sane desidero quid respondeam; satis enim defensa religio est in
secundo libro a Lucilio, cuius disputatio tibi ipsi, ut in extremo tertio scribis, ad
veritatem est visa propensior. Sed, quod praetermissum est in illis libris (credo,
quia commodius arbitratus es separatim id quaeri deque eo disseri), id est de
divinatione, quae est earum rerum, quae fortuitae putantur, praedictio atque
praesensio, id, si placet, videamus quam habeat vim et quale sit.
But certainly I am not lost for an answer to [Cotta’s] argument. For Balbus
gave religion a sufficient defence in the second book. To you yourself his
speech seemed closer to the truth, as you write at the end of the third book.
But—if you like—let’s see what the question of divination is like, and what
power there is in it. Divination is the prediction and foreknowledge of
states-of-affairs that are thought to be by chance. This question was passed
over in those books [in DND], I believe, because you decided it was more
convenient to investigate and to write about it separately. (Div. .)
At first sight Quintus’ proposal is surprising. Divination is not obviously
“passed over” in DND. It features prominently in Balbus’ arguments for

It is possible that illis libris, “those books” in which divination was passed over, refers to all three
books of DND. But Quintus has just said that he read through Book . He is reassured by Balbus’
defense of religion in Book . At no point has he referred to Book . (Div. .–)
 Quintus’ Stoic Case for Divination
two of the four heads of his speech: part [], that there are gods (DND
.–), and part [], that they care about us (DND .–,
–). (Cf. DND ., p.  above.) The latter head is, of course,
crucial for the Central Question.
To see where the people of DND really do skip over divination, we must
look to Cotta’s arguments against it in Book  of that dialogue. Two of
Cotta’s arguments survive. One asks whether divination would not be
useless, or even unkind to us, in an inexorably fated world like the Stoics’.
(DND .) Cotta’s second objection will be at issue in the rest of this
chapter:
unde porro ista divinatio, quis invenit fissum iecoris, quis cornicis cantum
notavit, quis sortis? quibus ego credo, nec possum Atti Navi quem commemor-
abas lituum contemnere, sed qui ista intellecta sint a philosophis debeo discere,
praesertim cum plurimis de rebus divini isti mentiantur. ‘at medici quoque’ (ita
enim dicebas) ‘saepe falluntur.’ quid simile medicina, cuius ego rationem video,
et divinatio, quae unde oriatur non intellego?
Moreover, where does this divination of yours come from? Who discovered
the cleft in the liver, who marked down the song of the crow, who marked
the lots? I myself trust these things, nor can I despise the staff of Attus
Navius whom you mentioned. But from philosophers I must learn how these
things were understood, especially when these diviners mislead in many
matters. ‘But doctors too are often mistaken’—for so you were saying. But
how are medicine (whose rationale I see myself ) and divination (I don’t
understand whence it springs) similar? (DND .)
A few sections later Balbus interrupts to complain that Cotta has not (as
was promised) allowed Balbus to defend himself on each point. Cotta
keeps changing the subject just when Balbus is about to jump in. Says
Balbus:
itaque maximae res tacitae praeterierunt, de divinatione de fato, quibus de
quaestionibus tu quidem strictim nostri autem multa solent dicere, sed ab hac ea
quaestione quae nunc in manibus est separantur; quare si videtur noli agere
confuse, ut hoc explicemus, hac disputatione quod quaeritur.
For that reason, matters of great import have passed by in silence: on
divination and on fate. You treat these questions in brief, but our Stoics
have much to say. Yet these subjects are distinct from the inquiry now at
hand. Therefore, if you please, stop going on in this mixed-up way, so
that we can make plain the matter under investigation in this discussion.
(DND .)
Here is the skipping of divination to which Quintus refers.
. Chrysippus and his Critics on Divination 
Why does Quintus want to fill this gap? It is because of the “power”
(vim) he thinks may lie in answers to the question of divination. (Div. .)
He says first that he reckons that if there are gods, then there is divination,
and that if there is divination, then there are gods. (Div. .) This, we will
agree, would be a remarkably powerful pair of conditionals, if they were
true. Marcus calls it an arx Stoicorum, “citadel of the Stoics.” (Div. .)
But Marcus also points out that the arx would be hard to establish: why
could there not be gods who allow us no divination, or why not prediction
of “events thought to be by chance” without any gods? (Div. .)
Quintus retreats to this:
mihi vero, inquit, satis est argumenti et esse deos et eos consulere rebus humanis,
quod esse clara et perspicua divinationis genera iudico.
For me there is sufficient evidence that there are gods, and that they care
about human affairs, because I judge that the well-known and obvious
kinds of divination are real. (Div. .)
It is important to see that Quintus has been reading Balbus carefully. For
his more modest position is that if there is divination, then there are gods
and they care for us. These are precisely the heads for which Balbus
recruited divination: that there are gods and that they care about us.
Quintus’ speech will fill in what the Stoics have to say for divination with
those theses, and with Cotta’s unanswered objections, in mind. How
precisely Quintus’ goal is only to continue the project of DND becomes
clear many pages later, when in section  of the second book of Div., the
startled reader learns that Quintus personally does not fully agree with the
Stoic position he has defended. For himself, Quintus accepts what he calls
“natural” divination, divination in dreams and by oracles, but he does not
accept “artificial” divination by augury, haruspicy, and the like. Much of
the speech, then, is a for-the-sake-of-argument defence of Balbus’ general
position in DND, and not precisely of Quintus’ own ideas. But, in my
view, Quintus defends Balbus’ general position by moving away from
Chrysippus’ position on divination. Let us first see what Chrysippus’
position was.
. Chrysippus and his Critics on Divination

Consider DL’s brief report of Stoic views on divination:
καὶ μὴν καὶ μαντικὴν ὑφεστάναι πᾶσάν φασιν, εἰ καὶ πρόνοιαν εἶναι· καὶ
αὐτὴν καὶ τέχνην ἀποφαίνουσι διά τινας ἐκβάσεις, ὥς φησι Ζήνων τε καὶ
Χρύσιππος ἐν τῷ δευτέρῳ Περὶ μαντικῆς καὶ Ἀθηνόδωρος καὶ Ποσειδώνιος
 Quintus’ Stoic Case for Divination
ἐν τῷ δυοδεκάτῳ τοῦ Φυσικοῦ λόγου καὶ ἐν τῷ πέμπτῳ Περὶ μαντικῆς. ὁ
μὲν γὰρ Παναίτιος ἀνυπόστατον αὐτήν φησιν.
Moreover, they say that every [sort of] divination is real, if providence too
exists; and they even declare that it is an art, on account of some outcomes,
as Zeno says, and Chrysippus in book  of his On Divination, and
Athenodorus, and Posidonius in book  of his Account of Physics, and in
book  of his On Divination. Panaetius, though, denies the reality of
divination. (., translation adapted from Inwood and Gerson ()
p. )
From this we can extract two claims.
C: If there is providence, every sort of divination is real.
C: We can see from some “outcomes” that divination is an art.
We should immediately note two distinctions between C and C. First,
C is a conditional, whereas C reports some looser form of inference.
Second, the conclusions are importantly different: in C, that divination is
real, in C, that it is an art. Since “providence” here must mean divine
providence, C seems to be a step towards establishing half of Quintus’
“Stoic citadel”: if there are gods, there is divination. But it is C that I shall
first take up here. What is the point of showing that divination is an art by
means of these “outcomes”?
When it comes to Chrysippus’ arguments for divination, we are in luck.
When in the fourth century the Christian Eusebius mounted his assault on
oracles in his Preparation for the Gospel, he turned the pagans against one
another by quoting at length the remarks against Chrysippus of a certain
Diogenianus. Diogenianus evidently launched a polemical attack on
Chrysippus’ On fate. In the passage at Preparation for the Gospel .,
Diogenianus discusses a proof of fate from Chrysippus, “stuffed with much
imbecility.” (πολλῆς εὐηθείας μεστόν, ..) Chrysippus, says

Who, and what, was Diogenianus? He is known to us only through Eusebius’ quotations, and even
the attribution of these sometimes depends only on the chapter headings surviving in the
manuscripts of the Preparation. On the question of what he was, see Gercke () –,
–, who first argued that Diogenianus was an Epicurean and not (as Zeller had thought) a
Peripatetic or the same Diogenianus we find in Plutarch’s On the Pythian oracle and Convivial
questions; Gottschalk ()  n.  with Mras () , for the argument that the chapter
heading to Preparation ., and Diogenianus’ willingness to countenance some idea of fate, suggest
that he was a Peripatetic and not an Epicurean; Hammerstaedt ()  n.  for a reply to
Gottschalk on behalf of the Epicurean identification. On balance, given his advocacy of Epicurus’
arguments about divination (Preparation .), his superior and abrasive tone in argument, and his
methodological remarks, it seems probable that Diogenianus was an Epicurean. His date is equally
unresolved. The (plausible) consensus is that he is of the second century AD, on the grounds of
parallels with Plutarch and Alexander (e.g. Hammerstaedt ()  n. ).
. Chrysippus and his Critics on Divination 
Diogenianus, put forward the following proof: “he says that the predic-
tions of diviners would not be true, if everything were not bound by fate.”
(..) The proof, evidently, rested on the falsity of the consequent – it is
not the case that the predictions of diviners are not true, so neither is it the
case that everything is not bound by fate. Diogenianus finds this risible
because, he says, it assumes that the truth of diviners’ predictions is either
evident or at any rate more likely to secure agreement than is the reality of
fate. So, how did Chrysippus support his silent premise, that “the predic-
tions of diviners are true”? Again, Diogenianus is scathing: Chrysippus
could only prove the reality of divination from the reality of fate. But
what manner of demonstration could be less sound than that? (..–)
Diogenianus insinuates with this that Chrysippus could not demonstrate
both fate from divination and divination from fate, because such demon-
strations advanced together would be circular, and viciously so.
Thus Chrysippus had reason to find some method to convince his
opponents, or at least those like Diogenianus, that there are relatively
evident reasons to grant the reality of divination. Luckily for us, Diogen-
ianus immediately proceeds to rubbish such a method – comparing
predictions with outcomes – and by doing so strongly implies that this
was Chrysippus’ own method:
τὸ γὰρ ἀποβαίνειν τινὰ κατὰ τὴν ἐνάργειαν, ὧν προλέγουσιν οἱ μάντεις,
οὐ τοῦ μαντικὴν ἐπιστήμην εἶναι σημεῖον ἂν εἴη ἀλλὰ τοῦ τυχικῶς συμ-
πίπτειν ταῖς προαγορεύσεσι συμφώνους τὰς ἐκβάσεις, ὅπερ οὐδεμίαν ἡμῖν
ἐπιστήμην ὑποδείκνυσιν.
For that some things evidently come true according to what the diviners
foretell is a sign, not of the existence of a divinatory science, but of the
chance concurrence of the outcomes with the predictions—a thing which
gives us no indication of any science. (Eusebius Preparation for the Gospel
.. = Diogenianus fr.  Gercke, translation based on Gifford ())

μὴ . . . ἂν τὰς τῶν μάντεων προρρήσεις ἀληθεῖς εἶναί φησιν, εἰ μὴ πάντα ὑπὸ τῆς εἱμαρμένης
περιείχοντο.

See Bobzien () – for a similar reading of this passage. The argument from divination to
providence at Sextus, Adversus mathematicos . has the same structure.

I take Diogenianus’ γάρ in line  to refer not to the immediately preceding question, but to his
claim that Chrysippus’ only available proof of divination would be from fate (–). What follows
is then intended to explain why another mode of support for divination, comparing predictions and
outcomes, is not legitimate. This suggests to me, as seems likely from the other evidence, that
Chrysippus employed this second mode.

“Us” here could be the Epicureans, or just people in general. An Epicurean point of comparison for
this passage as a whole could be Diogenes of Oenoanda fr.  Smith, but the text of the latter is
mostly conjectural.
 Quintus’ Stoic Case for Divination
So here we see criticized a demonstration of the sort that Diogenes Laërtius
called “on account of some outcomes” (., see pp. –).
Nor need we rely on Diogenes’ and Diogenianus’ mere implications
that Chrysippus proceeded in this way. For we know from Div. that he
collected two volumes of divinatory predictions, respectively from oracles
and from dreams. Nowhere in Div. is it explicit that Chrysippus system-
atically recorded outcomes along with the oracles and dreams, but it seems
very likely. For he provided the interpretations of the dreams, the most
likely use for which would be comparing prediction with outcome.
Moreover, at Div. . Marcus tells us a story:
defert ad coniectorem quidam somniasse se ovum pendere ex fascea lecti sui
cubicularis (est hoc in Chrysippi libro somnium); respondit coniector then-
saurum defossum esse sub lecto. fodit, invenit auri aliquantum, idque circum-
datum argento, misit coniectori, quantulum visum est de argento. tum ille:
‘nihilne’, inquit, ‘de vitello?’
Somebody told an interpreter that he had dreamt that an egg was suspended
from the upholstery of his bed (this is in Chrysippus’ book of dreams). The
interpreter responded that treasure was buried under the bed. The dreamer
dug, he found a certain amount of gold, and the gold was surrounded by
silver. He sent the interpreter what modicum of silver seemed right. The
interpreter said, “None of the yolk?”
The attribution of this story to Chrysippus is confirmed in a probably
independent source: Photius’ Byzantine Lexicon s.v. νεοττός, “yolk.”
Furthermore, Marcus will attack the worth of Chrysippus’ collection of
oracles in terms similar to Diogenianus’:
tuis enim oraculis Chrysippus totum volumen implevit partim falsis, ut ego
opinor, partim casu veris, ut fit in omni oratione saepissime.
For Chrysippus filled a whole roll with your oracles—some of them false,
I myself think, some true by chance, as happens very often in every kind of
speech. (Div. .)

Div. .: Chrysippus, qui totam de divinatione duobus libris explicavit sententiam, uno praeterea de oraculis,
uno de somnis, “Chrysippus, who set out his whole view on divination in two books, with an additional
one on oracles, and another on dreams.” Div. .: collegit innumerabilia oracula Chrysippus nec ullum
sine locuplete auctore atque teste, “Chrysippus collected countless oracles, not one of which lacked a
trustworthy source or witness.” Cf. Div. .. I suggest that the two supplementary volumes on dreams
and oracles which Cicero mentions in Div. . (uno praeterea . . . uno) consisted of Chrysippus’
collections of dreams and oracles with their interpretations and outcomes. Marcus says that
Chrysippus’ collection of oracles filled a totum volumen (Div. .) so if the supplementary volumes
were not these collections, we would have to suppose that there was yet another volume of collected
oracles which Cicero does not mention in Div. ..

The sentence continues: partim flexiloquis et obscuris, ut interpres eget interprete, et sors ipsa ad sortes
referenda sit; partim ambiguis, et quae ad dialecticum deferenda sit, “some of them are equivocal and
. Chrysippus and his Critics on Divination 
This also implies that Chrysippus gave the outcomes of the predictions
in his volume of oracles – why else would Marcus say that some were only
true by chance? Now, these testimonia from Div. do not refer to Chry-
sippus’ On fate, which is Diogenianus’ target. But we may assume that
Chrysippus’ procedure in On fate was consistent with his dedicated works
on divination. So it seems likely that in On fate Chrysippus presented some
predictions and their outcomes as empirical evidence for the reality of
divination, from which he then inferred the reality of fate.
But how could comparing some predictions and their outcomes give
Chrysippus his proof of the reality of fate? The premise he needs to secure
is that “the predictions of diviners are true.” If this is to be evidence for
Chrysippus’ all-encompassing determinism, the premise must in fact mean
that “all the predictions of diviners (qua diviners) are true.” How can
exhibiting some true predictions make the case, especially when it was true
(one would imagine) that in practice much divination went astray?
To see the answer, we must pay careful attention to what the outcomes
were supposed to establish. DL was explicit about this: Chrysippus
intended them to establish that divination is an art (pp. –).
Diogenianus is less helpful. He has already suggested two possible conclu-
sions in favor of divination, that the predictions are true and also that there
is divination, both of which lack DL’s precision. But I think this is just
carelessness on Diogenianus’ part. For he goes on to say that we do not
consider an archer “knowledgeable” or “scientific” (ἐπιστήμονα) when he
often misses but hits once, nor would we consider a doctor such if he kills
all but one of his patients, and that in general nothing is scientific that fails
in most of its proper tasks. (Eusebius, Preparation, ..) He goes on, of
course, to say that everyday life proves that divination mostly fails, and that
good evidence for this is the fact that those who “profess divination as an
art (τέχνην)” do not run their everyday lives on the advice of diviners.
opaque, so that the interpreter needs an interpreter; some of them are ambiguous, of the sort that
we should take to a logician.” The point is presumably that the alleged link, forged by interpretation
from an obscure oracle to an outcome, was suspicious. This would help to explain why Marcus
seems to think here that the truth of the predictions cited by Chrysippus was a matter of personal
opinion.

Bobzien ()  understands this to be Chrysippus’ commitment, and sees support for this
interpretation in the arguments around the necessity of divinatory theorems at Cicero, On fate
–.

There must have been more to Chrysippus’ argument that divination entails fate than has survived.
Even if all the predictions of diviners are true, that does not show that everything about the future is
fated now. It could just entail a very wide fatalism, whereby very many things but not everything are
fated beforehand. Presumably the argument was that this sort of divinatory fatalism implied a
degree of divine control of the future that could only be true if god in fact fated everything.
 Quintus’ Stoic Case for Divination
(Eusebius, Preparation, ..) These comments strongly suggest that
Chrysippus’ own claim was precisely that divination is an art or a science,
because what they imply against Chrysippus is that divination is not an art
or a science. I agree with Bobzien () p. , that “we can assume that
Chrysippus, if asked, would have explicated his claim [that the predictions
of diviners are true] as follows: the theorems and predictions of the seers
are true provided that they master their science and have not made
interpretational mistakes in the individual instances of prognostication.”
That is, all predictions successfully made by the science of divination will
be true. Pointing to unsuccessful predictions does not show that divination
is not a science, because in real life those who apply the science are fallible.
A prediction produced by an unsuccessful application of divinatory sci-
ence, or of a poorly constructed “divinatory” art, is not really a prediction
of divinatory science. So in this sense it is plausible that “all the predictions
of diviners qua diviners are true,” since those that are false are not properly
predictions of practitioners of divinatory science. Chrysippus catalogs his
outcomes, then, to show that there is a science of divination, even if a
science that is fallibly researched and applied. Note that Diogenianus’
reply is neatly targeted against this argument. He not only points out that
diviners sometimes go astray, he claims that they go wrong in the over-
whelming majority of cases. If their performance were indeed “no better
than chance” then Chrysippus would not even have a plausible empirical
case for the reality of his art.
There is further, positive evidence that Chrysippus considered divin-
ation in general an art in his definition of divination. I think that we have
three sources for definitions attributable to Chrysippus, although two do
not explicitly attribute their definitions to him. Marcus gives the earliest
version in Div. .:
Stoici autem tui negant quemquam nisi sapientem divinum esse posse. Chry-
sippus quidem divinationem definit his verbis: vim cognoscentem et videntem et
explicantem signa, quae a dis hominibus portendantur; officium autem esse eius
praenoscere, dei erga homines mente qua sint quidque significant, quem ad
modumque ea procurentur atque expientur. idemque somniorum coniectionem
definit hoc modo: esse vim cernentem et explanantem, quae a dis significentur in
somnis.
Your Stoics deny that anyone other than the sage is able to have the capacity
to divine. Chrysippus, indeed, defined divination with these words: a power
both recognizing, and seeing, and interpreting signs which are portended by
gods to humans; but he said that its role is to foresee how the gods are
disposed towards humans and what they are signifying, and how the things
. Chrysippus and his Critics on Divination 
signified should be attended to and expiated. The same Chrysippus
defined the interpretation of dreams this way: that it is a power discerning
and making clear things signified by gods in dreams.
A very similar definition appears in two Greek sources but is not
attributed to Chrysippus in particular. First Sextus Empiricus, Adversus
mathematicos .:
εἰ μὴ εἰσὶ θεοί, οὐδὲ μαντικὴ ὑπάρχει, ἐπιστήμη οὖσα θεωρητικὴ καὶ
ἐξηγητικὴ τῶν ὑπὸ θεῶν ἀνθρώποις διδομένων σημείων, οὐδὲ μὴν
θεοληπτικὴ καὶ ἀστρομαντική, οὐ λογική, οὐχ ἡ δι’ ὀνείρων πρόρρησις.
[According to the Stoics:] If there are not gods, divination is not real either
(divination is a science which can observe and interpret signs that are given
by gods to men), nor is [divination by] divine possession, or divination by
the stars, or by words, nor prediction from dreams.
Stobaeus gives two slightly different versions of a similar (but not identical)
definition:
εἶναι δὲ τὴν μαντικήν φασιν ἐπιστήμην θεωρητικὴν σημείων τῶν ἀπὸ θεῶν
ἢ δαιμόνων πρὸς ἀνθρώπινον βίον συντεινόντων· ὁμοίως δὲ καὶ τὰ εἴδη τῆς
μαντικῆς.
[The Stoics say that only the sage is a good diviner, orator, etc..] And they
say that divination is a science which can observe signs from gods or
daimones, which contribute to human life (and similarly for the kinds of
divination). (Stobaeus ..b = SVF ..–)
Καὶ μαντικὸν δὲ μόνον εἶναι τὸν σπουδαῖον, ὡς ἂν ἐπιστήμην ἔχοντα
διαγνωστικὴν σημείων τῶν ἐκ θεῶν ἢ δαιμόνων πρὸς ἀνθρώπινον βίον
τεινόντων. Δι’ ὃ καὶ τὰ εἴδη τῆς μαντικῆς εἶναι περὶ αὐτόν, τό τε ὀνειρ-
οκριτικὸν καὶ τὸ οἰωνοσκοπικὸν καὶ θυτικὸν καὶ εἴ τινα ἄλλα τούτοις ἐστὶ
παραπλήσια.
And [the Stoics say] that only the virtuous man has the capacity to divine,
in as much as he has a science which can distinguish signs reaching from

At first sight we might take procurentur and expientur to be two opposite reactions to whatever is
signified, the first referring to the promotion of welcome outcomes, the second to the expiation of
unwelcome ones. But in haruspicy procuro seems to be the term for the management of the right
reaction to prodigia, cf. Div. . with Pease (–) ad loc. Prodigia generally signal a negative
religious development requiring remedial action. So I think procurentur atque expientur both
describe the reaction to divine messages generally, and if anything to messages demanding action
to avoid a negative consequence.

coniectio is a puzzle. It must mean something like “interpretation,” but this use is unique (Pease
(–) ad loc.). It more often means “summary” or “summarizing” (see OLD or Lewis and
Short () s.v. coniectio). It presumably had some connection in Cicero’s mind with coniectura
(see pp. –), but what and how close this connection was is hard to estimate.
 Quintus’ Stoic Case for Divination
gods or daimones to human life. Hence also that the kinds of divination
concern him, the dream-interpretative kind, and the augural, and the
extispicine, and any others which there are in addition to these. (Stobaeus
..s = SVF .)
On the basis of Cicero’s explicit attribution to him, I think that all these
definitions are due ultimately to Chrysippus.
The obstacle to tracing the Greek definitions to Chrysippus is a key
difference between Marcus’ version in Div. Book  and those of Sextus and
Stobaeus. Marcus has Chrysippus call divination a vis, which should gloss
Greek δύναμις, “power” or “capacity,” while the Greek authors have “the
Stoics” call it an ἐπιστήμη, a “science.” But this is not so great a difficulty
as it at first appears. The point of Marcus’ passage is to ask Chrysippus
whether this “power” demands a sage or could be held by a fool. “What
then?” he asks, “Is there a need here for ordinary prudence, or for both
outstanding talent and perfect education?” (Div. .) Chrysippus,
I think, would be happy to take the latter option – that the only truly
successful, infallible diviner, possessor of the right science, would be a sage.
Marcus’ point is indeed that the real divinatory “power,” for Chrysippus, is
a science. So when Marcus has Chrysippus call divination a “power,” this
does not mean that for Marcus Chrysippus defined divination as a “power”
rather than a science. Rather, he called it a power and a science. Next,
why does Chrysippus define divination as a science, when in other places
we have seen him call it an art? For Chrysippus, the answer is that in an
ideal case, when it is known by the sage, the art (ars, τέχνη) will amount to
a science, a stable body of scientific knowledge (scientia, ἐπιστήμη). But in
the case of ordinary, fallible people, who have not achieved ideal scientific
knowledge, the same body of beliefs amounts only to what Chrysippus
would call an art. The art of divination, then, is what is relevant in
practice, and what might be established by its documented outcomes.
One consequence of Chrysippus’ definition is that divination is not only
an art of prediction. For it is concerned with any sign portended from the
gods, not only signs that predict the future. Indeed the egg dream turned
out to be a sign not of the future, but about the present: the diviner’s reply
was simply that was there treasure under the bed (see p.  above).
Quintus even gives an example, which I suggest is likely owed to Chry-
sippus, in which Sophocles solved the mystery of a theft from Hercules’

quid ergo? ad haec mediocri opus est prudentia an et ingenio praestanti et eruditione perfecta?

In thinking that we do not have to choose between a definition with power or one with science
I agree with Lévy () –, and Bobzien ()  n. .
. Chrysippus and his Critics on Divination 
temple when the god himself appeared in a dream and named the guilty
party. (Div. .) Sophocles had divinatory information about the past
theft. For Chrysippus, the purpose of divination is a sort of conditional
prescience (praenoscere, Div. ., pp. –). We need it to see how
the gods are disposed towards us now, so that we know how to attend to
their wishes. What we are to foresee is presumably what will happen if we
do, or do not do, what the gods wish. But in order to achieve this purpose,
divination is such that it can deliver news about the present and past as
well as about the future.
We know something about what Chrysippus’ ideal divinatory science
would have looked like. This is thanks to an argument in Cicero’s On fate,
sections –. For Chrysippus it was important that some future-tense
truths be true but possibly false, else we would lose responsibility for our
actions. For he thought that if I am contemplating, for example, a
donation to charity, my future decision on the matter is physically deter-
mined now. But he also thought that if it were not even possibly true that
I shall do otherwise than I shall do, then I would not be responsible for my
donation or refusal. But Chrysippus thought I am responsible for this
future choice. Thus, despite his physical determinism, he concluded that
true claims about our future actions could be false. The Marcus of On fate
tells us that Chrysippus fought with an opponent, Diodorus Cronus, over
whether divinatory arts threatened this intricate position on fate and future
contingency. For, says Marcus, a theorem of a divinatory art looks like this:
If somebody was born at the rising of the dog-star, he will not die at sea.
But this implies:
If Fabius was born at the rising of the dog-star, Fabius will not die at sea.
But Chrysippus allowed that all past truths were necessary. Therefore the
antecedent of the “Fabius” conditional was necessarily true. Now, says
Marcus, Chrysippus also thought that necessity was “transmitted” across a
conditional: if the antecedent is necessarily true, it follows for him that the

In making no reference to prediction of the future, Chrysippus’ definition of μαντική or divinatio
might seem oddly different from Cicero’s own working definition in the preface to Div. (.) as well
as Quintus’ own definition (see pp. –). But Santangelo ()  points out that although
the verb divinare is older, the noun divinatio itself does not appear before Cicero. See Santangelo
() – and () –, for more on the history of the term. Thus for Cicero the Latin
term may have been a flexible one.

For this argument, see Bobzien () Chapter .

Si quis oriente Canicula natus est, in mari non morietur.

Si Fabius oriente Canicula natus est, Fabius in mari non morietur.
 Quintus’ Stoic Case for Divination
consequent is necessarily true. So it follows for Chrysippus that it is
necessarily true that Fabius will not die at sea. Therefore Diodorus was
able to claim that, according to Chrysippus’ principles, it is not possibly
false that Fabius will not die at sea, which Chrysippus cannot accept.
Chrysippus’ solution, it appears, was to insist that the conditional of a
divinatory theorem was not, by his standards, a proper conditional. That
is, it should not (so to speak) transmit necessity from antecedent to
consequent. Rather the divinatory theorem should be analyzed as a neg-
ated conjunction, which elucidates the interrelation of the truth and falsity
of the two claims, but does not transmit necessity from one to the other.
Thus rewritten, the general conditional above would become:
It is not true both that a given person was born at the rising of the dog-star, and
that that person will die at sea.
Chrysippus’ analysis here would allow the truth of the proposition about
the time of his birth to be necessary, but also that the proposition about his
safety is both true and possibly false. Thus Diodorus could not use
divination against Chrysippus’ view that future truths could be false.
Presumably Chrysippus did not mean that real-life diviners ought to
rephrase their handbooks in his convoluted terms. Rather, he meant that
a wise diviner would understand that this, and not Chrysippan condition-
als, is what true divinatory theorems amount to.
This rather technical discussion gives us a window onto what the
practical use of Chrysippus’ ideal divinatory science would look like. We
saw that Chrysippus called divination a science for discerning and inter-
preting “signs” portended from gods to men. Now we can add that
according to the Stoics “signs” are “assertibles” – propositions with truth
value – like the one about Fabius’ birth at the rising of the dog-star. So in
the Fabius example, the astrologer can artfully tell that it is true that
“Fabius was born at the rising of the dog star” – that is, he discerns the
sign. Next, he finds a relevant general theorem. However this may be
phrased in practice, Chrysippus thinks that strictly speaking the theorem
amounts to, “It is not true both that a given person was born at the rising
of the dog-star, and that that person will die at sea.” From the combination
of the sign and the general theorem the diviner draws a sound inference,
that is, he interprets the sign. In Fabius’ case, our ideal diviner infers a
prediction: “it is not true that Fabius will die at sea.”

Non et natus est quis oriente Canicula, et is in mari morietur.
. Chrysippus and his Critics on Divination 
In sum, then, Chrysippus had reasons to make the limited claim that
divination was a successful science or art. First, he claimed that from the
reality of divination should follow two key Stoic theses: that everything
happens by fate, and that there are providential gods. Second, he could
avoid dialectically vicious circularities if he supported the claim that
divination is real with empirical evidence, rather than by theoretical
arguments. Accordingly, he collected this evidence, pairs of divinatory
interpretations and their outcomes. He argued that this empirical evidence
supported the limited thesis that divination was an art. By this he meant
that diviners can assemble collections of theorems connecting kinds of
signs and kinds of outcomes, of the form “Not both sign and not outcome,”
such that when the diviner observes that some such sign is true, he may infer
that a correlated outcome is true. Chrysippus then concluded that divin-
ation is real, and thus that there is fate and there are providential gods. It is
against some such strategy as this that Diogenianus seems to inveigh.
Diogenianus was not the only one. Along with a wider attack on
Chrysippus’ divinatory views (Div. .), in the century after Chrysippus
Carneades seems to have attacked the notion that divination was a science
or art. He pressed the objection “why do these things happen and by what
art can they be grasped”? (cur haec ita fiant aut qua arte perspici possint?,
Div. ..) Does Jupiter order that crows and ravens should croak on the
right or the left? (Div. .) Here Carneades demands that there be more
to Chrysippus’ defence of the divinatory art than the outcomes. For the
outcomes can show, at best, that the regularities specified by the predictive
theorems obtain. They cannot show anything about why these theorems
are true (“why do these things happen?”) or, epistemologically, how they
came to be cognised (“by what art can they be grasped?”). Carneades’
implicit premise seems to be that if divination were a real art or science it
should be able to supply higher-order theorems about the causes that
explain the truth of its general predictive theorems, and why those pre-
dictive theorems may be grasped cognitively and admitted to the art.
Marcus attributes to Carneades a second and more elaborate objection
to Chrysippus. (Div. .–) This asks what the subject matter of divin-
ation might be (quarumne rerum divinatio esset). I shall deal with this
objection in detail in my next chapter, when I study Marcus’ speech
(pp. –).
In DND, Cotta’s further question of where divination comes from is an
amplification of these sorts of objections. He asks who discovered the
defects of the liver that the haruspices used, who marked augural signs
and how these things were understood, especially when diviners tend to
 Quintus’ Stoic Case for Divination
mislead. (DND ., p.  above.) What he demands is not, like
Carneades, causal explanation for the predictive theorems. Instead he asks
for a developmental history of the divinatory art. This seems a telling
objection to divination in a way that it would not be against an art like
medicine. For divination, at least at Rome, was arcane. It is obvious why
somebody who cares for the sick might start to note correlations between
treatments and cures. But why on earth would anybody think that she
should attend to what follows upon the flight of birds, or to the condition
of the offal of a sacrifice? Further, how could such oddities come to be
understood precisely enough for the right divinatory theorems to be
adopted, especially in the absence of an undergirding causal theory?
We are now equipped with some helpful background to Quintus’ speech.
The Stoics, of whom Balbus is an example, thought that they could use
empirically plausible divination as evidence (among other things) for the
existence of gods, and for divine providence over human affairs. They
typically supported the claim that divination was an art by appeal to
“outcomes,” that is, by collections of divinatory predictions that came true.
In DND, Balbus gestures at, but does not give, this sort of support in his two
arguments from divination. Cotta in return poses a challenge that was well
rehearsed: why should he accept an art that can give no theoretical expla-
nantions of its subject, and what history could be given for such an “art” –
unde oriatur?, “where does it spring from?” – that would warrant thinking it
an art at all? (DND .) Balbus says that there remain extensive Stoic
answers to these questions (DND .), and it is some of these answers
which Quintus will give. Quintus’ answers represent, I think, not a rehearsal
of Chrysippus’ views as I have interpreted them. Rather, they reflect a new
approach, intended to reply to, or to avoid, the objections laid against
Chrysippus by critics like Carneades and Cotta. It is to this I turn next.
. Quintus’ New Stoic Case for Divination

In Div. at large there are two sorts of reason why one might believe in
divination. In speaking in his preface of the ancient acceptance of divin-
ation, Cicero says, “as I myself think, the ancients approved [divination]
more because they had been moved by the outcomes (eventis) than because
they had been informed by theory (ratione).” (Div. .) These two

atque haec, ut ego arbitror, veteres rerum magis eventis moniti quam ratione docti probaverunt. At least
for the case of divination, Cicero does not accept Balbus’ contention that the founders of Roman
cults worked from some understanding of physics (Chapter  section .).
. Quintus’ New Stoic Case for Divination 
modes of persuasion form one of the two main structuring principles for
Quintus’ speech. He makes grand claims at the outset that he need only
analyze outcomes to make the case for divination. (Div. .–) What he
does not need to analyze are the causes involved, and it is clear that
analyzing the causes would be a paradigm case of giving a theory.
quarum rerum eventa magis arbitror quam causas quaeri oportere. est enim vis
et natura quaedam, quae tum observatis longo tempore significationibus, tum
aliquo instinctu inflatuque divino futura praenuntiat.
I think it is more helpful to investigate the outcomes of this business [i.e.
the various kinds of divination] than the causes. For there is, at any rate,
some natural power that predicts the future, both by signs observed over a
long period of time, and by a sort of divine stimulation and inspiration.
(Div. .)
In the last sentence we see Quintus decline to give a causal explanation of
divination. If he is right, there must be some factor or factors in the natural
world that produce the connections between the future and today’s
birdsong, or last night’s dream, from which divinatory predictions are
made. If there were not, divination would not work as (according to
him) it evidently does. Given that Quintus thinks that the reality of
divination entails the reality of the gods, these factors in the natural world
must either be, or require, the gods. But it does not matter to Quintus
whether the Stoics know what those factors are.
Quintus explains how outcomes are to convince us of the reality of
divination, by appeal to analogies with the testing of herbal remedies, and
with prognostication of the weather.
mirari licet quae sint animadversa a medicis herbarum genera, quae radicum
ad morsus bestiarum, ad oculorum morbos, ad vulnera, quorum vim atque
naturam ratio numquam explicavit, utilitate et ars est et inventor probatus.
We may wonder at the variety of herbs that doctors have observed: which
roots are helpful for animal bites, which for diseases of the eyes, which for

eventa or “outcomes” is probably a translation of ἐκβάσεις – compare DL . (see pp. –)
with Diogenianius ap. Eusebius Preparation for the Gospel .. (see p.  above). It is Quintus’
term of art for the state of affairs which the divinatory prediction predicts. The careful use of this
term is intended to avoid the implication that the sign, prediction, and outcome are causally related
to one another.

Quintus is explicit that prognostication is not divination (ex alio genere, “of another kind”, Div.
.). This might seem puzzling, especially as he seems to think that prognostication is almost
equally hard to explain. The answer is that prognostic theorems are in principle explicable without
invoking the divine, even though in many cases we cannot see these explanations. For Boethus
made some progress in this regard, Quintus says. (Div. .)
 Quintus’ Stoic Case for Divination
wounds? Reason has never explained the power or nature of these plants.
Rather, their usefulness has proved each discoverer, and his art. (Div. .)
Or again:
ne hoc quidem quaero, cur haec arbor una ter floreat aut cur arandi matur-
itatem ad signum floris accommodet; hoc sum contentus, quod, etiamsi, cur
quidque fiat, ignorem, quid fiat, intellego. pro omni igitur divinatione idem,
quod pro rebus iis, quas commemoravi, respondebo.
Nor do I inquire even about this: why one tree flowers three times, or
why it adjusts its flowering as a sign of the right season to plough. I am
content with this: that, even if I do not know why each thing happens,
I understand what happens. So I shall give the same answer on behalf of
every kind of divination, as I have given on behalf of the facts I have just
mentioned. (Div. .)
Thus Quintus will follow something like the latter of the two strategies
that DL gives the Stoics on divination (see pp. –): he will support
divination with its outcomes, not by trying to give an explanatory theory.
Thus far, Quintus resembles Chrysippus.
But what has not been noticed, I think, is that, in other respects,
Quintus’ argument for divination is substantially different than Chrysip-
pus’. Quintus begins his speech with the distinction, on which he will
insist as the other structuring principle for the rest of his discussion,
between two kinds of divination: duo sunt. . . divinandi genera, quorum
alterum artis est, alterum naturae, “[t]here are two kinds of divining, of
which one is artificial, the other natural.” (Div. .) He gives examples of
each kind: the artificial kind includes the predictions of “entrail diviners
or of those who interpret prodigies or lightning [i.e. the haruspices], or
of augurs, or of astrologers, or of lots.” The natural kind, meanwhile,
includes the predictions of “dreams or oracles.” (aut somniorum aut vati-
cinationum, Div. .) Now, this distinction is at odds with Chrysippus’
case for divination. For Chrysippus argued that divination in general was
an art. But Quintus will argue that only one kind of divination is, or rather
uses, an art, and that the other is not and does not.

The word order of this clause is difficult, and the text may be corrupt (see Schultz  ad loc.).
I have translated to give what seems to be the intended sense.

The mastic tree, which was considered to flower at three times of year when ploughing was
desirable. For the evidence, see Pease (–) ad loc. Aratus included this prognostic sign in
his classic collection (ll. –), and Quintus has just quoted Cicero’s translation of
Aratus’ lines.

aut extispicum aut monstra aut fulgora interpretantium aut augurum aut astrologorum aut sortium.
. Quintus’ New Stoic Case for Divination 
It does not appear that Cicero invented this Stoic distinction into two
kinds of divination. For confirmation that there was a Stoic view that
some, “natural” divination requires no art, and thus that this was a Stoic
view distinct from Chrysippus’ own, comes from what appears to be an
independent source. This is the Essay on the life and poetry of Homer
spuriously attributed to Plutarch, a work “largely impossible to date”
according to its editors, but probably of about the third century AD.
The author of this text would have us think that Homer displays encyclo-
pedic learning on all subjects, to include divination, and remarks:
ταύτης μέντοι τὸ μὲν τεχνικόν φασιν εἶναι οἱ Στωικοί, οἷον ἱεροσκοπίαν καὶ
οἰωνοὺς καὶ τὸ περὶ φήμας καὶ κληδόνας καὶ σύμβολα, ἅπερ συλλήβδην
ὄτταν καλοῦμεν, τὸ δὲ ἄτεχνον καὶ ἀδίδακτον, τουτέστιν ἐνύπνια καὶ
ἐνθουσιασμούς.
The Stoics say that of [divination] there is on the one hand the artificial
[kind], for example extispicy, and birds, and that concerned with sayings
and proverbs and tokens, which together we call ‘voice omens[?],’ and on
the other there is the non-artificial and unteachable [kind], i.e. dreams and
ecstasies. (Ps.-Plutarch Essay on the life and poetry of Homer, )
“The Stoics” called “non-artificial and unteachable” the kinds of divination
that Quintus calls “natural” as opposed to “artificial.” But for Chrysippus
all species of divination were artificial. Thus the Stoics whom pseudo-
Plutarch claims to record seem to agree with Quintus, but they cannot
have included Chrysippus.

Div. . also gives a richer context for the distinction between artificial and natural divination in
earlier Stoic discussions, specifically in discussion of causal explanations of divination, in a way that
strongly suggests that Cicero took the distinction from some post-Chrysippan Stoic theory. The
Stoic(s) who made the distinction were not the first or only philosophers to see a difference between
artificial and inspired divination. Thus even in his own day, Chrysippus may have been unusual in
looking to provide a single account of all forms of divination. In his Phaedrus passage on μανία,
mania, “madness,” and μαντική, mantikē, “divination,” Plato distinguishes inspired divination from
some artificial kinds. (b–d) From Div. it seems that the Peripatetics supported natural but not
artificial divination, and thus invoked the distinction. From Aristotle’s own texts it is not quite clear
what his line was. Some sort of veridical dreaming is the only “divination” that receives his explicit
attention, and he hardly endorses it ringingly (see his On divination in dreams, cf. p.  above).
But he never condemns other sorts of divination out of hand, and when discussing the nose in the
History of animals he remarks that a sneeze is “the only breath that is a sacred and augural sign.”
(σημεῖον οἰονιστικὸν καὶ ἱερὸν μόνον τῶν πνευμάτων, b–)

Keaney and Lamberton () . It is of course possible that the author of the Essay owes this
account of Stoic divinatory theory to Cicero, whether directly or indirectly. The real Plutarch, after
all, must have read plenty of Cicero’s writings.

ὄτταν is an odd word in the MS which Keaney and Lamberton () leave untranslated. They
suggest, p.  n. , that it is related to ὄττεία, divination from sonic omens (cf. ὄσσα, “voice”).
 Quintus’ Stoic Case for Divination
At first sight the distinction between artificial and natural divination is
very puzzling. In what sense is divination by dreams more “natural” than
watching the flight of birds? Divinely inspired dreams might seem some-
what supernatural, while augury involves looking closely at the natural
world. Again, why would augury and the rest be more “artificial” than
divination by dreams? Divination by dreams often involved skilled inter-
pretation, as Chrysippus maintained and as we see, two centuries after
Cicero, in Artemidorus’ detailed manual, the Oneirocritica or Interpretation
of dreams. Further, what does Quintus mean when he makes “conjecture”
part of artificial divination? “Conjecture” appears by definition to be the
result of extempore guesswork and not of an art. And finally, in making
this distinction, does not Quintus start to give an explanation of the
mechanisms of divination, something he claims repeatedly to eschew in
the first part of his speech?
To find a satisfying answer to these puzzles we need to reflect on what
notion of an “art” Quintus has in mind. Although he disagrees with
Chrysippus about what a Stoic should say that divination is, Quintus
seems to share Chrysippus’ idea of what a divinatory art will look like
when written down. It will be a collection of theorems correlating kinds of
sign with kinds of outcome. Collections of such theorems indeed survive
from antiquity, like Artemidorus’ Interpretation, or the handbooks due to
the astrologers Dorotheus of Sidon and Vettius Valens. Dorotheus’ poem
(which survives in Arabic) consists predominantly of conditionals relating
astrological circumstances at a relevant hour to outcomes. Vetteius too
gives many passages of correlative theorems. These examples, admittedly,
are not part of Roman state divinatory practices, like augury or haruspicy.
But in fact Quintus mentions books associated both with the haruspices
and with the augurs, and these could easily have included such collections
of theorems, or something close enough for Quintus to make his point.
(Div. .)
Thus I think Quintus distinguishes between a kind of divination that
consists in the use of an “art” of this sort, and a kind of divination that does
not consist in the use of such an art, and another kind which is “natural.”
I suggest that the most satisfying way to understand this distinction is that
it is based on where predictive meaning is first encountered in the respective
kinds of divination. By “predictive meaning” I mean roughly “propos-
itional content about a future event.” In Stoic terms, this will be a future-
tense “assertible” (ἀξίωμα), on which more in a moment. Straightforward,
 
Pingree () passim. Pingree () passim.
. Quintus’ New Stoic Case for Divination 
spoken or written, future-tense language will be the most obvious example,
as in a prediction given by an augur, but many other kinds are possible,
like meaningful images experienced in a dream, or found figuratively in the
ravings of a frenzied prophet.
Consider how predictive meaning is found in artificial divination. An
event occurs, like the croak of a raven to your right when you are watching
for augural signs. This event has no intrinsic predictive content. There is
nothing in an ordinary bird singing that means anything about what will
happen to you tomorrow, no matter how closely you scrutinize it. But (in
principle) a true divinatory art might have a theorem that says “if a raven
croaks on the augur’s right, the city will lose the war,” or, analyzed more
precisely as Chrysippus would prefer, “not both a raven croaks on the
augur’s right and the city will not lose the war.” A diviner can take this
theorem, “plug in” the sign of the calling bird, and conclude that the
negation of the right-hand side of the theorem is true, too. Now she has
generated some predictive content – “the city will lose the war” – where
before there was none. The predictive content is first and only found once
the diviner has referred, literally or metaphorically, to her book, that is to
say, to the appropriate theorem of her art. The content is found artificially
because it is encountered first not in the sign, but on the right-hand side of
the relevant theorem of the art, which informs the prediction of the
outcome that the diviner should infer from the conjunction of the true
sign with the true theorem.
What about natural divination? In the natural case, natural phenomena
like dreams and the visions of the frenzied already contain the predictive
content. So to speak, they already mean the prediction. Sometimes, in the
case of dreams that require no interpretation, this is obvious. An example
in Div. is the dream of Gaius Gracchus, in which he was simply told by his
dead brother Tiberius that he (Gaius) must die the same death as him
(Tiberius). (Div. .)
Now Chrysippus will say, why then do dreams and oracles so often need
expert interpreters? Quintus’ reply will be that often the predictive mean-
ing of dreams and oracles is obscure or ambiguous. On these occasions,

On signs, see Allen () and (). In this account of the artificial divination I agree
approximately with Denyer’s assignment of “non-natural meaning” to artificial divinatory signs,
although I disagree with him that signs (qua signs) are outside the normal causal order of nature. On
the other hand, I do not agree that dreams and oracles also have non-natural meaning, as he seems to
think, since he does not limit his account to artificial divination (Denyer () –). I think the
point of calling natural divinatory events “natural” is that they have something like “natural
meaning.”
 Quintus’ Stoic Case for Divination
interpretation is needed. Quintus is happy to admit that this interpretation
is, or can be, artificial:
haec somniantis est divinatio. hic magna quaedam exoritur, neque ea naturalis,
sed artificiosa somniorum interpretatio eodemque modo et oraculorum et vati-
cinationum. sunt enim explanatores, ut grammatici poëtarum. nam ut
aurum et argentum, aes, ferrum frustra natura divina genuisset, nisi eadem
docuisset, quem ad modum ad eorum venas perveniretur, nec fruges terrae
bacasve arborum cum utilitate ulla generi humano dedisset, nisi earum cultus
et conditiones tradidisset, materiave quicquam iuvaret, nisi consectionis eius
fabricam haberemus, sic cum omni utilitate, quam di hominibus dederunt, ars
aliqua con iuncta est, per quam illa utilitas percipi possit. item igitur somniis,
vaticinationibus, oraclis, quod erant multa obscura, multa ambigua, explana-
tiones adhibitae sunt interpretum.
[The soul in sleep sees everything in nature.] This is the dreamer’s divin-
ation. At this point something important arises: interpretation that is not
natural but rather artificial, in the same way as interpretation of oracles and
prophecies. For there are explicators [of dreams and oracles] in the same
way as there are commentators on the poets. For just as divine nature would
have made gold and silver, copper or iron in vain, if she had not taught us
how we might come at the veins of these metals, nor would it have been any
use for her to have given the crops from the soil, or the timber of trees, to
the human race, unless she had handed down how to plant and look after
them . . . just so there was joined with every useful thing that the gods gave
to humans an art by which its utility could be grasped. In the same way, the
explications of interpreters are consulted for dreams, prophecies, and
oracles, many of which are opaque or ambiguous. (Div. .)
Obscure dreams or prophetic visions, then, are like a foreign language or
a difficult poem. They already mean something, just as metals have useful

Some editors have thought that this sentence (sunt. . . poëtarum) is an intrusive gloss. But this is not
convincing (see Schultz  ad loc.), and the sentence seems to me to be a particularly apt
comparison with the artificial interpretation of natural divinatory experiences, where the latter
already have meaning, even if difficult or “poetic” meaning.

Compare here some of Aristotle’s remarks. For him the most artful interpreter of dreams is one who
can “discern likenesses.” (On divination in dreams b–) Aristotle’s metaphor is the perception
of reflections in disturbed water. (b–, cf. On dreams a–) This suggests that he
thinks that dream-interpretation is a matter of discovering and making sense of confused content,
not generating predictive content on the basis of signs with no predictive content. Cf. further Allen
() : “[Dreams, frenzied pronouncements and certain kinds of oracles] differ from lightning
bolts, fissured livers, and stellar configurations studied by long observation, as well as the sweating
statues and hungry mice that occupy conjectural divination, by directly representing the events they
forecast. Someone carried away by divine afflatus predicts that an event will occur more or less by
saying that it will, a dream either by allowing a god to appear and tell the dreamer that it will or by
depicting its occurrence.” I agree with this sketch, except in that I disagree about the scope of
coniectura, see pp. –.
. Quintus’ New Stoic Case for Divination 
properties even before they are mined. If a dream or a vision needs
interpretation, the role of the interpreter is not to generate from the dream
or vision content it did not already contain. Rather, interpretation’s role is
to extract or to clarify the content that is already there. Thus the fact that
divinatory dreams and oracles often need an interpreter is no barrier to
Quintus’ distinction. For Quintus natural divination is natural because it
occurs when, without the intervention of human art, content predictive of
future chance events arises through natural processes in the soul of a
dreamer or of one in frenzy, even if that predictive content requires an
interpreter or, so to speak, a translator, in order to be understood.
Why would just these divinatory phenomena, dreams and the visions of
oracles and the frenzied, carry meaning? If they can, why could not the
flight of a bird? The distinction rests, I suggest, on the connection the
Stoics draw between reason and language. To become rational gives one
the ability to say “sayables,” the meaning or content of utterances. Among
sayables are “assertibles,” to include predictive propositions. Once we are
rational, all of our sense- and imaginative impressions have “sayable”
content (see p.  n. ). Thus our dream and frenzy visions naturally
“say” things too: they are images in a rational mind. But a bird, or its
flight, is not rational, and neither is nor produces speech or meaningful
images. It cannot say anything.
Now if I have interpreted Quintus’ distinction between artificial and
natural divination correctly, it is at odds with Chrysippus’ case for divin-
ation in such a way that Quintus needs a different definition of divination.
For Chrysippus defined divination as an art or (when the art is in the hands
of a sage) a science (see pp. –), while Quintus describes only part of
divination as scientific or artificial. Neatly, it turns out that Quintus does
indeed have a different definition of divination. Although we find other
versions of this definition in Div., the most careful version and Quintus’
own is at .: [divinatio], quae est earum rerum quae fortuitae putantur
praedictio atque praesensio, “[divination], which is the prediction and
foreseeing of those states of affairs that are thought to be by chance.”

Repici () – also distinguishes these two definitions of divination.

The other versions are given in Cicero’s authorial voice at Div. . and by Marcus at Div. .. At
.: . . . praesensionem et scientiam rerum futurarum, “the foreseeing and knowledge of future
matters”; the differences from Quintus’ version are easily explained by the place of this definition
in the first sentence of the treatise, where we would expect the subject to be defined loosely. At Div.
.: quam dicebas praesensionem esse rerum fortuitarum, “which you [Quintus] were saying was the
foreseeing of chance states-of-affairs.” In this context Marcus looks to put pressure on the Stoic
account of divination by pointing out that they thought nothing happened by chance.
 Quintus’ Stoic Case for Divination
There is no mention of an art or a science in this definition. Rather,
divination consists in a kind of praedictio atque praesensio, in “prediction
and foreseeing.” In the case of artificial divination, “prediction” happens
when the diviner derives his prediction from his art, but in the case of
natural divination, it occurs when the dreamer or seer forms predictive
content in his or her own dream or prophetic utterances. Artificial inter-
pretation is necessary for us to understand obscure content in divinatory
dreams or oracles. But under the new definition it is not part of divination
proper. It is the interpretation of some content already generated by a
completed divinatory event, the “foreseeing.” It is divination defined in
this new way that Quintus must show to be real. In the next two sections,
I shall examine Quintus’ arguments for the reality of artificial (section .)
and natural (section .) divination.
. Quintus’ Case from the Outcomes for Artificial Divination

In this section we shall see how Quintus argues for artificial divination,
while in the next I shall turn to the natural kind. When Quintus defends
his divinatory arts, he faces Cotta’s question – where did divination spring
from? So Quintus not only subjects Marcus to a deluge of examples of
predictions and outcomes. (Div. .–, –, –) He also gives
an idealized history of the divinatory arts. This is not a “theory” in the sense
of higher-order causal theorems that explain the if-then first-order the-
orems of the arts. Instead, it is a plausible story of how the predictive
theorems came to be adopted. It is in the context of this ideal history that
we will understand the role of Quintus’ apparently extra kind of divin-
ation, from coniectura, as a necessary part of the practice of artificial
divination.
Now it seems to me that we can understand Quintus’ strategy here
more clearly if we take a hint from many of our sources and draw the
following analogy with ancient debates about the theory of medicine.
Contemporary with Chrysippus in the second half of the third century
BC there began a long-running theoretical schism in medicine between the
Empiricist and the Rationalist doctors. Broadly speaking, the Empiricists
criticized medical theorists who claimed to be able to reason from obser-
vations to the underlying physical causes of medical problems and to
propose treatments on that basis. The Empiricists claimed that the medical

Date from Walzer and Frede () xx–xxii. For the comparison between Quintus’ artificial
divination and Empiricist medicine, see Hankinson ().
. Quintus’ Case from the Outcomes for Artificial Divination 
art could, and should, be derived purely from experience. Rationalists, in
turn, fought back on behalf of the use of reason and the construction
thereby of general causal theorems. Our primary source for the debate is
Galen. Chapter  of his On the sects for beginners (pp. – Helmreich) is an
introduction to the Empiricist view of medicine, and is, in effect, an
idealized history of an Empirical art. Let us examine this parallel, since it
will help to explain Quintus’ idealized history of divinatory arts (see
pp. – below).
According to Galen, in Empiricist medicine, there are three kinds of
“experience” in view: the “incidental kind,” the “extempore” kind, and the
“imitative” kind. (Sects p.  ll. –) Under the “incidental kind” fall
“natural” and “chance” experience, where an affection that harms or heals
the body comes about by obscure but natural causes (like an unexplained
nosebleed) or by chance (like an accidental cut), and is then observed.
(Sects .–) In neither case is the affection voluntary. Under the
extempore kind fall instances where something that harms or helps the
body is tried out for some less than scientific reason, for example, thanks to
a dream. (Sects .–) The imitative kind is the key to Empiricism. (Sects
.–) Under it, affections observed in experience of the other two kinds
are imitated in an effort to produce like effects. If such affections imitated
in this way prove helpful for the most part over many imitations, then a
theorem is added to the medical art. (Sects .–) It is easy to imagine
simple theorems of this sort – “when patients with a headache took an
aspirin, for the most part they reported less pain” – although most
theorems of a fully qualified doctor will doubtless be more complex. The
art itself simply consists of an “aggregation” or “heap” (ἄθροισμα) of such
theorems. (Sects .–) Rationalists, by contrast, would presumably
require that such theorems connecting treatment and outcome be properly
linked, by a rational structure, to higher-order theorems, that explained
why the treatments would work. The simple “heap” structure of the
Empiricists’ art therefore reflected their distinctive methods.
For our purposes, two further details of Empiricism are important. First,
Empiricists relied not only on direct personal experience, but also on the
recorded experiences of other doctors. The recording and passing on of
experiences was called ἱστορία, “history.” (Sects .–) History was

Galen On the sects for beginners Chapters I–IV with Frede’s introduction to Walzer and Frede
() xx–xxxi.

Here and in what follows I rely on the translation in Walzer and Frede ().

Translation from Walzer and Frede ().
 Quintus’ Stoic Case for Divination
presumably the source of most of an individual doctor’s theorems,
although Galen does not say this. Second, where an Empiricist doctor
faces a situation for which he has no experience or history to draw upon, he
can use the method of “transition to the similar” in which he evolves a new
theorem from experiences similar in many respects to his current case, but
not every respect. (Sects .–.) If a treatment devised in this way
should succeed, the respective new candidate theorem will of course be
subject to the full testing process, like any candidate theorem stumbled
upon by chance.
This conception of a medical art and its idealized history drew criticism
from the Rationalists. This was not, as Galen points out, because it led
Empiricist doctors to treat patients any differently than did their rivals.
(Sects .–) The dispute was over whether an art based on experience,
imitation, and transition to the similar could successfully generate all the
right theorems and the right treatments, or in the strongest form of
Rationalist objection, over whether such an “art” was an art at all. (Sects
.) To be properly comprehensive, perhaps an art should be able deduct-
ively to generate new theorems for unprecedented situations, as an art can
if equipped with a theoretical account of the causes of its subject matter,
for example about the biochemistry of headaches and aspirin.
The structural analogy of this medical debate, as reported by Galen,
with the debate over divination should be clear, as should some differences
between the two. A key difference is that the opponents of divination did
not, as Rationalists did for medicine, posit their own explanatory theory
for divination. They believed there is none. But a key similarity is that
these same opponents demand an explanatory theory (λόγος, ratio) from
those who claim that divination is an art, just as Rationalists insist that
medicine must have higher-order explanatory theory. Thus Rationalists
and the opponents of divination suppose the same constraints on what
counts as an art. This is what Cotta’s demand for a ratio or “theory” of
divination seems to amount to. We supposed that Carneades similarly
demanded to know why divination works (see p. ).
Quintus’ Stoic case for artificial divination seems to be designed with
such a debate in mind. By declining to define divination itself as an art, he
evades many of Carneades’ and Cotta’s strictures, which were targeted at
Chrysippus’ definition. Quintus claims that divination is the foreseeing
and prediction of states-of-affairs thought to be by chance. Whether or not
foresight or prediction really happens seems to be amenable to merely
empirical confirmation. Thus Quintus insists that he owes no theory,
despite Cotta’s demand. Whether his so-called divinatory “arts” are really
. Quintus’ Case from the Outcomes for Artificial Divination 
arts, by some substantial criterion of what counts as an art, does not matter
to him, nor does he need to equip them with higher-order theorems about
causes. All that matters is that their use seems to work.
Now Carneades might still try to press his more elaborate “subject-
matter” objection. Granted that Carneades can no longer demand a
clear subject matter or goal from divination on the grounds that it is an
art, he might ask now whether Quintus’ definition leaves us with any
clear concept of divination at all. Why do we need to propose a special
kind of prediction and foresight that comes from the gods? Sure
enough, we shall see Marcus press this objection in my next chapter
(pp. –).
The second part of Quintus definition, “of states-of-affairs thought to
be by chance,” helps to answer this evolution of the subject-matter objec-
tion. The Stoics held an epistemic view of chance: chance is “a cause
unclear to human reasoning.” (See SVF .–.) Furthermore, they
held that it is a cause unclear not just to some humans but rather simply to
humans, and that divination is a way of getting to know things that seem
to be unclear in this way. So the point of Quintus’ definition is that
divination gives us information about things that no human could predict
by conventional reasoning from causes about which he or she knows. God
knows all current causes and can predict their effects, so he is in a position
to tell us about them. Again, since Quintus’ definition (unlike Chrysip-
pus’) restricts divination to foreseeing (praesensio) and prediction (praedic-
tio), the information we get from divination in particular is always about
the future. The function of the various sorts of divination, then, is to be a
means by which a god tells us about some aspect of the future that we
think we cannot predict.
What, then, is Quintus’ ideal history of the divinatory arts? Broadly
speaking, Quintus says that the established divinatory arts of the Ciceros’
time were formed by recording the careful observation of signs and
outcomes over very long periods of time, just as Empiricist medics would
imitate an experience many times before a theorem was derived from it.
The process of making of these records he calls “observation,” observatio.
He says that observation works because,

See Alexander of Aphrodisias, Mantissa ..

Although Quintus defines divination as prediction of the future, many of his examples include
divination that is not precisely this – warnings (to Simonides in his dream, Div. .) or
information about the present or the past (Sophocles’ dreams about the identity of a notorious
thief, Div. .). I suggest that these examples come down from Chrysippus, for whom divination
was not necessarily prediction of the future.
 Quintus’ Stoic Case for Divination
observata sunt haec tempore immenso et in significatione eventis animadversa
et notata. nihil enim est autem quod non longinquitas temporum excipiente
memoria prodendis monumentis efficere atque assequi possit.
These things [= the crow calling on the left, the raven on the right] have
been observed over a vast length of time, and noticed because of their
outcomes, and then marked for what they signify. For there is nothing that
sheer length of time cannot bring about and carry through, when memory
collects and records are handed down. (Div. .)
This immediately precedes the Quintus’ own analogy with the medical art
that is probably an allusion to Empiricist medicine (see p. ). The
parallels are obvious. Signs and their correlated events are first of all
noticed (animadversa) and then marked down (notata). Over a sufficiently
great length of time, memory will collect these experiences. It will take
many lifetimes to build up these arts, so in addition to the use of memory
records must be kept. Quintus thinks that this process is capable of
producing any successful empirical art – no matter how complex, bizarre,
or seemingly unlikely. Astrology certainly depended, or claimed to depend,
on such long and painstakingly kept records. Cicero in his preface himself
remarks on the “daily observation” of the Egyptians who have pursued the
same art “by duration of time over almost uncountable ages.” (longinqui-
tate temporum innumerabilibus paene saeculis, Div. .) Quintus goes one
better and puts a figure on the amount of time the Babylonians have been
involved in astrology. He says that we should not accuse of falsehood those
“who have , years encompassed in their records, as they themselves
say.” With augury and haruspicy, there seems no reason to doubt that

All but one of the MSS have in significatione eventus, a crux for many editors, although Pease (–
) prints it untouched and hazards “in regard to the significance of their outcome.” Many
solutions are proposed, but Giomini prints eventis, the reading of H, a later MS. Giomini and Pease
both point to Div. ., eventis animadversa ac notata sunt, as a parallel in support of H. Giomini
notes with “fortasse rectius” the editions of Sturm () and Lambinus () who omit in. This is
indeed attractive, yielding a simple “noticed and recorded by their signification and their
outcomes.” At any rate, the approximate meaning is clear.

This is an overestimate, although not as great as that found in some other Greco-Roman treatments
of Babylon (see Wardle  ad loc.). But as a matter of fact the Babylonians did keep records
correlating celestial signs and earthly outcomes, with increasing detail and sophistication, for well
over a millennium down to the Hellenistic period. See the relevant part of the Enuma Anu Enlil for
the long older collection of astrological theorems, including material of many dates back into the
second millenium BC. (Reiner and Pingree –) The Enuma contains similar collections for
other forms of divination. For the later “astronomical diaries” see Sachs and Hunger (–).
Ptolemy claimed to have access to some of these records, going back to the reign of Nabonassar in
the eighth century BC (Almagest ..–). This is plausible, since Ptolemy cites records of lunar
eclipses at Babylon with some accuracy (Almagest IV.). Jacobs () makes an argument that
Cicero, directly or indirectly, was aware of some ancient near eastern divinatory signs, principally
. Quintus’ Case from the Outcomes for Artificial Divination 
the material in their books would have reflected a long tradition. So
Quintus concludes that the length of time over which the divinatory arts
have been tried and tested helps to answer Cotta’s challenge. Over very
long periods, it is plausible that somebody, for whatever reason, silly or
profound, with or without divine prompting, would start to examine the
predictive capacities of birds or offal, and that useful theorems could result
from many observations of such data.
Where, then, does coniectura, “conjecture,” come in? I suggest that for
Quintus, artificial diviners conjecture when they make a divinatory predic-
tion from what they guess is a sign, but a sign for which there is no existing
theorem in their art. They might do this, for example, if some new
phenomenon that seems to be within their purview arises: for the haruspex,
a prodigy or a deformity of the liver never before seen, an unprecedented
bird for the augur. If the resultant prediction is successful, then new
instances of the same type of sign can be examined over time, to see if
they have the same outcome. A special case of this would be at the
foundation of a new art, that is to say, when there is as yet no theorem
of the future art. In that instance an insightful soul guesses, for whatever
reason, that some event is a sign, and conjectures a prediction from it.
When this prediction is successful, he starts to look out for more instances
correlating the same sort of sign with the same sort of prediction, and a
candidate theorem is set on its way to confirmation by long observation. In
this way coniectura is a part of the practice of artificial diviners: not
divination by established theorems, but the process whereby artificial
diviners predict without a theorem and thus suggest a candidate theorem.
There is evidence that Quintus indeed considers that coniectura is a part
of the practice of artifical divination in the way I have suggested. When he
has finished his first catalog of outcomes for artificial divination, Quintus
says again that he agrees with those who say there are two types of
divination, “For there is art in people who pursue new matters by conjecture,
that which correlated the birth of lion to a woman with the fall and capture of the king (Jacobs
pp. –, cf. Div. .). The argument is plausible, but I am not yet convinced. While it is
plausible that the fall of the king became in Cicero the fall of a republic (res publica), it is also
plausible that for Cicero the fall of a republic would mean the rise of a king, and thus that Cicero has
in mind an outcome opposed to the near eastern example Jacobs cites. If Cicero derived this example
from the birth of Pericles, then indeed a dream of the birth of a lion would seem to presage the birth
of a politician of dangerous ability. In addition, if we suppose that Cicero’s example ultimately
derives from Jacobs’ near eastern text, it is sufficiently likely that this derivation would be indirect,
that it seems hasty to speak of Cicero’s “knowledge” of older near-eastern divinatory tradition
(p. ).
 Quintus’ Stoic Case for Divination
but have learnt old ones by observation.” (Div. .) In Quintus’ latter
situation, the diviner has learned theorems produced by past observation,
and can apply these. But when he comes to new matters, the diviner must
turn to conjecture. This strongly suggests that conjecture is part of the
same project as the established divinatory arts. Quintus later returns to this
theme:
quae vero aut coniectura explicantur aut eventis animadversa ac notata sunt, ea
genera divinandi, ut supra dixi, non naturalia, sed artificiosa dicuntur.
But, as I said above, those kinds of divining which are worked out by
conjecture or are noticed and marked down from the outcomes are not
called natural, but artificial. (Div. .)
Again, some part of the activity of these arts functions by “records and
doctrine” (monumentis et disciplina) as found in the books of the haruspices
and the augurs. (Div. .) But other aspects are worked out “suddenly
and ex tempore.” There follows from Div. . to . a litany of examples
of divination by sudden conjecture, predictions as ever correlated with
outcomes.
Some of these examples of coniectura are primordial, like Calchas who
predicted the length of the war at Troy from seeing a number of sparrows.
(Div. ., cf. Iliad .–) Some are historical instances of divination
by haruspices or augurs. It is not always easy to see why the diviner made
the conjecture – why Calchas, for example, connected sparrows and years
of war as he did. But as I understand Quintus’ story, this does not matter
too much. A diviner’s reasoning can be loose, or hard to understand, since
before any resulting candidate theorem is taken up with confidence, it will
be subject to much more testing.
A comprehensible example of conjecture from Quintus’ list is the
following. Just before the battle of Leuctra, some soon-to-be victorious
Boetians at distant Lebadia heard their cocks crowing insistently, which
their “augurs” interpreted as a sign of victory, since cocks crow when
victorious. (Div. .) Here we have some insight into a conjectural
process. The Boetian bird-diviners made their prediction by association
with a well-known behavioral fact about cocks. But it was not a prediction
from any established theorem of their art as such. Thus it constituted a
coniectura. Perhaps, after long enough study of the correlation between
assiduous crowing and victory, the conjecture would be confirmed as a
theorem.

est enim ars in iis, qui novas res coniectura persequuntur, veteres observatione didicerunt.
. Quintus’ Case from the Outcomes for Artificial Divination 
Now Quintus’ idea seems to be that the Boetian augurs’ conjecture was
right by something more reliable than chance, that conjecture is something
that expert or naturally talented diviners do. How can he square this with
the claim that he is not offering any “theories” underlying the predictive
theorems of artificial divination? His reply here might be similar to what he
said about the artificial interpretation of dreams. The heuristics which
seasoned augurs or haruspices might develop to guess at new theorems do
not have to reflect any (true) causal theory about why any divinatory
theorem is true. They are simply tentative interpretations of how we might
extrapolate the arbitrary relationships between already known signs and
outcomes. Since the gods want us to develop divinatory arts, it is plausible
that they might sometimes plan their signs and outcomes so as to provoke
the right conjectural prediction or to sustain a new candidate theorem. For
example, the gods behind the events of Lebadia might have chosen the
crowing of cocks as a sign because they knew that the people there
associated the crowing of cocks with victory. If the Lebadians then pro-
posed an associated candidate theorem (“if the cocks crow in such-and-
such a way, there will be victory”) the gods could choose to observe that
rule, so that it will eventually be confirmed as a theorem in Lebadian
augury.
The word coniectura features prominently in another part of Cicero’s
writing, on rhetorical theory. There it means argument over the facts in, for
example, a court case: did the accused do it or did he not? For example,
suppose that the evidence in court is that reliable witnesses report the
following: “Sextus was found dying from a dagger wound. Titus was
standing over the body, holding a bloody dagger.” The prosecution argues
that the jury should infer that Titus stabbed Sextus to death. The defence
argues that the jury should infer that Titus found Sextus stabbed and drew
the dagger from the wound in an effort to save Titus. Both conclusions are
consistent with the evidence. Neither is entailed by it in the way that a
conclusion in philosophy or mathematics might be entailed by the

A special category of coniectura emerges in Div. .–: fiunt certae divinationum coniecturae a
peritis, “sure conjectures in divining come about made by experts.” It does not seem likely that certae
here is to mean “certain,” rather “very reliable,” nor is it clear who the relevant peritis are since in
two of three examples they go unremarked and in the third they are simply haruspices. But the
predictions are all notably good: that Midas would be rich, that Plato would have sweetness of
speech, and that Roscius would be famous. So these predictions seem intended as examples of
spectacularly successful predictions from signs where no art can help, with the suggestion that such
spectacular success is achieved with the aid of plenty of practice in divining. These suggestions serve
as extra evidence that coniectura can sometimes be successful other than by mere chance.

See On invention ., .–; On the parts of rhetoric, –.
 Quintus’ Stoic Case for Divination
premises of a deduction. Thus a “conjectural” situation like this requires
the orator to draw on what the jury finds plausible in order to lead them to
one inference or the other. That sense of coniectura in rhetorical theory is a
parallel to the way Quintus uses coniectura in Div. As I interpret him,
Quintus thinks that artificial diviners usually take a sign, then deduce a
prediction by the combination of the sign with a relevant general theorem.
But in coniectura there is no relevant theorem. Rather, diviners take what
seems to be a sign, and try to find a plausible inference to a prediction. In
his reply to Quintus, Marcus will make a similar comparison between
divinatory and forensic coniectura. He says:
ut enim in causis iudicialibus alia coniectura est accusatoris, alia defensoris et
tamen utriusque credibilis, sic in omnibus iis rebus, quae coniectura investigari
videntur, anceps reperitur oratio.
For just as in cases at trial the accuser has one coniectura, the defender
another, and both coniecturae are believable, so in all of those matters that
seem right to investigate by coniectura we find two-sided rhetoric. (Div.
.)
We should conclude that coniectura plays a role in divination straddling
the respective roles in Empiricist medicine of extempore experience and
“transition to the similar.” It allows new arts to get underway and new
candidate theorems to be found for established arts. In this way it also
allows the practitioners of existing arts to deal with unprecedented situ-
ations. The examples Quintus gives show that coniectura can sometimes
produce true predictions, which is all that is necessary for his argument. Of
course, conjecture is not infallible, and perhaps not even reliable, and this
leaves arts whose theorems stem from conjectures fallible in principle, even
with the control of very long periods of observation. Quintus admits this.
In Div. .–, he considers the obvious objection that divination often
gets its predictions wrong. His first reply is to say that medicine, naviga-
tion, and politics get things wrong too, but are still considered arts. He
goes on:
similis est haruspicum responsio omnisque opinabilis divinatio; coniectura enim
nititur, ultra quam progredi non potest. ea fallit fortasse non numquam, sed
tamen ad veritatem saepe derigit; est enim ab omni aeternitate repetita, in qua
cum paene innumerabiliter res eodem modo evenirent isdem signis antegressis,
ars est effecta eadem saepe animadvertendo ac notando.
The advice of haruspices and all [merely] opinable divination is similar [to
arts like medicine and navigation], for it is founded on conjecture, beyond
which it cannot progress. Perhaps it plays us false sometimes, but on the
. Quintus’ Case from the Outcomes for Artificial Divination 
other hand it often directs us to the truth. For it has been derived from all
eternity, and it was made into an art, since over that eternity the same
events have come out preceded by the same signs an almost countless
number of times. (Div. .–)
This, I think, is Quintus’ account of the idealized history of artificial
divination in a nutshell (it immediately precedes his main catalog of
examples of successful artificial divination). It puts together the elements
we have discussed above, that divinatory arts are begun by conjecture and
confirmed by long observation. It makes clear that this process gives rise to
fallible, but reliable, arts. Here we see another difference between Quintus
and Chrysippus. Chrysippus seems to have thought that all the predictions
of diviners qua diviners are true (see pp. –), but Quintus thinks
that since artificial divination is fallible, even if practiced perfectly it will
not always produce true predictions.
You might object that this approach to artificial divination cannot
be Stoic. For, you might say, the Stoic sage is supposed to be
infallible, in that she is supposed not to form, or even to risk forming,
false opinions. But if she uses a fallible art, sometimes she will form,
or risk forming, false opinions. Thus, on this view, sages would not be
diviners. Indeed, according to the Stoics, nobody would be justified in
divining, because of the risk of false opinions. Thus (you would
conclude) the Stoics, who endorse divination both in principle and
in practice, would not take the position I have described. The answer
to this objection is to deny that a sage who knowingly uses a fallible
art must risk false opinions. For a sage may adjust the beliefs she
forms to fit her circumstances. To use artificial divination, she need
not believe that the theorems of the divinatory art are true. Rather, she
need believe only that it is reasonable to proceed as though they were
true, but that they might be false. Similarly, she will not believe
simply that a particular prediction made by use of the art is true.
Rather, she will believe that it is reasonable to suppose that the
prediction is true, but that it might not be.

If Quintus, although a Peripatetic at heart, assumes here the Stoic doctrine of the conflagration,
through which humans could not pass on their arts, then “all eternity” seems to be hyperbole.
Quintus seems to concede this with his paene innumerabiliter, “almost innumerable.” Alternatively,
perhaps he reflects that divinatory arts must be known to the Stoic god, in which case they do
survive eternally, and do reflect conjunctions of sign and outcome that this god has eternally
sustained.
 Quintus’ Stoic Case for Divination
. Quintus’ Case from the Outcomes for Natural Divination

Finally, we must see how Quintus hopes to show from outcomes alone
that natural divination is real. Here, of course, his strategy is not to argue
that there is any art that first arrives at predictive content. Instead, he tries
to show that the predictive content in divinatory dreams and in oracles
itself can be shown to be true by the outcomes. He gives his many
examples of such conjunctions of prediction and outcome at Div. .–.
Here Quintus faces the same obvious challenge that he faced in arguing
for artificial divination, that predictions drawn from dreams sometimes
seem to be wrong. But now he is without the same defensive option,
namely, he cannot argue that divination from dreams has proved to be a
fallible, but valuable, art. Quintus squares up to this challenge (at multa
falsa!, “But many are false!”) in Div. .. His immediate answer is,
“Rather, perhaps, obscure to us!” (immo obscura fortasse nobis.) For if the
predictive content in dreams and oracles is obscure, then to understand
this content will take interpretation. This interpretation can be fallible.
Thus in cases where a predictive but obscure dream was interpreted, but
the interpreter’s version of the prediction turned out false, the error can
always be attributed to the interpreter and not to the content of the dream.
Of course, there could also be cases where a predictive dream appears so
clear as not to need interpretation, and where its apparently clear predic-
tion turns out false. But even this case could be given the same defence.
For perhaps the dream’s predictive content seemed clear, but was in fact
obscure, so that it predicted obscurely something other than what it
seemed to predict clearly.
Nevertheless, it seems that the very content of a predictive dream could
itself be false. Quintus concedes this point, and therefore goes on to
consider such cases. Here he has a theory to suggest, with elements
borrowed from numerous philosophical and literary sources, which gives
reason to think that although dreams are not always predictive, they are
potentially reliable on occasion. What he argues is that humans have an
innate power to have dreams which are “veridical,” (veracia) (Div. .), to
“divine,” (divinare) (Div. .), to “presage,” (praesagire) (Div. .), and
the like. This power will yield true content in dreams and frenzies when the
soul is in the right state, for example when one dreams after living temper-
ately (Div. .–), or is close to death (Div. .–), or is in frenzy
(Div. .). For this reason dreams and frenzied predictions can be
reliable. But we can also see why they will often be unreliable. For we
are rarely in these states of exalted temperance or proximity to death.
. Quintus’ Case from the Outcomes for Natural Divination 
Now strictly speaking, this appeal to the states of soul appropriate to
natural divination may exceed Quintus’ rubric of arguing for divination
from outcomes and without causal explanation. But if so, it is not a serious
infraction. The theory is extremely abstract, and depends on no particular
account of body or soul. All Quintus aims to sketch is a world in which we
have some power that allows us to get at these truths, that sometimes is able
to work, and sometimes is not. He says nothing about what such a power
might be, or why proximity to death (and so on) might allow it to be
active. But to do this is hardly more than to make explicit some of the
assumptions behind the notion of “natural divination” itself. We can see
how abstract is Quintus’ theorizing here precisely from the eclectic range
of authorities he appeals to in its support: Plato (Div. .–), Posidonius
(Div. .), Plautus, wise old women (Div. .), tragic drama (Div.
.). We may compare the more usual role of an old woman in the
dialogues, as the paradigmatic victim of superstition (e.g. DND ., .,
Div. ., .). In short, Quintus’ claims here about the circumstances in
which natural divination is reliable are based on common sense, rather
than on technical theory.
Thus, more precisely stated, what Quintus hopes to show about natural
divination using the outcomes is this: we are sometimes able to have
reliably true predictive dreams or waking delusions, even if the predictive
content of these dreams or delusions is sometimes obscure. At Div.
.–, as he finishes his long catalog of examples of successful natural
divination, Quintus sums up this approach in what Schofield called “one
of the very rare philosophical arguments” in the speech ( p. ). For
this purpose, he follows Cratippus of Pergamon, a Peripatetic who gave
credence only to what Quintus calls natural divination. Cratippus com-
pares the ability to have veridical dreams and frenzies to the sensory power
of the eyes. Although eyes do not always function, if someone has seen real
things (vera) at least once, it follows that his eyes can see.
item igitur, si sine divinatione non potest officium et munus divinationis
exstare, potest autem quis, cum divinationem habeat, errare aliquando nec vera
cernere, satis est ad confirmandum divinationem semel aliquid esse ita divina-
tum ut nihil fortuito cecidisse videatur; sunt autem eius generis innumerabilia;
esse igitur divinationem confitendum est.
In the same way, then, if it is not possible for the purpose and function of
divination to come about without the [power of] divination, but if it is
possible that someone who has the [power of] divination to err sometimes
and not to perceive real things, then it is enough to support the [power of]
divination that something should once be divined in such a way that it
 Quintus’ Stoic Case for Divination
appears to have fallen out not at all by chance. But there are countless
examples of this sort. Therefore, it must be admitted that there is divin-
ation. (Div. .)
The argument makes two assumptions. The first is that the activity of
natural divination could not happen unless the human soul has some
power or means by which it is able to divine, just as seeing could not
happen without eyes’ ability to see. The second is that sometimes someone
whose soul has this means to divine can go wrong, just as someone with
eyes can sometimes fail to see. If there has been one incontrovertible
instance of natural divination, then there must be the power of divination
in the soul, since the function of divination could not be fulfilled without
the power, and since it is possible for a power to be used rarely or just once.
Meanwhile, any number of failures to predict, no matter how many, can
be explained as the person with the power to divine going wrong. So, just
one incontrovertible instance of divination would prove that the power to
divine exists, even in the face of very many false predictions. But Quintus
hopes to have shown us in the preceding sections of his speech that there
are many incontrovertible instances of natural divination.
Now this argument has a strongly Peripatetic flavor. The Stoics do not
seem to have posited some specific power or means of natural divination in
the soul, analogous to the eyes and their ability to see, as Cratippus does
here. But in this Stoic context, Quintus does not need to put too much
weight on whatever specific aspect of the soul Cratippus intended when he
compared our natural divinatory capabilities to the eyes, since, as I have
said, Quintus means only to argue that in the most general sense, we can
have reliably predictive dreams. On this basis, Quintus claims he can
appeal simply to the outcomes in order to prove the reality of natural
divination.
. Conclusion
I hope that this chapter has helped to clarify some of what seems so messy
about Quintus’ speech, or more specifically his main argument down to

There are numerous questions about what precisely Quintus thinks Cratippus means by the sensus
oculorum, or by what I have loosely called the “power” of divination. For in the hands of one or
another philosopher, there are substantial differences between (for example) technical notions of
potentialities, powers, or faculties, or again between the eyes’ sensitivity to light and the soul’s sense
of sight. But I do not think Quintus needs to specify his, or Cratippus’, point very precisely here,
since all Quintus requires is that there is some fact about the soul according to which it can
sometimes reliably produce predictive divinatory content.
. Conclusion 
Div. .. I have suggested that many of these problems are solved when
we see that Div. is written, at least in part, for learned readers, and features
characters who carry on the project of DND rather than start a discussion
of divination from scratch. Quintus replies to the brief exchanges on
divination between Balbus and Cotta in DND, and he does so by explain-
ing a new Stoic defense of divination that can evade Cotta’s charges against
Chrysippus’ position on the topic. In Chapter , I shall take up the next
stage in this back and forth, Marcus’ rebuttal to Quintus, in Div. Book .
In Chapter  section ., we saw Balbus’ world as a place of beauty,
redolent of the care lavished on us by a god who made everything else in
the cosmos for us rational creatures. Quintus’ arguments pursue his formal
goal of shoring up Balbus’ conclusion. If Quintus were to convince us that
divination were real, we would eo ipso concede that there are gods who care
for us. So in effect, Quintus cites empirical evidence for Balbus’ emphatic
“yes” to the Central Question. But we should also notice how Quintus’
speech enriches and makes more urgent Balbus’ portrait of humanity’s life
with the gods. It is not just that god fated a world for us and keeps it and
its parts in motion. He also communicates with us, both with societies and
with individuals. First, he says things to us in the most intimate way,
welling up from inside our souls at times, in dreams and delusions, when
our rational control of our imaginations is weakened. Second, he cooper-
ates with institutions, cities, and diviners, who can record the conjectures
and observations that make divinatory arts, through which the events of
the natural world can be made to yield messages, so that religion is not a
one way communication, from us to the gods, but rather a conversation.
God’s care for us is all the more lavishly apparent. This perspective is not
absent from Quintus’ speech. He quotes (and Cicero proudly translates
with precision) a long, complex Stoic syllogism proving the reality of
divination, used not only by Chrysippus, but also by later Stoic authors
on divination, Diogenes of Babylon and Antipater of Tarsus. As ever,
Quintus is at pains to point out that such arguments are not the basis of his
case. But even if the argument might not be convincing to non-Stoics, the
many premises exhibit how deeply divination was implicated in the Stoics’
beliefs about their caring gods: the gods love us, they are our benefactors
and friends, they know the future, it makes a difference to us to know the
future, and they have the means to tell us about it. (Div. .– cf.
.–) By paying attention to their divinatory signs and messages, we
pay attention to the loving divine friends who surround us.
 
Marcus’ Arguments against Divination
In Book  of DND, Cotta the pontifex undertook to face down Stoic

theology, on behalf of the skeptical Academy. In Book  of Div., Marcus
the augur takes on the Stoic theory of divination in the same cause. The
role of Marcus’ arguments in the overall project of DND and Div. is to offer
reasons why a Roman, even an augur like Cicero himself, need not think
that the gods hedge us about with messages and divinatory signs. According
to these arguments, Cicero need not worry that news from the gods
obtrudes in his auguries, or hounds him in his dreams. (Div. .)
Now Marcus’ speech presents some puzzles for my argument in this book.
For I have argued that Cicero’s dialogues are unified wholes (see p. ), and
also that Quintus’ speech is an idiosyncratic Stoic defense of divination (see
Chapter  section .). These two claims together entail a prediction about
Marcus’ arguments: he should respond to Quintus’ Stoic defense of divin-
ation in particular, not to Stoic views about divination in general. But
Marcus’ argument has seemed to some readers to fly wide of Quintus’ mark.
If it did, then I should have to give up at least one of the claims just
mentioned. In interpreting Marcus’ arguments in this chapter, I shall argue
that, for the most part, he sticks to the point and takes on Quintus’ speech in
particular, not divination in general. Quintus countered Cotta’s challenge to
the Chrysippan account, “where does this divination of yours come from?”
with his division into artificial and natural divination. Marcus, a good servant
of his own project as the author of DND, reasserts the challenge against, in
particular, this Stoic account of divination, which I hope to show in this
chapter. But in section ., I shall first discuss some of the charges of
unfairness to Quintus that scholars have laid against Marcus.
. Is Marcus’ Speech Suspiciously Unfair to Quintus?

The argument of Marcus’ speech has a clear and clearly sign-posted
structure. This draws admiration, especially as it follows Quintus’ messier

. Is Marcus’ Speech Suspiciously Unfair to Quintus? 
effort. “Nothing could be more straightforward than the structure of Book
II: any table of contents drawn up for Book I would be a fairly optimistic
and arbitrary construct.” (Schofield () ) The straightforward struc-
ture is as follows. Marcus first makes a preliminary “sally” (excursio, Div.
.), a sustained argument designed to tempt Stoics into a dilemma that
does for divination with a single blow. (Div. .–) Then he pursues
another, grand strategy (to “rout the wings” of Quintus’ position, cornua
commovere, Div. .), in which he runs through first the kinds of artificial
divination, and then those of natural divination, attacking the reality of
each in turn. (Div. .–)
But this clear structure masks some problems. As we shall see, the
dilemma of the preliminary “sally” is not finally compelling, least of all
to a Stoic. Marcus suggests some lack of confidence in it when he admits
that it was made “lightly armed.” (levis armaturae, Div. .) Thus much
of the heavy work of refutation must be done in the longer section that is
structured by type of divination. Marcus’ disciplined adherence to this
structure means that at no point does he set out what view (if any) he
advocates of Quintus’ position as a whole. But, as I have just pointed out, if
I am right in my general interpretation of DND and Div., then Marcus
ought to have in mind a strategy to dispose of Quintus’ Stoic defense of
divination in particular.
I will suggest later (sections .–. of this chapter) that Marcus does, in
fact, have a general approach for attacking Quintus. This is represented by
certain patterns of argument that recur in many of his respective criticisms
of the kinds of divination. But in this section, I first wish to make some
remarks about Marcus’ strategy, and in particular the following three
features of it. First, as Malcolm Schofield notes, Marcus’ approach is often
“rhetorical,” at least in the straightforward sense of employing the tech-
niques of forensic rhetoric. (Schofield () –) Marcus sometimes
calls attention to his use of rhetoric by using the language of the court-
room, and indeed Schofield provides some parallels between arguments in
Div.  and those of Cicero’s real-life defence speech Pro Roscio Amerino.
(Schofield () App. II, p. .) Second, Marcus routinely makes argu-
ments against the Stoics that at least appear to misinterpret to a significant
degree some salient part of the Stoic doctrine behind Quintus’ speech. As

testibus uti (Div. .), magnam iacturam causae (Div. .), omissis testibus (Div. .), quasi
quadam praevaricatione (Div. .).

Although I do not ultimately agree with his diagnosis that Marcus’ speech as a whole systematically
commits an ignoratio elenchi, Denyer () documents a number of possible straw men in the
 Marcus’ Arguments against Divination
I shall suggest, Marcus seems to argue against straw men, theories about
divination that his opponents, the Stoics, would not accept. Third, in
addition to his apparent misinterpretations, Marcus seems to depart from
the sort of strictly targeted arguments we might expect of him. That is, he
makes arguments founded on premises that Stoics will not welcome, and
that they may feel no pressure to accept. From these three features it would
be easy to form a negative opinion of Marcus’ speech, if what we expect (as
I do) is a philosophically rigorous reply to Quintus’ speech in particular.
Why, then, might Cicero have given Marcus such a speech? Three
possible answers to that question would threaten my project. One answer
would be that in fact Quintus has no systematic position in Book . On
that answer my reading of Quintus’ argument is ruled out of the text of
Book  by the text of Book . A second answer would be to say that the
mismatch is good reason to revisit the single-source hypothesis (see section
. of my introduction). One might favor this answer if one thought that a
mismatch suggests that Cicero used a source, or souces, for Book  which
argued against a different position than that held by the sources for Book
, and that he either did not spot the mismatch or made no sufficient effort
to correct it. If Cicero’s sources could dominate the parts of his text to
that extent then my thesis that he shaped the dialogues as literary unities
would look dubious. A third answer would be that Cicero was well aware
of the flaws he wrote into Marcus’ understanding of Book . He might
have written them so that (for example) some readers would see through
Marcus’ arguments and thus light on the stronger points of Quintus’ case.
Such a gambit is not out of the question for an author of Cicero’s
sophistication, but does not seem in keeping with his earnest use of
dialogue form to further his skeptical agenda, as I painted it in section
. of my introduction. But I shall now offer another interpretation of the
relationship between Marcus’ and Quintus’ speeches, one consistent with
my general position in this book.
First, on the issue of “rhetorical” tactics, we may note that in fact Cicero
would say that rhetoric is in theory the appropriate means by which to
reply to the sort of case Quintus has made on behalf of the Stoics. As
we saw in Chapter , Quintus carefully limited his case so that – he
speech. Hankinson ()  accepts Denyer’s conclusion, Timpanaro () – does not.
See also Repici ().

See e.g. Pease (–) on Div. . and the inconsistent definitions of divination between Books
 and : “This inconsistency is probably due to the use of different and unrelated sources for the two
books . . .”

Schofield () himself makes a similar point, pp. ,  n. , –.
. Is Marcus’ Speech Suspiciously Unfair to Quintus? 
claims – the burden of proof fell entirely on the empirical evidence of his
examples and to no extent on any physical explanation of divination.
Marcus certainly apprehends this aspect of Quintus’ defence. At a key
programmatic point in his speech – the outset of the main arguments
following his “sally” – he criticizes and thus implicitly acknowledges what
Quintus claimed was the only basis of his argument:
duxisti autem divinationem omnem a tribus rebus, a deo, a fato, a natura. sed
tamen cum explicare nihil posses, pugnasti commenticiorum exemplorum mir-
ifica copia. de quo primum hoc libet dicere: Hoc ego philosophi non esse
arbitror, testibus uti, qui aut casu veri aut malitia falsi fictique esse possunt;
argumentis et rationibus oportet, quare quidque ita sit, docere, non eventis, iis
praesertim, quibus mihi liceat non credere.
But you derived every sort of divination from three things, from god, from
fate, and from nature. [This is a reference to the physical theories Quintus
gave at Div. .–, cf. p.  above.] But since you were unable to
explain anything you went to battle using a wondrous host of confected
examples. About that I first want to say this: I do not think this appropriate
for a philosopher, to use pieces of evidence [lit. ‘witnesses’, testibus] that can
be true by chance, or can be false and concocted through ill will.
A philosopher should show why each thing is as it is by proofs and
arguments, not by outcomes, especially not by outcomes that I may dis-
trust. (Div. .)
Thus, on his own understanding of Quintus’ speech, Marcus’ goal in
refuting Quintus’ case for divination should be to show that the “host of
examples” is unable to establish what Quintus thinks it can. Let us see why
this makes the dispute a “rhetorical” one for Cicero.
Quintus asked that the question under examination be not why divin-
ation happens, but whether it happens. Marcus is well aware of this: “[you,
Quintus, said that] what happens, not why it happens, is what’s relevant.”
(quid fieret, non cur fieret, ad rem pertinere, Div. ..) To argue that
divination happens, Quintus advanced anecdotal examples of outcomes
that matched predictions. With a few exceptions, Marcus will not question
the truth of these anecdotes. In some cases he cannot question the
anecdote because he himself was its source. So what he will argue is that
Quintus was wrong to argue from the truth of the anecdotes to the reality
of divination. Thus, to be precise about what is at issue in Marcus’ speech,
he must dispute an inductive inference that was made from a fund of
evidence that the two sides mostly share. This evidence consists in particu-
lar examples, not general truths. As we saw in another context in my last
chapter (pp. –), from the perspective of Cicero’s rhetorical training
 Marcus’ Arguments against Divination
the question at issue is therefore a constitutio coniecturalis, that is to say, an
issue (constitutio) subject to “conjecture,” to arguments about the plausi-
bility of inductive inferences from agreed evidence to disputed facts.
The textbook advice on conjectural argument the young Cicero pre-
sented in On invention .– makes it clear that a great variety of tactics
was considered standard practice in the competition for plausibility at a
trial. In the case of actions done or not done, an orator will typically appeal
to a jury’s sense of character, passions, and normal behavior: would such a
person have done such a thing? By building a case on induction from
reputable anecdotes Quintus has invited a counter-argument analogous to
those forensic tactics. Marcus is not required by this sort of refutation to
dispute general theories or arguments (although he does, just as Quintus in
fact presents such theories and arguments alongside his formal case).
Instead, he is required to show Quintus (and us, the readers) that Quintus’
inferences from particular facts were less plausible, or at least no more
plausible, than some alternative inferences. To Cicero’s way of thinking,
this sort of dispute is rhetorical territory more than philosophical. Now, by
that I do not mean that the rhetorical arguments are disconnected from, or
ineffectual in, their philosophical context. The point is rather that faced
with the strategy of supporting a philosophically important thesis, that
“there is divination,” with particular pieces of empirical evidence, the
appropriate argumentative response for Cicero is to apply the tools of
rhetoric devised precisely for disputing inferences from that sort of evidence.
That is a philosophically legitimate response that, if successful, would yield
the philosophically important conclusion that there is no sufficient set of
empirical evidence for the reality of divination. Thus Schofield is right to
say that Cicero gives Marcus a forensic rhetorical bent in Div. because that
is what the subject matter demands: “a situation in which full-blown
rhetoric was exactly the right philosophical strategy—where philosophy
could with perfect propriety be rhetoric.” (Schofield () )
Second, I will now address the degree to which Marcus seems to
misinterpret, or does not argue against, Quintus’ specific position. In the
next two sections of this chapter (section . and .), we shall see cases
where Marcus proceeds on the basis of what we might call straw men,
claims that a Stoic supporter of Quintus’ speech would not accept, or
which seem to interpret Quintus’ arguments unsympathetically. For now,

There may be a pun to suggest this invitation in Marcus’ comment on Quintus’ “whether it happens,
not why it happens” tactic: sed te mirificam in latebram coniecisti, “you threw yourself [or,
‘conjectured yourself’] into a surprising refuge”, Div. ..
. Is Marcus’ Speech Suspiciously Unfair to Quintus? 
let us consider what we ought to conclude from these straw men about
Marcus’ general response to Quintus. Do they suggest that Marcus has
missed important features of Quintus’ speech, or of Stoicism more widely?
I think that such a conclusion would require another premise, that is, we
should have to suppose that in order to deliver a successful skeptical
response, Marcus ought to argue against Quintus’ position using only
claims Quintus has made, or that a Stoic supporter of the speech would
accept. Should we expect this?
One reason to expect it might be that Marcus is an Academic skeptic
and some historians have thought that the Academics tended to argue
using their opponent’s premises.
“[The Academics’] style of philosophising was ad hominem. Typically, they
would take hold of one of the doctrines of a dogmatic philosopher (the
Stoics were their usual target) and attempt to reduce it to absurdity. ‘If you
Stoics are right’, they would argue, ‘and such-and-such is the case, then we
cannot know the truth about so-and-so. You Stoics are committed by your
own principles to scepticism.’” (Annas and Barnes () )
“[O]ne possibility is that he [i.e. a sceptic] take his premises from a
dogmatic opponent. Then his arguments will have the following form:
supposing that something which the dogmatists assert is true, it can then
be shown that nothing can be known. . . . This is the typical strategy of the
Academics.” (Striker (c) )
That the Academics would argue in such a way might be important to
understanding an important question about the school, namely whether
they endorse their conclusions, for example the conclusion no wise person
should form opinions, or whether they merely aim to reduce their dog-
matic opponents to such an impasse. So, if these scholars are right about this
“typical strategy,” perhaps we should expect Marcus to pursue it. In which
case, his failure to do so properly would lead to the sort of conclusions
about Book  of Div. that I seek to avoid.
But that expectation would be mistaken. The importance of the ad
hominem strategy (as Annas and Barnes call it) to the Academics is
strongest in their famous argument addressed to the Stoics against catalep-
tic cognition and in favor of suspension of judgment. (Academica .–
cf. Sextus Empiricus, Adversus mathematicos .–) This argument is

Marcus’ Academic skepticism (for which see section . of my introduction) in Div. specifically is
clear from his remarks at . and . and can also be inferred from the continuity of his character
with the one “he” wrote for himself in DND. Cicero’s own Academic preferences are gestured at in
the preface, at Div. ..
 Marcus’ Arguments against Divination
meant to use premises all but one of which are welcome to the Stoics, and
another that is supposed to be evidently plausible (although the Stoics will
resist it). It is certainly an argument of central importance to the Academ-
ics, first because the Stoics were their major opponents, and second
because the argument, if successful, undermined all other Stoic claims to
epistemic success. But even if they thought they had succeeded in this
quick, epistemological way against Stoicism, the Academics still took it
upon themselves to argue against all dogmatic positions on non-
epistemological questions, in order to induce doubt on all the questions
that dogmatic thinkers claimed to have settled. This is the point of the
many speeches in Cicero’s dialogues. Of course, these refutations are
ultimately in the service of an epistemological outlook – skepticism –
but their immediate purpose is not to refute an explicitly anti-skeptical
claim (e.g., “knowledge is possible”) but rather to undermine some other
dogmatically held proposition (e.g. “the cosmos is rational”). A worthwhile
focus on the central, epistemological argument should not lead us overly to
limit our expectations of how the Academics will argue in other contexts.
Let us consider an example, from Marcus’ speech, of this sort of freestyle
approach to skeptical counter-argument. This is Marcus’ refutation of the
bare syllogistic argument for divination that Quintus attributed to
Chrysippus, Diogenes of Babylon, and Antipater of Tarsus (Div.
.– cf. .–, p.  above). Marcus’ approach is to call on a
principle that a compelling syllogism concludes something that was
doubted from premises that are not. (Div. .) This principle is prob-
ably Stoic, or welcome to the Stoics. But the ways that Marcus goes about
rendering each premise doubtful certainly are not welcome to the Stoics.
(Div. .–) He says that two premises are doubtful because
Epicurus, the poet Ennius, or the Peripatetic Dicaearchus asserted their
contraries, and that another premise is doubtful because “very learned
men” assert the contrary. He doubts still another premise because of the
implausibility of making gods look into every house; another, because the
gods might give us signs but no way to interpret those signs, just as they
gave the Romans themselves no way to interpret the signs used in Etruscan

Striker is clear about this. Her essays on skeptical strategies (c and d) explicitly focus on
epistemological strategies, and hence portray the Academics arguing against the Stoics. (Note that my
quotation of her above is her analysis of how the Academics show that “nothing can be known.”)

Cf. Sextus Empiricus, Outlines of Pyrrhonism .: “A proof, they say, is an argument which, by
way of agreed assumptions and in virtue of yielding a conclusion, reveals an unclear consequence.”
ἡ ἀπόδειξις λόγος δι’ ὁμολογουμένων λημμάτων κατὰ συναγωγὴν ἐπιφορὰν ἐκκαλύπτων ἄδηλον.
Translation from Annas and Barnes ().
. Is Marcus’ Speech Suspiciously Unfair to Quintus? 
haruspicy; and so he continues. Consequently, Marcus says, the argument
is not sound because its premises are doubted, in the sense that while Stoics
believe them, others do not believe them, or believe that they are false, or
that common sense finds them implausible. There is nothing illegitimate
about this procedure, even though none of Marcus’ reasons for doubting
the premises would not be appealing to a convinced Stoic who does believe
the premises. What he has shown is that the Stoic syllogism cannot
establish its conclusion to the satisfaction of somebody who is not already
a Stoic, because its premises are subject to counter-arguments that seem as
reasonable as the Stoic arguments in their favor. Similarly, we might say,
Marcus’ arguments against Quintus aim to question the plausibility of
Quintus’ inference from the outcomes that divination exists. There is no
reason to think that Marcus is obliged as an Academic to question that
plausibility in Stoic terms. He can use the more freestyle approach and
appeal to any source of plausibility that he can oppose to Quintus’ appeals
to plausibility. Thus when Marcus makes arguments from premises that a
Stoic would rubbish, perhaps rather than misunderstanding Quintus, or
introducing a straw man to gain an easy victory, Cicero wants us to see the
equal, or greater, plausibility of the alternative Marcus offers.
Is it not a problem that Marcus will fail to persuade any convinced
Stoics? It is a problem, so far as it goes. But it raises a question of audience:
is Marcus aiming to argue many convinced Stoics out of their position?
I think not, or at least not primarily. From Cicero’s point of view there are
two obvious “audiences” for Marcus’ speech. One audience is the readers,
most of whom were and are not Stoics (although of course some Roman
readers were, and some even might be today). What about Quintus? It
turns out that he is not a Stoic either, but rather a non-Stoic attracted to
Balbus’ position in DND and impressed by the possibility that the Stoics
have rich resources to support their claims about divination. (Div. .,
cf. p.  below.) As we saw in section . of my introduction, Cicero paid
close attention to how the choice of interlocutors in a dialogue would allow
him to modulate the drama. In DND he gave the Stoic speech to Balbus,
almost a cypher for “convinced Stoic.” In Div. he made a non-Stoic defend
the Stoic view, and his refutation perhaps proceeds accordingly.
In the next sections, we shall examine in detail Marcus’ arguments
against Quintus. I shall follow the highest-level structure of his speech:

In fact, I expect the Stoics were well aware of this fact when it comes to the syllogism in question.
Quintus puts no weight on it in his argument that there is divination. The syllogism probably gives
the structure of the Stoics’ own reasons for believing in divination, not a proof offered to outsiders.
 Marcus’ Arguments against Divination
I will discuss first the excursio, the “sally” by which he hopes to knock out
divination at a single blow (section .), and then his main set of argu-
ments that are directed at each kind of divination in turn (sections
.–.). But in the latter case I will depart from Marcus’ kind-by-kind
structure, and instead present two patterns of argument I think we can
detect across the arguments against the various kinds of divination.
. The “Sally”: Div. .–

The “sally” moves in three stages. First, Marcus attributes a question to
Carneades that tries to show that divination has no subject matter of its
own. (Div. .–) Second, a Stoic reply is worked out, whereby divin-
ation is said to be of future chance events. (Div. .–). Third, Marcus
confronts divination of future chance events with a dilemma, that claims to
conclude that there can be no divination of the sort Quintus described.
(Div. .–) The three stages of the sally seem to me to form one long
argument. The first two stages show why a Stoic must take a certain
position on divination, which is then shown to be vulnerable to the
dilemma of the third stage. Now the three-state argument of the sally is
of particular interest to me because, although I think the final dilemma is
not very compelling, it responds precisely to Quintus’ speech as I have
interpreted it in Chapter .
In Chapter  sections . and ., I proposed that Chrysippus defined
divination as the “science of observing and explaining signs portended by
gods to men.” (Div. ., cf. Sextus Empiricus, Adversus mathematicos
.) I pointed out that Quintus’ definition – “the prediction and foresee-
ing of future events thought to be by chance” – is different. In the sally
Marcus represents Quintus’ definition as though it had been provoked by
an attack on divination that he attributes to Carneades. For following his
account of Carneades’ subject matter objection to divination Marcus says,
“But I noticed, Quintus . . . that you defined divination as ‘the prediction
and foreseeing of those states of affairs that are by chance.’” (Div. . cf.
p.  above.) Thus Marcus’ reaction, and the structure of the sally in
general, lends more support to the history I proposed in my last chapter:
Carneades (and others) criticized Chrysippus’ version of divination, and the
new Stoic definition that Quintus adopts was worked out at least in part in

sed animadverti, Quinte, te . . . [sc. divinationem] ita definire: divinationem esse earum rerum
praedictionem et praesensionem, quae essent fortuitae.
. The “Sally”: Div. .– 
response to Carneades. Now, Cicero’s text does not tell this story expli-
citly. Marcus talks as though he is attacking one Stoic position, not two
successively. He does not present the “chance” definition of the second
stage as a development that responded to criticisms of the first. Further,
Marcus tells us where Carneades’ arguments begin, but not where they end.
So it is possible that the whole “sally” is owed to Carneades. Nevertheless,
I think that the first two parts of Cicero’s “sally” represent the historical
exchange between Chrysippus, Carneades, and Chrysippus’ successors, for
which I argued in my last chatper.
Carneades’ question about the subject matter of divination, which
Marcus seems to think helped to provoke Quintus’ definition, and with
which Marcus opens the “sally,” asks quarum rerum divinatio esset, “what
subjects is divination of?” (Div. .) Marcus then lists some subjects that
we know about by various means: those for which we use the senses (Div.
.), or arts like medicine, music, or astronomy (Div. .–), or philoso-
phy, to include ethics, logic, and physics (Div. .–), or matters of
politics and law (Div. .–). For each subject on the list, he shows by
example that we would prefer to discover the truth by the means he
mentions rather than by divination. Blind Tiresias could not tell black
from white by divination; we would go to an astronomer for facts on the
sun and moon rather than to a diviner; divination will be of less help in
deciding what behavior is just than will ethical philosophy; we would go to
a statesman for advice on political issues rather than to a diviner. So (if we
suppose that the list is exhaustive) there is no type of state of affairs of
which divination is the preferable method of investigation. At last to this
survey is added a premise:
nam aut omnium debet esse, aut aliqua ei materia danda est, in qua versari
possit.
For either [divination] ought to be about all states of affairs, or it must be
given some subject matter with which it can be concerned. (Div. .)

Timpanaro () also speculates that the new definition was prompted by Carneadean criticism
(lxiii). He proposes Antipater as the source of the version of the definition that Marcus gives (.,
.) – praedictio et praesensio rerum futurarum quae fortuitae [sunt] – and Posidonius as the source
of Quintus’ version, where putantur replaces sunt.

Reason to think that Carneades was responsible for the whole argument of the sally is that part of
the ultimate dilemma (the discussion of divine foreknowledge, Div. .) resembles explicitly
Carneadean material at On fate . But note that the dilemma also contains a relative of the Lazy
Argument that Carneades may have avoided, On fate .

As Hannibal said to his host who, warned off by extispicy, dared not take Hannibal’s advice and go
to war, An tu . . . carunculae vitulinae mavis quam imperatori veteri credere? (Div. .), “You would
rather trust some cutlets of veal, than a seasoned general?”
 Marcus’ Arguments against Divination
Since the survey is supposed to have shown that divination meets neither
of these criteria, there is probably no divination.
In one way, I think this is an incisive challenge. It gets at a sense we may
have that it is hard to pin down exactly what divination is. Of course, we
know roughly what its goal is – to receive communications from the gods –
but it is still hard to say what it consists in. The relationship between
astrology and astronomy illustrates the problem. In many ways the two arts
are concerned with the same subject, the movements of the heavens. But
one (astronomy) is the art whereby those movements are understood,
modeled, and predicted. The other (astrology) uses some of the results of
astronomy to produce analyses of events on earth. Even astrology’s advo-
cates could admit that its methods were less certain and exact than those of
astronomy. (Ptolemy, Tetrabiblos .) Marcus even says that astrology
ignores important astronomical results. (Div. .) Thus astrology looks
like an ersatz art, parasitic on astronomy. One could say something similar
about the relationships between augury and ornithology, haruspicy and
anatomy, and so on. Nowadays we might say that divination looks like a
pseudo-science. Carneades’ problem gets at this intuition quite precisely: is
there any type of question we would sooner settle by divination?
But it is harder to make out a compellingly rigorous version of Marcus’
argument. Why should we grant him his premise that divination must
have some type of state of affairs as its peculiar subject matter? It seems to
me that we can make the best sense of this premise if we think that
Carneades commented specifically on divination as defined by Chrysippus,
and that Marcus (although he does not say as much) repeats this criticism
in order to motivate the later Stoic view used by Quintus. One reason to
think so is that Marcus’ examples in pursuit of Carneades’ subject matter
objection are not especially concerned with the future. He challenges
divination to tell black from white, to teach us about music, to explain
the Liar Paradox, and to predict celestial events. This would be perverse in
reaction to Quintus’ definition of divination, in which divination is a
certain sort of prediction or foreseeing of the future. But on Chrysippus’
definition divination artfully observes and interprets signs from the gods
whether they be signs of the future, present, or past. As we saw in my last
chapter (p. ), Chrysippus captures the importance of prediction in
divination by making it divination’s “duty” (officium, Div. .), not by
making it the only thing divination could do.
I now turn to the second part of the “sally,” in which Marcus offers to
the Stoics a carefully supported version of Quintus’ definition of divin-
ation. (Div. .–) Marcus says that, according to Quintus, divination is
. The “Sally”: Div. .– 
earum rerum praedictionem et praesensionem quae essent fortuitae, “the
prediction and foreseeing of those states of affairs that are by chance.”
(Div. .) As we have seen, Marcus presents this new definition as a
noteworthy reaction by Quintus to the sort of problem he pointed to in
part one of the “sally.” But he has more to say about why Quintus may
have chosen the definition he did. He remarks that, in doing so, Quintus
kept divination away from “those conjectures that possess art or wisdom,
and from those states of affairs that are grasped by the senses or by the
arts.” (Div. .) Thus Marcus interprets the “chance events” definition
of divination as an attempt to mark off a set of states of affairs of which we
do not have a better method of investigation. Marcus immediately points
out a problem with this attempt: some other arts, like navigation or
weather forecasting, aim at “chance” outcomes. Even a perfect navigator
might be overtaken by a storm that nobody could have predicted, so that
there is a degree of chance either in the ship reaching port or in its failure
to do so. Thus perhaps the Stoics have failed to escape Carneades’ problem
with the “chance events” definition. Perhaps there will be other, preferable
arts to predict future chance events, as indeed we would rather watch a
weather forecast than consult a diviner on whether it will rain tomorrow.
But Marcus also supplies the Stoics with an answer to this challenge: the
sort of “chance” future events predicted by divination are those that none
of the non-divinatory arts could predict, even those that deal with chance
events like the weather. In fact, they are events no human art could predict.
“So it remains that those matters of chance can be divined which neither
any art nor any wisdom can foresee.” (Div. .) Marcus’ suggestion is
that there are some future events that not even a learned sage, equipped
with wisdom and with the most complete and pertinent of the other arts,
could predict other than by those arts associated with divination. For the
Stoics, that would be to say that no human, even in principle, could
reliably predict these events without help from something inhuman,
namely, from the gods. These future states of affairs, then, would be a
subject matter all divination’s own. This, then, is how Marcus suggests the
Stoics should understand the “chance events” definition of divination that
Quintus advanced.

Here we encounter an important problem when Marcus substitutes “are by chance” for Quintus’
“are thought to be by chance” in Div. .. I shall discuss this below, pp. –.

. . . ab iis coniecturis, quae haberent artem atque prudentiam, et ab iis rebus, quae sensibus aut artificiis
perciperentur . . .

reliquitur, ut ea fortuita divinari possint, quae nulla nec arte nec sapientia provideri possint.
 Marcus’ Arguments against Divination
Why might Marcus connect the “chance events” definition with an
attempt to elude all the predictive capabilities of the senses and the arts?
We think of chance events as unpredictable in one sense – nobody can be
sure whether a chance event will happen – but not in another – we may
well feel that we can assess how likely a chance event is, and sometimes
even that we can say with confidence, albeit not with certainty, what will
transpire. But Marcus’ interpretation of the “chance events” definition
makes better sense if he has in mind the Stoic notion of chance. The Stoics
defined chance epistemically. For them, chance is a “cause unclear to
human reasoning.” Thus defined, chance had a place in the Stoics’
determinist world. Of course, no cause was unclear absolutely, but only
to human reasoning. God, determining all things, was aware of all causes,
so in this sense “chance” could be called “divine.” Thus for a Stoic
“chance states of affairs” will be states of which we do not know the cause.
In one sense the sorts of events predicted by arts like navigation and
weather forecasting are Stoic chance events – we cannot assess the causes
of tomorrow’s weather competently enough to make very precise predic-
tions about it. We use the evidence of today’s weather to assess the current
causes of tomorrow’s weather, and thus whether it will rain. We might get
this prediction wrong, because it is hard to know all the relevant causes and
to make the right inferences from those we do know. Nevertheless, with an
art like meteorology, we try to use signs that are part of the causal system
leading up to the future outcome. If we suppose that all arts other than
divination in one way or another predict future events through such of
their causes as we can now make out, then absolutely chance events, those
absolutely whose causes are absolutely obscure to human reasoning, will
not be predictable by humans without divine help even by the other arts.
In this second part of the “sally,” Marcus indeed seems to read Quintus’
“states of affairs which are by chance” this way, because Marcus himself
suggests that chance states of affairs are supposed to be by chance not in
the sense that they require the meteorological or navigational sort of art to
predict them, but in the sense that no merely human art can predict them,
even in principle. In the second stage of the sally – but not subsequently –
Marcus works with a notion of chance as something relative to different
observers and predictors can know, that is, an epistemic notion. I think
he does this because he understands that the Stoic notion of chance is

Alexander, Mantissa . Bruns = SVF ., Aëtius, Placita .. = SVF ..

Simplicius, In physica, . Diels = SVF ..

Alexander of Aphrodisias, at his Mantissa .–, criticizes just this feature of Stoic chance.
. The “Sally”: Div. .– 
epistemic and because in this passage he is trying to “tempt” the Stoics into
a certain version of their view of divination.
Such, then, is the interpretation of Quintus’ definition that Marcus
offers the Stoics because (he implies) it marks off a subject matter that is all
divination’s own, and thus it is safe from Carneades’ criticism. In the third
part of the “sally” he tries to subject this “future chance events” account of
the subject matter of divination to a dilemma. The structure of the
dilemma is as follows. Either future chance events are fated beforehand,
or they are not. If they are not fated beforehand, they are not predictable,
even by the gods, and thus there is no divination. If they are fated
beforehand, then divination is neither useful nor advantageous, and there-
fore there is no divination. Therefore there is no divination. I shall
consider each horn of the dilemma in turn.
The first horn of the dilemma (Div. .–) aims to show that future
chance events are in principle unpredictable, even by god, and hence that
they are not amenable to divination. It is here that we encounter the first
notable straw man in Marcus’ speech. Marcus asserts that chance outcomes
are those that could have happened otherwise: “For what is fortune, what
is luck, what is chance, what is accident, except a case that falls out such
that, or happens such that, it could have fallen out or happened other-
wise?” (Div. .)
Now, you might say that this is not clearly a straw man, since Chry-
sippus was careful to leave room for the contingent in the fated world he
described, that is, for what is possible but not necessary. So there is a
sense in which a Stoic can agree that some events that will happen could
happen otherwise. She would say that the proposition “John can choose
toast over porridge tomorrow morning” was true yesterday, although it was
also true, and causally predetermined, that I would choose porridge this
morning, which I did. Hence you might think that what Marcus describes
as “by chance” is what a Stoic would call “contingent”: what will happen,
but might not happen.
But in fact Marcus’ notion of chance is not consistent with a Chrysip-
pan notion of what is possible but not necessary. For Marcus concludes
from it that god himself does not know whether a future chance event will
happen:

quid est enim aliud fors, quid fortuna, quid casus, quid eventus, nisi cum sic aliquid cecidit, sic evenit, ut
vel aliter cadere atque evenire potuerit?

On fate –, DL ., cf. Bobzien () –.
 Marcus’ Arguments against Divination
nihil enim est tam contrarium rationi et constantiae quam fortuna, ut mihi ne
in deum quidem cadere videatur, ut sciat, quid casu et fortuito futurum sit. si
enim scit, certe illud eveniet; sin certe eveniet, nulla fortuna est; est autem
fortuna; rerum igitur fortuitarum nulla praesensio est.
For there is nothing so contrary to reason and regularity as chance, so that it
seems to me that it does not fall even to god to know what will be by chance
and fortune. For if he knows, it will certainly happen. But if it will certainly
happen, there is no chance; but there is chance; therefore there is no
foreseeing of chance events. (Div. .)
A Stoic will not concede that there is chance of the sort implied by this
argument, nor that we could make the argument sound by substituting
“contingent” for “by chance.” A Stoic views contingency such that even if
god knows that some event will happen, and even if the event is deter-
mined now, it might still be contingent. So in On fate  Marcus addresses
Chrysippus:
tu, et quae non sint futura, posse fieri dicis . . . neque necesse fuisse Cypselum
regnare Corinthi, quamquam id millensimo ante anno Apollinis oraculo edi-
tum esset.
You say that things which will not happen, too, can happen . . . and that it
was not necessary that Cypselus reign in Corinth, although this had been
declared by the oracle of Apollo a thousand years before. (Translation from
Sharples ())
Chrysippus says that an event can be foreseen by god but still be contin-
gent, but Marcus (in Div .) says that what is foreseen by god cannot be
by chance. Thus Marcus’ version of chance in Div. . is not compatible
with Stoic contingency. As a result, a Stoic will not accept that a contin-
gent event which “could have happened otherwise” is a chance event in
Marcus’ sense of “chance.”
Furthermore, if we assume that Marcus uses “certainly” (certe) to mean
“necessarily,” then a Stoic will deny the second conditional premise of
Marcus’ argument in Div. . (quoted above), that if something will
certainly happen, then it will not happen by chance. For a Stoic, if an
event necessarily will happen it might still happen by chance, because an
event that necessarily happens might be obscure to human prediction. For
example, we might be ignorant of all or some of its antecedent causes, even
if they now obtain and somebody who knew about them could

The Stoic will also deny the first conditional premise, that if god knows that something will happen,
then it will certainly happen. Chrysippus’ views imply that god knows many future contingents.
. The “Sally”: Div. .– 
predict their effect. Marcus ought to know that this is the Stoic position.
Indeed it seemed in the immediately preceding section that he was aware
that for the Stoics chance is an epistemic notion. Admittedly, when he
offered the epistemic formulation in Div. . he did not explicitly include
the human element: “it remains that those chance things can be divined
which can be foreseen neither by art nor by wisdom.” But that these
“chance things” can be divined, according to the Stoic position he recom-
mended in Div. ., implies that the gods can foresee them though
humans cannot. Thus, in that passage, he implies that chance things
cannot be foreseen by human art or wisdom, but can be foreseen by the
gods. In sum, Marcus appears in Div. .– to flout what he understood
in Div. .–.
Even more seriously, Marcus implies that a chance event is one without
a cause:
qui potest provideri quicquam futurum esse, quod neque causam habet ullam
neque notam, cur futurum sit?
Who can foresee something that will be, which has neither any cause nor
mark of why it will be? (Div. .)
But for the Stoics, every event has antecedent causes. In the case of
chance events those causes are obscure to humans, but they are certainly
not obscure to god. Marcus says here that chance events are without
causes, and thus concludes that even a god cannot predict future chance
events. Thus, because of the notion of chance it requires, the final
dilemma of the “sally” attacks a straw man, that is to say, although
Marcus purports to argue against Quintus’ Stoic defence of divination,
in fact he argues against another position altogether, that no Stoic could
endorse. For it introduces a notion of chance foreign to the Stoics and
concludes from it that there cannot be divination of chance events. Yet,
as we have seen, earlier in the “sally” Marcus seemed to understand the
Stoic idea of chance.
What should we make of this? First, it is instructive to compare the later
Aristotelian Alexander of Aphrodisias’ approach to the Stoic definition of
chance. Alexander contends that the Stoics wished to keep the term
τύχη, “chance,” in their lexicon despite their determinism. He says that in
order to do so, they redefined the term in a way incompatible with any
common concept of chance.

Alexander, On fate .–, Mantissa .–.
 Marcus’ Arguments against Divination
τί γὰρ ἄλλο ποιοῦσιν οἱ τὴν τύχην καὶ τὸ αὐτόματον ὁριζόμενοι αἰτίαν
ἄδηλον ἀνθρωπίνῳ λογισμῷ, ἢ τύχης τι σημαινόμενον ἴδιον εἰσάγουσίν τε
καὶ νομοθετοῦσιν;
For what else do they do who define luck and chance as ‘a cause obscure to
human reasoning,’ than introduce by this decree some private signification
of ‘luck’? (Alexander, On fate .– Bruns, translation from Sharples
(), modified)
Alexander develops this point in various ways, but he seems inclined to
rule the Stoic definition out of court while he supposes that his own is
founded on “common and natural conceptions.” (On fate .)
Marcus lacks a developed alternative view of chance like Alexander’s,
but perhaps we can make best sense of Marcus’ decision to argue against a
straw man if we suppose that he takes a similar line. That is, Marcus opts
to insist on a “common sense” notion of fors, fortuna, casus, or eventus – all
words for something like “chance” or “luck” – whatever we want to call it.
(Div. .) This “common sense” notion regards a chance event as one
that might have happened otherwise, in the sense that it was not determined
beforehand. Marcus then argues against the Stoic account of divination
using this, non-Stoic notion of chance. The result is that Marcus argues
against a straw man so far as the Stoics are concerned, but in a way that
might appeal to the rest of us who rely on unreconstructed common sense:
if the Stoics wish to escape from the subject matter objection with the
“chance events” definition of divination, he says to us, they must use a
notion of chance the rest of us cannot accept.
If Marcus’ task is to give the most convincing answer to Quintus’
speech, why would he proceed in this way? I do not claim that there is a
smooth answer to this question. In particular, that Cicero writes the first
horn of the dilemma of the sally in this way without advertising the fact to
the reader must count, I think, as a failure of my principle of literary unity
(see p. ). But I would invoke section . of this chapter, where
I argued that an Academic skeptic, seeking to balance the plausibility of
the Stoic case rather than to refute it on its own terms, might give an
argument against the Stoics, targeted at least at the non-Stoic parts of his
audience, on a basis that no true blue Stoic could accept. Thus, perhaps a

Of course, this failure, or a few failures, on Cicero’s part to live up to this principle do not imply
that I am wrong to see it at work in the dialogues at large.
. The “Sally”: Div. .– 
non-Stoic audience, as Alexander did, will accept that we should find fault
with the Stoic notion of chance.
Let us turn now to the second horn of the ultimate dilemma of the
“sally” (Div. .–). In the first horn, Marcus tried to show that if the
events predicted by divination (i.e., future chance events on Quintus’
definition) are not determined beforehand, then they cannot be predicted.
In the second horn he tries to draw unfavorable consequences from the
premise that they are determined. Now, since he considers chance events
to be those not determined beforehand, he says, as he must:
aut si negas esse fortunam et omnia, quae fiunt quaeque futura sunt, ex omni
aeternitate definita dicis esse fataliter, muta definitionem divinationis, quam
dicebas praesensionem esse rerum fortuitarum.
But if you deny that there is chance, and you say that everything that is
happening, and which will be, was decided by fate from all eternity, change
your definition of divination, which you said was the foreseeing of chance
events. (Div. .)
This sentence presents an often remarked problem. For Marcus asks
Quintus to “change his definition,” yet Marcus has already misquoted
Quintus’ definition from Div. .. At Div. . Quintus defined divination
to be of future events which fortuitae putantur, “are thought to be by
chance” (see p.  above). But here in Div. ., Marcus says that
Quintus defined divination as of events that are by chance. The misquota-
tion looks significant. Marcus, in effect, accuses Quintus of asserting
contradictory claims on the Stoics’ behalf, both that there is divination of
events that are not fated and that all events are fated. But what Quintus
said is that divination is of events that are thought not to be fated. What
Quintus said is perfectly consistent with those events being fated, so that
Marcus’ accusation is flatly false. Or so say those who think this misquo-
tation amounts to another problematic straw man.

We should also recall that Marcus’ speech was not Cicero’s last word on the Stoic doctrine of fate.
That was still to come in On fate, as Marcus himself hints in Div.: “the Stoics say a lot of things
about this fate of theirs, about which, [more] elsewhere. For now, [let’s stick to] what we need.” (sed
tamen apud Stoicos de isto fato multa dicuntur; de quo alias; nunc quod necesse est. Div. ..) So
perhaps we are to suppose that Marcus can give more reasons to deny Stoic position on fate and
chance, but that such reasons would only be elucidated by the homonymous character in On fate.
But I would not put any weight on this, since it is my view that, at the time he wrote Div., Cicero
was not sure that he would have time to write On fate (see pp. –).

See Denyer () , Timpanaro () xciii–xciv, Schultz () , cf. Pease (–) ,
 n. , and on Div. . and ..
 Marcus’ Arguments against Divination
On the one hand, I am not much troubled by Marcus’ change to the
wording of the definition. For one thing, who is to say that Quintus has the
carefully worded version of the definition and Marcus the deviant one?
For another, according to me, the different wording of the definition is not
much of a distortion from Quintus’ point of view. When Quintus gave the
definition, I interpreted him to mean that divination is of states of affairs
whose causes are thought to be obscure to human reasoning, so that we
cannot predict them. This, I argued, was Quintus’ way of giving divinatory
prediction its own subject matter (see p. ). If that interpretation was
correct, whether divination is of events that are thought to be unpredictable
by human beings, or is of those that simply are unpredictable by human
beings, is not a crucial difference. On the other hand, Marcus does set up
another, more serious straw man in his criticism of Quintus’ definition.
This is that his criticism depends on the same problematic alteration to the
meaning of “chance” that I have just explored in the first horn of the sally’s
dilemma, since it requires that “by chance” does not mean what Quintus
meant by the term, but rather that it means “not fated.”
Let us then return to Marcus’ argument for the second horn of the
dilemma. He presents two problems for divination in a fated world. These
are, first, that the Stoics cannot coherently claim that divination is of any
use (Div. .–) and second that divination is, if anything, disadvan-
tageous for mortals in a Stoic world. (Div. .–) The former objection
turns on divination’s role as a way for the gods to give us advice. Augury is
a natural source of examples since it purports to give divine endorsement
or veto of proposed courses of action. Take the destruction of the Roman
fleets in the First Punic War (Div. .) that Quintus attributed to a
vitium, that is, to a contravention of auspices (Div. .). Marcus pre-
sumes that this means that had the auspices been obeyed, the fleets would
not have been lost. But, he says, if this conditional is true, then it was not
fated that the fleets should be destroyed. On the other hand, if it was fated
that the fleets should be destroyed, then they would have been destroyed
whether or not the auspices had been obeyed. In which case, the auspices
gave no useful information about the future (in whatever sense augury
gives information about the future) and hence there was no divination. So,
either the outcome was fated, or there was divination of it, but not both.

Cf. Balbus’ description of divination in DND .. From Quintus’ point of view, Balbus is certainly
careless with his formulation praedictiones et praesensiones rerum futurarum, “predictions and
foreseeing of future events,” since he makes no reference to chance.
. The “Sally”: Div. .– 
Marcus thinks that this result can be generalized to dispose of divination
entirely. (Div. .)
Here Marcus has made a very weak objection, and one with which the
Stoics had no trouble. Admittedly, it was important to the Stoics that at
least some divination could alter our behavior. In the argument Quintus
attributed to Chrysippus, Diogenes of Babylon, and Antipater of Tarsus
(Div. .–, cf. p.  above) the gods ought to tell us the future if they
can because it makes a difference to us, “for we will be more careful if we
know [what will happen].” But Marcus’ argument gets no purchase on fate
and divination as the Stoics understand them. The Stoics do not take a
view of fate whereby the destruction of the fleets is fated regardless of what
happened beforehand. This is clear in general because they hold that
everything is fated, and in particular because they were not troubled by
the Lazy Argument. The Lazy Argument aimed to reduce fate to an
absurdity, by alleging that, for example, if it is fated that you will recover
from your illness, you do not need to bother going to the doctor to be
treated, since you will recover in any case. (On fate –) But Chrysip-
pus’ claim is not that it was fated that Oedipus would be born to Laius
whether or not Laius slept with a woman, but rather that it was fated both
that Laius would sleep with Oedipus’ mother and that Oedipus would be
born as a result. According to Chrysippus, even though it was fated that
Oedipus was born it is also true that had Laius not slept with a woman,
then Oedipus would not have been born. This presumably applies to
courses of action advised by divination, too. It was fated that the fleets
be destroyed because it was fated that the auspices would be contravened,
but it is also true that had the auspices been obeyed, they would not have
been destroyed. Thus to claim that the auspices gave useful information
about the future is consistent with Stoic determinism.
Marcus’ second problem for divination in a Stoic fated world is that it
would not be a benefit to us. Implicitly, this argument relies on the Stoics’
belief in divine providence. Since they hold that the gods arrange every-
thing in the cosmos to our advantage, they could not accept that the gods
would arrange for divination to work if it were not to our advantage.
Marcus illustrates his point with some examples, specifically the bad ends
that met all of the First Triumvirate. Would it have benefited Caesar to
know that he would be murdered by his own associates, and that his body
would lie unattended in front of a statue of Pompey? (Div. .) We
should recall that this example was extremely visceral for Cicero. When he
wrote these lines, he had seen Caesar stabbed to death no more than a few
weeks before. In a Stoic world Caesar’s wretched end was ineluctably fated
 Marcus’ Arguments against Divination
to happen so, says Marcus, surely it would have been a trouble to Caesar to
know about it. Now, a Stoic might reply that if Caesar fulfilled his duty
to become a sage, and thus developed appropriate attitudes to the gods,
fate, and death, then he would find that such foreknowledge was not
troubling. But Marcus resourcefully quotes Zeus’ grief at the fated death
of his son, Sarpedon. If Zeus himself can be grieved at a fated death, surely
the Stoics are wrong to think that we could bear our own fated misfortunes
lightly? (Div. . cf. Iliad .–)
A reply to this argument is that it seems to ignore all the future except
death and disaster. Should we not weigh against the bad news the
uplifting, or just useful, information we would receive? But perhaps this
reply is too hasty. Imagine that we all received many accurate predictions
about good and bad aspects of our future, including the time and manner
of our death. It is at least plausible for Marcus to charge that this last fact
would be the one that weighed on our minds. Although the Stoics claim
that divination will make us “more careful” (cautiores, Div. ., cf. .,
.), not all divinatory predictions seem to work that way. Some of
them foretell simply that a given event will happen, and not necessarily a
welcome one. Phalaris’ mother learnt in a dream that her son would show
“brutal . . . cruelty,” Cyrus that he would reign for thirty years. (Div. .)
Perhaps in some cases such information might allow us to plan well. Cyrus
could plan his program in government neatly. But at other times the news
is simply distressing, and does not seem to open any avenue to improve
ourselves or our circumstances that we would not have had anyhow. One
thinks of characters in tragedy who struggle to void outcomes they believe
to be fated. We might well agree that it was better for Cicero as he wrote
Div. to be ignorant of the fate that awaited him in turn, less than two years
later.

Shakespeare’s Julius Caesar has given us the image of a Caesar who shrugs off endless portents of
doom. Shakespeare was no doubt inspired by the material in Plutarch, Life of Caesar . But already
in Div. Cicero left some evidence that the run up to Caesar’s death was “a strange disposèd time,” as
“he” calls it in Shakespeare’s play (Act , scene ), and suggested that Caesar ignored these
warnings. Compare Div. . with Div. ., .–.

I take this to be the strongest reading of the Stoic reply that Marcus gives at Div. .: omnia levius
casura rebus divinis procuratis, “everything that will befall us is easier to bear when we attend to
divine matters,” although rebus divinis procuratis also connotes “when the divine is rightly
propitiated.” According to my reading of Balbus on religion, a Stoic will attend to traditional
religious rites precisely to express the right attitude to the divine.

The Stoics, of course, would read this passage allegorically. Their gods do not grieve.

Or was he? Among the prodigies for  BC recorded in the text of Julius Obsequens we find that,
“an image that Marcus Cicero had set up in front of the shrine of Minerva the day before he went
into exile through a decree of the people was cast prone by the force of a whirlwind, its upper and
. The “Sally”: Div. .– 
But ultimately, this second argument against divination in a fated world
is not successful either. For, as Marcus admits, it too depends on the
success of the previous argument, that aimed to rule out conditional
predictions in divination. (Div. .–) Since the Stoics think that
divination can give conditional predictions, there is a way that divination
can give us obviously beneficial information even about death and disas-
ters, namely, how to avoid them: if you do this, you will avoid disaster. For
this reason, I think, the argument is not finally persuasive to a general
audience. In addition, a Stoic will not be moved by Homer’s attribution of
grief to Zeus. Which is to say, a Stoic sage could receive no accurate
prediction about the future that she would think bad or find emotionally
troubling, and therefore any Stoic would resist the urge to think any news
about the future was bad for him.
As a whole, then, the “sally” is not compelling. Marcus is right to say
that it is fought “lightly armed.” (Div. .) The first horn of the final
dilemma attacks a straw man in that it uses an anti-determinist view of
chance instead of an epistemic one. I have said that the appearance of such
straw men should not necessarily lead us to think that Marcus, or Cicero,
misunderstood Quintus’ speech in Book . There are ways in which they
can make for an effective reply from a skeptic. But the skeptic must finger
as rebarbitatively implausible only those Stoic views that might indeed be
rebarbitatively implausible. An epistemic notion of chance does not seem
to me to be one such. The second horn is ineffective, too. For the Stoic
theory of fate leaves room for contingency and thus for divination to be
helpful, but is not “fatalism” because it does not say that certain events are
fated while others are not. Yet even if the final dilemma falls flat, because
the “sally” tries to corner the Stoics in exactly the position that Quintus
took on their behalf, it shows that Marcus’ aim is to engage with divination
exactly as Quintus, as opposed to the Stoics in general, presented it. Cicero
is wrestling with two carefully chosen sides of the divination puzzle, not
slapping together unrelated speeches.
lower arms and head broken. It portended evil for Cicero himself.” (turbinis vi simulacrum, quod
M. Cicero ante cellam Minervae pridie, quam plebiscito in exilium iret, posuerat, dissipatum membris
pronum iacuit, fractis humeris bracchiis capite: dirum ipsi Ciceroni portendit. Prodigiorum Liber ,
Rossbach () .) It seems that Obsequens took his data from Livy, who could easily have had
reliable information about so recent a year. But the dire portent seems rather too perfect – the story
went that Antony had Cicero’s head and hands cut off and displayed on the rostra. (Plutarch, Life of
Cicero, .–.)
 Marcus’ Arguments against Divination
. The Main Strategy: Div. .–

Quintus’ case for the reality of divination was based on his many examples
and anecdotes, where outcomes followed successful predictions. Marcus
indeed takes on Quintus’ argument from the examples. There are a few
cases where Marcus doubts or denies the truth of Quintus’ anecdote: he
doubts Calchas’ conjecture from the sparrows as described by Odysseus at
Iliad .– (“were I to trust it,” si crederem, Div. ., cf. p. 
above); he doubts the stories of prodigies attendant on Plato’s and Midas’
boyhoods (Div. .); he denies that alleged rains of blood, or sweat on
statues, can consist of blood or sweat (Div. .); he describes the legends
of Romulus and Attus Navius as “confected myths”; he doubts the Delphic
oracles reported by Herodotus, whom he considers no more reliable than
Ennius (Div. .). But for the most part he is as committed to the truth
of the anecdotes, both prediction and outcome, as Quintus is. Marcus’
strategy, then, is not to reject the evidence. It is to reject the inferences
from the evidence to divination’s reality.
Now, Marcus does not state in one place his strategy against Quintus’
inferences. But we can discern patterns of argument. Against natural
divination Marcus argues that the apparent correlations between predic-
tion and outcome that Quintus cites could, and furthermore did, come
about just by chance, and that they should not convince us that there is any
underlying divinatory connection between predictions and outcomes. In
modern terms, he argues that the apparent correlations that Quintus can
exhibit are statistically insignificant, and hence not really correlations in
the sense of evidence for a connection. In the case of artificial divination,
Marcus uses the same general tactics as he uses against natural divination,
but adds another. This responds to an important part of Quintus’ answer
to Cotta’s challenge, in which Quintus gives an ideal history of divinatory
arts. Marcus tries to discredit every stage of that ideal history. I shall cover

Moatti ()  says that the difference between Quintus and Marcus is that between “two forms
of thought.” Quintus’ form of thought “lumps all domains of reality together and considers
prodigies as facts just as much as natural events,” such that “the discourse of the ancient records
is . . . held to be true, as if it actually created reality.” Marcus’ form of thought “picks out whatever
can be explained and rejects everything else as fiction.” I do not agree that the Stoicism that Quintus
represents in Div. was insensitive to the usual preference for rationally explicable facts, or that
Marcus in Div. argues from different assumptions about fact, fiction, and explanation than Quintus
does. Quintus has carefully limited his main argument to empirical evidence rather than causal
explanation because of the nature of the debate about divination in particular. Marcus chooses
largely to respond to this argument in kind, even if he also points out that philosophers like the
Stoics themselves usually demand explanations (e.g. Div. .).
. Marcus on the History of the Divinatory Arts 
Marcus’ treatment of the history of the divinatory arts first (section .).
Then I shall turn to Marcus’ arguments about chance (section .).
. Marcus on the History of the Divinatory Arts

Quintus’ ideal history of divinatory arts (on which see pp. –) was
summarized at Div. .–. According to Quintus’ history, theorems
must originally be formulated by coniectura, “conjecture,” induction of
predictions from single signs. In this sense artificial divination is “founded”
(nititur) on conjecture and can never produce a theorem that is, at bottom,
rationally justified by anything better than induction. (Div. .) But its
theorems are empirically tested by observation over a very long period,
which is reason to rate highly the probability of the theorems. (Div. .)
Meanwhile, new conjectures are made as new candidate signs are encoun-
tered (as Quintus makes explicit in Div. .), which may result in a
theorem that is eventually accepted into the divinatory art. I suggested that
this history was an answer to Cotta’s unde oriatur, “where does it spring
from?” that need make no appeal to a causal theory of divination (see
p. ). Since this ideal history was thus an important part of Quintus’
response to DND, we should expect Marcus to take it seriously. He does,
as we see in his pattern of efforts to undermine every aspect of the
Quintus’ story.
Cotta’s unde oriatur put the spotlight in particular on the origins of the
divinatory arts. How could anyone in the first place have supposed that
such arcane practices delivered truth from the gods? One role of conjecture
in Quintus’ history was to answer to that question. Let us therefore begin
with conjecture, since it was the first stage in his history.
Marcus takes on conjecture primarily in the section of his speech
devoted to the haruspices’ responses on prodigies. (Div. .–) He
approaches prodigies as claimed miracles, like the birth of a foal to a mule
(Div. .) or rains of blood (Div. .), or as startling events like the bees
on Plato’s lips (Div. .) or the crowing of the cocks at Lebadia (Div.
.). In Div., we see that the haruspices would (at least sometimes)
respond to these events by conjecture about the “meaning” of the remark-
able event rather than by the application of theorems – no doubt some of
the events in question were unique and as such could not satisfy an
established theorem of haruspicy. Quintus gives a small example of a
conjecture in haruspicy, where a snake appeared from under an altar where
the general Sulla was sacrificing. A haruspex conjectured that the sign
meant that Sulla should immediately march his army. Sulla takes the
 Marcus’ Arguments against Divination
advice, with good success. (Div. .) Sulla’s haruspex may have looked for
inspiration in similar signs about snakes theorems already on the books,
but, this being a conjecture, he must not have had a precisely relevant
theorem to work with, so that such conjectures are at best skilled induc-
tions that lack long empirical confirmation. Marcus says:
iam vero coniectura omnis, in qua nititur divinatio, ingeniis hominum in
multas aut diversas aut etiam contrarias partis saepe diducitur. ut enim in
causis iudicialibus alia coniectura est accusatoris, alia defensoris et tamen
utriusque credibilis, sic in omnibus iis rebus, quae coniectura investigari
videntur, anceps reperitur oratio.
But now, every conjecture—and divination is founded on conjecture—is

led by people’s wits [to conclusions] in many different or even contrary
directions. For just as in cases at trial the accuser has one conjecture, the
defender another, and both conjectures are believable, so in all of those
matters that seem right to investigate by conjecture we find that rhetoric
can go either way. (Div. .)
The point must have carried particular weight with an audience used to
pleading in court. In court, induction from limited evidence was often
obviously not sufficient to tell the true conclusion apart from other possible
inductions, so that the rhetorical skill of those pleading to the jury could be
decisive. Marcus points out new divinatory conjectures from remarkable
events have no more certitude than the forensic version. Now, this point is
certainly a score for Marcus against divinatory conjecture. It is also a
disquieting reminder that this sort of conjecture was the ultimate origin
of artificial divination, according to Quintus. But that reminder does not
critically weaken Quintus’ general defence of artificial divination. For a key
strength of Quintus’ ideal history is that he concedes that divinatory arts
might stem from some baffling or shabby original conjectures about livers,
lightning, and so on, and it is to this basis that he adds “long observation,”
the empirical testing that could fallibly, but reliably, sort the true theorems
from the false.
In criticizing their origins in conjecture, then, Marcus must give us
further reasons to think that it is a problem that the divinatory arts did not
begin in the way Quintus suggested. This he does, by rubbishing all stories
of the early beginnings of the arts, be they Quintus’ version of the
introduction and testing of theorems, or traditional myths about the
founding of the arts, or a mixture of the two. In each case, Marcus suggests
not merely that there was something problematic about the way that the
early theorems of an art were conjectured, which Quintus’ theory could
. Marcus on the History of the Divinatory Arts 
accommodate, but that the problem in question undermines, or threatens
to undermine, the whole history of the art. Here Marcus’ methods often
look “rhetorical” in that they amount to “is it really plausible that . . .?”
This is precisely appropriate in a philosophical context where Marcus tries
to persuade Quintus (and us) of what happened historically, on the basis of
little or no evidence (cf. pp. –). Let us examine three examples of
this pattern of argument.
First, in one version of the tactic, Marcus suggests that a divinatory art
was begun for inappropriate reasons. The clearest case is in Marcus’ discus-
sion of divination by lot: “the whole thing was invented by trickery, either
with monetary gain in mind, or superstition, or error.” (Div. .) For
Quintus assumes that the inventors of divinatory arts aimed to produce
true theorems, and started from some inductive conjecture about a sign.
Marcus suggests that the art does not have its roots in conjecture at true
predictive theorems, but rather was dreamed up in order to extract money
from the gullible, or to promote superstition. Now, even supposing that
this were right, Quintus might still be able to defend the usefulness of the
art. For again, in his model it may not matter why the theorems of arts
were proposed so long as they were then tested empirically by long
observation. But by casting doubt on the origins of the process Marcus
calls into question the commitment of diviners by lot to finding usefully
predictive theorems at all. Suppose that a significant proportion of diviners
by lot, not merely the founders of the field, were charlatans. That certainly
would undermine Quintus’ ideal history, since theorems would not gain
admission to art because long testing made them seem true. Rather, they
would be chosen because guesswork and experience suggested that they
could part fools from their money.
A second version of this pattern of argument is Marcus’ suggestion
about the origin of lightning as a divinatory sign in both haruspicy and
augury:
nonne perspicuum est ex prima admiratione hominum, quod tonitrua iactusque
fulminum extimuissent, credidisse ea efficere rerum omnium praepotentem
Iovem?
Surely it’s transparent that, as a result of people’s early wonder, because they
had been been terrified of thunder and lightning strikes, they [the first
haruspices] came to believe that Jupiter, mighty in all matters, made these
things happen? (Div. .)

tota res est inventa fallaciis aut ad quaestum aut ad superstitionem aut ad errorem.
 Marcus’ Arguments against Divination
Here the criticism is not that primitive lightning diviners were disingenu-
ous like the inventors of the lots. Rather, terror is proposed as their
motivation. In DND, such terror is connected with superstition, falsely
believing that the gods threaten more harm to us than in fact they do.
(DND ., ., ., .) That seems to be the situation Marcus
describes here: faced with thunder and lightning, the early lightning
diviners were scared of Jupiter. Once again, this suggestion alone is not a
threat to Quintus. Superstition might lead one to pursue earnestly the
“real” predictive power of lightning and to propose and test successful
theorems. The problem with a whole divinatory art that is motivated by
superstition is more insidious. For all parties to the debate in DND and
Div., superstition and excessive fear of the gods is to be excluded from
religion. To put the point in Stoic terms, diviners motivated by supersti-
tion are certainly fools and prone to error, especially (one presumes) in
matters to do with the divine. As a mechanism for generating successfully
predictive theorems Quintus’ ideal history survives these criticisms of its
early stages, but again Marcus insinuates that the real answer to “where
does it come from?” may be discreditable in ways that undermine the
whole process suggested in Quintus’ ideal history.
The third example of this pattern of argument is refreshingly different,
and finds its basis in a divinatory art that came attached to a myth of its
own founding. Marcus says that according to the Etruscans, a being called
Tages with the appearance of a boy was ploughed up from a field and
dictated the elementary set of theorems (disciplina) of haruspicy to a large
crowd. (Div. .) “It [the disciplina] has since grown as new things are
learnt and related to those same principles [i.e. what Tages dictated].”
Although Marcus does not say so, this is a history of divination import-
antly different from Quintus’. The original theorems were not conjec-
tured, but rather revealed by the mysterious Tages and thus, it seems, they
required no empirical testing. New theorems were added to this base by
some process of comparison with the existing theorems – presumably by
the same process of coniectura that the haruspices could call on in their
responses on Roman prodigies.
Although different from his own account, even if true, the myth of
Tages need not worry Quintus. If the original theorems of an art were
revealed rather than conjectured, so much the better. On the other hand,
there is no mention in this mythical history of long observation. The new
theorems seem to be admitted to the art by comparison with the old, and

eam postea crevisse rebus novis cognoscendis et ad eadem illa principia referendis.
. Marcus on the History of the Divinatory Arts 
not by empirical confirmation. Again, perhaps this would be acceptable
where the original theorems were indeed revealed, if there were some
reliable process of comparison. But Marcus, understandably, thinks the
story altogether absurd: “Is there anyone so mentally deficient that he
believes that there was ploughed up – shall I say a god, or a human?”
(Div. .) The problem for Quintus here is that the myth is what the
haruspices themselves report. “This is what we hear from them, they keep
this in written records, this is the source of their discipline.” (.) Thus
the haruspices themselves do not claim that their foundational theorems
were conjectured or confirmed by long observation, and the basis they do
claim, revelation, seems absurd. If the foundational theorems are untrust-
worthy in this way, the art as a whole seems open to question. Marcus says,
“We don’t need Carneades to refute this, do we?” (num ergo opus est ad
haec refellenda Carneade?, .)
This concludes our examination of Marcus’ pattern of argument against
conjecture at the origins of the divinatory arts. We now turn to the second
stage of Quintus’ ideal history, where theorems were confirmed by obser-
vation over very long periods, while further candidate theorems could still
be proposed by conjecture. Marcus also denies that this process happened
as Quintus described. On astrology, where Quintus’ claim of antiquity was
especially spectacular (Div. .), he says:
nam quod aiunt quadringenta septuaginta milia annorum in periclitandis
experiendisque pueris, quicumque essent nati, Babylonios posuisse, fallunt; si
enim esset factitatum, non esset desitum; neminem autem habemus auctorem,
qui id aut fieri dicat aut factum sciat.
For in that they say that the Babylonians had done trials and checks on [the
horoscope of] every child born for , years, they are deceived. For if
that had been [the Babylonians’] habit, they would not have ceased from it,
but we have no authority who either says that it goes on, or knows that it
was done. (Div. .)
Again, this is a “rhetorical” argument about plausibility. It is implausible
(Marcus claims) that the Babylonians would ever have lost the habit of
testing horoscopes. Thus, if they ever did so, they ought to do so now. But
there is no good evidence that they do so (or ever did so). Therefore, it
seems they never did. This is a tissue of probabilities and an argument
from silence, albeit a principled one. But how else can Marcus argue

estne quisquam ita desipiens, qui credat exaratum esse, deum dicam an hominem?

haec accepimus ab ipsis, haec scripta conservant, hunc fontem habent disciplinae.
 Marcus’ Arguments against Divination
against Quintus’ proposed history of Babylonian astrology? So far as the
Ciceros know, there is no good historical evidence about the matter, and
all they can do is scrap about plausibility. Now, as a matter of fact the
Babylonians did keep some astrological records. But certainly they did not
do so for hundreds of thousands of years, and probably not even for long
enough that we should find it plausible that they had any insight into
predictive horoscopy of individual lives (see p.  n. ). So we should
agree with Marcus, not Quintus.
Elsewhere, Marcus has more theoretical objections to “long observa-
tion.” At Div. ., in arguing against haruspicy, he compares the sort of
observation possible in the case of prognostication, that is, in activities like
weather forecasting (see pp. –), with that possible in divination.
The Stoics Posidonius and Boethus, he says, investigated the causal links
between prognostic signs and outcomes. Even if they could not understand
the causes linking the two, at least they could observe regularities of sign
and outcome. But with divination, “what does the statue of Natta have, or
the bronze tablets of the laws struck by lightning, that has [previously]
been observed or is ancient?” (Div. .) For the prognostic theorem
that “if herons fly inland vocalising, a storm is coming” (Div. .) one can
objectively mark down when the sign or the outcome have occurred. It is
thus possible to observe the relevant facts to test the theorem. But the
lightning strike on Natta’s statue was a unique event. What are the
pertinent facts about it to which a theorem should be applied, or by which
a theorem can be tested, or from which a coniectura can be made? The best
ways to generalize from it (Marcus suggests that the Nattae were nobles
and so danger was to be expected from the noble class, Div. .) seems
arbitrary and obscure – “How cunningly Jupiter thought that up!” says
Marcus, snidely. (hoc tam callide Iuppiter <ex>cogitavit!, Div. .) So, if
we suppose that divinatory signs and outcomes in general need this sort of
rather arbitrary interpretation in order to be related to theorems, then the
process of building up divinatory arts looks unamanageable.

Nattae vero statua aut aera legum de caelo tacta quid habent observatum ac vetustum? Lightning strikes
on Natta’s statue and the law tablets were endorsed by Cicero as a portent of Catiline’s conspiracy in
his De consulatu suo, cf. Div. ., .–.

Denyer () – criticizes Marcus’ “arbitrariness” objections on the grounds that with the Stoic
model signs can be arbitrarily connected with their correlated outcomes in the same way that words
can be arbitrarily connected to their meanings (he has in mind parts of Marcus’ speech – Div. .
and Div. . – that I discuss below and above, respectively). It is an important insight that signs
and outcomes can be arbitrarily connected in the Stoic model. But I do not think that Marcus
misunderstands the point. In the passages I am discussing here, he complains about necessary
arbitrariness in deciding which aspects of which events are significant and suggests that it is a barrier
. Marcus on the History of the Divinatory Arts 
Marcus chose a rather hard case, a lightning strike on a portrait statue,
to illustrate his point. In extispicy or augury it seems that the signs – flights
of birds, anatomical features on particular parts of the liver – would be
easier to mark down objectively. But elsewhere Marcus suggests that even
such apparently objective signs suffer from a subjective element, in that
the parameters for isolating them would originally have been arbitrary.
On the more objective aspects of the haruspices’ divination by lighting,
where the direction of the lightning was studied, he says:
valet autem in fulguribus observatio diuturna, in ostentis ratio plerumque
coniecturaque adhibetur. quid est igitur, quod observatum sit in fulgure?
caelum in sedecim partis diviserunt Etrusci. facile id quidem fuit, quattuor,
quas nos habemus, duplicare, post idem iterum facere, ut ex eo dicerent, fulmen
qua ex parte venisset. primum id quid interest? deinde quid significat?
But long-continued observation is effective with lightning, while with
prodigies reason for the most part and coniectura are used. So what is there
that has been observed with lightning? The Etruscans divided the sky into
sixteen parts. It was an easy thing to double the quarters we have, and then
to do the same again, so that they could say thereby from what direction the
bolt came. First, what difference does that [i.e. the direction] make? Next,
what does it signify? (Div. .)
These rhetorical questions are not answered. Marcus’ point seems to be
that the division of the sky adopted by the haruspices is based on the
cardinal points and thus, in a sense, is arbitrary. It was chosen before there
was any evidence that it was the right way to divide up the sky to find
useful divinatory theorems. Perhaps the gods use a system where the sky is
divided into sevenths, or pinwheeled nine degrees from the cardinal
orientation? If so, then the Etruscan system might help in some cases,
but in general will be misleading.
to the formation of successful divinatory arts. This involves no criticism of the arbitrary relation of
signs and outcomes.

A defender of Quintus’ outlook might be able to dispose of this and similar objections to early
diviners, using Denyer’s metaphor of the divinatory system as a “language” for communication
between gods and people. If the haruspices adopt a certain division of the sky, however arbitrary or
even perverse, could not the gods then adapt their signs to that division? If communication is the
goal it would make sense to change their “language” to suit the “listeners.” (Cf. Denyer ()  on
Marcus’ objection that there are local differences in the art of extispicy – Denyer argues that the
gods might use a different “language” to suit each locale, just as people do.) This is a reasonable
reply to Marcus, although I do not strictly agree with Denyer that artificial divination is susceptible
to the “language” analysis in the way that natural divination is, because I do not think the signs of
artificial divination (e.g. a crow croaking) “mean” the prediction derived from them (non-naturally
or otherwise).
 Marcus’ Arguments against Divination
Marcus’ most systematic attack on long observation is reserved for
Quintus’ ideal history of augury, in place of which Marcus offers an
ingenious history of his own. It is ingenious because in it Marcus walks
the tightrope between his position as an augur and his role as a debunker of
Stoic divination. Of course, it would be consistent for Marcus to play his
conversational role of Academic negative disputer while being personally
committed as an augur, or even to the reality of augury (which, as it will
turn out, he is not). Nevertheless, even while delivering his skeptical
speech, Marcus chooses to signal from time to time his political support
for the continuance of state divination even during his negative arguments
(e.g. Div. .). This applies most of all to augury, where he says that those
who disobey religion and ancestral custom by contravening the auspices
deserve “every torture.” (omni supplicio, Div. .). Marcus’ alternative
history of the discipline is consistent with this outlook. In it there are three
periods, of which the second and third are described here:
non enim sumus ii nos augures, qui avium reliquorumve signorum observatione
futura dicamus. et tamen credo Romulum, qui urbem auspicato condidit,
habuisse opinionem esse in providendis rebus augurandi scientiam (errabat
enim multis in rebus antiquitas), quam vel usu iam vel doctrina vel vetustate
immutatam videmus; retinetur autem et ad opinionem vulgi et ad magnas
utilitates rei publicae mos, religio, disciplina, ius augurium, collegii auctoritas.
For we [i.e. the Roman augural college] are not those augurs whose job is to
tell the future by observation of birds or the other signs. And yet I believe
that Romulus, who founded the city under auspices, had the opinion that
the science of augury lay in foreseeing facts (for antiquity erred in many
matters). We now see this science changed, whether by use, or by dogma, or
by age. But the custom, the religious duty, the teaching, and the law of
augury, and the authority of the college, were maintained with regard both
to public opinion and to their great advantages to the commonwealth. (Div.
.)
In Marcus’ second period of augural history, Romulus (and presumably
other early Romans) believed that augury was divination in the sense at
stake in Div., that is to say, they thought it really was a means by which the
gods could communicate with the augurs. That is why they included it in
the city’s institutions. But Marcus thinks they were mistaken about
augury’s success as a divinatory art. He must think that the error became

I interpret this passage differently than does Jerzy Linderski. He supposes that there were known to
Cicero two theories of augury within the augural college, () that augury was limited to allowing the
gods to warn against actions by particular magistrates on particular days, and that as such it did not
amount to a means of predicting the future, or () that a divine warning amounted to a prediction
. Marcus on the History of the Divinatory Arts 
known, because he says augury was retained other than for its divinatory
success: in the third period of divinatory history, augury was retained for
political reasons, not because it succeeded in divining the will of the gods.
There were changes in the art since the time of Romulus, and these can be
chalked up to changes of habit over time or to doctrinal changes in the
college, but not, it seems, to an evolving empirical art. This alternative
history already goes a long way to dispute Quintus’ ideal history in the case
of augury. The augurs of the second period, including Romulus, were just
mistaken about augury’s divinatory success. It follows that they were not
witnesses to the development of successful theorems. Quintus might reply
that, even if this were so, true candidate theorems dreamt up this second
period could have been properly tested by long observation during the
third period. But Marcus suggests that in the third period augury’s long
history was not devoted to improving the success of the art. Augury was
retained for reasons wholly other than the pursuit of successful divination.
Thus false theorems may have been allowed to stand.
The third period of Marcus’ history raises some obvious questions. First,
if augury is not divination, what are its advantages to the body politic at
Rome? It seems to me that an answer to this question is necessary for a full
understanding of Marcus’ position, and that we are especially entitled to an
answer when he says that those who frustrate these advantages should be
subject to torture. But it also seems to me that we are given no answer to
of disaster if the auspice was contravened, and hence that divination can predict the future.
Linderski says that in Div. . Marcus criticizes Romulus for taking augural view number (),
that is, for thinking that augury predicts the future. (Linderski () –) He writes, “It is
important to remember that here [i.e. in Div. .] Cicero speaks as an augur, and not as a sceptical
philosopher.” (p. ) So Linderski thinks that Marcus prefers one of the two augural views, number
(). I think that Marcus speaks as a skeptical philosopher and argues that augury is not divination at
all, not even of warnings against action. After all, the purpose of his speech is to argue that there is
no divination. If he were to concede that augury is divination, of any sort, then the speech would
fail. Thus when Marcus says that Romulus erred in believing that augury lies in providendis rebus he
just means that Romulus erred in believing that augury was divination. In the following paragraph
Marcus says, etenim, ut sint auspicia, quae nulla sunt, haec certe quibus utimur, sive tripudio, sive de
caelo, simulacra sunt auspiciorum, auspicia nullo modo (Div. .), “Anyhow, suppose that there are
auspices – which there are not – the ones that we use, whether the tripudium, or in the sky, are
certainly counterfeit auspices, and not auspices at all.” The clause underlined shows that Marcus
rejects the divinatory power of the auspices wholesale. Even if the art were properly practiced, there
would be no auspices, in the sense that there would be no real communication from the divine,
neither in the form of warnings against particular actions nor in the form of predictions of the
future. I think that Div. . is in the same vein. It says that there is no problem with a Roman
augur arguing against the reality of augural divination, because it is not really the job of the augural
college to divine anyway. (Santangelo  pp. – also says that Marus denies that augury is
divination, but by this he means that Marcus denies that it is prediction of the future.)
 Marcus’ Arguments against Divination
this question in Div.. Second, what is the meaning of ad opinionem vulgi,
“with regard to public opinion”?
We might try to answer these questions in a way that is charitable to the
honesty and intelligence of the augurs who began Marcus’ third period.
Perhaps they conceived augury as a system of religious performance that
could symbolize and foster a beneficial attitude to divine authority, namely
that the republic hoped to act in such a way that it won divine approval. In
this light, the system of consulting the sacred chickens before any state
action, where the chickens always delivered a purely symbolic “yes,” would
be just the sort of augural performance that was required. On this sort of
reading, the opinion that augury was supposed to promote among the
public was just this, that those in authority in the republic were concerned
to act in a such a way that the gods would be pleased.
But it seems to me that this is not the most natural reading of our
passage. The most natural reading is that augury’s advantage to the
republic simply was its effect on public opinion, namely, it convinced
the public that the gods approved of the actions of the authorities, even if
the augurs themselves were not convinced that this was so. Such an
attitude on the part of the augurs of the third period would at best be
thoughtless and cynical, and at worst amount to dishonest propaganda,
especially if, like Marcus, these augurs were prepared to resort to torture to
enforce their strictures.
Does Marcus, or Cicero himself, stand convicted of these sorts of
attitudes? In this passage, he (both character and author) certainly flirts
with them to a degree that seems objectionable. But we should be careful.
Public opinion in particular is a motivation for the practice of augury that
Marcus attributes to others. He never endorses it himself, even in this
passage where he speaks in the role of Quintus’ skeptical opponent. When
he gives his own views on the matter at the end of Div. (see pp. –),
he endorses traditional religious practices, but not with regard to public
opinion. Indeed, in that passage his call for the rejection of belief in
divination is to benefit et nobismet ipsis et nostris, “both ourselves and our
countrymen,” by the removal of a weight of superstition that has oppressed
nearly all human beings, both in Rome and elsewhere. (Div. .) It does
not seem that Marcus wants to restrict this change in attitude towards
divination to some class of priests, aristocrats, or intellectuals. It is instruct-
ive to compare this with the code of religious law that Marcus gives in the
Laws, which has a preamble intended to encourage the people to put their
religion in the context of just the sort of natural theology that Balbus
recommended, and Marcus partially preferred, in DND. (Laws .–,
. Marcus on the History of the Divinatory Arts 
cf. pp. –.) Thus although Cicero writes for a small class of educated
Romans, Marcus is not an esotericist and seems to intend, in principle,
that the change of attitudes resulting from his philosophical moderation of
religion should be open to the Roman people at large. This is so even
when, in practice, a plan of the sort that might occur to a modern “public
intellectual,” like an energetic program of education for those parts of
society beyond the readership of his books, seems alien to Cicero’s world.
Now, the second and third periods of augury that Marcus describes do
not include an account of its first origins. He says that Romulus dealt with
a “science,” apparently a preexisting science, of divination. But he alleges
that the traditional suggestions for an early, formative period are inconsist-
ent or implausible. A philosopher, he says, must offer an account of
augury’s inventio, the process of its discovery (Div. .):
quo modo autem haec aut quando aut a quibus inventa dicemus? Etrusci tamen
habent exaratum puerum auctorem disciplinae suae; nos quem? Attumne
Navium? at aliquot annis antiquior Romulus et Remus, ambo augures, ut
accepimus. an Pisidarum aut Cilicum aut Phrygum ista inventa dicemus?
placet igitur humanitatis expertis habere divinitatis auctores?
But how shall we say these things were discovered, and when, and by
whom? The Etruscans, meanwhile, have the boy who was ploughed up as
the author of their discipline—who do we have? Attus Navius? But Rom-
ulus and Remus, both augurs according to tradition, were a few years older

In DND ., Cotta says that in a contio, a speech before the people, it is difficult to deny that there
are gods, but that in a private conversation like the one we see in DND, it is very easy to do so. One
might interpret this as an esotericist remark, meaning that it is dangerous or disadvantageous to say
before a religiously credulous public what one can say freely among the sophisticated, freethinking
few. But I do not read the passage in this way. Rather, I think Cotta describes his experiences of his
own convictions. These experiences vary from one social context to another, a phenomenon to
which as a skeptic he may be especially attuned. When he, a priest and an aristocratic leader, speaks
about political or religious matters in front of his fellow Romans, he finds it natural to believe in the
gods. When he discusses philosophy with his friends, the arguments on all sides of the question
come to mind, and it is easy for him to see why one would deny that there are gods. See Wynne
(). Meanwhile, in Div. ., where Marcus embarks on arguments that haruspicy is not really
divination, he carefully asserts that it should be performed for political reasons and in order to
maintain the traditional religion. But he then seems to regret this reflex clarification and says, “But
we are alone. We may inquire without attracting resentment, even against me, as I doubt about a
number of points.” (sed soli sumus; licet verum exquirere sine invidia, mihi praesertim de plerisque
dubitanti.) Here it is clear that Marcus thinks that to say in public that haruspicy is not really
divination would attract a negative reaction, a reaction that would be relieved, or mitigated, by
emphasizing that it nevertheless be practiced for reasons of politics and correct religious
performance. But he does not say that in public he would argue that haruspicy is divination.
Rasmussen () and Beard () defend Cicero from charges of hypocrisy, on the basis that
ancient Romans were not concerned about religious belief.

Cf. the similar traditions at Laws ., where Marcus argues that augury must have been successful
among the Phrygians, Cilicians, and so forth.
 Marcus’ Arguments against Divination
than him. Or shall we say that these were discoveries of the Pisidians, or the
Cilicians, or the Phrygians? So do we hold then that people ignorant of
humane learning were authorities in divinity? (Div. .)
The result that Marcus draws from all this is shockingly strong, and
worded to call to Quintus’ mind his reading of DND. The myths of
Romulus and Attus Navius, ancestors who lend authority to augury, are
touchstones in Cicero’s more positive writing on the subject. Romulus, of
course, was said to have received auspices to found Rome by the observa-
tion of the flight of birds. (Livy ..–..) Attus Navius was said to be a
diviner under Tarquinius Priscus, the fifth king of Rome. As Quintus tells
the story, he came to prominence when, using his augural staff to invite
signs from the flight of birds, he was given divine direction to find the
biggest bunch of grapes in a vineyard. Tarquinius then consulted him as an
augur. To prove his power, he was able to cut through a whetstone with a
knife. (Div. .–) Romulus and Attus appear together in the Laws in
support of augury (.). In DND Balbus challenges Cotta: “Should we
despise the staff of Attus Navius?” (Atti Navi lituus ille. . . contemnendus
est?, DND .) Cotta bridles: “I cannot despise the staff of Attus Navius.”
(nec possumus Atti Navi. . . lituum contemnere, DND .) In Div., Quin-
tus had reminded Marcus of Romulus’ augury (.) when he tells Attus’
story (Div. .–). But now Marcus flatly rejects these stories (“con-
fected myths,” commenticiis fabellis, .) that all the other speakers in
both dialogues have treated as inviolable:
omitte igitur lituum Romuli, quem in maximo incendio negas potuisse
comburi; contemne cotem Atti Navii.
So get rid of Romulus’ staff, which you say the fiercest fire could not burn,
despise the whetstone of Attus Navius. (Div. .)
From Marcus the augur, this condemnation comes as something of a
shock. But it is consistent with his defence of his position as an augur.
His use of the word “confected” (commenticiis) is significant. In DND
Balbus, for example, used this term for regrettable aspects of the gods of
traditional myth (DND. . cf. p. ). Similarly, Marcus wishes to retain
the performances of augury, but to reject the notion that these performances
achieve divination, and thus also to reject the myths that attribute successful
divination, or miracles, to augurs like Romulus or Attus. Just as Balbus

With this phrase, Marcus turns Balbus’ Stoic criticisms of poetic myth back against the Roman
“history” used in the Stoic case for divination: Balbus said that the gods of poetic myth were
commenticios (DND .), cf. p. .
. “Chance Can Imitate Truth” 
would claim that he is pious to reject these myths, so Marcus would claim
that he is respectful both of augury and of his early Roman ancestors, when
(in this case unlike the Stoics) he rejects silly myths about them. In short,
Marcus’ position is that it is not a problem for an augur to argue that there is
no divination, because augury is not divination in the first place.
Just as he has done with the other divinatory arts, then, Marcus
systematically rubbishes every stage of Quintus’ ideal history in the case
of augury, and offers as the real reason for its development either that it is a
performance that confers political benefits on the state, or that at some
times augurs may have mistakenly believed augury was divination. He
claims that this rejection is consistent, in every sense, both with his defence
of his practice as an augur, and with his philosophical arguments against
divination.
. “Chance Can Imitate Truth”

I said that against Quintus’ support for all divination Marcus deployed the
argument that Quintus’ evidence could and did come about by chance,
and a further tactic (the questioning of Quintus’ ideal history) against
artificial divination specifically. Now I that I have surveyed the latter, I will
turn to Marcus’ tactics to do with correlation and chance.
The passage that comes closest to summarizing these tactics is Div.
.–. It is here that Marcus deals with Quintus’ use of examples of
prodigies for which Marcus himself is the source. In these cases it is most
starkly obvious that Marcus cannot question the truth of the “evidence,”
that is, the facts which Quintus cites as signs and outcomes. Instead,
Marcus tells us how he will question the statistical correlation, and the
divinatory connection, between these pairs of facts. He reminds us that
Quintus asked not why divinatory predictions happen, but whether they
happen, “As if I would either concede that [divination] happens or that it’s
appropriate for a philosopher not to inquire about the cause because of
which each thing happens.” (Div. .) Here, Marcus indicates first that

We are left to guess at what these political benefits might be. My best guess is that Marcus thinks
that augury is a system of gesture that expresses and reinforces Rome’s desire to act only in
conformity with the divine will, so as to promote thoughtful and just political decisions. But
perhaps he means that there are political benefits simply in obedience to ancestral authority as such.
In Div. . he says that the wise man carries on ancestral religion in order to protect the maiorum
instituta, “the institutions of his ancestors.” My thanks to an anonymous reader for the press for the
latter observation.

quasi ego aut fieri concederem aut esset philosophi causam, cur quidque fieret, non quaerere!
 Marcus’ Arguments against Divination
he denies that Quintus’ proposed correlations by sign and outcome
amount to evidence of a connection (he will not concede that it happens)
and second that, he will argue that in any case there could be no divinatory
connection of the sort Quintus claims (he will not concede that Quintus
can get away with not explaining why his signs and outcomes are correl-
ated, and that he will even assert that there is no divinatory explanation).
We find each of these two patterns of argument played out in many places
in Marcus’ speech. I shall take each pattern in turn, first the argument that
Quintus’ evidence from outcomes could have come about by chance, then
the argument that it did come about by chance.
Marcus elaborates the chance argument most fully in Div. .–.
Now in making this argument, Marcus takes on a question that Quintus
himself considered early in his speech. Quintus attributed it to Carneades.
In a manner reminiscent of Balbus’ arguments from beauty (DND .,
quoted p. ), Quintus recruited numerical probability to suggest that it
is an unreasonable charge:
quicquam potest casu esse factum, quod omnes habet in se numeros veritatis?
quattuor tali iacti casu Venerium efficiunt; num etiam centum Venerios, si
quadringentos talos ieceris, casu futuros putas?
What can be done by chance that has all the detailed properties of truth?
Four dice thrown make a Venus throw by chance; surely if you throw four
hundred dice, you don’t think that there’ll be one hundred Venus throws
by chance? (Div. .)
Quintus has a reasonable point. The theory of numerical probability is a
modern discovery, so Cicero and his readers could not use it to assess what
Quintus says. But, since we do not play ancient Roman dice games, we
should use some calculations to get a feel for the odds of what Quintus
describes. A talus was a “knucklebone,” in effect a die with four sides, with
its sides numbered one to four. In a Venus throw, four such dice are
thrown so that a different number appears on each die. (Pease (–
) ad loc.) If we assume fair and independent dice, the probability of
this outcome is !/ = / = /. Thus Venus throws will often
happen by chance. But the chance of one hundred consecutive Venus
throws is then (/) = .. . . x -. Obviously, that is a negligible
chance, the equivalent of  consecutive coin tosses coming up heads.

There may be a play here on omnes numeros habere, meaning “have all the detailed properties” (cf.
Pease (–) ad loc., On duties ., On ends ., DND .) and numerus meaning the
result of a throw at dice (OLD s.v. numerus .e).
. “Chance Can Imitate Truth” 
Therefore, as Quintus suggests, if one threw one hundred consecutive
Venus throws, one could reasonably conclude that the dice were loaded.
Quintus produces other examples that he says are like this: paint thrown
on a board might yield a rough outline of a face, but one would not think
Apelles’ Venus of Cos was the result of such a mess. A pig might by chance
scratch the letter A in the turf with his snout, but not Ennius’ Andromache.
Quintus, of course, supposes that the apparent successes of divination are
in the Venus or Andromache category – one should conclude that there is a
negligible probability that they have occurred by chance. Thus he thinks
that they are evidence divination happens, even if we cannot say how or
why it happens.
In many ways, Quintus makes an argument that is conceptually recog-
nizable from the point of view of modern statistics. He implies that the
observable divinatory outcomes are like the rolling of many dice. In that
sense he concedes that the divinatory outcomes can be analyzed as poten-
tially the outcome of chance. But the probability that they would fall out
as they did just by chance is so low that we can be reasonably certain that
there is some non-chance bias in the system – that there is divination.
On the other hand, Quintus runs the risk of drawing too strong a
conclusion from his scratching pig and paint flinging examples: “Plainly,
the fact is that chance never imitates truth perfectly.” (sic enim se profecto
res habet, ut numquam perfecte veritatem casus imitetur, Div. .) Quintus
avoids saying explicitly that chance cannot imitate truth perfectly, which
would be too strong a conclusion. But he comes very close when he says
that chance never imitates truth. A charitable reading would conclude that
Quintus is making a good bet. “Fling paint at a board until the end of
time, you’ll never reproduce Botticelli’s Venus.” But it is possible that the
Venus could come about by chance. Something similar ought to be
Quintus’ point about divination: “If you happen on the Venus, you would
be deeply unreasonable to conclude that it was produced by flinging paint
at a board.” Similarly, he could argue, the frequency and accuracy of the
truth of divinatory predictions ought to defy an attribution to “mere
chance” not because they cannot be the result of chance but because we
would be very unreasonable to think that they are. This is why Quintus
talks of chance imitating truth. The overall truth of the mass of divinatory
predictions is improbable in the way that the order of a great work of art is
improbable. But it is not clear that this charitable reading is the right

It is clear that Quintus’ point in Div. . is about the mass of predictions rather than individual
predictions because the immediate objection Quintus considers in Div. . is that divinatory
 Marcus’ Arguments against Divination
one. When Quintus says that chance never imitates truth he might mean
that it cannot.
There is another, clear problem with Quintus’ statistical argument.
What he lacks is any rigorous way to show that the deliverances of
divination are like the hundred Venus throws rather than the single Venus
throw, or like the masterpiece rather than the rough outline of a face. For
one thing, he lacks the sort of tools that modern statisticians have
developed to analyze and compare sets of empirical data. For another, he
lacks the sort of data that would be amenable to this analysis. His mass of
cherry-picked anecdotes are unsuitable.
In addition to this sort of general correlation between predictions and
outcomes en masse, Quintus sometimes has another sort of correlation
in mind. Perhaps, Quintus’ Peripatetic hero Cratippus argued, “it is
enough to make sure of divination that something should once be
divined in such a way that we think that it could in no way have fallen
out by chance.” (Div. .) Here the match between just one predic-
tion and its outcome would be so good that it seemed to be like an
artistic masterpiece rather than the outline of a face. We can imagine,
say, a dream, whose detailed and emotionally compelling adumbration
of some distant event, whether through straightforward foresight or in
some opaque way that required interpretation, is so overwhelming to
the dreamer that she can only conclude that the dream was divinatory,
just as one cannot look at the Venus without the overwhelming sense
that some artist made it. Indeed, this is the sort of example that seems
most amenable to comparison with the Venus or Andromache examples
and thus to Quintus’ statistical argument. But this potentially promising
line of argument faces a problem: there is not (or not obviously) any
such case among the examples that Quintus gives. Such a case would
presumably have to come from natural divination, since the predictions
of artificial divination are short propositions, insufficiently rich to have
the properties required. But Quintus describes none of his cases of
natural divination in this way.
How does Marcus reply to Quintus’ treatment of chance? In his
explicit discussion of Div. . (see p. ), he does not object that
Quintus failed to show that divination at large, or in specific instances,
predictions are sometimes false, which would be a fruitless objection if Quintus were making a
point only about some particularly good predictions.

satis est ad confirmandam divinationem semel aliquid esse ita divinatum, ut nihil fortuito cecidisse
videatur.
. “Chance Can Imitate Truth” 
is like his masterpiece cases. (Div. .–) Instead, he takes the unchar-
itable reading of Quintus’ conclusion. “You said that one hundred Venus
throws cannot coincide by chance.” (dixisti. . . centum Venerios non posse
casu consistere, Div. .) Marcus’ response, quite rightly, is to insist that
they can, or that flung paint could produce the Venus. He concludes, “So
chance can imitate truth, which you were just now denying.” (potest
igitur, quod modo negabas, veritatem casus imitari, Div. .) The imper-
fect of “you were just now denying” is significant. With it Marcus refers
not only to Div. . but also to Quintus’ whole speech, which
depended in general on his empirical, and thus statistical, approach.
Thus in Div. .– Marcus insists that even if the correlation between
predictions and outcomes were statistically like the order in a Botticelli or
in Dante’s Divine Comedy, that correlation could still be by chance. This
is true. Yet it is a lame argument against Quintus’ defence of divination if
we read the latter charitably, because we would still be unreasonable to
attribute the Venus to chance, which seemed to be Quintus’ point when
charitably understood.
But elsewhere Marcus takes the next step and attributes divinatory
successes to chance. That is, he denies that the successes of divination
are even statistically like the masterpiece cases. At the start of his main
arguments he accuses Quintus of using evidence “that can be either true by
chance, or false and concocted through malice.” (qui aut casu veri aut
malitia falsi fictique esse possunt, Div. .) In the case of artificial divin-
ation, he makes charges that follow this pattern against many pieces of
evidence that Quintus cited. He suggests that the signs of Catiline’s
planned coup that Quintus cited from Marcus’ poem De consulatu suo
might as well have been appropriate by chance as by the will of the gods
(Div. .). He says that many predictions of the haruspices are wrong
(Div. .). Thus he asks “How many things predicted by them actually
come out? Or, if something comes out, what can be said to show why it
did not come out by chance?” (Div. .). In Div. . he accounts for a
remarkably successful prediction of the haruspices with “nor is the haruspi-
ces’ misfortune so great that what they say will be so never happens, even by
chance.”

quota enim quaeque res evenit praedicta ab istis? aut, si evenit quippiam, quid adferri potest, cur non
casu id evenerit?

neque enim tanta est infelicitas haruspicum, ut ne casu quidem umquam fiat, quod futurum illi esse
dixerint.
 Marcus’ Arguments against Divination
For natural divination, Marcus takes on Cratippus’ argument that there
is some natural “power to divine” in humans:
adsumit autem Cratippus hoc modo: ‘sunt autem innumerabiles praesensiones
non fortuitae.’ At ego dico nullam (vide, quanta sit controversia); iam adsump-
tione non concessa nulla conclusio est. at impudentes sumus, qui, cum tam
perspicuum sit, non concedamus. quid est perspicuum? ‘multa vera,’ inquit,
‘evadere.’ quid, quod multo plura falsa? Nonne ipsa varietas, quae est propria
fortunae, fortunam esse causam, non naturam esse docet?
Cratippus takes the following as a premise: ‘But there are innumerable
instances of foresight that are not by chance.’ But I say there are none. See
how great our dispute is: now that the premise is not granted the syllogism
fails. ‘But we are shameless who will not grant this when it is so clear.’
What is clear? ‘Many things turn out true,’ he says. What about the fact
that many more turn out false? Surely this diversity [of outcome], which is
a property of chance, shows us that chance and not nature is the cause?
(Div. .)
As with haruspicy, Marcus argues that the mass of outcomes of natural
divination does not show a strong enough correlation to count as evidence
for a divinatory connection.
Now, Cratippus had argued that just one instance of successful divin-
ation was sufficient to show that we have the power to divine, just as, if a
man sees just once, we can say he has the faculty of sight. Marcus joins
Quintus in reporting this aspect of Cratippus’ argument. (Div. .). But
Marcus later argues that there is no one instance that cannot be attributed
to chance. (Div. .) We make innumerable conjectures from the
impressions of the drunk or insane and, like a man who shoots all day,
we sometimes hit the mark.
quid est tam incertum quam talorum iactus? tamen nemo est, quin saepe iactans
Venerium iaciat aliquando, non numquam etiam iterum ac tertium. num
igitur, ut inepti, Veneris id inpulsu fieri malimus quam casu dicere?
What is so uncertain as dice throws? Yet there is no one who, throwing
many times, does not sometimes throw a Venus, and sometimes even twice
and thrice in a row. So surely we don’t want to say, as silly people do, that it
happens by the intervention of Venus rather than by chance? (Div. .)
In other words, the visions of madmen coupled with conjecture can
occasionally produce striking results, but this is no evidence that anything
other than a lucky prediction has happened. There is no need to appeal to
divine agency, so there is no reason to think that there has been divination,
or that we have in our soul any power to divine that expresses itself in
. Conclusion 
dreams and frenzy. As with artificial divination, Marcus sees Quintus’
body of empirical data as insufficient evidence for the reality of natural
divination as Quintus, specifically, argued for it.
. Conclusion
In this and the preceding chapter, I have given an interpretation of Div.
whereby Cicero planned a unified dialogue to respond to the challenge
Cotta gave Balbus in Book  of DND, namely, where did divination come
from, and how did it come to be understood? (See Chapter  section ..)
The result is that Quintus defends a Stoic view (which he does not hold)
that is not intended to be the Stoic view of divination, but rather an
argument from empirical data, chosen to fit into Cicero’s overall project in
DND and Div. Marcus’ arguments in Book  are mostly devised to reply
to Quintus’ speech in particular. Much of what has seemed confusing
about Div. is explained when we see the two speeches as parts of the sort of
dialogues I described in my introduction, and as parts of the project
I explored in Chapter .
The upshot is the sort of skeptical “moderation” I suggested in Chap-
ter  (see pp. –): avoiding superstition by the suspension of judgment
in the face of dissension about the Central Question. It is tempting today
to see Marcus’ reply as Cicero’s sensible stance against the absurdities,
however subtly urged, of Quintus’ embrace of all divination. But we must
resist this temptation. Cicero’s plain objective is to balance the case for
divinatory phenomena as the decisive evidence for a “yes” to the Central
Question, against the case that they offer no such evidence.
But in my introduction I also suggested that even to a skeptic in
Cicero’s mould, one side or other of the question might, at any time,
seem more like the truth, and that, in addition to giving us the arguments
pro and contra, when he writes his dialogues Cicero might model this sort
of free reaction for us (see pp. –). I think that Cicero has Marcus take
an overall stance of that sort at the respective endings of DND and Div.. It
is to that stance I turn in my final chapter.

Cf. also Div. . on the oracle at Delphi, of whose predictions Marcus says that some are false
and some are true by chance. He thinks that still others are so obscure and ambiguous as to require
interpretation not just by diviners but by also grammarians. He implies that such ambiguity
increases the chance that a prediction will look true in the end.
 
Marcus’ Stance on the Central Question
I began Chapter  with the first sentence of DND, where Cicero called the
philosophical “inquiry about the nature of the gods” “most beautiful for
the mind to grasp and necessary for the moderation of religion.” Cicero
then invited us to his tribunal, where those on each side of the Central
Question speak, to make up our mind about how religion is made
moderate. In this final chapter I address the question: but what does
Cicero think?
Now there is good reason to wonder whether we should look for an
answer to this question at all. For in my opinion, as we saw in my
introduction (section .), Cicero is a radical Academic skeptic. That is,
he himself tries not to form dogmatic beliefs. Further, he writes dialogues
in the way he does exactly to avoid putting his, or anybody’s, authority
behind one answer or another. So why would he turn round and give us
“his” answer? Thus I must say on what basis I think he gives “his” answer,
and what he meant his readers to get out of it. In what follows, I first
present the textual data for Marcus’ views (section .). Then I examine
the skeptical basis on which Cicero might show us Marcus’ answer to the
Central Question (section .). Last, I present that answer (section .).
. The Evidence

Cicero shows the stance of each of his own characters – that is, of the
characters I have called “Marcus” – in the final passages of DND and of
Div..
First there is the last sentence of DND:
haec cum essent dicta, ita discessimus, ut Velleio Cottae disputatio verior, mihi
Balbi ad veritatis similitudinem videretur esse propensior.
When all this had been said, we left, with the result that Cotta’s speech
seemed more true to Velleius, but Balbus’ seemed to me to tilt closer to a
likeness of the truth.

. The Evidence 
What seems to Marcus to tilt closer to a likeness of the truth includes
Balbus’ argument that Stoic theology is adequate to render Roman religion
pious, which I examined in Chapter  sections . and .. But this
reaction is stated in a more than skeptical way. Consistent with his
Academic principles, Cicero could have written that Cotta’s speech seemed
veri similior, “more like the truth.” But he wrote that it “tilt[ed] closer to
likeness of the truth,” adding another qualification that uses the metaphor
of the movement of an arm of a balance. This suggests that Marcus does
not follow Balbus all the way or on every point. Now we are not told
Marcus’ views on Velleius’ speech, or on Cotta’s reply to it. But it is hard
to see how Marcus could, first, prefer Balbus’ speech in Book  over
Cotta’s speech in Book  but, second, also find much to like in Velleius’
arguments. Velleius’ own preference for Cotta over Balbus points to his
divergence from Marcus’ tastes. Thus we leave DND with the sense that
Marcus has a greater affinity for Balbus’ views than for Velleius’, or for
what Cotta had to say. We also have the sense that this affinity is partial or
qualified. But we are allowed no details. It seems to me that some of these
details are supplied in Div..
As we enter the conversation of Div. Quintus presents Marcus as the
author of DND. Quintus tells Marcus that Balbus’ speech “seemed to you
to tilt closer to the truth,” ad veritatem est visa propensior. (Div. .)
Marcus does not demur. Of course, Quintus is careful about the status
of a skeptic’s views. He says it seemed to tilt towards truth, not that Marcus
in Div. is still committed to his character’s view as found in DND.
Nevertheless Quintus associates Marcus in Div., the author of DND, with
Marcus his character in DND. Cicero thus licenses us to seek connections
between the views of these two characters.
The further evidence for these views comes from the closing paragraphs
of Div. . I have divided this passage with Roman numerals for
convenience:
(i) explodatur <igitur> haec quoque somniorum divinatio pariter cum
ceteris. Nam, ut vere loquamur, superstitio fusa per gentis oppressit omnium
fere animos atque hominum inbecillitatem occupavit. quod et in iis libris
dictum est, qui sunt de natura deorum, et hac disputatione id maxume egimus.
multum enim et nobismet ipsis et nostris profuturi videbamur, si eam funditus
sustulissemus. nec vero (id enim diligenter intellegi volo) superstitione tollenda

For some of the history of the interpretation of this sentence, see p.  n. .

Pease (–) has loquatur in his main text. But this seems to be no more than a typo. Even the
lemma in Pease’s own notes has loquamur.
 Marcus’ Stance on the Central Question
religio tollitur. nam et maiorum instituta tueri sacris caerimoniisque retinendis
sapientis est, et esse praestantem aliquam aeternamque naturam, et eam suspi-
ciendam admirandamque hominum generi pulchritudo mundi ordoque rerum
caelestium cogit confiteri.
(ii) quam ob rem, ut religio propaganda etiam est, quae est iuncta cum
cognitione naturae, sic superstitionis stirpes omnes eligendae. instat enim et urget
et, quo te cumque verteris, persequitur, sive tu vatem sive tu omen audieris, sive
immolaris sive avem aspexeris, si Chaldaeum, si haruspicem videris, si fulserit, si
tonuerit, si tactum aliquid erit de caelo, si ostenti simile natum factumve
quippiam; quorum necesse est plerumque aliquid eveniat, ut numquam liceat
quieta mente consistere. perfugium videtur omnium laborum et sollicitudinum
esse somnus. at ex eo ipso plurumae curae metusque nascuntur; qui quidem ipsi
per se minus valerent et magis contemnerentur, nisi somniorum patrocinium
philosophi suscepissent, nec ii quidem contemptissimi, sed in primis acuti et
consequentia et repugnantia videntes, qui prope iam absoluti et perfecti putan-
tur. quorum licentiae nisi Carneades restitisset, haud scio an soli iam philosophi
iudicarentur. Cum quibus omnis fere nobis disceptatio contentioque est, non
quod eos maxime contemnamus, sed quod videntur acutissime sententias suas
prudentissimeque defendere.
(i) Then let us say ‘boo!’ to this sort of divination, from dreams, just as
much as to the others. For, to speak frankly (ut vere loquamur), superstition
had spread throughout the nations, and weighed down nearly everybody’s
soul, and taken hold of people’s weakness. This was said in my books on the
nature of the gods, and in this discussion we have been concerned with it
especially. For I thought we would benefit ourselves and our countrymen if
we could take it away down to the foundations. But it is not the case (for
I want this to be understood carefully) that by taking away superstition we
take away religion. For, on the one hand, it is characteristic of the sage to
protect the institutions of his ancestors by maintenance of rites and cere-
monies. On the other, the beauty of the cosmos and the order of heavenly
things compels one to admit that there is some outstanding and eternal
nature, a nature that humans should look up to and wonder at.
(ii) For that reason, just as religion that is joined to some grasp of
nature should be propagated, in the same way every shoot of superstition
should be uprooted. For it presses in on us, and drives us, and wherever you
may turn it hounds us, whether you may make a sacrifice or look at a bird,
whether you see an astrologer or a haruspex, if there’s been lightning, or if
there’s been thunder, if something has been struck from heaven, if anything
like a prodigy has occurred, or has been born. It is necessary that much of
the time some one of these comes about, so that we are never permitted to

explodo = ex-plaudo, “to clap away” or, according to the OLD, “to drive offstage by clapping.” But
perhaps it has a sense of breaking the spell or driving away the bogeyman with a sudden clap. In
either case, “to say ‘boo!’,” either the drawn out version of the theatre or so as to startle, seems an
idiomatic equivalent.
. The Evidence 
rest with a mind at peace. Sleep seems to be the refuge from all troubles and
anxieties. But from sleep itself many cares and fears are born. These, at any
rate, would have less force and would be the more despised, were it not that
some philosophers had offered patronage to dreams, philosophers not to be
despised, but rather those especially sharp at seeing implications and con-
tradictions, who now are thought almost perfect and complete philosophers.
Had not Carneades [the Academic] stood against their licence, I think they
might be thought the only philosophers today. Nearly all of our dispute and
disagreement is with these Stoics, not because we especially despise them,
but because they seem to defend their views most sharply and with great
intellectual virtue. (Div. .–)
Now a skeptic assumes a mask when he argues. We would be wrong to
count his arguments as his own views. This is the end of a speech in which
Marcus has argued from behind the mask. So am I entitled to see Marcus
speaking his own mind in this passage? I take permission to do so from
the phrase ut vere loquamur, “to speak frankly.” In Div. ., Quintus
has said,
mihi vero . . . placet; his enim, quae adhuc disputasti, prorsus adsentior, et, vere
ut loquar, quamquam tua me oratio confirmavit, tamen etiam mea sponte
nimis superstitiosam de divinatione Stoicorum sententiam iudicabam
I agree, for I fully assent to what you have argued so far, and, to speak
frankly (vere ut loquar), although your speech has made me more sure, still,
even by my own choice, I judged the Stoic view of divination too
superstitious.
He goes on to say that he endorses natural but not artificial divination,
more as a result of Peritpatetic than Stoic arguments. The phrase “to speak
frankly” marks the point where Quintus takes off the Stoic mask he
assumed when he came to Balbus’ aid. Now Quintus will assert his own
view. When Marcus uses almost the same phrase (loquamur plural for
loquar singular) he signals that he, too, steps out of character to deliver his
own conclusion. I therefore think that Marcus himself asserts everything

There is consensus that some part of the ending of Div. is spoken in earnest. This consensus emerged
from Beard () and Schofield (). Wardle ()  n.  calls it the “‘Cambridge’ approach”.
Some followers are Scheid (-) –, Douglas (Douglas () –,  n. ),
Krostenko () , Leonhardt () –. This consensus looked to moderate earlier readings
which supposed that Cicero was a heroic champion of rationalism for his demolition of Quintus in
Div. Book  (Pease (–) –), or that the negative arguments in Div. Book  reveal
Cicero’s “real views” and their political edge, see Momigliano () , Linderski () –.
For this motivation, see Schofield ()  and Beard () –.
 Marcus’ Stance on the Central Question
from the sentence of Div. . in which “to speak frankly” (ut vere
loquamur) appears, to the end of his speech.
. The Epistemology of Marcus’ Stance

Here I shall present the evidence for the relevant features of Cicero’s
radical Academic skepticism, his general position described in my intro-
duction, section .. For not only did Cicero try to avoid dogmatic beliefs,
he also wanted to avoid the use of his authority to support any particular
view (see pp. –). But I shall argue that there is still room for Cicero to
show us the views of his characters, Marcus in DND and Marcus in Div..
Indeed Cicero has a reason to do exactly that.
In the preface to DND, Cicero has the following to say:
qui autem requirunt quid quaque de re ipsi sentiamus, curiosius id faciunt
quam necesse est; non enim tam auctoritatis in disputando quam rationis
momenta quaerenda sunt.
But those who ask what is our own view on each matter (quid de quaque re
sentiamus) do so more inquisitively than they need to. For the weight
(momenta) of reason, not of authority, must be sought in debate. (DND
.)
Compare this declaration with the skeptical end to Marcus’ speech in Div.
. (see pp. –). Neither Cicero in DND . nor Marcus in Div.
. denies that Cicero (or Marcus) has views (sententiae). Cicero tried
not to have dogmatic opinions about the truth, although we were told in
the Academica that he failed in this respect (Academica .). But even at
his most skeptical, at any time, the answer to a question may seem to him
one way rather than another.
So the reason not to expect an answer from Cicero is not that he would
never have a view he could report. It is rather, as DND . and Div.
. show, that he has a problem with authority (see pp. –). The
model of Academic teaching was to expose students to arguments against
one another or against some students’ thesis, removing the teacher’s

Among the members of the consensus there is disagreement over whether Marcus’ own thoughts
begin at the start of Div. . (Schofield () ) or at . cum autem proprium (Beard ()
–). Beard’s view is that the skeptical procedures subsequently stated in Div. . itself do not
allow Marcus to state the conclusions at Div. .– as his own assertions. But I would answer
that the skeptical procedures in Div. . are described in the future tense, which would allow that
the conclusions just stated did not follow these procedures – perhaps that is why the procedures need
to be reasserted. I argue below that .–. does not break these rules of skeptical discourse in
any case.
. The Epistemology of Marcus’ Stance 
authority from any particular argument. The student would, it was
expected, end up suspending judgment where in a dogmatic classroom
he would be required to assent even to arguments he could not understand
or did not really accept. But this procedure also leaves open how the matter
seems to the student. Each student may follow how it seems to him at the
time. This is how a radical skeptic may live.
Who plays which of these roles in the dialogues? Note again that in the
drama of Cicero’s dialogues there is no teacher. Professional Greek philoso-
phers are conspicuous for their absence. Instead, for each side of each
question, a Roman student will take up the case. Then each man present,
and each reader, is free to make up his or her own mind. Now some of
these characters have more philosophical expertise than others. Balbus, for
example, is supposed to have expertise in Stoicism comparable with the
leading Greeks (DND .). But even this expertise does not confer on
Balbus decisive authority over the other characters. Cicero does not mean
to confer authority over us, either. Thus if there is an analogy to the
relationship between teacher and students, it is that between Cicero the
author and his readers. Cicero orchestrated the topic, the characters, and
their speeches. It is Cicero whose name has persuaded the reader to pick up
the book. If so, then as an Academic he must be careful to withhold this
authority from arguments and conclusions.
Cicero accepts this danger in DND . (quoted on p. ). For he
acknowledges that people might want to know his view on each point. He
replies that they ought to use the weight of reason and not of authority.
This implies that it is his authority that his philosophical public might
want to follow. Of course, there is something conceited about this: was
there really an appetite for Cicero’s personal answers? Nevertheless, in the
epistemological scheme to which Cicero has dedicated himself, such
authority as Cicero the author could claim was indeed to be abrogated.
But I do not think that Cicero anywhere in DND or Div. interposes his
authorial voice in such a way that his authority is given to any view. Thus
he does not break his own rules in that way.
The views I examine in this chapter come from the mouths of the
characters I have called “Marcus” and not, so to speak, directly from
Cicero’s pen. In each dialogue, it is of course obvious that “Marcus”

The only exception is in the incomplete text we call Cicero’s Timaeus, the preface of which suggests
that Cicero planned to feature Cratippus of Pergamon in a dialogue.

At DND . Cotta remarks on Balbus’ auctoritas. Perhaps this is sarcastic, or perhaps he concedes
that Balbus does indeed have authority as an exponent of Stoic thought. If the latter, Cotta’s
subsequent speech shows that he is not influenced by such authority (cf. Academica .).
 Marcus’ Stance on the Central Question
represents Cicero, but readers also know that the conversation is a fictional
story about a particular occasion. Now, when Cicero tells us these stories
about how a question seemed to “him” at a particular, and fictional, time,
is that an infringement of the rule that he should not exert his authority?
Let us take the dialogues in turn. Marcus in DND is not even a party to the
debate in the dialogue. His view is merely the reaction of an attendee. It is
phrased carefully in the skeptical jargon of the Academica. This is as
carefully hedged a gesture as one could imagine. By contrast Marcus in
Div. is a participant. Further, his final view agrees with the arguments he
himself has just given, at length. He does not hedge it about with skeptical
qualifiers like “it seems to me.” His endorsement of the view seems much
more emphatic, and more like an infringement of Cicero’s rules. Thus
perhaps it tempts the reader to see Cicero’s own authority behind the
conclusion. Nevertheless, formally speaking, the reader would make a
mistake if he were to do so. Marcus is a skeptic who reports how the
question has seemed to him and in emphatic terms. But, as he immediately
reminds us in Div. ., his view is no more than that. He thinks we
would do well not to look for any authority in it.
To sum up, Cicero wishes his readers to be like students in an Academic
class. He wishes us to be able to judge freely without influence from
authority, where to judge is to see how things seem to us. But he is aware
that we can judge, and shows us how his characters use this freedom to
form views. In order to leave our judgment free he the author, by analogy
with an Academic teacher, withholds his authority from any argument or
conclusion. But in order to encourage us to think for ourselves, he gives us
models who do exactly that. In some cases we hear what other characters
think, like Velleius (DND .), Cotta (DND .–, .–), or Quin-
tus (Div. .). But sometimes it is his own character’s reactions of this
sort that we are shown. Since the author remains silent, showing these
reactions does not break Cicero’s rules.
Cicero’s hedged language in the last sentence of DND is designed to
assuage the worry of DND .. More obviously, Cicero contrasts Marcus’
criterion, what seems like the truth, with Velleius’, what seems true. This
keeps Marcus’ Academic credentials in order. A Clitomachean Academic
like Cicero does not take his beliefs to be true, but rather merely follows
what is truth-like. But less obviously, Marcus’ reaction in the last sentence
of DND is further qualified so as to recall DND .. Balbus’ speech
seemed to “tilt closer” to a likeness of the truth. That is, it seemed
propensior. This adjective is connected with the verb propendere, to hang
down, often used of the inclination of the arm of the scales in weighing.
. The Epistemology of Marcus’ Stance 
But in DND . Cicero used the same metaphor. For what is to be used
in debate is the momenta of reason, not of authority. A momentum is a
movement or an impulse, but the word is often used of the weight that
turns the scales. Thus when Cicero says at the end of the dialogue that
Balbus’ speech tipped the scales for Marcus more than did Cotta’s, he
shows us that Marcus assesses the speeches just as was prescribed in the
preface. We are to imagine that the weights that Marcus compares are
those of reason and not of authority.
Does Cicero risk the illusion that as an author he has given his authority
to the eventual views of his own characters? He does. For all that he has
warned us neither to look for his authority nor, if we do look, to find it in
his characters’ reactions, some readers might do either or both. Evidently
this is a risk that Cicero is willing to run. Now it would be silly to deny
that Cicero allows us to see some themes in his characters’ views, and even
in his own orchestration of the drama, across his many dialogues. On the
whole, the Epicureans are not taken as seriously as the Stoics. Cicero’s own
character, or a principal speaker like Laelius in On friendship, often seems
broadly sympathetic to Stoic or Antiochean positions, but with qualifica-
tions, exceptions, or points of uncertainty. One has the impression that
Cicero would be gratified if his readers’ thoughts tended to drift in the
same directions as his own. So it is plausible that he hoped, or part of him
hoped, that the readers of DND and Div. would broadly agree with
Marcus. But what is certain is that Cicero tells us that it would be a
misreading of the dialogues to agree with what Marcus happened to think
because Marcus thought it.
In what follows I consider what I call Marcus’ views as expressed in the
last sentence of DND and especially in Div. .– as quoted above
(pp. –). By this I mean only what seems most like the truth to him
following the discussion in either dialogue. We should see in these views a
model of how one can react skeptically to the dialogues, not a prescription of

For these meanings see OLD sv. propendere b and , propensus –, momentum b, , , TLL s.v.
propendere .–, propensus .–, momentum .–. A similar metaphor is
involved in the etymology of “pondering” in English.

In On ends we are left with the impression that over the various conversations Marcus tended to find
that both Stoic and Antiochean views are preferable to the Epicurean (and that they may or may not
come to the same thing), although in the Tusculans, Cicero chooses the proposita against which
Marcus argues so that a general picture of a Stoic-style approach to value and the emotions is built
up, but subjected to searching criticism. In On duties, Cicero prescribes a modified version of Stoic
ethics. In the remains of On fate, on the other hand, Marcus appears to argue against the Stoic view
of fate, while Academica is of course opposed both to Stoic and to Antiochean epistemology.
 Marcus’ Stance on the Central Question
what to think. Nevertheless, it is a coherent and distinctive perspective
on the Central Question.
. Marcus’ Stance at the End of Div.

I argued in Chapter  that Cicero’s project in DND and Div. was to
moderate religion by philosophical inquiry into the Central Question
about the nature of the gods: do they care for us? Religion I construed as
the set of performances prescribed by the traditional religious authorities at
Rome, like the pontifices. I said that such moderation was to impose two
modi, two limits or constraints, on religion. It had to do principally with
avoiding two sorts of false belief: that the gods care for us more than in fact
they do (superstition) or that they care less than in fact they do (impiety).
It could also involve reaching a view, however skeptical and qualified, on
whether and how they care for us. Religion would be moderate when
performed without the relevant false beliefs, and perhaps with carefully
formed theological views in their place. Here I show that Marcus takes a
stance on the outcome of this project at the end of Div..
At the end of DND Marcus broadly, but with some qualification, agrees
with Balbus. In Div. .– Marcus disagrees with the Stoic view of
divination. Again, there is no guarantee that things seem the same way to
Marcus on the day of Div. as they do to his character in DND. Yet it is
reasonable to suppose that when Marcus at the end of Div. rejects the Stoic
view of divination, but welcomes some other Stoic views about the religion
and the divine, we are shown one part of why he portrayed himself in
qualified agreement with Stoic theology as he did at the end of DND. The
agreement is qualified, at least in part, because Marcus does not agree with
Balbus about divination.
On what does Marcus agree with Balbus? I think this comes in what
I have identified as part (i) of my quotation from Div. .–
(pp. –). There Marcus says that his project, which he explicitly
links to the project of DND, is to uproot superstition while leaving religion
behind. He further says that the wise person will maintain the religious

Santangelo () – argues that, despite Div.’s structure of arguments on either side, and
despite Cicero’s avowed skepticism, Cicero intends the reader to see that Marcus’ arguments and
conclusions in Book  are the message of the work. I have already given my reasons to disagree.

Marcus does not mention impiety here because the specific threat in Div. is the Stoic view of
divination, which is unlikely to have underestimated the degree of divine care. If the Stoic view of
divination is wrong, then it is very likely superstitious, in that it would probably err in holding that
the gods care for us more than in fact they do.
. Marcus’ Stance at the End of Div. 
institutions of the ancestors. He says that the way to keep traditional
religion but also to uproot superstition is to join religion with a “grasp
of nature.” Balbus would agree with these key points.
So far so good – but I have not yet shown that (i) represents qualified
agreement with Balbus in particular. For Velleius would also accept the
key points I have just picked out in (i). He, too, wanted to uproot
superstition in favor of religion. In fact for him, impiety coincides with
superstition in the view that the gods are so unhappy (Velleius would say)
as to care about us. Not only that, but Marcus in (i) shares some terms
with Velleius. Velleius’ gods are also eternal and also have a “most
outstanding nature,” praestantissimam naturam (DND .). So whose
side is Marcus on?
The key to my reading of (i) is the sentence, “the beauty of the cosmos
and the order of heavenly things compels one to admit (cogit confiteri) that
there is some outstanding and eternal nature, a nature which humans
should admire and wonder at.” This sentence is supposed to be part of the
reason to preserve religion. Thus Marcus must mean that what we “should
admire and wonder at” is entitled to worship. Now Velleius could con-
cede, first, that the cosmos is beautiful (albeit not as beautiful as the gods,
DND .–). Second, he does concede there are some beings whose
eternal nature humans should admire and worship. But Velleius would not
concede that the former point “compels one to admit” the latter. On the
contrary, it is all-important to Velleius that the beauty of the cosmos
results only from the chance collision of atoms and not from any aspect
of the divine.
By contrast, we saw in Chapter  section .. that Balbus puts weight
on the movement of thought from the beauty of the cosmos to the
existence and to the active and caring nature of the gods. He says that
cosmic beauty and the order of the heavens help to give us our natural
preconception of the gods. Beauty and order also supply a key premise in
the second sort of argument Balbus gives for the thesis that “the cosmos is
governed by the gods.” When Balbus moves to his proof of the same thesis
from admiratione, “wondering” at the heavenly realm, this turns out to
mean contemplating its beauty (contemplari pulchritudinem, DND .,
.). Just so, in (i) it is the beauty and order of the cosmos that excites
wonder (admirandam). Now in Chapter  section .., I further argued
that beauty, for Balbus, is a feature god gives to the cosmos out of his
concern for us, his fellow rational beings. Cosmic beauty is in part
intended to help us to see that god cares about us. In sum, beauty was
key to Balbus’ experience of the world as a place made for us by a rational
 Marcus’ Stance on the Central Question
creator and thus to his affirmative answer to the Central Question. Marcus
agrees about that.
Now Marcus’ conclusion is circumscribed. He says only that “there is
some outstanding and eternal nature, a nature which humans should
admire and wonder at.” I should concede how much is not said explicitly.
For a start, Marcus does not use the word “god.” But I think divinity is
clearly implied. For the point of this conclusion is that we can have
religion without superstition. Thus this nature to admire, and to wonder
at, must be worthy of worship, that is to say, it must be a god. Both
Velleius and Balbus use “N/nature” to mean “the divine nature” and thus
god or the gods (e.g. DND ., .–). By calling this nature (a)
eternal and (b) outstanding, Marcus describes it so as to fit both the
Epicurean and the Stoic notions of the gods. (Cf. DND ., .–)
Thus we may securely infer that Marcus means that this eternal “nature”
is a god.
Very well, but what is this god like? Marcus does not say, for example,
that it is rational or that it cares about us. Perhaps he could mean by the
admiration due to nature something like a modern atheist’s attitude to the
natural world. An atheist today does not think the universe is rational, or
that it cares about us, but she might find it a source of awe and wonder.
Maybe in Cicero’s Rome such an atheist would have repurposed trad-
itional religion to express that attitude to uncaring, mindless, but magnifi-
cent nature. Thus I think that although Marcus finds at the end of Div.
that we should worship the god who made nature beautiful and orderly (he
is not an atheist), a careful reader could conclude that he does not think
that this god is rational, or that he does not think that it cares for us. But it
seems to me that the balance of probabilities is against either of these
conclusions. For Balbus there were two important consequences to the
beauty of the world. The first was that some rational creator made the
world. The second is that this creator made it for us, made it beautiful out
of concern for the sorts of creatures who can appreciate beauty. These are
important among the reasons for Balbus’ positive answer to the Central
Question. Now, the Marcus of Div. is discussing DND, in which his own
character found Balbus’ speech more congenial. He has also posed the
Central Question in the preface to DND. Thus – on balance – I think the
reader is led to the impression that Marcus here answers what he said was
the Central Question of the inquiry begun in DND, and does so by
pointing to the evidence offered by the speaker in DND with whom he
tended to agree. It seems to him that the Romans should worship this god
in the belief that it cares about us.
. Marcus’ Stance at the End of Div. 
Consider Marcus’ sequence of thought in the latter half of (i) and the
beginning of (ii). He insists that when superstition is uprooted, religion
need not be. He then gives us two points by way of proof: first, that the
wise man keeps the institutions of his ancestors and, second, that there is
a divine nature. Then he says that the religion to be propagated is the
kind that is joined with a grasp of nature. That is to say, for Romans,
right religion is the kind that joins ancestral institutions with suitable
reflections on natural science. This reiterates the project of DND and
Div. for which I argued in Chapter . But it does so in a way more
reminiscent of Balbus’ attitude to religion than of his opponent Cotta’s.
As we saw in Chapter , Balbus’ wise person is the one who defends
ancestral traditions by joining them with a grasp of nature. But Cotta
wanted to defend ancestral traditions by rejecting the threat of just such
Stoic reconsideration. Thus in this way, too, Marcus in Div. finds himself
in agreement with Balbus.
What about (ii)? Notice again that Marcus’ sequence of thought follows
the project of DND and Div.. He adverts to a series of traditional religious
performances: sacrifice (during which entrails would be examined), the
augural inspection of a bird sign, consulting the haruspex, apparent prodi-
gies that might have to be reported. But again he comments on what effect
beliefs will have on these performances. If a Roman believes that the
performances yield real divination, then these ever-present performances
become superstitious and persecute him. But if he does not believe in
divination, he can face them untroubled. The danger is that the dogmatic
philosophers who seem the best, whose general theology Marcus himself
has found plausible, argue for precisely the view that persecutes. But
luckily Carneades and the Academy can provide the remedy. Thus a
Roman who has read Div. should not believe that the Stoic view is true,
even if she finds it probable on balance. She should suspend judgment. As
for Marcus, he does not include the Stoic view of divination in his general
acceptance of the Stoic reinterpretation of religion. Rather, the Stoic
interpretation of divinatory performances, according to which the gods
really do send us news in all these myriad ways, is to be uprooted as
superstition. For by the Stoics’ own plausible rubric of re-interpreting
religion in the light of natural science (which I explored in Chapter 
section .), the Stoics’ own theory of divination (Chapter ) has seemed
to Marcus implausible. This is why Marcus emphasizes the dangerous
virtues of Stoic philosophy in (ii): what appears to be the well-earned
“authority” of those whose general theology he finds persuasive might have
led him into superstition about divinatory practices.
 Marcus’ Stance on the Central Question
It is enlightening to read the first sentence of DND alongside the last
sentence of (i) and the first sentence of (ii). In the first sentence of DND,
Cicero advertised that the inquiry into the nature of the gods is not only
“necessary for the moderation of religion” but is also “most beautiful for
the mind to grasp (ad cognitionem animi pulcherrimam).” Thus in (ii)
Marcus says “For that reason, just as religion which is joined to some grasp
of nature (quae est iuncta cum cognitione naturae) should be cultivated, in
the same way every shoot of superstition should be uprooted.” It is
attention to the beauty of the cosmos that compels us to a grasp of the
eternal nature behind it. That is why, properly done, the inquiry into
the nature of the gods for the moderation of religion is most beautiful for
the mind to grasp. It is the grasp of this eternal nature that makes religious
performances into pious actions. That is how religion is moderated.
Thus my interpretation of the view with which Marcus reports at the
end of Div. is as follows. The answer to the Central Question that at that
moment seems plausible to him is, “yes.” The beauty and order of the

As an Academic skeptic, Cicero is not convinced of the possibility of a cognitio, a successful mental
grasp in the technical sense used by his Stoics or Antiocheans. (Academica .) So he must mean
cognitio in some weaker sense here, such as a view that results from careful weighing of the different
arguments.

Since it was my task in this book to interpret DND and Div., I here give the stance that Cicero
constructs for “himself,” for Marcus, and for his narrator’s voice, in the dialogues. This stance was
not necessarily Cicero’s own. Thus it is natural to ask, how might Marcus’ position relate to the
historical Cicero’s religious or theological sentiments? In texts collected by Goar (), we find
two patterns of religious remarks in Cicero’s corpus beyond the philosophica. First, in his speeches,
Cicero refers often and extensively to the traditional religion. Some speeches, like De domo sua or De
haruspicum responsis, are chiefly concerned with religion. There is religious material in speeches on
other matters, too. (For some examples, see p.  n. .) But given their rhetorical context, these
references are not much help in understanding Cicero’s private views. Second, in his private
correspondence, Cicero makes references to the gods and to private religious performances like
prayer. But these references are infrequent and lukewarm, surprisingly so to a reader familiar with
the long attention to theology and religion in Cicero’s philosophical writings. Further, many of
these references can be interpreted as merely conventional turns of speech (see Goar 
pp. –). Most notable are two remarks that seem to indicate that Cicero’s wife Terentia
paid cult to the gods over private matters like his day-to-day health, and that Cicero did not do so.
The tone of these remarks may even be to mock Terentia’s pieties. (Letters to his friends ..=SB
and ..=SB) I agree with Goar (p. ) that, “religious ideas did not penetrate [Cicero’s]
private life very deeply, or have any great influence on it,” with perhaps two exceptions. One clear
exception was his unconventionally religious reaction to Tullia’s death, for which see Goar p. ,
Treggiari () –. A second possible exception is the praise that Marcus gives to the
Eleusinian mysteries as what we might call a “life experience” at Laws .. Clinton () ,
, – is cautious but concludes that Cicero was indeed an initiate. He writes that,
“Cicero departed from the Mysteries cum spe meliore moriendi . . . Cicero . . ., in my opinion,
show[s] signs of being interested in the Mysteries as a ‘religious’ act.” By contrast, the letters show
that philosophy much more frequently played a part in Cicero’s friendships and inner life (see
McConnell ). Yet perhaps my reading of DND and Div. suggests a possible way (no more) to
find Marcus’ estimation of religion and the gods in the philosophica, and Cicero’s omission of
. Marcus’ Stance at the End of Div. 
cosmos leads us to a “grasp” of nature, that it is eternal and admirable and
cares for us, a god worth worshipping. Further, Roman religion can be re-
interpreted as a set of performances expressing just such a view, as Balbus
suggested. But one aspect of traditional religion does not do what Stoic
interpretation purports: divination does not predict outcomes. Believing
the Stoics on that point would make one superstitious when one sacrificed,
or looked as an augur at a bird sign, and so on. Perhaps a Roman is to
think that divinatory performance is a good way to dramatize attention to
the wishes of the gods. But in order to avoid superstition, and thus to
complete the moderation of his religion, a Roman must see that through
these performances no divinatory signs or sayings will arrive.
As I have argued, by closing DND and Div. with a sketch of this view,
Cicero does not want to induce his readers to agree with Marcus. Marcus’
view of the moment is a model of how views by which to live can be
reached, not a view to follow on Cicero’s authority. Cicero invites his
fellow Romans to form their own, careful, and moderate views – as he will
continue to form his own views and to withhold his own assent. He
religion in the letters, less discrepant. I have suggested that Marcus is attracted to Balbus’ view that
Roman religion could be a way publicly to express philosophical admiration for a wise creator, but
that he is not impressed by the Stoic advocacy of continual divine interaction with us through
divination. Now Cicero does not express his philosophical, ethical, and political views in religious
terms in private letters, even though the political activities and personal friendships that these views
motivate could certainly have been justified in Balbus’ theological terms. But why should he? Balbus
could agree that that was not religion’s job. In the letters Cicero expresses these views in Latin, in
jokes, rhetoric, argument, and so on, the media that were indeed closer to his heart. But perhaps
Cicero thought he could express the same ideas through public religious performances, a medium of
gestures in some ways comprehensible to the public. Marcus’ theological reaction in DND was that
those performances could earnestly be done, however skeptically and tentatively, for a provident
divinity. Cicero’s discomfort with Terentia’s rituals need not have been that he found cult paid to
caring gods to be risible. Rather, his objection, if he had an objection, could have been theological.
For example, perhaps it seemed to him that a just god would not give advantages, like curing a bout
of vomiting, to individuals in return for prayers or sacrifices. If so, then it would have seemed to him
that Terentia’s religious actions were superstitious, if she believed (for example) that she owed a
pinch of incense to some god who had cured Cicero.

It is striking how consonant this part of Marcus’ view is with the one expressed by Marcus in Laws
.–, , , and .–.

In contrast to my last note, Marcus’ view of divination at the end of Div. cannot be reconciled with
the one expressed at Laws .–. One can come closer to reconciliation than at first appears. In
Div. . Marcus says that superstition spread through all peoples, implying there was a time when
this was not so. He does not say how things were before the spread of superstition. Similarly,
Marcus in the Laws implies that divination used to work properly, but no longer, because the
disciplina and ars have been lost (.). But Marcus’ endorsement in the Laws of the thesis there is
divination, and of a shortened version of the Stoic syllogism from Div. (Div. .–, .–
cf. Laws .), remain irreconcilable with the end of Div..

For another interpretation of the last sentence of DND in the light of a unified purpose for DND
and Div., see Lévy () –.
 Marcus’ Stance on the Central Question
expects that if they have followed the inquiry he has staged attentively,
they too will withhold the assent they would give to a proven truth from all
the claims about the gods they have read. For they have also read good
reasons why all the claims might be false. In this way, they avoid the
decisive sort of impiety and superstition that would come from acting on
rash and dogmatic but false beliefs about the gods. Nevertheless, on the
model Cicero shows them, readers can see how the question seems to each
of them for the moment.
Perhaps it seems to Cicero that the content of carefully weighed views of
the gods will tend to lead closer to piety. But I suggest that even if it does
not, then he hopes that it will at least help each Roman to find theological
meaning in the traditional religion. Either way, now that he has helped
them critically to liberate the philosophical riches of the Greeks, Cicero
hopes to be seen “to benefit our countrymen.” (nostris profuturi, Div.
.) Like Varro in the Academica, he has helped to lead them home,
when they were strangers in their own city (see pp. –). But in so doing
in well-written dialogues in his classical revival style, he has brought
philosophy itself to life. As he would say, he has illuminated it.
 
Terminology in DND and Div. for Religious Virtues

and Vices, and Greek Equivalents
Part : Stoic Religious Terminology in Greek
σοφία, wisdom
Sextus Empiricus, Adversus mathematicos . (= SVF .) σοφία,
ἐπιστήμη οὖσα θείων τε καὶ ἀνθρωπείων πραγμάτων, “wisdom, being
knowledge of matters human and divine.”
ὁσιότης, piety
Sextus Empiricus, Adversus mathematicos . (= SVF .) ἡ ὁσιότης,
δικαιοσύνη τις οὖσα πρὸς θεούς . . ., “piety, being a certain justice towards
the gods.”
Stobaeus, .. (= SVF .) τὴν γὰρ ὁσιότητα ὑπογράφεσθαι
δικαιοσύνην πρὸς θεούς. “For piety is justice towards the gods.”
[Andronicus], On Passions . (=SVF..) ὁσιότης δὲ ἐπιστήμη
παρεχομένη πιστοὺς καὶ τηροῦντας τὰ πρὸς τὸ θεῖον δίκαια. “But piety is
a science that produces people who are faithful and observant about what
is just towards the divine.”
εὐσέβεια, holiness
Sextus Empiricus, Adversus mathematicos . (= SVF .) ἔστι γὰρ
εὐσέβεια ἐπιστήμη θεῶν θεραπείας. “for holiness is science of the cult of
the gods.”
DL . (= SVF .) εἶναί τε τὴν εὐσέβειαν ἐπιστήμην θεῶν
θεραπείας, “holiness is science of the cult of the gods.”
[Andronicus], On Passions . (= SVF .) εὐσέβεια δὲ ἐπιστήμη
θεῶν θεραπείας.. “holiness is science of the cult of the gods.”

 Terminology in DND and Div.
Stobaeus, .. (= SVF .) εὐσέβειαν δὲ ἐπιστήμην θεῶν
θεραπείας “holiness is science of the cult of the gods.”
δεισιδαιμονία, fear of the gods

Stobaeus, . (= SVF .) δεισιδαιμονία δὲ φόβος θεῶν ἢ δαιμόνων,
“deisidaimonia [lit. fearing daemons] is fear of gods or daemons.”
Andronicus, On Passions  (=SVF .) δεισιδαιμονία δὲ φόβος τοῦ
δαιμονίου. [ἢ ὑπερέκπτωσις τῆς πρὸς θεοὺς τιμῆς. Von Arnim deletes this
phrase.] “deisidaimonia is fear of the daemonic, [or an excess of honor to
the gods].”
Clement, Stromateis . (= SVF .) ἡ γοῦν δεισιδαιμονία πάθος,
φόβος δαιμόνων οὖσα “deisidaimonia is an emotion, being fear of the
daemons.” (NB Clement’s Christian commitments lead him to call the
gods “daemons.”)
Stobaeus, .. (= SVF .) ὑπὸ δὲ τὸν φόβον . . . δεισιδαιμονίαι,
“under the genus fear . . . are deisidaimoniai.”
Part : Religious Terminology in DND
Religio and superstitio

Balbus:
et si conferre volumus nostra cum externis, ceteris rebus aut pares aut etiam
inferiores reperiemur, religione, id est cultu deorum, multo superiores. (DND .)
And if we wish to compare our affairs with those of outsiders, we shall find
that we are equal or inferior in other matters, but much superior in religion,
that is, in the cult of the gods.
videtisne igitur ut a physicis rebus bene atque utiliter inventis tracta ratio sit ad
commenticios et fictos deos. Quae res genuit falsas opiniones erroresque turbu-
lentos et superstitiones paene aniles. (DND .)
You see how, from physical facts whose discovery was good and useful, was
wrenched a plan for confected and fictional gods. This state of affairs gave
rise to false opinions, distressing errors, and superstitions almost like those
of an old woman.
cf. Cotta versus Velleius:
horum enim sententiae omnium non modo superstitionem tollunt, in qua inest
timor inanis deorum, sed etiam religionem, quae deorum cultu pio continetur.
(DND .)
Part : Religious Terminology in DND 
For the views of all these not only take away superstition, in which we find
empty fear of the gods, but also religion, which is comprised by pious cult
of the gods.
Pietas and Sanctitas (as defined by Cotta speaking against Velleius

DND .).
est enim pietas iustitia adversum deos
For piety is justice towards the gods.
sanctitas autem est scientia colendorum deorum
But holiness is knowledge of the cult of the gods.
Cf. Balbus: cultus autem deorum est optumus idemque castissimus* atque
sanctissimus plenissimusque pietatis . . . (DND .)
But the cult of the gods is best and least polluted and again most holy and
most full of piety . . .
(*castus=ἁγνός? DL says that Stoic sages are ἁγνοί, ..)
Summary: Cicero’s General Philosophical Terminology about

Religion in DND
Cicero’s Stoic
Latin Greek Meaning I translate
Virtues:
pietas ὁσιότης justice towards the gods piety
sanctitas εὐσέβεια science of the cult of the gods holiness
Vices:
superstitio – excessive religiosity, gives rise to superstition
δεισιδαιμονία, excessive fear of the gods.
Activities:
cultus θεῶν cult of the gods cult (of the
deorum θεραπεία gods)
religio – cult of the gods or sometimes the cult of the religion
gods done virtuously.
 
Velleius’ Strategies against his Opponents
Opponent Comments Strategy
. Thales The mind which made everything previously lacked S, F

sense, body.
Anaximander Gods not everlasting. E
. Anaximenes Formless air which began cannot be a god, must be F, E
mortal.
Anaxagoras The infinite has no sensation. Mind must have a body. S, F
. Alcmaeon Heavenly bodies are mortal. E
Pythagoras The portion of god in man is unhappy. H
. Xenophanes Mind must have a body, the infinite has no sensation. F, S
Parmenides Ring of lights in sky wrong form, no sense; others not F, S, E
eternal.
. Empedocles Elements not eternal, so sensation. E, S
Protagoras Not clear whether the gods exist, thus has no view of their *
nature.
Democritus Contradictory ideas; denies immutability, thus eternity. *, E
Diogenes Ap. Air no sense, wrong form. S, E
. Plato Inconsistent; holds god is incorporeal. *, F
. Xenophon Inconsistent. *
Antisthenes Incoherent. *
Speusippus Does away with our conception of god. N
. Aristotle Inconsistent; heavenly rotations not eternal and attack; *, E, S,
incorporeal has no sense of happiness. H
. Xenocrates Heavenly bodies have no sense of happiness. H
Heraclides Inconsistent; makes gods senseless and changeable in *, S, H
form.
. Theophrastus Inconsistent. *
Strato Nature has no sense or form. S, F
. Zeno Citium Law has no soul; aether has no sense, preconception. A, S, F
. Aristo Form of the deity inconceivable, no sense, god not F, S, A
ensouled.
Cleanthes Violates preconception on many counts. N
. Persaeus Dead discoverers of benefits we mourn and do not ?E, N
worship
. Chrysippus So against preconception as to be impossible to imagine. N
. Diogenes Bab. Ditto N
. Poets Make the gods vicious. H
*= Velleius alleges incoherence. For key to letters, see pp. –.

 
Balbus’ Classification of the Gods
Balbus’ qualified
What are Origin in human endorsement of
Level Type they like? Examples culture? divinity.
. Paradigmatic. . (–) Ensouled, better The cosmos. – –

than which
nothing
conceivable.
. (–) Most beautiful in Sun, moon, the – –
shape, helping other five
the cosmos, planets, the fixed
exhibiting stars.
mind and
reason.
. Cultural. . (–) “Natures of gods” Fides, Mens, The wisest of the “They were
yielding great Honos, Ops, Greeks and the established
benefits to Salus, Roman ancestors and named…
humankind. Concordia, named them for not without
Libertas, the benefits they cause.” ()
Victoria, bestow.
Cupido, Venus
Lubentina,
Voluptas.
. () The best and Hercules, Castor From human life and “They are rightly
immortal and Pollux, custom, because held as gods,
men, who Aesculapius, they excelled in since they are
gave benefits. Liber son of benefits. eternal and
Semele, ? the best.”
Romulus = ()
Quirinus.
. (–) Physical entities. Apollo, Ceres, “From physical “You see how,
Diana, Dis, theory,” but then from physical
Janus, Jupiter, “dressed in human facts whose
Juno, Mars, appearance, have discovery was
Minerva, supplied the poets good and
Neptune, the with myths,” and useful, was
Penates, “have crammed wrenched a
Proserpina, human life with plan for
Saturn, Venus, every confected and
Vesta. superstition.” () fictional
gods.” ()
Cf. pp. – above. Numbers in parentheses refer to sections of Book  of DND.

Bibliography
Algra, K. () ‘Stoic Theology’, in The Cambridge Companion to the Stoics, ed.
B. Inwood. Cambridge: –.
() Conceptions and Images: Hellenistic Philosophical Theology and Trad-
itional Religion. Amsterdam.
() ‘Stoic Philosophical Theology and Greco-Roman Religion’, in Salles
(ed.) (a): –.
Algra, K., Barnes, J., Mansfeld, J., and Schofield, M. (eds.) () The Cambridge
History of Hellenistic Philosophy. Cambridge.
Allen, J. () ‘Academic Probabilism and Stoic Epistemology’, CQ .: –.
() Inference from Signs: Ancient Debates about the Nature of Evidence.
Oxford.
() ‘The Stoics on the Origin of Language and the Foundations of Etymol-
ogy’, in Inwood and Frede: –.
() ‘Greek Philosophy and Signs’, in Annus: –.
Ando, C. () The Matter of the Gods: Religion and the Roman Empire. The
Transformation of the Classical Heritage . Berkeley.
() Review of Scheid () in JRS : –.
Annas, J. () ‘Plato’s Laws and Cicero’s de Legibus’, in Schofield (ed.):
–.
Annas, J. and Barnes, J. () The Modes of Scepticism: Ancient Texts and Modern
Interpretations. Cambridge.
() Sextus Empiricus: Outlines of Scepticism. Cambridge.
Annas, J. and Betegh, G. (eds.) () Cicero’s De Finibus: Philosophical
Approaches. Cambridge.
Annus, A. (ed.) () Divination and Interpretation of Signs in the Ancient World.
Oriental Institute Seminars . Chicago.
Atkins, J. () Cicero on Politics and the Limits of Reason: The Republic and the
Laws. Cambridge.
Auvray-Assayas, C. () ‘L’ordre du deuxième livre du De Natura Deorum de
Cicéro’, Revue d’Histoire des Textes : –.
() ‘Existence et providence des Dieux dans la théologie stoïcienne: remar-
ques sur l’ordre de l’exposé du De natura deorum (livre ) d’après la
traduction manuscrite’, Études Philosophiques .: –.

Bibliography 
Ax, W. () M. Tullius Cicero: De Natura Deorum. Scripta Quae Manserunt
Omnia . Stuttgart.
Baraz, Y. () A Written Republic: Cicero’s Philosophical Politics. Princeton.
Barnes, J. () ‘Medicine, Experience and Logic’, in Science and Speculation:
Studies in Hellenistic Theory and Practice, ed. J. Barnes, J. Brunschwig, M.
Burnyeat, and M. Schofield. Cambridge: –.
Barnes, J. (ed.) () The Complete Works of Aristotle: Revised Oxford Translation.
 vols. Princeton.
Beard, M. () ‘Cicero and Divination: The Formation of a Latin Discourse’,
JRS : –.
() ‘Priesthood in the Roman Republic’, in Pagan Priests: Religion and
Power in the Ancient World, ed. M. Beard and J. North. London.
Beard, M., North, J., and Price, S. () Religions of Rome.  vols. Cambridge.
Bénatouïl, T. () ‘How Industrious Can Zeus Be? The Extent and Objects of
Divine Activity in Stoicism’, in Salles (ed.), –.
Bett, R. () The Cambridge Companion to Scepticism. Cambridge.
() ‘Beauty and its Relation to Goodness in Stoicism’, in Ancient Models of
Mind: Studies in Human and Divine Rationality, ed. A. Nightingale and D.
Sedley. Cambridge: –.
() Sextus Empiricus: Against the Physicists. Cambridge.
Blank, D. () ‘Varro and Antiochus’, in Sedley (ed.): –.
Bobzien, S. () Determinism and Freedom in Stoic Philosophy. Oxford.
() ‘Did Epicurus Discover the Free-Will Problem?’, OSAP : –.
Boyancé, P. () ‘Les méthodes de l’histoire littéraire: Cicéron et son oeuvre
philosophique’, Revue des Études Latines .: –.
() Études sur l’humanisme cicéronien. Collection Latomus . Brussels.
Boys-Stones, G. () ‘Eros in Government: Zeno and the Virtuous City’, CQ
.: –.
(a) ‘The Stoics’ Two Types of Allegory’, in Metaphor, Allegory and the
Classical Tradition, ed. G. Boys-Stones. Oxford: –.
(b) Post-Hellenistic Philosophy: A Study of its Development from the Stoics to
Origen. Oxford.
() ‘Ancient Philosophy of Religion: An Introduction’, in Oppy and
Trakakis (eds.): –.
() ‘Antiochus’ Metaphysics’, in Sedley (ed.): –
Brennan, T. () ‘The Old Stoic Theory of the Emotions’, in J. Sihvola and
T. Engberg-Pedersen (eds.) The Emotions in Hellenistic Philosophy, The New
Synthese Historical Library (Texts and Studies in the History of Philosophy)
. Dordrecht: –.
() ‘Stoic Moral Psychology’, in Inwood (ed.): –.
() The Stoic Life. Oxford.
() ‘The Stoics’, in Oppy and Trakakis (eds.): –.
Briggs, W. and Calder, W. (eds.) () Classical Scholarship: A Biographical
Encyclopedia. New York.
 Bibliography
Bringmann, K. () Untersuchungen zum späten Cicero. Hypomnemata .
Göttingen.
Brittain, C. () Philo of Larissa. Oxford.
() ‘Common Sense: Concepts, Definition and Meaning In and Out of the
Stoa’, in Inwood and Frede (eds.): –.
() Cicero: On Academic Scepticism. Indianapolis.
() ‘Posidonius’ Theory of Predictive Dreams’, OSAP : –.
() ‘Cicero’s Sceptical Methods: The Example of the De Finibus’, in Annas
and Betegh (eds.): –.
Brouwer, R. () The Stoic Sage: The Early Stoics on Wisdom, Sagehood and
Socrates. Cambridge.
Brunt, P. () ‘Philosophy and Religion in the Late Republic’, in Philosophia
Togata: Essays on Philosophy and Roman Society, ed. M. Griffin and J. Barnes.
Oxford: –.
Burnyeat, M. () ‘Gods and Heaps’, in Language and Logos: Studies in Ancient
Greek Philosophy presented to G. E. L. Owen, ed. M. Schofield and
M. Nussbaum. Cambridge: –.
Clark, A. () The Descent of Manuscripts. Oxford.
Clinton, K. () ‘The Eleusinian Mysteries: Roman Initiates and Benefactors,
Second Century B.C. to A.D. ’, ANRW ..: –.
Cooper, J. () ‘Chrysippus on Physical Elements’, in Salles (ed.), –.
Couissin, P. () ‘Les sorites de Carnéade contre le polythéisme’, Revue des
Études Grecques : –.
Davies, J. () M. Tulli Ciceronis De Natura Deorum Libri Tres nd ed.
Cambridge.
DeFilippo, J. () ‘Cicero vs. Cotta in De natura deorum’, Ancient Philosophy
.: –.
Degraff, T. () ‘Plato in Cicero’, CPh .: –.
Denyer, N. () ‘The Case Against Divination: An Examination of Cicero’s De
Divinatione’, PCPS  n. s. : –.
Diels, H. () Doxographi Graeci. Berlin.
Dillon, J. and Long, A. (eds.) () The Question of Eclecticism. Berkeley.
Dougan, T. () M. Tulli Ciceronis Tusculanarum Disputationum Libri
Quinque: Volume I: Books I and II. Cambridge.
Douglas, A. () ‘Platonis Aemvlvs?’, Greece & Rome n.s. .: –.
() ‘Cicero the Philosopher’, in Cicero, ed. T. Dorey. London: –.
() Cicero: Brutus. Oxford.
() Cicero: Tusculan Disputations I. Warminster.
() Cicero: Tusculan Disputations II & V. Warminster.
() ‘Form and Content in the Tusculan Disputations’, in Powell (ed.):
–.
Dragona-Monachou, M. () ‘Posidonius’ “Hierarchy” Between God, Fate and
Nature, and Cicero’s De divinatione’, Filosofia : –.
() The Stoic Arguments for the Existence and Providence of the Gods. Athens.
Bibliography 
Drummond, W., and Walpole, R. () Herculanensia; or Archaeological and
Philological Dissertations, containing a Manuscript Found among the Ruins of
Herculaneum. London.
Durand, R. () ‘La date du “De Divinatione”’, in Mélanges Boissier: recueil de
mémoires concernant la littérature et les antiquités romaines dédié à Gaston
Boissier: à l’occasion de son e anniversaire. Paris: –.
Dyck, A. () Cicero: De Officiis. Ann Arbor.
() Cicero: De Natura Deorum: Book I. Cambridge.
Edelstein, L and Kidd, I. G. (eds.) (–) Posidonius.  vols. in .
Cambridge.
Erler, M. () ‘Epicurus as deus mortalis: homoiosis theoi and Epicurean Self-
cultivation’, in Frede and Laks (eds.): –.
Essler, H. () ‘Cicero’s Use and Abuse of Epicurean Theology’, in Fish and
Sanders (eds.): –.
Falconer, W. () ‘A Review of M. Durand’s La date du De Divinatione’, CP
.: –.
Feeney, D. () Literature and Religion at Rome: Cultures, Contexts and Beliefs.
Cambridge.
Festugière, A. () Épicure et ses dieux, nd edn. Paris.
Fish, J. and Sanders, K. (eds.) () Epicurus and the Epicurean Tradition.
Cambridge.
Fortenbaugh, W. and Steinmetz, P. (eds.) () Cicero’s Knowledge of the
Peripatos. Rutgers University Studies in Classical Humanities . New Bruns-
wick, NJ.
Fortenbaugh, W. and Pender, E. (eds.) () Heraclides of Pontus: Discussion.
Rutgers University Studies in Classical Humanities . New Brunswick, NJ.
Fox, M. () ‘Heraclides of Pontus and the Philosophical Dialogue’, in For-
tenbaugh and Pender (eds.): –.
Frede, D. and Laks, A. (eds.) () Traditions of Theology: Studies in Hellenistic
Theology, its Background and Aftermath. Leiden.
Frede, M. () ‘The Stoic Doctrine of the Affections of the Soul’, in
M. Schofield, and G. Striker, (eds.) The Norms of Nature. Cambridge: –.
() ‘Chaeremon der Stoiker’, ANRW II..: –.
() ‘The Stoic Conception of Reason’, in Hellenistic Philosophy vol. , ed. K.
Boudouris. Athens: –.
Furley, D. () Two Studies in the Greek Atomists. Princeton.
() ‘Aristotelian Material in Cicero’s De natura deorum’, in Fortenbaugh
and Steinmetz (eds): –.
Gercke, A. () ‘Chrysippea’, Jahrbücher für Classische Philologie suppl. :
–.
Gifford, E. () Eusebius: Preparation for the Gospel. Oxford.
Gildenhard, I. () Paideia Romana: Cicero’s Tusculan Disputations, Proceedings
of the Cambridge Philological Society supplemental volume . Cambridge.
() Creative Eloquence: The Construction Reality in Cicero’s Speeches. Oxford.
 Bibliography
() ‘Of Cicero’s Plato: Fictions, Forms, Foundations’, in Schofield (ed.):
–.
Giomini, R. () M. Tullius Cicero: De Divinatione, de Fato, Timaeus. Scripta
Quae Manserunt Omnia . Leipzig.
() Problemi cronologici e compositivi del “De divinatione” Ciceroniano. Rome.
Glare, P. (ed.) () Oxford Latin Dictionary. Oxford.
Glucker, J. () Antiochus and the Late Academy. Hypomnemata . Göttingen.
() ‘Chapter and Verse in Cicero’, Grazer Beiträge : –.
() ‘Cicero’s Philosophical Affiliations’, in Dillon and Long (eds.): –.
() ‘Cicero’s Philosophical Affiliations Again’, Liverpool Classical Monthly
: –.
() ‘Probabile, veri simile and Related Terms’, in Powell (ed.): –.
() ‘A Platonic Cento in Cicero’, Phronesis .: –.
() ‘The Philonian/Metrodorians: Problems of Method in Ancient Philoso-
phy’, Elenchos : –.
Goar, R. J. () Cicero and the State Religion. Amsterdam.
Gomperz, T. () Philodem: Über Frömmigkeit. Leipzig.
Görler, W. () ‘Silencing the Troublemaker: De Legibus . and the Con-
tinuity of Cicero’s Scepticism’, in Powell (ed.): –.
() ‘Cicero’s Philosophical Stance in the Lucullus,’ in Assent and Argument:
Studies in Cicero’s Academic Books: Proceedings of the th Symposium Helle-
nisticum, ed. B. Inwood and J. Mansfeld: –.
Gottschalk, H. () Heraclides of Pontus. Oxford.
() ‘Aristotelian Philosophers in the Roman World’, ANRW ..:
–.
Goulet, R. (ed.) (–) Dictionnaire des philosophes antiques. Paris.
Graver, M. () Stoicism and Emotion. Chicago.
() ‘Cicero’, in Oppy and Trakakis (eds.): –.
() ‘Honor and the Honorable: Cato’s Discourse in De Finibus ’, in Annas
and Betegh (eds.): –.
Griffin, M. () ‘Philosophical Badinage in Cicero’s Letters to his Friends’, in
Powell (ed.): –.
() ‘From Aristotle to Atticus: Cicero and Matius on Friendship’, in Barnes
and Griffin (eds.): –.
Griffin, M. and Atkins, E. () Cicero: On Duties. Cambridge.
Guillaumont, F. () Le De divinatione de Cicéron et les théories antiques de la
divination. Collection Latomus . Brussels.
Hammerstaedt, J. () ‘Das Kriterium der Prolepsis beim Epikureer Diogen-
ian’, Jahrbuch für Antike und Christentum : –.
Hankinson, R. () ‘Stoicism, Science and Divination’, in Method, Medicine
and Metaphysics: Studies in the Philosophy of Ancient Science, ed. R. Hankin-
son. Apeiron .. Edmonton: –.
Helmreich, Georgius (ed.) () Claudii Galeni Pergameni Scripta Minora vol. .
Leipzig.
Bibliography 
Hirzel, R. (–) Untersuchungen zu Ciceros Philosophischen Schriften. 
vols. Leipzig.
() Der Dialog: Ein literarhistorischer Versuch vol. . Leipzig.
Ierodiakonou, K. (ed.) () Topics in Stoic Philosophy. Oxford.
Inwood, B. () Ethics and Human Action in Early Stoicism. Oxford.
(ed.) () The Cambridge Companion to the Stoics. Cambridge.
Inwood, B. and Frede, D. (eds.) () Language and Learning: Philosophy of
Language in the Hellenistic Age. Cambridge.
Inwood, B. and Gerson, L. () The Stoics Reader: Selected Writings and
Testimonies. Indianapolis.
Jacobs, J. () ‘Traces of the Omen Series Šumma Izbu in Cicero, De divina-
tione’, in Annus (ed.): –.
Jones, A. () ‘The Stoics and the Astronomical Sciences’, in Inwood (ed.):
–.
Keaney, J. and Lamberton, R. () [Plutarch]: Essay on the Life and Poetry of
Homer. APA American Classical Studies . Atlanta.
Kearns, E. () The Heroes of Attica. BICS Supplement . London.
Kenney, E. () ‘Books and Readers in the Roman world’, in The Cambridge
History of Classical Literature vol. , ed. E. Kenney and W. Clausen (eds.).
Cambridge: –.
Konstan, D. () ‘Epicurus on the Gods’, in Fish and Sanders (eds.): –.
() ‘Epicurus’, in E. Zalta, (ed.), The Stanford Encyclopedia of Philosophy.
Fall  Edition. https://plato.stanford.edu/archives/fall/entries/
epicurus/
Krostenko, B. () ‘Beyond (Dis)belief: Rhetorical Form and Religious Symbol
in Cicero’s de Divinatione’, TAPA : –.
Kühner, R. () M. Tulli Ciceronis Tusculanarum disputationum libri quinque,
th edn. Hanover.
Lapidge, M. () ‘Arkhai and stoikheia: A Problem in Stoic Cosmology’,
Phronesis .: –.
Lemke, D. () Die Theologie Epikurs: Versuch einer Rekonstruktion. Munich.
Leonhardt, J. () Ciceros Kritik der Philosophenschulen. Zetemata .
Munich.
Levine, P. () ‘The Original Design and Publication of the De Natura
Deorum’, HSCP : –.
Lévy, C. () Cicero Academicus: Recherches sur les Académiques et sur la
philosophie Cicéronienne. Collection de l’École française de Rome .
Rome.
() ‘De Chrysippe à Posidonius: Variations Stoïciennes sur le thème de la
divination’, in Oracles et prophéties dans l’antiquité, ed. J. Heintz. Travaux du
Centre de Recherche sur le Proche-Orient et la Grèce Antique, Université
des Sciences Humaines de Strasbourg . Paris: –.
() ‘The Sceptical Academy: Decline and Afterlife’, in Bett (ed.): –.
Lewis, C. and Short, C. (eds.) () A Latin Dictionary. Oxford.
 Bibliography
Linderski, J. () ‘The Aedileship of Favonius, Curio the Younger and Cicero’s
Election to the Augurate,’ HSCP : –.
() ‘Cicero and Roman Divination’, PP : –.
Long, A. () ‘Cicero’s Plato and Aristotle’, in Powell (ed.): –.
() ‘Stoic Linguistics, Plato’s Cratylus, and Augustine De dialectica’, in
Inwood and Frede (eds.): –.
Long, A. and Sedley, D. () The Hellenistic Philosophers. Cambridge.
Mackie, J. () ‘Evil and Omnipotence’, Mind .: –.
Malcovati, H. () Oratorum Romanorum Fragmenta Liberae Rei Publicae, rd
edn.  vols. Turin.
Manuwald, A. () Die Prolepsislehre Epicurs. Bonn.
Mansfeld, J. () ‘Providence and the Destruction of the Universe in Early
Stoic Thought: With Some Remarks on the “Mysteries of Philosophy”’, in
Studies in Hellenistic Religions, ed. M. Vermaseren. Leiden: –.
Mansfeld, J. and Runia, D. (–) Aëtiana: The Method and Intellectual
Context of a Doxographer.  vols. Philosophia Antiqua , , and .
Leiden.
Mayor, J. B. (–) M. Tulli Ciceronis De Natura Deorum Libri Tres. 
vols. Cambridge.
McConnell, S. () Philosophical Life in Cicero’s Letters. Cambridge.
McKirahan, R. () ‘Epicurean Doxography in Cicero, De natura deorum,
Book ’, in Epicureismo greco e romano vol. , ed. G. Giannantoni and M.
Gigante. Naples: –.
Moatti, C. () The Birth of Critical Thinking in Republican Rome, trans. J.
Lloyd. Cambridge.
Momigliano, A. () ‘The Theological Efforts of the Roman Upper Classes in
the First Century BC’, CPh .: –.
Mommsen, T. () The History of Rome, trans. W. Dickson. London.
Mora, F. () ‘“Irrazionalismo” nazionalistico nel pensiero teologico di Cicer-
one’, Bollettino di Studi Latini : –.
Most, G. () ‘Cornutus and Stoic Allegoresis: A Preliminary Report’, ANRW
..: –.
Mras, K. () Eusebius Werke: . Band, Die Praeparatio Evangelica, vol. . Die
Griechischen Christlichen Schriftsteller der Ersten Jahrhunderte . Berlin.
Murphy, T. () ‘Cicero’s First Readers: Epistolary Evidence for the Dissemin-
ation of his Works’, CQ .: –.
North, J. () ‘The Limits of the “Religious” in the Late Roman Republic’,
History of Religions .: –.
Nussbaum, M. () ‘Eros and the Wise: The Stoic Response to a Cultural
Dilemma’, OSAP : –.
Obbink, D. () ‘The Atheism of Epicurus’, GRBS .: –.
() ‘What All Men Believe – Must Be True: Common Conceptions and
consensio omnium in Aristotle and Hellenistic Philosophy’, OSAP :
–.
() Philodemus On Piety. Oxford.
Bibliography 
() ‘The Stoic Sage in the Cosmic City’, in Ierodiakonou (ed.): –.
() ‘“All Gods are True” in Epicurus’, in Frede and Laks (eds.): –.
O’Keefe, T. () Epicurus on Freedom. Cambridge.
Oppy, G. and Trakakis, N. (eds.) () The History of Western Philosophy of
Religion vol. . Durham.
Pease, A. () ‘The Conclusion of Cicero’s de Natura Deorum’, TAPA :
–.
(ed.) (–) M. Tulli Ciceronis De Divinatione. University of Illinois
Studies in Language and Literature . and .. Urbana.
(–) Cicero: De Natura Deorum: Bimillennial Edition.  vols. Cam-
bridge, MA.
Penelhum, T. () God and Skepticism : A Study in Skepticism and Fideism.
Philosophical Studies Series in Philosophy . Dordrecht.
Penwill, J. () ‘Epicurus’, in Oppy and Trakakis (eds.): –.
Pingree, D. () Dorothei Sidonii Carmen Astrologicum. Leipzig.
() Vettii Valentis Anthologiae. Leipzig.
Poncelet, R. () Cicéron: Traducteur de Platon: L’expression de la pensée com-
plexe en latin classique. Paris.
Popkin, R. () The History of Scepticism from Savonarola to Bayle. Oxford.
Powell, J. (ed.) (a) Cicero the Philosopher: Twelve Papers. Oxford.
(b) ‘Introduction’ to Powell (ed.): –.
(c) ‘Cicero’s Translations from Greek’, in Powell (ed.): –.
Purinton, J. () ‘Epicurus on “Free Volition” and the Atomic Swerve’,
Phronesis .: –.
Rackham, H. () Cicero: De Natura Deorum, Academica. Cambridge, MA.
() Cicero: De Oratore, De Fato, Paradoxa Stoicorum, De Oratoria, Cam-
bridge, MA.
Ramelli, I. () Anneo Cornuto: Compendio di teologia greca. Milan.
Rasmussen, S. () ‘Cicero’s Stand on Prodigies: A Non-Existent Dilemma?’,
in Divination and Portents in the Roman World, ed. R. Wildfang and J.
Isager. Odense University Classical Studies . Odense: –.
Rawson, E. () Intellectual Life in the Late Roman Republic. Baltimore.
Reiner, E. and Pingree, D. (–) Babylonian Planetary Omens.  vols.
Malibu.
Repici, L. () ‘Gli Stoici e la Divinazione secondo Cicerone’, Hermes .:
–.
Rist, J. (ed.) () The Stoics. Berkeley.
Rossbach, O. () T. Livi: Periochae, Omnium Librorum Fragmenta, Oxy-
rhynchi Reperta, Iulii Obsequentis Prodigiorum Liber. Leipzig.
Rüpke, J. (ed.) () A Companion to Roman Religion. Oxford.
() Religion in Republican Rome. Philadelphia.
() ‘Historicizing Religion: Varro’s Antiquitates and History of Religion in
the Late Roman Republic’, History of Religions .: –.
Sachs, A. and Hunger, H. (–) Astronomical Diaries and Related Texts
from Babylonia.  vols. Vienna.
 Bibliography
Salles, R. () ‘Ἐκπύρωσις and the Goodness of God in Cleanthes’, Phronesis,
., –.
(ed.) (a) God & Cosmos in Stoicism. Oxford.
(b) ‘Chrysippus on Conflagration and the Indestructibility of the
Cosmos’, in Salles (ed.): –.
Santangelo, F. () ‘Law and Divination in the Late Roman Republic’, in
Tellegen-Couperus (ed.): –.
() Divination, Prediction and the End of the Roman Republic. Cambridge.
Schäublin, C. () Marcus Tullius Cicero: Über die Wahrsagung. Berlin.
Scheid, J. () ‘Polytheism Impossible; or, the Empty Gods: Reasons behind
the Void in the History of Roman Religion’, in The Inconceivable Polytheism:
Studies in Religious Historiography, ed. F. Schmidt. History and Anthropol-
ogy : –.
(–) ‘La parole des dieux. L’originalité du dialogue des Romains avec
leurs dieux’, Opus –: –.
() Romulus et ses frères: le Collège des Frères Arvales, modèle du culte public
dans la Rome des empereurs. Paris.
() An Introduction to Roman Religion, trans. J. Lloyd, Bloomington.
() Quand faire, c’est croire. Les rites sacrificiels des Romains. Paris.
Schofield, M. () ‘Preconception, Argument and God’, in Schofield, Bur-
nyeat, and Barnes (eds.): –.
() ‘Cicero For and Against Divination’, JRS : –.
() The Stoic Idea of the City. Cambridge.
() ‘Academic Epistemology’, in K. Algra, J. Barnes, J. Mansfeld, and
M. Schofield, (eds.): –.
() ‘Academic Therapy: Philo of Larissa and Cicero’s Project in the Tuscu-
lans’, in Philosophy and Power in the Graeco-Roman World: Essays in Honour
of Miriam Griffin, ed. G. Clark and T. Rajak. Oxford: –.
() ‘Ciceronian Dialogue’, in The End of Dialogue in Antiquity, ed. S.
Goldhill. Oxford: –.
() Aristotle, Plato and Pythagoreanism in the First Century BC: New Direc-
tions for Philosophy. Cambridge.
Schofield, M., Burnyeat, M., and Barnes, J. (eds.) () Doubt and Dogmatism:
Studies in Hellenistic Epistemology. Oxford.
Schofield, R. () Vitruvius: On Architecture. London.
Schultz, C. (). Commentary on Cicero, De Divinatione I. Michigan Classical
Commentaries . Ann Arbor.
Schütrumpf, E. () ‘Platonic Elements in the Structure of Cicero de Oratore
Book ’, Rhetorica : –.
Schütrumpf, E., Stork, P., Van Ophuijsen, J., and Prince, S. () Heraclides of
Pontus: Texts and Translations. Rutgers University Studies in Classical
Humanities . New Brunswick, NJ.
Sedley, D. () ‘The Ethics of Brutus and Cassius’, JRS : –.
() ‘The Origins of Stoic God’, in Frede and Laks (eds.): –.
() Creationism and its Critics in Antiquity. Berkeley.
Bibliography 
() ‘Epicurus’ Theological Innatism’, in Fish and Sanders (eds.): –.
() The Philosophy of Antiochus. Cambridge.
() ‘Cicero and the Timaeus’, in Schofield (ed.): –.
Shackleton Bailey, D. (–) Cicero’s Letters to Atticus.  vols. Cambridge.
() Cicero: Letters to Quintus and Brutus; Letter Fragments; Letter to Octa-
vian; Invectives; Handbook of Electioneering, Cambridge, MA.
Sharples, R. () Alexander of Aphrodisias: On Fate. London.
() Cicero: On Fate and Boethius: The Consolation of Philosophy IV.–, V.
Warminster.
Smith, P. () ‘“A Self-Indulgent Misuse of Leisure and Writing”? How Not to
Write Philosophy: Did Cicero Get it Right?’, in Powell (ed.): –.
Starr, R. () ‘The Circulation of Literary Texts in the Roman World’, CQ
.: –.
Steel, C. () Reading Cicero: Genre and Performance in Late Republican Rome.
London.
Steinmetz, P. () ‘Beobachtungen zu Ciceros philosophischem Standpunkt’,
in Fortenbaugh and Steinmetz: –.
Striker, G. () ‘Cicero and Greek Philosophy’, HSCPh : –.
(ed.) (a) Essays in Hellenistic Epistemology and Ethics. Cambridge.
(b) ‘Κριτήριον τῆς ἀληθείας’ in Striker (ed.), –.
(c) ‘Sceptical Strategies’, in Striker (ed.) –.
(d) ‘On the Difference between the Pyrrhonists and the Academics’, in
Striker (ed.): –.
Süss, W. () Cicero: Eine Einführung in seine philosophischen Schriften (mit
Ausschluß der staatsphilosophischen Werke). Akademie der Wissenschaften
und der Literatur in Mainz, Abhandlungen der geistes- und sozialwis-
senschaftlichen Klasse : –. Mainz.
Taran, L. () ‘Cicero’s Attitude Towards Stoicism and Scepticism in the De
Natura Deorum’, in Florilegium Columbianum: Essays in Honor of Paul Oskar
Kristeller, ed. K. Selig and R. Somerville. New York: –.
Tarrant, H. () ‘Recollection and Prophecy in the De Divinatione’, Phronesis
.: –.
Tarver, T. () ‘Varro and the Antiquarianism of Philosophy’, in Philosophia
Togata I, eds. J. Barnes and M. Griffin, Oxford, –.
Tellegen-Couperus, O. (ed.) () Law and Religion in the Roman Republic.
Mnemosyne Supplements: History and Archaeology of Antiquity .
Leiden.
Timpanaro, S. () ‘Alcuni fraintendimenti nel De Divinatione’, in Nuovi
contributi di filologia e storia della lingua latina. Bologna: –.
() Cicerone: Della divinazione. Milan.
Thesaurus linguae latinae (–) Leipzig.
Thorsrud, H. () Ancient Scepticism. Ancient Philosophies . Stocksfield.
() ‘Arcesilaus and Carneades’, in Bett (ed.): –.
Treggiari, S. () Terentia, Tullia and Publilia: The Women of Cicero’s Family.
Women of the Ancient World. Abingdon.
 Bibliography
Tsouni, G. () ‘Antiochus on the Happy Life’, in Sedley (ed.): –.
Usener, H. () Epicurea. Leipzig.
Van der Horst, P. () Chaeremon: Egyptian Priest and Stoic Philosopher.
Leiden.
Van Nuffelen, P. () ‘Varro’s Divine Antiquities: Roman Religion as an Image
of Truth’, CP .: –.
Vogt, K. () Law, Reason and the Cosmic City: Political Philosophy in the Early
Stoa. Oxford.
von Arnim, H. (–) Stoicorum Veterum Fragmenta.  vols. Leipzig.
Wallace-Hadrill, A. () Rome’s Cultural Revolution. Cambridge.
Walsh, P. () Cicero: On the Nature of the Gods. Oxford.
Walzer, R. and Frede, M. () Galen: Three Treatises on the Nature of Science.
Indianapolis.
Wardle, D. () Cicero: On Divination, Book . Oxford.
White, G. () “Copia verborum: Cicero’s Philosophical Translations.” Diss.
Princeton. ProQuest order no. .
Whitmarsh, T. () Battling the Gods: Atheism in the Ancient World. New York.
Wissowa, G., Kroll, W., and Mittelhaus, K. (eds.), (–) Paulys Realency-
clopädie der classischen Altertumswissenschaft. Stuttgart.
Woolf, R. () Cicero: The Philosophy of a Roman Sceptic. Abingdon.
Wynne, J. () ‘God’s Indifferents: Why Cicero’s Stoic Jupiter Made the
World’, Apeiron : –.
() ‘Learned and Wise: Cotta the Sceptic in Cicero’s De Natura Deorum’,
OSAP : –.
Yon, A. and Guillaumont, F. (eds.) () Cicéron: Traité du Destin. Paris.
Zarecki, J. () Cicero’s Ideal Statesman in Theory and Practice. London.
Zoll, G. () Cicero Platonis Aemulus: Untersuchung über die Form von Ciceros
Dialogen, besonders von De Oratore. Zürich.
General index
Academy, , , , ,  assent, 

action,  astrology, , , , , 
Aeneas, ,  ataraxia. See Epicureanism: happiness.
Aeschines the Socratic,  Atargatis, , 
Aesculapius, , ,  atheism, , , , , , 
afterlife. See heroes. Athenodorus of Tarsus, 
agnosticism, , , see also skepticism. Atlas, 
Alcmaeon of Croton,  Atticus, Titus Pomponius, , , , 
Alexander of Aphrodisias .,  Attus Navius, , , 
allegory, , , ,  auctoritas.  See authority.
altars, , ,  augury, , , , , , , –,
Amafinius, Gaius,  , , 
Amphiaraus,  auspices. See augury.
Anaxagoras of Clazomenae, ,  authority, , 
Anaximander, 
Anaximenes of Miletus, ,  Babylon, , , 
anger, , ,  Balbus, Quintus Lucilius, , 
anthropomorphism, , , , , –, beauty, , , , , , 
 belief, , 
Antiochus of Ascalon, , , ,  Boethus of Sidon, , 
Sosus,  Brutus, Marcus Junius (Iunius  RE), 
Antipater of Tarsus, , , ,  Brutus, Marcus Junius (the liberator), , , ,
On superstition,  
Antisthenes, , 
Apelles,  Cacus, 
Apis,  Caelum, 
apographa,  Caesar, Gaius Julius, , 
Apollo, , ,  Calchas, , 
apotheosis, – Carneades, , , , , , , , ,
Appius Claudius Pulcher (Claudius  RE),  
Ara Maxima,  Castor, –, , 
Aratus of Soli,  catalepsis. See grasp.
Arcesilaus,  Cato the Younger, Marcus Porcius, , 
Archelaus (the philosopher),  Catulus, Quintus Lutatius, 
Aristo of Chios,  Central Question, the, , , –, , , ,
Aristotle, , –, , , –, , , , , , , , , , ,
 , , , 
Arnim, Hans von, ,  Cepheus, 
art, ,  Ceres, , –, 
Artemidorus of Daldis,  Chaeremon of Alexandria, 
artificial divination, , , , ,  chance, , , , 

 General index
Chrysippus of Soli, , , , , , Coruncanius, Tiberius, , 
–, , , , , , , Cotta, Gaius Aurelius, , , , , 
, , , –, ,  Crassus, Marcus Licinius (the triumvir), 
On fate, ,  Cratippus of Pergamon, , , , 
Cicero junior, Marcus Tullius,  cult images of the gods, , , , 
Cicero junior, Quintus Tullius,  cultural gods. See classification of gods, Balbus’.
Cicero, Marcus Tullius Cupido, , 
Academic books,  Curio, Gaius Scribonius, 
Academica, , , , , , , , ,  Cyrus the Great, 
Consolation, 
De consulatu suo, ,  Dagon, 
De domo sua,  daimones. See heroes.
De haruspicum responsis,  definitions of divination, , , 
Hortensius, , , ,  Delphic oracle, , , see also oracles; frenzy.
Laws, , ,  Democritus, 
Letters to Atticus Diagoras of Melos, 
.,  dialogue form, , 
..,  Diana, , , 
Lucullus. See Academica. Dicaearchus of Messana, , 
Oeconomicus, ,  Diels, Hermann, , , 
On augury,  Diodorus Cronus, 
On divination,  Diogenes of Apollonia, 
On duties, , ,  Diogenes of Babylon, , , , , 
On ends, , –, –, , –,  Diogenianus (philosophical critic of divination),
On fate,  
On friendship, , ,  Dis, , , , 
On old age, , , ,  Dorotheus of Sidon, 
On the nature of the gods,  dreams, , , , , , 
On the orator, 
Pro Roscio Amerino,  Egyptian religion, , , 
Protagoras,  Empedocles, , 
religious views,  Empiricist and Rationalist doctors, , 
Republic, , ,  encyclopedia view of Late Sequence, 
Timaeus, , , ,  Ennius, Quintus, , , , , 
Tusculans, , , , , –, –,  Enuma Anu Enlil, 
Cicero, Quintus Tullius, ,  Epicureanism
classification of the gods, Balbus’, ,  canonics, 
cultural level,  divination, 
paradigmatic level,  Epicurus’ authority, 
type , ,  happiness, , , 
type , ,  Latin Epicurean authors, , 
type ,  piety, , 
type ,  religion, , 
type , ,  theism, 
Cleanthes, , , , ,  theology, , , , 
Clitomachus of Carthage, ,  view of rival theologies, , 
Codrus,  Epicurus, , , , , 
community, , , , , ,  epistemology
Concordia, , ,  Epicurean, 
conditionals, ,  Stoic and Academic, , 
coniectura. See conjecture. Erechtheus, 
conjecture, , , , , , ,  Eros, , see also Cupido.
consensus, arguments from, , ,  eternity of the gods, , , , 
Cornutus, Lucius Annaeus, , , –, Etruscan religion, , , 
 etymology, , , 
General index 
Euripides Kronos, 
Hercules furens, 
Eusebius of Caesarea Lactantius, Lucius Caelius Firmianus, 
Preparation for the Gospel,  Laelius Sapiens, Gaius, 
Evander, ,  Laius, 
extispicy. See haruspicy. Late Sequence, –, –, 
Latin language, , , 
faith, good,  Latinus, 
fatalism,  Lazy Argument, the, , 
fate, , , , , ,  learned reader principle, , , , , ,
fear of the gods (δεισιδαιμονία),  , 
fideism,  Lebadia, , 
Fides, , , ,  Liber (god of wine), , 
fides. See faith, good. Liber (son of Semele), , 
frenzy, , , , ,  Libertas, , 
literary unity principle, , , , , , 
Giants,  Livy (Titus Livius), 
Gracchus, Gaius Sempronius,  lots, divination by, , , 
Gracchus, Tiberius Sempronius,  Lucretius Carus, Titus, –, , , 
grace, ,  Lucullus, Lucius Licinius, 
grasp, , ,  Luna, , 
Lysias, 
Hades, , 
Hannibal Barca,  madness.  see also frenzy.
haruspicy, , , , , , , , , Marcellus, Gaius Claudius (Claudius  RE),
 
haste in writing Late Sequence, ,  Marcellus, Marcus Claudius (Claudius  RE),
Hera, ,  
Heraclides of Pontus, –, , , ,  Marcus, , see also Cicero, Marcus Tullius.
Hercules, , , ,  Mars, , , 
Herodotus of Halicarnassus,  meaning, 
heroes, , , , ,  Mens, , , 
Hestia, ,  Mercury, 
Hirtius, Aulus,  Midas, 
history of divinatory arts, , –, ,  Minerva, , , 
holiness, , ,  Mnesarchus of Athens, 
Homer, , ,  moderation of religion, , , , , ,
Honor, ,  , 
Hortensius Hortalus, Quintus,  Mommsen, Theodor, 
moon, , –, , , 
ignorance,  myths, , , –, , , , 
illumination (metaphor for writing of Late
Sequence),  natural divination, , , , 
impiety, , , see also moderation of religion. Neptune, , , 
in utramque partem form of dialogue, ,  Nigidius Figulus, Publius, , 
interpretation of dreams or oracles, –, Numa Pompilius, , , , 
–, 
Isis,  oaths, , , 
Obsequens, Julius, 
Janus, , , , ,  Odysseus, 
Judaism,  Oedipus, 
Juno, , , , , , ,  omens, 
Jupiter, , –, , , , , On piety, , , , , , 
–, , , ,  Ops, , , 
Juventas,  oracles, , see also frenzy.
 General index
orthodoxy,  Quellenforschung. See source criticism.
orthopraxy, –,  Quintus, , , see also Cicero, Quintus Tullius.
Osiris,  Quirinus, , 
outcomes, –, , , ,  Rabirius, 
Panaetius of Rhodes, , ,  ratio of divination. See theories of divination,

Parmenides of Elea,  causal.
Penates, , , ,  Rationalist doctors. See Empiricist and
performance, ,  Rationalist doctors.
Peripateticism, –, , , ,  reason
Persaeus,  in Epicurean theology, 
Persephone,  in Stoic theology, 
Phalaris,  religio. See religion.
Philo of Larissa, ,  religion, , , , , –, , –
Philodemus of Gadara. See On Piety. Remus, 
pietas. See piety. revelation, 
piety, , , , , , , see also rhetoric, , 
moderation of religion. rites, , , 
planets, , ,  Romulus, , , , , , , 
Plato, , –, , , , , , , Roscius, Quintus (the actor), 

Phaedrus,  sacra. See altars; images, cult; rites; sacrifices;
Timaeus, , ,  temple buildings.
plausible, ,  sacrifices, , , , , , , 
Plautus, Titus Maccius,  Sallustius
Pliny the Elder (Gaius Plinius Secundus),  author of Empedoclea, 
Pluto,  Gnaius, 
poets. See myths. Salus, , 
political context for dialogues,  sanctitas. See holiness.
Pollux, –, ,  Sarpedon, 
Pompey (Gnaeus Pompeius Magnus),  Saturn, , , 
pontifices, ,  sayables. See meaning.
portents. See prodigies. Scaevola, Publius Mucius (Mucius  RE), ,
Poseidon,  
Posidonius of Apamea, , , , , , , Scaurus, Marcus Aemilius (Aemilius  RE),
, ,  
prayer, , , ,  skepticism
preconceptions Academic, , , , , 
Epicurean, ,  and religion, , , , 
Stoic,  Pyrrhonist, , 
probabile. See plausible. Radical and Mitigated, –, , 
probability theory,  science, , 
problem of evil, the,  scientia. See science.
prodigies, , , ,  Scipio Nasica Corculus, Publius Cornelius, 
prognostication, , ,  Scot, Alexander, 
Prometheus,  Seneca, Lucius Annaeus (the younger), 
propositional content. See meaning. sensation, , 
propositum form of dialogue, , – Serapis, 
Proserpina, ,  Sextus Empiricus. See skepticism: Pyrrhonism.
Protagoras, , ,  Sibylline books, 
Ptolemy, Claudius, ,  Simonides, 
Pyrrhonism. See skepticism. single source hypothesis, 
Pythagoras, , ,  societas. See community.
Pythagoreanism, ,  Socrates, , , , 
General index 
Sol, ,  Titans, 
Sophocles, ,  Torquatus, Lucius Manlius, 
sorites,  tranquility. See Epicureanism: happiness.
sortes. See lots, divination by. Trophonius, 
soul, , , , , , , , –, Tullia, , , 
, , , , , ,  type , , , , or  gods. See classification of
source criticism,  gods, Balbus’.
Speusippus, 
stars, , , ,  Uranus, 
Stoicism Usener, Hermann, 
conflagration, 
corporealism,  Valens, Vettius, 
cosmopolitanism,  Varro Reatinus, Marcus Terentius, , , , ,
cosmos,  , , , , , 
creationism,  Velleius, Gaius, 
determinism. See fate. and On piety, 
epistemology,  Venus, , , 
ethics, ,  Lubentina, , 
religion, ,  Venus (throw at dice), 
style of writing. See Stoicism: wisdom. veri simile. See plausible.
superstition,  Vesta, , , , 
theology, ,  vices, worship of, 
wisdom, , , , , , ,  Victoria, , 
Strato of Lampsacus, –,  Virgil Maro, Publius
straw men in Div.,  Aeneid, 
sun, , –, , ,  virtue, , , 
superstitio. See superstition. Virtus, , 
superstition, , , , , , –, Vitruvius Pollio, Marcus, 
see also moderation of religion. Voluptas, , 
‘Syrian’ religion, , 
weather forecasting. See prognostication.
Tages, ,  wonder, , , 
temple buildings, , , –, , –,
, ,  Xenocrates, 
Terentia,  Xenophanes of Colophon, 
Terminus,  Xenophon, , 
Thales of Miletus,  Xerxes I, 
theodicy. See problem of evil, the.
Theodorus of Cyrene (‘the Atheist’),  Zeno of Citium, , , , , , ,
Theophrastus, ,  
theory of divination, causal, , –, , Republic, 
 Zeus, , , , –
Index locorum antiquorum
Aeschines ., 

Aspasia,  ., 
Aëtius ., 
Placita De dialectica
., ,  , 
.., 
Alexander of Aphrodisias Charisius
Mantissa Ars grammatica ed. Barwick
., ,  , 
.–,  , 
.–,  Cicero, Marcus Tullius
On fate Academica
. Bruns,  ., 
.–,  ., 
.–,  ., 
[Andronicus of Rhodes] ., 
On passions .–, 
,  .–, 
.,  ., 
Aratus ., 
Phaenomena ., 
–,  .–, 
Aristotle ., 
History of animals .–, 
b–,  ., 
On divination in dreams .–, 
b–,  ., 
b–,  ., 
b–,  .–, , 
On dreams ., , 
a–,  .–, 
Athenaeus .–, 
Deipnosophistae ., 
.a,  ., 
Augustine ., 
City of God ., 
.,  ., 
.,  .–, 
.,  .–, 
.,  .–, 
.,  .–, 

Index locorum antiquorum 
.,  .., , 
.,  .., 
.,  .., 
.,  ..–, 
.–,  ..–, 
.,  .., 
.,  ..–, 
.,  .a, 
Brutus ., 
,  .., 
De domo sua ., , 
,  ., 
,  .., 
In Verrem .., 
..–,  .., 
Laws .., 
.,  .., 
.,  .., 
.–,  ., 
.–,  .., 
.–,  Letters to his friends
.–,  .., 
.,  .., 
., ,  .., 
., ,  Letters to Quintus
.,  .., 
.,  .., 
.,  On divination
.–,  ., , 
., ,  ., 
.,  ., 
.,  ., , 
.,  ., , 
.,  ., , , , , , , 
.–, ,  .–, 
.–,  .–, 
.,  ., , , , , , 
.–,  ., , 
.,  ., , 
.,  .–, 
.,  .–, 
.–, ,  ., , , –, 
., – ., , 
.,  ., 
Letters to Atticus ., 
..,  ., 
..,  ., , –
.., – ., 
., ,  .–, , 
.,  ., 
..,  .–, 
..,  ., 
.,  ., 
..,  .–, 
.,  .–, 
 Index locorum antiquorum
Cicero, Marcus Tullius (cont.) ., , , 
.,  ., 
.–,  .–, 
.,  ., , 
.,  ., 
.,  .–, 
.,  ., 
.,  ., , 
., ,  .–, 
.–,  ., 
.,  ., , 
.–,  .–, , 
.–,  ., 
.,  ., , –
., – ., 
., – ., , 
., – ., 
., ,  ., 
.–,  ., 
.,  ., –
., , , , ,  ., 
.–, ,  ., 
.,  ., 
.–,  ., –
.–, , ,  ., , –
.–,  ., –
.–,  ., 
.,  ., 
.–, ,  ., –
.,  ., 
.,  ., 
., ,  ., 
.,  ., 
.,  ., 
.–,  ., , , 
.–,  ., , , , , , 
., , ,  .–, , 
.–,  .–, 
., , ,  ., 
.–,  ., 
.,  ., 
.–, ,  ., , 
.–,  ., 
.,  ., 
.–,  ., , –, , 
.–,  ., 
.–,  ., , , –
., ,  .–, , , , –
., ,  ., , , , , , 
.–,  On duties
.–, – ., 
.,  ., 
.,  ., 
., ,  ., 
.–,  ., 
Index locorum antiquorum 
.–,  On invention
.,  ., 
On ends .–, 
.,  .–, , 
.–,  .–, 
., ,  On the nature of the gods
.,  ., , , , , 
.–,  .–, 
.–,  ., , , 
.–,  .–, , , 
.,  ., , 
.,  .–, 
.,  ., , –
.–,  ., 
.–,  ., 
.,  .–, , 
.,  ., 
.–,  ., , , , 
.–,  .–, 
.–,  ., 
.,  ., , , , , 
.,  ., , 
.,  .–, 
.,  ., 
.–, ., 
 ., 
.,  ., , 
.,  .–, , 
.,  ., 
.,  ., 
.,  ., 
.,  ., 
.,  ., 
.,  ., 
.,  ., 
.–,  ., –
.,  ., 
On fate .–, 
, ,  .–, 
,  .–, 
–,  ., , , , , 
–,  .–, , 
,  ., , 
–,  .–, 
–,  ., 
,  .–, 
,  ., , 
–,  ., 
,  .–, 
–,  .–, 
fr. ,  ., 
On friendship ., , , 
,  ., 
–,  ., , 
–,  ., , 
 Index locorum antiquorum
Cicero, Marcus Tullius (cont.) .–, –
.,  ., , , –
.–,  ., , –, , 
.,  ., , 
.–, ,  .–, 
.,  ., , , 
.,  .–, , 
., ,  .–, , 
.,  ., 
.,  ., 
.–,  .–, 
.,  ., 
., ,  ., , , –, , , , 
.–, ,  .–, 
.,  ., , 
.–,  .–, , , 
., , ,  ., 
.–,  .–, 
.,  .–, 
.–,  ., , , , 
.,  ., 
.–,  .–, 
.,  .–, 
., –,  ., 
.–, ,  ., , 
., , –,  ., 
.–, ,  ., 
.,  ., , 
.,  .–, 
.–,  ., 
.,  ., , 
., ,  ., 
.,  .–, 
.–,  ., , –, , 
.–, ,  .–, 
.,  ., 
.–,  .–, 
., ,  ., 
.,  ., 
.–,  ., 
.–,  .–, 
.–,  .–, 
., ,  .–, 
.–, ,  ., –, 
.–,  .–., 
.,  .–, 
.–,  ., 
.,  ., , , , , 
.–,  .–, 
.–,  ., 
.–, – ., , 
.–, , ,  ., 
., ,  .–, 
.,  ., , , 
., , ,  ., 
Index locorum antiquorum 
., ,  .–, , 
.–,  .–, 
.–,  ., 
., ,  ., , –
.–,  ., 
., – ., 
., ,  ., 
.,  ., 
.,  ., , 
.–,  .–, 
., ,  ., 
.–,  .–, 
., – ., 
.–,  ., , 
.–,  ., 
.–,  ., 
.,  .–, 
.–,  ., 
.,  ., 
.,  .–, 
.,  .–, 
., ,  ., 
.,  ., –
.,  ., 
.,  ., 
., , , , , – .–, 
On the orator ., 
.,  ., 
.,  ., , 
.,  ., 
.–,  Cleanthes
.,  Hymn to Zeus
.–,  –, 
., ,  Clement of Alexandria
., ,  Stromateis
On the parts of rhetoric ., 
–,  Cornutus
,  Theology
Orator  Lang, 
,   Lang, 
Pro Cluentio  Lang, 
,  – Lang, 
, 
Pro Murena Diogenes Laërtius
–,  Lives of the philosophers
Pro Roscio Amerina ., 
,  ., 
Republic ., 
.–,  ., 
.–,  ., 
Timaeus ., 
, ,  ., 
–,  ., 
Tusculans ., 
.–,  ., 
 Index locorum antiquorum
Diogenes Laërtius (cont.) On the anger of God
., , ,  , 
.,  Livy
., , , ,  Ab urbe condita
.,  .–, 
.–,  ..–.., 
Diogenes of Oenoanda ., 
Fr.  Smith,  .–, 
Dionysius of Halicarnassus .–, 
Roman antiquities ., 
.–,  [Longinus]
Donatus De sublimitate
On Phormio , 
..,  Lucretius
De rerum natura
Epictetus .–, 
Dissertations .–, 
.–,  .–, 
Epicurus ., 
Letter to Herodotus ., 
,  .–, 
Letter to Menoeceus ., 
,  ., 
Principal Doctrines .–, 
, ,  .–, 
,  .–, 
Eusebius .–, 
Preparation for the Gospel .–, 
., ,  .–, 
.,  ., 
.., ,  Lydus, John
De mensibus
Galen ., 
On the doctrines of Plato and Hippocrates De ostentis
..,  , 
On the sects for beginners , 
- Helmreich, 
.–,  Macrobius
.,  Saturnalia
Gellius, Aulus .., 
Noctes Atticae
..,  Obsequens, Julius
..,  Prodigiorum liber
..,  , 
Origen
Homer Contra Celsum
Iliad ., 
.–, ,  Ovid
.–,  Fasti
.–, 
Lactantius .–, 
Divine institutes
..,  Philodemus
..,  On Piety
..–,  – Obbink, 
Index locorum antiquorum 
Photius Tetrabiblos
Lexicon ., 
Ν  (νεοττός), 
Plato Quintillian
Apology Institutes
d,  .., 
Epinomis .., 
–,  .., 
Phaedrus
b,  Seneca the younger
a,  De beneficiis
c,  .., 
b–d,  De constantia
Symposium .., 
–,  Letters
Timaeus , 
e–b,  ., 
Pliny the Elder Servius
Natural history In Aeneidem
.,  ., 
[Plutarch] Sextus Empiricus
Essay on the life and poetry of Homer Adversus mathematicos
,  .–, 
Plutarch .–, 
Life of Caesar ., 
,  ., 
Life of Cicero ., 
.–.,  ., 
On common conceptions ., , , 
c,  ., 
On Stoic self-contradictions ., 
c,  ., 
c,  ., 
d,  Outlines of Pyrrhonism
Posidonius ed. Edelstein and Kidd ., 
FC,  Simplicius
F,  In physica
FA,  . Diels, 
Fab,  Stobaeus, John
F,  ..b, 
F–,  ..i, 
T,  ..s, 
Priscian .., 
Institutes .., –
..,  .., 
..,  .., 
..,  ., 
Propertius .., 
Elegies Suetonius
.,  Lives of the Caesars
Ptolemy, Claudius ., 
Almagest
..–, Varro
 Antiquities human and divine ed. Cardauns
.,  Fr. a, 
 Index locorum antiquorum
Varro (cont.) Virgil
Fr. ,  Aeneid
Fr. ,  .–,
Fr. ,  
Fr. , ,  Vitruvius
Fr. ,  On architecture
Fr. ,  .., 
Fr. ,  .., 
On the Latin language
.–,  Xenophon
.,  Oeconomicus
.,  ., 

Wynne - Cicero On The Philosophy of Religion PDF

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Wynne - Cicero On The Philosophy of Religion PDF

Caricato da

Copyright:

Formati disponibili

CICERO ON THE PHILOSOPHY OF RELIGION

. . .  is Associate Professor of Classics in the Department

Cambridge University Press is part of the University of Cambridge.

Acknowledgments page viii

Introduction: Cicero and the Translation of Philosophy from

Appendix : Terminology in DND and Div. for Religious Virtues

This book began as a  dissertation at Cornell. I would like to thank

DL Diogenes Laërtius, Lives of the eminent philosophers

When I cite certain of Cicero’s works central to understanding his philo-

My label Latin title Other common titles

What I cite, following convention, as the Academica, is a portmanteau text.

. An Older View of the Late Sequence

. A More Positive View of Cicero’s Dialogues

. The Single Source Hypothesis

. Cicero and the Revival of the Classical Dialogue

. Cicero and the Skeptical Dialogue

Cicero explains that, following Caesar’s death, the consul-designate Hirtius

. Cicero’s Theological Trilogy

Cicero’s Project in On the Nature of the

. Action, Belief, and Roman religio

. Theological Facts and Conventional Piety

. How Philosophical Inquiry Can Moderate Religion

Cotta’s criticism uses the conceptual map of pious religion between

. Cicero’s Project in its Intellectual Context

. How the Project Shapes the Dialogues

Velleius the Epicurean

. Velleius’ Theology

. Velleius’ Treatment of his Opponents

. Religion and Epicureanism

Epicurus, then, did not wish to alter the performances of traditional

Cotta’s incredulity indicates that Epicurus claimed not to overthrow

Balbus the Stoic and Cotta the Skeptic

Nature, says Balbus, is not only craftsmanlike, but is actually a craftsman.

. Balbus versus Cotta on Religion

. Balbus and the Central Question

.. That God Governs the World and Cares about Us

Vitruvius’ three concerns of the human architect correspond to Balbus’

.. Beauty and the Central Question

. Balbus’ Stoic View of Traditional Religion

He sums them up thus:

In this passage, where Balbus has returned to more abstractly physical

. Cotta’s Response

.. Cotta and the Problem of Us

.. Cotta’s Reply to Balbus on Religion

Quintus’ Stoic Case for Divination

Consider again the following passage (cf. p.  above):

. The Occasion for Quintus’ Speech

. Chrysippus and his Critics on Divination

. Quintus’ New Stoic Case for Divination

. Quintus’ Case from the Outcomes for Artiﬁcial Divination

. Quintus’ Case from the Outcomes for Natural Divination

Marcus’ Arguments against Divination

In Book  of DND, Cotta the pontifex undertook to face down Stoic

. Is Marcus’ Speech Suspiciously Unfair to Quintus?

. The “Sally”: Div. .–

. The Main Strategy: Div. .–

. Marcus on the History of the Divinatory Arts

But now, every conjecture—and divination is founded on conjecture—is

. “Chance Can Imitate Truth”

Marcus’ Stance on the Central Question

. The Evidence

. The Epistemology of Marcus’ Stance

. Marcus’ Stance at the End of Div.