Journal of Philosophy, Inc.

A Critique of Philosophies of History

Author(s): Maurice Mandelbaum
Source: The Journal of Philosophy, Vol. 45, No. 14 (Jul. 1, 1948), pp. 365-378
Published by: Journal of Philosophy, Inc.
Stable URL: http://www.jstor.org/stable/2018939 .
Accessed: 29/08/2014 16:08

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Journal of Philosophy, Inc. is collaborating with JSTOR to digitize, preserve and extend access to The Journal
of Philosophy.

http://www.jstor.org

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
VOLUME XLV, No. 14 JULY 1, 1948

THE JOURNAL OF PHILOSOPHY

A CRITIQUE OF PHILOSOPHIES OF HISTORY'

0T HE purpose of this paper is to analyze the basic presuppo-
sitionsof all traditionalphilosophiesof historyand to suggest
that the often-notedconflictbetweenempiricalhistoriographyand
philosophiesof historyis in fact a necessary conflict. If this be
true it may, I think,be safely concludedthat the attemptto con-
structa philosophyof historyis a theoreticallyinvalid enterprise.
1. Definition of a Philosophy of History. If one were to
framea definitionof a philosophyof historyin terms,say, of Saint
Augustine,Bossuet, Condoreet,Herder, and Hegel, one could do
so by pointing out that every philosophyof historyattemptsto
sum up past history,to grasp it as a whole,and to decipherits ulti-
mate meaning. However,the monisticlinear interpretationof the
stream of human historywhich these accounts presuppose is not
to be foundin the morerecentmorphologicalphilosophiesof history
of Spenglerand Toynbee. Yet it appears to me that morphological
philosophies of history are in their most fundamentalrespects
similar to the earlier attempts. I thereforepropose to define a
philosophyof historyin such a way as to include both general
classes. As a definitionI offerthe followingstatement: a phi-
losophyof historyis any interpretationof historywhichpurports
to derive from a considerationof man's past a single concept or
principlewhich in itselfis sufficient to explain the ultimatedirec-
tion of historicalchange at every point in the historicalprocess.
Thus, any philosophyof historyconsistsin the formulationof a
law of historical change which explains the direction of flow of
concreteevents. This I take to be the essentialthesisof a philosophy
of history.
However,it mustbe notedthat almosteveryso-calledphilosophy
of history also contains other theses. Among these there are
usually to be found not only evaluative judgments of specific
civilizations,but also theoriesconcerningthe role of specificcausal
factorsin history-as-actuality.Thus, in Herder and Toynbee,for
example,much of the work is not concernedwith tracing an over-
all principleor law held to be explanatoryof the directionof his-
1 Paper read before the Conference.on Methods in Philosophy and the
Sciences, held at The New School for Social Research, December, 1947.
365

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
366 THE JOURNAL OF PHILOSOPHY

torical change,but is focusedupon the interplayof specificfactors

present at any one point in the historicalprocess. If, then, one
feels it legitimateto stress these aspects of Herder or Toynbee,
minimizingthose passages in which they express a convictionre-
gardinga law of historicaldirection,theirworkswould not be rep-
resentativeof what I have called "a philosophyof history." But
it is my interpretationof Herder and Toynbee,as well as of Saint
Augustine, Bossuet, Condorcet, Hegel, Spengler, etc., that the
searchfora principleof historicalchangeconstitutesthe framework
withinwhich theirevaluative judgmentsand theiranalyses of his-
toricalcausationare set,and thatthisframeworkis the mostfunda-
mental aspect of their thinking. Thus, though no philosophyof
historyconfinesitselfto tracingout a principlesufficient to explain
the ultimate directionof historicalchange, I find that the attempt
to discover such a principle is the essential common property of
those works which are usually denominated as "philosophies of
history.22
It must furthermore be noted that the term "a philosophyof
history,"as used in this sense, does not cover every concernwith
the problemof law in history. A law whichattemptsto formulate
invariant relations between historical events is not necessarilya
law designed to express the concretecourse of historicaldevelop-
ment. Therecan be laws concerninghistoryand theremay be laws
of history. To elucidate: if Marx be interpretedas holding that
there are inevitable developmentalstages of history which are
strictlydeterminedby the operation of a necessary sequence of
economicfactors,then he is giving us a law of history; if, on the
otherhand, he be interpretedonly as holding that at every point
in the historicalprocess economicfactorsoperate to determinethe
nature of otherfactorsin history,he is givingus a law concerning
history. It seems to me that one can findboth aspects in Marx's
thoughtconcerninghistoricallaws, and that some confusionhas
been born of this ambiguity. However,I do not wish to place any
stresson myuse of Marx as an illustration;the point whichI wish
to make is that there is a differencebetween attemptsto find in-
variant relationswithinhistoryand attemptsto show that thereis
2 I should like to point out that in what follows I am not denying the

legitimacy of " philosophies of history" in any sense other than that in

which I am using the term. I should in fact claim that a consideration of
the historical process raises many philosophic problems concerning the nature
of man, the nature of social institutions,the causes of changes in social insti-
tutions, and the relations between human norms and institutions. An ade-
quate consideration of these problems seems to me to be one of the most basic
and urgent philosophic needs of our time.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 CRITIQUE OF PHILOSOPHIES OF HISTORY 367

an invariant tendency of history. In speaking of philosophies

of historyI am referringonly to the latter theories.
This distinctionbetween two types of historical law may be
made moreexplicitby means of an analogy. A law of historymay
be comparedwith the second law of thermodynamics;a law con-
cerninghistorymay be comparedwith other laws of physics,for
example with the gravitationallaw formulatedby Newton. A law
of historyis analogous to the law of entropysince it states a direc-
tional tendencywithinthe systemunder consideration,a tendency
whichof itselfdeterminesthe generalcoursewhichthe transforma-
tions within that systemwill follow. Thus, a law of historynot
only has an explanatoryfunctionwhen we view any phase in the
whole series of transformations;it also has direct predictivesig-
nificance. However, a law concerninghistoryis analogous to the
law of gravitationin that it, consideredby itself,has an explana-
tory significance,but is not directlypredictive. A law concern-
ing invariantrelationshipsis only of predictivesignificancein an
indirect fashion: it permits us to calculate on the basis of a
knowledgeof initial conditions,and on the assumptionof the non-
interference of otherevents,whatwill happen next; it does not sum
up and definethe directionin whichchangetends to proceed. The
enormoussignificanceof these functionallaws for the prediction
of immediatelyfutureeventsarises out of the fact that theyenable
us to grasp how one stage of a process followsfromthe preceding
stage; however, their long-range predictive significanceis nil
exceptwhenwe are dealing withprocessesin whichwe may reason-
ably postulate the non-interference of other events. It is only in
such cases that we can calculate froma knowledgeof earlier con-
ditionswhat the conditionswhich we must take into account at a
later stage will be. For this reason the law of gravitation,which
has an extensive predictive significancein astronomy,lacks sig-
nificancefor long-rangepredictionsconcerningterrestrialevents.
And it appears to me fairlycertainthat in the fieldof historyany
laws concerningthe interrelationshipsof events would lack long-
range predictivevalue: there is no reason to believe that when we
know the initial conditionsat one point in time we can predict
from a knowledgeof invariant relationshipswhat the conditions
at a later point will be-in history,it seems,eventswhichwere not
part of the original initial conditionsalways impingeto alter the
conditionsat a later point in time. If, then,we wish to formulate
a historicallaw whichwill have long-rangepredictivesignificance,
it mustbe a law of historybased on the analogy of the second law
of thermodynamics;rather than being analogous to the law of
gravitation. And why,it may be asked, may we not believe it to

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
368 THE JOURNAL OF PHILOSOPHY

be possible to findsuch a law? The answerlies in the fact that to

do so we mustassumethat we are dealing withself-transformations
withina single closed system. The second law of thermodynamics,
it is to be noted,mustbe stated in theseterms. And in fact every
philosopherof historyin attemptingto state the law of history
assumes that he is dealing with a closed system. It shall later be
my aim to criticizethis presupposition.
2. Concerning the Nature of Empirical Historiography. It
would, I suppose,be grantedby all that the truthor the falsityof
a philosophyof historyis to be judged by the confirmation which
it receives when we consider the course of historical events. It
also appears to me to be true that it is throughthe methodsof
empirical historiographythat we ascertain (in so far as we can
ascertain) this courseof events.3 Therefore,if the presuppositions
of a philosophyof historyare necessarilyincompatiblewith the
results obtainable by empirical historiographywe have no choice
but to admit that the attemptto constructa philosophyof history
is a theoreticallyinvalid enterprise. In the attemptto show that
such an incompatibility existsI shall firstexamineone fundamental
fact concerningthe nature of empiricalhistoriography.
It is generallyconceded that the aim of historiographyis to
understandand depict the concretenature of human societiesand
the changeswhichtheyhave undergone.4 In dealing with societies
the historiographer is always concernedwith what mightbe called
their cultural components,i.e., with their political, economic,and
religiousinstitutions,with theirtechnology,their familyand class
organization,theirjuridical systems,theirmodes of thought,their
arts,and the like. All otherelementswhichenterinto a historical
account (e.g., single individual personalities,contacts with other
societies,or natural eventssuch as climaticchanges) do so only in
so far as theyare seen to be related to the nature and the changes
of these cultural components. To understand the nature of a
societyis to understandthese componentsand to apprehend their
3 For the sake of brevity I omit a defense of this thesis. It does not
appear to me to be possible to hold that a philosopher of history has special
access to history-as-actualityfrom which the empirical historiographer is
debarred. The philosopher of history may rewrite history, but in so doing
it is incumbentupon him to show that his reading of the past is more adequate
than other readings; this he must do in precisely the same manner as that
by means of which any historiographerwill argue for the relative adequacy of
his account of the past.
4 I am here excluding from consideration the whole field of "specialized"
histories, as not being germane to the contrast which I wish to draw between
empirical historiographyand philosophies of history. But what is here said
may also be applied, with suitable transformations,to the interesting and
important problems of specialized historiography.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 CRITIQUE OF PHILOSOPHIES OF HISTORY 369
interrelationships;to understandits changes is to understandthe
changes in their character or in their relationships. Thus, the
understandingof a societyis only gradually to be won; it emerges
out of an understandingof the nature and structuralrelations of
the cultural componentswhich make the society what it is. In
short,for the historiographera society is an interrelatedset of
cultural components. On the other hand, as we shall see, the
philosopherof historyis forcedby his search for a law of history
to treatsocietiesas unitaryentitieswhosenaturesare determinative
of the natures of their componentelements.
The issue is thus a metaphysicalone; and it mightbe thought
that the philosopherof historyhas a right to advance his meta-
physical hypothesiseven though it be contraryto the hypothesis
of the empiricalhistoriographer. However, as I shall now briefly
suggest,the metaphysicalhypothesisof the philosopherof history
rests upon a confusionbetween two differentsenses in which we
may speak of the "parts" of any process.
It would, I think,be agreed by both historiographers and phi-
losophers of historythat society is not "a thing," like desks or
chairs or buildings,but that it is "a process" in much the same
sense as a footballgame or a conversationis a process.5 When we
thinkof "a thing" (I am here speakingin strictlyphenomenologi-
cal terms) we think of an object whose parts and qualities are
copresentwith it; but "a process" is temporalin character,and
the relation between it and its parts and qualities is therebyaf-
fected. What, in the case of a process,may be called its "parts"?
It appears to me thatwe actuallythinkof the parts of a process
in eitherof two general ways. The firstis throughdistinguishing
eithersegmentsor phases withinthe process itself,and considering
these as its parts. Thus, the parts of a footballgame may be said
to be its four periods,or the parts of a conversation,or of a war,
the successivephases throughwhich it passed. But when we use
the conceptof "part" in this manner,we are merelysinglingout
one span fromthe larger process: the span is not an elementin
the process,but a portionof the process itself.
However, we may also speak of the parts of a process in the
sense of its elements. The constitutiveelementsof a footballgame
are, of course, the successive plays executed by the players, the
constitutiveelementsof the conversationare the remarkswhich
5 Societies are sometimescompared to organisms, and it might be claimed

that in avoiding this comparison I am guilty of neglecting the most relevant

metaphysical evidence. However, the metaphysical interpretation of organ-
isms contains precisely the same disputed problems with which we here have
to deal and can not therefore be fruitfully used in answering our question.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
370 THE JOURNAL OF PHILOSOPHY

were made. These elementsare "parts" of the process in a sense

quite differentfrom that in which a segment or a phase is "a
part." They are not merelythe processitselfseen as divided into
quantitativelyor qualitatively differentiatedportions; they are
the constitutiveelementswhich at each momentmake the continu-
ing processwhat it is.
The parts of a process (in the sense of its elements,its consti-
tutivecomponents)are not, of course,unrelated; were theytotally
unrelatedtheywould not be parts of the same process. A process
viewed in terms of its elementsis simply a set of interconnected
eventswhich,because of their interconnections, constitutea single
serial whole. Such a process may well be said to possess qualities
which the elements,taken singly,do not possess: not all qualities
of a whole need be qualities which reside in its individual isolable
parts. Yet such non-segregable, "over-all" qualities are always a
functionof relationshipsamong the parts. Therefore,if we wish
to understandthe over-allqualities of a process,for example,the
phases throughwhichit passed, we mustdo so on the basis of know-
ing the nature and relationshipsof its constitutiveparts. And
those over-all qualities which characterizethe process as a whole
are not determinativeof the elementswhichenterinto the process:
theycomeinto being onlybecause of the relationshipswhichobtain
betweenthese elements.
If the preceding general phenomenologicalremarks are true,
it is a confusionto hold that the over-allqualities which are pres-
ent in a given process are the constitutiveelementswhichmake it
what it is. On the contrary,they are qualities which can only be
predicatedof the processas a whole; theydo not in fact exist until
the process has been formedby the elementswhich,in their inter-
play, constituteit. Applied to the realm of historicaleventsthis
means that the over-allqualities attributedto any society (or other
continuingprocess) may trulybe descriptiveof its nature and the
courseof its changeswithoutbeing held to be factorswhichentered
intothe formationof that society. To attributea causal, formative
power to such over-allqualities is to confusethe aspects of a com-
plete process with the constitutiveelementswithin that process.
This confusion,if I am not mistaken,is an outgrowthof a confusion
betweenthe two meaningswhichthe term"part" may have when
applied to processes.
For the empirical historiographer,as I have suggested, the
parts of a process are its constitutiveelements:he seeks to obtain
knowledgeof the nature of a societythroughdiscoveringthe nature
and interrelationshipsof the changing institutions(the cultural
components) within it. For the philosopher of history,on the

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 CRITIQUE OF PHILOSOPHIES OF HISTORY 371

otherhand, those over-allaspects of a societywhich are held to be

descriptiveof its course as a whole are treated not as aspects of a
process but as determinativefactors within it. That this is
necessarilytrue of all philosophiesof historyI shall now attempt
to suggest, using as illustrative material Toynbee's attempt to
state and defend an empiricallyderived philosophyof history.
3. The Presuppositionsof a Philosophyof History. It will be
recalled that a philosophyof history,as I have definedthat term,
consistsin the attemptto finda law of history. If the philosophy
of historyis linear in character,the process with which the phi-
losopher of historyis concerned is the process of the whole of
human society; if it is morphological,the processes are those of
any and everyindividual society. If there is to be any such law
it must be based upon the fact that there is a necessary phase-
quality in these processes,i.e., a qualitative change characteristic
of it as a whole. If the philosopherof historyis not to findhimself
at odds with the historiographer,such a phase-quality must be
predicated of these processeson the basis of an adequate analysis
of the constitutiveelementsof the process,for any predicationof
a phase-qualitypresupposesa knowledgeof those interrelatedele-
ments which, taken together,are the process. With these state-
mentseveryphilosopherof historymightagree, claimingthat the
law of historywhich he has found is simply the generalizationof
a phase-qualitypresent throughouthuman society,or commonto
all societies.
We may firstnote two obvious difficulties in any such claim,
even thoughthese difficulties apply with equal force to those who
seek to establishnot a law of historybut a law concerninghistory.
The firstof these difficultiesis that it is doubtful whetherma-
terials exist on the basis of whichwe may justifiablyaffirm that we
have discovereda commonrepeated pattern of elementsor a com-
mon phase-qualityin all societies. The data at our disposal are far
too scattered and fragmentaryto afford us the possibility of
building up a concreteknowledgeof what was, say, the Minoan
society. At best we can do what Spengler and Toynbeehave done,
namely,attemptto findthe phase-qualitiesof those societies con-
cerning which we have a comparativewealth of knowledgeand
apply the hypothesisthat the same phase-qualitieswere presentin
other societies. If it can be shown that the fragmentarydata at
our disposal in the other cases fit with our hypothesis,then we
have some measure of validation of the general applicability of
that hypothesis. But the more fragmentarythese data are, the
more suspicious we should be concerningthe confirmation which
we extractfromthem.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
372 THE JOURNAL OF PHILOSOPHY
The second obvious difficulty, which holds equally of attempts
to findlaws of historyand laws concerninghistory,arises fromthe
fact that a phase-qualityor a patterningof eventswhichwas found
to hold in the past may not hold in the future. This difficulty is
recognizedby all, and is present,of course,in all attemptsto dis-
cover empiricallaws: I mightmerelyremark,as othershave often
remarked,that the difficulties inherentin assumingthe uniformity
of nature are probably slight in comparisonwith the difficulties
whichare involvedwhen one assumes the uniformityof history.
To these two points a further,and to my mind decisive,objec-
tion may be added whenone considersthe attemptswhichare made
to discover a law of history. In formulatingsuch a law, the
philosopherof historynecessarilyassumes that the process with
which he deals is a more ultimate,or "real," entitythan are the
elementswhich are contained within it. He must assume this if
he is to believe that there are necessary phase-qualities in all
societies; otherwisethe process would flow in whateverdirection
was made necessary by the addition of new elements which-
throughnecessityor historical accident-became connected with
those elementswhich it already contained.
That philosophiesof historysuch as thoseof Saint Augustineor
Hegel or Spengler ascribe an ontological priority to the whole
needs, I suppose, no demonstration. It is interesting,however,to
see how explicitlyToynbee admits that,fromthe point of view of
knowledge (and I suspect also fromthe point of view of being),
the wholemustbe consideredas havingpriorityoverthe parts. In
an apposite passage he says: "In order to understandthe parts,
we mustfirstfocusour attentionupon the wholebecause this whole
is the fieldof studywhichis intelligiblein itself." 6
Now, I need not point out that for the historiographersuch a
belief is unfounded. For him,a given historicalprocess is not in-
telligiblein itself: to understandwhat it was, and to know why it
was, we must trace out the nature and interrelationships of those
morespecificeventswfthoutwhichit would not have been anything
-let alone intelligible. This is the fundamentalclash betweenthe
historiographicalmethodand the presuppositionsof a philosophy
of history.
But we must not leave the matter here. The thesis which I
wish to defend is not merelythat there is a clash of methods,but
that the presuppositionsof the philosopherof historydemand that
he unwarrantablytransformthe results of historiographicalin-
quiry. To show this I shall brieflytrace out two of the implica-
tions of the fact that a philosopherof historytreats the process
6 Vol. I, p. 26.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 CRITIQUE OF PHILOSOPHIES OF HISTORY 373

with whichhe deals as being more ultimatethan the nature of its

constitutiveparts.
(1) The firstimplicationof this assumptionis that the phi-
losopher of historymust confinehimselfto those processeswhich
are self-containedentities. For if he attempts to deal with
processeswhich can be influencedby events which are the consti-
tutive parts of other processes,his law of phases will be broken.
Thus, he must select as the wholes with which he deals eitherthe
whole of humanity,or civilizations conceived as single, unified
self-containedsystems.7
If a philosophyof historyis of the linear type,seekingto treat
the historyof "Human Society" as a single unifiedwhole,it comes
into flagrant contradictionwith the results of historiography.
From the point of view of historiography we do not findthat there
is any single unitaryphenomenonwhichmay be termed"Human
Society" whichhas a single historyembracingthe historiesof all
individual societies. A linear philosophyof history,as is clear,
forexample,in the case of Hegel, is forcedto leave out of its account
the historyof manypeoples and to leave out large segmentsof time
in the historyof thosepeoples of whichit does treat. The histori-
cal processas thus viewed is a pure artifact; it is a constructmade
up of bits and pieces snatchedfromthe fabricof different histories.
But even a morphologicalapproach to historypresupposes a
unity beyond that which the historiographerfinds. This is, of
course,mostobviousin Spengler,but may even be illustratedwith
referenceto Toynbee,who is assuredlymore cautious than most in
the type of unity which he ascribes to a society. Toynbee,as is
well known, takes units as large as Hellenic Society (including
Roman historyas a phase of it) and WesternSociety (frombefore
700 A.D. throughour own day) and treats them as single self-
contained wholes. It is to these enormousunits that he applies
his law of phases,and it is difficultto see how a law of historycould
be applied to much smaller units. But the necessityfor treating
historicalprocesses of this magnitudeas single wholes leads to a
conflictwithhistoriography at at least two fundamentalpoints: (a)
it assumesthat thereis an overridingunity in societies,neglecting
those diversitieswhich are equally essentialfor our understanding
of the past, and (b) it assumes an unwarranteddegree of self-con-
tainednessin societies.8
7 If he admits influences at all, it must be on the assumption that the

influencingfactor is fundamentally transformed in being absorbed into the

host civilization, which is a frequent but by no means universal occurrencein
cases of cultural diffusion (cf. below, note 9).
8 A further point concerns Toynbee's manner of choosing his units. It

is a tacit premise of his method that there is always one cultural element-

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 Intencin y extensin de Berlin

374 THE JOURNAL OF PHILOSOPHY

(a) Historiographicalinquiries seem to reveal that at any one

time there is a great diversityas well as a similarityamong the
cultural elementsin any whole which is as extensiveas was, say,
Western Society in the year 1600, or 1700, or 1800. To suppose
thatWesternSociety (as thistermis denotativelyused by Toynbee)
must be a unitary phenomenonat all times in its historyis to
sacrificethe diversitiesamong its branches in favor of whatever
similaritymay be found. And such a sacrificemust lead to a dis-
regard of the events which transpiredin large areas of Western
Society at various times,much as a linear philosophyof historyis
forced to disregard a great deal of what is usually regarded as
history. But these disregarded areas may none the less be in-
fluentialin determiningthe course of Western Society, precisely
because they are dissimilarto, and may come into conflictwith,
otherareas which lie in what we are pleased to thinkis the main
streamof events. The historiographer who seeks a concreteknowl-
edge of the serial wholewhichToynbeedenotesas WesternSociety,
will build up such knowledgethroughan attentionto the intercon-
nectionsamong all the elementsin that whole. His resultswill be
less tidy,for diversityas well as unitywill have to be stressed,but
thoughno law of phases will result,an understandingof the process
itselfwill have been gained.
(b) Toynbee's assumptionof the self-containedness of a society
is also at odds withthe resultsof historiography. He is forcedby
his belief in a law of phases to assume that each societyis a self-
contained whole. In spite of his categoriesof affiliationand ap-
parentation,each societyis a wholewhichhas its own genesis. The
continuitieswhich he acknowledgesto exist between societies are
held to be so differentin degree fromthose whichexist withinone
society,that this differenceis in fact a differencein kind.9 But
religion-which is definitive for each society. (This leads to a situation
which would tend to bafflethe historiographer: two societies may share the
stage of history at one time and place-as is the case in India (Vol. I, p. 35,
note 2) and presumably elsewhere-and yet each society is a self-contained
whole with phases through which it is destined to pass.) For the historiog-
rapher there need be no single institution definitive of the nature of any
given society, nor need an institution which is basic within the fabric of one
society be basic within the fabric of all others.
However, I shall not discuss this point at length, since it is not necessary
for a philosopher of history to hold a position similar to that of Toynbee on
this point. We need merely note that it is not unusual to find such a position
adopted. (Cf. Hegel's emphasis on the civil constitution.)
9 Vol. I, pp. 45 f.; cf. pp. 36-44, and (on diffusion) pp. 427 ff. It is of
course difficultto deal with the problems which are here involved, due to the
absence of Parts IX and X of the work, but some clues as to Toynbee's views
are given in the published volumes, e.g., Vol. IV, p. 3.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 CRITIQUE OF PHILOSOPHIES OF HISTORY 375

thisthesisis suspecton a numberof counts: first,it is valid only on

the basis of Toynbee's assumptionregarding the unity within a
society,an assumptionwhichwe have seen reason for the historiog-
rapher to challenge; second, its apparent plausibility is probably
in part due to thefactthatwe lack the materialswithwhichto trace
the continuitieswhich existed between the primitiveantecedents
of some societiesand those societiesthemselves,rather than being
due to an absence of an actual continuity;and, third,while it may
be true that at certainpoints in the "life" of a society,this society
is extremelyresistantto influencesfrom without,it is doubtful
whetherat "birth" it would possess this characteristic-thus,there
is reason to suspectthat Toynbee's theoryof the genesisof a society
suffersfromthe mistakeof assumingthat what was true at a later
stage in a processmustalso have been true at an earlierstage in it.
To summarizethe discrepanciesbetweenthe historiographer's
view of the nature of a societyand the morphologicalview held by
Toynbee,we may say that the historiographer does not assume that
a long and expansivehistoricalprocesscan be consideredto be either
a unitary or a self-containedwhole: whateverdegree of unity or
self-containedness it possesses at any one time is a subject for in-
vestigation,not a postulateof his method.
(2) The second major point at which the aim of every phi-
losopherof historyforceshim to part companyfromordinaryhis-
toriographyis to be found in his necessaryrejectionof a pluralism
of efficient causes in history.10For the law of historywhich he
attemptsto establish,which is a law descriptiveof the necessary
phase-qualitiesin any historicalprocess,musteitherbe the sole type
of efficient cause, or must overrideall other causal factors. The
providentialteleologyof Saint Augustine,the natural teleologyof
Enlightenmentphilosophiesof history,the teleologyof the Spirit
in Hegel, all supplant or dominatethe lines of efficient causation.
The same is even true of Toynbee,whose fundamentalBergsonian
categories exemplifythemselvesin the concept of challenge-and-
responseand withdrawal-and-return, categoriesdesignedto explain
change in terms of the telos of the creative personality. When
these categories are applied by Toynbee to the specific events
within the historical process, they conformto the usual efficient
10 For the historiographerthere is always a pluralism of efficientcauses,
though he may seek to find some single type of causal relationship which is
an invariant factor in all historical processes. Thus, even Marx (when in-
terpreted as attempting to find a law of invariant relationships within history
rather than a law of history) believed in a pluralism of causes in understand-
ing the specific nature of historical events. E.g., cf. Feuer: "The Economic
Factor in History," in Science and Society, Vol. IV (1940).

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
376 THE JOURNAL OF PHILOSOPHY

causal explanationsgiven by historiographers. However,the con-

ventionalexplanationsare not ultimatelysatisfactoryfor Toynbee:
they are always treated merelyas exemplars of his general cate-
gories. They must be so treated or the concreteelementsin the
process, the specificevents in their relationshipsto one another,
would be intelligiblein themselvesand the indivisibleunity of the
whole would be lost. Thus, the challengeof the climate,the chal-
lenge of geography,the challenge of barbarians,etc., are all part
of one great Challenge-and-Response for the society.
On the other hand, for the historiographer,efficientcauses
within the historical process are specific. Whatever generaliza-
tions we may make concerningthem are only abstractionswhich
formulatethe concreteinterrelationshipswhich were found. To
him it mustappear as if the philosopherof historyhas hypostatized
a generalizationand has assumed it to be an operative entity
withinhistory. Nowherewithinthe fieldwithwhichhe deals does
the historiographer findevidenceadequate to confirmthe view that
behind all specificcausal factorsthere is a single overrulingprin-
ciple which determinesthe concretecourse of the whole historical
process. If such there be, it must be discoveredthroughmeans
otherthan those of empiricalhistoriography;and if such therebe,
what the historiographerfindsto be true in historyis overriden
and transformedby it.
I have now attemptedto trace two points at which the search
for a law of historyhas led philosophersof historyto offera trans-
formingreinterpretation of the resultsof historiographicalinquiry.
The firstof these was the necessityfor consideringevery civiliza-
tion as a self-containedunit; the second was the necessityfor in-
terpretinghistoricalcausation in termsdifferent fromthose which
are empiricallyfoundto be operativein the relationsof one element
to another. Both of these errorsseem to me to involve the con-
fusionwhich I have indicated: the substitutionof a consideration
of the phases, or of otherover-allqualities, of a historicalprocess
for an examinationof thoseelementswhichmake that processwhat
it actually is.
It now remainsforus to considerwhy,if a philosophyof history
is based on such a confusionand thereforecomes into conflictwith
empiricalhistoriography, philosophiesof historynone the less have
the hold whichtheydo on men's minds.
4. The Roots of Philosophiesof History. It appears to me that
if we examinethe whole historyof the successiveattemptsto con-
structphilosophiesof historywe findthat what oftengave rise to
them was a need to come to termswith social or political crises.
Since everylaw of historyentails some messageconcerninghuman

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 CRITIQUE OF PHILOSOPHIES OF HISTORY 377

value and destiny,howevercovertthismessagemay be, a philosophy

of historycan serve as a powerfulinstrumentfor propaganda as
well as serving as a means of obtaining consolation. By setting
offthe historyof our times against the panoramic backgroundof
the past, and by projectingthe supposed tendencyof historyinto
the future,we may move othersand bolster our own faith. But
neitherof these psychologicaleffectsattest to the truth of a par-
ticular philosophyof history,nor do theyserve as warrantfor the
validity of an attemptto finda law of history. What the solution
of a currentcrisis should be is a problemwhich demands an ac-
curate knowledgeof the specificnature of that crisis and a firm
moral belief. We can obtain neitherthe one nor the other from
any law concerningthe directionaltendenciesof historyas a whole.,
When interpretedas propagandisticweapons or as means of
consolation,philosophiesof history are importantcultural docu-
ments,and thus interpretedwe can also partially understandthe
hold whichtheyhave had overmen's minds. But thereis a further
reason why philosophiesof historyhave continuallyflourishedde-
spite the increasingrapiditywith whicheach has been supplanted
by the next. This reason lies in an intellectual rather than a
specificallysocial motive. As we see in primitivecosmogoniesas
well as in the most"advanced" philosophiesof history,it is a natu-
ral questionto ask, as Miltonhas Adam ask, " How came I thus,how
here?" The need to place one's self and one's society within a
larger context,to see it as part of some whole,is perhaps an ines-
capable need. Empirical historiography mightbe thoughtto satisfy
such a need. But where, as in primitivesocieties,historiography is
almostwhollylacking,or whereit is but fragmentary-perhapscon-
finedto a few eventsin the historyof only one society-it demands
supplementationby somehistoricalhypothesis. And when,finally,
a broader historiographicalknowledgeemerges,and it is seen that
othersocieties,as well as one's own,have theirhistories,undergoing
changes and perhaps sufferingwhat seem to be declines,then the
searchfor an explanationof these changesleads one to search for a
law of the process. It is out of the great ebbs and flowsin history
that a searchfora law of historyis born. And wheneverhistoriog-
raphytendsto neglectthe largerunits of history,confiningitselfto
small segmentsof time or areas of space, or wheneverit overlooks
what I have termedthe phase-qualitiesof a historicalprocess,focus-
ing whollyon isolated details,it will call forthin reactiona tendency
to strive for a broader understanding-an understandingwhich
may take the formof a philosophyof history. In Voltaire, and
the period of the Enlightenmentgenerally,this is what occurred.
In our own day, as is evidentfromToynbee,it is occurringagain.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions
 378 THE JOURNAL OF PHILOSOPHY

But the answer to a narrowhistoriographyis a broader historiog-

raphy,not a philosophyof history.
I submitthat there is no reason to believe that the methodof
historiography mustnecessarilyfail to affordus a knowledgeof the
larger-scaleeventsin the historyof a society-a knowledgewhieb
we earnestlyand legitimatelydesire. But the historiographer will
insist that such knowledgecan only come throughfollowingthe
multitudinousinterconnections among the eventswhich were con-
stitutive of that process. The most powerful tools for the ac-
complishmentof this purpose will be hypothesesconcerningthe
relationshipsbetweenspecificinstitutions,and betweenindividuals
and the institutionsunder which they live. These tools of under-
standing can only be forgedthroughan analysis of the concrete
cf. Escuela
elementswithinany historicalprocess. And when today,as prob-
de los anales ably neverbefore,historiographers are attemptingin theirpractice
(regardlessof the theorieswhichtheyhold) to findthe actual con-
nectionsbetween events,rather than merelyreportingthem,it is
a sign of weaknessratherthan of strengthto turn fromthe task of
seeking out the causal relationshipsbetween the events within
history,in favor of attemptingto buttressour hopes and stave off
our despair throughan appeal to a postulated law of history.
MAURICE MANDELBAUM
DARTMOUTH COLLEGE

THE HISTORIAN AND TRUTH'

H ISTORY is a science. It seeks the "truth." But science is
in quest of laws; historyis contentwith describingthe par-
ticular. What kind of particular? What is in the particular?
To the historianthe truth of a descriptionmeans in practice
its concordancewith a kind of reality, the "historical reality."
To himthe problemof truthis the problemof this historicalreality
and its specificcharacter. To him this question is independent
from and prior to the other problem-how to verify the truth.
The historiancan not detach fromhis subject-matter, ere he starts,
those areas or layers that lend themselvesconvenientlyto specific
demandsof a scientificmethodand forgetabout the others. Thus
he can not share the general belief in the sovereigntyof the scien-
tificmethodover the subject-matter. To him the methoddoes not
determinethe subject-matter;the subject-matterdeterminesthe
method. He can not assume beforehandthat the joints at which
the subject-mattershould be divided are those suggested by his
1 Paper read beforethe Conferenceon Methodsin Philosophyand the
Sciences,New Schoolfor Social Researeh,December,1947.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:08:38 PM

All use subject to JSTOR Terms and Conditions