MILL'S ARGUMENT FOR THE PRINCIPLE OF UTILITY 339

means to something admitted to be good without proof."4 It is not clear

what Mill intends to exclude, but several portions of his argument can be
recast into deductive form.5 He could, for instance, have argued as follows:
(1) If all nonultimate goods are means to some end, E, and if E is ad-
mitted to be good without proof, then E is an ultimate end.
(2) All nonultimate goods are means to the maximum happiness, and
the maximum happiness is admitted to be good without proof.
Therefore,
(3) The maximum happiness is an ultimate end.
Support for the premises in this argument would surely generate rational
support for Mill's utilitarianism. But he does not argue in this way; his proof
is more circuitous, more complex. He does not claim that the maximum
happiness is admitted to be good without proof. Instead, he provides a
proof for why it should be admitted to be good.
The key passage is the third paragraph:
The only proof capable of being given that an object is visible is that people actually
see it. The only proof that a sound is audible is that people hear it; and so of the other
sources of our experience. In like manner, I apprehend, the sole evidence it is possible
to produce that anything is desirable is that people do actually desire it. If the end
which the utilitarian doctrine proposes to itself were not, in theory and in practice,
acknowledged to be an end, nothing could ever convince any person that it was so.
No reason can be given why the general happiness is desirable, except that each per-
son, so far as he believes it to be attainable, desires his own happiness. This, however,
being a fact, we have not only all the proof which the case admits of, but all which it
is possible to require, that happiness is a good, that each person's happiness is a good
to that person, and the general happiness, therefore, a good to the aggregate of all
persons. Happiness has made out its title as one of the ends of conduct and, conse-
quently, one of the criteria of morality.6

4Mill, p. 7.
The philosophical difficulties in the argument are very great and can be highlighted with
this sort of reconstruction. And these most severe problems cannot be evaded by urging that
Mill was not trying to provide a strict proof. However this may be, Mill almost certainly
overstated the point when he said that ". . . questions of ultimate ends do not admit of proof,
in the ordinary acceptation of the term. To be incapable of proof by reasoning is common to
all first principles, to the first premises of our knowledge, as well as to those of our conduct."
(Mill, p. 44) It appears to be unquestionable that, whatever type of proof is being offered in the
famous chapter four, it is a proof by reasoning in which Mill was trying to establish ra-
tionally the principle of utility, the first principle of conduct. Several writers have stressed that
Mill was not offering a strict, direct deductive proof. See especially Hall, op. cit.; D. D.
Raphael, "Fallacies in and about Mill's Utilitarianism," Philosophy, vol. 30, (October, 1955),
pp. 344-57; and S. A. Moser, "A Comment on Mill's Argument for Utilitarianism," Inquiry,
vol. 6 (1963), pp. 308-18.
6 Mill, pp. 44-45.

This content downloaded from 157.39.76.95 on Thu, 10 Sep 2020 17:57:53 UTC
All use subject to https://about.jstor.org/terms
 340 PHILOSOPHY AND PHENOMENOLOGICAL RESEARCH

These legendary sentences express Mill's conviction that he has supported

two central claims: (a) Individual happiness is a good and (b) The general
happiness is a good.7
The utilitarian doctrine includes four other crucial elements: (c) Right
acts are conducive to what is good, (d) The good is the desirable, (e) Hap-
piness is the only intrinsically desirable human end, and (f) The general hap-
piness is the ultimate moral end. Mill's proof does not involve support for
(c). His discussion proceeds on the assumption that the correct moral theory
is a teleological one in which actions are to be evaluated according to the
supreme good.8 He writes also as though (d) is true; he does not argue that
'good' means "desirable," or that whatever is desirable must also be good.
I shall not question these features of utilitarianism. Mill argues for (e) with
the contention that the constraints of human nature require that anything
desired is desired as a means to or as a part of happiness.9 This is then used
to support (f). Near the end of chapter four, Mill reiterates the view that
t. . . nothing is a good to human beings but in so far as it is either itself
pleasurable or a means of attaining pleasure or averting pain," and imme-
diately adds, "But if this doctrine be true, the principle of utility is
proved. " '0

I want to show why Mill's argument cannot support the view that the
general happiness is the ultimate moral end. The next three sections are
devoted to the preliminary stage in which he argues (a) that the individual
happiness is good. I shall consider three interpretations of the startling
claim that ". . . the sole evidence it is possible to produce that anything is
desirable is that people do actually desire it." In sections V and VI I turn to
the argument that the general happiness is desirable. I argue that Mill's
evidentiary criterion is fatal to the defense of his utilitarian end.
' The term 'general happiness' will be interpreted to mean "maximum happiness." Mill
seems to be aware that he must show that the ultimate moral end involves the greatest at-
tainable degree of happiness. Some have interpreted Mill to have meant that the moral end is
the greatest average happiness, rather than the greatest absolute total. It is not necessary to
argue the matter here, since my evaluation of Mill's argument is not affected by it.
'I wish to avoid the controversy over whether Mill is an act or rule utilitarian. My argu-
ment is not affected by this issue. On these matters, see J. 0. Urmson, "The Interpretation of
the Moral Philosophy of J. S. Mill," The Philosophical Quarterly, vol. 3 (January, 1953), pp.
33-39; J. D. Mabbott, "Interpretations of Mill's Utilitarianism," The Philosophical Quarter-
ly, vol. 6 (1956), pp. 115-20; and Maurice Mandelbaum, "Two Moot Issues in Mill's
Utilitarianism," In Schneewind.
9 Mill, p. 48.
10 Mill, p. 51.

This content downloaded from 157.39.76.95 on Thu, 10 Sep 2020 17:57:53 UTC
All use subject to https://about.jstor.org/terms
 MILL'S ARGUMENT FOR THE PRINCIPLE OF UTILITY 341

II

Mill's argument that (individual) happiness is desirable has often been

thought to depend on a very strong link between desire and desirability. On
one view, something's being desired is a sufficient condition for its
desirability. The "sole evidence" that something is desirable is quite strong:
a thing cannot be desired unless it is also desirable. The analogy with
'visible' and 'audible' is thus quite apt, for being seen and being heard are
sufficient for things also to be visible and audible." This interpretation
allows a simple, direct argument:
(1) If something is desired, then it is desirable.
(2) Happiness is desired.
Therefore,
(3) Happiness is desirable.
This conclusion, together with two premises based on Mill's discussion of
the nature of human desire, quickly yield the conclusion that happiness is
the only desirable thing. There is evidence that Mill accepts
(4) If something is desirable, then it is capable of being desired,
and
(5) If something is capable of being desired, then it is happiness. The
latter claim is derived from Mill's view that happiness is the only thing per-
sons are capable of desiring intrinsically. 12 One can now conclude that hap-
piness is desirable, and that nothing else is desirable.
The construction of such simple proofs may cast doubt on this "suffi-
cient condition" interpretation (SC). It is rather surprising that Mill's seem-
ingly minimal initial moves, especially the mild-sounding claim that ". . .
the sole evidence it is possible to produce that anything is desirable is that
people do actually desire it," enable him to establish conclusively that hap-
piness is desirable. Interpreted according to SC, the argument is open to a
simple objection. The first premise appears to be obviously false. Mill must
Several writers have interpreted the text along these lines. Norman Kretzmann, while
acknowledging that he may be departing from Mill's actual intent, says that "If anything is
desired in such a way as to occasion some overt reaction on the part of the normal desirer for
the thing in question, then that thing is desirable." It should be noted that Mill places no such
restrictions or qualifications on the sort of desiring that is relevant; nor does he insist on any re-
quirements of accompanying reactions by normal desirers. Presumably, however, since he
believes human nature to be such that all persons desire their own individual happiness, almost
any individual's desire for his happiness would satisfy Kretzmann's conditions on the type of
desire which entails desirability. Somewhat different restrictions for the relevant desires are im-
posed by Carl Wellman. See Kretzmann, Gorovitz, p. 237. Cf. Carl Wellman, "A Reinter-
pretation of Mill's Proof," Ethics, 69 (1959), esp. pp. 270-72.
12 Mill, pp. 48-49.

This content downloaded from 157.39.76.95 on Thu, 10 Sep 2020 17:57:53 UTC
All use subject to https://about.jstor.org/terms
 342 PHILOSOPHY AND PHENOMENOLOGICAL RESEARCH

have recognized that a thing's being desired is not sufficient for it to be

desirable. His writings may be understood as aimed at getting DeoDle to stone
desiring undesirable things and to start desiring desirable ones.
If SC were correct, there would be immense conflict and confusion
among things warranting the adjective 'desirable.' Many human goods
would be achievable only at the expense of other goods. If A desires x and B
desires y, then x and y are desirable. It does not matter what x and y are,
how undesirable they are relative to other desires, or whether they conflict
with each other. But there is a good reason to believe that Mill would have
realized that either x or y could be undesirable though desired by A and B.
Many individuals' personal ends are not consonant with the ultimate end of
morality, the most important test of all human conduct (and thus of all
human ends). If they lack such consonance, surely they are not desirable.

III

There is an obvious alternative to the SC interpretation. The visibility

analogy does not commit Mill to the view that whatever is true of the rela-
tion between the seen and the visible is also true of the relation between the
desired and the desirable. He could have held that, though being seen is suf-
ficient for being visible, being desired is not sufficient for being desirable.
He might have agreed that, though the knowledge that something is seen
provides conclusive evidence that it is visible, knowledge that something is
desired provides only inconclusive evidence that it is desirable (even if this is
the only evidence it is possible to produce). According to this "inconclusive
evidence" interpretation (IE), the visibility analogy indicates some eviden-
tiary connection between desiredness and desirability. One may not con-
clude that an individual's happiness is desirable simply because he desires it.
The only evidence for its desirability would be just evidence; it does not
allow the use of conditionals, like (1) according to SC, with 'desires' in the
antecedents and 'desirable' in the consequents.
Mill's statement that ". . . the sole evidence it is possible to produce that
anything is desirable is that people do actually desire it" is comparable to
"The sole evidence it is possible to produce that a defendant is guilty is his
confession" or "The sole evidence it is possible to produce that a patient
has cancer is the blackened X-ray." The condition which as evidence is not
invariably accompanied by that for which it is evidence. Such an interpreta-
tion invites such questions as "How strong, or good, is such evidence?" and
"How great is the probability that that for which there is such evidence will

This content downloaded from 157.39.76.95 on Thu, 10 Sep 2020 17:57:53 UTC
All use subject to https://about.jstor.org/terms
 MiLL's ARGUMENT FOR THE PRINCIPLE OF UTrLITY 343

accompany it?" Here the text is unhelpful in determining, in particular

cases, the strength of the evidence and the resultant degree of desirability.
But there are more serious difficulties with IE. One is left with in-
conclusive evidence that each of a vast variety of things is desirable. Sup-
pose that each of several persons has confessed to a crime, but that there is
no other evidence against any of them. If one knows that not all are guilty
(suppose the crime to be such that only one person could have committed it),
then there will be only partial, perhaps very weak evidence that any par-
ticular individual is guilty. The case is analogous to those in which several
persons desire, as parts of their separately specified final goods, things not
all of which can be obtained. It is not incoherent to contend that each of
many conflicting goods is desirable; but one will have little guidance for
what to do if not all of them are obtainable, if the only evidence for each is
its being desired, if this evidence is inconclusive, and if there is no further
criterion to supply a basis for comparing them. It is tempting to suggest that
the principle of utility is such a criterion. But it would be circular to intro-
duce the principle at this preliminary stage of an argument designed to show
it to be true.
There is further reason to reject the IE interpretation. Mill states his in-
termediate conclusion much more strongly than the account allows. He says
that "Happiness has made out its title as one of the ends of conduct, one of
the ends of morality."'3 He has not yet argued that happiness is the only
possible object of human desire, but he writes as if he has shown much more
than that there is inconclusive, perhaps only weak, evidence that happiness
is good. His remark suggests that the reasons for thinking happiness to be
desirable are much stronger than the reasons for thinking the defendant to
be guilty or the patient to have cancer. If IE is correct, Mill greatly exag-
gerates his allowable conclusions.

IV

The third interpretation has recently been very popular. This view con-
siders the relation between the desired and the desirable to be weak, but uses
Mill's psychological doctrines to obtain a strong conclusion that happiness
is desirable. The argument is that since happiness is the only object of
human desire-the only thing persons can and do desire-it must be
desirable. The central claim is that happiness is the only legitimate can-
didate for what is desirable; it is the only thing which could be desirable.

13 Mill, p. 45.

This content downloaded from 157.39.76.95 on Thu, 10 Sep 2020 17:57:53 UTC
All use subject to https://about.jstor.org/terms