/. Value Inquiry 1 7 :3 3 -5 2 (1983).
© 1983 Martinus N ijh o ff Publishers, The Hague. Printed in The Netherlands.
UTILITY AND EQUALITY: SOME NEGLECTED PROBLEMS
TOM REGAN
N orth Carolina State University
I
U tilitarianism frequently is attacked on the grounds th a t is can sanction inequitable
distributions o f harm s and benefits. Since the end set by the th eo ry is aggregative,
critics argue th a t it could require th a t some few individuals be m ade to suffer a lo t
so th a t the rest o f hum anity m ight individually gain a little, th e aggregative gain by
the m any m ore than com pensating for the grievious losses o f the few. Those w ith
egalitarian m oral leanings find such an outcom e otiose and have sought a more
equitable principle o f distributive justice th an they th in k u tilitarian th eo ry p ro ­
vides. F or their part, utilitarians have devised various replies, and th e debate has
co n tinued unabated. A nd likely will. B ut there are oth er notions an d principles o f
equality in addition to, and m ore basic th an , those th a t figure in th e debate over
equal distribution. This latter debate concerns the justice o f distributing harm s and
benefits betw een various individuals, assuming th a t, prior to distributive concerns,
the affected individuals have been treated equitably. It is this predistributive sense
o f equality th a t B entham m arks by his fam ous declaration th a t ‘each is to co u n t for
one, no one for m ore than one,’ an understanding o f equality w hich, in Peter
Singer’s w ords, ‘utilitarians, from Jerem y Bentham to J.J.C . Sm art, take . . . as
axiom atic . . . in deciding m oral issues.’1 In their rush to air w hat th ey regard as the
unsavory m oral im plications o f utilitarianism concerning distributive justice, u tili­
tarianism ’s critics have failed to give equal tim e to th e problem th is Bentham ite
n o tio n o f equality poses for utilitarian theory. The present discussion attem p ts to
partially rem edy this oversight.
II
Singer should know w hereof he speaks since he is him self a self-avowed u tilitarian ,2
and the B entham ite n o tio n o f equality bulks very large in his th o u g h t. It will be in ­
structive, therefore, to begin by asking w hat equality m eans in Singer’s hands. It is
n o t factual equality. When, for exam ple, we say th a t all hum ans are equal, we do
34
n o t m ean th a t everybody has the same abilities, hair color, or num ber o f arms. If
this were w hat we m eant, th en the declaration o f equality w ould be paten tly false.
Instead, the kind o f equality Singer has in m ind is expressed by ‘th e basic m oral
principle o f equality,’3 a principle th a t is ‘n o t a description o f an alleged actual
equality am ong hum ans: it is a prescription o f how we should trea t hum ans’4 and,
Singer argues, m any nonhum an animals besides.5 W hat this principle prescribes is
th a t ‘the interests o f every being affected by an action are to be tak en in to account
and given the same weight as the like interests o f any oth er being.’6 For co n ­
venience’s sake let us refer to the principle ju st quoted as the equality principle.
T hat principle, then, according to Singer, is (a) prescriptive, n o t descriptive; (b)
basic; (c) m oral; (d) concerns the range o f interests to be considered (‘the interests
o f every being affected by an action are to be taken in to acco u n t’); and (e) pre­
scribes th a t equal interests be co unted equally.
Two options are available to Singer, given his view th a t the equality principle is a
basic m oral principle. E ither he m ay regard it as a basic m oral principle in the logical
sense o f ‘basic’, m eaning th a t it cannot be derived from any oth er m oral principle,7
or he m ay regard it as basic in a nonlogical sense, m eaning th a t th o u g h it is derivable
it is o f especially crucial m oral im portance. Surprisingly, some o f w hat Singer w rites
strongly suggest th a t he inclines tow ard the form er o p tio n . Thus, for exam ple, he
w rites the following:
The only principle o f equality I hold is the principle th a t the interests o f
every being affected by an action are to be taken in to account and given the
same w eight as the like interests o f any other being . . . (U) tilitarianism pre­
supposes this principle.8
This last sentence com m its Singer to the view th a t a logical relationship holds
betw een the principles o f equality and utility. When he says th a t utilitarianism
presupposes the equality principle he implies th a t unless we assume the validity o f
the equality principle utilitarianism has no m oral or logical footing. W hat the quoted
passage implies, in short, is th a t rather th an the m oral principle o f equality de­
pending on th a t o f utility , the reverse is true: The principle o f u tility depends on
the m ore basic m oral principle o f equality.
No consistent utilitarian can believe this. If utility is, as it m ust be for the
utilitarian, and as Singer says it is, ‘th e sole (m oral) basis o f m o rality ,’9 th en all
oth er m oral principles m ust be derivable from it and it in tu rn can n o t presuppose
any other m ore or equally basic m oral principle. For a utilitarian to argue otherw ise
is to render utilitarianism incoherent. Singer m ust, therefore, avoid th e first o p tion
at all costs.
The second option prom ises to be m ore attractive.10 If, as this o p tio n m aintains,
the equality principle, like all other m oral principles save th a t o f utility itself, is
derivable from the principle o f u tility , th en selecting this alternative a t least perm its
one to advance w hat looks to be a coherent version o f utilitarianism . B ut the a t­
tractiveness o f this second option is illusory. The attem p t to ground the equality
35
principle in th a t o f u tility involves a gross distortion o f th e n o tio n o f equality as it
applies to interests. The equality or inequality o f the interests o f tw o individuals, A
and B, depends on how im p o rta n t their respective interests are to th e m , A’s interest
being equal to B’s if their interests have like im portance to A and B, respectively,
A ’s and B’s interests being unequal if the case is otherw ise.T he equality or in ­
equality o f their interests cannot depend on how th e interests o f others will be
affected if A ’s and B’s interests are considered as equal or unequal. It this were so,
th en we w ould be free to regard th e same interests o f A and B as equal at one tim e
and as unequal at another, because the interests o f others happened to be affected
differently by regarding A’s and B’s interests differently at these different tim es.
This is to m ake a shambles o f the n o tio n o f equality as it applies to interests. A nd
y et this is precisely w here we are led in our understanding o f equal interests, if the
obligation to respect the equality principle is derived from the principle o f utility.
F or the u tility o f counting A’s and B’s interests as equal can vary from case to case,
even if their interests themselves do n o t. Thus, if u tility be our guide, we are per­
m itted to co u n t the same interests as equal in one case and as unequal in another.
This is to distort the concept o f equal interests beyond recognition.
The upshot o f th e foregoing does n o t bode well for a position like Singer’s. The
dilemma th a t m ust be faced is th a t o f either setting fo rth a view o f th e relationship
betw een the m oral principles o f u tility and equality th a t renders utilitarianism in ­
consistent (the argum ent given against the first option) or setting fo rth an account
o f the relationship betw een these tw o principles w hich avoids th e charge o f incon­
sistency b u t only at the price o f grossly distorting th e concept o f equal interests
(th e argum ent given against the second option). It should be em phasized th a t the
choice betw een these tw o alternatives cannot be avoided by a u tilitarian w ho, like
Singer, m aintains th a t the equality principle is a m oral principle. F or th en th a t
principle m ust either be conceived to be underived, in w hich case th e charge o f
inconsistency will apply, or derived, in w hich case the charge o f d istortion will p re­
vail. In either case, therefore, the equality principle can find no hom e w ithin u tili­
tarianism , if, like Singer, utilitarians view the equality principle as a m oral principle.
Ill
T here is an obvious way for utilitarians to try to avoid this outcom e. This is to
regard the equality principle as a fo rm a l m oral principle rath er th a n as a substantive
one — as a principle, th a t is, th a t does n o t itself lay dow n a m oral obligation co n ­
cerning w hat we are to do b u t one w hich sets fo rth a co n d itio n th a t m ust be m et by
any substantive m oral principle th a t does lay dow n such an obligation. On this view
o f the equality principle, in other w ords, this principle incorporates at least a partial
te st fo r determ ining w hen a principle is a m oral as distinct from a nonm oral principle.
Principles th a t pass this te st to th a t e x te n t qualify as m oral principles; those th a t
fail it lack th a t status. T hus, th e principle o f u tility qualifies, since it enjoins us to
consider the interests o f everyone affected and to co u n t equal interests equally.
36
E thical egoism, on the other hand, since it neither bids us to canvass th e interests o f
all affected parties nor to co u n t equal interests equally, fails to have the status o f a
m oral principle.
To regard the equality principle as a form al m oral principle does avoid the
dilemma fatal to those utilitarians w ho, like Singer, view th a t principle as a sub­
stantive m oral principle. The trouble for this defense o f utilitarianism is th a t it is
false th a t the equality principle is a form al m oral principle, in the sense explained.
Recall th a t th a t principle (i) concerns the range o f interests to be considered (‘the
interests o f every being affected by an action are to be taken in to ac co u n t’), and (ii)
prescribes th a t equal interests be co unted equally. I f th e equality principle is said to
be a form al m oral principle, th en no principle th a t fails to com ply w ith (i) and (ii)
could qualify as a m oral principle. This is n o t tru e. An extrem e deontologist, such
as K ant arguably is, who thinks th a t considerations ab o u t the interests Singer and
o th er utilitarians have in m ind — nam ely, w hat it is th a t individuals happen to w ant
o r desire — are irrelevant to the determ ination o f where our m oral d u ty lies, may
advance a defective vision o f m orality. But n o t even K an t’s harshest critics will
m aintain th a t the categorical im perative does n o t have th e status o f a m oral prin­
ciple: M istaken it m ay be, b u t at least it m ust be fo u n d to be m istaken as a m oral
principle. Since, then, principles w hich fail to com ply w ith (i) and (ii) m ight n o n e­
theless have the status o f a m oral principle, it is false th a t th e equality principle is a
form al m oral principle, in the sense explained.
IV
But if the equality principle is n o t a form al m oral p rin cip le; and if, for the reasons
given in th e earlier discussion o f Singer’s views, this principle cannot reasonably be
viewed as a substantive m oral principle w ithin utilitarian th e o ry ; th en w hat is a
utilitarian to do? The B entham ite proviso, ‘each to co u n t for one, no one for m ore
th an o n e,’ allegedly ‘axiom atic’ for utilitarianism , paradoxically seems n o t to be
able to find a place w ithin th a t theory.
S hort o f abandoning utilitarianism , there rem ains one final alternative. This is
for the utilitarian to regard the equality principle as a conditional form al principle.
To say it is conditional is to say th a t it becom es operative only if certain conditions
are m et. Singer provides a clue as to w hat these conditions m ight be w hen he
observes th a t each o f us has a ‘very natural concern th a t (o u r) ow n interests are to
be looked afte r.’11 Suppose we acknowledge this natural propensity. Then we shall
surely w ant others to take note o f our interests and weigh them equitably — th a t is,
n o t to discount th e im portance o f our interests ju st because they are n o t their
interests. If m y interest in x is the same as yours, th en I shall w ant y o u to give the
same w eight to m y interest as you give to yours. I f you do n o t, th en y o u will have
lapsed back into egoism, w hich is ta n tam o u n t to saying th a t you are n o t taking th e
m oral p o in t o f view. F or to take th a t p o in t o f view requires th a t one be willing to
universalize one’s judgm ents o f value. This requirem ent falls to m e, how ever, n o t
37
ju st to others. Thus, if I place a certain value on m y interest in x , I m ust recognize
the like value o f an yone’s similar interest; and if I w ould have others take n o te of
m y interests, because th ey are im p o rtan t to m e, th en I m ust take n o te o f theirs,
recognizing th a t they are im p o rtan t to them . In this way, th en , we arrive at the
equality principle, only now viewed as a conditional form al principle. L et us refer
to this way o f viewing equality as ‘conditional eq u ality ’ and express this principle
as follows: I f I w ould have others consider m y interests and co u n t th em equitably,
and if I am to take th e m oral point o f view, then I com m it m yself to considering
the interests o f all those affected and to counting equal interest equally.
K antians and other extrem e deontologists likely w ould look w ith disfavor even
on this way o f viewing the equality principle. If the interests we h appen to have are
irrelevant to the determ ination o f w here our m oral d u ty lies, th ey are apt to say,
th en we ought n o t to m ix considerations ab o u t interests w ith taking the m oral
p o in t o f view. But suppose we bypass this line o f debate, im p o rta n t th o u g h it is, on
the present occasion and assume th a t some individual (A) n o t only understands b ut
actually accepts conditional equality; th a t is, assume th a t A w o u ld have others
consider his interests and weigh them equitably, th a t A w o u ld take th e m oral p o in t
o f view, and th a t, as a consequence o f b o th these assum ptions, A does agree th a t
he, to o , m ust canvass th e interests o f others and co u n t equal interests equally.
G ranting all this, we m ay th en go on to ask ab o u t how A m ay view th e principle o f
u tility .12 Tw o options are again at hand. The first is th a t acceptance o f conditional
equality is consistent w ith utilitarianism . This is a w eak o p tio n because on its face
it provides no reason for selecting utilitarianism in preference to o th er substantive
views o f m orality th a t also are consistent w ith acceptance o f conditional equality —
e.g., Ross’ view th a t there are m any rules o f prim a facie duty. Weak though this
o p tion is, even it is to o strong for certain varieties o f utilitarianism , as will be shown
in section VI. The second, strong option is th a t acceptance o f th e conditional
equality principle logically com m its one to the principle o f u tility . This strong
o p tio n can itself take tw o form s, SOj m aintaining th a t acceptance o f con­
ditional equality com m its one to accepting the principle o f u tility to th e exclusion
o f any other substantive m oral principle, and S 0 2 m aintaining th a t acceptance o f
conditional equality com m its one to accepting utility as a m inim al m oral principle
while leaving open the possibility th a t there m ay be o th e r m oral principles besides
th a t o f utility w hich one m ay also consistently accept. R.M. Hare possibly accepts
S O j. I say ‘possibily’ because it is n o t altogether clear to me w hat H are’s considered
view is in this reg ard .13 Singer, how ever, evidently accepts S 0 2 . He w rite s:14
The utilitarian position is a m inim al one, a first base w hich we reach by
universalizing self-interested decision m aking. We can n o t, if we are to th in k
ethically, refuse to take this first step.
Here Singer allows th a t there m ay be other m oral principles in addition to u tility ,
while m aintaining th a t we m u st accept utility if, starting w ith our individual
interests, we take the m oral p o in t o f view. F or to take th a t p o in t o f view, starting
38
w ith our self-interest, com m its us to the conditional equality principle. A nd this in
tu rn com m its us to utilitarianism . At least this appears to be b o th a natural and fair
in terpretation o f the crucial argum ent for accepting utilitarianism in the following
passage.
Suppose I begin to th in k ethically, to the ex ten t o f recognizing th a t m y own
interests cannot co u n t for m ore, sim ply because th ey are m ine, th an the
interests o f others. In place o f m y ow n interests, I now have to take account
o f the interests o f all those affected by m y decision. This requires me to
weigh up these interests and adopt the course o f action m ost likely to m axi­
mize th e interests o f those affected. Thus I m ust choose the course o f action
w hich has the best consequences, on balance, for all affected. This is a form
o f utilitarianism . It differs from classical utilitarianism in th a t ‘best conse­
quences’ is u n derstood as m eaning w hat, on balance, furthers th e interests of
those affected, rather than m erely w hat increases pleasure and p ain .15
It is by m eans o f the argum ent ju st quoted, then, th a t Singer evidently believes we
are able to support S 0 2 — the view, again, th a t acceptance o f the principle o f con­
ditional equality com m its one to accepting utility at least as a m inim al, though n o t
necessarily as th e only, substantive m oral principle. Since it cannot be true th a t
acceptance o f conditional equality com m its one to accepting the principle o f utility
as the o n ly substantive m oral principle unless acceptance o f conditional equality
com m its one to accepting the principle o f utility as a m oral principle, to show th a t
S 0 2 is n o t true will be ta n tam o u n t to showing th a t SO x is n o t tru e either.
Now, before we can show why S 0 2 is false, it is necessary to m ake some pre­
lim inary rem arks ab o u t the need u tilitarian’s have for a th eo ry o f value and, more
particularly, about the role o f interests in setting fo rth such a th eo ry . These remarks
are offered in Section V. Though w hat is argued there will be used, in Sections VI
and VII, to close the door to some versions o f utilitarianism , th ey also leave a
possible opening for another version, w hich is discussed in Sections V III and IX.
The m aterial in Section V thus has both a critical and a potentially liberating
function to play vis-a-vis utilitarian theory.
V
U tilitarians, o f whatever stripe, are required to com plem ent their th eo ry o f obli­
gation w ith a theory o f value, since w ith o u t a theory o f the la tte r k ind the u tili­
tarian injunction to bring ab o u t ‘the best’ consequences will be em p ty and thus will
provide no m oral direction. As is well know n, the h isto ry o f u tilitarian th eo ry is
characterized by sharp disagreem ent over w hat has intrinsic value an d disvalue, the
classical utilitarians, B entham and Mill, for exam ple, holding th a t pleasure and pain
are the only intrinsic good and evil, respectively, while M oore’s so-called ‘ideal
utilitarianism ’ holds th a t there is a plurality o f intrinsic goods and evils w hich are
n o t reducible to pleasure and pain.16 O ur interest on th e present occasion does n o t
39
lie in the attem p t to decide who is right in this regard. R ather, the p o in t to be m ade
is th e uncontroversial one, th a t given any version o f utilitarianism we are en titled to
require th a t its advocates supply us w ith some account o f value.
The second po in t is this. I f acceptance o f conditional equality is supposed to
com m it one to acceptance o f u tility , and in view o f th e central im portance o f the
n o tion o f interests in the form ulation o f the conditional equality principle, one
w ould expect th a t it m ust be this same n otion th a t is central to the account of
value th a t form s the logical background o f this argum ent and inform s the utilitarian
injunction to bring ab o u t ‘the best’ consequences. This m ust be the case since, if
the needed account o f value tu rn ed o u t to be indepen d en t o f the n o tio n o f interests,
so th a t w hat co unted as ‘the b est’ consequences was quite ind ep en d en t o f anyone’s
having any interest whatsoever, then it w ould be altogether m ysterious, n o t to say
dow nright unintelligible, how anyone could th in k th a t by accepting th e conditional
equality principle we com m it ourselves to accepting th e principle o f u tility . In
order for this com m itm ent to seem even plausible, therefore, one m ust assume th a t
the co n ten t o f ‘the best’ consequences is to be cashed in term s o f the n o tio n o f
interests th a t figures so prom inently in the conditional equality principle. The
reasonableness o f this expectation is borne o u t by w hat Singer says, since ‘the b est’
consequences, in his view, are n o t to be identified w ith m axim izing pleasure and
reducing pain, say, b u t are to be ‘understood as m eaning w hat, on balance, furthers
the interests o f those affected.’
But the n o tio n o f having an interest is am biguous.17 To say th a t some individual
(A) has an interest in som ething (x ) m ight m ean either th a t (1) A w ants, desires,
prefers, lusts after, etc. x , or th a t (2) having x w ould p ro m o te A ’s welfare or w ould
co n tribute to A ’s good or well-being. Suppose we call th e form er sense o f having-aninterest the preference sense and the la tte r th e welfare sense. The tw o certainly are
distinct. An alcoholic m ay w ant a drink, and thus have an interest in a drink, in the
preference sense, despite the fact th a t having a drink w ould n o t be in his interests,
so th a t he has no interest in it, in the welfare sense, and a heart p atien t m ay have an
interest in a salt-free, low -fat diet, in the welfare sense, despite the fact th a t he dis­
likes this diet and so has no in terest in it, in the preference sense. Given the am ­
biguity o f talk o f ‘having an in te rest’; and given th a t, as ju st illustrated, individuals
can have an interest in som ething in one sense and n o t have an interest in the same
thing in th e o th er; the question arises, Which sense o f ‘having an in te re st’ is at work
in attem p ts to show th a t anyone who accepts the conditional equality principle,
w ith its em phasis on interests, is com m itted to the principle o f u tility , when th a t
principle is u n derstood as directing us to bring about ‘w hat, on balance, furthers the
interests o f those affected’? This question is o f fundam ental im portance since
w hich sense is involved will determ ine the account o f value we are being asked to
accept, either w hat we shall call a preference theo ry, according to w hich, roughly
speaking, value is the object o f any preference, o f a welfare th e o ry, according to
w hich w hat has value is w hat contributes to an individual’s good, w elfare, or
well-being.
The evidence at han d supports interpreting b o th Singer and Hare as exponents o f
40
a preference theory. In Singer’s case this finding is supported by th e very nam e he
gives to his version o f utilitarianism — nam ely, ‘preference utilitarianism ’18 — and
by w hat he states ab o u t interests generally, as w hen, for exam ple, he w rites, w ithout
dissenting, th a t we m ay ‘define “ interests” broadly enough, so th a t we count any­
thing th a t people desire as in their interests (unless it is incom patible w ith another
desire or desires) . . . ,19 In the case o f Hare, the evidence is less palpable, b u t it is
especially notew orthy th a t w hen he gives his argum ent for utilitarianism , an
argum ent w hich, he observes, ‘em phasizes th a t equal interests are to be counted
equally,’ his argum ent u ltim ately appeals to w hat the individuals affected by an
a ctio n ’s outcom e w ould w ish,20 w hich strongly suggests th a t the argum ent relies on
w hat th e affected individuals have an interest in, in the preference, as distinct from
the welfare, sense o f ‘have an in terest.’ M oreover, Singer, in com m enting on his
argum ent for utilitarianism , the one q uoted in the previous section, observes th a t
this ‘argum ent for utilitarianism based on interests or preferences owes m uch to
H are’s ‘Ethical T heory and U tilitarianism ’, although it does n o t go as far as the
argum ent in th a t article.’21 In interpreting Hare as viewing interests as preferences,
therefore, we reach the same interpretive finding as Singer, and though our agree­
m ent w ith Singer does n o t establish th a t we are right, it at least shows th a t in ter­
preting Hare as we do is n o t idiosyncratic.
There are familiar objections th a t can be raised against the adequacy o f a pre­
ference th eo ry o f value. Suppose we assume th a t a given individual (A) prefers
som ething (x). A t the very m ost w hat follows from this is th a t A places some
positive value on x . W hat does n o t clearly follow is th a t x is valuable. If I prefer
yo u r untim ely death, your public shame, or the ruination o f y o u r marriage or
health, then, let us assume, I place a positive value on any or all o f these o u t­
comes. But it is difficult to see how , assuming ju st this m uch, it follows th a t any or
all these outcom es are themselves valuable. To show their value, it is fair to say, will
require considerably m ore than showing th a t I personally prefer them , or even th a t
m ost people do. So there is, at the outset, a serious question to be raised ab o u t the
adequacy o f any preference theory o f .value. This question is m entioned b ut n ot
explored here, n o t because it is unim portant, b u t because our present interest lies in
another quarter. O ur present concern lies in asking about th e possible logical re­
lationship betw een accepting the conditional equality principle, on the one hand,
and, on th e other, accepting the principle o f utility , at least as a m inim al m oral
principle. A nd this is a question th a t can be exam ined independently o f our con­
testing the adequacy o f a preference theory o f value, considered in its own right. It
is this question to w hich we shall now direct our atten tio n , rem arking, before
turning to our exam ination, th a t w hen, in the n ex t tw o sections, the n o tio n o f
having an interest is used, it is the preference sense th a t is m eant. This will allow
econom y o f expression and should n o t be a cause o f m isunderstanding. The second
sense o f having an interest - the welfare sense - will be explored m ore fully in
sections V III and IX.
41
VI
There are alternative ways o f viewing the possible relationship betw een value and
preferences.22 In its m ost am bitious form th e preference th eo ry holds th a t (a) given
any x , x has value for a given individual (A) i f and o n ly i f A prefers x . Less
am bitious form s are (b) x has value for A i f A prefers x (th a t is, being the object o f
A’s desires, etc., is a sufficient condition o f x ’s having value for A), and (c) x has
value fo r A o n ly i f A prefers x (th a t is, being the object o f A’s desires, etc., is a
necessary condition o f x ’s having value for A). The im plications o f these views differ
in im p o rtan t ways, (b) allowing, for exam ple, whereas (a) and (c) do n o t, th a t some
x m ay have value for A independently o f A ’s preferring it. Because these views
differ, each m ust be exam ined as a possible w ay o f understanding value available to
those who seek to show th a t acceptance o f conditional equality com m its one to the
principle o f utility as a m inim al m oral principle. We shall p ostpone tem porarily an
assessment o f (b) and concentrate first on th a t feature com m on to b o th (a) and
(c) - th e feature, nam ely, th a t it is a necessary condition o f any x ’s having value for
A th a t A prefer x . F or convenience’s sake, let us refer to this as ‘th e preferencenecessary view.’
The preference-necessary view is a dem onstrably inadequate account o f value for
those who seek to convince us th a t we are com m ited to th e principle o f u tility , as a
m inim al m oral principle, if we accept the principle o f conditional equality. The
preference-necessary view holds th a t x has value for A only if A prefers x . Thus,
w hether A will place value on those consequences w hich, in Singer’s w ords, ‘m axi­
mize the interests o f those affected’ w ill d epend on w hether A happens to have an
interest in (i.e., prefers) these consequences. A nd it is perfectly possible th a t, though
A m ay have an interest in m any things, A does n o t have an interest in this.
In response it will be objected th a t if A fails to have an interest in maximizing
the interests o f those affected, th en A fails to fully accept the conditional equality
principle: Since th a t principle requires, am ong other things, th a t A consider the
interests o f everyone affected and co u n t equal interests equally, he m u st, if he is to
fulfill these requirem ents, also have an interest in, and th u s prefer, m axim izing the
interests o f everyone involved. This objection is specious and has apparent plausi­
bility only because it trades on an am biguity in the n o tio n o f ‘equal in terests.’ The
conditional equality principle requires th a t equal interests be co u n ted equally. It
does n o t require th a t one have an equal interest in equal interests. The difference
can be brought o u t by imagining th a t A and B b o th have an equal interest in a given
x (e.g., suppose x is the only available kidney and th a t b o th A and B have an equal
desire to have it). Then th e conditional equality principle requires th a t A n o t dis­
co u n t the im portance B attaches to getting x ju st because B is n o t A. If their
interests (preferences) are equal, then A m ust recognize th a t, lo o k ed at from B’s
p o in t o f view, getting x w ould be ju st as good for B, oth er things being equal, as A ’s
getting x w ould be, viewed from A ’s point o f view. But it does n o t follow from this,
and conditional equality does n o t itself require, th a t A m ust have an interest in B’s
getting x equal to A ’s in terest in his (A ’s) getting x - th a t A him self m u st prefer B’s
42
getting x as m uch as A prefers his (A ’s) getting x . From ‘I know th a t you w ant x as
m uch as I do,’ in other w ords, ‘I w ant y o u to get x as m uch as / w ant to get i t ’ does
n o t follow. Thus one can recognize the equality o f o th e r’s interests to one’s own
w ith out being com m itted to preferring th a t they further their interests as m uch as
one prefers furthering one’s own. A nd since one is n o t com m itted to this even when
one counts equal interests equally and considers the interests o f all who will be af­
fected by an action’s outcom e, one who accepts the conditional equality principle
certainly is n o t com m itted to having an interest in ‘maximizing th e interests o f
those affected.’
The situation, then, is this. If A’s having an interest in x is a necessary condition
o f A ’s valuing it, then A can accept the conditional equality principle and y et n o t
place, and n o t be com m itted to placing, any value on maximizing the interests of
everyone affected by an action’s outcom e. A is n o t co m m itted to this because A
m ay accept the conditional equality principle and yet n o t have, and n ot be com ­
m itted to having, an interest in maximizing the interests o f everyone affected. So,
the attem p t to generate a com m itm ent to utilitarianism , where this calls for m axi­
mizing the interests o f everyone affected, out o f acceptance o f conditional equality,
fails, if we assume th a t the account o f value on w hich the argum ent depends is the
preference-necessary view.
A n alternative account o f th e relationship betw een value and interest is the
preference-sufficient view: For any x, x has value for A if A prefers x . Since this
view o f value differs from the preference-necessary view, we need to ask w hether
this difference m akes any difference to the question at hand — the question of
w hether acceptance o f conditional equality com m its one to acceptance o f utility
as a m inim al m oral principle. The answer is no. What the preference-sufficient view
implies is th a t i f A happens to have an interest in m axim izing the interests of
everyone affected by an actio n ’s outcom e, then A will value th at outcom e. But
since A m ight ju st as well n o t have an interest in that outcom e and still accept the
principle o f conditional equality, it follows th a t the preference-sufficient view will
n o t suffice as account o f value th a t would necessitate the m ovem ent from accept­
ance o f conditional equality to acceptance o f the principle o f u tility , n o t even as a
m inim al principle.
Again it will be objected th a t if A fails to have an interest in m axim izing the
interests o f everyone affected by an action’s outcom e, th en it follows th a t A fails to
satisfy the requirem ents o f the principle o f conditional equality. B ut this reply has
no m ore force in the present case th an it had in the previous discussion o f the
preference-necessary view. Recognizing th e equality o f o th e r’s interests to o n e’s
ow n does n o t com m it one to preferring th a t their interests be fu rth ered as m uch as
one prefers th a t one’s ow n are. A m ust, it is true, recognize th a t, judged from B’s
p o in t o f view, it w ould be ju st as good (just as desirable, let us assume) if B’s
interests were furthered as it w ould be good, judged from his (A ’s) p o in t o f view, if
his (A ’s) were, assuming th a t they have equal interests and assuming th a t other
things are equal. B ut, in an interesting tu rn ab o u t o f Mill’s famous problem about
inferring ‘X is desirable’ from ‘X is desired,’ it sim ply does n o t follow th a t A m ust
him self desire th a t B’s interests be furthered, least o f all th a t A m ust desire this as
43
m uch as he desires th a t his own interests will be. Thus, since one is n o t com m itted
to preferring th a t others further their interests ju st because one recognizes th a t they
have interests some o f w hich are equal to o n e’s ow n, one w ho considers th e interests
o f others and counts them equally is n o t com m itted to the u tilitarian prescription
o f acting so as to maxim ize the interests o f everyone involved, given th e preferencesufficient view. One can accept the conditional equality principle an d n o t be com ­
m itte d to valuing th e m axim ization o f everyone’s interests as ‘th e best’ conse­
quences, i f we assume th a t one’s viewing these consequences as ‘th e best’ is co n ­
tingent upon one’s happening to prefer them .
Since, therefore, neither the preference-necessary nor the preference-sufficient
view com m its one who accepts th e conditional equality principle to having an
interest in, or thus, on either view, to valuing the m axim ization o f th e interests of,
those affected by an actio n ’s outcom e, neither view o f th e relationship betw een
value and interests will help validate the inference from (1) acceptance o f the
principle o f conditional equality to (2) acceptance o f the principle o f u tility . A nd
th e same is true when the preference-necessary and th e preference-sufficient views
are com bined, though a detailed supporting argum ent will n o t be ad u m b rated here.
It follows, therefore, assuming th a t the argum ent for utilitarianism based on ac­
ceptance o f conditional equality assumes one o f th e views regarding th e relationship
betw een interest and value exam ined in the preceding, th a t this argum ent fails to
establish th a t those w ho accept the principle o f conditional equality m ust therefore
accept th e principle o f u tility , at least as a m inim al principle. A nd since it fails to
establish this m uch on behalf o f utilitarianism , it also follows th a t this argum ent
cannot establish m ore than this. The argum ent in question, in o th er w ords, fails to
establish th a t those who accept conditional equality m ust therefore accept the
principle o f u tility as the o n ly substantive m oral principle, if value is und ersto o d in
term s o f one or another o f the versions o f the preference th eo ry exam ined in the
preceding. When value is understood in these ways, b o th SO : and S 0 2 , as charac­
terized earlier, are false.
VII
The following, then, has been argued to this point regarding the relationship be­
tw een the principle o f utility and the predistributive sense o f equality m arked by
B entham ’s declaration, ‘each to count for one, no one for m ore th an on e.’ In
section II it was argued th a t no utilitarian can reasonably h o ld th a t the equality
principle, w hen this is in terp reted in term s o f considering th e interests o f those
affected and counting equal interests equally, is a m oral principle, basic (w hich
w ould m ake utilitarianism inconsistent) or derived (w hich w ould d istort the concept
o f equality as it applies to interests). Singer in particular, th erefo re, m ust cease re­
ferring to equality as ‘a basic ethical principle,’ if he continues to subscribe to u til­
itarianism . Thus, on any possibly reasonable version o f utilitarianism , th e equality
principle m ust be regarded as a form al principle, n o t in th e sense th a t it sets fo rth a
44
necessary condition o f any principle’s having the status o f a m oral principle — this
it does n o t do, because, e.g., th e categorical im perative, as was argued in section III,
has this status b u t does n o t bid us consider the interests o f the individuals in ­
volved - b u t, at m ost, th a t it sets fo rth a conditional form al requirem ent, in a sense
explained in section IV. But because o f the argum ents ju st advanced in Section V,
there is no reason to believe th a t, viewed in this way, acceptance o f th e conditional
equality principle ‘leads to ’ utilitarianism , even as a ‘m inim al’ m oral principle, when
the account o f value th a t forms the backdrop is some version o f th e preference
theory.
There remains, th en , th e weak o ption characterized in section IV — the view th a t,
though acceptance o f conditional equality does n o t lead to utilitarianism , even as a
m inim al principle, one is at least consistent if one accepts b o th the principles o f
utility and conditional equality. Even this w eak o p tio n is too strong, how ever, if
the connection betw een value and interests is und ersto o d in any o f the ways
available to preference theorists, as these ways were characterized in the preceding
section. One thing utilitarians are com m itted to is th a t it som etim es is th e case th a t
either o f tw o outcom es w ould be as good as the other. For exam ple, if A j w ould
bring about as m uch good for A as A2 w ould bring ab o u t for B, th en , ceterus
parabus, utilitarians m ust m aintain th a t A i w ould be ju st as good as A 2 ; either
alternative, the utilitarian m ust say, w ould be equally good, equally desirable, no
m atter who is the beneficiary, A or B. T hat being so it is incum bent u p o n the
utilitarian to rely on a theo ry o f value th a t requires th a t, for cases like the one ju st
described, anyone w ho would judge one alternative to have a certain value m ust
also m ake (i.e., is rationally com m itted to m aking) the same ju d g m en t o f value
ab o u t the other. No version o f the preference theory can ground this requirem ent.
Given the preference-necessary view, y has value for A only if A has an interest in
y . L et us assume th a t A has an interest in y and th a t B does to o , B’s interest being
equal to A ’s .Utilitarianism requires a theory o f value th a t com m its A to making the
same judgm ent o f value, ceteris parabus, ab o u t B’s g ettin g y as he w ould m ake about
his (A ’s) getting it. The preference-necessary view fails in this regard. F or since A
can recognize th a t his and B’s interest in y are equal w ith o u t having, or being com ­
m itted to having, any interest in B’s getting y , it follows, given th e preferencenecessary view , th a t A can place a value on his getting y th a t differs from the value
he w ould place on B’s getting it. Thus, since utilitarianism requires a th eo ry of
value th a t com m its A to making the same value judg m en t w hether he or B g e tsy ;
and since the preference-necessary view does n o t carry this com m itm ent; u tili­
tarianism cannot rely on the preference-necessary view. Indeed, th e preferencenecessary view n o t only does n o t com m it A to making the same ju d g m en t o f value;
it positively justifies A in making different judgm ents, assuming th a t he him self
does n o t have an interest in B’s getting y . Thus, th e preference-necessary view
w ould authorize judgm ents o f value th a t are inconsistent w ith those judgm ents
utilitarians m u st m ake, given their theo ry . A nd since the preference-necessary view
has im plications th a t are inconsistent w ith the im plications o f utilitarianism , the
preference-necessary view o f value cannot consistently be held by utilitarians.
45
The same is true o f the preference-sufficient view. On this view, x has value for
A i f A has an interest in it. If we suppose th a t A and B have an equal interest in.y,
it does n o t follow th a t A m ust say th a t it w ould be ju st as good if B got y as it
w ould be if he (A) did, assuming the preference-sufficient view. Given this view, A
need n’t say this i f it happens to be the case th a t A has no interest in B’s getting ji.
For if A has no interest in B’s getting y , then A is n o t c o m m itte d to placing any
value w hatever on B’s getting it, assum ing th e preference-sufficient v ie w . Thus, the
preference-sufficient view is unable to account for the necessity, w ithin utilitarian
th eo ry , o f A ’s being com m itted to judging th a t it w ould be ju st as good, ju st as
desirable, if he got y or if B did. On the contrary, the preference-sufficient view
can allow A to deny the value o f B’s getting y , if it so happens th a t A has no interest
in B’s getting it. The preference-sufficient view, therefore, like th e preferencenecessary view, presents us w ith an account o f value th a t is logically at odds w ith
im plications o f utilitarian theo ry . A nd since these views o f value, considered either
singly or together, have im plications th a t are inconsistent w ith views th a t u tili­
tarians are com m itted to - utilitarians m u st judge th a t A ’s or B’s getting y w ould
be equally valuable, other things being equal - these versions o f the preference
th eo ry o f value are themselves inconsistent w ith utilitarianism .
This m uch established, we can now return to th e question o f th e consistency o f
th e principles o f conditional equality and utilitarianism . Earlier in this section it
was stated th a t the tw o principles are n o t consistent if the connections betw een
value and interests are understo o d in certain ways. W hat these ways are has now
been explained. I f th e conditional equality principle is understo o d in a way th a t
com m its one to the preference-necessary view, or to the preference-sufficient view,
or to the com bination o f these views, as previously characterized, th e n one assumes
an account o f value th a t is inconsistent w ith utilitarianism . Thus, if one interprets
the n o tio n o f interests, as this is used in the conditional equality principle, in one or
an o ther o f these ways, one interprets th a t principle in a way th a t proves to be in­
consistent w ith the principle o f utility. The fact th a t Singer does in terp ret interests
in this w ay, and the fact th a t Hare at least seems to (see section V ), is, if the
argum ent o f the present section is sound, absolutely fatal to their enterprise o f re­
quiring th a t those who accept conditional equality also accept th e principle o f
u tility . N ot only does acceptance o f th e conditional equality principle, w hen it is
u n d ersto o d in ways th a t com m it one to some version o f the preference th eo ry ,
n o t ‘lead to ’ the principle o f u tility ; w hen conditional equality is understo o d in
these ways it has im plications th a t are, if th e foregoing argum ent is sound, dem on­
strably inconsistent w ith the principle o f utility , viewed either as the only o r as a
m inim al m oral principle.
This last po in t can be m ade in a related way. U tilitarianism requires th a t one
take an im partial po in t o f view regarding the desirability o f th e outcom es o f alter­
native courses o f action. One m ust regard th a t outcom e as ‘the b est’ w hich brings
ab o ut the best consequences for everyone affected, w hether or n o t this outcom e
furthers one’s ow n interests m axim ally. W hat utilitarianism requires, therefore, is
an account o f value th a t illum inates w hy an individual, A, m u st recognize th a t the
46
consequences o f a given action are ‘the best’ even in cases where A’s ow n interests
are n o t furthered m axim ally. No version o f th e preference theory discussed in the
preceding is adequate in this regard. Since each version defines A ’s values in one
way or another in term s o f A’s preferences, these accounts o f value m ake A ’s judg­
m en t o f value contingent upon w hat A happens to prefer. T hus, since A m ight n ot
prefer consequences w hich are ‘the best’ for others b u t n o t for him self, and though
A m ay n o t have this preference while at the same tim e he recognizes th a t others
have preferences th a t are o f equal im portance to them as A’s are to him , the versions
o f the preference theory canvassed in the foregoing can n o t provide the theoretical
underpinning for the im partial point o f view utilitarianism requires. On the
co n trary, all versions o f the preference theory offer accounts o f value th a t stand in
th e w ay o f grounding th e necessity o f taking the im partial p o in t o f view utilitarian21
ism requires.
Perhaps it will be replied th a t anyone who recognizes the equality o f the interests
o f others to one’s own but who does n o t have an interest in having these interests
fu rthered equal to one’s own interest in one’s own case fails to display, to use.
w ords o f H are,24 ‘Christian hum ility and agape and the hum anist co u n terp arts.’ The
p o in t is certainly arguable. But whatever its m erits it is dem onstrably irrelevant to
the issue at hand. The issue at han d is w hether, given A ’s acceptance o f the con­
ditional equality principle, he is com m itted to the utilitarian prescription o f acting
so as to maximize the interests o f everyone involved. The argum ent advanced in the
above is th a t he is n o t com m itted to this, given the versions o f the preference
th eo ry exam ined. To reply th a t, though he is n o t com m itted to this, he nonetheless
fails to display ‘Chirstian hum ility and agape and the hum anist co u n terp arts’ if he is
unwilling to act to forw ard everyone’s interests misses the logical target and also
leaves undisclosed how the virtues o f Christian hum ility and the like could th e m ­
selves possibly be accounted for by some version o f th e preference theory. To
assume th a t they could be accounted for in this w ay w ould be to beg the question
and thus could n o t be used to prop up any version o f th e preference th eo ry on
behalf o f utilitarians.
VIII
It is im portant to emphasize w hat does and does n o t follow from the argum ent ad­
vanced in sections VI and VII against S 0 2 and, by im plication, S O j. What follows,
if th a t argum ent is sound, is th a t neither SOi nor S 0 2 can possibly be true if u tili­
tarians accept some version o f the preference theory o f value. W hat does n o t follow
from this, how ever, is th a t it m u st be inconsistent to accept b o th th e principles of
conditional equality and utility . This does n o t follow because, as n o te d in section V,
the n otion o f having an interest is im portantly am biguous, having b o th a preference
sense and a welfare sense. Now, given th a t th e welfare sense o f having an interest is
n o t synonym ous w ith the preference sense, it does n o t follow th a t th e principles o f
conditional equality and utility are inconsistent when th e n o tio n o f having an
47
interest is u n derstood in the welfare sense, even if it is tru e, as has been argued in
section V II, th a t these tw o principles are inconsistent if the n o tio n o f having an
interest is understo o d in th e preference sense. Indeed, n o t only does this n o t follow ;
for reasons about to be explained, the tw o principles arguably are n o t inconsistent,
when th e welfare in terp retatio n o f having an interest is applied.
The explanation o f this possibly is as follows. When interests are und ersto o d in
the welfare sense, w hat an individual has an interest in is n o t contingent upon w hat
th a t or another individual happens to prefer (e.g., w ant, desire). F o r exam ple, some­
thing (x) could be in A ’s interests and A m ight n o t w ant to have it, and the same
thing could be in B’s interests and A m ight n o t w ant B to have it either. Now
imagine th a t A judges th a t x is in his interests in the welfare sense. T hen A is com ­
m itted to the view th a t x will contribute to his welfare independently o f his hap ­
pening to w ant x . T hat being so, A cannot consistently go on to deny th a t B’s
getting x w ould be equally good, assuming th a t he and B have equal interests, on
th e grounds th a t he (A ) happens n o t to have an interest in (e.g., n o t to w ant) B to
get x . This liberty w ould be available to A i f interests were u n d ersto o d in the p re­
ference sense, b u t it is denied him w hen th e welfare sense is u nderstood. If A is to
co u n t equal interests equally, as conditional equality requires, th en A can n o t count
his ow n interests as welfare interests and deny th a t status to the like interests o f
others. Interests equal to his welfare interests are equal welfare interests, no m atter
whose th ey are and independently o f w hether he personally happens to have an
interest in seeing th a t others have their welfare interests furthered. Thus, if A judges
th a t x is in his interest in the welfare sense, and if he concedes th a t B’s interest in x
is equal to his ow n, th en A m ust acknow ledge th a t B’s getting x w ould benefit B as
m uch as his (A ’s) getting x w ould benefit A. A nd since the value o f A ’s getting x is
in dependent o f his w anting x and depends instead on how m uch having x will co n ­
trib u te to his w elfare, the value o f B’s having x m ust tu rn on these same consider­
ations. Given the welfare sense o f interests, th en , it is arguable th a t, A n o t only can,
A m u st acknow ledge th a t it w ould be ju st as good, ju st as desirable, if B g o tx as if
he (A) did, even if he does n o t prefer b o th outcom es equally. Since this is the view
o f this m atter utilitarians m ust accept and seek to illum inate by m eans o f a th eo ry
o f value, it is possible th a t, w hen interests are viewed in the welfare sense, and value
is u n derstood in term s o f w hat contributes to an individual’s w elfare, utilitarians
here have a theory o f value whose im plications are consistent w ith th e im plications
o f utilitarianism itself. This question can n o t be pursued fu rth er on th e present o c­
casion, except to n o te again th a t n either Singer no r Hare, for exam ple, understand
the relationship betw een interests and value in the w ay utilitarianism w ould seem to
require. Judged on the basis o f the evidence at hand (see section V ), th ey u n d er­
stand the n o tio n o f having an interest and the allied n o tio n o f having value in term s
o f individuals having preferences, w hen they should un d erstan d this fam ily o f ideas
in term s o f individuals having a welfare.
48
IX
Suppose the point m ost recently argued is true: One is consistent if one accepts
b oth the principles o f conditional equality and utility , if the n o tio n o f interests is
u n d erstood in the welfare sense. O ur final question, th en , is w hether one who ac­
cepts th e conditional equality principle m u st accept the principle o f u tility , either
as the sole or as a m inim al m oral principle, w hen interests are u n d ersto o d in the
required way. Because the welfare sense o f having an interest is distinct from the
preference sense, one cannot dispute an affirm ative answer to this question in the
way an affirmative answer was disputed, in section V I, w hen the preference sense
was involved. Since w hat is in A ’s interests, in the welfare sense, is n o t contingent
on w hat A happens to prefer, A cannot avoid viewing those consequences w hich
optim ize everyone’s welfare interests as ‘th e best’ by insisting th a t he happens n o t
to prefer this outcom e. Thus, if those consequences utilitarianism views as ‘the best’
are those which optim ize the welfare interests o f everyone affected, rather than
their preferences; if, th a t is, the form o f utilitarianism we are to consider is w hat
m ight be called welfare utilitarianism , n o t preference utilitarianism ; th en is it n o t
the case th a t A is com m itted to the principle o f u tility i f A accepts th e principle o f
conditional equality? T hat is the question at hand.
A negative answer is once again the correct answer. To m ake this clearer, suppose
th a t A accepts th e conditional equality principle: A is co m m itted to considering the
interests o f others, as he w ould have them consider his, and he is co m m itted to
counting equal interests equally, as he w ould have others do. Now, th a t really is all
th a t acceptance o f conditional equality com es to. As I have argued elsew here,25
this principle provides us w ith no direction regarding w hat we are m orally obligated
to do a fter interests have been canvassed and weighed equitably; all th a t it requires
is th a t we do the canvassing and weighing. The principle o f u tility , on the other
h an d, does provide post-canvassing and post-weighing direction. It lays it down, to
repeat Singer’s characterization, th a t we ‘m ust choose the course o f action w ith the
best consequences, on balance, for all affected.’ But if the conditional equality
principle does n o t provide us w ith the kind o f direction provided by the principle
o f u tility , how could it be th a t one who accepts th e form er principle m u st accept
the latter? It is difficult to see how accepting a principle w hich is void o f the kind
o f direction provided by the principle o f utility could com m it one to th a t latter
principle.
This difficulty can be highlighted by rem inding ourselves o f the fundatm entally
different status the tw o principles in question are supposed to have. The principle
o f conditional equality is a fo rm a l principle, one w hich specifies w hat one is com ­
m itted to , given th a t one w ould have others consider o n e’s own interests and weigh
them equitably. The principle o f u tility , on the oth er hand, is a substantive m oral
principle, specifying those actions one has an obligation to perform . It is funda­
m entally unclear how it could be true th a t, by accepting a form al principle, one is
thereby com m itted to a m oral one. It is n o t unclear how , by accepting one m oral
principle, one m ight also be com m itted to another m oral principle. B ut, for reasons
49
offered in the opening section o f this essay, there are overwhelming reasons against
a utilitarian’s regarding th e principle o f equality as a m oral principle, either as basic
(underived), w hich w ould render utilitarianism inconsistent, or as nonbasic (d e­
rived), w hich w ould distort the n o tio n o f equal interests.
This last po in t bears on a final attem p t the utilitarian m ight m ake to explain
how accepting conditional equality com m its one to the principle o f u tility , as­
suming, as we have th ro u g h o u t the present section, th a t th e n o tio n o f having an
interest is u nderstoo d in the welfare sense. F or suppose the utilitarian argues as
follows: ‘To accept th e conditional principle o f equality is n o t m erely to com m it
oneself to considering th e interests o f others, as one w ould have th em consider
on e’s ow n; nor does it m erely com m it one to counting equal interests equally,
w hom ever’s interests they m ay be. To accept this principle also includes o n e’s ac­
cepting th e injunction to act so as to fu rth e r the interests o f o thers, as one w ould
have others act so as to further o n e’s ow n. B ut,’ th e utilitarian goes on, ‘since one
accepts this injuction as p art o f conditional equality, it follows th a t one m ust also
agree to act as the principle o f u tility directs us. F or how else can we act so as to
forw ard th e interests o f all concerned save by acting as the u tilitarian principle
directs us? In this w ay, then, it is possible to show th a t anyone w ho accepts the
conditional equality principle is and m ust be com m itted to the principle o f u tility ,
at least as a m inim al m oral principle, even if n o t the only o n e.’
U nfortu n ately for utilitarianism , this argum ent, how ever plausible it m ay appear,
will n o t stand a m o m e n t’s critical reflection. One can n o t have it b o th ways. One
can not hold b oth th a t th e conditional equality principle is a form al principle and
th a t th a t principle directs us to act so as to fu rth er th e interests o f all concerned,
since to in terp ret th a t principle in this latter way is to alter its status. It is to in ter­
p ret it as a substantive m oral principle. M oreover, fo r reasons only recently adduced
for a second tim e, no utilitarian can reasonably ho ld th a t the eq u ality principle is a
substantive m oral principle. Thus, on any reasonable version o f utilitarianism , the
equality principle m ust be viewed as a form al principle and, m ore particularly, for
reasons given in section IV, as a conditional form al principle. The argum ent ju st
given on behalf o f the atte m p t to show th a t one com m its oneself to u tility if one
accepts conditional equality is therefore doubly wrong — w rong, first, because no
form al m oral principle can be a substantive one, and wrong, secondly, because no
u tilitarian can consistently regard th e equality principle as a m oral principle from
w hich the principle o f utility can be derived. D espite illusions to the contrary,
therefore, this last defense o f utilitarianism proves to be deficient, an d the conclu­
sion we should reach is thus th a t, even w hen th e notions o f having an interest and
value are u n derstood in ways th a t at least m ake them consistent w ith utilitarian
th e o ry , it is n o t true th a t one w ho accepts the conditional equality principle is
co m m itted to the principle o f uiltity , either as th e sole or as a m inim al m oral
principle.
50
It has n o t been the burden o f the present essay to argue th a t utilitarianism is a false
or otherw ise unacceptable m oral theory, though, for reasons I have offered else­
w here,26 I believe it is. My interest th oughout has been to focus on a different and
a surprisingly neglected range o f problem s th a t utilitarian th eo ry m ust deal with:
those problem s associated w ith th e predistributive, B entham ite n o tio n o f equality,
‘each to co u n t fo r one, no one for m ore th an one.’ I have sought to determ ine how
this n o tio n o f equality, especially w hen it is in terpreted in term s o f th e equality o f
interests, can find a hom e w ithin utilitarian thought. I f I am correct, this hom e can
be provided o n ly i f th e B entham ite principle o f equality is regarded by utilitarians
as a conditional form al principle, a nd o n ly i f the central n o tio n s o f having an
interest and value are understood in term s o f w hat co n trib u tes to an individual’s
w elfare, as distinct from w hat individuals prefer. In terp reted in these term s, the
B entham ite principle o f equality is at least consistent w ith utilitarianism . But th a t is
all it is. The atte m p t to show th a t it is m ore - in particular, the atte m p t to show th a t
acceptance o f conditional equality com m its one to the principle o f u tility , at least
as a m inim al m oral principle - fails, ju st as Mill’s attem p ted ‘p ro o f’ o f u tility does.
The failure laid bare in the present essay does n o t show th a t utilitarianism is false or
otherw ise unacceptable. W hat it does show is th a t, co n trary to th e argum ents of
Singer and Hare, and short o f m aking Sidgwick’s last-ditch appeal to ‘in itu itio n ’ as
the vehicle by w hich we recognize the principle o f u tility ’s claim on our rational
assent, we rem ain as far as we ever have been from having any com pelling reason for
accepting th a t principle as our m oral guide.27
NOTES
1. Peter Singer, Practical Ethics (Cambridge: Cambridge University Press, 1979), pp. 10—11.
2. See, for example, his ‘The Fable o f the Fox and the Unliberated Animal,’ Ethics (January
1978): 122.
3. ‘All Animals Are Equal, ‘Philosophical Exchange (Summer 1974), reprinted in Anim al
Rights and Human Obligations, ed. Tom Regan and Peter Singer (Englewood Cliffs, NJ:
Prentice-Hall, 1976), p. 148. Page references are to the latter appearance o f this essay. In
his most recently published work, Singer continues to characterize equality as a moral or
ethical principle. See his Practical Ethics, p. 18, where he writes that ‘(e)quality is a basic
ethical principle, not an assertion o f fact.’
4. Singer, ‘All Animals Are Equal,’ p.152.
5. See the previously cited works as well as A nim al Liberation (New York: New York Review
Book), 1975.
6. Singer,‘All Animals Are Equal, p. 152.
7. I assume that the problem o f the logical relationship between the principles o f utility and
equality can and will arise for ‘cognitivist’ and ‘noncognitivist’ utilitarians alike, and thus
that this problem could not be avoided by a Singer-type utilitarian who advocates a non­
cognitivist meta-ethic. In assuming that noncognitivist utilitarians can and will accept that
logical relations hold between different moral prescriptions, I assume what noncognitivist
utilitarians themselves profess. On this point, see R.M. Hare, The Language o f Morals
51
(Oxford: Clarendon Press, 1952), esp. chs. 2 and 3; and J.J.C. Smart’s contribution to
Utilitarianism: For and Against (Cambridge: Cambridge University Press, 1973), esp. pp
8.
9.
10.
11.
12.
7 -9 .
‘Utilitarianism and Vegetarianism,’ Philosophy and Public Affairs (Summer 1 9 8 0 ):3 2 8 29. For my criticisms o f Singer’s attempt to ground the moral obligatoriness o f veg­
etarianism on utility, see my essay, ‘Utilitarianism, Vegetarianism, and Animal Rights’ in
this same issue o f Philosophy and Public Affairs, and my ‘Animal Rights, Human Wrongs,’
Environmental E thics (Summer 1980):99 —120.
Singer, ‘Utilitarianism and Vegetarianism,’ p. 329.
Though a possible option, I know of no utilitarian who has chosen it. Certainly Singer,
much to his credit, does not. To do so would, among other things, undermine his case for
animal liberation, since the utility o f counting the interest animals have in avoiding pain as
equal to the interest humans have in avoiding it might differ quite markedly.
Singer, Practical Ethics, p. 12.
In discussions o f an earlier draft o f the present essay, W.R. Carter made the point that the
principle of conditional equality itself does not ‘lead to ’ any substantive moral principle
because it cannot do so. Carter’s point is that since the conditional equality principle is ex­
pressed hypothetically (‘If so and so, then such and such’); since there is no nontautological, nonhypothetical proposition that can follow necessarily from any hypothetical
propostion; and since whatever else we may want to say o f the principle of utiliity, all at
least are agreed that it is not tautological and nonhypothetical; it follows that one cannot
derive the principle o f utility from the mere statement o f the principle of conditional
equality. What one must assume, and what is assumed in the argument that unfolds in the
following pages, especially the argument o f sections V I-IX , is that some individual (A)
does accept the conditional equality principle; that is, A would have others consider his
interests and to count them equitably, etc. The logical form o f the argument to be ex­
amined in subsequent sections thus is not ( [ p & q ] -*r)->-s, where p symbolizes ‘A
would have others consider his interests and count them equitably,’ q symbolizes ‘A takes
the moral point o f view ,’ r symbolizes ‘A commits himself to considering the interests of
all those affected and to counting equal interests equally,’ and s symbolizes ‘A commits
himself to the principle o f utility.’ This cannot be the logical form o f the argument since
l [ p &,q] - * r ) - * s is not an argument. Rather, the form o f the argument must be assumed
to be as follows:
( [ p & q ] -*r) ^ s
([P &<?] ~*r)
s
13.
On one occasion Hare observes that utilitarianism is ‘consistent’ with his views on the
meanings of the moral use of terms like ‘right’ and ‘ought’, thus suggesting that he would
accept what I have termed ‘the weak option.’ See his ‘Principles,’ Proceedings o f the
Aristotelian Society (1 9 7 2 -1 9 7 3 ), p. 15. On another occasion he argues that what I have
termed the principle o f conditional equality ‘leads to ’ or ‘yields’ utilitarianism, thereby
suggesting that he would accept ‘the strong option.’ See his ‘Ethical Theory and Utili­
tarianism,’ in H.D. Lewis, ed., Contemporary British Philosophy (London, 1976), pp.
1 1 6 -1 1 7 . I believe, though I shall not argue the point here, that Hare’s arguments are
open to all the same (and more) objections raised in the present essay against Singer’s
attempt to generate utilitarianism out o f the conditional equality principle. I discuss this
and related matters in an unpublished essay, ‘Hare on Justice, Utility and the Appeal to
Intuition.’ It is also worth noting that Singer, for one, though acknowledging a debt to
Hare for the use o f the idea o f getting utilitarianism from conditional equality, makes it
clear that he does not think the argument shows as much as he thinks Hare supposes. “The
tentative argument for a utilitarianism based on interests or preferences owes most to
52
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Hare’s ‘Ethical Theory and Utilitarianism’, although it does not go as far as the argument
in that article’ (Singer, Practical Ethics, p. 222). Singer, then, evidently interprets Hare as
selecting what I have labeled ‘S O ,’ - i.e., ‘Strong O p tion ,.’ I believe this is the most
reasonable interpretation o f Hare.
Singer, Practical Ethics, p. 13.
Ibid., pp. 1 2 -1 3 .
This does not mean that utilitarianism has clear sailing if only the preference theory of
value is avoided. Serious problems remain even for a Moorean position regarding intrinsic
value. On this, see my ‘Moore’s Accounts o f ‘Right’, Dialogue (October 1972), and ‘A
Refutation of Utilitarianism’ Canadian Journal o f Philosophy (forthcoming).
I note and explore this distinction in my ‘McCloskey on Why Animals Cannot Have
Right,' Philosophical Quarterly (October 1977).
Singer, Practical Ethics, p. 12.
Ibid.
R.M. Hare, ‘Ethical Theory and Utilitarianism,’ p. 117. It is worth noting that at one
point Hare suggests an argument for utilitarianism that makes use o f the notions o f harm
and benefit, thereby suggesting an argument that involves understanding interests and
values in the welfare sense. It is highly significant, however, that Hare decides ‘n o t . . . to
put my theory in terms o f benefits and the reduction o f harms, because this leads to dif­
ficulties that I wish to avoid’ (ibid., emphasis added). If this means that Hare prefers to
rest his argument or acceptance o f some version o f the preference theory o f value, as the
available evidence suggests he does, then, if the argument o f section VII is sound, Hare has
avoided some (unspecified) difficulties only to embrace the quite specific difficulty of
relying on an account o f value which is inconsistent with the very principle (utility) he
seeks to support.
Singer, Practical Ethics, p. 222, emphasis added.
The distinctions drawn and addressed in what follows apply both to ‘crude’ and more re­
fined types o f preference theories o f value. A crude form would be that x has value for A
if A Oust happens) to desire x . A more refined form would be that x has value for A if, on
balance and after reflection on all the relevant facts, A then happens to desire x . Singer
sometimes writes as if he is willing to endorse a crude form o f preference theory (cf.
Practical Ethics, p. 12) and, at other times, he writes as if only a more refined form is
acceptable (cf. ibid., p. 80). The criticisms launched against S 0 2 and, by implication, S O , ,
in sections VI and VII, apply with equal force to both the crude and the more refined
understandings o f the connection between value and preferences.
Thus, the common argument that preference utilitarianism has serious problems of
measuring or weighing the preferences o f the individuate whose interests are affected by an
action’s outcome, though certainly a not unimportant ground for objecting to preference
utilitarianism, actually concedes more to this variety o f utilitarianism than should be con­
ceded to it. That objection assumes that a preference theory o f value is hospitable to util­
itarianism, the problem being that it is difficult to work out a credible, reliable method of
weighing and ranking preferences. The more fundamental difficulty, however, and the one
argued for in the above, is that a preference theory o f value actually is inconsistent with the
principle o f utility. For a perspicuous discussion of the ‘measuring’ problems preference
utilitarians face, see Dan W. Brock, ‘Utilitarianism,’ in Tom Regan and Donald Van De
Veer, eds., Individual Rights and Public P olicy (Totowa, NJ: Rowman and Littlefield),
1982.
R.M. Hare, ‘Ethical Theory and Utilitarianism,’ p. 20, fn. 11. I do not mean to imply that
Hare would object in this way at this point, only that someone might.
See my ‘Utilitarianism, Vegetarianism, and Animal Rights’, pp. 3 1 2 ff .
See, in particular, ‘Utilitarianism, Vegetarianism, and Animal Rights’ and ‘A Refutation of
Utilitarianism.’
Work on this essay was assisted by a grant from the National Endowment for the Human­
ities. It is a pleasure to acknowledge this support. I also want to thank Dale Jamieson,
Barbara Levenbook and especially W.R. Carter for helpful criticisms o f earlier drafts of this
paper.