-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 337
Ethical Theory and Moral Practice 2: 337363, 1999. 1999 Kluwer
Academic Publishers. Printed in the Netherlands.
ADAM SWIFT
PUBLIC OPINION AND POLITICAL PHILOSOPHY:THE RELATION BETWEEN
SOCIAL-SCIENTIFIC AND
PHILOSOPHICAL ANALYSES OF DISTRIBUTIVE JUSTICE1
ABSTRACT. This paper considers the relation between
philosophical discussions of,and social-scientific research into
popular beliefs about, distributive justice. The firstpart sets out
the differences and tensions between the two perspectives,
identifyingconsiderations which tend to lead adherents of each
discipline to regard the other as irrelevantto its concerns. The
second discusses four reasons why social scientists might benefit
fromphilosophy: problems in identifying inconsistency, the fact
that non-justice considerationsmight underlie distributive
judgments, the way in which different principles of justice
canyield the same concrete distributive judgments, and the
ambiguity of key terms. The thirdpart distinguishes and evaluates
three versions of the claim that normative theorising aboutjustice
can profit from empirical research into public opinion: that its
findings are food forthought, that they amount to feasibility
constraints, and that they are constitutive ofnormatively justified
principles of justice. The view that popular opinion about
justicehas a strongly constitutive role to play in justifying
principles of distributive justicestricto sensu is rejected, but it
is argued that what the people think (and what they canreasonably
be expected to come to think) on distributive matters can be an
importantfactor for the political theorist to take into account,
for reasons of legitimacy, or feasibility,or both.
KEY WORDS: desert, distributive justice, public opinion, John
Rawls, social justice,Michael Walzer
1. INTRODUCTION
What is the proper relation between social-scientific research
into peoplesbeliefs about distributive justice and normative
philosophising on thesubject? Sociologists and psychologists have
produced an extensiveliterature attempting to describe and explain
peoples beliefs aboutdistributive matters, while the normative
investigation of what justicemeans and what it requires of us has
received no less attention from political
1 An ancestor of this paper was given at a Conference on
Social-Scientific and Nor-mative Analyses of Justice at the
European University Institute, Florence, in 1993. Partsof it
subsequently appeared in Swift et al. (1995). I am grateful to my
co-authors and toAldine de Gruyter for permission to reproduce
those parts here. Work on it was com-pleted during a stay as
Visiting Fellow in the Social and Political Theory Group in
theResearch School of Social Sciences at the Australian National
University, where itbenefitted from discussions with Richard
Holton, Susan Mendus, and Michael Smith,and from comments at
presentations both there and at the University of Melbourne.
-
ADAM SWIFT338
philosophers. Despite occasional calls for the enrichment of the
one bythe other, and even more occasional suggestions as to how
exactly thismight be brought about and to what effect, there have
been few systematicattempts to address the question of what these
two approaches can do forone another (notable exceptions are Miller
(1994) and Elster (1995)). Thispaper, reflecting my involvement in
the International Social Justice Project(see Kluegel et al. (1995)
for details), is an attempt to tackle that questionhead on.
Much of what follows is sceptical. I argue that social
scientists have muchto learn from philosophers, the main lesson
being how difficult it is for themever really to get at peoples
beliefs about principles of justice. Thisscepticism is tempered,
however, by an acknowledgement that there are otherkinds of
distributive belief that such research can tell us about and that
maybe more important. A similar claim emerges from my discussion of
whatnormative theorists have to learn from the findings of
empirical research.Here I reject the view that popular opinion
about justice has a stronglyconstitutive role to play in justifying
principles of justice stricto sensu, andshow that some of those who
might be thought to accord it such a role donot in fact do so. But
I acknowledge a weaker constitutive role and claimthat what the
people think (and what they can reasonably be expected to cometo
think) on distributive matters can be an important factor for the
politicaltheorist to take into account for reasons of legitimacy,
or feasibility, or both.
2. DIFFERENCES AND TENSIONS
The view that social science and political philosophy have
little to say toone another, in this area at least, is easily
stated. The justification ofprinciples of justice seems simply to
occupy a different logical space fromthe description and
explanation of the principles to which people actuallysubscribe.
Indeed, and here the difference begins to turn into anopposition or
tension, since philosophers reasons for favouring aparticular set
of principles are likely to be rather complicated and tohave taken
a great deal of thought to arrive at, it would perhaps besurprising
if ordinary people, not necessarily less intelligent but in
allprobability having devoted less time to the exercise of that
intelligenceon this particular question, agreed with them.
Normative politicalphilosophy understands itself as providing
reasons why people shouldendorse particular conceptions of justice,
and the conceptions that peoplehappen empirically to espouse would
appear not only to be of littlerelevance but also likely to lack
the very sophistication that philosophersseek to achieve.
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 339
This was certainly the view of one of the anonymous referees who
wasasked to comment upon Gordon Marshalls request for funding for
theBritish leg of the ISJP:
Surveys of popular opinion on these topics seem to me of little
academic value...Thegreat debate about justice that has been in
progress since the time of Plato has thrownup many difficulties.
But we will not be helped in the least in the resolution of
thesedifficulties by a knowledge of the quirks of public opinion.
Justice is, one might almostsay, a semi-technical notion. It is the
topic discussed by Plato in the Republic, byAquinas in Summa
Theologia, and by Mill in Utilitarianism... . Someone who
knowsnothing of this material is hardly in a position to contribute
to the resolution of ourproblems. It would seem therefore a waste
of time to survey the views of people whoare not in a position to
judge the issues...2
From this perspective, there is no problem in concluding that
people aresimply wrong in advocating distributive principles that
more carefulthought, or perhaps even superior moral sense, would
lead them to disavow.
The flip side of these philosophical doubts about the value of
social-scientific research is the social scientists tendency to
mistrust thephilosophers search for rational justification and
truth. That Justice isin the eye of the beholder is widely
recognized by social psychologists,reports Tornblom (1992, p. 177),
and the fallacious inference from theexplanatory importance of
subjective perceptions to the meta-ethical claimthat moral
judgments are merely subjective and not susceptible to criteriaof
rational assessment is widespread. Quite understandably struck
notonly by the sheer variety of views about justice espoused by lay
socialactors but also by the myriad of competing conceptions of
justice putforward by academic philosophers, the social scientist
is likely to dismissthe task of answering the question Just what is
just? as hopeless andpompous (Reis 1984, p. 38). This is again a
bad inference, since the mereexistence of disagreement about which
conception of justice is rightcannot warrant the conclusion that no
such conception could be. It is moreplausible, perhaps, to doubt
that even if the, ex hypothesi, true (or reasonable)moral
distributive theory were to be arrived at philosophically, its
advocateswould be unable to secure acknowledgement of its truth (or
reasonableness)and agreement on its substance. This is not to deny
the very idea of anormative philosophical position being right but
simply to assert theempirical predictive claim perhaps itself of
great importance to thenormative theorist that people are unlikely
to agree on any position,whether right or not.
2 The Economic and Social Research Council took a different view
and supported theresearch with grant number R000232421.
-
ADAM SWIFT340
Fuelling the philosophical scepticism characteristic of social
scientists,and another source of tension between the two
approaches, is the fact thatthe social-scientific project is
explanatory as well as descriptive. Socialscience seeks to provide
causal explanations of those very justice judgmentsfor or against
which the political philosopher seeks to supply reasons. Fromthe
explanatory perspective, this search for causes is likely to tell
againstthe significance of philosophical argument. If a persons
class origins,gender, personal history of social mobility,
experience of solidaristicrelations, and so on are regarded as
playing a causal role in forming herbeliefs about distributive
justice, then it is not clear what role is left forthe rational
argument that it is the philosophers project to provide. Fromthe
other side, of course, those committed to the practice of
reason-givingare likely to regard the search for causal
explanations of beliefs asessentially irrelevant. The reasons for a
belief and its causal origins aredifferent in kind, and it seems
open to the normative philosopher to arguethat nothing to do with
the explanation of why someone holds a particularconception of
justice tells for or against that conceptions being
morallyjustified.
A distinct, but again misplaced, source of social-scientific
scepticismabout normative philosophising derives more specifically
from the socialscientists awareness of the extent to which, as a
matter of empirical fact,peoples distributive judgments are complex
and highly context-specific(Tornblom 1992). But this only
constitutes an objection to the search for avalid normative theory
if one assumes both that that the judgments peopleordinarily make
must in some way or other inform ones view about thejudgments that
they have reason to make and that philosophical normativetheories
of justice are by their very nature insensitive to the
complexityand context-specificity of justice judgments. This latter
assumption seemsmistaken. It is true that philosophers tend to
formulate their positions interms that abstract from complexity and
transcend particularity of contexts,to seek the deeper structures
of thought, the more general and widelyapplicable reasons, that
ought to guide our concrete judgments. Accordingto Walzer (1983, p.
4), the first impulse of the philosopher is . . . to searchfor some
underlying unity: a short list of basic goods, quickly abstractedto
a single good; a single distributive criterion or an interconnected
set;and the philosopher himself standing, symbolically at least, at
a singledecision point. Perhaps this is the philosophical impulse.
But we shouldnot confuse Walzers pluralistic substantive theory
with the anti-philosophical methodology he claims to use to arrive
at it. There is nothingin principle incoherent about the idea of a
full-bloodedly philosophical
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 341
theory of justice that was itself pluralistic and acknowledged
that one couldnot expect to make correct judgments about particular
distributive issueswithout knowing a great deal of detail about the
particulars of the situationone was being asked to judge. It may
simply be right that what should countas a just distribution
depends on various characteristics of the good beingdistributed,
the individuals among whom it is being distributed, the
relationbetween those individuals, and so on. A wholeheartedly
philosophicalnormative conception of justice could consist of just
such a complex andplural set of considerations.
So: philosophers tend to dismiss the significance of empirical
researchbecause what people believe and what they ought to believe
are differentin kind, and philosophy is a quasi-technical
discipline. Social scientiststend towards moral scepticism, partly
because they observe not onlyordinary people but also professional
philosophers disagreeing about whatjustice demands, partly because
causal explanations seem in some sensecompetitive with the rational
justification of justice beliefs, and partlybecause social
scientists are so aware of the multiplicity, complexity
andcontext-specificity of peoples justice judgments. Indeed, they
are alsoaware of the extent to which the judgments of individual
lay actors arecontradictory and inconsistent. The inference from
people disagree aboutthe truth of what is just to there is no truth
about what is just perhapsgains plausibility from the observation
that people seem to disagree notonly with one another but also with
themselves. That inference remainsinvalid. But one can understand
why those who spend their time wadingthrough the contradictions and
confusions in ordinary minds should growdoubtful of the claim that
there is a coherent, albeit possibly complex,philosophical truth to
be apprehended.
3. SOCIAL SCIENCE NEEDS PHILOSOPHY?
This issue of individual inconsistency raises the question of
the extent towhich social-scientific research necessarily requires
philosophical analysis.As my colleagues and I have argued elsewhere
(Burgoyne et al. 1993), thecontext-specificity of distributive
judgments, the possibility of sophisticatedcontextual reasoning,
and the different interpretations to which principlesof justice are
reasonably subject make it very difficult to ascertain forcertain
whether judgments that appear inconsistent really are so. This
isnot, of course, to deny that people often are inconsistent, and
for a varietyof reasons ranging from intellectual incompetence to
self-servinghypocrisy. Indeed, even if it cannot warrant the
conclusion that there is no
-
ADAM SWIFT342
such thing as justice, the extent of confusion and contradiction
in ordinaryminds may quite reasonably lead one to doubt any
justificatory strategythat claims to give a significant role to
empirical opinion. But often theevidence adduced to support the
charge of inconsistency simply fails toestablish the case. It is
not necessarily inconsistent to believe in the justiceof a more
equal society but be unwilling oneself to pay more taxes:
perhapsthe respondent giving these apparently contradictory views
believes thatshe is already paying more than her fair share, or
that revenue currentlyspent on defence should be diverted to social
welfare. Similarly, thosewho advocate the abolition of private
education but send their children toprivate schools need not be
hypocritical or suffering from moral weakness:perhaps they think
that were private education to be abolished the qualityof state
education would increase sufficiently for them to discharge
throughstate schools that parental obligation that they believe
currently requiresthem to go private. Finally, and here we have an
example not of contextualreasoning but of straightforward
difference in interpretation, it is notnecessarily inconsistent to
believe that justice demands equality ofopportunity but permits
people to pass their wealth on to their children;for example if
equality of opportunity is taken to mean only that peoplesrace or
gender should be irrelevant to their employment prospects.
These examples suggest that there is a philosophical component
to theinterpretation of empirical findings, and this alone would
seem sufficientto persuade the social scientist of her need for
philosophical assistance.The empirical researcher, however
sceptical of the philosophers normativeproject, cannot escape the
demands of analytical precision and clarity, forher very claim that
her subjects do not themselves possess precise and clearviews about
justice presupposes her ability to identify what they do thinkabout
justice and to recognize what a clear and precise view about
justicelooks like and this in turn requires awareness of the
theoretical complexitiesthat are the stuff of philosophical
analysis. (See also Miller 1994, pp. 169174) and Bell and
Schokkaert (1992, p. 247)).
Let me illustrate this claim further by reference to my
experience in theISJP. One of the guiding theoretical ideas
underlying the construction ofthe questionnaire was the distinction
between peoples judgments aboutwhat is the case and their judgments
about what ought to be the case: thedifference, for example,
between the belief that people do get rewardedfor effort and the
belief that they ought to be so rewarded, or between viewsabout
what factory workers actually earn and views about what they
oughtto earn. An immediate problem raised by the distinction in
this form is thatpeoples beliefs about what ought to be the case
are not necessarilyequivalent to their beliefs about justice. This
is because it is conventional,
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 343
at least among philosophers, to distinguish between the demands
of justiceand the demands of morality more broadly understood:
perhaps justicerequires a certain distribution but some other moral
consideration precludesit, or perhaps justice does not require a
certain distribution but the moralclaims of humanity or charity
(Campbell 1974), or an independent moralideal such as social
equality (Miller 1994, pp. 17374), do. We could thusfind our
respondents saying that people ought to get what they need, orthat
managing directors should only earn five times as much as
assemblyline workers, without knowing whether this was something
that theyregarded as a principle of justice.
One problem, then, is that the researcher needs to take care
that she isgetting at what people think about justice.3 Philosophy
is needed to identifythe appropriate subset of moral concerns. A
second concerns the way inwhich one cannot directly infer peoples
beliefs about principles of justicefrom their beliefs about the
justice of particular distributions, or even fromtheir beliefs
about the justice of particular distributive principles. Supposeone
found (as we did) that more than 90% of people believe that a
justsociety is one whose members have equality of opportunity. Is
this becausethey subscribe to a principle of desert and hold that
such a society is likelyto give people what they deserve? Because
they believe that people equallyhave a right to self-realisation
and regard equality of opportunity as theprecondition of its
realisation? Or because they have a Rawlsian concernfor the
worst-off and believe that such a society will make better use
ofthe pool of ability, hence be more productive, and hence work to
thatgroups long-term advantage? Suppose one found that (as we did)
thatpeople think that the income disparities that they believe to
exist betweenfactory workers and managing directors are unjust, and
(as we did not) thatthe just incomes attaching to those two
occupations would be equal. Is thisbecause they believe that people
deserve to be rewarded for hard work (andfor hard work alone) and
they think that factory workers work as hard as
3 The obvious way to do this is to ask them questions
specifically about justice, and sowe did. The success of this
strategy will, however, depend on the extent to which re-spondents
are sensitive to the precise wording of the questions put to them.
Asking themabout what they think justice requires does not
guarantee that they will answer in thoseterms. A further problem is
that in many areas it seems natural to use the term fair todenote
this stricter domain, and ISJP pilot studies suggested that that we
risked not mak-ing sense to our respondents if we insisted on the
philosophically precise point thatfairness is but one way of
thinking about justice. Here we were left in an
impossiblesituation: the only way to defend ourselves against the
charge of prejudicing our findingsby equating justice with fairness
(rather than, for example, with entitlement) was to usean adjective
just that was, in some contexts, not part of the ordinary language
ofour respondents and so risked the unintelligibility problem that
I discuss below.
-
ADAM SWIFT344
managing directors? Or is it because they reject the view that
people deserveto be rewarded for hard work but reject also any
other attempts to justifyinequality as just? Again, philosophy is
needed to tell the social scientistthat particular judgments as to
the justice or otherwise of particulardistributions, or even of
particular distributive principles such as equalityof opportunity,
cannot license inferences about which particular principlesof
justice are being endorsed.
Researchers can of course try to combat this problem by ensuring
thattheir research instruments are formulated as carefully as
possible to capturefoundational principles of justice rather than
more superficial distributivejudgments. The ISJP questionnaire had
distinctions between desert,entitlement, fairness, functional
inequality (i.e. inequality conducive to thelong-term advantage of
the worst-off) built into its questions and doingthis inevitably
involved awareness of the philosophical literature.We didnot,
however, and I will argue a survey questionnaire cannot, go far
enough.
Consider the question of desert. It is clearly important to
distinguishbetween the idea that people deserve to be rewarded for
effort and the ideathat they deserve to be rewarded for possessing
valuable skills orintelligence. The statements we put to
respondents in an attempt to get atthis distinction were People who
work hard deserve to earn more thanthose who do not and Its just
luck if some people are more skillful orintelligent than others, so
they dont deserve to earn more money. Butthe second of these
contains two distinct claims. Accordingly, when peopledisagree with
the statement, this might be because they do not accept
thatintelligence and skill are a matter of luck (perhaps they
believe that apersons skill is a result of their efforts in
developing their natural abilities)or because they do regard them
as chance attributes but think themdeserving of reward nonetheless.
To see the depth of the problem, noticethat even if we had been
able to establish that it was the latter, this wouldstill leave
indeterminacy. Respondents might be understanding the desertclaim
as one about legitimate expectations, given current institutions
andpractices, rather than as a claim about desert in a stronger,
pre-institutionalsense. Or it might even be that they are using the
notion of desert as ashorthand for a consequentialist claim as when
somebody says that shebelieves that brain surgeons deserve to earn
more money than hospitalorderlies because she thinks that if brain
surgeons did not earn more thennobody would want to be one. Unless
one knows what respondentsunderstand by questions about desert, one
simply cannot tell which of thesevarious quite distinct positions
they are endorsing.
It is tempting to think that this kind of indeterminacy could
have been,perhaps should have been, removed. If one really wants to
find out what
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 345
people think about such fine-tuned issues, then one should ask
appropriatelyfine-tuned questions. The objection to this suggestion
comes in twovariants. The first is that such questions will involve
a considerabledistancing from the cruder ways in which
non-philosophers think aboutdistributive matters and will thereby
be open to the charge of imposing ontheir respondents categories
that are quite alien to them, and thereby ofcreating data that fail
to correspond to the mind-sets of lay social actors.This is
something that social researchers are generally supposed to
avoid.But it is obvious that finding out what people believe about
difficultphilosophical issues will involve getting them to think in
ways that theyare not used to. So unless one is opposed to the
whole project of seeingwhat people think about such issues after
reflection of a kind they maynot previously have undertaken one
must surely accept the offering, ifnot the imposing, of categories
with which they are not already familiar.4
More worrying is the variant of the objection that points to the
sheerimpracticability of a philosophically more precise
questionnaire. A surveycomposed of questions unintelligible to many
of its respondents and asurvey attempting to get at distinctions of
the kind outlined above willsurely consist of such questions will
not be a good way of finding outwhat its respondents think about
anything. This objection seems to mecompelling. It is simply
impractical to attempt by means of a surveyquestionnaire to
identify what a representative sample of the populationthinks about
the kind of issue that divides philosophers, because one cannotget
sensible responses to questions that use formulations
unintelligible tothe majority of those being surveyed, and the
issues that divide philosopherscannot be formulated in any other
way.
What method would do better? Confidently to identify peoples
beliefsabout philosophical issues one will need, I think, to teach
them somephilosophy. So the proper investigative method is
something like a seriesof tutorials, or at least a series of
lectures or classes explaining thesedistinctions, illustrating them
by means of examples, allowing people toask questions, and so on,
followed by a now intelligible questionnaire. Onewill not, perhaps,
know whether one is identifying the beliefs that peopleheld prior
to the research process, causing them to revise their beliefs
in
4 We countered the reasonable element of this objection as best
we could, whilst main-taining our philosophical purity, by
ensuring, first, that respondents were able to answerdont know as
well as neither agree or disagree, and second, in the case of the
Britishsurvey alone, that we had an open-ended question in which
respondents were invited tosay what they thought about social
justice - and particularly encouraged to mention ideasthat were
different from those in our questionnaire - with our interviewers
instructed towrite down whatever they said.
-
ADAM SWIFT346
the course of that process, or simply leading them to have
beliefs on issueswhere previously they had none. But only in this
way can the problemsdiscussed above be overcome.5
I have argued that social scientists need philosophy if they are
to makegood their claim to be investigating lay actors beliefs
about principles ofdistributive justice. This started as a modest
claim about difficulties ininterpreting findings and ended with the
bolder but in fact rather obvioussuggestion that one needs to teach
people some philosophy before one canfind out what they think about
philosophical matters. But why should thesocial scientist be
concerned to investigate such matters in the first place?If
ordinary people do not always distinguish between what justice
requiresand what ought to be the case, or between desert and
legitimate expecta-tions, does this not show that the
considerations outlined above are simplyirrelevant to the
social-scientific endeavour? The nice distinctions that arethe
philosophers stock-in-trade would seem to have no explanatory
ordescriptive implications if they do not correspond to the ways in
whichsocial actors themselves conceive distributive issues this is
what motivatesthe worry about imposing categories on respondents.
Perhaps socialscientists do not particularly care about what people
believe aboutprinciples of justice stricto sensu. They are
interested in the vaguer, lessprecise judgments that ordinary
people ordinarily make, and if philosophersrefuse to grant these
the status of principles of justice then that is theirproblem. On
this account, empirical researchers often are not researchingwhat
philosophers would consider ordinary peoples principles of
justicebut this is something that they can happily concede.
To be sure, insofar as she is concerned to explain and predict
what peoplebelieve and are likely to believe about the justice of a
distribution inchanged circumstances or with new information, it
may well be that thesocial scientist cannot afford to ignore some
of the distinctions outlinedabove. The significance of the finding
that, in a particular case, incentivepayments are unnecessary or
ineffective will be different depending onwhether peoples support
for the inequality in question is justified on desert-based or
incentive-based grounds (Miller 1994, p. 175). Similarly,
withindesert, it will make a good deal of difference to their
particular judgments
5 There are some similarities between this proposal and Fishkins
(1995) deliberativeopinion polls. Fishkin, however, is driven to
endorse deliberative polls because he thinkswe should care about
peoples informed views on matters of political concern aboutwhich
they are likely already to have uninformed views. My proposal is
motivated bydoubt that members of the public have views on, or can
understand questions about,principles of justice as those are
discussed by philosophers.
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 347
as to the justice of particular policies whether they believe
that peopledeserve to be rewarded for effort but not natural
talent, or for both. If youare interested in predicting justice
judgments, then doubtless you needcarefully and correctly to
identify the principles that generate thosejudgments.
But why should one be interested in justice judgments
specifically? Theobvious answer is that they matter because they
explain what people do.As Miller (1994, p. 176) puts it, citing
Barrington Moore (1979) in support,It has often been maintained
indeed it is almost a truism thatexperiencing a situation as unjust
is what leads people to protest against itand to take action to
alter it, and the literature, both social-scientific andnormative,
abounds with claims to this effect. (For examples of each seeGoode
(1978, p. 329) and Nozick (1974, p. 158)). Despite such
widespreadagreement, however, a sense of injustice is surely
neither a necessary nora sufficient condition for such action.
Put to one side more generally sceptical doubts about the
significanceof peoples moral beliefs, and certainly about the
significance of their moralbeliefs as stated to an interviewer, for
what people do. These doubts playa major role in explaining
social-scientific mistrust of research into justicebeliefs, but I
am here considering a more specific possibility, one that hasforce
even on the assumption that people are typically motivated to act
inaccordance with their moral judgments. This is the possibility
that peoplemight believe a distribution to be unjust but still not
believe that it shouldbe altered. The point noted above, that we
cannot take peoples views aboutwhat ought to be the case as
equivalent to their views about justice, makesit an open question
how much weight and significance people accord theirviews on
justice.6
One might, for example, hold it to be unjust that the wealthy
should beable to convert their greater ability to pay into superior
health care, oreducation, or to bequeath their wealth to their
children, but nonethelessbelieve that the incentive structure
necessary for a productive economy,given people as they are and the
international economy as it is, requires,as a moral demand, that
such conversions be permitted. Similarly, one mightthink it unjust
that people should command superior rewards for possessingscarce
natural talents but accept, again on incentive grounds, the
on-balanceadvantages of a system that rewards their exercise. If
people can acceptthat justice would require one course of action
but nonetheless regardreasons of justice as trumped by moral
reasons that require another, then
6 I am grateful to Volker Schmidt whose question What do we know
when we knowwhat people believe about justice? alerted me to this
possibility.
-
ADAM SWIFT348
justice need not be overridingly motivating. In which case,
social scientists,supposing them to be concerned ultimately with
the description, explana-tion and prediction of action, should not
necessarily care about justice ratherthan the more general
distributive judgments, and rather vaguer ways ofconceiving
distributive issues, that they are accustomed to investigate.This
possibility suggests that the question of how people conceive
therelation between their beliefs about justice and their other
moral beliefsis one that would itself repay empirical
philosophically informed investigation.
A summary of the argument of this section may be helpful. There
arefour reasons why social scientists might benefit from
philosophy: problemsin identifying inconsistency, the fact that
non-justice considerations mightunderlie distributive judgments,
the way in which different principles ofjustice can yield the same
distributive judgments, and the ambiguity of keyterms such as
desert. The standard survey questionnaire is incapable
ofidentifying popular beliefs about principles of justice specified
with theprecision that interests philosophers, although one
administered after a doseof philosophical education might be able
to do so. The social scientist can,however, reply that she is not
particularly interested in principles of justicestricto sensu, a
reply that gains force from the observation that principlesof
justice are not necessarily decisive for action, even among those
forwhom moral considerations all things considered are indeed
overridinglymotivating.
4. PHILOSOPHY NEEDS SOCIAL SCIENCE?
Despite the prima facie irrelevance of empirical opinion to
normativejustification outlined earlier, and forcibly voiced by our
anonymous referee,recent years have witnessed a variety of
developments on the philosophicalside that seem to indicate an
increasing appreciation of the normativerelevance of empirical
work. According to Miller (1992, p. 556), and citingRawls and
Walzer in support, few contemporary political theorists wouldwish
to draw such a sharp line between common opinion and
theoreticaltruth. Most would claim in one way or another to
incorporate andsystematize existing beliefs about justice in their
theoretical constructions.Indeed for Miller himself, empirical
evidence should play a significantrole in justifying a normative
theory of justice, or to put it another way . . .such a theory is
to be tested, in part, by its correspondence with our
evidenceconcerning everyday beliefs about justice (1994, p.
177).
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 349
I can think of three ways in which empirical research into the
justice-beliefs held by lay social actors might be thought
important to politicalphilosophy. In increasing order of strength,
they are first, that such beliefs,and knowledge of their causal
determinants, provide food for thought;second, that they constitute
feasibility constraints on the realisation of thephilosophically
justifiable distribution; and third, that they are constitutiveof
that distribution itself.
The first, so weak as to be uncontroversial, is simply that
knowingthat others, perhaps the great majority of others, think
differently givesthe philosopher grounds for caution. Elster (1992,
p. 193) observes thatthe knowledge that others hold or practice
very different conceptionsshould make him scrutinize his own
opinions with extra care, and haselsewhere (1995, pp. 9495)
suggested that Frolich and Oppenheimers(1992) finding that
truncated utilitarianism maximising total welfaresubject to a floor
constraint on individual welfare is overwhelmingly themost popular
distribution actually chosen by people in a situationsimulating
Rawls veil of ignorance, gives philosophers reason to accordit more
careful consideration than it has received hitherto. On this
account,they may still end up rejecting popular opinion, but at
least it will havebeen treated with appropriate respect.
If descriptive studies can play this cautionary role, then
research intothe factors that explain why people believe what they
do about distributivejustice can and should be similarly
thought-provoking. For example, wheresociological research reveals
that people are inclined to favour principlesthat correspond to
their self-interest, then philosophers whose own viewsseem to
reflect this process should think again. Of course, they may cometo
the same conclusions: as Elster (1995, p. 94) acknowledges,
correlationsof this kind are not sufficient to invalidate
arguments. But one does nothave time to question all of ones
intuitions equally carefully, and empiricalresearch of an
explanatory kind can help the philosopher identify thosemost worthy
of suspicion.
Descriptive and explanatory empirical research can, then, be
helpful forthe normative project but this role remains merely
external. There is nosuggestion here that what other people think,
or even why they think it,can do more than give the philosopher
reason cautiously to reconsider herown arguments and intuitions.
The second reason why empirical researchmight be of interest to the
political theorist continues to adopt this externalattitude, but
holds that lay beliefs matter insofar as they constitute the
causalfield that confronts his or her attempts to guide action.
Conceptions ofjustice that fail sufficiently to correspond to
ordinary thinking are doomed
-
ADAM SWIFT350
to failure, however sound they may be in philosophical terms,
for, as Dunn(1980, p. 247) observes, if historical agents are to be
provided with reasonsfor acting, they must be furnished with
reasons which are reasons for them(italics in original). The bounds
of political possibility are to a great extentset by popular
opinion, so that judging what can be done politically
requiresknowledge of that opinion. Add to this the claim that it
makes little senseto advocate that which it is impossible to
achieve, or that there is reason tobe concerned with the set of
feasible outcomes given the status quo, andone has a variety of
moral arguments for why political theorists shouldtake lay beliefs
into account. We need not regard those beliefs as makingany
difference to the truth about justice, for our conclusion may well
besimply that justice is unattainable and will remain so while
popular opinionremains as it is, but there is reason to take them
into account when offeringprescriptions here and now.
The third argument as to why popular beliefs might bear on
politicalphilosophies of justice is the rather stronger one that
the distributiveprinciples that are justified for the society in
question may be internallyrelated to lay beliefs themselves. Where
the second argument regardsthose beliefs merely as constraints upon
the feasibility of achieving a justsociety, and the justification
of principles of justice as occurring quiteindependently of popular
beliefs, this third argument claims that suchbeliefs are in some
sense constitutive of a proper understanding of whatjustice
demands.
Three versions of this third line of argument can be identified,
rangingfrom the weak and unobjectionable, through the substantial
and interesting,to the strong and mistaken. The weak and
unobjectionable version saysthat the philosopher who talks about
justice needs to make sure that she istalking about that which is
conventionally referred to by the word justice.This is indeed an
empirical issue, decidable only by attention to the wayin which the
term, and its derivatives, are actually used, in particular
byattention to those commonplaces that are so commonplace as to
becandidates for a priori truths (Jackson and Pettit 1995). A
philosopher whooffers her theory of justice with no claim that it
should motivate those towhom it is addressed, or who offers it as a
theory of how one should behaveto avoid embarrassment, for example,
has not understood what the termmeans. There is, of course, room
for philosophical argument as to whichcommonplaces are indeed a
priori truths, but the question of whichstatements are commonplaces
would seem to be an empirical matter. Thisclaim, then, makes
popular beliefs about justice constitutive of thephilosophically
correct view about justice in the sense that they indicate
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 351
what it is for a belief to be a belief about justice, and not,
for example,about charity, or etiquette. One way of seeing how weak
and unobjection-able this sense is is to notice that both Nozick
and Rawls, despite the hugedifferences in the content of their
theories, satisfy the requirement to attendto the grammar of
justice.
The substantial and interesting version of the constitutive
claim notesthat principles of justice may give some weight to what
people think isjust even where the principles themselves have been
justified in ways thatmake no reference to what people think. For
example, it is plausible to hold,as a principle of justice, that
peoples legitimate expectations constitutevalid claims to reward,
or at least that they should be given some weightin judging what
distributive outcomes are just. Even someone who
regardsconventional desert claims as mistaken may acknowledge that
the fact thata persons actions have been informed by the belief
that they are valid, ina context where all or most other relevant
actors believe the same, ispertinent to the question of what she
should get as a matter of justice. Tothe extent that judging a
persons legitimate expectations depends onknowing what she and
others believe about justice, justice beliefs willclearly be
relevant information for anyone deciding what distributiveoutcomes
are just. What people think is just, then, may enter into
whatjustice requires, not via the claim that what they think is
constitutive ofcorrect principles of justice but via the claim that
there is independent moralreason - a principle of justice justified
independently of popular opinion -to give those views some
weight.
The first two versions of the constitutive claim gives popular
beliefs arole, first in determining what it is for something to be
a theory of justice,and, second, when combined with a principle
giving some weight to popularbeliefs, in making a difference to
what justice requires. The third, andstrongest, version, gives
popular opinion a role in determining the contentof principles of
justice themselves. This is the claim that part of the answerto the
question of what principles should govern the distribution of
goodsin a society is to be found by looking at the way that
ordinary people thinkthat they ought to be distributed. It is in
this direction that recent work inpolitical philosophy has taken a
turn.
Or so it might seem. In fact, as I will now argue, although the
justificatorystrategies adopted by theorists such as Walzer and
Rawls are rathercomplicated, it is fairly clear that neither gives
any genuinely constitutiverole to popular beliefs about
distributive justice. Miller is the theorist whoclaims the greatest
weight for popular opinion, and indeed for publicopinion as that is
revealed by surveys, but even his stated methodology
-
ADAM SWIFT352
seems to me to leave more of a gap between the philosopher and
the publicthan he himself recognises.
Let me start with Walzer, whose claim (1983, p. 9) that all
distributionsare just or unjust relative to the social meanings of
the goods at stake, andwhose stated aim of interpreting to his
fellow citizens the meanings thathe and they share, are probably
the best-known examples of this apparentlyempirical turn. When we
look at Walzers justification of his proclaimedmethod, we find two
distinct arguments. One is quasi-conceptual andsomewhat
relativistic: goods mean different things in different
societies,the proper distribution of a good is one in accordance
with its socialmeaning, and there is no cross-cultural basis on
which to criticise a societysunderstandings of its goods. The other
is democratic: it is respect for laybeliefs that requires the
philosopher to accord them moral weight. If oneunderstands politics
and political theory as in the business of respondingto citizens
wills, then what matters is not so much the
independentphilosophical validity (rightness) of citizens beliefs,
but the fact that theyare theirs (Walzer 1981). Although Walzer
does not seem to notice this,the quasi-conceptual and somewhat
relativistic argument and the democraticargument are quite
different. It is one thing to take shared meanings to
beconstitutive of justice because one lacks a basis for criticism,
quite anotherto give them constitutive weight out of a respect for
those who share them.7
Now it is noticeable that in Spheres of Justice Walzer puts this
secondargument somewhat differently. There (1983, p. 314, p. 320)
he does notmention democracy, but says that what guides his
approach is a decentrespect for the opinions of mankind, and he
grounds this respect in theclaim that it is as culture-producing
creatures that we are one anothersequals. This change of
formulation is significant, for it suggests that Walzeris
uncomfortable using an argument familiar from democratic theory
insupport of a method supposed to yield principles of justice. And
un-comfortable he should be, for making the democratic case for
respectingthe views of ones fellow citizens, or those of another
culture, even wherethose views are mistaken, is surely different
from regarding those viewsas constitutive of justice itself. One
may very well accept the democraticlegitimacy of distributions that
reflect popular opinion without being at
7 Rorty is similarly ambiguous. Although he argues that there is
no way to arrive atprinciples of justice other than by what he
calls (1990, p. 287) a historico-sociologicaldescription of the way
we live now, he does so in a paper asserting the priority
ofdemocracy to philosophy, which suggests that he too can plausibly
be read as arguing thedemocratic, rather than the
anti-foundationalist, case for heeding the beliefs of his
fellowcitizens. For a fuller discussion of Rortys views on these
issues, see Mulhall and Swift(1996, pp. 259275)
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 353
all tempted to regard that opinion as constitutive of the right
answer as towhat distributive justice demands.8 (For discussion of
the so-called paradoxof democracy see Wollheim (1962) and Estlund
(1989)). Research intopublic opinion might thus be important on
democratic grounds the firstopinion pollsters conceived their work
in just such terms, as do those whoadvocate more sophisticated
methods (Fishkin 1995) and in this way thefindings of empirical
research (though not necessarily of social surveys)might indeed be
of significance to the normative political theorist concernedwith
legitimacy. But this need not be to grant its findings any
constitutiverole in a theory of distributive justice. 9
This crucial distinction between justice and legitimacy also
helps inthe understanding of Rawls attitude to popular opinion,
though here thingsare yet more complicated if perhaps less confused
than they were inWalzers case. Where Rorty interprets Rawls claim
that his theory ofjustice is political and not metaphysical as
showing how liberal democracycan get along without philosophical
presuppositions, requiring only historyand sociology (1990, p.
284), in fact Rawls adopts a variant of the positionthat holds that
we have moral reason to give weight to the beliefs thatcitizens
hold, or can hold, in common. Rawls claim is that we should
takeseriously certain fundamental ideas seen as implicit in the
public politicalculture of a democratic society, not simply because
they represent the
8 It may be the right answer as to what procedural justice
demands. If one holds thatwhat makes a decision legitimate is that
it is the outcome of a just procedure, then onewill accept the
legitimacy of a distribution precisely because it satisfies the
demands ofthis kind of justice. But it remains the case that the
answer can be the wrong one in thesense I am discussing. Just as
the verdict of a jury has legitimacy even if it gets it wrong,so an
ideal democracy might generate decisions on distributive matters
that are legiti-mate (because the outcome of a just procedure)
while being quite mistaken about thedemands of distributive
justice. This paper is concerned with beliefs about, and
theirrelation to the justification of, principles of distributive
justice stricto sensu. It is, how-ever, worth pointing out that
peoples beliefs about legitimacy - or about what makes aprocedure
just - can themselves be mistaken. If we have reason to regard
citizens mis-taken decisions about distributive justice as
legitimate, this is because of a correct analy-sis of what makes a
decision legitimate and not because of what they think makes
adecision legitimate. I am grateful to Michael Smith for discussion
on this point.
9 Two further questions cannot be pursued here. First, what kind
of public opinion -how post-deliberation? how philosophically
educated? - is relevant for legitimacy? Sec-ond, to what extent
(and hence by what institutional means) does legitimacy
requirepopular opinion to be accommodated? One might think that
conventional liberal demo-cratic processes give sufficient weight
to the right kind of popular opinion, with politicalparties being
effectively constrained by their desire for re-election to attend
to the find-ings of opinion polls and focus groups. Or one might
argue for a greater role for refer-enda, or for citizens juries and
deliberative polls, and so on.
-
ADAM SWIFT354
way we live now, but rather because we have reason to value a
societythat is publicly justifiable to its members. Baldly
summarising a complexargument, what Rawls (1993) actually argues is
that, since political poweris the exercise of power held by free
and equal citizens, that exercise isonly legitimate when it is used
in ways that can be justified by appeal topublic reason; the
importance of public justifiability leads the politicaltheorist to
societys main institutions, and their accepted forms
ofinterpretation . . . seen as a fund of implicitly shared ideas
and principles(1993, p. 1314); the idea of society as a fair scheme
of cooperationbetween free and equal citizens is such an idea,
which is articulated byRawls in terms of the imaginative construct
of the original position; andfrom that device of representation
emerges a substantive theory ofdistributive justice.
This, then, is a distinct, and much more fully articulated,
variant of thelegitimacy argument strands of which we detected in
Walzer. It too can bethought of as invoking a proper respect for
peoples beliefs, but here therespect is not shown simply by giving
the people what they will as it ison the democratic view but by
insisting that the use of state coercion ispublicly justifiable to
them. This is the liberal principle of legitimacy.As Rawls puts it
(1993, p. 137), our exercise of political power is fullyproper only
when it is exercised in accordance with a constitution
theessentials of which all citizens as free and equal may
reasonably be expectedto endorse in the light of principles and
ideals acceptable to their commonhuman reason. Rawls is thus led to
espouse a methodology constraininghim to the working up of ideas
implicit in our public political culture byhis distinctively
liberal understanding of the proper relation between theindividual
and the state. And unlike Walzer, Rawls does not confuse thedemands
of legitimacy with those of justice. That is why he can continueto
argue for the justice of justice as fairness, difference principle
and all,whilst recognising that other liberal political conceptions
of justice cansatisfy the demands of public reason and hence be
legitimate (Estlund1996).10
So: it is one thing to regard the beliefs of a societys members
aboutprinciples of justice as constitutive of the kinds of
principle that may
10 There is another strand of argument in Rawls that emphasises
the importance ofstability, and this might seem to give him a
further reason to care about popular opinion.The proper role of
stability considerations in Rawls argument is more complicated
thancan adequately be discussed here (see Mulhall and Swift 1996,
pp. 240242), but forpresent purposes it is sufficient to note that
to the extent that they constitute a genuinelydistinct strand of
reasoning they have to do with feasibility.
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 355
legitimately regulate their society, but it is quite another to
regard thosebeliefs as constitutive of the right answer as to what
principles of justiceare correct for that society. This
distinction, between principles beingjustified as legitimate
principles to govern the distribution of benefits andburdens in a
society and their being justified as correct principles of
justice,may seem pedantic, but it is one recognised by all who
acknowledge thejustifiability, on grounds of legitimacy, of laws
that they themselves regard,and would vote against, as unjust.
Walzer veers between a conceptual anda democratic version of the
legitimacy argument, and seems, mistakenly,to regard both as
implying that his method of attending to shared socialmeanings is a
way of justifying principles of justice. Rawls is clearer onthe
distinction between justice and legitimacy, and offers a version of
thelatter that attends less to what people currently believe than
to what can bejustified to them in appropriate terms.
Despite this analysis, it remains worth investigating how these
positionsrelate to empirical research of the kind discussed above.
To what extentare the things that Rawls and Walzer would have
political philosophersconsider really empirical peoples empirical
beliefs about distributivejustice? The answer is very little.
In Walzers case, it is crucial to his overall position though
notobviously consistent with his own methodological injunctions
that hecan distinguish between the social meaning of a good, and
the opinionsthat people currently have about how it should be
distributed. That is whyhe can defend his approach as compatible
with radical social criticism. Heknows, for example, that many of
his fellow Americans think it quite properfor health care to be on
sale in the market, and his argument against thisview in fact
relies upon a claim about the distributive logic of the practiceof
medicine and the inconsistency involved in a societys permitting
amarket whilst simultaneously recognizing health care to be a
need,evidenced by the use of communal funds to finance research,
buildhospitals, underwrite the treatment of the very old, and so on
(Walzer 1983,p. 86, p. 90). As Warnke (1992, p. 18) puts it social
meaning is not a matterof the opinion individuals may have about
goods, institutions and practices;it is a matter of the goods,
institutions and practices themselves about whichindividuals have
opinions. Walzers decent respect for the opinions ofmankind turns
out to have little to do what social scientists usually regardas
public opinion.
Where Walzer claims to be providing a reading of the way we
understandparticular goods in our society, Rawlss text is that much
vaguer thing hecalls the public political culture, and he is less
concerned with peoplescurrent beliefs than with what can be
justified to them by appeal to
-
ADAM SWIFT356
common human reason. While knowledge of others intuitions
aboutjustice can perhaps help in the process of reaching reflective
equilibrium,by providing food for thought as discussed above, and
while it matters thatjustice be presented as an articulation of
implicitly shared ideas andprinciples, it is clear that he takes
his commitment to public justifiabilityto be consistent with
conclusions that are a long way from the beliefs ofmost Americans.
This is most obvious in the case of desert, to which Rawlsis
notoriously hostile, but which the great majority of US citizens
endorse.11
Does Rawls fail to follow through on his own methodological
in-junctions? In Millers view (1994, pp. 181182), fulfilling the
Rawlsiancommitment to public justifiability requires greater
attention to empiricalevidence than is paid by Rawls himself. While
it is possible to distinguishbetween what people now believe about
justice and that which can bejustified using only commonly accepted
modes of argument, it is im-plausible to think that the two will
differ radically. For these kinds ofargument are going to be just
the familiar correction of empirical error, offaulty inference, of
the distorting effect of self-interest, which cannot beexpected to
lead to beliefs that contrast markedly with those already
current.For Miller, if we really care about producing a theory that
can be justifiedto all citizens, then we have to know what those
citizens currently believe,and our conclusions can only be their
beliefs corrected in commonsensicalways. 12
It is not obvious, however, that restricting the philosopher to
the publiclyjustifiable correction of popular opinion entails that
his conclusions willbe similar in content to that opinion, as
Miller seems to suggest, for it is atleast possible that that
opinion itself rests upon faulty inference of the kindthat can be
demonstrated publicly. We can grant that a theory of justiceshould
be understood as bringing out the deep structure of a set of
everydaybeliefs which, on the surface, are to some degree
ambiguous, confused,contradictory (Miller 1994, p. 177) but insist
that this may lead us a long
11 As has been argued elsewhere (Mulhall and Swift 1996, pp.
242245) there is acrucial ambiguity in Rawls notion of public
justifiability. In one sense, what can bepublicly justifiable is
that which appeals to the public political culture as
implicitlyshared. But in another, it amounts to that which, given
the burdens of judgment, itwould be unreasonable to reject. I focus
on the first of these two senses in the text be-cause it is that
which would seem to accord greater importance to public
opinion.
12 Where Miller criticises Rawls for not paying sufficient
attention to popular opinion,Gaus (1996, pp. 131136) regards Rawls
endorsement of the accessibility condition -that public
justifications must rely on methods of reasoning accessible to
others, in par-ticular commonsense reasoning - as committing him to
an excessively populist theory ofjustification. For Gaus,
commonsense reasoning is prone to error, and indeed seems to
beinconsistent with the attitude to disagreement that Rawls himself
regards as reasonable.
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 357
way from the content of mass empirical opinion. Empirical
opinion aboutdistributive justice might be respected, on this view,
in the sense thatarguments are presented as publicly justifiable
corrections to it, but anyplausible account of what public
justification amounts to will leave thephilosopher plenty of room
to argue against public opinion as that isconventionally
understood.
5. THE EXAMPLE OF DESERT
Perhaps the best way to illustrate this point is to consider the
most glaringcase of a discrepancy between lay beliefs about justice
on the one hand,and those endorsed by political philosophers on the
other. As severaltheorists have recently noted, it is on the
question of desert that (to quoteMiller again) much of the recent
theorizing about justice appears to beout step with popular opinion
(1991, p. 372). According to Scheffler(1992), for example, it is
the rejection by contemporary philosophicalliberalism of
traditional notions of desert and responsibility that remainwidely
held in the general population that goes at least some way
towardsexplaining the failure of liberalism as a concrete political
programme.Galston (1991, pp. 159162) too has argued that Rawls
account of justicefails to provide the most plausible description
of the shared understandingof Americas public culture, since there
is in the contemporary UnitedStates a broad consensus that, amongst
other things, regards desert as aproper basis of distribution, and
ability, effort and self-denial as the basesof desert. Again this
is thought to have political implications, since theinadequacies of
Rawls dangerously one-sided reconstruction of the liberaltradition
are mirrored in the national electoral disasters of
contemporaryliberalism.
On the first of three possibilities outlined above, this
discrepancybetween public and professional opinion is no more than
a reason for thephilosopher to think again, and to check his or her
arguments. Respect forlay beliefs might perhaps lead to a
reconsideration of intuitions, or a lookat the variables with which
the rejection of desert commonly correlates inorder to consider
whether they might explain the philosophical positionin a way that
competes with its validity. For Galston (1991, p. 159) thematter
should end here, for his view is not only that the Rawlsian
analysisfails to fit public opinion but also that it does violence
to a reasonableaccount of the moral point of view. Suppose,
however, that, havingreconsidered, the philosopher continues to
reject that traditional thinkingabout responsibility and agency
that underlies the belief in desert. Scheffler(1992, p. 319),
indeed, is keen to make clear his own view that liberals
-
ADAM SWIFT358
may well be right to be sceptical of such thinking. This leaves
the moreinteresting second and third reasons to take such (ex
hypothesi mistaken)opinion into account. On one, the people may be
wrong about desert, butare ignored at the cost of morally serious
political isolation and in-consequence. On the other, the popular
affirmation of desert helps to justifydistributions that accord
with its requirements.
Taking the latter first, no position we have discussed holds
thatdistributions in accordance with traditional desert claims are
just simplybecause people think they are. The theorist who would
have us pay greatestattention to empirical opinion is Miller, for
whom taking public justifiabilityseriously means that justice can
only be popular opinion corrected bycommonly accepted modes of
reasoning. But how should we apply thatclaim to this case? It is at
least arguable that the philosophers rejection oftraditional desert
claims, or at least some of them, itself depends only uponforms of
argument that are widely regarded as acceptable. The idea thatwe
cannot deserve greater (or lesser) rewards than others for the
exerciseof abilities for the greater (or lesser) possession of
which we are notresponsible is commonsensical, and the claim that
at least some of ourdifferential talents are ours by chance is
surely uncontroversial. Some ofcontemporary liberalisms hostility
to desert claims - that component which,in denying even that we
deserve to be rewarded for our efforts, perhapsdoes rest upon a
controversial claim about human agency - may not survivethe filter
of public justification. But there are other purported desert
bases,such as innate differences, that could be rejected on the
basis of nothingmore than common sense.
In the case of desert, then, it seems that even the philosopher
limited tothe task of the publicly justifiable correction of
popular opinion may, paceMiller, argue for a position that differs
radically from it. The demands ofpublic justifiability may rule out
certain kinds of philosophical argument,and what the people think
may enter into the process of justification in thatthe philosopher
must understand, and present herself to be systematising
andbringing out the deep structure latent within popular beliefs.
But the extentof the ambiguity, confusion and inconsistency in
those beliefs is sufficientto leave a great deal of work to be
done, work that may well lead toconclusions differing markedly in
content not only from empirical publicopinion understood as those
beliefs that survey research shows to beregularly and consensually
endorsed by the vast majority of ordinary peoplebut also from the
more reflective beliefs that would result from the researchmethod
advocated earlier in this paper.
Clearly there is a great deal more to be said on the questions
of whatpublic justification amounts to and whether we should indeed
care that our
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 359
theory satisfies its demands.13 Leaving behind such
complexities, let meturn finally to the alternative, strategic,
suggestion that arguing the truthabout desert where this conflicts
with popular opinion may produce, indeedaccording to Scheffler and
Galston has produced, sub-optimal politicaloutcomes. Here it is
tempting to distinguish between the role of thephilosopher, who can
properly transcend such consequentialist con-siderations, and that
of the political actor, who, on some views at least, isirrevocably
caught up in them. (For the classic treatment of this issue,
seeWeber (1948)). While it is a nice question for a political
party, say, whetherit should include in its manifesto commitments
to policies that it believesjust but unpopular, or argue for what
it knows to be unjust in one sphere inorder to gain the power to
bring about justice in others, the philosophermight claim that such
problems are not hers. Her trade is the pursuit andadvertisement of
truth, and the likelihood of others coming to agree withher
conclusions matters not one jot. Or, if she is unwilling to
abjureconsequentialist considerations altogether, she can point out
that, while onecan imagine scientific cases where the consequences
of knowing orpublishing truths were so likely to be so harmful that
the scientist might,on balance, have reason to desist from their
pursuit or publication, it seemsless plausible that the same should
apply to moral truths. (See Greenawalt(1995) for an interesting
discussion).
That said, philosophers may wish to justify their philosophising
on themore positive consequentialist ground that arguing for the
moral truth islikely, in the long run, to bring about better
outcomes than would otherwisebe achieved. Rather than taking the
feasible set as given, it might be thoughtthat it is precisely the
task of the philosopher to change that set, by changingthe content
of those popular beliefs that do so much to determine it. Thisis at
the very least a widespread understanding of the proper role of
thepolitical philosopher, and from this perspective the finding
that most peoplemistakenly believe in conventional desert claims
does not give one reasoneven of a consequentialist kind to pander
to that opinion.
13 What of the philosopher who believes that conventional desert
claims are mistakenbut who accepts that this cannot be demonstrated
in publicly acceptable ways - that itrelies on controversial but
valid forms of argument? Presumably she can, in private lifeas it
were, preach and seek to persuade others of the truth; but, on the
position presentedhere, cannot consider this part of a conception
of justice or vote for it. It should be clearthat this is not Rawls
own position. He thinks that the fact that conventional
desertclaims are mistaken can be shown by appeal to public reason,
and hence voted for, butaccepts that other conceptions of justice,
more sympathetic to such desert claims, canalso be justified in
this way and so must be regarded as legitimate.
-
ADAM SWIFT360
Of course, this particular defence of philosophical immunity
fromempirical opinion relies upon a social-scientific prediction
about the extentto which arguing for the truth will have the effect
of bringing others tobelieve it. Clearly time frames are crucially
important people may believein 200 years what it would be crazy to
expect them to come to believe inthe next decade and the
philosopher defending her radical detachmenton consequentialist
grounds will need to take seriously the moral con-siderations that
they throw up. But the philosopher who cares aboutchanging the
minds of those she addresses, rather than those generationsdown the
line, may well feel it necessary to heed the findings of
empiricalresearch, both descriptive and explanatory. For what
beliefs can plausiblybe changed to depends both on what they
currently are and on the causalprocesses that determine their
formation. The philosopher must knowfrom where it is that her
strategy of change is to start and worry aboutthe causal factors
that help to determine its likely success. What if, forexample,
people would not (and could not) stop believing in
traditionalnotions of desert even when their mistake was pointed
out to them(Strawson 1974)? What is needed, on this view, is a
research programmethat seeks to assess the efficacy of (good)
reasons in the formation ofpeoples justice beliefs or, if that
sounds an implausibly Utopian project,at least one that focusses on
the mechanisms by which such beliefs areformed and can be
changed.
Notice, moreover, that the strategic attitude to empirical
beliefs aboutdistributive justice takes it for granted that such
beliefs really are part ofthe causal field that is to be confronted
and negotiated. Unlike theconstitutive accounts, which give moral
weight to citizens beliefs or whatcan be justified to them simply
in virtue of their being believed orjustifiable, the strategic
perspective only gives one reason to care aboutwhat people think if
that influences their actions. If people do not act tosupport (or
resist) what they believe to be just (or unjust), then there canbe
little reason, from this perspective, to take their declared views
intoaccount. This particular consequentialist defence of
philosophical lack ofattention to current public opinion requires
not only that the philosopherbe able to convert disbelievers, but
also that the truths to which they areconverted are such as to
motivate action. To be sure, some may choose toignore
consequentialist considerations altogether, and others may care
onlythat people come to believe the truth about justice, not that
they aremotivated to act on that truth. But empirical
investigation, not only of thelikelihood of people coming to
believe the truth about justice but also ofthe extent to which
their beliefs about justice have motivational force,
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 361
would seem to be of interest to all but the most other-worldly
of philo-sophers concerned with the demands of distributive
justice.
6. CONCLUSION
If this paper has a single overall message, it is that deciding
the properrelation between social-scientific and philosophical
analyses of distributivejustice requires us to be very careful
about what is meant by principles ofdistributive justice. Such
principles are a rather specific and abstract subsetof moral
considerations. This is a point against those social scientists
whoequate beliefs about justice with all normative beliefs about
distributivequestions, or who fail to see that different principles
of justice may yieldthe same judgment as to the justice of a
particular distribution. Becausethey are such distinctive things,
social scientists may not, on reflection,consider them particularly
important. For example, it is an open empiricalquestion to what
extent peoples actions are best explained by their beliefsabout
principles of justice.
But it is also a point against those political philosophers who,
failing todistinguish between justice and legitimacy, overstate the
constitutive roleof popular beliefs in relation to the former. What
the people think (or arelikely to come to think) justice requires
is important for reasons offeasibility at least to the extent that
what people think about justice affectswhat they do. It matters
also constitutively only if and because we holdmoral principles
that give weight to what they think in deciding whatdistributions
are legitimate. The principle of legitimate expectations cap-tures
this at one level, the principle of democratic legitimacy does so
atanother. In both cases these principles are not justified by
appeal to popularopinion. Popular beliefs about distributive
justice are indeed importantfactors for the political philosopher
to take into account, but for reasonsof feasibility or legitimacy,
not because they play any role in the justificationof principles of
distributive justice.
REFERENCES
Bell, J. and E. Schokkaert, Interdisciplinary theory and
research on justice, in K. R.Scherer (ed.) Justice:
Interdisciplinary Perspectives, Cambridge: Cambridge Univer-sity
Press, 1992, pp. 237253.
Burgoyne, C., A. Swift, and G. Marshall, Inconsistency in
Beliefs about DistributiveJustice: A Cautionary Note, Journal for
the Theory of Social Behaviour 23, (1993),pp. 327342.
-
ADAM SWIFT362
Campbell, T., Humanity before Justice, British Journal of
Political Science 4, (1974),pp. 116.
Dunn, J., Political Obligation in its Historical Context,
Cambridge, Cambridge Univer-sity Press, 1980.
Elster, J., Local Justice, New York: Russell Sage Foundation,
1992.Elster, J., The Empirical Study of Justice, in D. Miller and
M. Walzer (eds.) Justice,
Pluralism and Equality, Oxford: Oxford University Press, 1995,
pp. 8198.Estlund, D., The Persistent Puzzle of the Minority
Democrat, American Philosophical
Quarterly 26, (1989), pp. 143151.Estlund, D., The Survival of
Egalitarian Justice in John Rawlss Political Liberalism,
Journal of Political Philosophy 4, (1996), pp. 6878.Fishkin, J.,
The Voice of the People: Public opinion and Democracy, New Haven:
Yale
University Press, 1995.Frolich, N. and J.A. Oppenheimer,
Choosing Justice: An Experimental Approach to Ethical
Theory, Berkeley: University of California Press, 1992.Galston,
W., Liberal Purposes, Cambridge: Cambridge University Press,
1991.Gaus, G.F., Justificatory Liberalism, Oxford: Oxford
University Press, 1996.Goode, W.J., The Celebration of Heroes:
Prestige as a Social Control System, Berkeley:
University of California Press, 1978.Greenawalt, K., Truth or
Consequences, in I. Shapiro and J. W. DeCew (eds.) Nomos
XXXVII Theory and Practice, New York: New York University Press,
1996, pp. 386399.
Jackson, F. and P. Pettit, Moral Functionalism and Moral
Motivation, PhilosophicalQuarterly 45, (1996), pp. 2040.
Lerner, M.J., The Belief in a Just World: A Fundamental
Delusion, New York: PlenumPress, 1980.
Miller, D., Review Article: Recent Theories of Social Justice,
British Journal of Politi-cal Science 21, (1991), pp. 371391.
Miller, D., Distributive Justice: What the People Think, Ethics
102, (1992), pp. 555593.
Miller, D., Review of K. Scherer (ed.) Justice:
Interdisciplinary Perspectives, SocialJustice Research 7, (1994),
pp. 167188.
Moore, B., Injustice: The Social Bases of Obedience and Revolt,
London: Macmillan,1979.
Mulhall, S. and A. Swift, Liberals and Communitarians, (2nd ed.)
Oxford: Blackwell,1996.
Nozick, R., Anarchy, State and Utopia, Oxford: Blackwell,
1974.Rawls, J., A Theory of Justice, Cambridge MA: Harvard
University Press, 1971.Rawls, J., Political Liberalism, New York:
Columbia University Press, 1993.Reis, H.T., The multidimensionality
of Justice, in R. Folger (ed.) The Sense of Injustice:
Social Psychological Perspectives, New York: Plenum Press, 1984,
pp. 2561.Rorty, R., The Priority of Democracy to Philosophy, in A.
Malachowski (ed.), Reading
Rorty, Oxford: Blackwell, 1990, pp. 279302.Scheffler, S.,
Responsibility, Reactive Attitudes, and Liberalism in Philosophy
and Poli-
tics, Philosophy and Public Affairs 21, (1992), pp.
299323.Strawson. P., Freedom and Resentment and Other Essays,
London: Methuen, 1974.Swift, A., G. Marshall, C. Burgoyne, and D.
Routh, Distributive Justice: Does It Mat-
ter What the People Think?, in J. Kluegel et al (eds.) Social
Justice and Political
-
PUBLIC OPINION AND POLITICAL PHILOSOPHY 363
Change: Public Opinion in Capitalist and Post-Communist States,
New York: Aldinede Gruyter, 1995, pp. 1547.
Tornblom, K., The social psychology of distributive justice, in
K. Scherer (ed.) 1992, pp.177236.
Walzer, M., Philosophy and Democracy, Political Theory 9,
(1981), pp. 379399.Walzer, M., Spheres of Justice, New York: Basic
Books, 1983.Warnke, G., Justice and Interpretation, Oxford:
Blackwell, 1992.Weber, M., Politics as a Vocation and Science as a
Vocation, in H. H. Gerth and C. W.
Mills, (eds), From Max Weber, London: Routledge and Kegan Paul,
1948, pp. 77156.Wollheim, R., A Paradox in the Theory of Democracy,
in P. Laslett and W. G. Runciman
(eds.) Philosophy, Politics and Society, Second Series, Oxford:
Blackwell, 1962, pp.7187.
Balliol College OxfordOxford, OX1 3BJUK
-
ADAM SWIFT364