5 Relativism Cognitive and Moral

8/12/2019 5 Relativism Cognitive and Moral

1/44

Relativism: Cognitive and Moral

Author(s): Steven Lukes and W. G. RuncimanReviewed work(s):Source: Proceedings of the Aristotelian Society, Supplementary Volumes, Vol. 48 (1974), pp.165-189+191-208Published by: Blackwell Publishingon behalf of The Aristotelian SocietyStable URL: http://www.jstor.org/stable/4106865.

Accessed: 01/06/2012 14:22

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

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Blackwell Publishingand The Aristotelian Societyare collaborating with JSTOR to digitize, preserve and

extend access to Proceedings of the Aristotelian Society, Supplementary Volumes.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=blackhttp://www.jstor.org/action/showPublisher?publisherCode=aristotelianhttp://www.jstor.org/stable/4106865?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/4106865?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=aristotelianhttp://www.jstor.org/action/showPublisher?publisherCode=black


2/44

RELATIVISM:COGNITIVEANDMORALStevenLukesand W. G. RuncimanI-StevenLukes

. . . on nevoit rien de juste ou d'injustequi ne changedequaliteenchangeantde climat.Trois degresd'elevationde pole renversent oute la jurisprudence; n meridiendecide de la verite; en peu d'anneesde possession, eslois fondamentales hangent; le droit a ses epoques,I'entree de Saturne au Lion nous marque l'origined'un tel crime. Plaisante ustice qu'uneriviere borneVerite au dea des Pyrenees, rreurau dela(Pascal,Pensees,V, 294).

In this paper I want to considerhow seriouly hesewords ofPascalshould be taken. Howfar-reachingre the implicationsof the relativism hey express?Do they commit us simply toaccepting he empiricalclaim that culturesand their compo-nents areremarkably iverse,or do they commitus further osome philosophical r normativedoctrine? Morespecifically,are truthand logic,morality,evenrationality tself,ultimatelycontext-or culture-or theory-dependent,elative to particularand irreducibly arious formsof life' or systemsof thought?And how wide-rangingre their implications? nsofaras theyreach, do they do so equally to what we call knowledgeandwhat we call morality to 'truth'and to, say, 'justice'Relativismhashad a considerableogue n recent imes,andmany thinkers n different fields have, in varying degrees,yielded to its temptations.For Quine, 'Where t makessenseto apply "true" s to a sentencecouched n the termsof a giventheory and seen from within this theory, completewith itspositedreality'(Word ndObject, echnologyPressand Wiley,New York, I960, p. 24). For Wittgenstein, All testing, allconfirmation nd disconfirmationf a hypothesis akes placealreadywithina system'and the system s 'the element n whichargumentshave their life' (On Certainty, lackwell,Oxford,I969, I05). Under Wittegenstein'snfluence,Peter Winchappliesthis idea to the philosophyof social scienceand D. Z.

I65


3/44

166 I-STEVEN LUKESPhillips to the philosophy of religion. For Winch, 'our idea ofwhat belongs to the realm of reality is given for us in thelanguage that we use' and 'logical relations between proposi-tions themselvesdepend on social relationsbetween men' (TheIdea of a Social Science,Routledge, London, 1958, pp. 15, 126).According to Phillips, 'Religious language is not an interpre-tation of how things are, but determines how things are forthe believer. The saint and the atheist do not interpret thesame world in different ways. They see different worlds '(Faith and Philosophical nquiry,Routledge, London, 1970, p.132.) In the history and philosophy of science, Kuhn andFeyerabend make similar claims. Kuhn says that 'in a sensethat I am unable to explicate further, the proponentsof com-peting paradigms practise their trade in different worlds':in paradigmchoice, 'there is no standardhigher than the assentof the relevant community' (The Structuref ScientificRevolu-tions,University of Chicago Press,Chicago, 1964, pp. 149, 93)-Feyerabend denies that it is 'possible to make a judgment ofverisimilitudexcept within the confines of a particular theory'('Consolations or the Specialist'in I. Lakatosand A. Musgrave,eds., Criticism ndthe Growth f Knowledge,C.U.P. Cambridge,1970, pp. 227-8) and calls for an 'anarchistic epistemology'.Within linguistics,the so-called 'Sapir-Whorfhypothesis'posits'the relativityof all conceptual systems,oursincluded, and theirdependence upon language . . . 'and maintains that 'allobserversare not led by the same physicalevidence to the samepicture of the universe, unless their linguistic backgroundsaresimilar, or can in some way be calibrated'. (J. B. Carroll, ed.,Language,ThoughtndReality,M.I.T. Press,Cambridge,Mass.,1964, p. 214). Within social anthropology Lucien Levy-Bruhlarguedthat primitives'live, think, feel, move and act in a worldwhich at a number of points does not coincide with ours':their reality is itself 'mystical', their logic is 'strangeand evenhostile' to 'our conceptual and logical thought' and they havea view of causality 'of a type other than that familiar to us'(La Mentalit6primitive,Alcan, Paris, 1922, pp. 47, 520, 85),while Ruth Benedict saw in different cultures 'equally validpatterns of life which mankind has created for itself from theraw materials of existence' (Patterns of Culture,Routledge,London, 1935, P. 201). And within the sociology of knowledge,


4/44

RELATIVISM: COGNITIVE AND MORAL 167Mannheim saw the identificationof socially-locatedperspectivesashavingabearingonvalidityand the truthof what men believe.

The temptationof relativism is a powerfuland all-embracingone. If formsof life or systemsof thought are inescapablycon-stitutive of men's perception and their understanding, thensurely their moralities, their religious and their aestheticprinciples will be as relative as their knowledge? Indeed, thesocial anthropologist Mary Douglas, a Durkheimian muchinfluenced by Quine, links the social construction of realitywith boundary-maintaining moral iules and the divisionbetween sacred and profane; conversely, she writes that 'themoral order and the knowledgewhich sustains t are createdbysocial conventions. If their man-made originswere not hidden,they would be stripped of some of their authority' (RulesandMeanings, Penguin Education, Harmondsworth, 1973, P. 15).Thus knowledge, morality, and religion are closely interlinkedand mutually sustaining, and relative to particularsocial con-texts. But equally, there are those who resist the temptation ofsuch ideas by proclaimingobjectivismin morality, religion andknowledgealike. Roger Trigg concluded his recent book ReasonandCommitmentCambridge University Press,Cambridge, 1973,p. 168) by assertingthat without the notion of objectivity,there could be no criteriato distinguish knowledge fromignorance, and human reason becomes impotent. Withit, the claims of religion, the discoveries of science, theassumptions of moral argument, and much else, takeon the importance they deserve.The purposeof this paper is to expressa perplexing sense ofintellectual discomfort at my inability to accept either of theseall-embracingpositions. To put the matter sharply, I can seegood reasons for rejecting cognitive relativism but no over-whelmingly good reasons for rejecting moral relativism. Thisstance is, of course, not unfamiliar, among both social scien-tists and philosophers.Durkheim was firmly committed to thecognitive supremacy of science, while adhering to a certainkind of moral relativism,accordingto which a moralityis a setof 'moral facts', that is socially-given ideals and imperatives,characteristicof a given society of a given type at a given stageof development,which individualscan (cognitively) graspmoreor less adequately. On the other hand, both Max Weber and


5/44

168 I-STEVEN LUKESmost contemporary Anglo-Saxon philosophers tend towardsupholdingthe cognitivesupremacyof the scientific method andthe non-cognitive status of moral judgment based on choicebetween principlesor ideals that are irreduciblyat war. I aminclined to this latter position, though it strikes me as certainlyover-simpleand perhaps ultimately untenable, for the sorts ofreasons that are suggested in Section IV. What follows, then,is a kind of dialogue between the case for combining cognitiveanti-relativism with moral relativism, advanced in the firstthree sections,and two counter-arguments o that case, adum-brated in the fourth.

IBy cognitive relativismI do not mean the empiricalthesis thatthere is a diversity of world-views, theories, forms of explana-tion, modes of classificationand individuation, etc., but ratherthe philosophicalthesis that truth and logic are always relativeto particularsystemsof thoughtor language: on thisview, whatis true and how successfullyto ascertainit, and what is a validor consistentargument are always internal to a system, whichis itself one among others and relative to a particular socialgroupor context or historicalperiod. I and others have arguedthe case against this view elsewhere1and I will here merelybriefly recapitulate its main lines and implications.First, and negatively, the influence, however deep, oftheories,systems,paradigms,perspectives,and so on upon men'sperceptions and understanding is one thing: the relativistclaim that there are no theory-independentobjects of percep-tion and understandingis another. Similarly, the influence oftheoriesupon what men may count as valid or consistent is onething; the relativist claim that validity and consistency aretheory-dependent s another. Moreover,it does not follow fromthe diversityof theories,or indeed from the existenceof differentconcepts and criteriaof truth and validity, that there may notbe some such concepts and criteria which are invariable be-cause universaland fundamental.

(It is, incidentally, strikingthat few relativistsseem able, inthe end, to take the theory-dependenceof their worlds,and thepluralisticsocial solipsismit entails, really seriously.A familiar


6/44

RELATIVISM: COGNITIVE AND MORAL 169pattern of retreatis discernible: witness Kuhn's 'ReflectionsonMy Critics' (Lakatos and Musgrave, op. cit., p. 263), where,for instance, he speaksof the puzzles of normal science being'directly presented by nature', or Mannheim's assertion thatthe 'ultimate criterionof truth or falsity is to be found in theinvestigation of the object' (Ideologyand Utopia,Routledge,London, 1960, p. 4), or Winch's recent denial that he was everarguing 'absurdly, that ways in which men live together cannever be criticised, nor even that a way of living can never becharacterised as in any sense "irrational"' (EthicsandAction,Routledge, London, 1972,P. 3), or Levy-Bruhl'sabandonmentin his late notebooks of the notion of the all-pervasivenessofthe 'mystical',and that of 'pre-logicalmentality'on the groundthat 'the logical structure of the mind is the same in all knownhuman societies' (Les Carnetsde Lucien Livy-Bruhl,P.U.F.,Paris, 1949, p. 61), or Quine's recent claim that 'logical truthis guaranteed under translation'and his proposal to ban 'anymanual of translation that would represent the foreigners ascontradicting our logic (apart perhaps from corrigible confu-sions in complex sentences)',so that, e.g., the Law of ExcludedMiddle is no longer seen as revisable (Philosophy f Logic,Prentice Hall, Englewood Cliffs, 1970, p. 83, et. sq.).Second, and positively there are grounds for supposing thatthere are conceptsand criteriaof truth and of logic that arenottheory-or context-bound,but universaland fundamental. Thetruth of a community'sbeliefs and the validity of their reason-ing cannot be entirely up to them, a function of the norms towhich they conform, the language games they play, the lin-guistic dispositions they exhibit, the paradigms to which theysubscribe.

Briefly (and exceedingly summarily),we can only justifiablyclaim that a community holds beliefs (propositionsaccepted astrue) on the assumptionthat we can translate their language,and we can only do that(on the assumptionthat the meaningof a sentence is given by its truth conditions) if a number ofcircumstances hold. First, they must have beliefs about theworld whose truth conditions are the same for them and for us,since only if this is so can we identify those beliefs. In otherwords, though we and they need not agree about all 'the facts',we must correctly assume that we and they share a reality,


7/44

17o I-STEVEN LUKESwhich is independentof how it is conceived. Secondly, we andthey must share certain forms of behaviour-specifically, theactivitiesof assertingand describing,as opposedto, say, betting,objecting, questioning, etc. Thirdly, not all their logical rulescan be matters of pure convention, since, unless they possess,say, the concept of negation and the laws of identity and non-contradiction,we could never understand their putative beliefs,inferences or arguments.Indeed, we could not then even creditthem with the possibilityof holding beliefs,inferringor arguing,and we could never find their equivalentsof 'and', 'not', 'or','if. . . then. .' etc, whose meaning is in part given by thelogical truths.Of course,they may violate the logical laws withwhich they ordinarily operate, say in ritual contexts, but thespecial mystery or paradox of what they then say gains itsforce from that very fact; and there may be certain limitedlogical divergences (e.g. they may be intuitionists who rejectp v ~p as a logical truth), but these cannot go too far withoutincomprehensibilitysetting in.I further maintain that there are criteria of rationality(specifically,principlesspecifyingwhat counts as a good reasonfor believing something-especially what counts as verifica-tion and falsification)which simply arecriteria of rationality,not merely criteria of rationality within a certain context orsystem, though there may well be alternativeways of arrivingat truths (the oracle, for instance, may be wise). Only on thatassumption is it possible to account for the reasons whichjustify commitment to a belief system as such. On the reverseassumption, there are no reasons which are not internal andrelative to the system itself. The system determines what is areason,and one cannot give a reasonfor accepting,or rejecting,the system. (What makes a Kuhnian 'anomaly' intolerable?The answerto thisquestioncannotbe internal o the paradigm.)To assert the existence of non-context-dependent criteria ofrationality is not, however, to be ethnocentric, since we maywell misconceive or misapply these criteria, and, withincertain domains, such as the traditional diagnosis of disease,they say, traditionalor tribalsocieties)may in these respectsbemore rational than we are, or they may apply alternativecriteria that are as successfulas, or even more successfulthan,ours. Nevertheless, some ways are better (in a non-context-


8/44

RELATIVISM: COGNITIVE AND MORAL 171

dependentsenseof 'better') than othersfor arriving at truths-and, unless we assumethis, we could not satisfactorilyexplainhow belief systemshold togetheror how they change. Thus theweb of Zande witchcraftbeliefs holdstogetherin part by shield-ing its adherentsfromthe perceptionof falsificationand contra-diction (' . . a Zande cannot get out of its meshes because it isthe only world he knows. He cannot think that his thought iswrong'2); likewise, superstitions have been abandoned andscientific theories supersededin consequence,at least in part,of confrontationwith evidence and logical criticism. The ulti-mate consequencesof denying this view have been admirablyexpressedby Michael Dummett, in criticism of Quine's modelof language, as follows:At the worst, it is irremediably conservative, becausethere can be no base from which to criticisewhatever is

generally accepted: we do not really know any of thelanguage unlesswe know all of the language; and we donot know the language until we accept as true every-thing that is so accepted by its speakers,since, until wewe do, we cannot have the same linguistic dispositionsas they. At the best, it is simply defeatist: it rendersinprincipleinscrutablethe lawswhich governthe commonacceptanceof a statement as trueor its laterdemotion.Ineither case it is, in effect though not in intention, anti-intellectual; for it stigmatizesas misguided any attempteither to discover or to impose such laws. (Frege,Duckworth, London, 1973, p. 627).In summary, then, I claim that there are conditions oftruth, rules of logic and criteria of rationality which are uni-versal and fundamental. They are universal, in that they existand are operative within all languages and cultures. They arefundamentalin two senses.First, they specify the ultimate con-straintsto which all thought is subject. Thus all societies, withlanguages expressingbeliefs,mustapplythem in general,thoughthey may violate them in particular;indeed, it could be arguedthat they are basicadaptive mechanismsforany human society.But they are also fundamental in a second sense: namely, thatit can probably be shown that those concepts of truth, rules ofreasoningand criteria of rationalitywhich are at variance withthese (above all, in ritual and ideological contexts), are in fact7


9/44

172 I-STEVEN LUKES

parasitic upon them. That is, where there are second-ordernative beliefsabout what counts as true or valid or what countsas a good reason for holding a belief which are at odds withthese basic principles, then those beliefs can only be renderedfully intelligible against the backgroundof such principles.3It has been argued against this general position that it al-together begs the question, since it assumes relativism to beincorrect: in particular, it assumes that we can successfullyidentify their beliefs, follow their reasoning and understandtheir reasons.To this there are three replies. First, to developand pressas far as they will go transcendentalargumentssuchas those suggested here, which seek to establish the precondi-tions for transcultural(and, by extension, intracultural) com-munication. Second, if this still fails to convince the reallyhard-boiled relativist, stubbornly committed to philosophicdoubt, then he must be made to acknowledge the solipsisticconclusionsto which his doctrine, taken seriously,finally leads.Third, I would stress the sociological significance of denyingcognitive relativism. Only such a denial makes it possible toexamine-indeed acknowledge the possibility of-false con-sciousness, where men's beliefs about their own and othersocieties can be characterised as mistaken or distorted or em-pirically inadequate, and, in virtue of these features, havesignificantsocial and political consequences.Only by assumingthat one has access to a reliable, non-relativemeans of identi-fying a disjunction between social consciousness,on the onehand, and social realities, on the other, is it possible even toraisequestionsabout the waysin which misperceptionsand mis-understandingsof all kindsarise and play theirpart in prevent-ing, or promoting, social change. (Of this, more later). Simi-larly, only the application of non-context- or non-system-dependent rules of logic allow one to investigate the social r6leof absurdity.' Finally, only a denial of cognitive relativismallows one to raise questions about the differential cognitivesuccess of different societies in different domains, and seek toexplain these.5


10/44


III now turn to moralor ethical relativism. The followingdistinctdoctrineshave all been brought under this label:(i) the empirical thesis that moral values and principles conflictin a fundamentalway,that is,that they are not merelydifferent,but incompatible: the conflict between them is not resolvablebybeing reduced to a factualdisputethat is in turnresolvable,orbysubsumingone principleunder the other, or both undera third.When the conflict is seen as occurring across cultural linesbetween moral codes, this becomes the familiar doctrine of'culturalrelativism'. In general, it is an empirical thesis aboutthe diversityof moralsand the nature and distributionof moralconflict.(ii) the philosophicalthesis that there is in principle ultimatelyno rational way of resolving fundamental conflicts betweenmoral values, beliefs, principles, codes, systems-that there isno warrant (or no warrantnot itself internal and relative to aparticularmoral system)for counting a particularset of moralvalues, beliefs, etc. as true, valid, correct, objective, etc. AsWestermarckwrote, this thesis impliesthat there is no objective standard of morality, andobjectivity presupposesuniversality. As truth is one itit has to be the same for any one who knows it, and ifmoralityis a matter of truth andfalsity,in the normativesenseof the terms,the same must be the case with moraltruth (Ethical Relativity, Kegan Paul, London, 1932,p. 183).The thesis that it is not has been variously called ethicalsubjectivism,meta-ethical relativism and axiological or valuerelativism: it is with this that I shall be concerned.(iii) the normativethesis that an act is, say right or wrong, goodor bad, or a person, say, praise-or blameworthy if and onlyif he sojudges-or, in the culturalform, if his society sojudges.This thesisamounts to the making f moraljudgments, by syste-matically adopting the values and principles of the actor orhis society. Bernard Williams calls a versionof this 'the anthro-pologist's heresy, possibly the most absurd view to have beenadvanced even in moral philosophy' and characterises it 'inits vulgar and unregenerate form', which is also 'the most


11/44

174 I-STEVEN LUKESdistinctiveand the most influential',as advancingthreeproposi-tions: 'that "right"means (can only be coherentlyunderstoodas meaning) "right for a given society"; that "rightfor a givensociety" is to be understoodin a functionalist sense; and that(therefore) it is wrong for people in one society to condemn,interfere with, etc., the values of another society' (Morality,Penguin,Harmondsworth,1973,P-34).Williamseasilyshowstheinconsistencyof this composite doctrine: the third propositionuses a non-relativesense of 'wrong' not allowed for in the firstproposition. But in a simpler and not inconsistent form, thisversion of relativism is just the first-order (ultra-liberal orromantic) moral position of systematic acceptance: 'When inRome judge as the Romans judge'. This is sometimes callednormativeelativism:t follows from neither of the other twotheses and is certainly absurd.These various theses must be distinguishedfrom a furtherview, which is not relativist at all but often supposed to be:that moral judgments must take account of context and con-sequences. This is what John Ladd calls applicationalelativity:it 'entails that a certain act which might be wrong in one set ofcircumstances could be right in other circumstances' (TheStructuref a MoralCode,HarvardUniversityPress,Cambridge,Mass., I957, p. 327)-or, one might add, that it might be(equally wrong, but) blameworthy in the one and not in theother.

Now, it is clear that the first, negative, argument which Iused when discussing cognitive relativismhas an exact parallelin relation to moral relativism(by which I henceforthmean thesecond, philosophical, thesis distinguished above). The in-fluence of moral codes, ethical systems, ways of life, etc. onmen's actual moral judgments and actions is one thing; theclaim that there are no correct moraljudgments or objectivelyright actionsis another. It does not, in otherwords,follow fromthe diversityof morals,or indeed from the existenceof differentmoral concepts and criteria, that there are no such conceptsand criteriawhich are invariable becauseuniversaland funda-mental. In other words, descriptiverelativism does not entailmeta-ethical relativism.The problemis to discoveranalogous argumentsof a positivenature which show that therearesuch invariable,universaland


12/44


fundamental concepts and criteria. Without such arguments,meta-ethical relativism-the view that there are no extra- ortrans-systemic roundsforcriticisingmoral beliefs and actions orsystemsas a whole-appears convincing, or at least unrefuted.But, in the firstplace, it does not seem to be the case that wemust assume a common set of objective moral truths in orderto translateand identifyalien moral beliefs. Can we not identifythese, as John Ladd suggests (op. cit., pp. 107-8) by dis-covering the prescriptionsfor conduct that have an especial'superiorityand legitimacy' in a culture, such as those con-tained in 'the relatively formal discoursesgiven by leaders totheir people on the conduct expected of them', or 'in talks tochildren ... conductedin a moreseriousatmosphere',or in thewords andjudgments of respected'wise men' who give 'formalmoral talks at importantgatherings: at weddings, curing cere-monies, beforeand aftera person'sdeath, as well as during theairingof disputes'?Canwe not ascertainacommunity'sorgroup'smoral concepts by discovering the regulative concepts whichits members use to characterise and evaluate the activities andrelationsof central concern to them and to guide their actionsin respect of those activities and relations? Some of these con-cepts may be identical to ours: others may be alternative,perhaps incompatible, interpretationsof what are recognisablythe same concepts, and thereforetranslatable,with caution, byour terms (Italian 'pride',Japanese 'honour',Barotse'justice');and yet others may be distinct concepts for which we have noequivalentsin our own societies.The point is that there appearto be no determinable theoretical limits to what concepts asociety or group, let alone an individual, can employ in thisway. Some might suggest the requirementsset by the need forsocietal survival.But it is not obviousthat all societies, let alonegroups and individuals within them, give priority to this need;and, in any case, as ColinTurnbull's TheMountain eople Cape,London, 1973) horrifyinglyshows,a societyof sortscan surviveby following codes of behaviour that are far indeed from thoseto which we commonly attach value.Nor, secondly, does there appear to be a distinctively morallogic, analogous to but distinct from the basic rules of logicdiscussedin Section I, which we must assume to exist in com-mon between 'us' and 'them' in order that we should be able


13/44

176 I-STEVEN LUKESto identify their moral beliefsand arguments.And finally, thereappear to be no distinctivelymoral criteria of rationality, thatis, principlesspecifyingwhat are to count as good reasons formoraljudgmentsoractions,whicharenotinternal to a particularmoralbelief system.There are of course,thoseprinciplesalreadyconsidered under the heading of cognition-specifically theprinciplesof logic and of verificationand falsification.But theseare not distinctive of moral thought, but rather of thought ingeneral-which is not to deny that it is both possible andimportant to apply them in criticism of moral codes. Muchsocial criticism within our societies takes the form of showingup inconsistenciesbetween principles and attitudes or policies,as, for example, in the debate over capital punishment or theattack on racial discrimination in purportedly egalitariansocieties.6Similarly, moral codes can be shown to be based ondistorted or mistaken empirical beliefs (e.g. about race) oron false assumptions about the efficacy of certain practices(e.g. about the deterrent effects of certain forms of punish-ment).But where are the further limits to what may count as agood reason for reaching moral conclusions? In a theocraticculture divine commands will provide such reasons, in agerontocratic culture the authority of the elders and in abibliocentricone that of the Book,undernationalismthe higherends of the collectivity and under official Communism thoseof the Party, in a market society individualistic utilitariancalculation, in a Mediterraneansociety the fear of shame or thedesire for honour, etc. Nevertheless, it is often suggested thatthere are certain universal rational principles which governmoral discourse-e.g. that moraljudgments be universalisable,or impartial, or critical, or reflective, or that they be directedto the maximisationof human welfare,or that they ministertohuman needs or answer human interests,or that they respecthuman beings as persons or rational agents or ends in them-selves, etc. But it is not difficultto producehistoricaland ethno-graphic evidence of putative moral systemswhich violate eachof these principles (and are these latter not, in many cases,empty, since what counts as, say, 'criticalness'or 'needs' or'interests'willbe largely,perhapsentirely,context-dependent?).And is it not too easy a solution to say that these are not then


14/44


genuine moralities, since that is simply to refute moral relati-vism by definition? Still, it could be argued that some subsetof the rational principles cited above, and others, must beappealed to by judgments or imperativesif they are to count asrecognisablymoral, though none need be common to all (i.e.that there is a 'family resemblance'between moralities). But,even if this argument be accepted, there would still be veryconsiderable scope, not merely for the interpretation of, butfor the differentialweighting of these supposedly distinguishingmarksof morality.

Furthermore,this seeminglyirresolvablepolyarchyof moral-ities is an inherent feature of the moral world we inhabit: asMax Weber observed'the ultimately possibleattitudes towardslife are irreconcilable' (FromMax Weber, d. Gerth and Mills,Routledge,London,1948,p. 152). Of course,as I have suggestedabove, the mere fact of moral diversitydoes not of itself entailthat fundamental moral conflictsare not rationally resolvable:one of the contending moralprinciplesorjudgments may, afterall, be 'correct' or 'valid' and others 'incorrect' or 'mistaken'.But in the absence of convincing arguments (that is, argumentsthat are not themselves rationally contestable) for this con-clusion, or even for the pragmatic necessityof assumingit, thecase for meta-ethical relativismsurvives.

Indeed, it may seem to be supportedby a notable feature ofmoral concepts,at least in modern, morally polyarchic societies-that they are 'essentially contestable', inevitably involving'endless disputes about their proper uses on the part of theirusers'7,suchthat to engage in such disputesis itselfto engage inmoral argument. I think it is possible, and would be highlyinstructive, to identify distinct and systematically conflictingmoralities within contemporary societies which are, in aKuhnian sense, 'incommensurable' (this being an area ofwhich both moral philosophers and sociologists are curiouslyshy). Recall Auden the Arcadian's response to the Utopian:' . .. between my Eden, and his New Jerusalem, no treaty isnegotiable' ('Vespers' n Collected horter oems,Faber, London,1966, p. 334); and consider the fundamental conflictsoccurringat the present time within Britain over the permissible limitsof social inequality and the requirements of social justice.These moralities are incommensurablein that adjudicationsin


15/44

178 I-STEVEN LUKESfavour of one rather than another are themselves alwaysrationally contestable and are always made from within aparticularmoralview, though, of course,moralviewscan them-selves be changed from within, and indeed in the direction ofwhat we call enlightenmentand progress.But that udgment isitself system-dependent.It is one morality among others whichcounts universalisability,impartiality and concern with whatit conceives as welfare, respect for persons, even criticalnessorreflectivenessas commendable or important features. Despitethe rationallycontestableclaims of that morality, such featuresappear to have a transcendental status only from within itsperspective. As Westermarckshowed, this is a claim charac-teristic of all moral systems, whose proponents 'believe thatmoral judgments possess an objective validity which none ofthem has been able to prove' (op.cit., pp. 44-5). In the absenceof such proof, one can only conclude that in morality there isno Archimedeanpoint.

IIII have arguedthat with respectto our knowledgeof the world,truth is distinguishable from error because there are non-relative truth conditions, non-relative principles of reasoningand ways ofjustifying claims to such knowledgethat are objec-tively better than other ways. By contrast, moral judgmentsmay be incompatible but equally rational, because the criteriaof rationalityandjustificationin morals are themselves relativeto conflicting and irreconcilable perspectives. I now proposeto illustrate these contentions by taking a much-discussedsociological topic, about which Mr. Runciman has written aninteresting and challenging book (RelativeDeprivationandSocialJustice,Routledge, London, 1966), namely the questionof social inequality.If we ask, first, the question, "What forms of inequality existin a given society?", there will inevitably be a host of complexconceptual and methodological difficulties.Nevertheless, thequestion is in principle an answerable one. Given all the prob-lems of how to classify and compare data, how to interpretofficial statistics, etc., it is not impossible to document in-equalities of income, wealth, education, status, etc., though of


16/44


course the sociological and political significance of these dataand the interrelationsbetween them will be matters of dispute.What is striking is that these sociological findings are at vari-ance with the systematically-patternedsocial perceptions ofgroups or strata within the society. Runciman seeks to show,among otherthings,that the referencegroupswith whichpeopleidentify themselves strongly affect their perceptions of thesocial order, leading them to misperceive it in the sense ofhaving empirically nadequatebeliefsaboutit. Thus, Runcimanwrites:Given the actual distribution of wealth, the answer tothe question ["Whatsortof people doyou thinkaredoingnoticeably better than you and your family?'] whichwould most obviously and naturally accord with thefacts of inequality would be a referenceby both manualand non-manual workers to those in business or the

professions.Although a few manual workersare earningmore than some non-manual, the incomes of very manymembersof the non-manual stratum are far above thoseof even the most prosperousmanual workers. But whenaskeda question directly tied to inequalitiesof class,fewmembers of the manual stratum drew a comparisonfromtheothersideofthemanual/non-manual ine (p. 196).'Not only,' writes Runciman, 'are comparative referencegroups not chosen in accordance with the facts of inequality,but such a correspondencewith the facts is least likely of allamong those who are in fact most unequally placed' (p. 21o).In otherwords,Runciman claims to have shownthat manualworkers,for historical reasons and specificallywith respect toincome and wealth, make comparisons which serve to limitand distort their awareness of the structuredinequality of thesocial order as a whole. (This is, in fact, only one possible inter-pretationof Runciman's findings.Another is that they have thesame awarenessof inequalities, but only regard some as salientto them. Unfortunately the survey on which Runciman reliesis too crude to distinguish between these two interpretationsand he does not distinguishbetween them in his analysis.)Thissupposedlylimited or distortedawareness is false consciousnessin a very simplesense, though it is not for that reasonsociologi-cally or politically unimportant.To take another example, the8


17/44

180 I-STEVEN LUKES

Polish sociologist Stanislaw Ossowskiwrites of the 'difficultieswith which Communistideology has to cope in connexionwiththe changes which have taken place in the socialist society' asbeing 'no less than thosewhich the AmericanCreed has encoun-tered in its collision with the American reality': in 'bothcountriesthe view of their own society is based on the assump-tion that even widely ranging sharesin the national income arenot sufficient to establish social stratification, nor do theynecessarilycause either class antagonismsor other symptomscharacteristic of a class structure'. But Ossowski clearly be-lieved that such 'ways of viewing concrete societies'amount to'stereotypes and social fictions'-to ideology in the Marxistsense. Forthe objective reality with which these ways of viewingare concerned may impose an interpretationwhich isvery far removed from that which a classless societywould require.But from the viewpoint of the interests ofprivileged and ruling groups the utility of presenting

one's own society in terms of a non-egalitarianclasslesssociety is apparent. In the world of today, both in thebourgeoisemocraciesand the people'sdemocracies,sucha presentation affords no bases for group solidarityamongst the underprivileged; it inclines them to en-deavour to improve their fortunes,and to seekupwardsocialmobility by means of personaleffortand their ownindustry, and not by collective action (ClassStructurenthe Social Consciousness,outledge, London, I963, pp.115-1I7, 154).Here, then, are two relatively simple examples of false con-sciousness (which both have significant effects in helping toprevent social change). In Runciman's case it is a matter of alimited and distorted view of the pattern of inequalities; inOssowski's case it is a systematically propagated ideologicalview, which is objectively mistaken (at variance with what'objective reality imposes') about the causes and consequencesof such inequalities.More complexformsof false-consciousness,as found in the writings of Marx, involve an empirically in-adequate understandingof the deeper structuresand processesunderlyingsocial and economicrelationships,and a consequent'reification' of those relationships and blindness to historical


18/44

RELATIVISM: COGNITIVE AND MORAL 181possibilitiesof social change. Thus, for Marx, nineteenth-cen-tury political economy gave a superficial explanation of theworkingsof capitalism, and therebyprecludedthose convincedby it from conceiving of the possibility of capitalism's super-cession by socialism.Let us return to Runciman's book. In it he writes that once'the structureof a society has been examined and its patternofinequalities mapped out, two questions at once arise, eitherof which leads in turn to the other: first, what is the relationbetween institutionalized inequalities and the awareness orresentment of them? And second, which, if any, of these in-equalities ought to be perceived and resented-whether theyare or not-by the standards of socialjustice?' (pp. 3-4). Thelast part of his book is devoted to trying to answer this secondquestion. In it he usesRawls'stheoryof socialjustice to 'classifyfeelingsof relativedeprivationas "legitimate"or "illegitimate"by the standardsof social justice' (p. 284): to demonstrate inprinciple what kinds of grievancescould be vindicated as legi-timate and what reference group choices could therefore bedescribed by this standald as 'correct', and to reveal 'falseconsciousness' in the form of'attitudes to social inequality'which are 'restrictedor mistaken' (p. 292). This requiresthatone establish which inequalities are vindicated by 'the canonsofjustice'; these are establishedby a theoryofjustice whichcan provide an adequate assessmentof relative depriva-tion, and in so doing restate the 'false consciousness'argument in an appropriateform. Once given a theoryof justice, there is a valid sense in which the perceptionor resentment of inequalities can be described as mis-guided ovel and above the sense of ignorance of observ-able facts, or expedient means. This is not becausepeople's interests can be shown to be other than theythink-because, for example, their location in societyinhibits them from accepting the Marxist theory ofhistory and thereby modifying their idea of what is totheir advantage. The perception of inequalities can beshown to be misguided only in the different sense thatif people resent inequalities which are not unjust, theyare illegitimately resenting them; and if they accept orare unaware of inequalities which are unjust, they are


19/44

182 I-STEVEN LUKES

waiving, as it were, a right to resent them (pp. 25 I-2).So Runciman tries to showhow the inequalities of class and status for which someempiricalevidence has been presentedcould be assessedin the light of principles agreed under the conditionsofRawls's model. Are the inequalities which have beendiscussedsuchthat the referencegroupschosenrepresenta correctassessmentof entitlement, or are these choicessymptomaticof an inhibited or distorted recognition ofhow far the social structureis unjust? Ought more, orless, manual workersand their wives to feel that theyare justly rewardedby comparisonwith others, or doesthe disapprovalfound among some members of the non-manual stratum express a legitimate grievance againstnarrowing differentials of class? Should manual andnon-manual work be accorded equal prestige, or doesjustice permit certain kinds of talent or position to bemore highly regardedthan others? (p. 260).

He concludes his discussion by sketching a picture of 'a justsociety with the social and economic lineaments of twentiethcentury Britain', in which there would be less inequality ofwealth, no inherited privilege, no educational discriminationagainst social groups or the economically disadvantaged, nounearnedincome except on the basisof need, no inequalitiesofreward except those based on need, merit and contributionto the common good, equality of opportunity,no deference notbased on praise, the authority of positions to be mutuallyagreeable in advance of their being occupied, maximum con-sultation before administrativedecisions, and unlimited com-parisons between social positions in the bringing of claimsagainst one another (p. 291).Now, one strikingfeature of this picture is its essential con-testability. It may share with Rawls's own applications of histheory the feature of being in accord with what Rawls calls'our' intuitions and 'our considered udgments'-namely, thoseembodied in a wide liberal-social-democraticconsensus-butit is no less contestable for that. Ultra-conservatives,clericalauthoritarians, Empire Loyalists, fascists, racial separatists,Saint-Simonian technocrats, individualist liberals, anarchists,


20/44


radical egalitarians derive from 'the canons of justice' signifi-cantly different images of 'thejust society'.Yet Rawls's aim is, precisely, to eliminate this very contest-ability. The original position affords a standpoint that is'objective and expresses our autonomy . . . to see our placefrom the perspectiveof this position is to see it subspecie eterni-tatis: it is to regard the human situation not only from allsocial but also from all temporal points of view' (A TheoryfJustice,Clarendon, Oxford, 1972, p. 587). In A Theory fJusticehe claims to have found 'an Archimedeanpoint forjudging thebasic stucture of society' according to the principles of justice(p. 584). His central assumptionis that these principlescan bespecified and rendered determinate through rational inquiry:that an intensive processof ratiocination based on 'knowledgeof the generalfactsabout human society' (p. 137) can lead oneto a single structured set of principles, whihc underlie 'our'sense of justice ('the unique solution to the problemset by theoriginal position' (p. I 19). He seeks to validate his theory of

justice by a controlled thought experiment which he believescan establishthe 'basic principles'which determine our every-dayjudgments of what isjust. But Rawls's thought experimentcannot establish the truth of the theory it advances. Thematerials used in that experiment are 'theoretically definedindividuals' (p. I47) rationally pursuing their potential in-terests under hypothetical constraints.These 'individuals' aresupposedly abstracted from any social and historical context.But such abstract individuals are literally inconceivable: withall historicallyand socially specificdeterminantsremoved,theybecome what F. H. Bradley called 'a theoretical attempt toisolate what cannot be isolated' (EthicalStudies,Clarendon,Oxford, 1927, p. I71). It is, therefore, not merely contingentlybut necessarily the case that, in practice, Rawls endows the'individuals' in the Original Position with historicallyspecificand socially located features. (And this is so with all social con-tract theorists,and indeed all those who base their theoriesonthe nature of abstract (pre-social, trans-social or non-social)'individuals'. Such 'human nature' always turns out to belongto a particularkind of social man8).Thus the motivations, beliefsand indeed the very rationalityof Rawls's 'individuals' are recognisably those of a subclassof


21/44

r84 I-STEVEN LUKESrather cautious, modern, Western, liberal, democratic, indi-vidualistic men. They are 'committed to differentconceptionsof the good' (p. 327), so they 'put forwardconflicting claims'(p. 128) and are not preparedto abandon theirinterests p. 129),they have 'general desires'for 'primary goods' (p. 92) (whichare clearly culture-specific), they 'tend to love, cherish, andsupportwhateveraffirms heir own good' (p. i77), theydemandequality of opportunity, but regard unequal rewards as neces-sary incentives (see p. 315), and their rationality consists inacquiring the means of furthering their ends, and, morespecifically, in a safety-first policy of maximising the benefitsof the worst possible outcome. They 'understand politicalaffairs and the principles of economic theory; they know thebasesof social organizationand the laws of human psychology'(p. 137)--but are thesenot culturallyspecificbeliefs? And why,for example, should they, as Rawls maintains, rule out a hier-archical society as potentially just, or regard Aquinas's in-tolerance of heresy as irrational (see p. 215)? Because theirvalues and their conceptions of rationality are distinctivelythose of certain modern Western men. In general, Rawlsappealsto 'our' intuitionsand claimsthat his theoryis 'a theoryof our moral sentiments as manifested by our consideredjudgments in reflective equilibrium' (p. 120). Yet 'we' mani-festlydo not all agreeand are in any caseonly a tiny segmentofthe human race. So, in the end, Rawls's Archimedean pointfor 'judging the basic structureof society' (necessarily)eludeshim, and his achievement is narrower than and different fromhis aim-namely, to have produced a liberal democratictheory ofjustice. Justice is an essentiallycontestedconcept andevery theory of justice ariseswithin and expressesa particularmoral and political perspective.As a matter of fact, Runciman makes little use of Rawls'sthought experiment of the Original Position and seeks ratherto treat the question of what set of arrangementswould beregarded as just by disinterested persons 'in a deliberatelyempirical manner' by postulating what actual contemporarymen, supposed 'temporarily amnesiac' about their interests,would be likely rationally to agree about: i.e. he asks 'whatprincipleswouldinfacthave been selected...in sucha situation?'('False Consciousness' n Sociologyn its Place and OtherEssays,


22/44

RELATIVISM: COGNITIVE AND MORAL 185C.U.P., Cambridge, 1970, Pp. 220-1). But without some im-plausible assumptionof a shared 'moral sense', the hypothesisof such an agreement is quite arbitrary. More relevantly, itlacks even the surfaceplausibility of Rawls's heroic attempt toattain an Archimedean point with impossible hypotheticalabstract 'individuals', since it involves postulating what actualconcrete contemporary individuals would agree on underhypothetical conditions in the light of their given, specificmoral perspectives.Runciman seems to acknowledge at a number of points theessential contestabilityof the concept of justice. There are, hewrites, 'rival theoriesof socialjustice' and 'alternativeviews ofhow [the relation between inequality and grievance] should beassessed by the standard of justice' (RelativeDeprivation ndSocialJustice, pp. 295, 8). But he also writes of that assessmentas an 'answerable question' (p. 295) and, as we have seen,characterises certain attitudes to inequalities as 'false con-sciousness',as 'incorrect','restricted', mistaken',etc. But if themoral relativistpositiontakenabove is inescapablycorrect,thenRunciman's attempt to assess feelings of relative deprivationin the light of social justice fails to reveal 'false consciousness'.It simply shows that from within a certain moral and politicalperspectivesuch feelingswill be evaluated according to certainprinciples.But that perspectivehas no privilegedstatus.

IVI have argued that there is (or rather that we must assumethere to be) an Archimedean point in matters of knowledgebut that there appears to be no such Archimedean point inmattersof morality.Why, then, do I find this double claim un-comfortable?

Because, in the firstplace, it appears to rest on too simple adistinction between fact and value. The problem here arisesespecially acutely with respect to the identification of socialfacts. Is it not possible that, within certain ranges, this is (per-haps must be) always done fromwithin a particularmoral andpolitical perspective? If so, then, within those ranges, everydescription of a social phenomenon, every identification of asocial fact will be value-laden, at least in the sense of ruling


23/44

186 I-STEVEN LUKEScertain moral and political evaluations out of court. Perhaps,in other words, certain theory-laden identifications of socialfacts are inextricable fusions of descriptionand evaluation, inthe sense of presupposinga 'given framework'which 'restrictsthe range of value positionswhich can be defensiblyadopted'.Let us take as an example the identificationof an exercise ofpower within a society.10A concept of power very widely usedin contemporary political science is the following: that AexercisespoweroverBwhenAaffectsB in a mannercontrary o B'sinterests.Now the notion of 'interests'(like that of 'needs' is anirreducibly evaluative notion: if I say something is in yourinterests, I imply that you have a primaacie claim to it, and ifI say that 'policy x is in A's interest', this constitutes a primafacie justification for that policy. In general, talk of interestsprovides a licence for makingjudgments of a moral and poli-tical character.So it is not surprisingthat differentconceptionsof what human interests areare associatedwith different moraland political positions.One can distinguish (somewhatcrudely)between the following three conceptions of what interests are:(I) the liberalconception, which relates men's interests to whatthey actually want or prefer, to their policy preferences asmanifested by their political participation; (2) the reformistconception which, deploringthat not all men's wants are givenequal weight within the political system, also relates their in-terests to what they actually want and prefer, but allows thatthis may be revealed in the form of deflected, submerged orconcealed wants and preferences; and (3) the radicalconcep-tion, which maintains that men's wants may themselves be aproduct of a systemwhich worksagainst their interestsand, insuch cases, relatesthe latter to what men would want and pre-fer, were they able to make the choice. In other words, each ofthese conceptions of interestspicks out a certain range of theentire class of actual and possible wants as the relevant objectof moral appraisal;and that selection is itselfa matterof moraland political dispute.In the light of this, and of the concept of power as definedabove, it will be clear that differentconceptionsof what are tocount as interestswill yield differentwaysof identifying power.And this is indeed what one observes in practice. A politicalscientist operatingwith a purely liberal conception of interests,


24/44


will only see power where there is a conflict of overt pre-ferences between A and B, and A prevails.Another,who allowsthat B's preferencesmay be submerged,will cast his net wider.A third, who is ready to allow that power can be exercisedagainst B's real interests (which may not be manifest in andmay even conflict with his actual wants), will see power whereneither of the other two see it. Moreover, these differencesofempirical scope are essentiallylinkedto differentvalue assump-tions: in each case these latterpredeterminethe concept'srangeof empirical application. From which I conclude that the con-cept of power too is essentiallycontested,and that what, on theface of it, looked like an empiricallydecidable matter (answer-ing the question: 'is this an exercise of power ?') turns out oninspection to be ineradicably evaluative-and necessarily so,since it appears that any way of identifying power rests uponsome normatively specific conception of interests,and conflictswith others.Thus the first cause for discomfortin being a cognitive anti-relativist but a moral relativist is that there may be, at leastwithin certainranges,no morally (and politically) neutral formof cognition of social facts: the concepts available for identify-ing them may be as essentially contested as I have claimedmoral concepts to be.The second reason for discomfortmay be seen as arisingfromthe fact that contests over the latter are, after all, contests oversomething: essentially contested concepts must have somecommon core; otherwise, how could we justifiably claim thatthe contestswere about the same concept? Implicit in the posi-tion I have taken above is the idea that the concept of moralityis itself essentially contestable: that the criteria determiningwhat counts as 'moral', the objects of moral judgment, theformsof moraljustification, etc., are to be seen in a pluralisticmanner as irreducibly and indefinitely diverse. But, how, inthat case, can one identify a particular principleorjudgment orbelief as moral atherthan somethingelse ? I cited with approvalLadd's suggestion that one can look for those prescriptionsforconduct which have a special superiority and legitimacy in aculture and I also suggested looking for those regulative con-cepts which the members of a community apply to activitiesand relationsof central concern to them. But how can we rule


25/44

188 I-STEVEN LUKESout the possibility of a given culture's so applying non-moralconcepts, according superiority and legitimacy to non-moralprescriptions, or at most to an attenuated and degradedmorality? This seems, indeed, to be exactly the case of the Ik,the wretched starvingtribe describedin Turnbull's TheMoun-tainPeople:ashe saysof them, they 'havesuccessfullyabandoneduselessappendages, by which I refer to those "basic" qualitiessuch as family, cooperative sociality, belief, love, hope and soforth, for the very good reason that in their context these mili-tated againstsurvival' (p. 289). Does not the very act of identi-fying a set of principles,judgments, action, etc., as moral om-mit us to makingassumptionsabout the content of morality, itsr6le in organisingand regulating social life and its relation tohuman needs, wants, interests, purposes, virtues, excellences,defined somehow-but how?-independently of any particularmoral perspective? And if that is so, are we not thereby com-mitted to a non-contestable definition of morality, and therebyto setting limits-but how narrow?-to moral relativism by,at the very least, ruling out certain judgments and actions(such as those of the Ik) as candidatesfor morality?

I began this paper by referringto the temptations of rela-tivism. These can be overcome either by resistingthem in totoor by giving in to them with abandon. The situationsof the con-sistent Puritan and of the uninhibited voluptuary are at leastunambiguous. It is the partial resistance to temptation thatcauses anxiety and a lingering sense of dissatisfaction.

REFERENCESSee especiallySteven Lukes, 'Some ProblemsaboutRation-ality' and Martin Hollis, 'The Limits of Irrationality' and'Reason and Ritual', all reprinted in Bryan Wilson (ed.),Rationality, Blackwell, Oxford, 1970), Martin Hollis, 'Witch-craft and Winchcraft', Philosophyof Social Science,2, 1972,pp. 89-103 and Steven Lukes, 'On the Social Determinationof Truth' in Robin Horton and Ruth Finnegan (eds.), Modesof


26/44

RELATIVISM: COGNITIVE AND MORAL 189Thought:Essayson Thinkingn Western ndNon-Westernocieties(Faber, London, I973).

2 E. E. Evans-Pritchard,Witchcraft,Oracles ndMagic amongtheAzande (Clarendon, Oxford, 1937), P. 194-3 See Lukes, 'On the Social Determination of Truth', pp.240-I.4 See Ernest Gellner, 'Concepts and Society', reprinted inWilson (ed.), Rationality.5See the essays in Horton and Finnegan (eds.), ModesofThought,sp. that by Horton, and also his essay 'AfricanTradi-tional Thought and Western Science', reprinted in Rationality.6 See, e.g., Gunnar Myodal, An American ilemma:TheNegroProblemndModernDemocracy,vols. (Harper,New York, 1944).'W. B. Gallie, 'Essentially Contested Concepts', Proceedingsof theAristotelianociety, 6, 1955-6, p. 169.8 For a fuller development of this argument, see the presentauthor's Individualism(Blackwell, Oxford, 1973), chs. I I and 20.9 Charles Taylor, 'Neutrality in Political Science' in PeterLaslett and W. G. Runciman (eds.), Philosophy,Politics andSociety,Third Series (Blackwell,Oxford, 1967), PP. 56-7.10 The argument of this and the next paragraph is morefully set out in the present author's Power: a Radical View,(Macmillan, London, 1974).


27/44

RELATIVISM: COGNITIVE ANDMORALSteven Lukes and W. G. Runciman

II--W. G. RuncimanOn what groundscould it plausibly be arguedthat the validityof a propositionis a function of the circumstancesunder whichit is conceived ? Mr Lukes's answer is, in effect,"none"; and ifI dissent from him it is only because I think he makes a littlelight of one ancillary argument sometimes advanced on theother side. But I do not, on the other hand, share his uneaseabout the contrastbetweenthe natureof ethical and of scientificjudgements. I make no claim to know the answers to thequestionsabout ethical theory which he raises, and I willinglyagree that the merits of the notion of a social contract as thebasisfor a theory ofjustice are disputable.But I cannot see thatthis implies anythingone way or the other for his answer to theprior question with which this paper begins. Those who accepta cognitivist ethical theory are, self-evidently, anti-relativistsin this too; but those who reject cognitivism in ethics are notthereby required to embrace relativism in science. The moralto be drawn from what Mr Lukeshas to say in the concludingsection of his paper is, I suggest, a differentone. To elucidateit, I shall first run brieflyover the salient points in the relativistcase; I shall then outline what I think is the more difficultproblem which underlies much of what is apt to be discussedunder the heading "sociology of knowledge"; and I shall endby putting forward some tentative suggestions, supported byone or two brief examples, about the lines along which thisproblem can in practice best be overcome.

INo one will question the well-documented fact that what isbelieved to be true varies from both place to place and time totime. But from this it has, unfortunately, seemed to somephilosophers and/or sociologists to follow that the notion ofscientificprogressmust itself be suspect. Are we not free, they

'9'


28/44

192 II-W. G. RUNCIMANhave asked,in any field of enquiryto rejectthe tacit agreementamong the practitionersof research that their chosen standardsof validity transcendthe boundariesof professionand culture?This questionis best answeredby another,as Mr Lukesamongothers has already done elsewhere (in Horton and Finnegan,ed., Modesof Thought, . 236): on what groundsdoes it followfrom the fact that the beliefsof scientistsaresociallyand psycho-logically determinedthat no criterion of validity is itself cross-culturally valid? Once we talk of knowledge at all we pre-suppose the possibility in principle of the detection of error(cf. Plato, Euthydemus87a), and to infer relativismin mattersof knowledgefrom the undisputedfactsof culturaldiversitycanonly be self-defeating.What is more, the fact of diversity is ifanything less relevant to the logic of scientific discovery thanthe fact of convergence.Our chemistry, geology and mechanicsmay not be those of the Azande or the Ndembu. But it issurely a safe prediction that they (or their grandchildren)willcome to use our textbooks,not we theirs.The progressof sciencemay be tentative and haphazard,and its resultsalwayssuscep-tible to revision. But it still does not follow from suuscuiquemosthat achacuna verite.The aspect of cultural diversity on which the attention ofphilosopherscan moreprofitablybe concentrated s the relationbetween varying criteria of validity and varying contents ofbelief. People vary not merely in what constitutes, for them,adequate evidence to confirm what they believe, or claim toknow, but also in what they think it possible to have beliefsabout in the firstplace. (Cognitivisttheoriesof ethics furnishasgood exampleasany). The point may seemfamiliarenough,butit is too often forgotten in discussionsabout the language andritual of alien cultures. As I have argued elsewhere (ArchivesEuropeennese SociologieX, 2), the first question to which thesociologist and/or philosopher needs to address himself iswhether the speech or action which puzzles him is in fact ex-pressiveof a belief at all, even if at firstsight it seems to be so.Indeed, the advice holds equally within a common culture.Many Englishmen can be observed in churches saying (orsinging) thingsof the grammaticalform of empiricalstatementswhich they do not themselvesbelieve to have truth-valueat all,just as many members of so-called "primitive" societies are


29/44


well aware of using ostensiblyliteral language in a symbolic ormetaphoricalsense. The distinctionholds, moreover,whatevermay be the answer to Wittgenstein's question "Is belief anexperience?" and whatever may be the difficulties posed byborderline cases or by the relation of the concept of belief tothat of knowledge. It is indeed necessary to remember thatwhat can count as a matter for belief is both inter-culturallyand intra-culturallyvariable. But the conclusion to be drawnis not that even a suitably cautious doctrine of scientific pro-gress is untenable; it is merely that the practising sociologistmust beware of too easily doubting other people's claims thatthey hold as a belief something which he himself does not.It is not, however, so easy to rebut the argumentfor conceptualrelativism,and it is here that I think Mr Lukes is a little hasty.It is one thing to say that the idea of negation, and thereforeof falsehood, is a cultural universal. It is quite another to saythat the content of belief,whetherscientific or of any otherkind,is translatable without loss of meaning between one culture(or sub-culture) and the next. To concede that criteria ofvalidity are transcultural is not to concede that criteria ofidentity are. Indeed this is, I think, the lesson to be learnt fromthe writings of Durkheim and his followers, who for all theerrorsof both ethnographyand logic of which their commenta-tors have convicted them were right to point out that the wayin which the world is classifiedand labelled is differentlydeter-mined in different cultures at a very fundamental level. Seenfrom this angle, the argumentdeployed by Mr Lukesplaces atonce more emphasis than is needed on truth and less than isneeded on synonymy. To answer the Durkheimians on thispoint, it is not enough simply to say that the best way to deter-mine the meanings of sentences in a language is to ascertaintheir truth-conditions and truth-conditions are not context-dependent in the manner that Durkheim appears to havebelieved.But once again, the question to be askedfor the purposesofthis paper is not how the problems inherent in the notion ofsynonymyare to be resolved butwhat conclusionis to be drawnfrom the fact that they arise in the way they do. Just as somephilosophersand sociologistshave been misledby the anomaliesin the historyof western scienceinto questioningthe underlying


30/44


31/44


IIA pertinent clue to the nature of the problem is the fact thatthe classicalliterature,if it may be so called, in the sociologyofknowledge has largely been preoccupiedwith the epistemologyof the human as opposed to the natural sciences. This pre-occupation is echoed, and even amplified, by Mr Lukes in hissuggestion that "certain theory-laden identifications of socialfactsare inextricableconfusionsof descriptionand evaluation".On this view, if I understand him correctly, there is still nohint that the validity of a proposition is ever a function of itsorigin; but there is more than a hint that the terms in whicha social theory is couched reflect the culturalmilieu in which itoriginates in such a manner as to render acceptance or rejec-tion of the theory as stated conditional on acceptance or rejec-tion of the political and moral standards dominant in thatmilieu. This view I believe to be mistaken.But the mistakeis asubtler one than the mistakesof the vulgar relativists who arecriticized in the earlier section of Mr Lukes'spaper.The mistake restson an implicit confusion between the opentexture of theoretical terms and the contestable nature ofevaluative terms. Mr Lukesis rightly wary of taking too simplea view of the distinction between fact and value. But he is stilltoo much its prisoner,for he seems to assume that if statementsabout such things as power or interests are somehow notstraightforwardlyfactual they must be implicitly evaluative.But why? Some further argument must be adduced to showthat the difficultiesconfronting the political scientist when heuses the concept of power are different in kind from those con-fronting the physical scientist when he uses the concept ofgravity. No doubt the first are, at least for the present, moreintractable. But that is because we do not have even a rudi-mentary scientific theory about power, whereas we do have afully developed scientific theory about gravity. We all, admit-tedly, have views about how power ought or ought not to beexercised of a kind which we do not have about the workingsofgravity. But this does not mean that a well-tested and wide-ranging theory of power, when or if we ever have one, will belogically tied to our judgements about the morality of its usefor one or another set of ends. On the contrary: one of its


32/44

196 II-W. G. RUNCIMANcharacteristics will be that the confirmed hypotheses derivedfrom it will be as compelling to one school of political moralistsas to another.There still is, I agree, a difficultyabout 'power' which thereis not about 'gravity'. But it is of a kind which has nothinglogically to do with questionsof ethics. It is well illustratedin,for example, Alasdair MacIntyre's paper "Is a Science ofComparative Politics Possible?" (in his AgainsttheSelf-Imagesof theAge, ch. 22), where he cites a sample survey which wasclaimed by its authors to show that Italians take less pride intheir national political culture than Germans or Englishmen,but which made no attempt to take account of the differenceinthe meaningof the conceptof 'pride'between the threecultures.It is a truism,as MacIntyre himselfobserves,that "institutionsand practices are always partially, even if to differing degrees,constitutedby what certainpeople think and feel about them".The problem, however, is to say just what follows from thistruism both for the practice of the human as opposed to thenatural sciences and for the philosophical questions to whichMr Lukes and I have been asked to addressourselves;and ittakes what I believe to be its most illuminating form in thequestionswhen,and on what grounds,the observer'sconceptualcategoriestake priorityover those of his subjects.These questionsare sometimes so put as to be answeredbyreference to the difficultiesof translation between the subjects'and observer'srespective languages: thus in the example justcited, it ought to have been specified by the authors concernedwhat they take to be the difference between "orgoglioso"nd"proud", and the termsrelated to them in Italian and Englishrespectively.But although this is indeed a legitimate require-ment to impose, it would be misleading to imply that theapparentdiscretionwhich the sociologisthas in adoptingone oranother set of terms for the construction of a would-be theoryof some aspect of human behaviour derives simply from dif-ferences of language. It is true that special difficulties attachto the task of the sociologistwho requires, in Quine's phrase,"an entering wedge into a strange lexicon". But all languagesare both learnable and teachable; were they not, they wouldn'tbe thereto perplexthe sociologist n the firstplace. Once, there-fore, the language has been learned, the sociologist is no


33/44


34/44

198 II-W. G. RUNCIMANbetween the conceptual schemata adopted or retained by thesubjects studied and their structural, cultural and historicalcontext.Let me cite an example already familiar from extensivediscussionby sociologistsand philosophersalike: the theologyof the Nuer as described by Evans-Pritchard.It is a particu-larly difficult example, since it raises the problemsnot only ofselecting the most accuraterenderinginto English (and Evans-Pritchard is on record as saying that nobody ignorant of theNuer language should venture to express an opinion on themeaning of Nuer statements) but of ascertainingthe point atwhich beliefs with specificable propositionalcontent shade into"belief in": indeed, the two compoundeach otherin the centralquestion whether "gainliakaiKuoth" is strictly equivalent to"I believe in God". But these problems, as I have already re-marked,do not affect the argumentI am seekingto put forwardhere. Assume that Evans-Pritchard'sdescriptionis acceptableto a bilingual Nuer and that he is able satisfactorilyto discri-minate between statements about the attributespredicated ofthe Nuer God and statements about the attitudes consideredappropriateon the part of his worshippers.How then is he toanswer the questionswhich this raises in their turn about thenature and significance (to them)of the aniconism which is(to him)the most strikingfeature of their religion?In very broad terms, I think the answer is: by dissolvingthisquestion into two separate, albeit complementary, questions.The firstof these is the Durkheimian,by which I mean that thesociologist will ask himself how the aniconism of the Nuerrelates to their form of social organization as well as to theirritual practices and (if he can discover it) to the history oftheir now established terminology relating to a withdrawnHigh God. The second question, on the other hand, which hewill ask himself is the comparative question: how does theaniconismof the Nuer relate to that of other peoples, whatevertheir culture, structure or ecology, whose religion can alsoplausibly be classed as aniconistic? This will, no doubt, raiseagain the problem of translation; but again it is a problem,which, however intractable, is an incidental one. Readerswhofind more illuminatinga paralleldrawn to a culturewith whichthey are already to some degree familiarmay preferto explore


35/44

RELATIVISM: COGNITIVE AND MORAL 199such suggestionsas that "God is perhaps no more withdrawnfrom the outlook of some Africanpeoples who have beencreditedwitha DeusOtiosus han he is fromthe consciousnessof the average MediterraneanCatholic" (M. Singleton inSinger and Street, ed., ZandeThemes,p. I45). But whatevercomparisonis chosen, its purposewill be to elucidate in a waythat reportagealone can never do those featuresof the religionof the Nuer which are distinctivelycharacteristicof it; and thecontrastsneeding to be drawn will be in such termsas the Nuer'sgroundsfor dispensingwith liturgy, their relative emphasisonthe greatnessof God as opposed to his indifference to humanconcerns,and the extent of their fear of interventionby God inthe natural order as distinct from the guilt which may be in-duced in them by reflection on God's supernatural purity.Only when the relevant questions of both kinds have beenanswered can the sociologist claim to be able properly to saywhat sortof areligiontheNuerreligion s;and itisthisunavoidablequestion which bringsboth him and his readers ace to face withthe problemof conceptual, as opposedto scientific, relativism.Before I passto the concludingsection of the paper, it may beworth my anticipating two possibleobjectionswhich my discus-sion so farmaybe thoughtto invite. The firstis thatby suggestingthat the concepts of one culture can be better understood byinvokingcontrastswith those of another,Iam alreadyoverridingthe priority which ought to attach to the subjects' ownaccount of what their beliefsand attitudesmean to them. Is notthis, it might be asked, the very approach which has so oftenled to ethnocentrismand misunderstanding? To this the reply isthat these dangers are indeed always present; but it would beas implausible to suggest that such contrasts are inherentlymisleading as to suggest that alien languages are in principleunlearnable. No doubt there is a sense in which the observer'scategoriesmust be strictlyirrelevant: the Englishhistorian whodraws a parallel between the doctrine of the Elizabethan magithat the discovery of the "form-number" which signifies aparticular thing gives power over that thing and his ownrecognition that the formulae of modern science do indeedgive power over nature is saying something which no Eliza-bethan maguscould have understood.But what historian wouldever suppose otherwise? All he claims is that the parallel is in


36/44

200 II-W, G. RUNCIMANsome sense illuminating, whether in helping the reader to seemore clearly the plausibility of the now outdated belief or inshowinghim morepreciselywhat it implied. An argumentthatno such parallel can possiblybe illuminating at all would haveto be sustainedon some separateground of its own.The secondobjectionis that what I am sayingis not so muchmistaken as unnecessary, and unnecessary not because theproblem which I have posed is an unreal one but it is becauseit has alreadybeen worked out in the writingsof Dilthey, Croce,Husserl, Collingwood, Schutz and their successors.To this thereply is simply that despite the obvious overlapin interest,andto some degreein approach,I have not found the problem any-where stated by either Phenomenologistsor so-called "Ethno-methodologists" n the formwhich seemsto me required.This,I suspect, is because the writers most concerned to insist thatthe experiencesof otherpeople shouldbe recounted n theirownrather than the observer'sterms have been seekingto refute arival doctrine (for which Behaviourismis probably the mostusefullabel) about the practiceof the humansciencesgenerallyrather than to develop a rigorous account of the criteria bywhich the sociologistis to assess the descriptivecontent of theevidence presented to him. However necessary it may be tocorrect the excesses of Behaviourismby the reminderthat theostensible actions of self-conscioushuman agents can all tooeasily be misunderstood,such correctionstill leaves unclear thegrounds on which the subject'saccount of his actions can, ormust, be said to overrule the observer's. Until these groundscan be set out not only generally but with some fair degree ofprecision-which as far as I am aware they have not-thereremains the risk that even the best-foundedand most salutaryof the Phenomenologists'criticismswill tend towards a sub-jectivism as extreme and as unfruitful as the exaggeratedBehaviourismagainst which they were rightly directed.

IIIAny attempt to advance the solution of the problem of con-ceptual relativism necessarily presupposes answers to severalrelated questions which are themselvescontroversial, includ-ing questionsabout the nature of meaning. As to this, it will be


37/44


enough for my purpose to assume no more than that, as it isput by Michael Dummett (Frege,p. 85), "Meaning, under anytheory whatever, cannot be in principlesubjective becausemeaningis a matter of what is conveyedy language"(Dummett'sitalics). But there are two further,and perhapsmore question-able, assumptionswhich I need also to make. The first concernsthe distinction between explanation and description, and thesecond that between intention and motive.If it follows, as it seems to me it must, from the rejection ofrelativism in science that every action could, in principle, bebetter explained by some qualified observer than by the agent,then it follows also that the area of discretion within which theagent's account has priority is only that of description as dis-tinct from (although of course related to) explanation. I hereuse "description"in the sense which I have tried to elucidateelsewhere (see Mind, LXXXI, No. 323) whereby it is markedoff at once fromreportageon the one hand and from evaluationon the other. The issues which it raises have, in my view,received less attention from philosophers than they deserve.Indeed, I would go so far as to predict that in due course theywill generate as voluminous and as interesting a literature inthe philosophyof the social sciencesas that which now existsonthe r6le of laws in sociological explanation. But however thatmay be, the point which I propose to assumewithout furtherargument is that an agent's account of what he is doing is inprinciple privileged as against that of any observer whereashis account of whyhe is doing it is not. Some philosopherswill,I realize, wish to dispute this on the grounds that the agentknows his reasons,but the observerknows, at best, only whatthe agent tells him of his reasons. But from the fact that anagent has knowledge of what he thinks o be his reasons of akind which the observerself-evidentlycannot have it does notfollow that the agent cannot be mistaken about them. Agentsoften are. In disputed cases, the evidence may well be lackingwhich would conclusively vindicate the counterfactual con-ditional which the observerwishes to maintain even in the faceof the agent's denial. But the observer will always be able tospecify a possiblealternativeexplanation, however far-fetched,which if vindicated wouldrequire the agent to concede to him.What he cannot do is in the same way to suggest to the agent


38/44

202 II-W. G. RUNCIMANthat he is not doingwhat he says (without oking or dissembling)that he is doing, since what the agent is doing is constitutedbywhat he thinks he is doing in making the sounds and gestureswhich the observerhears and sees. The only sense in which anagent can be mistaken about what he is doing is where he ismistaken about one or more of the relevant ancillary facts, aswhen, say, he believes that he is signing a cheque but is writingon the wrong bit of paper or believes that he is saying gracebut is rememberingthe wrong bit of Latin.This leads directlyto my assumptionabout intentions,whichI take to be constitutive of actions whereas motives are not.Imputations of motives, accordingly, have explanatory force;indeed, where "reasons" are invoked to explain behaviourthey very often, although not always, turn out to be motives.Whether they are than to be called causesdependson the sensegiven to "cause".I find no particular difficulty in so callingthem, provided that "cause" is not too narrowly or mechanisti-cally defined, but I do not think the argument of this paperdepends on it. The more importantpoint is that the practisingsociologist is well advised to follow the rule that the actionswhich he seeksto explain shouldfirst be definedby reference tothe intentionwhich constitutesthemand only thereafterrelatedto the motive(s) which may or may not have prompted them.Thus when presenting the finished account to his readers, heshould first make clear what he holds his subjectsto have beendoing in (as opposed to by) making their speechesand gesturesand then set out that selected aspect of their antecedent bio-graphieswhich, in his view, was decisive in causing them to dothis rather than something else. Typically, he will show thatit was an act of (purchase,worship, murderetc.) prompted by(avarice, fear, vindictivenessetc.) as a result of specifiedeventsand influences in the agents'past which implanted the motiveswhich any one of a predictablerange of stimuli would trigger.Actions can, of course,also be performedwitha (further)inten-tion which may in part explain them. But in these contexts thenotion of intention elides into the notion of plan, and plans arenot, I takeit, constitutiveof actionsany more than the programof a purposive mechanism is constitutive of its behaviour inhoming to its goal. It is the intention which constitutes anaction which the observer must make sure he does not


39/44


misdescribe,even where he is confident of having explainedit correctly.The sociologist, therefore, who is called on to defend hisaccount of what his subjectsare doing is requiredto show thathis account, however differently phrased, is still compatiblewith theirs and capable in principle of being acknowledgedtobe compatible by them. It is not a matter of his disputing withthem about the correctnessof the facts-or at least, not in theordinary sense; it is rather a matter of his establishing theappropriatenessof the terms in which his account of the factsis given. The sociologist who claims, as Evans-Pritchardoncedid (Witchcraft,Oracles ndMagic amongheAzande,p. I I), that"the facts will be the same without their labels" is right in onerespect but wrong in another. He is right if he is saying thatboth his observations and his explanations of them can bereported under one label as well as another without theaccuracyof the observations or the validity of the explanationsbeing necessarilyaffected thereby. He is wrong if he is sayingthat a change of label makes no difference to the meaningattributed to the behaviour which he has observed and nowseeks to explain. From this standpoint, the appositenessof theterms which he employs is and must be relative to the culturewithin which that behaviour takes place.I do not know whether the kind of example of most interestto sociologistswill be of most interest to philosophersalso. Butone of the topics on which the sociologist's discretion in hischoice of terms is notoriously problematical is the topic ofdeviance-a term which is itself often a subject of controversy.What are the right terms in which to characterize behaviourone of whose essential features is that it transgressesone or moreestablished institutional and/or societal norms? Part of thedifficulty lies in the need for the sociologist to disclaim anyjudgement of his own about the moral legitimacy of thesenorms. But this danger, once recognized, can be surmounted.Although Mr Lukes will say that whatever description thesociologist chooses to give rules some, but not other, moralevaluationsout of court, I cannot see that this is so. If the term"deviance"is deliberatelysubstitutedby him for the term"sin",this betraysan evident concern on his part to evade the possibleimputation to him of a willingness to accept the institutional


40/44

204 II-W. G. RUNCIMAN

standardof morality as his own. But he is perfectlyfree to usethe institutional term without sharing the institutional evalua-tion: nothing prevents him from approving of what "they"call "sin" and disapproving of what "they" call "virtue".What is much more difficultis to find the termswhich will notcommit his subjectso an account of their actions which woulddistort theirprobably inarticulate and possibly even incoherentdisagreementwith the institutional norm they are consciouslyviolating; and for thispurpose,"deviance"may be no improve-ment on "sin", but rather the contrary.

It is not even enough to say merely that the sociologist'sterms must be acceptable to those whose behaviour he usesthem to describe, for this raises in turn the question of thegroundsfor such acceptance: the native informant is not neces-sarily right about such things any more than the sociologist isnecessarily wrong. In addition, the sociologistmust set out hisanswersto both the two types of questionswhich I listed in theprevioussection. In the case of "deviance", he must firstof allbe able to document the scope and applicationsof the sanctionswhich are enforced to sustain the institutional norms, theoccasion and frequencyof typical violationsof these normsandthe use, according to context, of the terms in which violationsare reported both by those with the authority to impose therelevant sanctions and by those upon whom the sanctions areimposed. Second, however, he must be able to place what hesees as the intention constitutive of the act of violation in thebroader context of human intentions and social institutionsgenerally; and it is here, not in the matter of moral and poli-tical judgements of value, that his account will, in Mr Lukes'sphrase,necessarilyrule someotherpossiblealternative accountsout of court.I do not deny that in any sociologist'saccount of "deviance"in the culture he has chosen to study explanation, descriptionand evaluationare likely to be interwoven,often so closelyas tobe difficult to disentangle. But I do maintain that such dis-entanglementis possiblein principle, even if, in someinstances,it cannot be done at the time. Consider,to take a differentsortof example, the social historianof Victorian Britain seeking todescribe the significance of the doctrines of classical politicaleconomy to the statesmenand administratorsof his period. He


41/44


has the familiarproblem of distancinghimselfsufficientlyfromthe moral assumptions underlying the formulation of thosedoctrines not to allow, as the political economists may them-selves have allowed,judgements of fact to be confused with, ordistorted by, judgements of value. But he has also both toassess the validity of those doctrines viewed as hypotheses insocial science and to understand how those who accepted themat the time viewed them. For this latter purpose, I think therecan be no doubt that anachronisticparallels are helpful. Thetwentieth-century sociologist aware, as

5 Relativism Cognitive and Moral

Documents

Transcript of 5 Relativism Cognitive and Moral