SOCIALEPISTEMOLOGY , 2002, VOL. 16, NO . 3, 267–281

Dilemmas of objectivity

MARIANNEJANACK

Objectivityis a virtuein most circles.As weevaluateour students (even those we don’t like)we tryto be objective. As wedeliberateon juries,we tryto be objective about the accused and thevictim, the prosecutor and thedefence. As wetoteup theevidence for and againsta certainposition or theory, we try to separate our personal preferences fromthe argument, or we try to separate the arguer (and ourevaluation of him/ her) fromthe case presented.Sometimes we do wellat this,sometimes we do lesswell at it,but we recognize something valuable intheeffort. Considerthe following cases inwhich objectivity or its failure is at issue:

(1)A professorgives his favourite student an Ainaclass inwhichthe student has done onlymediocre work. (2)A scientistoverlooks evidence that would call intoquestion a theorythat she has goneout on alimbto defend,and has investedmuch timeand energyin pursuing. (3)A scientistskews data inorder to bolster support for a favouritepolitical cause. (4)A particularphilosophy journal, which does not use a blindreview process, only publishesarticles written by men. (5)A memberof a trialjury questions evidence presented in atrialbecause it conicts withracist stereotypes.

Some ofthese invocations of objectivity imply that a separationshould exist between thesource or origin of the theory or argument and theargument itself. Some imply thatour deliberations should be answerable only to ‘ theevidence’ or ‘ thefacts’ and notsimply to our own preferences and idiosyncrasies.Some implythat our emotionsor biases (broadly understood to include emotions) should not interfere withour reasoning. Inits ontological guise, the appeal toobjectivity is a wayof appealing to the way thingsare, or to the world as itis independently of our desires about howwe want it tobe. So, while it might seem reasonable to think of objectivity as an epistemic virtue,the term also carriesmetaphysical overtones. It would seem to be an uncontroversialepistemic virtue premised on a fairlycommon-sense realism: there is awaythe world is, and theway the world is isin principle distinguishable from the way we wish it were.

Author : MarianneJanack is Assistant Professorof Philosophy atHamiltonCollege in Clinton, NY,13323, USA.She teachesclasses in epistemology,philosophy of scienceand feminist philosophy .She ispresently workingon alongerproject on rationalityandfeminist theories of emotion.

Social Epistemology ISSN0269-1728 print/ISSN1464-5297 online # 2002 Taylor &FrancisLtd http://www.tandf.co.uk/journals DOI:10.1080/0269172022000025624 268 MARIANNEJANACK

Theconnection between objectivity and truthhas beenan importanttool for feministand otherlibratory projects, but failures of objectivity are not always or onlyepistemic failures. The claim thatthere is still sexism in the world can only bedenied by someone who fails to be objective. This is a failurethat has two differentand separableaspects to it. It is an epistemicfailure, in so faras it seemsto involve a wilfulavoidance of evidence that is all tooclear (that women ofall racesstill suffer the greater burden of social ills like poverty, spousal abuse, discriminatorypay practicesand discriminatoryhiring practices, for example, and enjoyfewer of the social goods likefreedom of self-determination, respect, etc. thando men).It is also an instanceof a theoryor claim thatfails to correspond tothe facts. That is, this case involvesboth bad reasoningand falsehood.Would soundreasoning (we’ ll call thisobjectivity r)have guaranteedcorrespondence to reality (call this objectivity c)?Notnecessarily, buthad our claimant beenobjective r she mighthave beenmore likely tobe objective c.Yet, as we all know,objectivity c doesnot require objectivity r—peoplecan getthe facts rightaccidentally (i.e.by guessing) orthrough faulty reasoning. Objectivity c is notsecured byobjectivity r, and objectivity r is neither necessary norsufŽ cient forobjectivity c. Comparethis situation, however, to the situation described in (4)above. Of course,it ispossiblethat no women write articles that are as good as thosewritten by men, and so thelack ofpublishedarticles by women is dueto the inferior quality of articleswritten bywomen, but thefact that the review process is notblind leads oneto suspect that the qualityof the manuscript is notthe only factor being used in the evaluations. We think thatit seems unlikely that all themanuscripts submitted by women would be unacceptableby the journal’ s standards,unless, of course, the author’ s sexenters into theevaluation either implicitly or explicitly. Let’ s say thatthe author’ s sexis explicitlyconsidered in the evaluation process, that in fact the journal is dedicated to publishingarticles written only by men. Would thisbe a failureof objectivity? Not necessarily;it might be fairly explicit sexism, but not necessarily a failureof objectivity,since the reviewers do notclaim tobe impartial; in this instance they simplydo notwant to publish articles by women. Were we to question their motivationfor this stance, we might eventually encounter a failureof objectivity— thatis, the justiŽ cation for this practice might ultimately lead us back toa failureof objectivityon the part of the publishers. The beliefs upon which the practice is based mightbe subject to the same criticismsdetailed in the case ofthe sexism-denier describedabove. But thesimple fact that a journaldoes not publish articles by womendoes not lead to theconclusion that the journal is notobjective. Yet,if the reviewers claim notto be using sex as acriterionof evaluation, yet are implicitlydoing so, then I thinkwe have something more like a failureof objectivity. If,that is, the reviewers claim thatthey will publish all and onlythose articles they judgeto be well-argued, for instance, and measureup tocertain standards of scholarship(e.g. carefully researched, well-written, etc.) then the implicit operation of sexas awayof evaluating quality seems to be a signiŽcant failure of objectivity. But isitthe same kindof failure as thatdescribed in the case of the sexism-denier? There seemsto be a certain‘ familyresemblance’ but noshared characteristics here. Case 4 seemsto be a case inwhich a certaintenet of informal reasoning has beenviolated: namelythat an argumentis valid independently of who offers it. While this seems to bea canonof good reasoninggenerally, it is different from the canon of good reasoningthat would have us lookat all theevidence available for a certainclaim. Objectivityin this sense is something more like impartiality with respect to the source DILEMMASOFOBJECTIVITY 269 ofan argument or position, and itis atenetwe learn(or teach) in informal logic classes. Yet,the demand thatwe look for all theevidence for or against a givenposition is a much moregeneral and unreective epistemic practice— one in which, in fact, the reliabilityor credibility of a claimantmight enter as an importantpiece of evidence. 1 Ina similar,yet slightly different case, if one of our colleagues were to give the studentshe liked grades thatfailed to re ect the quality of their actual work,we wouldthink that his objectivity as ajudgewas compromised. Under these circumstances,we would want to say thathis personal attachments had interfered withhis professional judgment, and whilehe might not be epistemically culpable, wewould probably thinkthat he was morally culpable. Take, for instance, the case inwhich the professor in question gives in ated grades towork he knows to beinferior. This seems to be a moral failing,rather than an epistemicfailing. Doesthe analysis change if we tweak the story a bit?Consider the more likely scenarioin which the professor’ s evaluationof the student’ s workis in uenced by hispersonal attachments. Does this constitute a differentkind of failure of objectivity?It does seem to be different from the scenario in which the professor knowinglygives his favourite students better grades thanthey deserve. But inthe case inwhich the professor Ž ltershis evaluations through his personal attachments,the areas of moral and epistemicculpability become murkier. We mightbe tempted to say thatin this case thefailure is more epistemic than moral,yet it seems that in ascribing epistemic culpability we would need to assume thatthe professor could reasonablyhave disentangled his personal preferencesfrom his belief-forming processes. And it’ s notclear that our beliefs and theprocesses by which we go about formingthem can beso easilyseparated intothe ‘ true’and the‘ subjectivelyenhanced’ . Considera variationon this scenario: instead of giving his favourite students better grades thantheir work warrants, he perversely gives them worse grades. He might do thisbecause, he reasons, he knows these students so well,he likes them so much and heknows that receiving a bad grade ona paperwill compel themto work much harderthan they would have if they’ d gottena mediocregrade. He might thinkthat they will respond to a lowergrade byre-doubling their efforts, and will vastlyimprove their papers as aresult.He might also judgetheir work more harshlybecause he knows what they are capable ofat theirbest— that is, his beliefsabout thequality of his favourite students’ work would, in this case, also beŽ lteredthrough his personal attachments. The point I wishto make hereis thatit is not just that the evaluation is Ž lteredthrough personal attachments that makes ita failureof objectivity, but rather that the consequences of such attachmentsis a certainkind of evaluation. The case inwhich the professor gives hisfavourite students better grades thanthey seem to merit is morally discriminablefrom the case inwhich the professor gives his favourite students worsegrades thanthey deserve, but theydo notseem to be epistemically discriminable—in each case the beliefs the professor has areŽ lteredthrough his personalattachments. If we think that the real epistemic crime is allowing personalpreference to in uence the ways in which we go about formingbeliefs about studentwork, then these cases arenot at all differentin epistemic terms. But theydo seemto be different morally. I thinkthe ways in which these cases ‘partways’ morally but not epistemically shows that it is not simply that one’ s beliefsare Ž lteredthrough one’ s attachments,but thatsuch attachments might leadto a certainkind of favouritism. 270 MARIANNEJANACK

Casesof evaluation seem to be different in important ways from cases inwhich mattersof fact are at stake.Evaluations do notseem to involve a simple correspondenceto the facts of the matter, and yetwe do thinkthat such evaluations admit ofmore or less objectivity. But theobjectivity of evaluations does not seem to dependso much ongetting it right, or taking the perspective of the ‘ ideal’ disinterestedobserver. Rather, judgments about therelative objectivity of evaluations seemto depend on a notionof appropriateness. In this respect, the case ofthe professorwho gives his favourite students better grades thanthey deserve overlaps withthe case ofthe journal reviewers. The failure of objectivity in these cases isdue tothe use of inappropriate and ‘secret’criteria of evaluation, which enter the process ofevaluation as hiddenfactors. Thescenario in which the professor gives his favourite students worse grades than theydeserve to motivate them seems to be a differentmatter— unlike the case in whichhe gives his favourite students better grades thanthey deserve, which is afairly uncontroversialexample of a failureof objectivity, the case inwhich he gives them worsegrades ismore puzzling. Should this,too, be characterized as afailureof objectivity?While both cases involvea situationin which the professor’ s beliefsabout thequality of the work is Žlteredthrough his personal attachments, the case inwhich hegives them worse grades thanthey deserve is less clearly characterized as afailure ofobjectivity. The difference might be in the fact that we expect emotions and emotionalattachments to track in some predictableway: classic failures of objectivity arethose in which we treat those we dislike badly (a professorwho gives students he dislikesworse grades thanthey deserve, for instance), and wetreat those we like betterthan we probably oughtto. We expect failures of objectivity to follow this kind ofpredictable pattern. Theother difference might be this: the scenario in which the professor gives students worsegrades thanthey deserve in orderto motivate them seems to bea case inwhichhe appears tobeusinggrades notreally as evaluations,but as motivators.Evaluations can bemore or lessobjective but strategic choices about howto motivate students seem to bea differentmatter. Undeniably, such strategic choices are related to ‘ mattersof fact’— itis either true of false that giving students lower grades thanthey deserve will motivatethem to work harder and do betterwork. But theprofessor who gets it wrong—who chooses the wrong strategy— does not fail to be objective; he just turns outto bewrong. However,again, notice that how (or why) one ends up gettingit wrong makes agreat dealof difference in some cases. In fact, in such cases howone gets to a conclusionis moreimportant than whether or not one’ s conclusionturns out to be true. A jury memberwho failed to be convinced by the evidence in a case,but ratherinsisted on thetruth of racist stereotypes (case 5) seems to be both epistemically irresponsible and immoral.Racist stereotypesare not ‘ private’— theyare all toopublic— but in a casein which one is asked to determine guilt or innocence, we tend to think that suchstereotypes should not be taken into account. Not onlythat, we think that anyonewho actually believesthem to be appropriate bases forsuch determinations is immoral.Indeed, even if such a jurymember were right in her determinations of guiltor innocence, we would still think her both epistemically irresponsible and immoral. Ascientistwho interpreted all theevidence in a waythat supported her preferred theorywhile ignoring disconŽ rming evidence would clearly be epistemically irresponsiblebut mightor might not be morally irresponsible. Would ourevaluation DILEMMASOFOBJECTIVITY 271 ofher epistemicvirtue change if hertheory turned out to beright? I thinkit is altogether possiblethat, were she to turn out to be right, we would think her less epistemically irresponsible,although we might not revise our judgment about herlack of objectivity.That is, were she to get it right, even though she had ignored disconŽrming evidence, we might think that while she still failed to be objective, neverthelessthe truth of her conclusion mitigates her epistemic irresponsibility. In sucha case,objectivity cannot be simply identiŽ ed with epistemic virtue, since judgmentsof epistemic virtue are sometimes made inlight of the truth of one’ s conclusion.We can imaginea storyabout thescientist who just knows she is right, eventhough her colleagues disagree with her and theevidence seems to be piling up thather intuition in this case is wrong— and yetshe perseveres, driven by her commitmentto an idea.If the idea turns out to be true in spite of the countervailing evidence,she is an epistemichero; if it turns out to be false, she is an epistemicgoat. Inthe case inwhich she is a hero,objectivity is irrelevant and, in fact, would have beena detriment;in the case inwhich she is a goat,we attribute her failure to a lack ofobjectivity or an overweening ambition, or both. However,note the Ž nedistinctions between this case (case2) and case 3,where a scientistintentionally skews data tosupport a favouredpolitical cause. Intentionally skewingdata seemsto be a moreegregious failing than the failure described in case 2.Giving the available evidence the best possible spin or ignoring disconŽ rming evidence(as thescientist in 2 does)seems to admit ofan interpretationin which the epistemicagent is just sloppy, perhaps, but notvicious, whereas the case in which someoneknowingly deceives others requires a wilfulact ofepistemic vice. Case 3 wouldsurely provoke greater censure than would case 2,and itis not clear that we wouldrevise our judgment even if the scientist in case 3wereto turn out to be right—either about thelegitimacy of the political cause orthe truth of the theory. Therefore,in thiscase (unlikecase 2),truth would not rehabilitate her epistemic status. Whilethe cases Isketchedin thumbnailform at thebeginning of the paper seemed to beclear-cut cases ofa failureof objectivity, the devil, as theysay, is in thedetails. Most ofus do notencounter situations in thepat formin whichI presentedthem in my list; most ofour grapplings withobjectivity in everyday life are much messier.Decisions about howto be objective, whether and whenwe have failed to be objective, whetherand whenothers have failed to be objective (and whetherit matters if we or theyhave not been objective) are often fraught. So whileit seems as if,on the face of it,everyone knows what it means to be objective and howto go about beingso, at thesame timewe often Ž ndit difŽ cult to determine what counts as objectivityand whatcounts as afailureof objectivity. What seems to be unquestioned is thevalue of the ideal. Yet,in spite of both the ubiquitous nature of dilemmas ofobjectivity in ordinary experienceand thecentrality of appeals toobjectivity in different libratory projects, feministshave nevertheless engaged inan animatedcritique of that ideal. The ideal ofobjectivity is variously accused of:codifying a certainkind of masculine psychologicalapproach tothe world (Bordo 1987, 1999); not really safeguarding the truthand integrityof scientiŽ c inquiry(Harding 1993); of resting on an old and unphilosophicalemotion/ reasondualism (Lloyd1994); and ofbeing a mischaracterizationof the ways in which good scienceis practiced (Keller 1983, 1985).As a result,much ofthe feminist project has takenthe form of re-deŽ ning or re-envisioningobjectivity. Despite their differences, Longino (1990), Harding (1993), Nelson(1990), and Heldkeand Hellert(1995) have all worked,more or less, within 272 MARIANNEJANACK theterms set by this problematic. They have variously tried to establish procedures that wouldenhance objectivity, strengthen it, or actualize the epistemic ideals it embodies butfails to meet. Thefeminist theoretical engagement with the concept of objectivity in both epistemologyand philosophyof science has also made feministwork a targetfor criticism.Nussbaum argues thatthe feminist attempt to critique ‘ objectivity’is wrong- headedand thatit in fact undermines feminist efforts. Nussbaum pointsout that appeals toobjectivity and reasonhave been essential to feminist challenges to sexist hiringpractices, sexist evaluations of female faculty and unreasonableprejudice againstwomen’ s abilities(Nussbaum 1994,pp. 59– 60). Haack seesthe recent explosionof work in feminist epistemology as intellectuallyquestionable, arguing that feministpolitical philosophy seems appropriate, but that‘ feminist’epistemology and ‘feminist’philosophy of science are examples of a kindof ‘ imperialist’version of feminismthat has wanderedfar from its rightful place as amovementbased onan emphasison the ways in which women and menare alike, rather than different, and thatemphasized the value of equal opportunity(Haack 1998,p. 124). Thus, she claims thatcriticisms of objectivity that invoke masculine cognitive style and experience,or which appeal toaccounts ofidentity formation that emphasize gender differentiationundermine the egalitarian grounds that seem to be essential to counteringclaims thatwomen are less responsible epistemic agents than are men. Itis temptingto see the differences between feminist critics of objectivity and feminist defendersof objectivityas differencesbetween the Marxist and post-modernbranches of feminism(who critique the ideal) and itsliberal adherents (who defend the ideal). This is certainlypart of the issue. But thedeeper issue is related to the very concept of objectivityitself— or so Ishallshow in this paper. Section 1 lays outthe terrain with respectto objectivity and thefeminist engagement with the concept. Here I tryto showwhy ‘ objectivity’as aconceptand an idealis such a centralconcern for feminists,and whyit is central to the debate about thestatus of the feminist philosophicalproject. In section 2, I showthat the concept of objectivity is, in fact, a hodgepodgeof a varietyof different ideals and thatsome ofthe attempts to redeŽ ne objectivityhave obscured that point. Section 3 buildson section 2 toshow how this insightinto the concept of objectivity can beused to de-fuse particular challenges posedto feminist epistemology and feministphilosophy of science in particular, and feministphilosophy more generally. In this section, I also showhow the debate about thevalue of the ideal of objectivity is not really about thedifferences between ‘liberal’and ‘postmodern’feminists, but is in fact a relicof the hodgepodge nature of theconcept ‘ objectivity’.

1. Feminists and objectivity:can’ t live with it,can’ t live withoutit?

‘Objectivity’, understoodas acertainkind of metaphysico-epistemic ideal, seems to presentspecial challenges to feminists in epistemology and philosophyof science. As Antonyremarks, feminists seem to be caught inthe ‘ paradox ofbias’ : exposingbad biasin some places whilearguing, on the other hand, that bias is inevitable and somebiases are better than other biases. But withoutthe regulative ideal of an unbiasedperspective, Antony (1995) remarks, the grounds for such claims seemto be undercut:which biases are acceptable and whichare not? And what kind of authorityattaches to feminist perspectives? DILEMMASOFOBJECTIVITY 273

Morethan that, however, ‘ objectivity’in this sense seems to be tied in a sortof Gordianknot with philosophy. It is often thought to be synonymous with ‘ reason’or ‘rationality’in both moral and scientiŽc discourse.The presumed universality of the ‘viewfrom nowhere’ is at theheart of philosophical and scientiŽc authority,re- enactedand re-enforcedin the ways in which philosophical and scientiŽc articlesare written.As a result,critiques of objectivity as an idealseem to be synonymous with attackson rationality and philosophicalthought itself. Giventhis, the ideal of objectivity seems to be on a collisioncourse with the feminist project,if we understand by that a projectwhich takes gender/ sex/class/raceto benot (ornot simply) irrelevant and contingentfacts about personsand theirepistemic situations,but insome importantways constitutive and inescapable—to be is to be a gendered/raced/classed epistemicagent. From the perspective of those who take objectivityto be synonymous with philosophical thought, the term ‘ feminist philosophy’seems to be oxymoronic, synonymous with something like ‘ perspectival aperspectivalism’or the view of some peoplein particular. Part ofthe feminist claim, ofcourse, is that the ideal of objectivity as ithas beenlionized in philosophy and sciencehas alwaysbeen something oxymoronic in this way— itis just that people whoseviews have passed as theview of no-one-in-particular have not been aware thatthere might be other views that had beenexcluded. So, the argument goes, the viewof some-people-in-particularhas infactpassed as theview of no-one-in-particular. However,the nuances of the critiques of objectivity presented by feminists mean that thispithy statement of thefeminist position is overly simplistic. Compare, for instance, thevarious ways in whichLongino, Bordo, Harding, and Scheman havediagnosed the ‘problemof objectivity’ . Longino(1990) argues thatit is social epistemic engagement, ratherthan individual epistemic practices, that does and can workto correct for the tendencyof all ofus tomiss the ‘ pre-understandings’that operate in our theory construction.Bordo (1987,1999) uses a psychoanalyticmodel to illuminate the contoursof the shift to Cartesian rationalism. Harding (1993)argues thatthe way to make sciencemore objective is to ensure that social/ politicalpositions can bemade moretransparent in the scientiŽ c process.Finally, Scheman (1995) argues thatthe identitiesof individual knowers are not contingent facts about knowers,but are, in fact,irreducibly involved in epistemicpractices. 2 Thevariety has oftenbeen lost on critics of feminist philosophy’ s engagementwith theideal of objectivity, however. Haack argues that,in order for the critique to make senseas aspeciŽcally feminist critique,it must assume thatwomen have a unique culture,underwritten by speciŽ cally femaleways of knowing (Haack 1998,p. 125, p.138). Nussbaum claims thatfeminist discussions of objectivity do notshow that it is,in principle, impossible, only that some instances of supposed objectivitywere, in fact,instances of bias. Nussbaum argues thatobjectivity as atypeof impartiality has beenuseful in helping women to win court cases and equal rightsin a varietyof instanceswhere they were unfairly victimized by male bias(1994, p. 59). But itis not clearthat Haack and Nussbaum havethe same thingsin mind here: Haack seemsto equatethe feminist discussion of objectivity with individual ways of thinking; Nussbaum seemsto have in mind an idealof equal treatmentunder the law. This shouldnot be surprising, however, given the nature of the concept being defended fromattack. The fact that Haack and Nussbaum defendtwo different things from attack—that is, they see two different things as beingjeopardized by the feminist critiqueof objectivity— ispart and parcelof the fact that ‘ objectivitythe ideal’ is not conceptuallycoherent. I willdefend this claim inthe next section. 274 MARIANNEJANACK

2. Metaphors ofobjectivity:crazy-quil ting ourway to an ideal

Theideal put forwardas theideal of objectivity is, as Finehas remarked,a hodgepodge (Fine1998, p. 14). As my examplesat thebeginning of the paper show, what is called ‘objectivity’or what is demanded in the name of objectivity, or what seems to be missing indifferentfailures of objectivityis variable.A recentarticle in abookentitled Objectivity and ItsOther remarkedthat ‘ objectivity’doesn’ t havejust one oppositional term, but ratherit has many (Weber1995) testifyingboth to its richness as aconceptand its importanceas an ideal.Yet, as clearas ourcommonsense intuitions are with respect toobjectivity in the ordinary course of events, the case turnsout to be much more complicatedwhen the concept is subjected to closer scrutiny, as my discussionof the cases offailures of objectivity showed. Just whatobjectivity is,as Ihaveindicated, is hard to say. Megill (1994) cites four differentsenses of the term, while Fine (1998) outlines a widevariety of conceptions ofobjectivity— objectivity as astyleof thought, scientiŽ c method,procedural, the viewfrom nowhere, the view of no one in particular, the fortiŽ cation against relativismand irrationalism,impersonal product, unbiased product, and as the really real—someof which overlap with Megill’ s categories,some ofwhich do not.Fine arrivesat theconclusion that in spite of this apparent diversity, the issue of objectivityis, at bottom,an issueabout trust.By contrast,Nozick (1998) equates objectivitywith invariance, using ‘ objective’as amodiŽer for ‘ truth’and ‘fact’(which, Itakeit, means that‘ objective’is not synonymous with either ‘ truth’or ‘ fact’— some truthsor facts, by this reckoning, are ‘ objectivetruths’ or ‘ objectivefacts’ while other truthsor facts are not objective). Invariance, according to Nozick, is what we get whenwe abstract from the other qualities of objectivity, which are:

(1)accessibility from different angles; (2)the possibility of intersubjectiveagreement; and (3)the independence of a giventruth or fact ( p)frompeople’ s ‘beliefs,desires, hopes and observationsor measurements that p’(1998,p. 21).

Nagel has famously coinedit ‘ theview from nowhere’ (1986) while Williams has termed itthe‘ absoluteconception’ (1985). Perhapswhat is most strikingin these discussions (with the exception of Fine’ s) is the invocationof the ideal of perspective to explain something that is the very antithesis of perspective.Indeed, the heavy use of metaphors in general in these philosophical discussionsis what I Žndintriguing. While philosophers have often distrusted metaphors—suspecting that there is something smuggled inwith the laundry, so to speak,or that some vaguenessor ambiguity is being left in place whereit should be exorcisedwhen metaphors are invoked— the literature on objectivity seems to use metaphorsto do conceptualwork. More than that, however, philosophers and scientistswriting on objectivity seem to abandon themselvesto this ‘ driveto metaphorize’with nary a blink. 3 Thatmetaphors are essential to thoughtis not a newclaim. 4 Butthat they areso openly anddeliberately employed in philosophical analyses ofobjectivityis worthy ofremark. For notonly doesthe metaphorof perspective implicitly framethe concept ofobjectivity,it is alsoexplicitly invokedto explain what is meant bythe concept. Rather than beingsimply an heuristic device,the metaphorof perspective is doingdouble duty. It is boththe ‘cognitiveframe’ for the concept andthe explanationof the concept. Wittgenstein gives DILEMMASOFOBJECTIVITY 275 ussome reasonto think that the driveto metaphorizeis away ofsaying that ‘myspadeis turned’— that is,there is something which is unexplainable,or, at least, unexplainablein the terms set bythe problem.It can alsomean that we have come tothe Žnal justiŽcation upon which all otherjustiŽ catory moves depend.Perhaps the seemingly irresistibleurge to explain objectivity in terms ofa metaphorthat is,in fact, not very explanatory,just means that we have reached conceptual bedrockin this case. However,I thinkthe issue is more philosophically signiŽ cant. A certainkind of instabilityseems endemic to an analysisof the concept. We use the idea of perspective toexplicate the ideal of perspectivelessness. As a result,the ‘ frame’undermines the ‘target’of the metaphor. The frame of the metaphor is the notion of perspective or view,while the target is the idea of objectivity. But insofaras themetaphor attempts to captureobjectivity as aperspectiveor view that is not aperspectiveor view, the metaphorworks against itself. Metaphors serve explanatory functions insofar as weare ledto see something new in the target of the metaphor by its juxtaposition with the frame.The metaphorical deŽ nitions of objectivity depend on the frame of ‘perspective’ or‘ view’to explain a ‘viewfrom nowhere’ and aperspectivethat is not a perspectiveat all,and isnot even the sum ofall perspectives,but is, in fact, aperspectivalness. I think theconceptual instability created by the metaphorical deŽ nitions of objectivity are unavoidable,however. This is my senseof the issue: that metaphors are invoked to explainand captureobjectivity because there are so many differentmeanings to the termthat they can onlybe captured if wehaverecourse to a metaphorthat invites itself tobe understood in a varietyof ways— some ofthem competing, some ofthem non- competing.Those meanings of the term that are non-competitors are so becausethey are,in fact,irrelevant to eachother. ‘ Objectivity’refers to so many differentthings that itcannot be captured in a purelydescriptive and literalform. In addition to the meaningslisted by Megill, Fine and Nozick,my surveyof the literature on objectivity has turnedup thefollowing uses of the term: 5

(1)objectivity as valueneutrality; (2)objectivity as lack ofbias, with bias understood as including: (a) personalattachment; (b) politicalaims; (c)ideological commitments; (d) preferences; (e)desires; (f)interests; (g) emotion. (3)objectivity as scientiŽc method; (4)objectivity as rationality; (5)objectivity as an attitudeof ‘ psychologicaldistance’ ; (6)objectivity as ‘world-directedness’; (7)objectivity as impersonality; (8)objectivity as impartiality; (9)objectivity as havingto do withfacts; (10)objectivity as havingto do withthings as theyare in themselves; objectivity as universality; (11)objectivity as disinterestedness; (12)objectivity as commensurability; (13)objectivity as intersubjectiveagreement. 276 MARIANNEJANACK

Whilesome of these meanings seem to be related, there are slight differences in shadesof meaning, and oneuse does not necessarily equate with any of the otheruses of the term. ‘ Universality’, forinstance, is not necessarily equivalent with‘ valueneutrality’ nor is it equivalent with ‘ havingto do withthe world of facts’because what is universal might be value-laden, and so on.And one shouldnot be misled into thinking that the uses of the term within a particular domain (e.g.scientiŽ c discourse)are at leastinternally consistent; the fact of the matteris that these meanings all circulatefairly promiscuously over the whole rangeof domains. The discussion of objectivity in the sciences invokes all 14 of thesemeanings, and discussionsof bias (or lack thereof)in the sciences invoke eachof the categories listed under meaning 2. The terms that seem to draw on legal orpolitical ideals of objectivity pop up justas oftenin scientiŽ c discourses as do theepistemic ideals. But thereis a tendencyto run all theseterms together, ignoring the fact that some of theseare epistemic ideals, some aremoral ideals,some arerelated to thepolitical realm, someto the legal realm,some re ect a preferencefor a certainkind of ‘ coolness’and judiciousness,some encapsulate ontological commitments, and someare psychologicalattitudes. The metaphorical nature of ‘ objectivity’and itsreasonably commonsenseand everydayapplications, disguises the fact that, if we try to Ž gure outwhat is really meant by ‘ objectivity’we Ž nda dizzyingarray of different kinds of virtues,ideals, metaphysical positions and psychologicalstates. The philosophical tendency,though, seems to push for some unifying virtue under which all these differentmeanings might be grouped, one virtue or ideal that underwrites them all or capturesmost orall ofthem. Philosophers—feminist and non-feministalike— have rushed to provide ‘ successor virtues’or have tried to show that, in spite of the diversity of meanings, what all uses ofthe term ‘ objectivity’have in common issome other moral-epistemic virtue. The idealof objectivity has beentranslated or re-conŽ gured variously as epistemic responsibility,re exivity, dialectical responsiveness, democratic openness, thinking fromthe lives of the marginalized, intersubjective agreement, and trustworthiness. Thejob is, in a way,a reclamation.Philosophers who have tackled this issue have triedto identify what it is about objectivitythat we think is so valuable,without takingon or assuming thosemeanings of the term that seem troublesome. It is an attemptto reclaim the term as useful,in spite of some ofits drawbacks— one of whichwould seem to be its amorphous nature. But whatis itabout thenotion of ‘ objectivity’that makes itbothso ‘commonsensical’ and yetso hardto explain, theorize or analyse? Toparaphrase Oliver Wendell Holmes, wemay notbe able to deŽ ne it, but wethink we know it (or, rather, we recognize its failure)when we see it. ‘ Objectivity’seems to be a collectionof a varietyof different kindsof ideals, captured by varying terms, whose outlines seem to come intofocus themore they are absent in a particularinquiry, decision or theory. Furthermore, we seemmost ableto see what is missing from a particularfailure of objectivity only in comparisonwith other similar cases; when we think of objectivity and itsfailures, it becomesstrikingly easier to talk about whatkinds of things we value in action and judgmentby comparing differentkinds of scenarios. There does not seem to be any particularvirtue or cluster of related virtues absent from cases inwhich we determine thatsomeone has failedto be objective. Rather, we make ourjudgments on the basis ofcomparisons, and thecollection of things that are identiŽ ed as lackingin particular failuresof objectivity seems to evade the philosophical attempt to reduce them to a DILEMMASOFOBJECTIVITY 277 singleideal. Any ideal that can bemade toŽ tisin dangerof being true only in virtueof itsover-generality. Ideals thatevade formalization, analysis or deŽ nition might nonetheless serve differentkinds of purposes; indeed, it is often their very  exibilitythat makes themso useful,since they can readilybe adapted tothe needs at hand.But feministand non- feministattempts to replace objectivity with other kinds of epistemic virtues (e.g. trust,responsibility, etc.) both show the  exibilityand proteannature of ‘ objectivity’ while,at thesame time,covering it over. In fact, it seems that rather than having ‘objectivity’we have ‘ objectivities’, and theattempt to reduce ‘ objectivity’to some otherpolitical, moral, or epistemic virtue disguises that fact. Doesthis mean that the attempts to re-deŽne objectivity by philosophers of science and epistemologists(both feminist and non-feminist)have been in vain? The answer here is ‘no’. Theattempt to re-deŽ ne objectivity has beenimportant work. Indeed, it has beenessential work, and Imyselfhave been working on itboth in my researchand in theseminars I havetaught onthe problem. However, I do notthink it can continue toserve as an endin itself, nor should it continue in isolation from the attempt to exacerbateand call forththe ways in which the ideal has beencobbled togetherout of avarietyof disciplinary discourses and boundarydisputes, political movements, and philosophicalworries, leaving us withan idealthat is a variegatedmotley. 6 The projectof re-deŽ ning objectivityhas hadthe unfortunatesecondary effect ofdisguising andpapering over the faultlines that exist between the variety ofdifferent uses ofthe term. Even within aparticulardomain (e.g. science), the senses ofthe term vary considerably.For Nozick, the term means ‘invariance’and for Fine itmeans ‘trust’; Schefer uses itto mean, variously, ‘ scientiŽc method’, ‘rationality’, ‘impersonality’and ‘intersubjective agreement’. Clearly,rationality is notsynonymous with scientiŽc method,nor does scientiŽ c methodyield intersubjective agreement in any interesting or reliableway; there mightbe said to be more intersubjective agreement aboutastrology orpsychic healingthan there is about,for example, any particulartheory ofhigh- density physics. Hadit not been forvarious attempts tore-deŽ ne ‘objectivity’, the hodgepodgewould not have beenrevealed. Nevertheless, the fact that hopesprings eternal that the successor virtuefor ‘ objectivity’will befound means that such attempts have alsoserved toavert oureyes fromthe hodgepodgenature of the ideal. Thegaping voidthat opens up oncewe put theconcept of ‘ objectivity’to arigorous analyticaltest is philosophically instructive as wellas usefulfor feminist purposes. Rather thantrying to eliminate or overlookthe tensions created by the various meanings of the term,we needto emphasize them, both as amatterof philosophical correction and as a pointof feminist theoretical engagement. We need to point out the ways in which philosophershave had asortof blind spot where the concept of ‘ objectivity’is concernedwhich has ledto a lack ofanalytical rigour. But wecan also usethe tensionsand inconsistenciesto address different challenges to feminist epistemology and philosophyof sciencemore generally.

3. ‘Isfeminist philosophy really philosophy?’and other commonchallenges tofeminist workin epistemologyand philosophy ofscience

Thekinds of questions that feminists are asked when we propose different kinds of ‘procedural’objectivities are often questions about themetaphysical senses of objectivity.So, for instance, when feminists propose more democratic and open 278 MARIANNEJANACK processesof scientiŽ c inquiryas awayof makingscience more objective, we areasked, ‘But willthat make forbetter science?’ 7 The presumedconnection between procedural objectivity(understood as scientiŽc method)and metaphysical objectivity(understood as ‘the way the worldis’ ) is the unanalysed andunannounced premise withoutwhich the questioncannot operate.By focusingon the differentkinds of‘ objectivities’, we can show betterthe ways in which such questionsare vague and ambiguous. As Fine notes, the connection between proceduralobjectivity and metaphysical objectivityis ‘tenuous atbest’ (1998, p. 16). As aresult, such questionsare loaded, playing upon the ambiguitybetween the proceduraland the ontologicalsenses ofthe term ‘objectivity’to criticize feminist philosophicalclaims aboutscience andepistemology. Takingthe concept of objectivity apart toexpose the ways in which it encompassesa vastand disparatecollection of epistemic goods, moral virtues,political goods, psychologicalattitudes, and ontologicalcommitments has afurtherbeneŽ t, too. It allowsfeminists to respond to Antony’ s ‘paradox ofbias’ and toshow the ways in whichcharges of unfairness and bias againstwomen are not inconsistent with claims thatperspective is unavoidable. For we can say that—in bringing charges of unfairnesswhen we say thatreviewers or search committees are biased against feministwork— we are invoking the ‘ common’everyday notion of objectivity as fairness.When we make apointabout theinescapability of perspective, we are callingattention to the fact that the metaphorical/ conceptualunderstanding of objectivitydepends upon thevery thing it is trying to eliminate (i.e. perspective). To call bothof these ‘ biases’is to miss the important differences between ‘ bias’as ‘unfair partiality’and ‘bias’as ‘perspective’. Thus,it seems to me that we can make appeals toobjectivity as fairnessin our critiques of unfair practices, while, at thesame time, invokingthe inescapability of perspective. Thus, tensions between feminist approachesto the issue of objectivity are not really about thefeminist epistemological project,as Antonyseems to imply, nor are they about aschismbetween ‘ liberal’and ‘postmodernist’feminisms, as Nussbaum and Haack seemto imply. The tensions arisebecause of the con icted and proteannature of objectivityas an ideal. However,suppose weinterpret the ‘ paradox ofbias’ as aparadox about perspective—that is, as theclaim thatperspective is inescapable, and thatthere are betterand worse‘ perspectives’. This,however, does not seem to be a paradox at all, but rathera fairlytrivial truth. In fact, philosophy invokes the idea of the superiority ofsome perspectives over others all thetime, even within the constraints of the traditionalunderstanding of objectivity as the‘ viewfrom nowhere’ — theperspective ofnatural science is sometimes taken to be such a perspective,as isthe perspective of theideal observer. Feminists can simplyrespond that a perspectiveinformed by feministanalysis is superiorto perspectivesnot so informed. Similarly,the feminist critique of objectivity as acertainkind of psychological attitude ofdistance does not entail anything in particular about therelative merits of objectivity as fairness.It is only if we thinkthat taking up acertainsort of psychological attitude is necessaryfor any of the other kinds of objectivities that we will mistake the feminist critiqueof such an attitudeand itsvalue for an attackon all thedifferent kinds of virtues—epistemic, political, moral— that sit uneasily grouped togetheras ‘objectivity’. 8 Feministscan respondsimilarly to Haack’ s chargethat the feminist critique of objectivitymust bepremised on a peculiarly‘ female’culture or way of knowing. Emphasizingthe variety of goods and virtuesthat march underthe banner of ‘objectivity’shows that we need not assume thata critiqueof any particular ideal of objectivitycommits us tothe assumption of ‘ women’s waysof knowing’ . Haack DILEMMASOFOBJECTIVITY 279 presupposesthat objectivity is about individualreasoning, so thatwomen would have to ‘thinkdifferently’ than do menin somenon-trivial way that was also inherentlyfemale. But thefeminist discussions that Haack singlesout for criticism— particularly Harding’ s work—focus not on objectivity as an individual’s wayof thinking or psychological attitude,but onthe idea of procedural objectivity. Thus, feminists need not make politicallyand scientiŽcally questionableclaims about individualreasoning that presumea biologicallyhard-wired ‘ femaleway of knowing’ and thatgive rise to ‘women’s culture’. Theyneed only emphasize the ways in which experience is (or failsto be) integrated into theory in procedural objectivity and theways in which scientiŽc methodencourages or fails to encourage re ection on pre-theoretic commitmentsand sociallyand historicallyspeciŽ c patternsof experience. 9 Thispaper began withdifferent ‘ dilemmas ofobjectivity’ — thatis, different puzzles about what’s missingwhen objectivity is saidto be missingfrom a particularaction or judgment.While the thumbnail sketches seem clear-cut enough as failuresof objectivity,establishing what exactly is missing when we are presented with a failureof objectivity poses its own problems, especially when we try to do so by comparing different,but similar,cases. And we all facedilemmas ofobjectivity on aregularbasis— sorting out what is required of us ifwe are to be ‘ objective’in a givencase is the stuff of agonized discussions, re ection and consultation.That the processis often arduous, and thatthe right approach isoften not clear is partly dueto the variety of ideals we juggle in our re ection on our moral, political and epistemicpractices. My emphasis on the variegated motley that makes up theideal of‘ objectivity’is intended as aphilosophicaland feministclariŽ cation, which I thinkis required to counterbalance the conventional wisdom in philosophical and feministcircles that we can isolatesome one or two different virtues or meanings forthe term that cover all ormost ofour uses of the term. The attempts to deŽ ne ‘objectivity’in terms of other epistemic or moral virtueshas had,as Ihavesaid, theunfortunate secondary effect of masking the instability of the concept as wellas disguisingthe philosophical signiŽ cance of the fact that a vastand disparatearray ofvirtues, goods and proceduresare collected under the umbrella ofthe concept. Thatthey all belongunder the umbrella ofthe term ‘ objectivity’seems to have led us ona ratherlong route of identifying the common ground,or cluster of commonalities,that would tie them all together.Not onlyis this project not promising,but also Ithinkthe important philosophical and feministwork at this pointis the work of taking apart theconcept of objectivity, showing not what holds ittogether, but how great the fault lines between its different senses might be. To thisend, I havetried to show how criticisms of feminist epistemology and philosophyof science can beanswered by emphasizing the hodgepodge natureof theconcept rather than its coherence. Is my emphasison taking the concept of ‘ objectivity’apart an attackon the concept?The answer is complicated. If ‘ objectivity’is taken to be a singleuniŽ ed conceptand itsphilosophical value depends on it being coherent, then my claims constitutean attack.On theother hand, we can also understandthis as philosophicalprogress. Feminist and non-feministanalyses have shown the ideal of objectivityto be a largeand unwieldycollection of possibly competing, possibly non-competingideals and metaphysicalpresuppositions at workin philosophical and scientiŽc discourse,and thisseems to me to constitute a clariŽcation. As a clariŽcation, the feminist discussion of objectivity also seemsto me to exemplify oneof the highest callings of philosophical discourse. 280 MARIANNEJANACK

Notes

1.It isimportant to note, however, that manyfeminists and non-feminists alike have argued that the requirement that ourevaluation of the arguerbe separated from our evaluation of the argumentpresupposes a particular theoryabout identity and its relationshipto knowledge, as well as a particulartheory of what counts as knowledge.Further, it seemsthat in somecontexts epistemicefŽ ciency requires that weuse the ethosof a particulararguer as a factorin evaluatinghis/ herclaims. Interesting discussionsof this issuearise in the literatureon testimony as well as in socialstudies of science.See, for instance, Code (1995), Shapin (1994)and Janackand Adams (1999). 2.I realizethat thesethumbnail sketches do not dojustice to the intricaciesof the positionsglossed here. They aremeant to be illustrations only, and not analyses,of the varietyof positionsheld by feministsworking on this problem. 3.Again, Fine is an exceptionhere. He notices quite clearly the waysin which metaphorworks in these discussions. 4.See, for instance, Lakoff and Johnson (1980,1999), Haack (1998, pp. 69– 89) and Davidson (1984) for a small butdiverse collection of discussionsof metaphor. 5.In additionto the worksmentioned earlier, the list comesfrom a surveyof the following:Moser (1993), Rescher(1997), Sacks (2000), Schef er (1982), Rorty (1991), Bordo (1987, 1999), Keller (1983, 1985), Longino (1990),Harding (1993), Nussbaum (1994) and Haack (1998). There is no one-to-onecorrespondence between the senseslisted and the workssurveyed; often two or three of the differentsenses of ‘objectivity’were used in the samework. So, for instance, Schef er shifts his meaningsof the termamong senses 3, 4, 7 and14, sometimeswithin the sameparagraph. 6.See, for instance, Daston (1992) and Shapin (1994). 7.I owethis exampleto Elizabeth Potter, whose re ection on such conversations,along with that ofLinda Alcoff andTrish Glazebrook,helped me sort out these issues. 8.An instructivecase here is that ofNicholas Rescher, who seems to both misunderstandthe feministcritiques of objectivitypresented by Susan Bordo and Evelyn FoxKeller, and then run togetherobjectivity as lack of emotionalattachment with objectivityas ‘the waythe worldis’ (see Rescher 1997). 9.Haack insists that the onlysense that canbe made of the claimspresented as feminist epistemological claims is that they areeither (a) true but not feministor (b)feminist but false.For a detaileddiscussion of this argument andwhy it fails,see Janack and LaRocque (2001). Her argument is especially puzzling, given her claim to be the kindof feministwho emphasizes the similaritiesbetween men and women, since this wouldseem to imply that the convergencebetween feminist and non-feminist theories of ‘scienceas social’ , asHaack terms them, shouldbe a welcomesupport for her understandingof feminismas concernedwith the erasureof gender as a signiŽcant category.Yet, the conclusionthat she drawsis that this convergencemeans that such theoriesare not ‘feminist’. If ‘feminist’does not mean‘ comingfrom women’ or ‘basedon women’s experience’but rather meansthat menand women are essentially the same,then it seemsthat Haackshould conclude that the convergencesupports her pointthat genderis irrelevant, and she shouldembrace feminist theories of ‘science associal’ . What she canmean by sayingthat such theoriesare true but not feministis unclear.

Acknowledgement

Iwouldlike to thankRaja Halwaniand twoanonymous reviewersfor SocialEpistemology , all ofwhom provided important and good feedbackas Iworkedon this paper.

References

ALCOFF, L. and POTTER, E., 1993, Feminist Epistemologies (NewYork: Routledge). ANTONY,L.,1995,Comment on Naomi Scheman. Metaphilosophy ,26,191– 199. BORDO, S., 1987, The Flight to Objectivity (Albany,NY: SUNY Press). BORDO, S., 1999, Re-reading the Canon: Feminist Interpretations ofRene Descartes (UniversityPark, PA: Penn State UniversityPress). CODE, L., 1995, Rhetorical Spaces: Essayson Gendered Locations (NewYork: Routledge). DASTON,L.,1992,Objectivity and the escapefrom perspective. Social Studiesof Science ,22,597– 618. DAVIDSON, D., 1984, Inquiriesinto Truth and Interpretation (NewYork: Oxford University Press). FINE,A.,1998,The viewpointof no one in particular.Presidential Address of the Central Divisionof the AmericanPhilosophical Association. Proceedings and Addresses ofthe APA , 72, 9–20. HAACK, S., 1998, Manifestoof a Passionate Moderate (Chicago:University of ChicagoPress). DILEMMASOFOBJECTIVITY 281

HARDING,S.,1993, What is‘ StrongObjectivity’ ? In L.Alcoffand E. Potter (eds), Feminist Epistemologies (New York: Routledge). HELDKE, L. and KELLERT,S.,1995, Objectivity as responsibility. Metaphilosophy ,26,360– 378. JANACK, M. and ADAMS, J., 1999, The Changing Tradition: Women in the History ofRhetoric (Calgary:University of CalgaryPress). JANACK, M. and LAROCQUE,M.,2001, Fashion emergency: feminists accused of trendiness. Philosophy Now , 33, 9– 11. KELLER,E.F.,1983, AFeeling for the Organism: the Life and Work ofBarbara McClintock (San Francisco:W.H. Freeman). KELLER,E.F.,1985, Reections onGender and Science (NewHaven: Yale University Press). LAKOFF, G. and JOHNSON, M., 1980, Metaphors We Live By (Chicago:University of Chicago Press). LOHNSON, G. and JOHNSON, M., 1999, Philosophy in the Flesh: the Embodied Mind and ItsChallenge to Western Thought (NewYork: Basic Books). LLOYD, G., 1984, The Man ofReason: ‘Male’and ‘Female’in Western Philosophy (Minneapolis:University of Minnesota Press). LONGINO, H, 1990, Science asSocial Knowledge: Values and Objectivity in ScientiŽc Inquiry (Princeton, NJ: Princeton UniversityPress). MEGILL,A.(ed.),1994, Rethinking Objectivity (Durham,NC: DukeUniversity Press). MOSER, P., 1993, Philosophy After Objectivity: Making Sensein Perspective (NewYork: Oxford University Press). NAGEL, T., 1986, The View from Nowhere (NewYork: Oxford University Press). NELSON,L.H.,1990, Who Knows?From Quine to aFeminist Empiricism (Philadelphia:Temple University Press). NOZICK,R.,1998, ‘ Invarianceand Objectivity’ , PresidentialAddress of the Eastern Divisionof the American PhilosophicalAssociation. Proceedings and Addresses ofthe APA , 72, 21–48. NUSSBAUM,M.,1994, Feminists and philosophy. New York Review ofBooks ,20October,pp. 59– 62. RESCHER, N., 1997, Objectivity: The Obligations ofImpersonal Reason (NotreDame, IN: Universityof Notre Dame Press). RORTY, R., 1991, Objectivity, Relativism, and Truth (NewYork: Cambridge University Press). SACKS, M., 2000, Objectivity asInsight (Oxford:Clarendon Press). SCHEFFLER, I., 1982, Science and Subjectivity (Indianapolis:Hackett Publishing). SHAPIN, S., 1994, ASocial History ofTruth (Chicago:University of ChicagoPress). WEBER,S.,1995, Objectivity Otherwise. In W.Natter,T. Schatzkiand J. P.Jones,III (eds) Objectivity and ItsOther (NewYork: Guilford Press). WILLIAMS, B., 1985, Ethics and the Limits ofPhilosophy (Cambridge,MA: Harvard University Press).