You are on page 1of 36

Hypatia, Inc.

Feminist Epistemology: An Interpretation and a Defense


Author(s): Elizabeth Anderson
Source: Hypatia, Vol. 10, No. 3, Analytic Feminism (Summer, 1995), pp. 50-84
Published by: Blackwell Publishing on behalf of Hypatia, Inc.
Stable URL: http://www.jstor.org/stable/3810237
Accessed: 27/03/2010 09:42
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=hypatiainc.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content 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 support@jstor.org.

Hypatia, Inc. and Blackwell Publishing are collaborating with JSTOR to digitize, preserve and extend access to
Hypatia.

http://www.jstor.org

FeministEpistemology:
An Interpretation
anda Defense
ELIZABETHANDERSON

has oftenbeenunderstood
as thestudyof feminine"ways
Feministepistemology
is
better
understoodas the branchof
But
feminist epistemology
of knowing."
that
the
various
studies
social
naturalized,
influencesof normsand
epistemology
and
on theproductionof
and
interests
experiences
conceptionsof gender gendered
claims
about
avoids
dubious
This
femininecognitive
understanding
knowledge.
differencesand enablesfeministresearchin variousdisciplinesto pose deepinternal
critiquesof mainstreamresearch.

Feministepistemologyis about the waysgender influences what we take to


be knowledge.Consider impersonaltheoretical and scientific knowledge,the
kind of knowledge privilegedin the academy.Westernsocieties have labeled
this kind of knowledge"masculine"and preventedwomen fromacquiringand
producingit, often on the pretext that it woulddivert their vital energiesfrom
their "natural"reproductivelabor (Hubbard1990; Schiebinger 1989). Theoretical knowledge is also often tailoredto the needs of mostly male managers,
bureaucrats,and officials exercising power in their role-given capacities (H.
Rose 1987; Smith 1974; Collins 1990). Feminist epistemologistsclaim that
the ways gender categorieshave been used to understandthe characterand
status of theoretical knowledge, whether men or women have producedand
applied this knowledge, and whose interests it has served have often had a
detrimental impact on its content. For instance, feminist epistemologists
suggest that various kinds of practical know-how and personal knowledge
(knowledgethat bearsthe marksof the knower'sbiographyand identity), such
as the kinds of untheoretical knowledge that mothers have of children, are
undervaluedwhen they are labeled "feminine."Given the androcentricneed
to representthe "masculine"as independent of the "feminine,"this labeling
has led to a failure to use untheoreticalknowledge effectively in theoretical
reasoning(Smith 1974; H. Rose 1987).
Hypatiavol. 10, no. 3 (Summer1995) by ElizabethAnderson

ElizabethAnderson

51

Traditionalepistemologyfinds these claims of feminist epistemologyto be


highly disturbing,if not plainly absurd.Some feminist epistemologistsin turn
have rejected empiricism(Harding1986) or even traditionalepistemologyas
a whole (Flax 1983) for its seeming inability to comprehendthese claims. I
argue,contraryto these views, that a naturalizedempiricistepistemologyoffers
excellent prospects for advancing a feminist epistemology of theoretical
knowledge.
The project of feminist epistemologywith respectto theoretical knowledge
has two primaryaims (Longino 1993a). First, it endeavors to explain the
achievements of feminist criticism of science, which is devoted to revealing
sexismand androcentrismin theoreticalinquiry.An adequatefeministepistemology must explain what it is for a scientific theory or practice to be sexist
and androcentric,how these featuresare expressedin theoretical inquiryand
in the application of theoretical knowledge, and what bearingthese features
have on evaluatingresearch.Second, the projectof feministepistemologyaims
to defendfeministscientificpractices,which incorporatea commitmentto the
liberation of women and the social and political equality of all persons.An
adequatefeminist epistemologymust explain how researchprojectswith such
moral and political commitments can produce knowledge that meets such
epistemic standardsas empirical adequacyand fruitfulness.I will arguethat
these aims can be satisfiedby a branchof naturalized,social epistemologythat
retainscommitmentsto a modestempiricismand to rationalinquiry.Feminist
naturalizedepistemologiststhereforedemand no radicalbreakfrom the fundamental internal commitments of empirical science. They may propose
changesin ourconceptions of what these commitmentsamountto, or changes
in our methods of inquiry.But these can be derived from the core concept of
reason,conjoined with perhapssurprisingyet empiricallysupportedhypotheses about social or psychologicalobstacles to achieving them, and the social
and material arrangementsrequiredfor enabling better researchto be done.
To see how such derivationsare possible, modest conceptions of empiricism
and reason must be explained before I outline a feminist epistemology that
employsthese notions.
A MODESTEMPIRICISM

I shall call "empiricism"the view that experience ultimately providesall


the evidence we have about the world (Nelson 1990), or more modestly,that
observationprovides the least defeasibleevidence we have about the world
(Longino 1993a). No thought processoperatingindependently of empirical
evidence can rule out any conceivable hypothesis about the world. I believe
that empiricism,so understood,is congenial to the puzzlingand seemingly
bizarrehypothesesof feministepistemologybecauseit impliestwo things.First,
for all we know, anythingcan cause anything,and anythingmight provide an

52

Hypatia

illuminating fruitfulmodel for any otherphenomenon.There are no sound a


priori restrictions on the concepts or vocabularywe use in describing and
explaining the world,so long as these concepts "turnwheels"in theories that
have empiricalimplications.Second, empiricismimpliesthat the discoveryof
the best theories demandsthe fullest and freest development of our imaginations. There is no reasonto think ourpresentlycrampedand stunted imaginations set the actual limits of the world, but they do set the limits of what we
now take to be possible. We can never know what further stretch of the
imagination might uncover and explain what furtherexpanse of the world.
Since feminist epistemology and feminist criticism of science contain many
empiricalclaims about the influence of gender on science that appearat first
glance to be unimaginable,it is importantto note that nothing in empiricism
justifiesdismissingsuch claims out of hand.
Empiricismis commonly taken to mean something else: a doctrine that
imposesa priorisubstantiverestrictionson the kinds of entities and concepts
that can ultimately figure in science. Variousself-describedempiricistshave
tried to eliminate from science referenceto unobservables,and use of intentional, modal, and evaluativeconcepts, or to reducethese to concepts thought
to be more "naturalistic."These substantivecommitments are simply bets as
to how empiricalscience will actuallyturn out. Transformedinto restrictions
on the permissiblecontent of theories, they are attempts to win the bets by
rigging the game in advance, preventing the exploration of hypotheses that
might show them wrong.This contradictswhat I take to be the fundamental
commitmentsof modestempiricism.Since feministepistemologyand feminist
criticism of science contain many empirical claims couched in unreduced
social, intentional, and evaluative vocabularies,it is importantto note that
modestempiricismis not committedto eliminatingsuch claimsfromscientific
theories.
I take modest empiricism,then, to be a purely methodological doctrine,
which rejects a priori commitments to what the content of our theories and
models must be. Empiricismis promiscuousin its permissibleontology and
opportunisticin its methodsand models.Any hypothesisor method is permitted that advances the goals of discoveringand explaining novel phenomena
consistent with the constraint that the theories produced seek empirical
adequacy.
RATIONALITY
AS REFLECTIVE
ENDORSABILITY

Reason is the power to change our attitudes, intentions, and practices in


response to reflection on the merits of having them or engaging in them.
Theoretical reason is the power to acquire,reject, and revise our cognitive
attitudes (beliefs and theoretical commitments) and our practices of inquiry
through reflection on our reasonsfor holding them and engaging in them-

ElizabethAnderson

53

that is, throughreflectionon argumentsand evidence for ourbeliefs and about


the consequencesof our practices.Reflective endorsementis the only test for
whether a considerationcounts as a reasonforhaving any attitudeor engaging
in any practice of inquiry:we ask, on reflectingon the waysthe consideration
could or does influence our attitudesand practicesand the implicationsof its
influencingus, whether we can endorseits influencingus in those ways.If we
can reflectively endorse its influence, we count the considerationas a reason
for our attitudesor practices (Anderson 1993, 91-98).
This conception of reasonas reflective self-governmentrejects the ideal of
individualisticself-sufficiency,which somefeministshave arguedis androcentric, or expressiveof specificallymale needs (Bordo 1987; Duran 1991). Rational inquiryis a social enterprise(Longino 1990;Nelson 1993). Anything that
counts asevidence fora theorymustbe publiclyaccessible,and in experimental
contexts, replicable by others. Individuals must use tools, methods, and
conceptual frameworks developed by others in order to get their own
inquiries under way. They must rely on the testimony of others to get
evidence that is too costly or difficult for them to gather on their own, and
even to interpret the evidence of their own senses (Coady 1992). Thus it
is impossible for individuals to rely only on themselves, for the very reason
and interpretations of their experience on which they rely and which seems
most to be their own, is a social achievement, not an individual endowment
(Nelson 1990; Scheman 1983).
The social character of rational inquiry suggests two things. First, the
theoriesproducedby our practicesof inquirymay bearthe marksof the social
relations of the inquirers.To the extent that conceptions of gender inform
these social relations, we might expect these conceptions to influence theoretical inquiry. Second, insofar as we reflectively reject certain ways that
gender influences the practices and productsof inquiry,we need not try to
correct these problemsby demandingthat individual investigatorssomehow
abstractfrom their gender or gender-relatedvalues and commitments. Each
individual might be subject to perhaps ineradicable cognitive biases or
partiality due to gender or other influences. But if the social relations of
inquirersare well arranged,then each person'sbiases can check and correct
the others'. In this way, theoretical rationality and objectivity can be
expressed by the whole community of inquirerseven when no individual's
thought processes are perfectly impartial, objective, or sound (Longino
1990; Nelson 1990; Solomon 1994).
FEMINIST
EPISTEMOLOGY
AS A BRANCHOF
SOCIAL
EPISTEMOLOGY
NATURALIZED,

Many theorists have proposedthat we think of feminist epistemologyas a


social branchof naturalizedepistemology(Nelson 1990; Harding1986; Potter

54

Hypatia

1993; Tuana 1992; Antony 1993; Duran 1991). Naturalizedepistemologists


considerknowledgeproductionas an activity in which inquirersaresubjectto
the same causal forces that affect their objects of study (Quine 1969). They
askof science that it providean account of its own activity.This point of view
enables us to investigate empiricallyhow knowledge changes as we change
factorsconcerning the inquirers.Social epistemologyis the branchof naturalized epistemologythat investigatesthe influence of specificallysocial factors
on knowledgeproduction:who gets to participatein theoretical inquiry,who
listens to whom, the relative prestigeof differentstyles and fields of research,
the political and economic conditions in which inquirersconduct their investigations, the social settings in which they interactwith the subjectsof study,
their ideological commitments,the availabilityof models and narrativeforms
in the culture that can be used to structurescientific observationand explain
phenomena,and so forth.Feministepistemologycan be regardedas the branch
of social epistemology that investigates the influence of sociallyconstructed
andnormsof genderandgender-specific
interestsandexperiences
on the
conceptions
productionof knowledge. It askshow the historicalexclusion of women from
theoretical inquiryhas affectedthe direction and content of researchin fields
such as anthropology,philosophy, and psychology; how the use of gender
metaphorsin biology has made some phenomena more salient than others;
how history,economics, and medicine wouldchange if we viewed phenomena
from the standpoint of women's rather than men's lives; how the feminist
movement has changed our data, our ways of describing the data, and our
theories about differencesbetween men and women.
These are all empirical questions. By framing the questions of feminist
epistemologyas empiricalones, feminist theorists can challenge mainstream
theorists, who are largely empiricists,in a way that they cannot responsibly
ignore or dismiss. This way of framing feminist epistemology also enables
feministsto make argumentsfor reformingtheoreticalpractice in termsinternal to the self-criticalcommitmentsof science itself. Feminist criticismsand
remediescan be seen as particular,if surprising,instances of general types of
criticism and remedyalreadyacknowledgedand accommodatedby scientific
practice. For naturalizedepistemology, considered as a tool for improving
scientific practices, is already incorporatedinto the self-critical and selfreforminginstitutions of science.
How can naturalizedepistemology,which studieshow knowledgeclaimsare
actually produced,supportnormative views about how we ought to produce
knowledge claims? This gap between "is" and "ought" is bridged by the
reflective self-endorsementtest. Naturalizedepistemologyconsidersinquirers
in their social relationsassystemsof belief-formationprocesses,and theoretical
inquiryas a social practice that uses these processesto generate new beliefs.
These beliefs in turn are related to one another through variousexplanatory
theories, models, or narratives that aim to produce understandingof the

ElizabethAnderson

55

phenomena being studied.This two-level representationof theoreticalinquiry


suggeststwo ways naturalizedepistemology can get critical leverage on our
knowledge practices. First, we can examine our belief-formationprocesses.
Some of these processesare such that, once we reflect on how they work or
what they do, we lose confidence in the beliefs to which they give rise, since
they do not reliably lead to true beliefs (consider optical illusions). Other
processessatisfythe reflective endorsementtest: reflectingon how they work
or what they do leads us to endorse them and the beliefs to which they give
rise (consider deductive inference). A knowledge practice is rational to the
extent that it promotessuch critical self-reflectionsand respondsto them by
checking or canceling out the unreliablebelief-formationmechanisms and
enabling the reliableones.
The institution of placebo-controlled,double-blind,multi-center trials as
the standardfor testing drugsrepresentsan exemplarycritical achievement of
naturalized epistemology. Each feature of this experimental standard was
institutedin responseto the discoveryof an unreliablebelief-formationmechanism that had to be checked. The well-known placebo effect, in which
subjects report symptom improvementwhen they receive any intervention
they believe mayhelp them, is checkedbyrequiringthat the therapeuticeffects
of drugsbe measuredagainsta control groupwhich is administereda placebo,
and by requiringthat subjectsnot know whetherthey belong to the control or
the experimentalgroup.Wishfulthinking on the partof experimenters,which
leads to exaggeratedreportsof the therapeuticeffectivenessof drugson trial,
is checked by making the tests double-blind,so that even the experimenter
does not know which groupsubjectsbelong to. Multi-centertrialsensurethat
experimental outcomes are not merely an artifact of the micro-cultureof
researchersat a single site. These are all reformsscientific institutions have
made in the past few decades, in response to scientific studies of its own
practice.
The normative implications of much feminist epistemology and feminist
criticism of science can be modeled on the case of double-blindtesting. If a
genderednorm is found to influence the productionof knowledge claims in
ways that cannot be reflectivelyendorsed,then we have epistemic reasonsto
reformour knowledgepracticesso that this norm is changed or its effects are
blocked. Feminist empiricist epistemology thus produces argumentsof the
same logical type as those alreadyaccepted by our knowledgepractices.
Feminist empiricistepistemologycan generate normative implicationsfor
theory in a second way.The model of double-blindtesting worksonly at the
level of weeding out false beliefs. But getting an adequateunderstandingof
phenomena is not simplya matterof removingsexist or androcentricbiasfrom
factualclaims so as to allow scientiststo see unvarnishedtruth.Fortheoretical
inquirydoes not aim simplyto generatetrue beliefs. One can add to the stock
of true beliefs without the aid of systematic theorizing.Although empirical

56

Hypatia

adequacyposes a fundamentalconstrainton theorizing,the point of theory is


to organizebeliefs and generate understandingthrough models that explain
phenomena that people find significant,important,or fundamental,and that
abstractfromphenomena thought to be unimportant.But whether a phenomenon is consideredimportantor fundamentaldependson practicalneeds and
interests, which may be gendered or staked in other socially constructed
positions such as class (Tiles 1987). Theories or models offer us only partial
mapsof the world.Thus differentpeople mayfind differentmodelssatisfactory,
depending on which aspects of the world the models highlight (Longino
1993b, 114-16).
This relativity of the value of a model to the socially conditioned interests
and experiences of the people to whom it is offered does not imply that
theoreticalexplanationsmustbe false,or that all areequallygood, or that there
is no common basis for comparingtheir merits.Empiricaladequacyprovides
the fundamentaland commonstandardforcomparingall theories.But a theory
can be empiricallyadequatewithout being interestingor useful.
Thus, feminist naturalizedepistemologyuses reason both to constrain and
to expand the range of acceptabletheories, given what we know about how
theories are formed. By raisingthe standardsfor evaluating methods of data
collection and interpretationin the light of the reflective endorsement test,
feministepistemologistslimit the field of credibletheories. By legitimizingthe
explicit introduction of feminist interests to justify the choice of different
models, feminist epistemologistsuse reasonin its permissivemode to open up
space for alternative theories oriented towardliberatoryends and to contest
theoriesthat close offpossibilitiesforsocialchangebyrepresentingthe subjects
of studyas if they had no room to maneuver(Longino 1989, 210-13).
Such moves to multiply available explanatorymodels, like the moves to
reformscientific practices on the lines of the double-blind experiment, are
internal to the practices of science. These two types of critical activity
correspondto the two goals of the feministepistemologyof theoreticalknowledge: to legitimate science oriented towardfeminist ends and to underwrite
feminist criticism of sexist and androcentric science. The fact that these
activities can be situated inside science does not mean that the changes
feminist epistemologyrecommendsfor science must be modest. The sorts of
criticismsthat generate internal reformof scientific practices today focus on
such mattersas improvingdata-gatheringinstrumentsand technical features
of experimental method. Feminist epistemology and feminist criticism of
science focus on changing the backgroundsocial conditions in which science
is practiced. It is therefore an explicitly political enterprise,but one that is
justifiedby epistemic values, such as reasonand empiricaladequacy,to which
science alreadydeclaresits allegiance.
The varietyof claims made by feminist epistemologistsand feminist critics
of science is bewildering.Without attemptingto account for or endorseall the

ElizabethAnderson

57

conflicting claims made in the name of feminist epistemology,I shall follow


the strategyof readingmost of them as contributionsto a researchprogramin
naturalizedsocial epistemology.I proposethat we can sort most of them into
four categories, each specifying a particular type of gender influence on
theoretical inquiry.Feministepistemologyhas generallybeen better at identifying the ways gender is implicated in our knowledge practices than at
explaininghow these findingsshouldaffectour evaluationsof the practicesor
the theoriesthey produce(Longino 1993a).Naturalizedepistemologyprovides
a frameworkfordevelopingsuch explanations.So I suggestsome questionsthat
probe the normative implications of each category of gender influence on
theorizing.
First, studies that investigate genderstructuresfocus on the ways gender
normsstructurethe divisionof laborin society,includingthe divisionsbetween
intellectual and manual and service labor, and within the academy,among
differentdisciplines and subfields,and among primaryresearchers,teachers,
and assistants.These studies consider how the content of theories has been
affectedby historical discriminationagainstwomen entering the sciences, by
the difficultieswomen scientists have getting their work recognized,and by
the wayswomen have changedthe orientationof fieldsof studyonce they have
entered the elite ranks in significantnumbers.These studies seek to answer
the question, What difference does, or would, an equal representationand
statusof women researchersmake to theoretical inquiry?
Second, some studies consider the uses of gendersymbolism,which occurs
when we representnonhuman or inanimate phenomena as "masculine"or
"feminine"and model them aftergender ideals or stereotypes.Feminist epistemologists have found gender symbolism to be pervasive in theoretical
inquiry.It is used to representthe relations of scientists to their subjectsof
studyand the relationsof differenttypesof knowledgeor of differentdisciplines
and subfieldsto one another,to describethe characterof scientific objectivity,
and to model nonhuman and inanimate phenomena. These studies seek to
answer the question, What differencedoes it make to our theories and our
scientificpracticesthat we conceive of theoreticalinquiryitself and its subjects
of study as gendered phenomena?How would our theories and practices of
inquiry change if we altered our conceptions of the "masculine"and the
"feminine,"or ceased to employ gendersymbolismin understandingour own
theorizingor inanimate objects?
in biology, the social sciences,
Third, some studies focus on androcentrism
and cultural and literarystudies. Androcentrismoccurs when theories take
males, men's lives, or "masculinity"to set the norm for humans or animals
generally,with female differenceseither ignored or representedas deviant;
when phenomena are viewed from the perspective of men's lives, without
regard to how women see them differently;and when male activities or
predicaments are represented as the primarycauses or sites of important

58

Hypatia

changes, without regardto the roles of females in initiating or facilitating


changes or the ways the situation of females has been crucial to determining
structuralconstraintsand potentialsfor change. These studiesask,How would
the content of theoriesbe differentif we viewed phenomenafromthe perspective of women'slives, or refusedto accept either "masculinity"or "femininity"
as setting the norm for humansor animalsgenerally?
Fourth,some studies focus on sexismin theory,which can appeareither in
practicesthat applythe theoryor in the content of the theory itself. Sexism is
evident when theoriesareappliedin waysthat underminewomen'sinterestsor
that reinforcetheirsubordinationto men. The contentof a theoryis sexistwhen
it assertsthat women are inferiorto men, justlyor inevitably subordinatedto
men, or properlyconfined to gender-stereotypedroles, or when it judges or
describeswomen accordingto sexist idealsor doublestandards,or when it uses
such claims as backgroundassumptionsto secure an evidential link between
observations and theoretical claims. Feminist studies of sexism in theories
explore the prospectsfor alternative scientific theories that meet criteria of
empiricaladequacywhile seeking to serve women'sinterests and to promote
universalequality.
THE GENDERED
DIVISIONOFTHEORETICAL
LABOR

Feministcriticsof science have carefullydocumentedthe historyof women's


exclusion from theoretical inquiry (Rossiter 1982; Schiebinger 1989).
Although formalbarriersto women'sentry into variousacademic disciplines
arenow illegal in the United States, informalbarriersat all levels remain.Girls
are socializedby parentsand peers to avoid studyingor excelling in subjects
considered"masculine,"such as mathematicsand the naturalsciences. Teachers and school counselorsactively discouragegirlsfrompursuingthese subjects
(Curran1980, 30-32). The classroomclimate in mixed-genderschools favors
boys. Teachers pay more attention and offer more encouragementto white
boys than to girls,solicit their participationmore,and expect them to achieve
more, especially in mathematicscourses (Becker 1981; AAUW 1992). Boys
marginalizegirls in class by interruption and sexual harassment (AAUW
1992). These behaviors in mixed-genderschools have a detrimental impact
on girls'academicambitionsand performance.Girls in all-girlschools express
a wider diversityof academic interestsand performbetter academicallythan
girlsin mixed-genderschools (Curran1980, 34). The disadvantageto women's
academic performanceand interests from attending mixed-gender schools
extends to college. The predominantlymale faculty in mixed-gendercolleges
supportwomen students'academicambitionsless than male andfemalefaculty
at women'scolleges. Women'scolleges produce50 percent more high-achieving women relativeto the numberof theirfemalegraduatesthan coeducational
institutions (Tidball 1980). Graduateschools present women with informal

ElizabethAnderson

59

barriersor costs to advancement, including sexual harassmentand exclusion


fromnetworksof male mentorsand colleaguesoften vital to the advancement
of aspiringacademics(Reskin 1979; S. Rose 1989).
Womenwho overcome these obstaclesand obtain advanceddegreesarenot
treated as equals once they enter academicpositions. Women whose qualifications are comparableto their male colleagues get lower pay, less research
support, jobs in less prestigious institutions, lower-rankingpositions, and
positions that assign more and lower-level teaching (Astin and Beyer 1973;
Fox 1981). The prestigeof the graduateinstitution, publications,and having
one's work cited aid men's career advancement much more than women's
(Rosenfeld 1981). Women in scientific and engineering professions with
publicationratesequalto those of their male peershave higherunemployment
rates, lower starting salaries, and lower academic rank than men. These
differencescannot be explained by the greaterimpact on women of marriage
and children (Vetter 1981). The National Science Foundation (1984) found
that afteradjustingforfactorssuch as women interruptingtheir careersto take
careof children,half the salarydifferentialbetween male and female scientists
could be explained only by sex discrimination.
The gendereddivision of theoretical labordoes not simplyprevent women
from doing research or getting published. It fits into a broader gendered
structureof epistemic authoritywhich assignsgreatercredibility,respect, and
importance to men's than women's claims. Laboratory,field, and natural
experiments alike show that the perceived gender of the author influences
people's judgments of the quality of research, independent of its content.
PsychologistsM. A. Paludi and W. D. Bauer (1983) found that a grouptold
that a paper'sauthorwas "JohnT. McKay"assignedit a much higher average
rankingthan a grouptold that the same paper'sauthorwas "JoanT. McKay."
A grouptold that its author was "J.T. McKay"rated the paper between the
other groups'evaluations,reflectingthe suspicionthat the authorwasa woman
trying to conceal her gender identity. Academics are no less disposed than
others to judge the quality of workhigher simplybecause they believe a man
has done it. L. S. Fidell (1970) sent vitae identical in all but name to heads of
psychology departmentsthat advertisedopen rank positions. The jobs the
psychologists said they would offer to the purportedlymale applicant were
higher-rankingthan those they were willing to offerto the purportedlyfemale
applicant.When the ModernLanguageAssociation reviewedpaperssubmitted for their meetings with authors'names attached, men's submissionswere
accepted at significantlyhigherratesthan women's.After the MLA instituted
blindreviewingof papers,women'sacceptanceratesroseto equalitywith men's
(Lefkowitz1979).
The concerns raised by the influence of sexist norms on the division of
theoretical labor and epistemic authority are not simply matters of justice.
Feminist epistemology asks what impact these injustices towardwomen stu-

60

Hypatia

dents and researchershave had on the content, shape, and progressof theoretical knowledge. In some cases, sex discrimination in the academy has
demonstrablyretarded the growth of knowledge. It took more than three
decades for biologists to understandand recognizethe revolutionaryimportance of BarbaraMcClintock'sdiscoveryof genetic transposition.Her attempts
to communicate this discovery to the largerscientific community met with
incomprehensionand disdain.This failurecan be partlyexplained by the fact
that no biology departmentwas willing to hire her for a permanentposition
despite her distinguishedrecordof discoveriesand publications.Lackingthe
opportunitiessuch a position wouldhave providedto recruitgraduatestudents
to her researchprogram,McClintock had no one else doing researchlike hers
who could replicateher resultsor help communicatethem to a widerscientific
community (Keller 1983).
Cases such as McClintock's demonstrate that the gendered structureof
theoretical labor and cognitive authority sometimes slows the progressof
knowledge. But does it change the content or shape of knowledge or the
directionof knowledgegrowth?If the genderof the knoweris irrelevantto the
content of what is investigated, discovered,or invented, then the impact of
removingsex discriminationwouldbe to add to the pace of knowledgegrowth
by adding more inquirers and by raising the average level of talent and
dedication in the research community. Feminist epistemology would then
recommend strictly "gender-blind" changes in the processes by which
research jobs get assigned and epistemic authority distributed. The MLA's
adoption of blind reviewing of papersto reduce cognitive bias due to sexism
in the evaluation of research represents an exemplary application of this
side of feminist epistemology. It is logically on a par with the institution of
double-blind testing in drugresearchto reduce cognitive bias due to wishful
thinking.
But if the gender of the inquirermakesa differenceto the content of what
is accepted as knowledge, then the exclusion and undervaluationof women's
participationin theoreticalinquirydoes not merelyset uprandomlydistributed
roadblocksto the improvementof understanding.It impartsa systematicbias
on what is taken to be knowledge. If the gender of the inquirer makes a
differenceto what is known, then feministepistemologywould not confine its
recommendationsto purelygender-blindreformsin our knowledgepractices.
It could recommend that these knowledge practices actively seek gender
diversityand balance among inquirersand actively attend to the genderof the
researchersin evaluating their products.
The genderof the researcheris known to makea differenceto what is known
in certain areas of social science. In survey research,subjects give different
answersto questions depending on the perceived gender of the interviewer
(Sherif 1987, 47-48). The perceived race of the interviewer also influences
subjects'responses.It is a highly significantvariableaccounting for subjects'

ElizabethAnderson

61

responsesto questionsaboutrace relations (Schuman and Hatchett 1974). In


anthropology,informantsvarytheir responsesdependingon the genderof the
anthropologist. In many societies, male anthropologistshave less access to
women's social worlds than female anthropologistsdo (Leacock 1982). The
race of the researcheraffectsaccess to social worldsas well. Native Americans
sometimes grant Asian anthropologistsaccess to religiousritualsfrom which
they ban whites (Pai 1985).
Where the perceived gender and race of the researcherare variablesinfluencing the phenomena being observedor influencing access to the phenomena, soundresearchdesignmustpay attention to the genderand racialmakeup
of the researchers.In surveyresearch,these effects can be analyticallyexcised
by ensuring a gender balanced and racially diverse research team and then
statistically isolating the variations in responses due to factors other than
subjects'responsesto the characteristicsof the interviewers.In anthropology,
the method of reflexive sociology,insteadof attemptingto analyzeawaythese
effects,treatsthem as a subjectof studyin theirown right.It advisesresearchers
to interpretwhat informantstell them not as straightforward
native observation reportson their own culture, but as reflections of a strategicinteraction
between informant and researcherand between the informant and other
members of the community being studied (Bourdieu 1977). To obtain a
complete representationof informants'reportstrategieswith respectto gender,
both male and female researchersmust interact with both male and female
informantsand consider why informantsvaried their responsesaccordingto
their own and the researcher'sgender (see Bell, Caplan, and Karim[1993] for
exemplarycases of feminist reflexive anthropology).Similarreasoningapplies
to factorssuch as race, class, nationality,and sexual orientation. So reflexive
sociology, like survey research, requires a diversity of inquirers to obtain
worthwhileresults.
The phenomena just discussedconcern the causalimpact of the gender of
the researcheron the objectof knowledge.Manyfeministepistemologistsclaim
that the gender of the inquirerinfluences the characterof knowledge itself by
anotherroute, which travelsthroughthe subjectivityof the researcherherself.
The gender of the researcherinfluences what is known not just throughher
influence on the object of knowledge but by what are claimed to be genderspecificor gender-typicalcognitive or affectivedispositions,skills,knowledge,
interests,or methods that she bringsto the studyof the object. The varietyof
claims of this type must be sorted through and investigated with great care.
Some are local and modest. No one disputesthat personalknowledgeof what
it is like to be pregnant,undergochildbirth,suffermenstrualcramps,and have
other experiences of a female body is specific to women. Gynecology has
certainly progressedsince women have entered the field and have brought
their personalknowledgeto bearon misogynistmedicalpractices.The claims
get more controversialthe more global they are in scope. Some people claim

62

Hypatia

that women have gender-typical"waysof knowing,"styles of thinking, methodologies, and ontologies that globallygovern or characterizetheir cognitive
activities acrossall subjectmatters.Forinstance, variousfeminist epistemologists have claimed that women think more intuitively and contextually,
concern themselves more with particularsthan abstractions, emotionally
engage themselves more with individual subjects of study, and frame their
thoughts in termsof a relationalratherthan an atomistic ontology (Belenky,
Clinchy, Goldbergerand Tarule 1986; Gilligan 1982; H. Rose 1987; Smith
1974; Collins 1990).
There is little persuasiveevidence forsuch globalclaims (Tavris1992, chap.
2). I believe the temptation to accept them is based partly on a confusion
between gendersymbolism-the fact that certain styles of thinking are labeled
"feminine"-and the actual characteristicsof women. It is also partlydue to
the lack of morecomplex and nuancedmodelsof how women enteringcertain
fields have changed the courseof theorizingfor reasonsthat seem connected
to their gender or their feminist commitments. I will propose an alternative
model toward the end of this essay, which does not suppose that women
theorists bring some sharedfeminine differenceto all subjectsof knowledge.
Controversiesover supposedglobal differencesin the ways men and women
think have tended to overshadowother highly interesting work in feminist
epistemologythat does not depend on claims that men and women think in
essentially different ways. The influence of genderedconcepts and norms in
ourknowledgepracticesextends farbeyondthe waysmale and female individuals are socialized and assignedto differentroles in the division of labor.To
see this, consider the role of gendersymbolismin theoretical knowledge.
GENDER SYMBOLISM(I): THE HIERARCHYOF KNOWLEDGE

It is a characteristicof human thought that our concepts do not stay put


behind the neat logical fences philosopherslike to erect for them. Like sly
coyotes, they slip past these flimsy barriersto range far and wide, picking up
consorts of all varieties, and, in astonishinglyfecund acts of miscegenation
shocking to conceptual purists,leave offspringwho bear a disturbingresemblance to the waywardparentand inheritthe impulseto roamthe old territory.
The philosophical guardiansof these offspring,tryingto shake off the taint of
sexual scandalbut feeling guilty about the effort,don't quite know whether to
cover up a concept's pedigree or, by means of the discovery/justification
distinction, deny that it matters. The latter strategycan work only if, like
keepersof a zoo, the philosopherscan keep their animalsfenced in. Feminist
epistemologists track these creaturessneaking past their fences while their
keepersdreamof tamed animalshappyto remainconfined.
The most cunning and promiscuouscoyotes are our gender concepts. In a
manner befitting their own links to sex, they will copulate with anything.

ElizabethAnderson

63

Feminist epistemologistsnote that there is hardlyany conceptual dichotomy


that has not been modeled after and in turn used to model the masculine/feminine dichotomy: mind/body, culture/nature, reason/emotion,
and so forth.These scandalobjective/subjective, tough-minded/soft-hearted,
of
ous metaphorical unions generate conceptions knowledge, science, and
rational inquiry,as well as conceptions of the objects of these inquiries,that
areshapedin partby sexist views aboutthe properrelationsbetween men and
women. Feminist epistemologists investigate how these conceptions are
informedand distortedby sexist imagery.They also considerhow alternative
conceptions aresuppressedby the limitsimposedbysexismon the imagination,
or by the sexist or androcentricinterestsservedby their presentsymboliclinks
to gender (Rooney 1991).
Gendersymbolismappearson at least two levels of ourknowledgepractices:
in the construction of a hierarchyof prestigeand authorityamong kinds and
fields of knowledge and in the content of theoretical inquiryitself. Consider
firstthe waysdifferentkindsand fieldsof knowledgearegendered.At the most
general level, impersonaltheoretical knowledge is coded "masculine."Personalknowledge-the kind of knowledgethat is inseparablefromthe knower's
identity, biography,and emotional experiences-is coded "feminine."Theoretical knowledge is thought to be masculinein part because it lays claims to
objectivity, which is thought to be achieved through the rigorousexclusion
from thought of feminine subjectivity-of emotions, particularity,interests,
and values.These usesof gendersymbolismhave epistemicimportbecausethey
structure a hierarchy of prestige and cognitive authority among kinds of
knowledge,and hence of knowers,that is homologouswith the genderhierarchy. As men in sexist society expresscontempt for women and enjoy higher
prestige than women, so do theoretical knowersexpresscontempt for those
with "merely"personal knowledge of the same subject matters, and enjoy
higher prestige than they. Echoing the sexist norms that women must obey
men but men need not listen to women, the gender-coded hierarchy of
knowledge embodies the norm that personalknowledge must submit to the
judgmentsof impersonaltheoreticalknowledge,while theoretical knowledge
has nothing to learn frompersonalknowledgeand may ignore its claims.
These epistemic norms cannot withstand reflective scrutiny. Successful
theorizingdeeplydependson personalknowledge,particularlyembodiedskills,
and often depends on emotional engagement with the subjects of study
(Polanyi 1958; Keller 1983, 1985). Cora Diamond's(1991) insightfuldiscussion of Vicki Hearne'spersonal knowledge as an animal trainer provides a
particularlyfine illustrationof this point. Hearne'swritings(1982) expose the
failuresof knowledgethat occur when theoristsignore the experiences,skills,
and language of animal trainers. In her animal training classes, Heame
observed that people's success in training their pets was inversely related to
their training in the behavioral sciences. The anthropomorphicand value-

64

Hypatia

laden languageof animal trainersenables them to understandwhat animals


are doing in ways not readily accessible to the impersonal, behavioristic
languagefavoredby most behavioralscientists. And their skills and personal
knowledge of the animals they work with empower trainers to elicit from
animalsconsiderablymore complex and interestingbehaviorsthan scientists
elicit. These powersarenot irrelevantto theorizingabout animals.Reflecting
on Heame's story about the philosopherRay Frey,Diamond writes:
[Frey]attempted to set up a test for his dog'scapacity to rank
rationaldesires.When, in orderto see how the dog would rank
desires,he threw a stick forhis dog... and at the same time put
food before the dog, the dog stood looking at him. Freycould
not see that the dog wanted to know what Freywanted him to
do; Frey'sconception of the dog as part of an experimental
set-up (taken to include two possible desiredactivities but not
taken to include queerbehaviorby the dog'smaster),with Frey
as the observer,blocked his understanding.Frey'spast experience with his dog did not feed an understandingof how the dog
saw him; he could not grasphis own failure,as the dog'smaster,
to make coherent sense, so could not see the dog as responding
to that failureto make sense. (Diamond 1991, 1014 n. 15)
Diamond diagnosesthis epistemic failureas the productof Frey'sattachment
to a theory of knowledgethat distrustspersonalexperience on the groundthat
it is distorted by the subject's emotional engagement with the object of
knowledge.The theorysupposesthat we can't achieve objective knowledgeof
our object throughsuch engagementbecause all it will offer is a reflection of
the subject'sown emotions. Subjectivity merely projects qualities onto the
object and does not reveal qualitiesof the object. But the theory is mistaken.
Love and respectfor anotherbeing, animalor person,and trustin the personal
experiencesof engagementthat are informedby such love and respectmaybe
essential both for drawingout and for graspingthat being'sfull potentialities.
One of the reasonswhy behavioriststend to elicit such boring behaviorfrom
animalsand humans is that they don't give them the opportunitiesto exhibit
a more impressiverepertoireof behaviorsthat respectfor them would require
them to offer.
The gender-codedhierarchyof knowledgeextends to specific subject matters and methods within theoretical knowledge. The natural sciences are
"harder,"more like the male body and hence moreprestigious,than the social
sciences or the "soft"humanities,supposedto be awashin feminine emotionality and subjectivity.Mathematicsis coded masculineand is the languageof
physics, the most prestigiousscience. Through their closer association with
physics, quantitative subfieldsof biology and the social sciences enjoy higher
prestigethan subfieldsof the same discipline or branch of science employing

ElizabethAnderson

65

a qualitative,historical,or interpretivemethodology.Experimentationasserts
more control over subjectsof study than observationdoes. So experimental
subfieldsin biology and psychologyare coded masculine and commandmore
cognitive authoritythan observationalsubfieldsof the samedisciplines.Values
are designatedfeminine. So normativesubfieldsin philosophy such as ethics
and political philosophy enjoy less prestige than supposedlynonnormative
fieldssuch asphilosophyof languageandmind.Social interpretationis thought
to be a feminine skill. So interpretiveanthropologyis designatedless masculine, scientific, and rigorousthan physical anthropology,which deals with
"hard"facts like fossilbones. In each of these cases,the sociallyenforcednorm
forrelationsbetween fields of knowledgemirrorsthat of the relationsbetween
husband and wife in the ideal patriarchalfamily: the masculine science is
autonomousfromand exercisesauthorityover the feminine science, which is
supposedlydependenton the former'spronouncementsto know what it should
think next.
This genderedhierarchyof theoreticalsubfieldsproducesseriouscognitive
distortions. Carolyn Sherif (1987) has investigated how the hierarchy of
prestigegeneratescognitive biasesin psychology.Fortyyearsago,experimental
psychology dominated developmental and social psychology.The gendered
character of this difference in cognitive authority is not difficult to read.
Experimentalpsychologists,by imitating the methods of the "hard"sciences
through manipulatingquantifiedvariables,claim some of the prestigeof the
naturalsciences. Developmentaland social psychologistsengage in laborthat
looks more like the low-status labor conventionally assigned to women.
Developmental psychologists work with children; social psychologists deal
with human relationships,and forty years ago usually did so in settings not
under the control of the researcher.Following the norm that "masculine"
sciences need not pay attention to findingsin "feminine"sciences, which it is
assumedcannot possiblybear on their more "fundamental"research,experimental psychology has a history of constructing experiments that, like Ray
Frey's,ignore the ways the social context of the experiment itself and the
social relation between experimenter and subject influence outcomes. The
result has been a history of findings that lack robustness because they are
mere artifacts of the experimental situation. In experimental research on
sex differences, this error has taken the form of ascribing observed differences in male and female behavior underexperimental conditions to innate
difference in male and female psychology rather than to the ways the
experiment has socially structured the situation so as to elicit different
responses from men and women.
The notorious claim in experimental psychology that women are more
suggestiblethan men offersan instructiveillustrationof the perilsof ignoring
social psychology (Sherif 1987, 49-50). The original experiments that confirmedthe hypothesisof greatersuggestibilityinvolved male researcherstrying

66

Hypatia

to persuademen and women to change their beliefs with respect to subject


mattersoriented to stereotypicalmale interests.Unaware of how their own
gender-typical interests had imparted a bias in the selection of topics of
persuasion,the predominantlymale researchersconfidently reportedas a sex
differencein suggestibilitywhat was in fact a differencein suggestibilityowing
to the degree of interest the subjects had in the topics. Differences in the
gender-typed cognitive authority of the researcher also affect subjects'
responses.Men are more open to the suggestionsof a female researcherwhen
the topic is coded feminine, while women are moreopen to the suggestionsof
a male researcherwhen the topic is coded masculine.
Cognitive distortionsdue to the gender-codingof typesand fields of knowledge are strictly separablefrom any claims about differencesin the ways men
and women think. Although it is true that the "feminine" sciences and
subfieldsattract more women researchersthan the "masculine"sciences do,
the differencesin cognitive authoritybetween the varioussciences and subfieldsweremodeledon differencesin socialauthoritybetween men andwomen
beforewomen constituted a significantportion of the researchersin any field.
Men still predominateeven in fields of study that are designated feminine.
And scientists'neglect of personalknowledgedeprivesmanymen who engage
in stereotypicallymale activities of cognitive authority.For example, animal
behavioristsignore the personalknowledgemale policemen have about their
police dogs (Diamond 1991). For these reasons, Diamond and Sherif have
questionedhow genderfiguresinto the cognitive distortionsinstitutedby the
hierarchy of knowledge and by scientistic conceptions of objectivity.1By
shifting our focus from gender structureand supposedgender differencesin
waysof knowing to gendersymbolism,we can see how ideasabout gendercan
distortthe relationsbetween formsof knowledgeindependentlyof the gender
of the knower.In the light of the cognitive distortionscaused by the gendercoding of typesanddomainsof knowledge,feministnaturalizedepistemologists
should recommend that we no longer model the relations between different
kinds of knowledge on a sexist view of the authorityrelations between men
and women.
GENDERSYMBOLISM
(II):THECONTENTOFTHEORIES

Gender symbolism figures in the content of theories as well as in their


relationsof cognitive authoritywhenever conceptions of human genderrelations or gendered characteristicsare used to model phenomena that are not
gendered. Biology is particularlyrich with gender symbolism-in models of
gamete fertilization,nucleus-cell interaction, primatology,and evolutionary
theory (Biology and Gender Study Group 1988; Haraway1989; Keller 1985,
1992). Evelyn Fox Keller,a mathematicalbiologist and feminist philosopher
of science, has exploredgendersymbolismin evolutionarytheory most subtly

ElizabethAnderson

67

(Keller 1992). Consider the fact that evolutionarytheory tends to delineate


the unit of naturalselection, the entity accordedthe statusof an "individual,"
at the point where the theoristis willing to use complex and cooperativerather
than competitive models of interaction. Among individuals, antagonistic
competition predominatesand mutualisticinteractionsare downplayed.The
individualis considered"selfish"in relation to other individuals.Thus, theories that take the gene to be the unit of selection characterizethe gene as a
ruthlessegoist readyto sacrificethe interestsof its host organismfor the sake
of reproducingitself (Dawkins 1976). Where the organismis taken to be the
unit of selection, it is representedas selfishlycompetitive with respectto other
individualorganisms.But within the individual,cooperationamongconstitutive partsprevails.Cooperationis modeledafterthe family,often a patriarchal
family.The cells of an individualorganismcooperatebecause of the bonds of
kinship:they sharethe same genes. The constitutivepartsof an individualcell
cooperate because they are ruled by a wise and benevolent patriarch,the
"mastermolecule" DNA, which autonomouslytells all the other partsof the
cell what to do, solely on the basis of information it contains within itself.
Thus, evolutionary theory models the biological world after a sexist and
androcentric concepition of liberal society, in which the public sphere is
governed by competition among presumablymasculine selfish individuals
and the private sphere of the family is governed by male heads of households
enforcing cooperation among its members (Keller 1992, chap. 8). This
model is not rigidly or consistently applied in evolutionary theory, but it
does mark theoretical tendencies that can be traced back to the fact that
Darwin modeled his theory of natural selection after Malthus's dismal
model of capitalist society.
Takenby itself, that evolutionarytheoryemploysa sexist ideology of liberal
society to model biological phenomena does not have any straightforward
normative implications. Defenders of the theory can appeal to the discovery/justificationdistinction here: just because a theory had its origins in
politically objectionable ideas or social contexts does not mean that it is false
or useless. Evolutionarytheory is extraordinarilyfruitfuland empiricallywell
confirmed. The model-theoretic view of theories, widely used by feminist
empiricistsand feminist postmoderniststo analyzethe roles of gender in the
constructionof theoreticalknowledge,affirmsthe epistemic legitimacyof any
coherent models, hence of any coherent sexist models, in science (Longino
1993b;Haraway1986).
In the model-theoreticview, scientific theoriesproposeelaboratemetaphors
or models of phenomena. Their virtues are empirical adequacy,simplicity,
clarity,and fruitfulness.Theories are empiricallyadequateto the extent that
the relations among entities in the model are homologous with the observed
relations among entities in the world. Empiricallyadequate models offer a
satisfactoryexplanationof phenomena to the extent that they model unfamil-

68

Hypatia

iarphenomena in waysthat aresimple,perspicuousand analyticallytractable.


They arefruitfulto the extent that they organizeinquirers'conceptions of their
subjects in ways that suggest lines of investigation that uncover novel phenomena that can be accommodatedby further refinements of the model.
Empiricistsplace no a priori constraintson the things that may constitute
useful models for phenomena. Anything might be an illuminating model for
anything else. So, empiricistscan offer no a priori epistemic objections to
modeling nongendered phenomena after gendered ones, even if the models
are overtly sexist or patriarchal.Such models may well illuminate and effectively organizeimportantaspectsof the objects being studied.
So the troublewith using sexist gendersymbolismin theoretical models is
not that the models are sexist. The trouble lies rather in the extraordinary
political salience and rhetoricalpowerof sexist gender ideology,which generates numerous cognitive distortions. Keller has carefullydelineated several
such distortionsin evolutionarytheory,especiallywith respectto its privileging
of models of competitive over cooperative or mutualist interactions among
organisms.First, to the extent that political ideology incorporatesfalse conceptualidentities and dichotomies,a scientific model borrowingits vocabulary
and structureis likely to overlook the alternativessuppressedby that ideology
or to elide distinctionsbetween empiricallydistinct phenomena.The ideology
of possessive individualismfalsely identifies autonomy with selfishness and
falselycontrastsself-interestwith cooperation.When used to model phenomena in evolutionarybiology,it leadsto a false identificationof peaceful,passive
consumption activity with violent, competitive behavior,and to a neglect of
mutualist interactions among organisms.Thus, the mathematical tools of
population biology and mathematicalecology are rarelyused to model cooperationamongorganismsalthoughthey could do so; in contrastwith sociobiology, these mathematicalsubfieldsof biology have even neglected the impact
of sexualintercourseandparentingbehavioron the fitnessof organisms(Keller
1992, 119-21). Although the technical definition of competition avoids false
identities and dichotomies, biologistsconstantly turn to its colloquial meanings to explain their findings and frameresearchquestions. In this way, "the
use of a term with established colloquial meaning in a technical context
permits the simultaneoustransferand denial of its colloquial connotations"
(Keller 1992, 121). When the languageused in a model has particularlystrong
ideological connotations, the cognitive biases it invites are particularlyresistant to exposureand criticism.
The symbolic identification of the scientific with a masculine outlook
generatesfurthercognitive distortions.The ideology of masculinity,in representing emotion as feminine and as cognitively distorting,falsely assimilates
emotion-laden thoughts-and even thoughts about emotions-to sentimentality.In identifyingthe scientificoutlookwith that of a man who has outgrown
his tutelage, cut his dependence on his mother, and is preparedto meet the

ElizabethAnderson

69

competitive demands of the public sphere with a clear eye, the ideology of
masculinitytends to confuseseeing the naturalworldas indifferentin the sense
of devoid of teleological lawswith seeing the social worldas hostile in the sense
of full of agents who pursuetheir interests at others' expense (Keller 1992,
116-18). This confusion tempts biologists into thinking that the selfishness
their models ascribeto genes and the ruthlessstrategicrationalitytheir models
ascribeto individualorganisms(mere metaphors,however theoreticallypowerful) are more "real"than the actual care a dog expressestowardher pups.
Such thoughts also reflect the rhetoricof unmaskingbase motivationsbehind
policies that seem to be benevolent, a common if overused tactic in liberal
politics and political theory.The powerof this rhetoricdependson an appearance/realitydistinction that has no place wherethe stakesarecompetingsocial
modelsof biological phenomena, whose merits depend on their metaphorical
ratherthan their referentialpowers.Thus, to the extent that the theoretical
preference for competitive models in biology is underwritten by rhetoric
borrowed from androcentric political ideologies, the preference reflects a
confusion between models and reality as well as an unjustified intrusion of
androcentricpolitical loyalties into the scientific enterprise.
These arenot concerns that can be relieved by deployingthe discovery/justification distinction. To the extent that motivations tied to acquiring a
masculine-coded prestige as a theorist induce mathematical ecologists to
overlook the epistemic defects of models of natural selection that fail to
consider the actual impact of sexual selection, parenting, and cooperative
interactions,they distortthe context of justificationitself. Some of the criteria
of justification, such as simplicity, are also distorted in the light of the
androcentric distinction between public and private values. For example,
simplicity in mathematical biology has been characterizedso as to prefer
explanations of apparentlyfavorablepatterns of group survival in terms of
chance to explanations in termsof interspecificfeedbackloops, if straightforward individualistic mechanisms are not available to explain them (Keller
1992, 153). Finally,to the extent that genderideologiesinformthe context of
discoveryby influencing the direction of inquiryand development of mathematicaltools, they prevent the growthof alternativemodels and the tools that
could make them tractable,and hence they bias our views of what is "simple"
(Keller 1992, 160). The discovery/justificationdistinction, while usefulwhen
consideringthe epistemic relation of a theory to its confirmingor disconfirming evidence, breaksdown once we considerthe relative meritsof alternative
theories. In the latter context, any influence that biases the development of
the field of alternatives will bias the evaluation of theories. A theoretical
approachmay appearbest justifiednot becauseit offersan adequatemodel of
the world but becauseandrocentricideologieshave causedmore thought and
resourcesto be invested in it than in alternatives.

70

Hypatia

So feminist naturalizedepistemologistsshould offer a complex verdict on


gender symbolism in the content of theories. They should leave open the
possibility that gendered models of ungenderedphenomena may be highly
illuminatingand successful,and hence legitimatelyusedin theoreticalinquiry.
The impressiveexplanatorysuccessesof evolutionarytheorydemonstratethis.
At the same time, the ideologicalpower of gendersymbolismsometimesgets
the better of otherwisecarefultheorists.It can generateconceptual confusion
in ways that are hard to detect, and obscuretheoretical possibilitiesthat may
be worth pursuing.The most reliable way to tell when the use of gender
symbolismis generatingsuch cognitive distortionsis to critically investigate
the gender ideology it dependson and the role this ideology plays in society.
In other words,theoristswho use genderedmodels would do well to consider
how feminist theory can help them avoid cognitive distortion. Feminist
naturalized epistemologists therefore should recommend that theorists
attractedto genderedmodelsof ungenderedphenomenaproceedwith caution,
in consultation with feminist theorists. It recommendsan importantchange
in the cognitive authority of disciplines, through its demonstration that
biologistshave something to learn fromfeminist theory after all.
ANDROCENTRISM

A knowledgepractice is androcentricif it reflects an orientation gearedto


specificallyor typicallymale interestsor male lives. Androcentrismcan appear
in a knowledge practice in at least two ways: in the content of theories or
research programsand in the interests that lead inquirers to frame their
researchin certain termsor aroundcertainproblems.Feministsin the natural
and social sciences have advanced feminist epistemology most fully and
persuasivelyby exposing androcentrismin the content of social-scientificand
biological theories.
The content of theories can be androcentricin severalways.A theory may
reflect the view that males, male lives, or "masculinity"set the norm for
humans or animals generally.Fromthis point of view, females, their lives, or
"feminine"characteristicsarerepresentedas problematic,deviationsfromthe
norm, and hence in need of a type of explanation not requiredfor their male
counterparts.Androcentrismof this sort often appearsin the waystheoretical
questionsareframed.Fordecades,psychologicaland biological researchabout
sex differenceshas been framedby the question, "Why are women different
from men?"and the presumedsex difference has cast women in a deviant
position. Researchershave been preoccupiedwith such questions as why girls
are more suggestible,less ambitious,less analyticallyminded, and have lower
self-esteem than boys. Let us leave aside the fact that all these questions are
basedon unfoundedbeliefs aboutsex differences(MaccobyandJacklin 1974).
Why haven't researchersasked why boys are less responsive to others, more

ElizabethAnderson

71

pushy,less syntheticallyminded, and more conceited than girls?The framing


of the problemto be investigatedreflectsnot just a commitment to asymmetrical explanation of men's and women'scharacteristics,but to an evaluation
of women'sdifferencesas dimensionsof inferiority(Tavris1992, chap. 1). It is
thus sexist as well as androcentric.
Another way in which the content of theories can be androcentricis in
describingor defining phenomena from the perspective of men or typically
male lives, withoutpayingattention to how they wouldbe describeddifferently
if examined fromthe point of view of women'slives. Economistsand political
scientists have traditionallydefined class and socioeconomic status from the
point of view of men'slives: a man'sclassor socioeconomic statusis definedin
termsof his own occupation or earnings,whereasa women'sstatus is defined
in termsof her father'sor husband'soccupation or earnings.Such definitions
obscure the differencesin power, prestige, and opportunitiesbetween male
managersand their homemakerwives, and between homemakerwives and
female managers(Stiehm 1983). They also prevent an analysisof the distinctive economic roles and statusof full-time homemakersand of adult independent unmarriedwomen. The distinction between laborand leisure,central to
standardeconomic analyses of the supply of wage labor, also reflects the
perspective of male heads of households (Waring 1990). Classically, the
distinction demarcatesthe publicfromthe privatespheresby contrastingtheir
characteristicactivities ashavingnegativeversuspositive utility,or instrumental versusintrinsic value, or as controlledby others versusfreelyself-directed.
Fromthe standpoint of the lives of women with husbandsor children, these
demarcationsmake no sense. These women are not at leisurewhenever they
are not engaged in paid labor.Professionalwomen often find much of their
unpaid work to constitute a drudgeryfrom which paid labor representsan
escape with positive intrinsic value. Middle-classand working-classwomen
who engage in paid laborand who cannot affordto hire othersto performtheir
householdtasksand child carearebetterrepresentedas engagedin (sometimes
involuntary) dual-careeror double-shift labor than in trading off labor for
leisure.Full-time mothersand homemakersoften view what some considerto
be their leisure activities as highly importantwork in its own right, even if it
is unpaid.
The androcentrismimplicit in the standardeconomic definition of productive labor has profound implications for national income accounting, the
fundamentalconceptual frameworkfor defining and measuringwhat counts
as economically relevant data for macroeconomic theory. It effectively
excludes women's gender-typicalunpaid domestic labor from gross national
product (GNP) calculations, making women's work largely invisible in the
economy. In the advanced industrializednations, economists explain this
omission by arguingthat GNP figuresproperlymeasureonly the economic
value of productionfor marketexchange. In developing nations, where only

72

Hypatia

a modest proportion of productive activity shows up in market exchanges,


economists have long recognized the uselessness of measures of national
production that look only at the market;so they impute a market value to
variousunmarketeddomesticproductionactivitiesassociatedwith subsistence
agriculture,home construction, and the like. But which of these household
activities do economists choose to count as productive?In practice, they have
defined the "productionboundary"in such societies by imposingan obsolete
Westernandrocentricconception of the household.They assumethat households consist of a productiveprimaryproducer,the husband,who supportsa
wife engagedin "housework,"which is assumedto be economically unimportant or unproductive."Housework"has no clear definition in societies where
most production takes place within the household. So economists apply the
concept of "housework"to whatever productive activities a society conventionally assignsto women. Thus, women'sunmarketedlaborin these societies
counts as productive only if men usually perform it too, whereas men's
unmarketedlaboris usuallycounted in the national income statisticsregardless
of its relation to women's labor (Waring 1990, 74-87). The result is that in
Africa, where women do 70 percent of the hoeing and weeding of subsistence
crops,80 percent of crop transportationand storage,and 90 percent of water
and fuel collecting and food processing,these vital activities rarelyappearin
the national income accounts (Waring1990, 84). Here, androcentrismis built
into the very data for economic theorizing, in a such a way that women's
gender-typicalactivities become invisible.
Even when a theory does not go so faras to define the phenomena in a way
that excludes female activities, it may still be androcentricin assumingthat
male activities or predicamentsare the sole or primarysources of important
changesor events. Until recently,primatologistsfocusedalmostexclusivelyon
the behaviorof male primates.They assumedthat male sexual and dominance
behaviorsdeterminedthe basic structureof primatesocial order,and that the
crucial social relationshipsamong troop-dwellingprimates that determined
the reproductivefitnessof individualsand maintainedtrooporganizationwere
between the dominant male and other males. The assumptionfollowed from
a sociobiological argumentthat claimed to show that females of any species
will typically be the "limiting resource"for reproduction:most females will
realizean equal and maximumreproductivepotential, while males will vary
enormouslyin their reproductivefitness. Natural selection, the driving force
of evolutionarychange, would thereforeoperateprimarilyon male characteristics and behavior (Hrdy 1986).
These assumptions were not seriously challenged until women, some
inspiredby the feminist movement, startedentering the field of primatology
in substantial numbers in the mid-1970s. Many studied female-female and
female-infantinteractions,female dominance and cooperativebehavior,and
female sexual activity. By turning their focus from male to female behaviors

ElizabethAnderson

73

and relationships,they found that infant survivalvariedenormously,depending on the behaviorand social statusof the mothers,that troop survivalitself
sometimes depended on the eldest female (who would teach others the
location of distant waterholes that had surviveddroughts),and that femaledirectedsocial and sexual behaviorsplay key roles in maintainingand changing primate social organizations (Hrdy 1981; Haraway 1989). Today the
importanceof female primatesis widely recognizedand studiedby both male
and female primatologists.
What normative implicationsshould be drawnabout the epistemic status
of androcentrictheories?Some feminist epistemologistspropose that theory
can proceed better by viewing the world through the eyes of female agents.
Gynocentrictheorycan be fun. What could be a moreamusingretortto a study
that purportsto explain why women lackself-esteemthan a studythat explains
why men are conceited?It can also be instructive.RichardWrangham(1979)
has proposed a gynocentric model of primate social organizationthat has
achieved widespreadrecognition in primatology.The model assumes the
centralityof female competition for food resources,and predictshow females
will space themselves (singly or in kin-related groups) according to the
distributionof the foods they eat. Males then space themselves so as to gain
optimumaccessto females.The model is gynocentricboth in definingthe core
of primate social groups around female kin-relations rather than around
relations to a dominant male and in taking the situation of females to
constitute the primaryvariablethat accountsforvariationsin male and general
primate social organization.According to the feminist primatologistSarah
Hrdy (1981, 126), Wrangham'smodel offersthe best availableexplanationof
primatesocial organization.
The three androcentric theoretical constructs mentioned correspondto
three differentwaysin which a theorycould be "gender-centric":
in takingone
sex or genderto set the normforboth, in definingcentralconceptswith respect
to the sex- or gender-typicalcharacteristics,behaviors,or perspectivesof males
or females alone, and in taking the behaviors,situation, or characteristicsof
one sex or gender to be causallycentral in determiningparticularoutcomes.
These logical differencesin gender-centrictheorizinghave differentepistemic
implications. As Wrangham'stheory shows, gynocentric causal models can
sometimes be superiorto androcentricmodels. Whether they are superiorin
any particulardomain of interest is an empirical question. It can only be
answeredby comparingrival gender-centricmodels to one another and to
models that do not privilege either male- or female-typical activities or
situationsin their causalaccounts,but ratherfocuson activities and situations
common to both males and females. An importantcontribution of feminist
scholarship in the social sciences and biology has been to show that the
activities and situations of females have been far more causallyimportantin
variousdomains than androcentrictheories have recognized.

74

Hypatia

The other two types of gender-centrismare much more problematicthan


this causal type. A theory that takesone genderto set the norm for both must
bear an explanatory burden not borne by theories that refuse to represent
difference as deviance. It must explain why an asymmetricalexplanation is
requiredfor male- and female-specificcharacteristics.Given the dominant
backgroundassumptionof modernscience that the cosmos does not have its
own telos, it is hardto justifyany claim that one gendernaturallysets the norm
for both. Claims about norms must be located in human value judgments,
which is to say that the only justificationfornormativegender-centrismwould
have to lie in a substantive sexist moral or political theory. As we shall see
below,empiricismdoes not rule out the use of value judgmentsas background
assumptionsin scientific theories. Nevertheless, this analysis of normative
gender-centrismsuggestswhy feminists should not be satisfied with a tableturning, "why men are so conceited" type of gynocentric theorizing.Posing
such questions may expose the androcentrismof standardways of framing
researchproblemsin sex-differencesresearchto healthy ridicule. But because
feminists are interested in upholding the equality of all persons, not the
dominationof women over men, they have no interestin claimingthat women
set the norm for humansgenerally.
Theories that tailor concepts to the activities or positions specific to or
typical of one gender only and then apply them to everyone are straightforwardlyempiricallyinadequate.As the case of androcentricdefinitions of class
showed, they obscure actual empiricaldifferencesbetween men and women
and between differently situated women. As the case of the labor/leisure
distinction showed, they overgeneralizefrom the typical situation of one
genderto that of both. When conceptuallyandrocentrictheoriesguidepublic
policy, the resultingpolicies are usuallysexist, since theories cannot respond
to phenomena they make invisible. Thus, when GNP statistics fail to count
women's labor as productive, and public policies aim to increase GNP, they
may do so in ways that fail to improve the well-being of women and their
families and may even reduce it. In Malawiand Lesotho, where women grow
most of the food fordomestic consumption,foreignaid projectshave provided
agriculturaltraining to the men who have no use for it, and offered only
home economics education to women (Waring 1990, 232, 234). In the
Sahel, a USAID drought-reliefproject forced women into economic dependency on men by replacing only men's cattle herds, on the androcentric
assumption that women did not engage in economically significant labor
(Waring 1990, 176-77).
Feminist naturalizedepistemologiststhereforepass differentjudgmentson
differentkinds of gender-centrismin theoretical inquiry.Conceptual gendercentrism is plainly inadequate in any society with overlappinggender roles,
because it leads to overgeneralizationand obscuresthe differencesbetween
empiricallydistinct phenomena. It could work only in societies where men

ElizabethAnderson

75

and women inhabit completely and rigidly segregatedspheres, and only for
concepts that apply exclusively to one or the other gender in such a society.
Normative gender-centrismeither dependson a problematiccosmic teleology
or on sexist values. This does not automaticallymake it epistemicallyinadequate,but it does requirethe assumptionof an explanatoryburden(why men's
and women's traits do not receive symmetricalexplanatorytreatment) that
non-gender-centrictheories need not assume.In addition, its dependence on
sexist values give theoristswho repudiatesexism sufficientreason to conduct
inquirythat is not normativelygender-centric.Finally,causalgender-centrism
may or may not be empiricallyjustified.Some events do turn asymmetrically
on what men or women do, or on how men or women are situated.
The chief trap in causal gender-centrismis the temptation to reify the
domain of events that are said to turn asymmetricallyon the actions or
characteristicsof one or the other gender.The selection of a domainof inquiry
is alwaysa function of the interestsof the inquirer.2Failureto recognizethis
maylead androcentrictheoriststo constructtheir domainof studyin waysthat
confine it to just those phenomena that turn asymmetricallyon men'sactivities. They may thereforedeclare as an objective fact that, say, women have
little causal impact on the "economy,"when all that is going on is that they
have not taken any interest in women'sproductiveactivities, and so have not
categorizedthose activities as "economic."Feminist naturalizedepistemologistscaution againstthe view that domainsof inquirydemarcatenaturalkinds.
FollowingQuine, they questionsupposedconceptualbarriersbetween natural
and social science, analyticand synthetic knowledge,personaland impersonal
knowledge,fact and value (Nelson 1990, chap. 3). Their empiricistcommitments enable them to uncover surprisingconnections among apparently
distantpoints in the web of belief. If naturalizedepistemologistsuse space-age
technology to explore the universeof knowledge,feminist naturalizedepistemologists could be said to specialize in the discovery of wormholes in that
universe. Gender and science are not light-years apart after all; subspace
distortionsin our cognitive apparatuspermitsurprisinglyrapidtransportfrom
one to the other, but feminist navigatorsare needed to ensure that we know
the route we are travellingand have reasonto take it.
SEXISMIN SCIENTIFIC
THEORIES

One frequentlytraveledroute between genderand science employsnormative assumptionsaboutthe properrelationsbetween men and women, or about
the respective characteristicsand interestsof men and women, in the content
or application of scientific theories. When a theory assertsthat women are
inferiorto men, properlysubordinatedto men, or properlyconfined to genderstereotypedroles, or when it judgesor describeswomen accordingto sexist or
doublestandards,the content of the theoryis sexist. When people employsuch

76

Hypatia

assumptions in applying theories, the application of the theory is sexist.


Naturalizedfeminist epistemologyconsidershow our evaluations of theories
should change once their sexism is broughtto light.
The application of theories can be sexist in direct or indirect ways.
Theories may be used to provide direct ideological justification for patriarchal structures.Steven Goldberg (1973) uses his theory of sex differences
in aggression to justify a gendered division of labor that deliberately
confines women to low-prestige occupations. More usually,the application
of theories is indirectly sexist in taking certain sexist values for granted
ratherthan trying to justify them. Forexample, researchon oral contraceptives for men and women uses a double standardfor evaluating the acceptability of side effects. Oral contraceptives for men are disqualified if they
reduce libido, but oral contraceptives for women are not rejected for
reducing women's sexual desire.
In a standardpositivist analysis,neither form of sexism in the application
of theories has any bearingon the epistemic value of the theories in question.
That a theory is used to supportunpopularpolitical programsdoes not show
that the theory is false. At most, it reflects a failureof the proponentsof the
programto respect the logical gap between fact and value. But opponents of
the programfail to respect this gap in attacking a theory for the uses to
which it is put. According to this view, theories supplyfacts that all persons
must accept, regardless of their political commitments. That a theory is
indirectly applied in sexist ways provides even less ground for attacking its
content. The question of truth must be strictly separated from the uses to
which such truths are put.
Naturalized epistemology does not support such a sanguine analysis of
theoriesthat areappliedin sexist ways."Successful"technologicalapplications
of theories are currentlytaken to provide evidence of their epistemic merits.
If knowledge is power, then power is a criterion of adequateunderstanding.
The prevailinginterpretationof this criteriondoes not considerwhose power
is enhanced by the theory and whose interestsare servedby it. Feministsurge
that these considerationsbe taken explicitly into account when one evaluates
whethertechnological applicationsof theoriessupplyevidence of an adequate
understandingof the phenomena they control (Tiles 1987). It maybe truethat
certain drugswould be effective in controlling the phenomena of women's
hormonal cycles that are currentlydesignatedas pathologies constitutive of
premenstrualsyndrome.Such control may come at the expense of women's
interests, not just because of undesirableside effects but also because the
legitimation of drugtreatmentreinforcesthe medicalizationof women'scomplaints, as if these complaintswere symptomsto be medicatedratherthan as
claimson othersto change their behavior(Zita 1989). Doctorsmaybe satisfied
that such a "successful"drug treatment of PMS supplies evidence that the
theory it applies provides them with an adequateunderstandingof women's

ElizabethAnderson

77

menstrual cycles. But should women be satisfied with this understanding?


Suppose the phenomena associatedwith PMS could also be eliminated, or
revalued,by widespreadacceptanceof feministconceptions of women'sbodies
or by egalitarianchanges that would make social arrangementsless frustrating
to women. (This would be possible if women'ssymptomsof distressin PMS
were partly caused by misogynistsocial expectations that representwomen's
menstrualcycles as pathological.) Such a successful"technological"application of feminist theory would providewomen with an understandingof their
own menstrualcycles that would empowerthem. Where the sexist medical
technology would enable women to adapt their bodies to the demands of a
sexist society, the feminist technology would empower women to change
society so that their bodies were no longer considered "diseased."Thus,
applications of theories may influence the content of theories whenever
"success"in application is taken to justify the theory in question. Sexist or
feministvalues mayinformcriteriaof successin application,which mayin turn
informcompeting criteriaof adequateunderstanding.The epistemic evaluation of theories thereforecannot be sharplyseparatedfrom the intereststheir
applicationsserve.
Feminist naturalizedepistemologyalso rejects the positivist view that the
epistemic merits of theories can be assessed independently of their direct
ideologicalapplications(Longino 1990;Antony 1993;Potter 1993). Although
any acceptableideologymust makesurethat it does not fly in the face of facts,
theories do not merely state facts but organizethem into systemsthat tell us
what their significance is. Theories logically go beyond the facts; they are
"underdetermined"by all the empirical evidence that is or ever could be
adduced in their favor (Quine 1960, 22). The evidential link between an
observedfact and a theoreticalhypothesiscan only be securedby background
auxiliaryhypotheses.This leaves open the logical possibilitythat ideological
judgmentsmay not be implicationsof an independentlysupportedtheorybut
figure in the justification of the theory itself, by supplying evidential links
between empiricalobservationsand hypotheses.
A particularlytransparentexample of this phenomenon may be found in
theories about sex differencesin intelligence. Girls scoredsignificantlyhigher
than boys on the firstStanford-BinetIQ tests developed by LewisTerman.To
correctforthis "embarrassment,"
Termaneliminatedportionsof the test where
scored
than
and
inserted questions on which boys scored
girls
higher
boys
than
The
substitution
was considered necessary to ensure the
higher
girls.
of
the
test
school
validity
against
grades, the only available independent
measuresof children'sintelligence, which did not differby gender.ButTerman
did not adjusthis test to eliminate sex differenceson subtestsof the IQ, such
as those about quantitative reasoning. These differences seemed
unproblematicbecausethey conformedto prevailingideological assumptions
about appropriategenderroles (Mensh and Mensh 1991, 68-69). Today,that

78

Hypatia

IQ scoresare good predictorsof a child'sschool gradesis still taken to provide


key evidence for the claim that differencesin IQ scoresmeasuredifferencesin
children'sinnate intelligence. But the evidential link tying school gradesto
this theoreticalclaim dependson the backgroundvalue judgmentthat schools
providefair educationalopportunitiesto all children with respect to all fields
of study.Those schools that discouragegirls frompursuingmath and science
assume that girls have inferior quantitative reasoning ability; they do not
recognizethat lack of encouragementcan cause relatively lower performance
on math tests.
From a positivist point of view, this reasoning is defective on two counts.
First,it is circularto claim that IQ tests demonstrateinnate sex differencesin
quantitative reasoningability when the assumptionof innate sex differences
is built into the backgroundhypothesesneeded to validate the tests. Second,
no reasoningis scientificallysoundthat incorporatesvalue judgmentsinto the
backgroundassumptionsthat link observations to theory. The salience of
positivist views of science as well as their usefulnessto feminists in criticizing
researchaboutsex differenceshas temptedsome feministsto use the positivist
requirementthat science be value-freeto discreditall scientific projectsthat
incorporatesexist values in the explicit or implicit content of their theories.
But this appropriationof positivism puts at odds the two aims of feminist
epistemology-to criticize sexist science and to promote feminist science. If
incorporatingsexist values into scientific theories is illegitimate on positivist
grounds, then so is incorporating feminist values into scientific theories
(Longino 1993a, 259).
Feminist naturalizedepistemologistsoffer a more nuanced responseto the
presence of value judgmentsin scientific inference. Even "goodscience" can
incorporatesuch value judgments.The logical gap between theory and observation ensuresthat one cannot in principlerule out the possibilitythat value
judgments are implicit in the backgroundassumptionsused to argue that a
given observationconstitutesevidence fora given hypothesis(Longino 1990).
Fromthe perspectiveof an individualscientist, it is not unreasonableto use
any of one's firm beliefs, including beliefs about values, to reason from an
observation to a theory. Nor does the prospect of circularitythreaten the
scientific validity of one's reasoning,as long as the circle of reasoning is big
enough. In a coherent web of belief, every belief offerssome supportfor every
other belief, and no belief is perfectlyself-supporting.Theories that incorporatevalue judgmentscan be scientificallysound as long as they areempirically
adequate.
This reasoning underwritesthe legitimacy of feminist scientific research,
which incorporates feminist values into its theories. Such values may be
detected in the commitment of feminist researchersto regard women as
intelligent agents, capable of reflectingon and changing the conditions that
presentlyconstrain their actions. This commitment tends to supporta theo-

ElizabethAnderson

79

reticalpreferencefor causalmodelsof femalebehaviorthat highlight feedback


loops between their intentional states and their social and physical environments, and that resist purely structuralistaccounts of female "nature"that
leave no room for females to resist their circumstancesor maneuver among
alternate possibilities (Longino 1989, 210-13; Haraway 1989, chap. 13). In
contrast,most behavioristand some sociobiologicaltheoriesfavormodels that
highlight linear causal chains from fixed physiologicalor physical conditions
to determinate behaviors, and that emphasizethe structuralconstraints on
action. The epistemic values of simplicity,prediction,and control might seem
to supportlinear, structuralcausal models. But we have seen that control at
least is a contested value; the kinds of control taken to warrantclaims of
adequate understandingdepend on substantive value judgments about the
importanceof particularhuman interests.Is adequateunderstandingachieved
when a theory empowers scientists to control women's lives, or when it
empowerswomen to control their own lives?Rival interpretationsof the other
epistemic values also depend on contested nonepistemic values. The kind of
simplicity one favors depends on one's aesthetic values. In any event, other
epistemic values, such as fruitfulness,appear to favor complex, nonlinear
causal models of human behavior. Such models support experiments that
generate novel behaviors disruptive of presumedstructuralconstraints on
action.
Naturalized feminist epistemology thus permits scientific projects that
incorporatefeminist values into the content and application of theories. It
does not provide methodological arguments against the pursuit of sexist
theories.It does claim, however,that it is irrationalfortheoriststo pursuesexist
researchprogramsif they do not endorse sexist values. Moral and political
argumentsabout the rationality of particularvalues may therefore have a
bearingon the rationalityof pursuingparticularresearchprograms.In addition,
the objectivity of science demands that the background assumptions of
researchprogramsbe exposed to criticism.A scientific communitycomposed
of inquirerswho sharethe samebackgroundassumptionsis unlikelyto be aware
of the roles these assumptions play in licensing inferences from observations to hypotheses, and even less likely to examine these assumptions
critically. Naturalized epistemology therefore recommends that the scientific community include a diversity of inquirerswho accept different background assumptions. A community of inquirers who largely accept sexist
values and incorporate them into their background assumptions could
enhance the objectivity of the community's practice by expanding its
membership to include researcherswith feminist commitments (Longino
1993a, 267-269).

80

Hypatia

OFNATURALIZED
THELOCAL
CHARACTER
FEMINIST
EPISTEMOLOGY
In readingthe projectof feministepistemologyalong naturalized,empiricist
lines, I have tried to show how its interest and critical power do not depend
on the global, transcendentalclaims that all knowledge is gendered or that
rationality as a regulatoryepistemic ideal is masculine. Naturalizedfeminist
epistemologistsmay travel to distant locations in the universe of belief, but
they alwaysremain inside that universe and travel from gender to science by
way of discrete, empiricallydiscoveredpaths. They have an interest in constructing new paths to empirically adequate, fruitful, and useful forms of
feminist science and in breakingup other paths that lead to cognitively and
sociallyunsatisfactorydestinations.All the paths by which naturalizedepistemologistsfind genderto influencetheoreticalknowledgearelocal, contingent,
and empiricallyconditioned. All the paths by which they proposeto change
these influences accept rationality as a key epistemic ideal and empirical
adequacyas a fundamentalgoal of acceptabletheories.This ideal and this goal
are in principleequallyopen to pursuitby male and female inquirers,but may
be best realizedby mixed-genderresearchcommunities.Naturalizedepistemologistsfind no persuasiveevidence that indicatesthat all women inquirersbring
some sharedglobal feminine differencein ways of thinking to all subjectsof
studynor that such a feminine differencegives us privilegedaccess to the way
the world is.
In rejectingglobal,transcendentalclaimsaboutdifferencesin the waysmen
and women think, naturalizedfeminist epistemologistsdo not imply that the
entry and advancement of significant numbers of women into scientific
communitiesmakesno systematicdifferenceto the knowledge these communities produce.But, following their view of inquiryas a social, not an individual, enterprise, they credit the improvements in knowledge such entry
producesto the greaterdiversityand equalityof membershipin the scientific
community rather than to any purportedlyprivileged subject position of
women as knowers (Tuana 1992; Longino 1993a). Men and women do have
somegender-specificexperiencesandpersonalknowledgedue to their different
socializationand social status.We have seen that such experiences and forms
of knowledge can be fruitfullybroughtto bear upon theoretical inquiry.So it
should not be surprisingthat women researchershave exposed and criticized
androcentrismin theories much more than men have. The diversity and
equalityof inquirershelp ensurethat social models do not merelyreflect or fit
the circumstancesof a narrowdemographicsegment of the population when
they are meant to applyto everyone.They correcta cognitive bias commonly
found among inquirersbelonging to all demographicgroups,located in the
habit of assumingthat the waythe worldappearsto oneself is the way it appears
to everyone.

ElizabethAnderson

81

This surveyof some findings of naturalizedfeminist epistemologyhas also


identifiedimprovementsin knowledgethat have or wouldcome aboutthrough
the entry of feministtheoristsinto variousfields, and throughrevisions in the
systemsof cognitive authorityamong fields that would bring the findings of
feminist theoriststo bearupon apparentlydistantsubjects.3We have seen that
the use of gender symbolismto model nonhuman phenomena is fraughtwith
cognitive traps.So it should not be surprisingthat feminist researchers,who
make it their business to study the contradictions and incoherences in our
conceptions of gender,can improve theories by exposing and clearingup the
confusions they inherit from the gender ideologies they use as models. By
pursuingfeminist research in the humanities, social sciences, and biology,
feminist researchersalso pose challenges to prevailing theories. Here again,
the kindsof changeswe shouldexpect in theoreticalknowledgefromthe entry
of feminist researchersinto various fields do not typically consist in the
production of specifically feminist ontologies, methodologies, standpoints,
paradigms,or doctrines.Feministcontributionsto theorizingaremoreusefully
conceived as altering the field of theoretical possibilities (Haraway1986, 81,
96). Research informed by feminist commitments makes new explanatory
models available, reframesold questions, exposes facts that undermine the
plausibility of previously dominant theories, improves data-gatheringtechniques, and shifts the relationsof cognitive authorityamong fields and theories. In these and many other ways, it reconfiguresour assessmentsof the
prospects and virtues of various research programs.Without claiming that
women, or feminists, have a globally differentor privilegedway of knowing,
naturalizedfeminist epistemologyexplains how feminist theory can productively transformthe field of theoreticalknowledge.

NOTES
I wish to thank Ann Cudd,Sally Haslanger,Don Herzog,David Hills, PeterRailton,
Justin Schwartz,MiriamSolomon, and the faculties at the Law Schools of Columbia
University,the University of Chicago, and NorthwesternUniversity for helpful com-

mentsandcriticisms.

1. Diamond (1991, 1009) writesthat the exclusion of animal trainers'knowledge


fromthe realmof authoritativeknowledge"cannotin any verysimplewaybe connected
to gender."Pointingout that the terms"hard"and"soft"as appliedto formsof knowledge
are used by "men tryingto put down other men,"Sherifarguesthat for this reasonit is
misleading"to inferthat these termssymbolize"masculine"and "feminine"
"particularly
(1987, 46-47). I wouldhave thoughtthat her observationsupportsthe genderedreading,
since a standardwayformen to putdownothermen is to insinuatethat they arefeminine.
2. The interestsat stakeneed not be self-interestsor even ideologicalinterestsof a
broadersort. One might just be curiousabouthow rainbowsform,without seeking this

82

Hypatia

knowledgefor the sake of finding out how to get the proverbialpot of gold at the end.
Curiosityis one kind of interestwe can expressin a phenomenon.
3. The questionof the impactof feministtheoristson knowledgeis distinctfrombut
relatedto the questionof the impactof women theoristson knowledge.Not all women
theoristsarefeminists,andsomefeministtheoristsaremen.At the sametime, therecould
be no genuinefeministtheorythat wasconductedby men alone. Feministtheoryis theory
committed to the liberationand equalityof women. These goals can only be achieved
throughthe exerciseof women'sown agency,especiallyin definingand coming to know
themselves.Feministtheoryis one of the vehicles of women'sagency in pursuitof these
goals, and thereforecannot realizeits aims if it is not conductedby women. So it should
not be surprisingthat most of the transformationsof knowledge induced by feminist
theorywere broughtaboutby women.

REFERENCES
New York:
Alcoff, Linda, and Elizabeth Potter, eds. 1993. Feministepistemologies.
Routledge.
American Association of University Women. 1992. The AAUW report:How schools
shortchange
girls.Preparedby WellesleyCollege Center for Researchon Women.
Anderson,Elizabeth.1993.Valuein ethicsandeconomics.Cambridge:HarvardUniversity
Press.
Antony, Louise.1993. Quine as feminist:The radicalimportof naturalizedepistemology.
In A mindof one'sown:Feministessaysonreasonandobjectivity.
See Antony andWitt
1993.
Antony, Louise,and CharlotteWitt. 1993. A mindof one'sown:Feministessayson reason
andobjectivity.
Boulder,CO: Westview.
Astin, Helen S., and Alan E. Beyer.1973. Sex discriminationin academe.In Academic
womenon themove, ed. Alice. S. Rossi and Ann Calderwood.New York:Russell
Sage Foundation.
Becker,JoanneRossi. 1981. Differentialtreatmentof femalesand males in mathematics
classes.Journalfor Researchin Mathematics
Education12(1): 40-53.
Belenky,Mary,Blythe Clinchy,Nancy Goldberger,and Jill Tarule.1986. Women'sways
of knowing.New York:BasicBooks.
Bell, Diane, Pat Caplan,and WazirKarim,eds. 1993. Genderedfields:Women,men, and
New York:Routledge.
ethnography.
Gender
and
Biology
StudyGroup.1988.The importanceof feministcritiqueforcontemporarycell biology.Hypatia3(1): 61-76.
and culture.Albany:
Bordo,Susan. 1987. The flightto objectivity:
Essayson cartesianism
State Universityof New YorkPress.
Bourdieu,Pierre.1977. Outlineof a theoryof practice.Cambridge:CambridgeUniversity
Press.
Coady,C. A. J. 1992. Testimony.Oxford:ClarendonPress.
andthepolitics
Collins,PatriciaHill. 1990.Blackfeminist
consciousness,
thought:
Knowledge,
Boston:Unwin Hyman.
of empowerment.
Curran,Libby.1980. Science education:Did she drop out or was she pushed?In Alice
ed. BrightonWomenand Science Group.London:Virago.
throughthemicroscope,
Dawkins,Richard.1976. The selfishgene.New York:OxfordUniversityPress.

Elizabeth Anderson

83

Diamond, Cora. 1991. Knowing tornadoesand other things. New LiteraryHistory22:


1001-15.
Totowa,NJ:Rowmanand Littlefield.
Duran,Jane. 1991. Towarda feministepistemology.
Fidell, L. S. 1970. Empiricalverification of sex discriminationin hiring practices in
25(12): 1094-98.
psychology.AmericanPsychologist
Flax, Jane. 1983. Political philosophyand the patriarchalunconscious.In Discovering
reality.See Hardingand Hintikka 1983.
Fox,MaryFrank.1981.Sex segregationandsalarystructurein academia.Sociologyof Work
andOccupations8(1): 39-60.
Gilligan, Carol. 1982. In a differentvoice.Cambridge:HarvardUniversityPress.
New York:Morrow.
of patriarchy.
Goldberg,Steven. 1973. The inevitability
to
Haraway,Donna. 1986. Primatologyis politicsby other means.In Feministapproaches
science,ed. Ruth Bleier.New York:Pergamon.
.1989. Primatevisions.New York:Routledge.
Harding,Sandra.1986. Thesciencequestioninfeminism.Ithaca:CornellUniversityPress.
Harding, Sandra, and Merrill B. Hintikka, eds. 1983. DiscoveringReality.Dodrecht,
Holland:D. Reidel.
Heare, Vicki. 1982. Adam'stask.New York:Vintage.
Hrdy,Sarah. 1981. The womanthatneverevolved.Cambridge:HarvardUniversityPress.
. 1986.Empathy,polyandry,andthe mythof the coy female.In Feministapproaches
to science,ed. Ruth Bleier.New York:Pergamon.
Hubbard,Ruth. 1990. The politicsof women'sbiology.New Brunswick,NJ: Rutgers
UniversityPress.
New York:Freeman.
Keller,EvelynFox. 1983. A feelingfor theorganism.
. 1985. The forceof the pacemakerconcept in theoriesof aggregationin cellular
on genderandscience.New Haven:YaleUniversityPress.
slime mold. In Reflections
. 1992. Secretsof life, secretsof death.New York:Routledge.
Leacock,Eleanor.1982. Mythsof maledominance.New York:Monthly Review Press.
Lefkowitz,M. R. 1979. Educationfor women in a man's world. Chronicleof Higher
Education,6 August,p. 56.
andreality,
Longino,Helen. 1989. Can therebe a feministscience?In Women,knowledge,
ed. Ann Garryand MarilynPearsall.Boston:Unwin Hyman.
.1990. Scienceas socialknowledge.
Princeton,NJ: PrincetonUniversityPress.
.1993a. Essentialtensions-Phase two:Feminist,philosophical,andsocialstudies
of science. In A mindof one'sown.See Antony and Witt 1993.
.1993b. Subjects,power,andknowledge:Descriptionandprescriptionin feminist
See Alcoff and Potter 1993.
philosophiesof science. In Feministepistemologies.
Maccoby,Eleanor,and Carol Jacklin. 1974. The psychologyof sex differences.Stanford:
StanfordUniversityPress.
Carbondale:SouthernIllinois
Mensh,Elaine,and HarryMensh. 1991.TheIQ mythology.
UniversityPress.
in scienceandengineering.
National Science Foundation.1984. Womenandminorities
Nelson, Lynn.1990.Whoknows?FromQuinetoafeministempiricism.
Philadelphia:Temple
UniversityPress.
. 1993. Epistemologicalcommunities.In Feministepistemologies.
See Alcoff and
Potter 1993.
Pai, HyungIl. 1985. (Anthropologist,Universityof California,Santa Barbara).Personal
communication.

84

Hypatia

Paludi,Michele Antoinette, and WilliamD. Bauer.1983. Goldbergrevisited:What'sin


an author'sname. Sex Roles9(3): 287-390.
Chicago:Universityof Chicago Press.
Polanyi,Michael. 1958. Personalknowledge.
See
Potter,Elizabeth.1993. Genderand epistemicnegotiation.In Feministepistemologies.
Alcoff and Potter 1993.
Quine, W. V. 0. 1960. Wordandobject.Cambridge:MIT Press.
. 1969. Epistemologynaturalized.In Ontologicalrelativityand otheressays.New
York:ColumbiaUniversityPress.
Reskin,Barbara.1979. Academic sponsorshipand scientists'careers.Sociologyof Education52(3): 129-46.
Rooney, Phyllis. 1991. Gendered reason: Sex metaphor and conceptions of reason.
Hypatia6(2): 77-103.
Rose, Hilary. 1987. Hand, brain, and heart: A feminist epistemologyfor the natural
sciences. In Sex andscientificinquiry,ed. SandraHardingand JeanO'Barr.Chicago:
Universityof Chicago Press.
Rose, Suzanna. 1989. Women biologistsand the "old boy" network. Women'sStudies
International
Forum12(3): 349-54.
Rosenfeld, Rachel. 1981. Academic career mobility for psychologists.In Womenin
scientificand engineeringprofessions,ed. Violet Haas and Carolyn Perrucci.Ann
Arbor:Universityof MichiganPress.
Rossiter,Margaret.1982. Womenscientistsin America:Strugglesand strategiesto 1940.
Baltimore:Johns HopkinsUniversityPress.
Scheman, Naomi. 1983. Individualismand the objects of psychology.In Discovering
reality.See Hardingand Hintikka1983.
Schiebinger,Londa. 1989. Themindhasno sex?Cambridge:HarvardUniversityPress.
Schuman,Howard,and ShirleyHatchett. 1974.Blackracialattitudes:Trendsandcomplexities.Ann Arbor:Universityof MichiganPress.
Sherif, Carolyn. 1987. Bias in psychology.In Feminismand methodology,ed. Sandra
Harding.Bloomington:IndianaUniversityPress.
Smith, Dorothy.1974.Women'sperspectiveas a radicalcritiqueof sociology.Sociological
inquiry44(1): 7-13.
Solomon, Miriam.1994. Social epistemology.Nous 28: 325-343.
Stiehm, Judith. 1983. Our aristotelianhangover.In Discovering
reality.See Hardingand
Hintikka 1983.
Tavris,Carol. 1992. Themismeasure
of woman.New York:Simon and Schuster.
Tidball,M. Elizabeth.1980.Women'scollegesand womenachieversrevisited.Signs5(3):
504-17.
Tiles, Mary.1987. A science of Marsor of Venus?Philosophy
62(July):293-306.
Tuana,Nancy. 1992. The radicalfutureof feministempiricism.Hypatia7(1): 100-14.
Vetter, Betty. 1981. Changing patternsof recruitmentand employment.In Womenin
scientificand engineeringprofessions,ed. Violet Haas and Carolyn Perrucci.Ann
Arbor:Universityof MichiganPress.
Waring,Marilyn.1990. If womencounted.San Francisco:HarperCollins.
Wrangham,Richard.1979. On the evolutionof apesocialsystems.BiologyandSocialLife:
SocialSciencesInformation18:335-68.
Zita, Jacquelyn.1989. The premenstrualsyndrome:"Dis-easing"the female cycle. In
Feminismandscience,ed. Nancy Tuana.Bloomington:IndianaUniversityPress.