Gender, Politics, and the Theoretical Virtues

Author(s): Helen E. Longino

Source: Synthese, Vol. 104, No. 3, Feminism and Science (Sep., 1995), pp. 383-397
Published by: Springer
Stable URL: http://www.jstor.org/stable/20117439 .
Accessed: 11/01/2011 13:57

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=springer. .

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.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Synthese.

http://www.jstor.org
 HELENE. LONGINO

GENDER, POLITICS, AND THE THEORETICAL VIRTUES*

ABSTRACT. Traits like simplicity and explanatory power have traditionally been treated
as values internal to the sciences, constitutive rather than contextual. As such they are
cognitive virtues. This essay contrasts a traditional set of such virtues with a set of alternative
virtues drawn from feminist writings about the sciences. In certain theoretical contexts, the

only reasons for preferring a traditional or an alternative virtue are socio-political. This
undermines the notion that the traditional virtues can be considered purely cognitive.

In the 1970's Thomas Kuhn responded to critics of The Structure of Sci

entific Revolutions who claimed that Kuhn's analysis made theory choice
entirely subjective. In his essay 'Objectivity, Values, and Theory Choice',
Kuhn discussed five values that scientists use to guide their judgments
in choosing between competing theories. These are accuracy, simplicity,
internal and external consistency, breadth of scope, and fruitfulness. Kuhn
had a lot to say about these values and how they functioned; his overall
claim was that they constituted objective grounds for theory choice. My
point inmentioning them here is that most of the elements on the list (with
the possible exception of fruitfulness) are the sorts of considerations that
end up in philosophers' lists of what, besides agreement with observational
and experimental data, counts for the truth (or acceptability) of a theory
or hypothesis.1 Indeed, when drawing a distinction between what I called
constitutive and contextual I used items like empirical
values, adeguacy,
simplicity, explanatory power (a cognate of breadth of scope) as paradig
matic examples of constitutive values. In this I was simply following in
the footsteps of other philosophers of science.
The items on Kuhn's list have qualified as constitutive values because
they have been considered characteristics enhancing the likelihood of the
truth of a theory or hypothesis. My point in Science as Social Knowledge
(Longino 1990) was to complicate
the distinction, mainly by arguing that
-
what we values - social or practical
identified as contextual interests
could and did function as constitutive values. That is, social or practical
interests function as do so-called cognitive values in determining what

Synthese 104: 383-397,1995.

? 1995 Kluwer Academic Publishers. Printed in the Netherlands.
384 HELEN E. LONGINO

counts as good or acceptable scientific judgment. Which interests do so

and how depends on other features of the overall context. I have been

rethinking this distinction and am no longer convinced that what I was

treating as paradigmatic constitutive values have a solely epistemic or

cognitive basis. I hope in what follows to shake your confidence (if you
have any), too.
My strategy will be crablike: moving sideways rather than forward. I
want first to examine an alternative set of values to the traditional ones.
I will then use that examination as a lens through which to reexamine
elements of the traditional set. To place this in a more general philosophical
context, I want first briefly to summarize relevant bits of the account of
scientific inquiry that forms the background to my thinking.
-
I've argued for a view I call contextual empiricism while experience
(experiment, observation) constitutes the least defeasible legitimator of
knowledge claims in the sciences, the evidential relevance of particular
elements of experience to hypotheses ismediated by background assump
tions operating at many levels. What controls the role of background
assumptions is interaction among scientists, interaction consisting in criti
cism of assumptions involved in observation, of assumptions involved in

reasoning, of assumptions involved in thinking a given hypothesis plau

sible, of assumptions involved in the application of particular methods to
the solution of particular problems. To be successful in uncovering such
assumptions, criticism must proceed from a variety of points of view,
ideally as many as are available.
This account, I maintain, has at least two consequences. 1) It allows
us to see that the same process accounts for both the suppression and the
expression of social values, interests and ideology in the sciences. Idiosyn
cratic values are suppressed, while values held by all members are invisible
(as values, interests, or ideology). These are, therefore, not available for
control by discursive interactions and are free to be expressed in the con
tent of theories by those members.
accepted 2) It identifies the producer of
scientificknowledge, the knower, as the community rather than the indi
vidual scientist. This means that certain features of community structure
are important to the knowledge productive capacity of a community. I've
discussed four such features. There must be:

(a) avenues for the expression and dissemination of criticism;

(b) uptake of, or response to, criticism;

(c) public standards by reference to which theories, etc. are

assessed;

(d) equality of intellectual authority.

 GENDER, POLITICS,AND THE THEORETICALVIRTUES 385

This last condition is intended to correct for the fact that imbalances in
the social composition of the scientific community mean that certain sets
of values will escape criticism. The extension of equality of intellectual
authority to all qualified participants is intended to require representative
diversity in the community. I originally thought of this in terms of repre
senting the diversity of substantive assumptions that could play a role in
evidential reasoning, and would hence be among the public standards men
tioned in the third condition. But, as I said, I want now to raise questions
about the more usual candidates for cognitive standards or values.

ii

In claiming that public standards are required for a knowledge productive

community, I am not claiming that there is a single set of standards that
characterizes all scientific communities. I'm claiming instead that there is
a pool of standards - formal, substantive, and practical
- that such com

munities draw upon in regulating themselves. Criticism and endorsement

of theories and explanatory models, as well as the profferring of alterna
tives, are made germane to a given community by appeal to some one or
more of the standards it recognizes. Satisfaction of a standard is a prima
facie reasonfor accepting a model or theory. As prima facie, its probative
value, of course, can be overridden by failure to satisfy another standard
assigned greater weight in that context. Different, but overlapping, sets
from this pool characterize different communities. Sets of standards of
different communities are related by family resemblance, one might say,
rather than by identity. Kuhn's values constitute one (or part of one) such
set. I want to think about a different
set, drawing on work that has been
done by feminist scientists and feminist historians and philosophers of
science. Here one finds empirical adequacy (a.k.a. accuracy), but also nov

elty, ontological heterogeneity, complexity of interaction, applicability to

human needs, diffusion or decentralization of power. There are undoubt
edly others, but (as Kuhn said about his list) this list is enough to make the
points I want to make.
The are generally
traits listed invoked singly or in groups of two or
three and for the most part become evident as values in the context of their
use.2 Like the elements in Kuhn's list, they function as virtues, qualities
of a theory, hypothesis, or model that are regarded as desirable and hence

guide judgments between alternatives. I shall refer to them as virtues, val

ues, standards, criteria, ignoring the differences between those concepts
for purposes of this discussion. One might say in defense of the Kuhnian
set that they conduce to truth. Kuhn didn't say so, but less cautious advo
386 HELEN E. LONGINO

cates have. But if an alternative set offers grounds for accepting theories
or models that do just as well as those validated by traditional standards at
organizing and generating explanations of the phenomena, then this argu
ment is shown to be hollow. Let me begin by offering some interpretation
of the elements of this alternative set based on the contexts inwhich they've
been deployed. Then I shall offer some reflections on their status.
1) Empirical adequacy. Empirical adequacy generally means agree
ment of the observational claims of a theory or model with observational
and experimental data, present, retrospective, or predictive. A good deal
of feminist effort has gone into discrediting research programs that pur
port to show a biological etiology for differences ascribed on the basis
of sex. The (feminist) scientists involved in this effort - scientists such
as Ruth Bleier, Anne Fausto-Sterling, -
Richard Lewontin, Ruth Doell
have concentrated on showing that such research fails minimal standards
of empirical adequacy, either through faulty research design or improper
statistical methodology. I take their appeal to empirical adequacy in the
context of their critiques to constitute an implicit endorsement of the stan
dard. Empirical adequacy is valued for, among other things, its power when
guiding inquiry to reveal both gender and gender bias. It is, of course, a
standard shared with race and class sensitive research communities as well
as with most mainstream communities. Failure to meet the standard in a
strong sense, i.e. the generation of statements about what will or has been
observed that are incompatible with what has actually been observed, is

grounds in most cases for rejection of the hypothesis or theory in ques

tion. Application is not, as I shall argue below,
of the standard always a
straightforward matter
and empirical adequacy is not a sufficient criterion
of theory and hypothesis choice. So, other values come into play in theory,
hypothesis and model assessment.
2) Novelty. By novelty, Imean models or theories that differ in signifi
cant ways from presently accepted theories, either by postulating different
entities and processes, adopting different principles of explanation, incor

porating alternative metaphors, or by attempting to describe and explain

phenomena that have not previously been the subject of scientific inves

tigation. Several thinkers have endorsed the novelty of a model or theory

as a value. Sandra Harding seems to do so explicitly when she calls both
for "successor science" and for "deconstructing the assumptions upon
which are grounded anything that resembles the science we know" (Hard
ing 1986b). And she has interpreted Donna Haraway as supporting "an

epistemology that justifies knowledge claims only insofar as they arise

from enthusiastic violation of the founding taboos of Western humanism"

(Harding 1986b, p. 193). Without going that far, certainly one can read
 GENDER, POLITICS,AND THE THEORETICALVIRTUES 387

Haraway's invocation of the visions of certain science fiction writers as

an appeal for or endorsement of a departure from entrenched assumptions,
for the sake of a new framework (or new frameworks). Nothing less, she
suggests, will be appropriate for the new circumstances of twenty-first
century life (Haraway 1992).
Treating novelty as a virtue reflects a doubt that mainstream theoretical
frameworks are adequate to the problems confronting us, as well as a sus
picion of any frameworks developed in the exclusionary context of modern

European and American science. Since mainstream traditional frameworks

have been used in accounts that either neglect female contributions to pro
cesses biological and social or that treat as natural alleged male superiority
in various dimensions, something new will be required to address phenom
ena in a non-androcentric way. Novelty could, of course, have stronger and
weaker interpretations. The strong interpretation demands new frameworks
and theories to replace current ones in the domains in which they are cur
rently employed. On the weaker interpretation, new frameworks are to
be sought in satisfying a demand for scientific understanding of hitherto
neglected phenomena.
3) Ontological heterogeneity. Any theory posits, implicitly or explicitly,
an ontology, that is, it characterizes what is to count as a real entity in
its domain. A theory characterized by ontological heterogeneity is one
that grants parity to different kinds of entities. Ontological homogeneity,
by contrast, characterizes theories that posit only one sort of causally
efficacious entity, or that treat apparently different entities as versions of a
standard or paradigmatic member of the domain, or that treat differences as
eliminable through decomposition of entities into a single basic kind. This
criterion is found in two quite different sorts of discussion in the feminist
literature on the sciences. Feminists writing about biology have urged
that we take account of individual
difference among the individuals and
samples that constitutethe objects of study.3 Although she was not herself a
feminist, Barbara McClintock's attention to the individual kernels of a cob
of corn (which helped her to recognize an underlying pattern of mutability)
has been taken as a paradigm of what a feminist attitude to nature ought to
be. Primatologist Jeanne Altmann has insisted on methods of observation
that descriptively preserve the differences among the primates and groups
of primates that she studies. (Altmann 1974)
Treating individual as important and not to be elided
differences in
abstractions or idealizations smooth out heterogeneity
which is valuing

heterogeneity, taking it as a basic aspect, if not of the natural world, of

one's theories of it.4 One may have a variety of reasons for so valuing
models that preserve heterogeneity. The reason feminists have embraced
 388 HELEN E. LONGINO

this aspect of the work of McClintock, Altmann, and others is connected, I

think, to the second discussion I draw on here: the rejection of theories of

inferiority. of inferiority are supported

Theories in part by an intolerance of

heterogeneity. Difference must be ordered, one type chosen as the standard,

and all others seen as failed or incomplete versions. Theories of inferiority
which take the white middle-class male (or the free male citizen) as the
standard grant ontological priority to that type. Difference is then treated
as a departure from, a failure to fully meet, the standard, rather than
simply difference. Ontological heterogeneity permits equal standing for
different types, and mandates investigation of the details of such difference.
Difference is resource, not failure.
4) Complexity of Relationship. While the prior criterion values theories
that are pluralist with respect to entities, this criterion values theories that
treat relationships between entities and processes as mutual, rather than
unidirectional, and as involving multiple rather than single factors. Many
feminist scientists have taken complex interaction as a fundamental princi
ple of explanation. Evelyn Keller's (1983) account of the work of Barbara
McClintock and her defense of an interactionist perspective in her (1985)
may provide the best known examples, but scientists from icons like Ruth
Bleier and Anne Fausto-Sterling to much less well known practitioners
have eschewed single factor causal models for models that incorporate
dynamic interaction, models in which no factor can be described as domi
nant or controlling and that describe processes in which all active factors
influence the others. This perspective has been employed in areas rang
ing from neuroscience to cell biochemistry by scientists self-consciously
practicing science as feminists as well as, of course, by non-feminists. It
has also been endorsed in texts devoted mainly to reflections about the
sciences.

The rationales offered for embracing this criterion have ranged from a
metaphysical certainty that this is the way the world is to the notion that
the criterion expresses a femalequality of apprehension. Some rationales
are less antecedently problematic than others. In particular, one might note
that simple models of single factor control often makes one party to an
interaction a passive object rather than an agent. This has been the fate
of female gametes in accounts of fertilization and of female organisms in
accounts of social structure. Asymmetry of agency in the physiological
context is used to naturalize asymmetry in the social. Replacing simple
models of single factor control in social contexts with more complex mod
els of social interaction makes visible the role of gender in the structure of
social institutions and the role of private, domestic (traditionally, women's)
work inmaintaining the activity and institutions of the "public" sphere.
 GENDER, POLITICS,AND THE THEORETICALVIRTUES 389

5) Applicability to Current Human Needs. This and the next are more

pragmatic criteria, and relevant to decisions about what theories or theoret

ical frameworks to work on. This criterion favors research programs that
can ultimately generate applicable knowledge. Many, but not all, feminists
in the sciences have stressed the potential role of scientific understanding
in improving the material conditions of human life, or alleviating some
of its misery. Scientific inquiry directed at reducing hunger (by improving
techniques of sustainable agriculture, soil preservation, etc.), promoting
health, assisting the infirm, protecting or reversing the destruction of the
environment, is valued over knowledge pursued either for political domi
nation, i.e., science for "defense", or for knowledge's sake. As expressed
in feminist contexts, this is not just a call for more applied science, but for
research that can be directed towards meeting the human and social needs

traditionally to
ministered by women. The applicability criterion could be
understood, then, as requiring research into hitherto neglected areas and
hence triggering the novelty criterion in its weaker interpretation.
6) Diffusion of Power. This criterion is the practical version of the
fourth criterion, the one favoring models that incorporate interactive rather
than dominant-subordinate relationships in explanatory models. This one

gives preference to research programs that do not require arcane exper

tise, expensive equipment, or that otherwise limit access to utilization
and participation. This feature has emerged as a value in a number of
different contexts. Feminists in engineering and in economics have con
demned requirements of mathematical achievement far beyond what is

required for successfully engaging in these fields. Other feminists, such as

Hilary Rose (1983) and Ruth Ginzburg (1987), have urged a revamping
of traditional distinctions to include widely
distributed practices such as
midwifery as scientific practices. They urge that such practices be used as
models for feminist science practice. Feminist health professionals urge a
preference for medical practices and procedures that empower the individ
ual woman either to make decisions about her health or to retain control
over her own body. And ecofeminists and feminists in developing regions
urge the development of technologies that are accessible and that can be

locally implemented (Sen and Grown 1987). Diffusion or decentralization

of power interprets the above cited elements of the applicability criteri
on as knowledge of soil conservation, intensive small-scale sustainable

agriculture, promoting health by preventive measures such as improved

hygiene rather than high-tech interventive measures available only to the

few, protection of the environment by conservation and widely dispersed
renewable energy technologies.
390 HELEN E. LONGINO

The various proponents of these standards have had different ideas about
how they work or ought to work in inquiry. If we treat them as components
of a community set of public standards as I am suggesting, we take them as
criteria proposed for the assessmentof theories, models, and hypotheses,
guiding their formulation, acceptance, and rejection (or perhaps in the
case of the last two, what Allen Franklin calls theory pursuit). As Kuhn
noticed for the values he discussed, these require further interpretation to be
a
applied in given research context, they are not simultaneously maximally
satisfiable, and they are not subject to hierarchical ordering or algorithmic

application.

in

Since empirical adequacy is almost universally recognized as a value, and

since others of these characteristics have been endorsed as virtues by non
feminists, one might well wonder what about these standards is specifically
feminist.5 Several answers to this question can be discerned in the texts in
which these virtues have been endorsed.
One approach holds that these
characteristics express a feminine or
female orientationto the world, i.e. that women either because of biology
or social experience are more likely to understand the world via theories
characterized by these traits. This is said primarily of the substantive and
pragmatic virtues. Women are said, for example, to be more inclined to
perceive mutual influence and interaction than unidirectional single factor
control, and to be more interested in research that will help others. What
would be feminist, then, would be treating as theoretical virtues charac
teristics of women's ways of thinking about the natural and social worlds.
The problems with this approach are, first, that there's no evidence that
women are inclined biologically or culturally to understand the world in
these ways, and, second, that even if they were, we'd still need an argument
that these were traits that ought to be valued in theory construction and
assessment. Of course, if one is antecedently convinced, as some advocates
of these virtues are, that the world
really is constituted of heterogeneous
entities that interact in complex ways, the need for such an argument will
be much less apparent than it is to one less certain. But if the world is such
as to be more adequately understood via theories exhibiting these virtues,
then they ought to be promoted as general theoretical virtues and not just
as feminist theoretical virtues.
A secondapproach suggests that women are more likely to value the
characteristics of theories because they are outsiders tomainstream science
and so less likely to be acculturated to the values of the mainstream. But this
 GENDER, POLITICS,AND THE THEORETICALVIRTUES 391

is an even less plausible candidate for grounding the claim that the virtues
would be feminist.6 Neither female biology nor feminine conditioning, but

marginality explains the appeal of these virtues. Marginality, however, is

common to any group excluded from the practice of science and so not

specifically feminist. Furthermore, while marginal status may alienate or

free those marginalized from mainstream values, in some cases preference
for alternative values may be the basis of marginalization. And, as is the
case for the previous approach, the empirical data supporting the view that

marginalized groups are likely to endorse these virtues in particular has

yet to be brought forward.
Rather than look to sociological facts uses them, I have
about who

suggested that we look to the work

these virtues can do for specifically
feminist inquiry.7 In the account given above of each of the virtues, I

suggested how inquiry guided by them would be thought to reveal gender,

either in the form of bias about the phenomena or as a phenomenon in
the domain itself, or to reveal the activities of women or females in the
domain. The aim of revealing gender and/or the activities of women is, I
propose, what makes inquiry feminist. Feminist theoretical virtues will be
those that serve this aim. Thus, satisfying it is a bottom line requirement
of theoretical standards. I should emphasize that I am not arguing here
that the virtues I have discussed so far are the theoretical virtues feminists
should adopt. I think such a claim needs further discussion and argument.
What I do propose is that the basis on which such a claim should be argued
and disputed is the contribution any proposed virtue can make to furthering
feminist goals in inquiry. If the virtues that have been discussed here are
feminist, it is because they satisfy this bottom line requirement, and not
because of any intrinsic, statistical, or symbolic association with women
or cultural femininity.

IV

My reasons for thinking about the elements and structure of this list are
several. One of course, is to contribute to the feminist discussion of scien
tific knowledge. In that spirit, I've argued that the bottom line requirement

gives us a basis for critique of the standards as I articulated them amoment

ago. Secondly, and relatedly, I've argued that we should understand the
values not as absolutes, but as contextually mandated at particular moments
in the framework of a kind of bootstrapping provisionalism.8 But my pur
pose here is to use this example of an alternative to the traditional set of
theoretical virtues as a lens through which to focus on some of the latter's
less-remarked aspects.
392 HELEN E. LONGINO

One way to begin is to pair elements from the two sets.

Feminist list Traditional list

Empirical adequacy Accuracy
Novelty Internal /External Consistency
Ontological heterogeneity Simplicity
Complexity of interaction Breadth of scope
Applicability to human needs
Diffusion of power Fruitfulness

There isn't a single neat pairing, but some rough links can be made. Of
these only the pair made by the first in each list tend in the same direction,
while the other pairs include elements pulling in opposite directions. As you
will I see, I don't think this is a reason for thinking that empirical adequacy
or accuracy - or scientific - value
is the real constitutive in scientific

judgment, while the rest are all contextual. I want to start, however, by
exploring the more clearly contrastive pairs.
Novelty and external consistency are quite starkly opposed. The nov
elty criterion recommends theories and models that depart from accepted
theories, while the criterion of external consistency recommends those
that do not contradict them. Different interpretations of the two criteria
can produce different articulations of the contrast, but what interests me
here is their socio-political valence. Obviously, the socio-political basis for
the criterion of novelty is the need for theoretical frameworks other than
- - in
those that have functioned directly or indirectly gender oppression.
External consistency, in a context in which theories have had that func
tion, perpetuates it. Those satisfied with the status quo will endorse this
criterion, and the effect of its endorsement is to keep from view the ways
in which currently accepted theories are implicated in the legitimation of
gender oppression.
Donna Haraway (1986) has pointed out, for example, how the reten
tion of a socio-biological framework in Sarah Blaffer Hrdy's feminist
primatology replicates problematic moves in liberal feminism, which per
petuates the framing assumptions about individualist and self-regarding
human nature of liberal political theory. In both cases the feminist turn is
limited to claiming for females what has been reserved for males with
out challenging the deeper assumptions about human nature involved in
both the scientific and the political program. And Susan Sperling (1991)
develops a similar argument with respect to the functionalist and socio

biological frameworks she identifies in most of the feminist primatology

of the last twenty-five years. Her point is that it preserves essentialist and
 GENDER, POLITICS,AND THE THEORETICALVIRTUES 393

determinist concepts of gender, its feminism being restricted to revaluing

the roles of females in primate evolution. The models advanced by these

primatologists thus satisfy the mainstream virtue of external consistency.

Paying attention to females, making them more central to the analysis,
corrects omissions of androcentric field work, but by leaving the theoret
ical scaffolding in place, the work under discussion fails to challenge the
ways in which
socio-biological analysis naturalizes the social relations of
capitalism. While a few women may benefit in such a system, the vast
majority are impoverished. Even though it has been resisted in certain
quarters, one reason the feminist primatology has been taken seriously is
its conservativism with respect to basic theory. According to Haraway and
Sperling, its exemplification of this traditional virtue is also a cause of its

political regressiveness.
Pursuing contrast, we can see how certain interpretations
yet another of
the simplicity criterion are laden with socio-political values. The interpre
tation that contrasts with ontological heterogeneity is an ontological one:
the simpler theory is the one positing the fewest different kinds of funda
mental entities (or of causally effective entities). This encourages us to find
ways of treating putative entities which are not members of the privileged
class either as epiphenomena, as constructions that can be disassembled
into collections of entities of the privileged class (cells into molecules,
molecules into atoms, etc.), as parts of members of the privileged class,
or as variants whose deviations from the standard can be disregarded. To
suppose the social world is composed of just one or a few kinds of basic
entity (e.g. rational self-interested individuals in neoclassical economic
theory) erases the differences among persons that are fundamental to how
they act. Economics, for example, treats the head of household as the main
economic actor - in the household - and
assuming its (his) dominance
that the interests of other members of the household -
assuming spouse,
- are identical with those of the
partner, children, elderly parent head. By
erasing the independent interests of other household members from the
oretical view, these models prop up an oppressive family structure (one
- - to make
person "the benevolent patriarch" is supposed the decisions)
and indirectly legitimate the assumption by welfare policy makers, family
policy makers, etc., that this structure is the primary and appropriate family
structure in our society.
One of the other interpretations of simplicity is in the form of a contrast
with the virtue
of complexity of interaction. A model involving causal
relations going only in one direction is simpler than one in which elements
in the model interact to produce an effect or in which elements mutually
influence each other. As I suggested above, the former, which mimic social
394 HELEN E. LONGINO

relations of domination, naturalize these relations. That is, the more the
natural orderappears to be one in which factors can be said to control
processes or events the more such an organizational form seems to be
-
natural the way of nature. To treat this form of simplicity as a theoretical
virtue is to incorporate its socio-political valence into the justificatory
processes of science.
Heterogeneity and complexity of interaction can also be contrasted
with the virtue of breadth of scope. The former validate models contain
ing different sorts of fundamental
entities in complex relations of mutual

dependence. This
is compatible with only loosely related multiple models
rather than a quantity of models or phenomena derived from (explained
by) one or a very few basic principles, which is the situation envisioned
by breadth of scope. I shall leave the political interpretations here to the
reader.

Finally, the feminist practical virtues can also be used to demonstrate

the political valence of their counterpart.
Fruitfulness, for Kuhn, referred
to the capacity of a theory to generate problems or puzzles demanding
solutions and to provide the resources with which to solve them. This, of
course, means more opportunities to articulate connections between the
theory and putatively established phenomena as well as other theories.
The feminist practical virtues favor theories and models that can be used
to improve living conditions in a way that reduces inequalities of power.
Taking them
seriously requires looking beyond the immediate (internal)
context of research to the ways in which itmight or might not be developed.
This in turn requires taking stock of the social, political, and economic
context in which development might take place. Fruitfulness is by contrast
conservative in that it is inward looking. It directs attention away from
the social and technological applications of research, whether they be
beneficial or harmful.
Well, what of empirical adequacy and accuracy? Why can't we treat
them as the same virtue and the fact that it appears in both lists as evi
dence that here at last is a possible candidate for a purely cognitive virtue?
As someone who is by instinct some kind of empiricist this is an appeal

ing move. Let me explain why I hesitate to make it. Regina Kollek has
challenged (in discussion) the appropriateness of citing empirical adequa
cy as a feminist criterion. Don't the data change in different discursive
and research contexts, she asks? If so, doesn't empirical adequacy beg
the question with which data and with whose data the observational ele
ment of a model or theory ought to be in agreement? Shouldn't a feminist
criterion consist in specific methods rather than the demonstrably vague
notion of empirical adequacy? Iwould not go so far as to specify particular
 GENDER, POLITICS,AND THE THEORETICALVIRTUES 395

methods as inherently feminist or not, as I think how a method is used

is more important than the method itself. But the question draws atten
tion to ways in which empirical adequacy needs further interpretation or

specification.
One of the critiques of modern experimental methods is that they
involve what Ruth Hubbard calls "context stripping". When we detach
a factor from the contexts in which occurs, we are hoping
it naturally to
achieve understanding of that factor's precise contribution to some process.
But by taking it out of its natural context we deprive ourselves of under
standing how its operation is affected by factors in the context from which
it has been This is, of course, a crucial aspect of experimen
removed.
tal method. I suspect that it's not (or not always) the decontextualization
that is to be deplored, but the concomitant devaluation as unimportant
or ephemeral of what remains. The decontextualization of experimental
variables is analogous to the way in which activity in the public domain
of modern industrial societies is analytically detached from the material
conditions of its possibility in the private domain. Resistance to the con
stant marginalization of domestic (and female) activity has made feminists
sensitive to the processes of exclusion and devaluation. These are prob
lematic not only in our understandings of the social world but also in our
understanding of the natural world. The failure to attend fully to the inter
actions of the entire social group, including its females, in studying the
males of a species has led to distorted accounts of the structure of animal
societies, including male-male interactions. In toxicity studies, the focus
on a single chemical's toxic properties fails to inform us how its activity
is modified, canceled or magnified by interaction with other elements in
its naturalenvironments. Focus on gene action has blinded us to the ways
in which the genes must be activated by other elements in the cell. These
models may well be empirically adequate in relation to data generated in
laboratory experiments, but not in relation to potential data excluded by a
particular experimental set up.
Comparable remarks must be made accuracy. A model may be
about
accurate with respect to a narrow
range of possible data. But why should
the data in that band be the relevant ones? Empirical adequacy and accu
racy (treated as one or separate virtues) need further interpretation to be

meaningfully applied in a context of theory choice. Those interpretations

are likely to import the socio-political or practical dimensions that the
search for a purely cognitive criterion seeks to escape. At the very least the
burden of argument falls on those who think such an escape possible.
396 HELEN E. LONGINO

I've argued that by identifying values of a scientific community other

than the traditional ones we can get insight into important features of
the latter. In particular, I've tried to give some reasons for thinking that
those traditional are not purely epistemic
values (if at all), but that their
use in certain contexts of scientific judgment imports significant socio

political values into those contexts. Even the apparently neutral criteria of
accuracy or empirical adequacy can involve socio-political dimensions in
the judgment of which data a theory or model must agree with.
I do not want to claim that the virtues or criteria I've discussed have
fixed and absolute socio-political meanings. the very least, whatever
At
valence they have will be modified by their interaction with whatever other
criteria are brought to bear in a given situation and the relative priorities

assigned to the different members of a given set. And the social context
in which they are used will also make a difference. Thus, it is not clear
that treating simplicity as a theoretical virtue would have the same socio

political resonance in a social-political context which values diversity and

equality. But in our context, in which diversity and equality are granted
lip service but made to defer to more important social values like order
and economic competitiveness, and in which the physical and life sciences
possess a greater cognitive authority than other intellectual sources of
value, it does serve anti-progressive ends. Similarly, heterogeneity could,
in a context other than our own, fail to be a theoretical virtue with a
liberatory potential.
Finally, a note of caution must be sounded. Simply articulating those
theoretical virtues that could count as feminist does not mean it is possible
to develop, pursue, and establish theories and models that exemplify them.
The virtues to which they are opposed are mainstream precisely because
they stand in reinforcing relations with values and relations in the social
context of science. There may be pockets, "niches", in which it is possible
to practice alternative science that satisfies at least some feminist criteria
(the ones discussed or others). But these must be deliberately created and
will stand in complicated relations to other sites of research. The practice
of is too materially
science dependent on its socio-political context for

significant change to be possible without changes in that context.

One cannot claim by looking at the theoretical virtues themselves that
they are liberatory, oppressive, feminist, masculinist, or neutral. One must
look instead at the grounds that are offered for treating them as virtues
and the ways in which their deployment in particular scientific arguments
and research programs resonates with conditions in the social and political
context of the research. One must look as well at the consequences of
 GENDER, POLITICS,AND THE THEORETICALVIRTUES 397

employing their contraries in a given context, and the grounds for doing
so or not. Placing the virtues in this fuller setting of use and defense
undermines the idea that we could separate out the purely cognitive from
the political in any absolute or final sense. It depends on the context.

NOTES
*
I am grateful to participants in the Indiana University Workshop on Social Values in the
Context of Justification for their comments on an earlier, spoken, version of this paper.
1
See Quine and Ullian (1978) for simplicity and external consistency, i.e., consistency with

presently accepted theory. Philip Kitcher (1993) has argued for unification as a scientific
desideratum. This, for present purposes, can be considered as a variation of breadth of

scope. Kitcher, it might be argued, has a more precise measure in mind than the notion of
breadth of scope allows. I will treat differences in a later paper.
2
I first discussed what I have called the feminist theoretical virtues in (Longino 1993a).
I used them again in a discussion of the possibility of feminist epistemology in (Longino

1994). The exposition of the next several pages borrows from those earlier publications.
3
See Bleier (1984), Keller (1985), Fausto-Sterling (1985).
4
Stephen F. Kellert has suggested that the value at work here is particularism. I shall defer
consideration of the relation between particularity and heterogeneity to another occasion.
5
For example, Levins and Lewontin (1985) embrace both heterogeneity and a strong form
of interaction they label "dialectical" as features of dialectical biology. Literary scientists

Stephen J. Gould and Lewis Thomas endorse interaction as a principle of explanation.

Whatever sympathies with feminism they may have, it is not feminism that leads them to

heterogeneity or interaction. Indeed Gould explicitly says that gender or feminism have

nothing to do with it. It's just a matter of good science.

6
Ian Barbour suggested this point to me in conversation.
7
Cf. Longino (forthcoming).
8
Cf. Longino (forthcoming).

University of Minnesota

Minneapolis, MN 55455
U.S.A.