科学哲学ニューズレター |
No.18, September 1997
Tetsuji Iseda: Can epistemology be social?: A review of recent attempts in social epistemology
Tetsuji Iseda
University of Maryland at College Park
The recent development of Science Studies has stimulated many people to
do interdisciplinary studies on science, and philosophers are no exceptions.
Especially since the late eighties, philosophers inspired by SSK (sociology
of scientific knowledge) works have been attempting to incorporate sociological
insights into epistemological arguments. The purpose of this paper is to
review some of these attempts. First, I analyze attempts in socialized epistemology
as an extension of naturalized epistemology. A.I. Goldman and Philip Kitcher
are representative people in this trend, and I also include David Hull in
this category despite some differences between his approach and others'.
Then, as an alternative to the above approach, I summarize Steve Fuller's
social epistemology ('socialized epistemology' and 'social epistemology'
are usually used interchangeably in the literature, but in this paper I
use these terms to distinguish Fuller's view from the other approach). My
conclusion is that both of these attempts have problems to overcome.
1. Socialized epistemology
1-1. From naturalized to socialized epistemology
Many philosophers have been trying to extend naturalized epistemology by
incorporating sociological (and sometimes economic) insights into it. Naturalized
epistemology was proposed by Quine in his paper "Epistemology naturalized"
(1969) and has been developed by A.I. Goldman and others. There are many
different trends in naturalistic approach, but several typical features
should be mentioned here. First, naturalized epistemology regards the failures
of traditional epistemology as the failure of foundationalism that seeks
a firm foundation of knowledge. proposes to give up foundationalism itself.
Rather, according to this view, we need to start by taking a large par t
of science (especially psychology) for granted to do epistemology. Thus,
Quine proposes that epistemology should be a behavioristic study of the
relation between theory and evidence, and most naturalized epistemologists
after him emphasize the importance of cognitive science in epistemology.
Since epistemology is supposed to assess the validity of scientific knowledge,
this means a circular relationship between science and epistemology. Nevertheless,
if foundationalism is really hopeless, then maybe circularity is not that
bad after all. Second, in naturalistic approach, justification is defined
in term s of psychological states or processes (Kim 1988, 48). A typical
definition is reliabilism: our belief is justified if it results from a
reliable belief forming process, and a process is reliable if the process
tends to produce tru e beliefs (Goldman 1992, 112-125). Our scientific knowledge
is used to distinguish true beliefs from false ones. Finally, even though
many naturalize d epistemologists, unlike Quine himself, recognize the importance
of the normative task of epistemology, they still try to restrict the scope
of normative judgment to actual psychological processes in various ways
(Kornblit h 1994a nicely summarizes these various arguments) . A good example
is what Fuller calls the "ought implies can" principle (Fuller
1992, 442). That is, when it is hard for people not to use a certain psychological
process, they do not have the obligation not to use it, whether it is problematic
or not. For example, many psychologists have studied the so-called "availability
heuristic," namely the phenomena that people tend to assign a higher
frequency to things which come to mind more easily than they do to other
things (Tversky and Kahneman 1973, Nisbett and Ross 1980). This heuristic
sometimes leads us t o an erroneous assessment of frequency, but Goldman
argues that we don't have th e obligation to avoid it, if the heuristic
is as stubborn as psychological studies suggest (Goldman 1992, 173-174).
Naturalized epistemology usually deals with individual psychological processes
, but it is not hard to extend the scope of naturalistic arguments to the
consensus formation process of a scientific community, and this is what
socialized epistemology is about. The following subsections deal with three
proponents of such social approach.
1-2. Goldman's 'social epistemics'
A.I. Goldman is not only a leading proponent of naturalized epistemology,
but also a pioneer in the field of socialized epistemology (which he calls
'social epistemics'). He argues that there are a couple of reasons why epistemologists
should be interested in sociology (Goldman 1992, 179-183). The first is
that there are many potential cases of belief formation in which a social
explanation is appropriate. Even rational beliefs call for a social explanatio
n when we need to look at interpersonal influences among scientists. Second,
sociological considerations are relevant to the normative aspects of epistemology,
when we want to evaluate the epistemic merits of different socia l institutions.
Goldman considers four alternative bases for such an evaluative task, namely
relativism, consensualism, expertism, and veritism. He finds difficulties
with the first three, and endorses veritism, according to which social practices
or institutions should be evaluated in terms of production of true beliefs.
Obviously this is an extension of his reliabilism, but in this case Goldman
thinks that justification is too narrow a concept to evaluate a social practice.
He would rather talk about the "intellectual strength" of a practice
, namely its reliability, power, fecundity, speed, and efficiency, in producing
true beliefs (ibid. 195; a similar, but more clear-cut reliabilist approach
is advocated in Kornblith 1994b). To apply these criteria, we need to assess
the truth value of beliefs, and this assessment is done through established
scientific procedures (205). Again he does not seem to think that this circularity
is problematic (206).
Beside the above meta-level argument, Goldman did an interesting study on
the belief formation process with an economist, M. Shaked (1991). This is
an analysis of the credit-seeking activity of scientists. Goldman and Shaked
firs t give a formal definition of credit (232). A contribution gets credit
in accordance with the degree to which it changes other scientists' degree
of belief that a certain theory is true. If scientists seek credit in this
sense, then their activity can lead to more true beliefs when several further
assumptions are met: such as that scientists are generally Baysians, and
the likelihood assessment of the scientific community is objective enough.
They obtain several interesting results with this model. For example, they
show tha t the credit-seeking activity of scientists can lead to more true
beliefs than genuine truth-seeking activity, under some specific conditions.
What is the point of such an analysis? Obviously their definition of credit
is too artificial to be regarded as a description of actual credits recognized
by scientists. On the other hand, they seem to avoid making normative judgments
i n this paper. It seems to me that their purpose is to reconcile two views
of science, namely scientists as those motivated by non-epistemic goals
(as sociologists claim) and science as an efficient epistemic enterprise
(as philosophers tend to believe).
1-3. Kitcher and division of cognitive labor
Philip Kitcher's elaborate analysis of the social dimension of scientific
communities is another representative work in this field (Kitcher 1990,
1993).
Kitcher uses a variant of reliabilism as his standard of rationality, accordin
g to which the rationality of a procedure of conceptual change is judged
in term s of the ratio of progressive change, and progress is in turn judged
in terms of the attainment of significant truth (his own formulations
are fairly complicated; see Kitcher 1993, ch.4, ch.6). This standard is
applied to both individual cognitive processes and scientific communities,
but here we are concerned with the latter.
Kitcher is especially interested in the division of cognitive labor. The
problem he deals with is as follows (Kitcher 1990, 5-8): if each scientist
makes an ideally rational judgment about plausibility of rival theories
or research programs, then all scientists choose the theory or program best
supported by available empirical evidence. In an ideal situation, all scientists
will pursue the same program. On the other hand, from the community's point
of view, if there are rival theories or research programs, the best thing
is to allocate scientists to each theory or program in accordance with the
relative plausibility of those alternatives. Thus individual rationality
and communal rationality seem to contradict with each other.
He tries to solve this predicament by introducing a further assumption that
scientists are epistemically sullied agents, namely they seek not
only truth, but also priority (Kitcher 1993, ch.8). If a scientist pursue
the same program as many other scientists, the probability that this scientist
will mak e the first discovery becomes quite low. On the other hand, even
if a program is unlikely to succeed under available evidence, if it is actually
a correct one and no one else pursues it, the scientist can become the first
discoverer without competition. Therefore, if scientists behave as rational
sullied agents, the community attains better allocation than the community
of pure epistemic agents. Kitcher also analyzes the roles played by authority,
tradition, local autocracy etc. in great mathematical detail. These seemingly
non-epistemic elements of scientific practice are found to have certain
epistemic functions.
Kitcher is fairly clear in that he is doing a normative task of identifying
successful (i.e. progressive) social systems (Kitcher 1993, 303). But it
seems to me that he fails to establish that the systems he endorses are
more progressive than the alternatives. Even if scientists' efforts are
allocated properly, if the communal criteria of acceptance of a discovery
are bad ones, this community is unlikely to make progress. On the other
hand, if the communa l criteria are good, this community is likely to make
progress in the long run, whether or not the efforts are allocated properly.
In this sense, Kitcher seem s to fail to establish the relevance of his
argument to his own standard of rationality.
1-4. Hull's evolutionary account
Even though David Hull is not usually classified as a naturalized epistemologist,
his position is even closer to Quine than that of Goldman or Kitcher. He
abandons traditional epistemology because he cannot take the "mindless
doubt" of skeptics seriously (Hull 1988c, 243). He sees no possibility
of answering the skeptics. Hull also dismisses normative epistemology because
there is no justification of knowledge claims that epistemologists have
been looking for (Hull 1988a, 13). The project to which h e commits himself
is, thus, purely descriptive and non-skeptical. What he tries to show is
that "science is so organized that self interest promotes the greater
good," namely the increase of "our knowledge of the empirical
world" (Hull 1988a, 357). Notice that he presumes that science does
produce knowledge of the empirical world.
Hull's analysis of scientific practice is marked by his extensive use of
analogy with the evolutionary process. He also tries to support his argument
with his own case study of the history of taxonomic systems in biology (Hull
1988a, chs. 2-7).
Hull applies notions in evolutionary biology, such as replicators, interactors
, selection etc. to scientists themselves (Hull 1988a, 434). Replicators
in science are beliefs, goals, methodologies, and so on. Interactors are
scientists. Scientists act for their conceptual inclusive fitness,
namely, so as to encourage other scientists to use their work. For example,
scientists give credit to other scientists in their own work because this
increases the credibility of the work. This will in turn increase the probability
that the work will be cited by other scientists (Hull 1988a, 310). Similar
relationships exist between a scientist and his/her own graduate students
(Hull 1988b, 127-128). The scientist should balance these considerations
to maximize his/her own conceptual inclusive fitness. This make s a scientific
practice a kind of competitive cooperation (Hull 1988a, 319).
Let us suppose that his descriptions of scientific practices are accurate
(and surely he can cite many actual cases to support his discussions). Does
this lead to an increase of knowledge about the world? Not necessarily.
Hull notice s this, and adds an assumption that erroneous claims are likely
to be discovered and punished severely (Hull 1988a, 320). But which claims
are erroneous? If he simply means claims which are recognized as erroneous
by the scientific community, the first half of his assumption becomes trivially
true. If not, he needs some independent criterion of error, and this seems
to require an involvement in traditional epistemology.
1-5. A critical assessment of these arguments
Naturalized and socialized epistemology are supposed to be alternatives
to traditional epistemology. But even if foundationalism has serious problems,
this does not automatically warrant to accept a large part of science. We
need an argument here, and such an argument should fall in the domain of
traditiona l epistemology. If this is so, I do not see any incompatibility
between these alternatives. The models by Goldman and Shaked, Kitcher, and
Hull can be seen as applied epistemology which bridges traditional
epistemology and actual scientific practice. Their rejection of an a priori
normative argument may have come from their misconception of normativity.
As Kim (1988) argues, more sophisticated meta-ethical considerations are
likely to resolve the seeming conflict between traditional and naturalized/socialized
epistemology. But this paper is not the place to pursue that possibility.
2. Fuller's social epistemology
2-1. Fuller's argument
Steve Fuller also takes sociological findings seriously, but his position
show s remarkable contrasts with that of naturalized epistemologists, so
it is worth discussing separately.
Fuller proposes "social epistemology" as science policy-making,
which is a normative enterprise to look for better scientific institutions,
informed by psychology and sociology (Fuller 1988, 3). The starting point
of social epistemology is the worry that "left to its own devices,
science will not necessarily produce the sort of knowledge that we are interested
in having" (Fuller 1992, 392). Usually we want knowledge not for its
own sake but for som e independent purposes like social welfare, and we
should judge the efficacy of science from this outside point of view. For
such judgments, Fuller proposes t o use an experimental method. For example,
we can conduct experiments taking independent variables such as group size
and communication constraints, and dependent variables such as the sort
of knowledge produced (Fuller 1992, 405). Of course the experimental results
are artificial in a sense, but this does no t bother social epistemologists.
If we find effective institutional structures i n experimental settings,
what social epistemologists may want to do is make "the world behave
more like the laboratory than vice versa" (Fuller 1992, 406).
One of the peculiarities of Fuller's argument is that he thinks that the
normative assessments cannot be purely epistemic -- it is always science
policy-making at the same time. He has several arguments why epistemic issues
and policy-making issues are not separable. First, even if scientists want
tru e information about their research objectives, this does not mean that
they want it above all else (Fuller 1996, 162). Any epistemic activity has
associated costs, and any normative argument about science need to consider
these costs. He also argues that we are in a post-epistemic world in which
mutual critical scrutiny of scientists can not work properly (Fuller 1994).
The major problem is the size of scientific activity. Some time ago, when
scientific enterprise was smaller, "it was reasonable to presume that
whatever one wrote would be subject to critical scrutiny by the relevant
disciplinary community" (43). But now, since the number of scientists
and their publications have increased remarkably, we cannot expect such
ideal readers. Now a scientific research project is really scrutinized when
the funding is determined, and the scrutiny is relative to the funder's
interest. But if science is conducted to meet such interests, research is
no longer evaluated as products of inquiry. It is just another kind of business.
2-2. A critical assessment of Fuller's argument
I think in some points Fuller's argument is better than those by naturalized
epistemologists. I agree with him that epistemology is an essentially normativ
e enterprise, and that the scope of normative judgments should not be restricted
to actual processes. I also think that his emphasis on an experimental method
suggests a better relationship between philosophers and sociologists. But
what I can never agree with him on is his claim about inseparability of
epistemic and non-epistemic judgments. It is true that when we make a final
assessment about a scientific institution we need to count everything in.
But this does not mean that we cannot distinguish epistemic considerations
and non-epistemic ones in the course of the assessment. I think the problem
is his naive conception of normativity. Again, sophisticated meta-ethics
can help here. In this respect, Miriam Solomon's "social empiricism"
(1994) proposes an interesting revision of Fuller.
3. Conclusions
Since the purpose of this paper is to review these attempts, I have provided
only sketchy assessments of them. They ask many interesting questions which
ar e not asked in traditional epistemology, and some of them are surely
useful if w e want to apply traditional epistemology to the evaluation of
actual scientific practices. I think that social/socialized epistemologies
are not conceived in this way mainly because the proponents of them have
only naive understanding o f normativity.
(c) Tetsuji Iseda