Bradie Assesing-1-50

ResearchGate
See discussions, stats, and author profiles

for this publication at: https:/fwwnw.researchgate.net/publication/227025801
Assessing Evolutionary Epistemology

Article in Biology £ Philosophy - December 1986
DOL:101007/8F00140%2
ciTarions READS
189 1472
1author:
Michael Bradie
Bowling Green state University
83 PUBLICATIONS 412 CITATIONS
SEE PROFILE
All content followingthis page was uploaded by Michael Bradie on 21 April 2016
“The user has requested enhancementofthe dovntoadefle
Assessing Evolutionary Epistemology*
MICHAEL BRADIE
Department af Philosaphy
Bowling Green State University
Bowling Green OH 43403, U.S.A.
ABSTRACT: There are two interrclated but distinct programs which go by the name
“evolutionary epistemology.” One attempts to account for the characteristics of cognitive
mechanisms in animals and humans by a straightforward extension of the biological theory
of evolution to those aspects or traits of animals which are the biological substrates of
cognitive activity, e.g., their brains, sensory systems, motor systems, etc. (EEM program).
The other program attempts to account for the evaluation of ideas, scientific theories and
culture in general by using models and metaphors drawn from evolutionary biology (EET
program). The paper begins by distinguishing the two programs and discussing the
relationship between them. The next section addresses the metaphorical and analogical
relationship between evolutionary epistemology and evolutionary biology. Section IV treats
the question of the locus of the epistemological problem in the light of an evolutionary
analysis. The key questions here involve the relationship between evolutionary epis-
temology and traditional epistemology and the legitimacy of evolutionary epistemology as
epistemology. Section V examines the underlying ontological presupposition" and impli-
cations of evolutionary epistemology. Finally, section VI, which is merely the sketch of a
problem, addresses the parallel between evolutionary epistemology and evolutionary ethics.
KEY WORDS: Evolution, epistemology, selection, analogy, metaphor, norms.
Mr. Drummond, I hope [ haven't said anything to offend

you. You see, 1 haven't really thought very much. I was
always afraid of what I might think — so it seemed safer
not to think at all. But now I know. A thought is like a
child inside our body. It has to be born. If it dies inside
you, part of you die, too! (Pointing to the book) Maybe
what Mr. Darwin wrote is bad. [ don't know. Bad or
good, it doesn't make any difference. The ideas have to
come out — like children. Some of 'em healthy as a bean
plant, some sickly. I think the sickly ideas die mostly,
don't you Bert?
Jerome Lawrence, Inherit the Wind
L INTRODUCTION
This paper is a work in progress. At this point [ am not completely sure

how it is going to turn out. The aim is to examine some recent work in
Biology and Philosophy 1 (1986) 401—459.

© 1986 D. Reidel Publishing Company.
402 MICHAEL BRADIE
what is called evolutionary epistemology in order to determine what, if

anything, evolutionary biology has to tell us about central philosophical
problems in epistemology and metaphysics. Along the way we shall
consider the extent to which “evolutionary epistemology” is either “evolu-
tionary™ or “epistemology.”
In order to keep the project within reasonable bounds, certain limita-
tions have been imposed. The basic goal is a taxonomic survey of problem
areas and suggested insights which have been claimed for evolutionary
approaches to epistemological issues. The extension of evolutionary bio-
logical considerations to questions in epistemology has roots in the 19th
century. But, aside from a few scattered remarks no attempt has been
made here to treat these 1 9th century contributions in any detail
Evolutionary epistemology as it now stands, at least in some of its
formulations, is part of what are now known as naturalized epistemologies.
Apart from suggesting how the two integrate with one another, no attempt
has been made to explore the general topic of naturalized epistemology.
That is subject for a treatise in itself. Evolutionary epistemology also
charts courses which lead directly into the deep contemporary discussions
on the topic of rationality and relativism. Again, outside of some brief
remarks which try to establish the general shape of the problems and the
alleged implications of evolutionary epistemology for them, these topics
have not been further pursued. More generally the questions which
evolutionary epistemology addresses are one facet of the much larger and
more complicated set of questions comprising the nature of sociocultural
evolution and its relation to biological evolution. These wider concerns
are, for the most part, beyond the scope of this paper.
The plan of this paper is as follows. In section II, I distinguish and
discuss the relationship between two different but interrelated programs
that both go under the name of “evolutionary epistemology.” Section III
addresses the metaphorical and analogical relationship between evolution-
ary epistemology and evolutionary biology. Section TV treats the question
of the locus of the epistemological problem in the light of an evolutionary
analysis. The key questions here involve the relationship between evolu-
tionary epistemology and traditional epistemology and the legitimacy of
evolutionary epistemology as epistemology. If its legitimacy, in some sense,
is granted then what can we learn about knowledge and knowing via this
approach? Section V examines the underlying ontological presuppositions
and implications of evolutionary epistemology. Finally, section VI, which
is merely the sketch of a problem, addresses the question of the parallel
between evolutionary epistemology and evolutionary ethics and what this
parallel can or cannot tell us about the relevance of evolutionary and
biological considerations in ethics and epistemology.
EVOLUTIONARY EPISTEMOLOGY 403
II. THREE DISTINCTIONS
1. There are two interrelated but distinct programs which go by the name
“evolutionary epistemology.” One is the attempt to account for the charac-
teristics of cognitive mechanisms in animals and humans by a straight-
forward extension of the biological theory of evolution to those aspects or
traits of animals which are the biological substrates of cognitive activity,
e.g., their brains, sensory systems, motor systems, etc. The other program
attempts to account for the evolution of ideas, scientific theories and
culture in general by using models and metaphors drawn from evolu-
tionary biology. Both programs have their roots in 19th century biology
and social philosophy, in the work of Darwin, Spencer and others. There
have been a number of attempts in the intervening years to develop the
programs in detail (see the bibliography in Campbell 1974). Much of the
contemporary work derives from work by Lorenz, Campbell, Popper and
Toulmin. In this section, I want to briefly describe these two programs and
establish their interrelationship.
For the sake of convenience, [ shall refer to the first attempt to extend
evolutionary theory to the explanation of the development of cognitivie
structures as the evolution of cognitive mechanisms program (EEM). I
shall refer to the other program, the attempt to analyze the growth of
knowledge using evolutionary models, drawn from biology, as the evolu-
tion of theories program (EET).
Development in biology is either ontogenetic or phylogenetic. The
development of knowledge or of knowing mechanisms exhibits a parallel
distinction. One might expect, since the biological processes of ontogenesis
are different from the biological processes of phylogenesis, that evolu-
tionary epistemologies would reflect this difference. Curiously enough,
however, for the most part they do not. I shall return to this point
below.
Let us turn to the characterization of EEM and EET. A clear statement
of the EEM program can be found in Vollmer (1975, p. 102):
Our cognitive apparatus is a result of evolution. The subjective cognitive structures are
adapted to the world because they have evolved, in the course of evolution, in
adaptation to that world. And they match (partially) the real structures because only
such matching has made such survival possible (quoted by Bunge 1983, p. 8).
Lorenz expresses similar sentiments in his 1977 book, Behind the Mirror:
T consider human understanding in the same way as any other phylogenetically evolved
function which serves the purposes of survival, that is, as a function of a natural
physical system interaction with a physical external world. (Lorenz 1977, p. 4)
In an earlier paper, Lorenz had endorsed the ‘biologizing of Kant'. The a
priori categorical structures which organisms use to form their cognitive
404 MICHAEL BRADIE
pictures of reality are to be understood as the a posteriori evolutionary

products of phylogenetic development. Thus,
One familiar with the innate modes of reaction of subhuman organisms can readily
hypothesize that the a priori is due to hereditary differentiations of the central nervous
system which have become characteristic of the species, producing hereditary dis-
positions to thiak in certain forms. (Lorenz 1982, p. 122)
A recent turn among biologists, which Popper and Campbell endorse
as well, not only sees cognitive structures as evolutionary products but
attempts to analyze all biological evolutionary development as the evolu-
tion of ‘knowledge’ structures (Plotkin 1982). Plotkin sees contemporary
evolutionary epistemology, as he understands it, as comprising a set of
philosophical issues as well as a set of biological issues. The philosophical
issues include
(1) questions about the validation and limitations of human knowl-

edge in the light of the supposed evolutionary development of
knowing structures,
(2) the recognition that the cognitive abilities of humans are crucial
to their survival and evolution, and
(3) the implications of the fact that the ability to know is an
evolved biological trait.
On the biological side, the issues include
(1) the recognition that biological systems are knowledge systems,

(2) the recognition that evolution itself is a process of gaining
knowledge, and
(3) the search for the features allegedly shared in common by all
forms of knowledge gain.
I should remark in passing that Plotkin's formulation of the problem is

colored by his view that evolutionary theory must provide a complete
account of all aspects of biological systems. The axe he is grinding here is
the “inadequacy” of the “synthetic” theory of evolution, as it now stands,
to do this. Thus, the articles that are collected in Plotkin (1982) are
addressed, for the most part, to the biological issues (although not entirely,
e.g. Hull 1982, Lewontin 1982). Also, I want to draw your attention to
what appears to be a curious reversal in the basic plan of the program. As
originally conceived, it was the straightforward application of evolutionary
considerations to the development of the biological structures involved in
cognitive activity. In Plotkin (and Campbell and Popper as well), an
understanding of how knowing works is being used to try to analyze the
fundamental adaptive processes of the evolution of all biological struc-
tures. I shall have more to say about this question in section III, where we
consider whatis serving as metaphor for what.
Sir Karl Popper is a well known advocate of both programs in evolu-

tionary epistemology. In a very recent paper, Popper has reduced his view
on evolutionary epistemology to five theses. The first thesis is an endorse-
ment of the EEM program:
The specifically human ability to know, and also the ability to produce scientific
knowledge, arc the results of natural selection. They arc closely connected with the
evolution of a specifically human language. This first thesis is almost trivial. (Popper
1984, p. 239).
Perhaps what helps save the first thesis from utter triviality is the
extravagant ontology which Popper has constructed upon this premise in
the form of a series of interacting independent realities. I take it that we
can all agree with Popper's first thesis without being thereby committed to
a Popperian trialism, or even dualism for that matter. However, even so
we might wish to pause and reconsider. Given the well known fact that
natural selection is not the only driving force in evolution, Popper's thesis,
far from being trivial, might, in fact, be false. Sober has explored a similar
theme in his 1981 paper on the evolution of rationality. Be that as it may,
we may use this opportunity to remind ourselves that ‘biologically evolved
and 'biologically evolved through means of natural selection' are by no
means equivalent. This, in turn, implies there might (must?) be more to
what drives evolutionary epistemology than natural selection.
Tn his ‘Reply to My Critics', Popper notes some further consequences of
evolutionary theory. From the fact that man is an animal and that animal
senses have evolved from primitive beginnings, it follows, Popper thinks,
that human knowledge is almost as fallible as animal knowledge, and that
human senses, like animal senses, are part of a “decoding mechanism”
(Schilpp 1974, p. 1059). Elsewhere, in the same volume, he remarks that
our sense organs have many subtle decoding and interpreting devices built into them —
that is, adaptations, or theories [!].... These are ... conjectures[!] (Schilpp 1974,
p. 111)
I draw your attention to this passage because it illustrates the sense in
which Popper is a fellow traveller with those who seek to interpret
biological structures in the light of our understanding of cognition.
Donald Campbell, with whose views Popper has expressed “almost
complete agreement,” (Schilpp 1974, p. 1059), has developed an evolu-
tionary approach to epistemology in great detail. In his masterful survey
of this subject, which I will not attempt to summarize here, Campbell
endorses a number of points which are characteristic of the EEM pro-
gram. In particular, he notes, with approval, Lorenz's biologizing of Kant
and the implication that the categories, etc., are to be read ‘descriptively’
and not “prescriptively” He also advocates the view that
- evolution — even in its biological aspects — is a knowledge process, and ... the
natural-selection paradigm for such knowledge increments can be generalized to other
epistemic activities, such as learning, thought and science. (Schilpp 1974, p. 412)
406 MICHAEL BRADIE
In a 1977 paper, he sees himself in agreement with Richards (1977) in

holding that the categories of perception and thought have a biological
evolutionary origin (via mutation and selection) (Campbell 1977, p. 506).
It is well known that Campbell is among the hardliners in evolutionary
epistemology in arguing for the applicability of a ‘blind-selection-and-
retention’ model to explain not only the evolution of all biological
structures (and not merely cognitive ones) but also the growth of scientific
knowledge which is more properly viewed as part of the complementary
EET program.
We could continue in this vein for some time, but let me bring this part
of the discussion to a close with one final example. Richard Blackwell, in a
1973 paper, has developed an evolutionary epistemology which is distinct-
ly Piagetian. Piaget himself deserves consideration under the program, and
in a fuller treatment will be accorded such, but Blackwell will serve to
illustrate the general tenor of their views. The underlying theme of what he
calls the adaptational model of knowingis that
the processes of human knowing as well as the objects of cognition constructed by them
are ultimately to be understood as the instruments of evolutionary adaptation between
man and environment. (Blackwell 1973b, p. 334)
Also,
The adaptational model considers cognition as an extension of the pre-cognitive realm
of evolution with both realms governed by the same generic types of laws and
processes. (Blackwell 1973b, p. 334)
Finally, Blackwell argues that appeals to evolutionary evidence indicate
that there is a more or less progressive development of cognitive powers
from the lower animals to human beings (Blackwell 1973b, p. 321; cf.
Ruse 1984 for a similar sentiment).
Blackwell's model, and Piaget's, as far as [ understand it, introduce the
possibility that feedback mechanisms allow for environmental intervention
in the genome d /a Waddington's ‘genetic assimilation’ as a contributing
factor in the evolutionary development of cognitive mechanisms. Of
course, the standard Darwinian response is that such effects insofar as
they are real can be understood, in principle, using natural selection
models. I am not going to try to resolve the issue here on way or another.
Suffice it to say that several of the papers in Plotkin (1982) take such
possibilities seriously.
Let us turn to a brief consideration of models proposed under the EET
program. As I suggested earlier, the two programs are interrelated and one
often, but not always, finds the same authors arguing for both. Thus,
Lorenz (1977) endorses Campbell's extension of the natural selection
paradigm to thinking, learning and the development of science.
The method of the genome, perpetually making experiments, matching their results
against reality, and retaining what is fittest, differs from that adopted by man in his
scientific quest for knowledge in only one respect, and that not a vital one, namely that
the genome learns only from its successes, whereas man learns also from his failures.
(Lorenz 1977, p. 24)
Although this is somewhat outside the scope of this paper, note that
Lorenz is of a mind to extend the analysis to the analysis of cultural
change in general.
.. the parallels between [the] . . . historical development [of such cultural products as
the railroad coach and military uniforms| and the phylogenetic evolution of organs
make one suspect that analogous forces are at work — in particular, that it is natural
selection and not rational planning which is the dominant factor. (Lorenz 1977, p. 235)
Popper's views on this are well known. The second thesis of his version of
evolutionary epistemology in his 1984 paper is
The evolution of scientific knowledge is, in the main, the evolution of better theories.
This is, again, a Darwinian process. The theories become better adapted through
natural selection: they give us better and better information about reality (they get
nearer and nearer to the truth). All organisms are problem solvers: problems arise
together with life. (Popper 1984, p. 239)
In the section of his autobiography entitled Darwinism as a Metaphysical
Research Program, Popper claims that the method developed in his Logic
of Scientific Discovery is a method of Darwinian selection as opposed to
what he calls “Lamarckian instruction” (Schilpp 1974, p. 133). T won't
pause to sort this out here, except to remark that when Popper wrote this,
he believed that Darwin's theory of natural selection was not a testable
scientific hypothesis but was instead a metaphysical (i.e. untestable but
nonetheless viable) research program. He has since recanted on biological
Darwinism (Popper 1979). I am not sure what the implications of that
reversal are for his understanding of epistemological Darwinism (cf.
Schilpp 1974, p. 135).
Campbell, in his contribution to the Schilpp volume on Popper, en-
dorses Popper's treatment of the succession of theories in science as due
to a selective elimination process analogous to the eliminative role of
natural selection in biological evolution. In addition, trial and error
learning by animals, including man, brings the evolutionary model to the
ontogenesis of knowledge (Schilpp 1974, p. 415f). More on this below. In
his 1977 paper, Campbell indicates again that the view that competition
among scientific theories is analogous to natural selection is a second
fundamental point on which he and Richards (1977) are in agreement
(Campbell 1977, p. 506).
The core thesis of Stephen Toulmin's provocative Human Understand-
ingis a commitment to what Toulmin considers a form of epistemological
Darwinism.
Darwin's populational theory of variation and natural selection” is one illustration of a
408 MICHAEL BRADIE
more general form of historical explanation: and . this same pattern is applicable also,
on appropriate conditions, to historical entiti and populations of other kinds.
(Touimin 1972, p. 135)
In an earlier paper, Toulmin developed a preliminary version of this view,
and argued that science develops in a two-step process analogous to
biological evolution. At each stage in the historical development of
science, a pool of competing intellectual variants exists along with a
selection process which determines which variants survive and which die
out (Toulmin 1967, p. 465). Despite intense criticism of both the general
and specific features, Toulmin was still defending the general model as late
as 1981. However, it must be pointed out that Human Understanding was
the first volume in a projected trilogy. Neither of the planned subsequent
volumes have yet appeared.
One philosopher who is working on a Toulminesque project is David
Hull (1982). Instead of arguing by analogy from biology to culture or
attempting a literal extension of biology to culture, Hull prefers to develop
a general analysis of “evolution through selection processes which applies
equally to biological, social and conceptual evolution” (Hull 1982, p. 275).
Finally, consider Rescher's methodological turn.
. we shall deal with the role of biology-analogous selection phenomena in the

development of the materials of human thinking. This epistemological approach
envisages the operation of evolution-like processes in the historical development of the
substantive content of man's thinking, and does not limit its perspective to the positive
role of the capability of thought in biological evolution [ie., EEM]. The evolution of
thoughts rather than that of thinkers is the issue. The program of “evolutionary
epistemology” as standardly conceived envisages the development and transmission of
man's beliefs, ideas, and theories as itself representing an evolution-like process. Our
present theory endeavors to give a methodological twist to this idea. (Rescher 1977, p.
128)
This brief and by no means exhaustive survey should serve to illustrate
that there are, in fact, two quite distinct programs parading under the
name “evolutionary epistemology” which are, nevertheless, interrelated.
They are not, however, identical. I can well imagine that the EEM
program will turn out to be true, while the EET program may never turn
out anything more than programmatic notes. Indeed, there is a sense in
which some version of the EEM program must be true if our current
understanding of evolutionary process is anywhere near correct. What
remains to be seen is what useful insights, if any, will be forthcoming from
it about the evolution of the cognitive mechanisms of organisms. A further
question is what, if anything, any such results have to do with epistem-
ology, either narrowly or broadly conceived. I return to this latter question
in section IV below.
Given the distinctness of the EEM and EET programs, how are they
connected? E.O. Wilson (1975) urges that
The biologist, who is concerned with questions of physiology and evolutionary history,
realizes that self-knowledge is constrained and shaped by the emotional control centers
in the hypothalamus ard limbic system of the brain. These centers flood our con-
sciousness with all the emotions — hate, love, guilt, fear, and others — that are
consulted by ethical philosophers who wish to intuit the standards of good and evil.
What, we are then compelled to ask, made the hypothalamus and limbic system? They
evolved by natural selection. That simple biological statement must be pursued to
explain ethics and ethical philosophers, if not epistemology and epistemologists, at all
depths.
One apparently plausible route connecting the two programs lies in the
attempt to biologize the Kantian categories. There is a slippery slope
leading from a central part of the EEM program (the attempt to under-
stand the phylogenetic development of the biologically material cognitive
apparatus) through the claim that all organisms and lineages have “built
in” specific cognitive apparatus characteristic of their place in the phylo-
genetic tree to the claim that each organism has its own characteristic “a
priori” Kantian categories and finally to the central claim of the EET
program, viz., that the content of knowledge as shaped, in part, by the
“a priori” categories itself undergoes some form of evolutionary develop-
ment. Schematically
evolution of biological substrate EEM. evolution of brain

EEM . N M . 1. _EEM
—EEM. evolution of mind FE evolution of categories ———
evolution of human knowledge
T have not found this line of reasoning explicitly spelled out anywhere
but it seems like a seductive path to follow and may well be the basis of
the confidence, in those who believe it, that the evolutionary model works
both at the level of material mechanisms and cognitive content. The
crucial link is the biologizing of the categories for it is at this stage that
they assume their normative force. Seen from the perspective of the EEM
program, the categories are just another feature among many of highly
developed organisms. Seen from the perspective of the EET program, the
categories are not only descriptive of how organisms cognize and evaluate,
but prescriptive as well. Thus, if the train of thought as outlined above
were to work, then one would have a “biological” analysis and explanation
of some normative features of the behavior of organisms. While this line of
reasoning may account for why some people may think that an evolu-
tionary explanation of the brain and its functions means there is an
evolutionary explanation of knowledge, it is not logically persuasive.
An even more blatantly fallacious argument linking the two programs
can be found in Wuketits (1984b, p. 8f). He argues as follows:
Since the human mind is a product of evolution [an EEM claim| — and any opposite
view such as that of classical daulism means a kind of 'obscurantism' — the evolutionary
410 MICHAEL BRADIE
approach can be extended to the products of mind, thatis to say to epistemic activities
such as science [an EET claim|
In support of this contention, Wuketits cites the obvious fact that science
has changed over the centuries. From this he concludes that it is evident
that science “has undergone many changes and intricate developmental
processes . .. history of science means evolution of science.” Perhaps so, if
by “evolution” one means any change whatsoever. But the crucial ques-
tion, which is implicitly begged in Wuketits' argument, is whether the
evolutionary processes and mechanisms which gave rise to the mind
(EEM mechanisms) are the same or significantly similar or analogous to
the processes and mechanisms which underlie the changing content of
accepted human knowledge (EET mechanisms). The resolution of the
question of the connection between the two programs hinges on a detailed
analysis of the EEM program, which is beyond the scope of this paper. 1
hope to deal with these issues on a later occasion.
At this point, I want only to mention that, as might be expected when
philosophers dispute about favorite themes, there are questions about how
serious each of these programs are and which deserve to be considered
“genuine” evolutionary epistemology.
Popper, Campbell, Toulmin, Hull and Rescher, among others, clearly
take EET seriously but others do not. Mario Bunge, for one, thinks that
the attempt to develop analogies between biological evolution and the
history of ideas is ill-conceived. It passes for evolutionary epistemology,
but is not (Bunge 1973, p. 58). The analogies, he alleges, are superficial,
and the disanalogies loom large (cf. section III for further discussion of
these claims). For Bunge,
Genuine evolutionary epistemology takes organic evolution seriously, deals with the
evolution of cognitive abilities as an aspect of brain evolution, and takes the social
matrix into account. (Bunge 1983, p. 59)
EEM programs are genuine, EET programs are not. Skagestad (1978)
agrees. Evolutionary epistemology, he claims, is not advanced by looking
for analogies between biological evolution and knowledge processes.
Taking evolution seriously, he contends, involves recognizing that humans
are animals, as such subject to biological evolution, who have generated a
novel means of evolution — cultural evolution (including the evolution of
science) — which proceeds by mechanisms other than natural selection
(Skagestad 1978, p. 620). As he sees it,
The crucial cuestion of evolutionary epistemology is the question of how evolution by
natural selection was able to generate, in one biological species, a mode of evolution
not operating through natural selection, and yet contributing to the survival of the
species in question. (Skagestad 1978, p. 620)
Even Michael Ruse, who has been a persistent critic of the EET
programs of Popper and Toulmin, has recently come around to the view
that “Epistemology needs Darwinism” (Ruse 1986). Ruse thinks that a

sociobiologically inspired evolutionary epistemology may shed light on
some traditional epistemological problems. I take this as an endorsement
of some version of the EEM program as T have characterized it. See
Vollmer (1984) for a discussion of some of the problems allegedly solved
by evolutionary epistemology developed along the lines of an EEM pro-
gram.
My own viev is that both programs are worth pursuing, if only for the
reason that by recasting old problems in new ways we cannot help but
learn something of interest whether we succeed or fail. In this vein, an
appropriate way to bring this discussion to a close is with a point made by
Tennessen about knowledge versus survival. In endorsing the view of
Peter Zapffe, Tennessen summarizes his position as follows:
Man's in principle unlimited capacity for insights and fore-sights — his ability to find
out how things around him and in him operate, which is undoubtedly responsible for
his present privileged place in the scheme of organic creation — this capacity, when
extrapolated beyond its biological efficacy, may yield the most horrifying, vertiginously
pernicious, unendurable insights into the fatuous futility and monstrous absurdity of the
totality of human existence. This uniquely human ability to 'stand out' ('ek-sistere”) and,
while still breathing, examine his own ‘total situation' is, as Zapffe sees it, an insanely
haphazard 'short-circuit' in nature, a prerogative with which no other being has hitherto
been blessed or cursed. Man's choice is: either to abdicate from his humanness, give up
the prerogative of unlimited knowing, his intellectual honesty, his search for truth, his
demand for order, justice, meaning ... in short, to give up all that makes man
essentially different from a (happy) pig; or: to face his fate viz. his cameo-appearance in
that wild, banal, grotesque, and loathsome carnival in the world's graveyard which life
is. But were man to choose the latter alternative, the only decent and dignified
response, that Zapffe sees, is for man to choose to die out completely, to deliberately
leave the Earth deserted behind him . . . . (Tennessen 1973, p. 408)
2. T want to briefly address the question of the application of the evolu-

tionary model to what corresponds to ontogenetic processes of knowledge
growth as well as phylogenetic processes. Campbell, as was noted earlier,
endorses Popper's “extension” of the trial and error model to include
learning processes by individual organisms. In ‘Of Clouds and Clocks,’
Popper had claimed that organisms as well as phyla were “problem
solvers.” In fact, he puts forward there the curious analogy that as the
actions of an individual organism are tentative solutions to problems faced
by it in its environment, so individual organisms are tentative solutions to
problems faced by the phylum of which they are members (Popper 1972,
p. 243). In another paper, while arguing that science is the extension of
common sense, he draws the connection between the two in the following
way. From an ontogenetic point of view, scientific explanations rest on the
expectations of the newborn child. That is, the child grows into critical
adulthood by the well known Popperian process of conjecture and refuta-
412 MICHAEL BRADIE
tion, with the inital conjectures formed on the basis of innate expecta-
tions. These innate expectations are the result of evolutionary develop-
ment. So, from a phylogenetic point of view, today's science rests on the
“expectations” of ancestral unicellular organisms. This is epitomized by the
quip that “There is, as it were, only one step from the [ancestral] amoeba
to Einstein” (Popper 1972, p. 347; compare the similar sentiment express-
ed in Campbell's pre-Popperian 1960). This way of putting the point
mixes the two programs. The phylogenetic development of expectations in
organisms in a lineage is a question appropriate to EEM. The ontogenetic
development of knowledge in an individual (as opposed to the ontogenetic
development of the biological structures necessary for the individual to
become a competent critical adult) is a question in the EET program. The
corresponding phylogenetic EET question concerns the historical evolu-
tion of science from, say, Aristotle to Einstein. Toulmin characterizes the
distinction within the EET program in a clearer way.
We ... face questions about the social, cultural, and intellectual changes that are
responsible for the historical evolution of our various modes of life and thought — our
institutions, our concepts, and our other practical procedures. (These questions
correspond to questions about phylogeny in evolutionary biology.) Individually
speaking, we .. . face questions about the manner in which maturation and experience,
socialization end enculturation shape the young child's capacities for rational thought
and action — how the child comes to participate in his native society and culture.
(These questions correspond to the questions about ontogeny in developmental
biology.) (Toulmin 1981, p. 26)
Restricting our attention to the growth of knowledge, it is clear that
Popper and Toulmin both endorse selectionist models of both processes
(although their models are significantly different). Hull holds a similar
view (Hull 1982, pp. 304ff).
1 do not want to stop to evaluate these programs here, but merely wish
to point out that, at the biological level, phylogenetic processes seem quite
different from ontogenetic processes. I realize that we know very little
about ontogenetic processes in biology (and, some might say, not much
more about phylogenetic ones except in general outline) but the tradition,
I take it, is to treat them as distinct. In the pre-Darwin days, when people
thought they knew more about ontogeny, the tendency was to treat
evolutionary processes as basically more of the same but on a larger time
scale. Darwin put an end to that. Now we see the trend reversing, and
some people are beginning to try to understand ontogenetic processes in
terms of models borrowed from phylogenetic considerations. [ don't know
whether these efforts will prove fruitful or not. The point I want to make is
that there is, at the present, a prima facie difference between the two sorts
of biological processes. To the extent, therefore, that we are inclined to
use biology as our guide to understand knowledge in all its aspects we
should be somewhat wary of the glib assumption that the phylogenesis of
knowledge and the ontogenesis of knowledge should be amenable to the

same model adapted from phylogenetic considerations in biology.
3. To close this section, we should take note of yet a third distinction.

Many have argued that neither program, and certainly not the EEM
program, has anything at all to do with epistemology, properly under-
stood. The basis for that contention is that epistemology, properly under-
stood, is a normative discipline, whereas the EEM and EET programs are
concerned with the construction of causal and genetic (i.e., factual) models.
No such models, it is alleged, can have anything important to contribute to
normative epistemology. We shall return to this question in section IV in
great detail. Here [ only mention that, outside of trying to derive norms
from facts (which is what sociobiologists are sometimes trying to do), the
standard way out is to distinguish between descriptive and normative
epistemology and argue that evolutionary epistemology is relevant to the
former but not the latter. To the objection that “descriptive epistemology”
has nothing to do with epistemology, we shall also return in section IV.
For now, consider Riedl's characterization of evolutionary epistemology.
In contrast to the various philosophical epistemologies, evolutionary epistemology
attempts to investigate the mechanism of cognition from the point of view of its
phylogeny. It is mainly distinguished from the traditional positions in that it adopts a
point of view outside the subject and examines different cognitive mechanisms
comparatively. It is thus able to present objectively a series of problems [including the
problems of traditional epistemologies] that are not soluble on the level of reason alone
[but, which are soluble from the phylogenetic point of view] . . . (Ried] 1984, p. 220)
In effect, Riedl is claiming that evolutionary epistemology, as he under-
stanás it, is capable of solving traditional problems of reason and epistem-
ology by appealing to “ultimate” evolutionary explanations rather than
“proximate” traditional analyses.
On the other hand, Hull (1982), in addressing the issue at hand, agrees
“with the critics that a purely descriptive epistemology is 'epistemology' in
name only” (Hull 1982, p. 273). He does not, however, take this as a
significant criticism of Campbells EET program but rather as an indi-
cation that such efforts, his own included, should be seen not as attempts to
understand epistemology from an evolutionary point of view, but rather to
produce a “scientific theory of socio-cultural evolution.” With this we have
come up against the constraining boundaries of the scope of this paper
and so it is time to turn to other things.
III. EXPLORING THE METAPHOR
In this section I want to develop in some greater detail the constraints and
considerations that have been thought relevant in constructing an evolu-
414 MICHAEL BRADIE
tionary model of conceptual development along biological lines. The

projects we will be here considering, thus, belong to the EET program.
1. The Making of the Analogy
Campbell (1960, p. 381) argues that underlying both trial and error
problem solving and natural selection in evolution is a general model for
“inductive gains.” He sets three conditions for such models. They must
incorporate (1) “a mechanism for introducing variation,” (2) “a consistent
selection process,” and (3) a preservation and reproduction mechanism.
Bechtel (1984) echoes these views. An evolutionary model of conceptual
change must find analogies for biological variation, selection and retention.
Toulmin argues that the key feature of evolutionary models of conceptual
change based on Darwinian biology as opposed to evolutionary models
not so based is that the Darwinian models are “populational” rather than
“providential” (Toulmin 1972, pp. 322, 325f). As such, the “only feature
common to all populational changes is, precisely, the general form of
“... [the] ... dual process of variation and selection” (Toulmin 1972,
p.337). For Toulmin,
What links the historical development of intellectual disciplines to populational

processes of other kinds is ... no specifically biological analogy, but simply the general
pattern of development by innovation and selection. (Toulmin 1972, p. 141)
The general consensus seems to be that any biologically derived
evolutionary model of conceptual change must exhibit at least these fea-
tures. It left open whether such models are required to find specific analo-
gies for all the key concepts in evolutionary biology, ie. “gene,”
“genotype/phenotype,” “organism,” “species,” and the like. (See below for
a discussion of some of these.)
The question then is whether the analogy is superficial (as argued,
e.g., by Thagard 1980, p. 187, Kary 1982, Lewontin 1982, Bunge 1983,
p. 58), or whether it is to be taken seriously. In effect, is “evolutionary”
epistemology a “mere” metaphor? Bechtel (1984, 315f) seems to argue
that it does not much matter. The point to making such analogies, he
claims, is not to prove that conceptual evolution has occurred one way or
another, but merely to serve as device for focussing attention on con-
ceptual change by raising interesting questions. As such, substantial
disanalogies, even if they exist (as many claim they do), are not necessarily
disabling. I, for one, am sympathetic with this point of view. Cohen (1973,
p. 49), however, in criticizing Toulmin's 1972 model argues that if, as he
claims, the parallelism between Darwinian evolution and conceptual
evolution is not close, then Toulmin's “Darwinian” analysis is “nothing but
metaphor and the light it sheds is rather a dim one.”
Toulmin (1967) and Popper (1972) disagree. In his 1967 paper,

Toulmin says
In talking about the development of natural science as “evolutionary.” T have not been
employing a mere facon de parler, or analogy, or metaphor. The idea that the historical
changes by which scientific thought develops frequently follow an "evolutionary”
pattern needs to be taxen quite seriously; and the implications of such a pattern of
change can be, not merely suggestive, but explanatory. (Toulmin 1967, p. 470)
Toulmin's (1981) current view seems similar. What the alleged explana-
tory virtues of the model are we shall see in section IV. Here [ only want
to note my reservation about Toulmin's lumping together facons de parter,
metaphors and analogies as all, somehow, not serious, merely suggestive
and non-explanatory. While facons de parler may, by definition, be so
construed, a strong case can be made out for the explanatory role of
models and analogies (see, e.g., Hesse 1966, Ruse 1973a, 1973b, McMullin
1974, 1976, and Bradie 1984a, 1984b).
Popper (1977) argues as follows:
While animal knowledge and pre-scientific knowledge grow mainly through the
climination of those holding the unfit hypotheses, scientific criticism often makes our
theories perish in our stead, eliminating our mistaken beliefs before such beliefs lead to
our own elimination.
This statement of the situation is . . . not meant metaphorically, though of course it
makes use of metaphors. The theory of knowledge which I wish to propose is a largely
Darwinian theory of the growth of knowledge. From the amoeba to Einstein, the
growth of knowledge is always the same: we try to solve our problems, and to obtain,
by a process of elimination, something approaching adequacy in our tentative solutions.
(Popper 1972, p. 261)
The problem with Popper's program is that, despite his longstanding
commitment to it, it remains programmatic and, in details, somewhat
incoherent. The central general claims (1) that natural selection is an
example of a trial and error process and (2) evolutionary processes can be
viewed (metaphorically!) as problem solving processes vill be examined
below. Here I address two points of detail. Recall that Popper extends the
trial and error model to include both ontogenetic and phylogenetic
processes of knowledge growth. He then suggests that
The tentative solutions which animals and plants incorporate into their anatomy and
their behavior are biological analogues of theories; and vice versa. Theories correspond
to endosomatic organs [my emphasis] and their ways of working. (Popper 1972, p. 145)
The problem with identifying organs as the analogues of theories is
precisely that mentioned earlier, namely, that there is a prima facie
difference between the processes involved in the phylogenetic evolution of
biological organs and their ontogenetic development. As such, although in
one sense theories may be construed as exosomatic “organs,” it is not clear
what we are entitled to infer about their development on the basis of this
analogy.
416 MICHAEL BRADIE
A related point involves Popper's use of the branching tree (phyloge-

netic) model of evolutionary biology to characterize the growth of knowl-
edge as similarly evolutionary (Popper 1972, p. 262). The problem, which
Popper recognizes but which does not deter him, is that the two trees are
mirror images cf one another. Whereas the phylogenetic tree of biological
growth starts from a single branch and becomes a bush, the growth of
pure knowledge, on Poppers view, is inverted, with many branches
converging on a single unifying, all-encompassing theory. For the moment,
merely note how this begs some important epistemological questions
concerning the convergence of knowledge and the (implicit) commitment
to the idea of global scientific “progress.”
Popper does allow that the “evolutionary tree of our tools and instru-
ments” follows the pattern of the biological tree, e.g. from a single stone
branching into specialized forms. But, surely, this too is wrong. Not all
tools can be traced back to a single origin, even if all hacking tools can be.
The use of fire, the discovery of the wheel, and the use of water power
(perhaps) represent distinct sources. Rather than a tree or bush, a better
metaphor would be a wire fence with many initial points interconnecting
with other points through the crisscrossing of the wires.
If two quasi-trees were not enough, Popper allows for a third “tree of
applied knowledge,” allegedly similar in structure to the tool/biology tree.
From one applied piece of information, many branching applications can
be expected. But, again, surely the more appropriate analogue is a “tool-
fence” not a “biological tree.” The serious point underlying these images is
that the extent to which the structural development of tools, pure knowl-
edge or applied knowledge is not the same as biological development
counts against thinking there is any deep relationship between the pro-
cesses by means of which they develop or evolve.
In addition, throughout his discussion, Popper shifts back and forth
between phylogenetic and ontogenetic considerations using first phyloge-
netic arguments to bolster the ontogenetic analogy and then ontogenetic
considerations to bolster the phylogenetic analogy. Note, also, that even on
Popper's own grounds, no tree or tool use or applied knowledge vis d vis
individual developmentis going to be an adequate model, since an organism
is born with a serof dispositions (each of which will be a distinct “root”).
It is not my intention to offer an extended discussion of all of the
proposed evolutionary models of conceptual change. Instead, I offer the
following chart, sketchy as itis, to indicate some of the various alternatives
Which have been proposed. The list is by no means complete and reflects
just those schemas with which I am familiar. In addition, I have not spelled
out all the analogical details that can be found in many of them.
In constructing an analogy between evolution in biology and the evolu-
tion of science or knowledge in general, one should, it seems, identify
appropriate epistemological analogues for key biological concepts. Thus.
we might expect to find epistemic analogues for “organism,” “species,”

“population,” “variation,” “mutation,” “drift,” “adaptation,” “environment,
“selection,” “phenotype,” “genotype,” “fitness,” as well as some clear
indication of intended transmission mechanisms, hereditary principles and
criteria of success. None of the evolutionary models of conceptual change
with which 1 am familiar provides this much detail.
is a
All evolutionary models seem to have three components: (1) there
there are
source of variation; (2) there is a selection mechanism; (3)
mechanisms for the transmission and retention of information. However,
not all of the surveyed models represented in Table 1 supply even this
minimal characterization.
2. Critique of the Models
(i) Common Problems

In any early statement of his position, Campbell (1960) noted five
common problems shared by the ‘blind variation and selective retention’
theory of conceptual development and the biological theory of evolution.
1. The basic insight, evolution by natural selection, is tautological
(Campbell 1960, p. 396). Popper, before he changed his mind, was of
the same opinion (Schilpp 1974, p. 134; 1972, p. 69f).T take it that
it has been well established that this objection is without force. The
theory of evolution by natural selection may be many things, but it is not
tautological.
2. There is a need for an explicit account of the source of variations
which would enable us “to predict the generation of heterogeneous
thought trials.” Something akin to the “detailed deterministic explanations
of the mutation of genes” is needed (Campbell 1960, p. 396). In the
intervening years, the natural selection models have dealt with this to
some degree, but nothing approaching the kind of explanation Campbell
thought necessary is yet forthcoming. We do not yet understand enough
about the psychological and sociological dynamics of discovery.
3, Both involve post hoc reconstructions of past historical lineages; any
general regularities are or will be, at best, probabilistic (Campbell 1960, p.
396). Hull (1982) addresses this issue and tries to spell out just what we
can and cannot expect from a general theory of sociocultural evolution.
4. Both theories suffer from what we would today call an ‘adaptationist
bias”, the production of just so’ stories as Gould has called them (Camp-
bell 1960, p. 397).
5. An independent characterization of the relevant environment within
which changes take place is often absent (Campbell 1960, p. 397).
Toulmin (1972) and Hull (1982) seek to address this issue, but it is a
difficult one. Lewontin (1978) discusses some of the problems with the
biological concept of environment and environment-organism interaction,
(29°d) eu
-o1d, 001 a1 YoM asOUI 10 SONrOSI
204 PP MO Paam 01 oAros YOIYM UOTEA
-19su00 wipered jO suopesApIUO] “7 (9d) sorroay
(197d) Amaeyisrea “T auynuaros jO SUONENUIOJ SANEUIOYE OL6T UeLIOyO y
(L07 ) uray aa
1e9p 0) djay i suonvaouur pesodoxd
24) mOy pue e paanp ae suonsog
-ns ay swejqord ayrads ea smoys
BRADIE
(97 STT ‘161 de -anjono 01 ajgen Antondun vossmasip jO ¡XoIIO o,

18 SOAJSULOUT YIMM sprepuers pepos se mjosur Ao seniigissod oumuas
(91 “d) dsoonuarddr pdnd-1orsmpy PUE [ENIO3[[21UT [€20] ‘[BULIOJIT SROLIEA QW029q YOWM sanjañou - penooperu TL61 UTINOL
(99 d) “yrys suorysey pue sonr
MICHAEL
-femor [BIL0ISIY se UONN[0A? 0) 19afgns

918 ENNO asau[ “uonsonb uy sspuaros
JO Aunumios sy jO sompea oy JO
uoNDayal B ae YA BUALD [200T T
uoneagisiey, pue
(6sbd) spdnd xroy 07 way ‘woneaguaa, “nn se yons suonend
MusUen Vy OyM SISHUAS jO sanmu -se Jo suonoajjar (Ieurapco) reojo “T
-nunos Kq paties o1e seapr oyhuaros snouen (TLp"d) ammiafuoo
201y ¿OGT urwynoy,
uonuaj9y 7y VoIssItUSURI | uonDajs UONeNEA siony
TaTal
418
419
(182 d) swsneyoour
uonajos oanuñoo asayl JO soueiiod
Wi Yy 20pi0 yuel jOU $I0P 19peros
(92 “d) -uoneorununuos ¡eadde - oanimur pue “ssou-voy p
teuosadiarur pue oysuen prdnd-raset -u0u “worsppoxd “Amassmejd “Aronduns
e (zg7 “d) vorfenoo ydnoiyr sinodo “oualayoo pue aouaysuoo “Aypunsas (062
vowsssuer] (/7 ‘) sasoy ome si “uognjosa1 wajgo1d “1amod Anoyeuejdxo d) 10119 pue “uopimur soueyo opnjpur
-uajos pue saysered ale sarioau] "ISO ‘ssooons aanorpald “oddns ¡eouiduro uoneuEA JO $201M0s 2y, "SarDads aAmru
-800 Jo suonvLIvA ome SUOISI9A ÁOMJ 0861 12PENPS
EVOLUTIONAR Y EPISTEMOLOGY
pue dusered jO JEYI SI [apoWI 1Aq YL Suipnppur suonouny aanrudoo sOMEA

(oz
TI7 “de) “estains 9u23, o) ajgronpar
(907 “d) ."uonemur parreo 10u pue 1aya ¡rofoj0ypísd y ed
2q ueo - yomm ssa01d y vi WIVIQ ul “st atuawr e JO ON[BA [BAIAINS YL
o1 ueiq woy Suidesy Aq jood awraur (g07 “d) .:siayio ueya ¡ood-urmaur ay
ay ur saapsway arededod sawau, ur ¡mjssa0ons 210W aY SAWAW JUIOS, (907 “d) souraw sunmed
(g6¢ d) suoneandde 1ny
JO Ánuea 10219 oy Ul ssaumast puy
w sidoouoo pue sarsoosip asoui
jO [eAAIns oyy Áq oouors ur peror
-1 ST SUONIpUOJ 9SIOAPE pUE ISIIAIP
1sour 9yl 9pun JUBLEA E JO [BAIA
-ins oy, "mowuoJALO SujSueyo e 01
(z6¢ “d) “aida0u0o wenod y0si 1depe 01 sowads e siuod yorga
-u jO omeu oanemuenb pue ¡euon ameu ua ur oovajsrro 107 Iy
-e1ado ayrads oy uo paseq si “aduajas auy woy maryIp Apeaid jOU seapy
ur s Ammuos yO adowud y Fuowe 1o1U0D B, st AIOY DS UN (Z6£ ") ssousapusaur uewny £L6T VONOH
uoNuajy 7y UOISSIWSURL, uopajs UONELIZA siouny
(parunmo?) | 1QU7
(L *9 siaideyo)
swstueyoowr - Aeuonnjos3-02 aImind
-9u39 Jo 1s xajdwo e Áq 19u10Ue
0) vonerauad auo wOy ponitusuen a1
sua9minz “woiseaur pue vonesrum
JO japour jeamydesSoaforq y jO saui (£ 11deu9)
ay duoje - Apevonerauadenur paru esrenpiapur yo Supyew vorspop pue
-suen pue suresq uwwnY ul somonns uopEnfeA Jo sonsumay oyl adeys yorga 1861 UOSIIA.
yuN-pou ul papoquia are sua8ininz semi onouadida snid uonoajes [enIEN susfamno pue uopsumy
(rs-d)
swowuonaus Swifueyo Ajsnonunuod
10 SJUIWUOIIAUD 19PIM 0) 109dSal M
Aey pue uewuonauo {eao] 0 (05 d) nusumionana
BRADIE
wadsar yua sonpoedeo Jurajos-Wajqo1g 0u pim SunEop 107 s100) se saMoau, 1861 10A0UNA
(a
65 “dd) yom A 965 “dd) -om Koy yorga uuaa sonru
MICHAEL
uNj UN SIUIUNWIUIOD yHUaros ay -Nuwos 9y7 pue sisyuaros ¡enprarpur JO

pue sistuaros renprarpur aujO 1usumor surewop [emdoouod ay JO XSO SY)
uS urjgosd oy ym oouepiosos Uy Ul parerousf seapr jo SUONEMQUION g T86T SPIEUINY
(££ “d) -aseastp snonsayur ue yo peards
ay o1 snoSojuue Ajesojo st vorsspu
-suen (muozuoy (ps -d) ‘uonessuss (1SE
awes ay) JO sIaquiaur 0XI AUT UD9ANAd d) say uwrmaueg Jo sulo) ul oy
uorsstwsuen ¡euOZHOy pue Suldsgo -9ajas permeu snjd sjenprarpur £q suors (se 1861 UeLPI:4
oj juamd wos uorsstusurn [2MIAA -19ap saAjoAur SNA vonoojos jerMO d) 'sioa Suikdoo pue suoneaouu] pue e7103S-NTARJ
(voza)
“Sunapow pue (Z67 :d) voneminoug (20£ d) "uonoeyspes 10 ¡uawoo10jUI2y (16 d) "uonoz pue ninowyp Jo sQeH 0861 uoiBue
UOHURIRY % UOISSIUSUEL], uonaaleg voneLEA s1ogny
420
(panumuoo) 7 1407
421
vopearumuioj sainpaoo1d Junsa y, S861 INH

Gyrsz -de)
“(29pajmouy z-pprom) iuapms o) jusp
-ms Wiy BuuIwd| BIA porqusues pue
EPISTEMOLOGY
(a8papmouy E-puom sipzo 1addog 1eym)

219 “sixo) Ul payndalqo 'siviaq uewny (orz-d)
Áq peiear ae ‘sejoure se ‘sauoay], “Aupiqeujisres pue Ay Surajos wojgo1d {ovz d) souoayy, 861 19ddog
(15¢ “d) -Murquidasoe o ooy
aanoalgo ue opiod Jutajosa 2q 0) jou
parwnsse 918 Yary Smeu JO sMEL YL
1ey) Jurensuoo ) 01 199fqns (yETE "dd)
afueyo SANUNWWOD USRS JO $IS
EVOLUTIONARY
-191ur pue spaau ay) se aajona 01 190fqns (STE-9)

ae yomm sanpedeo Juiajos wajgorg “seapt pro JO sSudno19as pue seapt moN V86T PINPI
(£ T£ dd) ‘[jom se anjono siurens
-09 asay) “saajosa Suipueisiopun yn
“uonea -uaps sy “Munumioo oypuaras y) Áq
-nuwwoo [zuosiadsur pue pdnd-1oisejar paujensuoo Ajjeaoy si ssa1do1d oypuars (pLz “d) suorsioa AroauL 861 INH
uopuajy Y VoNSSIUSUBLL uon2aps UONELEA siouny
(pamuunoo) [ ajguL
422 MICHAEL BRADIE
and Lewontin (1982) considers some alleged analogous difficulties in

conceptual evolution. Again these are issues on which more work in the
psychology and sociology of science is needed.
(ii) Blind Variation

The feature of Campbell's model which has come under most attack
is his insistence that the variation in conceptual evolution is blind or
unjustified. That it is unjustified, ie., that there is no before the fact
guarantee that any of our conjectures will in fact do the job they are
expected to do, is palatable. That the variation is blind, i.e. that our
conjectures are not made in response to some pressure from the problem
environment, is less palatable. Bunge (1983, p. 58), siding with Piaget and
against Campbell, holds that
The search for knowledge and possession of it modifies behavior, and behavior is not
only an outcome of evolution but also a motor of it.
Campbell himself notes that Simon (1969, p. 95) rejects the blind trial and
error point of view. Simon argues
The more difficult and novel the problem, the greater is likely to be the amount of trial
and error required to find a solution. At the same time, the trial and error is not
random or blind; it is, in fact, highly selective. (Campbell 1974, p. 430)
Campbell attempts to rebut this line of attack by arguing that what
selectivity exists is the result of phylogenetically prior blind trial and error
searches whose results have been codified in biological structures, mem-
ory, habit, instinct, and so on. Skagestad (1978, pp. 613ff) considers these
lines of defense and finds them wanting.
Even so, Skagestad argues the analogy breaks down on the blindness
issue because
no scientific progress would be possible if hypotheses were to be proposed ‘blindly’,
.. [that is]... independently of the problems for which they eventually provide
solutions, the way biological variations are independent of the adaptive needs of the
organism.
Peircean worries about the possibility of getting any program off the
ground if there are infinitely many initially equaliy plausible conjectures
at each choice point are the grounds for Skagestad's criticism. Rescher,
in criticizing Popper's commitment to the same idea of progress through
blind trial and error conjectures and refutation echoes similar Peircean
concerns (Rescher 1977, p. 162). Rescher concludes that the crucial
disanalogy is that the trial and error search for these is not
a blind groping amongst all conceivable alternatives, but ... [is] ... a carefully guided
search among the really promising alternatives. (Rescher 1977, p. 157)
Rescher's solution is to abandon what he calls “thesis Darwinism,” the
view that individual hypotheses are subject to blind trial and error selec-
tion in favor of “methodological Darwinism,” the view that methods are

subject to blind trial and error selection. In this way
one can imbed the *vitalistic” features of epistemological evolution at the thesis level
within an orthodoxy randomized and blindly unguided evolutionary model at the
methodological level. (Rescher 1977, p. 157)
Hull, however, objects to Rescher's initial point as misguided. Biological
evolution is not a completely random trial and error process, if by
“completely random” one means that each phylogenetic change is one
of an infinite number of equally likely possibilities. Selection in biology,
he points out, “takes place only among the actual combinations which
appear .. .” (Hull 1982, p. 307).
The crucial disanalogy for Hull concerns the role of intentions (Hull
1982, pp. 307, 312—322; cf. Hull 1985). In this vein, Thagard claims that
the lack of intentional design in nature, which is the great virtue of the
Darwinian model, is the great flaw in evolutionary epistemology (Thagard
1980, p. 188). Campbell (1960, p. 396) sees the issue differently. The
great virtue of evolutionary epistemology, at least as he saw it then, was
that, like Darwin's model, it could show how apparent purpose and design
could be the result of arandom and blind variation and selection regime.
Hull’s tack is not to deny that the development of science involves
intentional considerations but to deny that the difference between inten-
tional and non-intentional activity is as great as some think. Thus, Elster
(1979, 1983) has argued that there is an “unbridgeable gap” between
biological systems, which are to be explained in functional terms, and the
subject matter of the social sciences, such as the growth of knowledge,
which is to be explained in intentional terms. Hull tries to bridge that gap
by arguing that to the extent that scientific communities and conceptual
systems are individuals and historical entities just as are biological species,
their evolution can be explained in terms of mechanisms which are similar
to those at work in the evolution of species. Again, this is not to deny that
the evolution of scientific communities and conceptual systems involves
intentions in a way that the evolution of biological systems does not. The
issue is whether the role of intentions in conceptual evolution makes it
irredeemably different from biological evolution. That is what Hull denies.
It remains, of course, to spell out the details of how this process works
and this Hull does not claim to have accomplished.
Bechtel (1984, p. 316) makes the similar point that a complete analysis
of scientific change “must examine the consequences of intentionality for
creating conceptual variation” But, left untouched, he argues, is “the
crucial similarity upon which the use of an evolutionary model depends,
namely that in both cases variations are proposed without guarantees that
they will be successful” (Bechtel 1984, p. 316). That may be, but some
might argue that the really crucial point is that biological variations arise
424 MICHAEL BRADIE
spontaneously and at random with respect to the needs of organisms,

whereas conjectures and theories, although they may be blind with respect
to their potential success are not randomly generated with respect to the
needs of knowers but are developed in response to a problematic envi-
ronment.
William James, for one, had no trouble even in this regard.
It might ... seem as if the distinction between the agency of “spontaneous variation,” as
the producer of changed forms, and the environment, as their preserver and destroyer,
did not hold in the case of mental progress . . .. But ... [ have no hesitation whatever
in holding firm to the Darwinian distinction even here - [ can easily show that
throughout the whole extent of those mental departments which are highest, which are
characteristically human ... the new conceptions, emotions, and active tendencies
which evolve are originally produced in the shape of random images, fancies, accidental
outbirths of spontaneous variation in the functional activity of the excessively unstable
human brain, which the outer environment simply confirms or refutes, adopts or
rejects, preserves or destroys — selects, in short, just as it selects morphological and
social variations due to molecular accidents ofan analogous sort. (James 1880, p. 456)
Others are less sangunie about this point. Toulmin (1972, p. 328) finds
the source of the apparent design in conceptual and scientific progress in
the fact that conceptual variation and selection are “coupled,” with specific
variations appearing in response to environmental pressures. This con-
trasts with the situation in biological evolution where the source of
variation, genetic mutation, is “uncoupled” from the selection processes
effected by the environment. Toulmin, in effect, abandons “blindness,” but
argues that this does not destroy the validity of the analogy because the
central point is that both conceptual and biological evolution involve some
process of variation and selection (Toulmin 1972, p. 337).
L. J. Cohen (1973, p. 48; 1974, p. 324) thinks otherwise. He sees the
appeal to coupling as a fatal flaw in Toulmin's scheme, fatal, that is,
insofar as it can claim to be properly evolutionary.
(iii) Selective Retention

In biological evolution the name of the game is survival, not of the
individual organism, but of the lineage to which it belongs. The organism
contributes to the survival of the lineage by means of its reproductive
success which is a measure of its fitness. What corresponds to “conceptual
fitness?” Truth? Darwinian fitness is relational and admits of degrees.
Truth does not. Perhaps empirical adequacy is a better analogue. Even so,
empirical adequacy or truth is not the only goal of conceptual activity.
Whereas in biological evolution, there is ultimately only one problem, in
conceptual evolution there are many (Popper 1972, p. 244). The cor-
relative question of how of characterize the selection criteria in concep-
tual evolution is troublesome and difficult (cf. Lewontin 1982, p. 164;
Hull 1973; Suppe 1977, pp. 677ff).
Toulmin argues that two basic qualifications are necessary with respect
to the evolutionary model of human understanding. First
. the goals of adaptation vary substantially from one context to another and from one
enterprise to another. Different arts and sciences, different cultural enterprises and
modes of social organization are adaptive in different respects. In the continuing
competition between rival scientific ideas, the variants that are selected for are,
manifestly, quite other than these that are selected for in the refinement of techniques in
the fine arts or the development of more effective procedures of public administration.
(Toulmin 1981,p. 33)
Secondly, within each of the activities, e.g., in the "continuing competition
between rival scientific ideas,” the “processes and mechanisms of human
adaptation” are of different kinds and include (1) conscious deliberations,
(2) populational processes involving rival variants, and (3) procedures for
encoding the adoptive forms in a manner that will allow for their trans-
mission from one generation to the next.
Popper also is concerned to refute the “view that human knowledge can
only be understood as an instrument in our struggle for survival.” The
fittest hypothesis, he argues, “is the one which best solves the problem it
was designed to solve, and which resists criticism better than competing
hypotheses” (Popper 1972, p. 264). As theorists, “we do not only look for
biological or instrumental success. In science we search for truth” (Popper
1972,p.69).
Richards, in his suggested modification of Campbell's view of “blind
variation and selective retention,” holds that
. one must suppose that selection components operate in accord with certain
essential [as opposed to “contextual”?] criteria: logical consistency, semantic coherence,
standards of verifiability and falsifiability, and observational relevance. (Richards 1981,
p-5)
Neither Toulmin nor Popper nor Richards take these differences to be
telling against the analogy. Skagestad, however, thinks that the analogy
breaks down because biological evolution and the evolution of intellectual
traditions occur in different ways. As he sees it, natural selection works by
the “proliferation of favorable variations at the expense of unfavorable
ones.” On the other hand, “an intellectual tradition guides intellectual
evolution ... by narrowing the range of permissible guesses and excluding
a wide range of guesses, some of which might be true” (Skagestad 1978, p.
615). I do not find this particularly compelling because the “intellectual
tradition” is more akin to the population as which natural selection works
than it is a natural selection itself. That true guesses may be lost in the
shuffle is no more surprising than the fact that “fitter” organisms may not
reproduce in the local environment in which they might have the mis-
fortune to be.
Others who have objected to the legitimacy of “evolutionary” epistem-
ology on the grounds that the selection procedures in the growth of
426 MICHAEL BRADIE
knowledge are different from those which occur in biological evolution

include Dobzhansky (1973) and Crittenden (1977). My inclination is not
to regard that line of objection as, in principle, defeating. Of course, it
remains to establish in detail what the selection mechanisms may be and
how they operate in particular cases. But a similar problem faces biologists
attempting to account for population changes in the field. Carte blanche
appeals to selection are no more satisfactory in the one case than in the
other.
(iv) The Problem of Progress

Almost all critics and defenders alike agree that in one important
respect conceptual evolution differs from biological evolution. In science,
it is claimed, there is progress toward a goal; in biology, there is none.
Kuhn is one of the few who sees a virtue where others see a flaw.
The analogy that relates the evolution of organisms to the evolution of scientific ideas
can easily be pushed too far. But with respect to ... [the idea of progress through
scientific revolutions] ... it is very nearly perfect ... The resolution of revolutions is
the selection by conflict within the scientific community of the fittest way to practice
future science. The net result of a sequence of such revolutionary selections, separated
by periods of normal research, is the wonderfully adapted set of instruments we call
modern scientific knowledge, Successive stages in that developmental process are
marked by an increase in articulation and specialization. And the entire process may
have occurred, as we now suppose biological evolution did, without benefit of a set
goal, a permanent fixed scientific truth, of which each stage in the development of
scientific knowiedge is a better exemplar. (Kuhn 1962, pp. 172f)
Kuhn's views have been subject to extensive criticism which I do not
intend to review here (cf. Lakatos and Musgrave 1970 for some relevant
papers on the issue). Under pressure for clarifying the sense in which his
position did or did not avoid relativism, Kuhn restated his position as
follows.
T believe it would be easy to design a set of criteria — including maximum accuracy to
predictions, degree of specializations, number (but not scope) of concrete problem
solutions — which would enable any observer involved with either theory to tell which
was the older, which the descendant. For me, therefore, scientific development is, like
biological evolution, unidirectional and irreversible. One scientific theory is not as good
as another for doing what scientists normally do. In that sense I am not a relativist.
(Lakatos and Musgrave 1970, p. 264)
Toulmin (1972, pp. 322ff) objects to this reformulation as illegitimately
attributing unidirectionality to biological evolution and as a throwback to a
providential view of evolutionary theory rather than a populational view
which, Toulmin argues, is more in line with Darwinian philosophy. In
distancing his own view from this later Kuhnian position, Toulmin argues
that his own view is evolutionary only insofar as “changes from one
temporal cross-section to the next involve the selective perpetuation of
conceptual variants” (Toulmin 1972, p. 323). No unidirectionality is
implied. I will not stop to assess Toulmin's strategy for avoiding relativism
(cf. Toulmin 1972, pp. 495—500; Blackwell 1973a; Cohen 1973 and Hull
1973; Kordig 1982, 1983).
It is one short step to the morass of contemporary debates on rationality
which T want to studiously avoid. Let me just register here my feeling that
there is something right about Kuhn's 1960 view and Toulmin's 1972 view
and if they are taken to lead to the view that science is irrational, then
something needs to be done about the concept of “rationality.” Since I do
notknow thatitis, [do not intend to try to do it here.
Indeed, Toulmin argues for a “local” or “ecological” concept of con-
textual rationality. The recognition of the populational nature of concept
change leads us to the conclusion that there are no universal criteria for
rationality or “global” selection criteria (Toulmin 1972, p. 317; 1981, p.
31; cf. Blackwell 1973 and Hull 1973).
That the progress of science implies some such “global” criteria
whereas natural selection does not has, in fact, been taken as a strong
point of disanalogy between conceptual and biological evolution (e.g.
Elster 1979; Thagard 1980; and Blackwell 1973). Hull (1982) and
Bechtel (1984) have argued that scientific progress need not involve a
commitment of global criteria if it is recognized that the regularities of
nature and the “laws of nature’ exert a transcontextual constraint on the
development of scientific theories.
My own view is that the progress issue is a problematic test for the
adequacy of evolutionary models of scientific development. If the views of
Kuhn and Toulmin are correct then our notion of progress needs to be
re-evaluated (and T think this was Kuhn's primary intention in his 1970
“recantation”). For all we know progress in science may turn not to be as
chimerical as we now take immanent purpose in nature to be. Leading
intellectuals 300 years ago may have disagreed. Realism in and of itself
does not seem to me to be sufficient to guarantee that science can be
progressive. Popperian realism does, but that leads to will-o'-the-wisp
chases after measures of progress such as “verisimilitude.” The fact of the
matter is that judgments of progress in science are always "local” judg-
ments. This is because even if “global” criteria do exist we are never in a
position to know what they are, except by a local presumption or fiat. The
history of science and the history of ideas gives almost all of us pause
when we contemplate reifying a contemporary standard as an inviolable
canon. Rather than assume progress does or does not exist, a more useful
exercise for those interested in constructing evolutionary models would be
to take alleged cases of indisputable progress and see to what extent they
can be ‘explained away' as the vicissitudes of historical fortune and
changing local fashions.
428 MICHAEL BRADIE
(v) Miscellaneous Objections

a. Cohen (1973) in his review of Toulmin's (1972) has alleged that a
fundamental disanalogy between intellectual disciplines and species is that
whereas species are classes with organisms as members, intellectual
disciplines are wholes with concepts, practices, etc. as parts. Hull (1974,
1982, et al) has taken up the challenge and argued that, in fact, species are
historical individuals and individual organisms are their parts.
b. Dawkins (1976) introduced the “meme” as the conceptual analogue
of the gene (cf. Lumsden and Wilson's “culturgen” in their 1981). For
Dawkins, the gene is an analogy and nothing more. lts special feature is
that it is a replicator (Dawkins 1976, p. 205). Hull accepts the utility of
Dawkins' concept and argues that the chief difference between the two is
that genes play two roles in biological evolution (autocatalysis or self-
replication and heterocatalysis — the production of organisms) whereas
memes play only one role in sociocultural evolution-autocatalysis. As a
result, for Hull, the genotype-phenotype distinction, which is central in
biological evolution, is not central and perhaps not all that important in
“memetic evolution” (Hull 1982, p. 307). As far as I understand him, it is
this difference which is the key to why Hull thinks that sociocultural
evolution can at best be only metaphorically Darwinian or Lamarckian
(Hull 1982, p. 310; cf. Hull 1985). Bunge (1983, p. 58), on the other hand,
sees a fundamental problem since whereas genes are in some sense alive
and self-replicators, ideas or “memes” are not. Kary (1982) makes a similar
point.
c. Finally, Ruse (1983, pp. 144—146) makes an interesting observation
with respect to what he sees as deficiencies in both Popper's model and
Toulmin's model. As he sees it, Toulmin's model does not pretend to be
normative but only to be descriptive of the evolution of knowledge.
Insofar as it is, Ruse argues, it fails to address any central epistemological
issue, since these are, for the most part, normative. Popper's evolutionary
epistemology at least claims to be normative and so can be said to be
addressing epistemological issues. However, Darwin's theory, on Popper's
model, is a descriptive and not a normative theory, and so Popper's
program does not qualify as truly “evolutionary.” We shall return in
section IV to address in greater detail the relationship or lack of it
between evolutionary epistemology and “real” epistemology.
(vi) Isit Darwinian?

If we waive, for the moment, the question of whether “evolutionary
epistemology” is really “evolutionary,” and grant that it is, in some sense,
a further question has arisen about whether it deserves to be called
“Darwinian.” Many who would admit that there is something evolutionary
about conceptual change deny that it does although there is no consensus
about why not.
Blackwell (1973, p. 65) and Cohen (1973, p. 48; 1974, p. 324) argue
that Toulmin's model is not really “Darwinian” because of the “coupling”
of selection and variation. Cohen urges that
Toulmin's claim to be using the term 'evolutionary' in the precise and strict neo-
Darwinian sense (pp. 134—135) seems hardly more accurate than the claim of some
cultural relativists to be generalizing from relativity physics [a claim Toulmin rejects on
pp. 89—91 of his 1972 book]. (Cohen 1973, p. 49)
Thagard (1980, p. 188) also thinks that Toulmin's coupling of variation

and selection indicates a serious flaw in the analogy. Lewontin (1982, p.
154), who urges that the key Darwinian contribution is that variation is
random with respect to environmental challenge, would presumably agree
as well.
Kary (1982, p. 367) argues that some notion of common descent is
required for any evolutionary theory, scientific or otherwise to qualify as
Darwinian. While recognizing that Toulmin sees the need for some
mechanism of descent, she objects that Toulmin's account of that mecha-
nism is extremely vague. Concepts, unlike genes or organisms are not
self-replicating. They do not contain requisite causal mechanisms of
inheritance features which she holds are crucial to any proper Darwinian
evolutionary theory (Kary 1982, pp. 364, 367). She points out that
in Toulmin's evolutionary account concepts do not produce new concepts at all. Instead
they are either modified through time, or else abandoned. And the source of new
concepts is not located in other concepts, but in the mental activity of some human
being.
The net resultis
We should hardly expect the situation in which Toulmin's concepts exist to yield proper
Darwinian results. For in the absence of some causal link among these concepts which
would account for the production of new concepts by prior concepts through descent,
we cannot expect new incividuals (disciplines) to be produced from the parts (concepts)
of prior individuals. (Kary 1982, p. 364)
Skagestad (1978, p. 615) and other argue, in effect, that since the
mechanisms of conceptual selection are not those of biological selection,
conceptual change is either not properly evolutionary or not Darwinian.
Toulmin, as was noted earlier, sees the key issue to be whether the
evolutionary model is providential (non-Darwinian) or populational (Dar-
winian). As I understand him, any populational approach which involves
variational mechanisms and selection mechanisms qualifies as Darwinian.
For Popper and Campbell, the key feature is the presence of evolution by
blind trial and error selection, with the apparent non-blindness (which
drove Toulmin to postulate some coupling between variation and selec-
tion) being handled by appeals to inherited feedback mechanisms which
themselves are the product of earlier blind trail-and-error selection.
Ernst Mayr (1982, p. 505) has pointed out that Darwin's “theory” in
430 MICHAEL BRADIE
the Origins is really a collection of five distinct theories. In addition, no

one of the key “Darwinians” held all five of the theories. Hull, in a series
of papers, has made a similar point. To qualify as a Darwinian, one has to
accept a certain portion of a cluster of views and (for Hull, at least,
although not for Mayr) be intellectually connected with mainstream
Darvinism in appropriate ways.' The same, I would think, should hold for
evolutionary epistemologists who claim the heritage of Darwin. If we agree
with Toulmin that the key feature is a populational approach then his
model qualifies as Darwinian. If we agree with his critics that decoupled
variation and selection is the key, then his model is not Darwinian. This
makes the issue a terminological point. It is not, thereby, unimportant or
insignificant. It is just a reminder that we need to be clear about defining
exactly that we mean by calling our favorite models “Darwinian” or “non-
Darwinian” (cf. Mayr 1983). This will serve to avoid needless controversy
and also prevent us from succumbing to or being taken in by implicit
claims to the effect that since such and such a model is Darwinian, some-
how the success which that model (or models under that name) enjoy in
the biological realm has somehow been transferred (with sometimes little
work) to the realm of conceptual change.
3. What Is the Metaphor?

a. It is often assumed that evolutionary epistemology is metaphorically
modelled on biological evolution. But, in fact, there is a good deal of
transfer the other way around as well. Rather than biological growth
serving as a model for the growth of knowledge, often the growth of
knowledge is used as a model for trying to understand biological evolu-
tion. Thus, Campbell (1960, p. 380, fn. 2) allows that
. any process providing a shared program [!] for organismic adaptation in external
environments is included as a knowledge process and any gain in the adequacy of such
a program is regarded as a gain in knowledge. If the reader prefers, he can understand
the paper adequately regarding the term “knowledge” as metaphorical when applied to
the lower levels in the developmental hierarchy.
Popper (1972, p. 145) alleges that
The tentative solutions which animals and plants incorporate into their anatomy and
their behavior are biological analogues of theories; and vice versa: theories correspond
to endosomatic organs and their ways of functioning,
Crittenden, in criticizing Popper's program suggests that
the terminology of trial and error as applied [to] evolutionary processes is simply a
convenient and arresting metaphor. In this case, the claim that the growth of knowledge
mirrors the process of natural selection in this particular way seems to collapse. It is
not that biological data at this point reveal something about the nature of the growth of
knowledge. Rather, a claim about knowledge — that it is the product of trial and error
gambits — has been applied to the evolutionary process. (Crittenden 1977, p. 231)
Ruse, in commenting on a different aspect of Popper's program, his

attempt to reconcile the random variation in biological evolution with the
apparent directed variation in the growth of knowledge, notes that Popper
goes so far as to see a directedness in the biological phenomenon to
mimic the directedness he finds in the growth of knowledge. (Ruse 1983,
pp. 146, 148)
This is not merely a modern tendency. Somenzi (1980) credits Samuel
Butler with first introducing the metaphor of organisms as problem-sol-
vers. Skagestad, in discussing Peirce's views argues that insofar as Peirce
accepts the ‘trial and error’ view, earlier advocated by Whewell, he, along
with others recognized “the obvious analogy between the process of
scientific growth and that of biological evolution” (Skagestad 1979, p. 94).
Now, clearly, analogies work both ways. Metaphors, perhaps, are slightly
less flexible. In fact, though, the passage Skagestad cites from Peirce
suggests that the metaphor goes from science 10 biology.
[In science we] ... proceed by experimentation. That is to say, we guess out the laws bit
by bit. We ask, What if we were to vary our procedure a littie? Would the result be the
same? We try it. If we are on the wrong track, an emphatic negative soon gets put upon
the guess, and so our conceptions get nearer and nearer right. The improvement of our
inventions are made in the same manner. The theory of natural selection in nanure
proceeds by similar experimentation to adapt a stock of animals or plants precisely to its
environment, and to keep it in adaptation to a slowly changing evironment. (Skagestad
1979, p. 94; the quotation is from CP, II 86)
Plotkin (1982), far from seeing a difficulty here, endorses the idea that
biological adaptations are, in some sense, knowledge. The direct point of
contact between biology and evolutionary epistemology, as he sees it, is
the way in which the organization of the adaptation is related to com-
ponents or aspects of the environment
... information or knowledge, in biological terms, describes a relationship between the
order of the world, whatever that order is, and the answering and reciprocal
organization of an organism. (Plotkin 1982, p. 6)
Biological evolution, insofar as it results in the emergence of adaptations
at all levels of organization is similarly alleged to be a knowledge process
(Plotkin 1982, pp. 7—9).
Piaget, for one, sees nothing amiss in this.
. the language of the biologists and that in vogue among psychologists or

epistemologists in the various fields of learning is today converging at many points. The
language of “information” is, in fact, now in common use among biologists as it is also
among those who study language and intelligence ... people are always talking about
the information of the genome or about transmission of information aimed at securing
the action of the genome on ontogenetic growth. Schmalhausen taiks of information
coming from the environment through the channel of selection, and so on. Now this
sort of language is no more anthropomorphic or psychomorphic than that of
“programming,” introduced, as it constantly is, into the analysis of the genetic program
432 MICHAEL BRADIE
which is enforced in the course of embryo-genesis, Such language applies to

programmed machines and “information” they manipulate. (Piaget 1971, p. 57; cf. von
Schilcher and Tennant 1984, p. 171)
While reserving judgment on Plotkin's position. it seems to me that the
reciprocal transfer of metaphors from biology to knowledge and back
again (or vice versa, depending on how you read the scenario) is not
necessarily detrimental to understanding either. In fact, I rather think it is
common practice, as [ have suggested elsewhere (Bradie 1984a, 1984b).
b. At a more specific level, Hull (1982, p. 318) argues that the proper
analogue for science is artificial selection, not natural selection. Bechtel
(1984, p. 316) seems to be endorsing a similar point when he argues that
one difference between biological evolution and theory development is
that biological evolution in nature occurs as the result of an interaction of
forces (of selection, drift, etc.) but humans choose the theories they
embrace and develop. Both processes, he notes, however, are similar in
that they both involve selection criteria. Darwin, Hull reminds us, saw
natural selection as distinct from artificial selection and on that basis he
could use the latter as a model or metaphor for the former. (As is well
known, Wallace did not use the analogy to artificial selection in his formu-
lation of the theory of natural selection. In his 1889 book, Darwinism,
Wallace pointed to this dependence as a weakness in Darwin's work
(Beddal 1968, p. 239).) Nowadays, however, artificial selection is treated
as a special case of natural selection. This is a specific example of the
general tendency to use metaphors from one domain or area in order
to understand a second domain and then to ‘re-export’, so to speak,
metaphors based on the second domain to understand the first. It rein-
forces or weakens the analogy between biological evolution by natural
selection and theory development depending on whether you see artificial
selection as a special case of (modern view) or as distinct from (Darwin's
view) natural selection.
c. Lewontin (1982) and Toulmin (1967, 1972, 1981) have both noted
that contemporary evolutionary models of the growth of knowledge based
on a “Darwinian” model do not represent the introduction of biological
considerations into epistemology, but rather a shift from one underlying
metaphor to ancther. Lewontin sees two central growth metaphors per-
meating biology and the social sciences: unfolding (a transformational
model characteristic of embryology) and trial and error (a variational
model charactertistic of Darwinian evolution). The distinction between
transformational models and variational models of development corre-
sponds roughly to Toulmin's distinction between providential and popula-
tional processes. Lewontin explicitly (and Toulmin, at least implicitly)
argues that “unfolding” is as much a metaphor for embryology and “trial

and error adaptation” is as much a metaphor for biological evolution, as
they are allegedly metaphors for the evolution of knowledge or culture.
Lewontin goes on to argue that both metaphors are bad metaphors for
embryology and organic evolution, so, he thinks, “little wonder that they
have failed to resolve the contradictions in the theory of cognition”
(Lewontin 1982, p. 155). Embryological development is not a “mere”
unfolding of potentialities in the genome, because of the fact that the
phenotypes are produced through an interactive process between geno-
types and environment (P = G X E). In general, he concludes (against,
e.g., Piagetian schemes)
the relation between genes, environment, and organism is extra-ordinarily diverse from
one species to another, from one organ or tissue or enzyme to another, from one
genotype within a species to another ... [and, thus] ... there is no model that is not
trivially generalized, that can serve as a useful metaphor for psychic development
(Lewontin 1982, p. 156)
The argument against treating organic evolution as a process of trial
and error adaptations follows the familiar line which includes pointing out
the limitation of viewing organisms as optimality seeking problem-solvers,
a reminder that frequency dependent selection means natural selection
does not always maximize or even increase the fitness of organisms and
reminders about the effects of pleiotropy, “hitch-hiking,” random drift, and
the like (Lewontin 1982, pp. 157—159.)
I am not persuaded by the argument that since these metaphors are
imperfect models of organic development they cannot be expected to cast
any significant light on the evolution of culture or knowledge. Indeed, their
very imperfection in this respect may be a hidden blessing in disguise for
whenever biological data suggests the basic metaphor must be modified
the needed modifications can serve as a guide to rethinking our views
about how knowledge develops. Thus, e.g., the fact that organisms are not
optimizing problem-solvers does not, in itself, undermine the view that
they are problem-solvers using, in some sense, a trial and error approach.
They may be “satisficing” trial and error problem solvers. Similarly, even if
we admit that not all organic evolution is adaptational, surely some is, and
we are driven to consider how to formulate and solve the correlative
problems in cultural or epistemic evolution.
In any case, Lewontin is not arguing for a metaphor-free biology or
social science. He just wants more appropriate metaphors. Lewontin sees
both unfolding and trial and error adaptation metaphors as resting on
a deeper metaphor that pervades all of contemporary Western thought
and ideology — the alienated mechanistic world of Cartesian dualism.
In its stead, Lewontin proposes a metaphor of “dialectical interdepend-
ence’ which involves the “interpenetration of organism and environment”
434 MICHAEL BRADIE
(Lewontin1982. pp. 159—163). On this view, the metaphor of “adaptation”

itself becomes suspect relying as it does, according to Lewontin, on the
view of “pre-existing niches” waiting to be occupied by organisms capable
of exploiting them. This, he holds, is the Cartesian vision in evolutionary
theory
. an environment causally prior to, and ontologically independent of organisms

[t]he world is divided into causes and effects, the external and the internal,
environments and the organisms they “contain'. (Lewontin 1982, p. 159)
In place of this, Lewontin suggests that “organisms and environment are
dialectically related.” Niches exist and can be identified only in virtue of
the organisms that occupy them. Organisms and environments are con-
stantly shaping and reshaping each other. The separation of organism from
its environment can be, one is led to think, only conceptual and not real
(cf. also Lewontin 1978).
What are the implications of this metaphor for epistemology? Potenti-
ally, I think, they are very profound, so profound, that one (not me,
though) might be tempted to say that if this is the way biology “really is,”
then it cannot serve as a useful model for conceptual change. For one
thing, if niches are “created” by organisms, and do not exist independently
of them, this suggests that problems may not properly exist independently
of their tentative solutions. To accept this would be to throw a serious
roadblock in front of the retreat from positivism. The positivists, recall, at
least some of the early logical positivists, were inclined to ground the
objectivity and rationality of science on a sharp distinction between
theories and observational data. When this distinction proved difficult to
characterize with any precision, it was suggested that theories and obser-
vations might be co-dependent. This threat to objectivity was responded
to, at least by some, by arguing that ‘problems’ remained constant and
could serve the objectifying role that had once been served by the idea of
a theory-independent observation base. Taking Lewontin's dialectical
metaphor seriously suggests that the theory-problem distinction may turn
out to be as problematic as the theory-observation distinction. More
radical consequences follow. If the organism-environment coupled pair is
the analogue of the knower-known pair, then the idea of a constant regular
nature, ie., many forms of realism, are in danger of going under. I
suspect Lewontin does not want to push the metaphor to this extreme.
But, he does conclude by suggesting that
The fundamental error of evolutionary epistemologies as they now exist is their failure
to understand how much of what is “out there' is the product of what is ‘in here”
(Lewontin 1982,p. 169; cf. Fleck 1979)
IV. WHAT IS EVOLUTIONARY EPISTEMOLOGY SUPPOSED TO DO FOR US?
1. Preamble
In this section, I address two connected issues which deal with the
epistemological implications of the evolutionary turn in the theory of
knowledge. First, there is the question of the locus of the epistemological
problem. What is the relationship between evolutionary epistemology and
traditional epistemology? What is an adequate theory of knowledge
supposed to do for us? Is evolutionary epistemology a “genuine” epistem-
ology or is it, as some critics have alleged, “epistemology” in name only?
Second, granting the legitimacy issue, what then is biology supposed to tell
us about knowledge and knowing? Of course, the EEM program promises
to provide a detailed explanation of the evolutionary development of the
cognitive mechanisms of organisms. Here we are concerned with the
further alleged implications of a biological understanding of knowledge for
the content of knowledge.
2. The Locus of the Epistemological Problem

For Campbell, evolutionary epistemology minimally takes cognizance of
and is compatible with “man's status as a product of biological and social
evolution” (Campbell 1974, p. 413). Campbell characertizes his approach
to epistemology as “descriptive” rather than “analytic.” A descriptive
epistemology is “descriptive of man as knower.” An analytically consistent
epistemology, for Campbell, is a theory of knowledge, truth and evidence
which accounts for how humans know, what they know, etc. There are
many such epistemological theories. A “correct” analytic epistemology
must agree with or be compatible with the descriptive epistemology which
follows from the evolutionary view of human beings. Campbell rejects the
views, (1) that truth is divinely revealed to humans; (2) direct realism,
which assumes viridical visual perception; and (3) epistemologies based on
ordinary language analysis, as all being incompatible with the evolutionary
picture of human development.
Descriptive epistemology, Campbell allows, is more a branch of science
than it is of traditional philosophy (Campbell 1974b, p. 141). As such, it is
both hypothetical and contingent (Brewer and Collins 1981, p. 12).
Dretske (1971, p. 586) agrees that evolutionary considerations might
be construed as telling us that
more often than not, or perhaps in the vast majority of cases, the normal members of 2
well-adapted species (such as you and I) are right in their perceptual judgments about
their surroundings.
However, Dretske argues, this has no particular epistemological signifi-
cance, just as the fact that well trained rats invariably find their way
through mazes does not entail that the rats know their way through the
436 MICHAEL BRADIE
maze. The difference between us and the rat, and this Dretske allows,
might be argued to be epistemologically significant, is that
we have developed so that our perceptual responses are partially determined by those
elements in our surroundings about which our judgments are made, (Dretske 1971,
p. 586)
Human beings, unlike rats, make judgments (cf. Dretske's similar recent
remarks in his 1984 Presidential Address to the APA Western Division
Meetings in Chicago). Evolutionary theory, and by implication, evolution-
ary epistemology, has nothing of interest to say about typically epistem-
ologically significant questions such as the right to be sure, the question of
what counts as adequate evidence, what counts as a good or the best
explanation, and how to distinguish between conclusive and inconclusive
reasons (Dretske 1971, p. 586).
Campbell does not disagree with this assessment. Descriptive epistem-
ology, in his view, is trying to do something different from traditional
epistemology (Campbell 1974b, p. 140). These are three possible con-
figurations of the relationship between descriptive epistemology and
traditional epistemology.
(1) Descriptive epistemology as a competitor to traditional epistem-
ology. On this view, both are trying to address the same concerns and
offering competing solutions. Dretske argues that descriptive epistemology
in this sense fails to touch the traditional questions and is, thereby,
epistemologically irrelevant.
(2) Descriptive epistemology might be seen as a successor discipline to
traditional epistemology. On this reading, descriptive epistemology does
not address the questions of traditional epistemology because it deems
them irrelevant or unanswerable or uninteresting. I take it that some
naturalized epistemologies fall into this camp (e.g., Quine's).
(3) Descriptive epistemology might be seen as complementary to
traditional epistemology. This, I take it, is Campbell's view.
As such, Campbell admits that descriptive epistemology, in his sense,
does beg the traditional epistemological question of how knowledge is
possible. It attempts to explore the problem of knowledge “within the
framework of contingent knowledge, and by assuming such knowledge”
(Campbell 1974b, p. 141). Campbell, in a 1977 lecture held
While [ want descriptive epistemology to deal with normative issues, with validity, truth,
justification of knowledge — that is, to be epistemology — descriptive epistemology can
only do so at the cost of presumptions about the nature of the world and thus beg the
traditional epistemologist's question. (Quoted by Shimony 1981, p. 99)
Brewer and Collins characterize Campbell's descriptive epistemology as
“hypothetically normative.” As they see it, it seeks to explain why science
works (?) if and when it works to produce "valid knowledge.” It seeks to
explain why science fails, if and when it does. Finally, it seeks to determine
how to go about acquiring “valid knowledge.” It does all this, given some
presumptive general truths about the world (Brewer and Collins 1981, p.
12). Among those presumptive truths is an ontological position which has
come to be known as “hypothetical realism.” (See section V below for
further discussion.)
With respect to the traditional epistemological question (how is knowl-
edge possible?), descriptive epistemology must, Campbell says, be what he
calls an “epistemology of the other one.” Such a perspective abandons the
justification of first person knowledge and works instead “on the problem
of how people in general, or other organisms, come to know” (Campbell
1974b,p. 141).
Hull (1982, p. 273) maintains that evolutionary epistemology, in addi-
tion to seeking to extend a biological theory of evolution “to include social
and cultural traits,” seeks “to supply an epistemic justification of our
knowledge of the external world.” Dretske, however, has doubts about the
ability of an “evolutionary view of man's perceptual powers” to satisfy the
sceptic (Dretske 1971, p. 588). In this connection, Stroud, in criticizing
Quine's naturalized epistemology agrees that, in effect, all we get is an
“epistemology of the other” But, the question of how knowledge is
possible at all goes unanswered (Stroud 1981b, pp. 463—466). In re-
sponding to Stroud's objections to attacking epistemological questions by
“projecting ourselves into the other's place,” Quine argues that “this
projection must be seen not transcendentally but as a routine matter of
analogies and causal hypotheses within our scientific theories” (Quine
1981, p. 474; cf. Olding 1983, p. 2).
In such a way, he thinks, we get from an epistemology of the other to
traditional epistemology. On the surface, it would seem that this just begs
the question against the sceptic again. But, perhaps one can defuse, in
part, the sting of traditional scepticism by dividing sceptics into two
groups: those who will not accept anything and those who argue for
scepticism on the basis of illusion or the fallibility of science, etc. To the
former, we say nothing and leave them at the crossroads. To the latter,
Quine's point seems more telling for, in effect, it is a rejection of the
sceptics’ move to “transcendentalize” objections which, after all, are
derived from intersubjective comparisons and errors in the first place. On
this reading, both the sceptic and the epistemologist of the other start from
the same considerations but the sceptic is the one who gives the argument
a transcendental turn and then complains that appeals to intersubjective
experience are question-begging. The epistemologist of the other need
only block the initial turn to thwart this line of argument. The wrong move
would be to accept the problem as posed by the sceptic in its transcen-
dental form and then try to argue back to intersubjectivity. This latter
strategy runs a foul of all the ‘veil of illusions’ objections. The trick is to
avoid being seduced behind the veil in the first place.
438 MICHAEL BRADIE
This line of reasoning will not, I suppose, quell sceptical qualms which
start from “in principle” objections based on “logical possibilities.” How-
ever, these sceptics are really no better than the ones we left at the
crossroads since the “in principle” objections can be raised, in principle,
forever. Life is too short to deal with such people.
Olding (1983, p. 6) considers an argument to the effect that evolution-
ary explanatiors of the origins of our cognitive capacities, in themselves,
lead to scepticism. The argumentis this:
(1) If Darwinism is true, then evolution would not produce rational
creatures capable of arriving at justified true belief about the
world.
() Evolution has produced such creatures (i.e., us).
Therefore,
(3) Darvinism is false.

The rationale for premise (1) is that Darwinian evo. tion promotes the
development of abilities that enhance survival value not truth. Although
what has survival value and what is true may at times coincide, they do not
always. We shall return to this argument below. For now, assume premise
(1) to be unproblematic. If we accept (2) then, we are driven to the
conclusion that Darwinism is false. If we reject (2), by admitting that even
human beings do not have the capacity to arrive at justified true beliefs
about the world, we are driven to scepticism. Even if Darwinism were
true, we would have no good reason to believe it. Premise (1), then, must
be false. Olding goes on to argue that the standard evolutionary strategy
for explaning the origin of complex adaptations can work here as well,
while at the same time recognizing that much needs to be said about how
exactly this takes place if we are to reconcile survival value with truth and
avoid relativism at the same time. In this connection, Skagestad has urged
that
The crucial question of evolutionary epistemology is the question of how evolution by
natural selection was able to generate, in one biological species, a mode of evolution
not operating through natural selection, and yet contributing to the survival of the
species in question. (Skagestad 1978, p. 620; cf, e.g., von Schilcher and Tennant 1984,
p.200; Dawkins 1976; Lumsden and Wilson 1981)
Another line of argument suggests that evolutionary accounts are
somehow incompatible with justificatory or normative considerations. An
evolutionary account of knowledge, being genetic and causal, might be
argued not to touch the central normative questions of epistemology or to
commit the fallacy of deriving “ought” from “is.” (See, e.g,, Crittenden
1977, p. 229; Putnam 1982, Giere 1985).
Rescher puts the potential objection as follows:
Evolutionary epistemolcgy commits a mode of genetic fallacy. The “genetic fallacy”

confuses the course of historical development with one af probative justification, for
example, by arguing from the fact that a doctrine has a somewhat reputable (or
disreputable) origin that it must be tenable (or untenable). The Darwinian epistem-
ologist in effect argues in just his way, moving from historical survival to the
presumptive correctness of methods. Surely no such transcategorical inference from the
factual issue of historical considerations to conclusions regarding the issue of the
normative validation of a method can possibly be valid. One cannot move from the
historical order of temporal development to the logical order of probative concate-
nation. (Rescher 1977, p. 135)
Rescher rejects this argument on the grounds that one of the explicit
premises of the argument, as he lays it out, is a value-laden premise to the
effect that “the historical process is merit-indicative for methods because
methodological survival in a rational community indicates a continued
adoption that betokens ‘fitness’ where fitness reflects a warranted prefer-
ence on grounds of demonstrated adequacy” (Rescher 1977, p. 135). One
escapes the genetic fallacy, for Rescher, by moving to the methodological
level. Incidentally, this move also seems to defuse the survival/truth value
issue as Rescher sees it (Rescher 1977, p. 141).
Shimony (1981; cf. also 1971) proposes integrating factual and norma-
tive considerations in epistemology by reconstructing epistemology as a
“dialectical” rather than a foundational enterprise. The resulting program,
and Shimony emphasizes that this is just a program, he calls ‘integral’
epistemology. It would combine descriptive, analytic, and methodolgical
components in a naturalistic and dialectical framework.
The problem of reconciling an evolutionary approach with the norma-
tive concerns of traditional epistemology is a deep and vexing one. I am
not persuaded that the difficulties warrant ruling out the “genuineness” of
evolutionary epistemology. My guess is that some version of an evolu-
tionary-naturalistic epistemology which incorporates biological, social,
cultural and logical considerations will give us a reasonable handle on the
nature and role of norms in knowing. If this entails that some traditional
concepts of value cannot be accommodated, so much the worse for those
traditional concepts. Here I am in full agreement with Campbell's view
that, at the very least, evolutionary considerations place constraints on
what can count as an adequate theory of epistemic value.
3. The Virtues of Evolutionary Epistemology
a. Rationality
Ackermann (1970, p. 65f) suggests than an evolutionary account of
scientific practice focuses attention of scientists as members of a popula-
tion. He goes on to suggest that suitable populational concepts of objectiv-
ity and rationality can be developed which are superior to those which
440 MICHAEL BRADIE
attempt to define rationality and objectivity in terms of the individual

scientist.
In order to compute the probabilities relevant to a reasonable description of adaptation
and fitness, the differential success of suitably similar individuals must be taken “into
account.” We may then suppose by the analogy that the biographies of individual scien-
tists can be regarded as illustrating objectivity [and rationality| only in the context of a
consideration of the differential success of all of the other individuals who are attempting
to survive and reproduce in the same data environment. (Ackermann 1970, p. 66)
Toulmin as well has developed a populational or “ecological” concept of
rationality which he claims provides “a healthy antidote both the exces-
sively systematic and ahistorical analyses of developing systems, and also
to excessively sweeping ‘evolutionist’ doctrines” (Toulmin 1972, pp. 329,
337). Toulmin's claims to have provided an evolutionary account of
rationality which avoids relativism have been challenged. (See, e.g., Black-
well 1973; Cohen 1973; Hull 1973.) One particularly troubling point is
the claim that Toulmin's scheme and others like it suffer from self-refer-
ential inconsistency (Blackwell 1973a, p. 67; Kordig 1982, 1983). Black-
well puts the problem as follows. On Toulmin's view, there are no
invariants either in nature or knowledge. What then of selection and
variation, which on Toulmin's view, are the engines that drive both
biological and conceptual evolution? If they are held to be invariant, then
Blackwell sees Toulmin as being driven into the camp of the ahistorical
formalists, If they are variants then, Blackwell urges, Toulmin's account of
rationality is threatened, and he is driven into the camp of the historical
relativists. The problem is put more formally by Kordig who claims that
Toulmin's view and others like it are self-referentially inconsistent. In this
regard, Hull, in his review of Toulmin's book makes the point that
Toulmin' view need not be intrinsically self-referential.
From the fact tnat the units with which a theory is concerned evolve it does not follow
that the theory itself evolves . . From the fact that species evolve it does not follow
that the synthetic theory of evolution evolves. Of course, on independent evidence, it is
quite clear that theories about the evolution of biological species have evolved. (Hull
1972,p.1124)
Similarly, pushing the argument up one level, from the fact that scientific
theories evolve it does not follow that the constraints which guide the
choice of theories evolve, although there is some independent evidence to
suggest that they do. But, suppose, for the sake of argument, that Toulmin's
view is self-referential. Kordig's argumentis basically as follows:
(1) Toulmin's theory T has the property E of being evolutionary.
(2) The theory T holds that all properties of all biological systems,
natural and intellectual, evolve.
Therefore,
(3) The property E evolves.

But
(4) Tf E evolves, it changes and becomes whatit now is not.

Therefore,
(5) T must eventually possess property E.
(6) Atthat point, T will not be evolutionary.
But,
(7) (6) contradicts (1)
Therefore,
(S) T must be necessarily false.
I find this argument striking but not compelling. Since I think there is
something basically right about Toulmin's “ecological” concept of rational-
ity and the attendant consequence that the very standards of rationality
themselves are subject to change, I am persuaded that there must be some-
thing fundamentally amiss with Kordig's argument. I am not, however, sure
about what it might be. Kordig, himself, suggests one way out. He rejects
the idea that the liar paradox shows the self-referential inconsistency of
ordinary language by arguing that what it shows, rather, is that “I am now
lying” does not express a proposition. Similarly, one might argue, E is nota
“real” property. This is not, however, completely compelling.
The problem of rationality is a deep problem for evolutionary epistem-
ologists, but not, as far as I see, a fatal one. In passing, note that Sober
(1981) argues that situating the problem of rationality in an evolutionary
context sheds light on a number of philosophical problems.
b. The Historiography of Science

A number of authors have claimed that evolutionary accounts of the
development of science have virtues over alternative historiographic
approaches. Thus, Ackermann (1970, p. 69) holds that the evolutionary
perspective helps explain the simultaneous existence of classical and non-
classical theories (e.g,, relativity theory/Newtonian mechanics; Ptolemaic
models/Keplerian models, etc.). The idea is that each fits into different
“data environments’ which do not necessarily disappear with the introduc-
tion ofa new theory.
Richards' evolutionary model of the development of science is moti-
vated by a desire to preserve what he sees as the insights of the social
psychological and Gestalt models, i.e., that science is part of a larger social
complex (Richards 1981, p. 48) and that norms are determined, not by
nature, but by human decisions, without accepting the destructive rel-
ativism that they both entail. The model he favors, which he calls the NSM
(“natural selection model”), has a number of claimed advantages including
its stress on the need to integrate philosophical and historical analyses of
442 MICHAEL BRADIE
science with psychological, sociological, economic and political studies

(Richards 1981, pp. 70f, 71). The integrationist emphasis of evolutionary
accounts is noted also by Blackwell (1973b), Shrader (1980) and Toulmin
(1967, 1972), among others
Toulmin also thinks that such models are meritorious insofar as they
focus attention on questions such as who carries the tradition, who is
responsible for innovation, and who determines the manner of selection
(Toulmin 1967, p. 468). Hull, of course, is the leading spokesman in
developing this approach. (See, e.g., Hull 1982, 1983, et al
c. Pure and Applied Knowledge

The demarcation between pure and applied science disappears in the
wake of an evolutionary analysis, or so claims Ackermann (1970, p. 70).
In the final analysis . . . the [theoretical] analogue to food is control of the environment,
and not understanding. No line of demarcation between pure and applied science is
consequently to be sought or to be found. Should a theory not make a difference in
some data environment leading ultimately to better control of some process with
consequential potential benefits for human life, it will not survive the adaptive struggle
of scientific lopment. By means of tracing out this tenuous path, the ultimate
relevance of scientific practice to human life can be made out on a philosophically
coherent basis.
Blackwell (1973b, p. 330) endorses the view that knowing has a practical
dimension. In addition, he argues that the concept of adaptation supplants
the concept of explanation as central. (Cf. Karl Popper on this point!)
Blackwell also urges that an evolutionary perspective requires a pragmatic
theory of truth where pragmatic value is to be defined as “adaptional
advantage.” Popper, Campbell, Hull and others do not accept this radical
interpretation. We shall explore the implications of the evolutionary
perspective for realism in greater detail in the next section. For the
moment, note that Blackwell's position is close to the 19th century view
that Toulmin has called “Mach's error,” namely the failure to recognize the
possibility that the “intellectual selection-criteria” of science need not be
the same as those operative in biology (Toulmin 1972, p. 321). The force
of this point, regardless of whether or not one accepts Toulmin's inter-
pretation, is that natural selection appears to have produced an adaptive
strategy, natural science, which itself need not be ruled by the same
considerations as biological evolution. (Skagestad 1978, and Dawkins
1976 have made the same point.)
d. The Mind-Body Problem and Other Metaphysical Issues

Karl Popper, in his paper ‘Of Clouds and Clocks, argues that evolu-
tionary considerations resolve what he calls “Compton's problem” of how
the non-physical (aims, purposes, etc.) can affect the physical. He sees this
as a version of Descrates’ classical mind-body problem. The quick and
dirty resolution is that higher levels of language, in terms of which aims
and purposes can be formulated, have evolved under selection pressures

to exert better control of both a) lower levels of language and b) the
organism's adaptation to the environment.
Nonmaterial aims and purposes exert influence on material organisms
because it is the function for which they evolved. Well, “just-so” stories
aside, this is all well and good but we still may be vaguely dissatisfied. If
we go back and consider Descartes' original formulation and attempted
solution of the problem, we can see why. What Descartes and latter day
Cartesians are after is a proximate solution in terms of structures as they
exist in minded creatures today (or circa 1630). What Popper offers, at
best, is a hypothetical ultimate solution. Even if we are not completely
satisfied with Popper's proffered solution, the evolutionary perspective as
Popper uses it here, reminds us that biological systems and their prop-
erties (whatever their level) have a two-fold aspect, ultimate and proxi-
mate. A complete understanding of biological phenomena (in the light of
evolutionary theory) requires both a proximate and an ultimate analysis.
Thus, Popper's (proximately) unilluminating analysis is somewhat, albeit
dimly, illuminating because it remains us of the historical/contingent
dimension of classical metaphysical problems. The illumination is, admit-
tedly, in this case at least, quite dim. It remains an open question whether
socio-cultural considerations in problems such as these wash out any but
the barest biological considerations. After all, one can just as easily argue,
as many do, that ‘consciousness’ is an evolutionary adaptation which
enhances the survival and reproductive prospects of organisms which
possess it. This hardly “solves” the problem of consciousness. All the
difficult questions, how it works, etc., are proximate and remain to be
solved. (Popper, in a note added in 1974 to the revised edition of
Objective Knowledge, admits as much (p. 255). The proximate problem he
regards as insoluble.) Nevertheless, even here, the evolutionary point of
view demystifies consciousness to the extent that it is plausible to construe
it as just another handy adaptive organ. Still, there remains much work to
be done even from the point of view of ultimate analyses. If all ultimate
analyses turn out to be “just so” stories then perhaps not much can be
gained by this line of inquiry. If not, then an evolutionary perspective can
be expected to yield some understanding of these and similar questions,
however crude. A mere sanguine view of the efficacy of evolutionary
epistemology to resolve fundamental philosophical issues is offered by
Riedl (1984).
The above are only a sampling of the many claimed virtues of evolu-
tionary epistemology. Although I am not convinced of the merits of all of
these claims, neither am T convinced by the claim (e.g., by Dretske) that
an evolutionary approach to epistemology is not relevant at all to basic
epistemological issues. But, evolutionary explanations should not be ex-
pected to do everything or condemned if they do not. Here, the distinction
introduced by Mayr between ultimate and proximate accounts of biolog-
444 MICHAEL BRADIE
ical questions can be fruitfully extended to epistemological questions as

well (as I tried to suggest above). For the present, this remains a program,
awaiting its Darwin.
V. THE PROBLEM OF REALITY
Campbell argues that descriptive epistemology, which studies how orga-

nisms acquire and process knowledge, leads to “debunking of the value of
‘hard facts” (Campbell 1979, pp. 195—196). Nevertheless, he attests, the
ideology of ‘stubborn facts’ has a functional truth. Evolutionary epistem-
ology in turn, abandons “literal truth” while retaining the “goal of truth.”
For Campbell the central epistemological issues concern how organisms
interact with their environments to produce knowledge. Evolutionary and
processing considerations suggest that knowledge is analogous to a mosaic
mural, “a compromise of vehicular characteristics and of referent attrib-
utes” (Campbell 1979, p. 183). We find ourselves committed to an
“epistemological relativity.” Amoebas know what they know, frogs know
other things, humans still more. Each kind comes to know what it does
through processing by cognitive structures which are the product of
evolutionary development. We may know more things than other crea-
tures or different things but each kind of organism constructs, as it were,
an image of reality based on its own needs and capacities. However,
Campbell, like Popper, is unwilling to push this epistemological relativism
to an ontological relativism even though, as he says,
The language of science is subjective, provincial, approximative, and metaphoric, never
the language of reality itself. (Campbeil 1975, p. 1120)
That reality itself does have a “language” is embodied in Campbell's
presumptive ontological view which he calls “hypothetical realism.” The
basic postulate of hypothetical realism is that there is an objective world of
objects and relations which exists independently of any knowing and
perceiving organisms. The organisms which inhabit and interact with this
world, however, have only indirect, fallible knowledge, which is “edited”
by the “objective referent” (Brewer and Collins 1981a, pp. 2f).
Clark (1983, pp. 24f) sees the evolutionary epistemologist as committ-
ed to both a hypothetical realism, insofar as old organisms including
human knowers are assumed to be evolutionary products of an “indepen-
dent, pre-existing reality,” and a “transcendental realism” as well. Evolu-
tionary epistemologists are transcendental realists in that organismic
minds, as functions of an evolved organ, the brain, are limited by adaptive
considerations presumably to ‘understand’ only what is useful to the
survival of the organism. Clark suggests
The limits implied by the adaptive model thus suggest that what is real will extend
beyond what we can ever know to be real. (Clark 1983, p. 25)
Clark's problem, which is not ours here, is to reconcile this result with the
theme from anti-realist sematics that meanings are determined by assert-
ibility rather than truth conditions (cf. also Clark 1984).
The question of whether all minded organisms are limited by the kinds
of adaptive considerations that Clark alludes to is, of course, problematic.
As we have already seen, a key question for understanding human
knowledge and human reason is the extent to which the human mind,
once evolved, has overcome its biological heritage and broken free from
the “leash” of the genes. Perhaps, in fact, the way to distinguish humans
from “lower” organisms is the extent to which human minds are not and
the minds of lower animals are subject to the constraints of adaptive
considerations.
Still, the point Clark makes about the commitment to a transcendental
realism, as understood above, seems well taken. It is important, also,
to realize that the commitment to such a transcendental reality and the
commitment to the “goal of truth” are not identical. More on this anon.
Lorenz (1977, pp. 6f) too, is committed to hypothetical realism. He
characterizes it as follows:
The scientist sees man as a creature who owes his qualities and functions, including his
highly developed powers of cognition, to evolution, that age-long process of genesis in
the course of which all organisms have come to terms with external reality aud, we say,
‘adapt’ to it. This process is one of knowledge, for any adaptation to a particular
circumstance of externai reality presupposes that a measure of information about that
circumstance has already been absorbed. (Lorenz, 1977, p. 6)
What we experience [when we look through the 'spectacles' of our modes of thought
and perception which ae *functions of neurosensory organization that has evolved in
the service of survival] is indeed a real image of reality — albeit an extremely simple
one, only just sufficing for our practical purposes; we have developed ‘organs’ only for
those aspects of reality of which, in the interest of survival, it was imperative for our
species to take account, so that selection pressure produced this particular cognitive ap-
paratus. (Lorenz 1977, p. 7)
Hypothetical realism, he continues, is the view that
the categories and modes of perception of man's cognitive apparatus are the natural
products of phylogeny and thus adapted to the parameters of external reality in the
same way, and for the same reasons, as the horse's hooves are adapted to the prairie, or
the fish's fins to the water, (Lorenz 1977, p. 37; cf. Lorenz 1982, pp. 124£)
In his 1941 paper, “Kant's doctrine of the a priori in the light of
contemporary biology,” Lorenz tries to show how the Kantian a priori
categories and forms of intuition can be given an evolutionary rendition.
The basic idea is that what Kant took to be a priori (for all time and all
rational creatures) forms and categories are, in fact, the result of biological
evolution. If we grant that all organisms with a sufficiently sophisticated
nervous system are capable of some degree of cognitive ability, then we
are forced to the conclusion that each species has its own version of the
forms and categories. On Lorenz's somewhat peculiar reading of Kant, he
446 MICHAEL BRADIE
argues that since the a priori cognitive apparatus of organisms evolve, “the
boundaries of the transcendental begin to shift [and] ... the boundary
separating the experienceable from the transcendental must vary for each
individual type of organism” (Lorenz 1982, p. 123). The transcendental
world, on this reading, is a reality which lies just beyond the grasp of
cognizing organisms, who, as their lineage evolves, may converge upon it.
This may not be Kant, but it does fit in well with Clark's view of the kind
of transcendental realism to which evolutionary epistemologists are com-
mitted. Ruse (1986), if [ read him right, sees this Lorenzian argument as
a [failed] attempt to block Humean scepticism. Insofar as we come to
see that different organisms “perceive” the world differently in the light
of our perceptions of them, it is reminiscent of the limitations of an
“epistemology of the other.” All the different bounderies are being drawn,
as it were, in the context of the phenomenal worid as experienced by
humans. This is, of course, quite right and Kant, were he alive and
recalcitrant to dialectical evolutionary implications to his view of reason,
would no doubt argue as much. If we start with a Kantian commitment to
the distinction between the Phenomenal and Noumenal world, then
evolutionary considerations (being all Phenomenal) do not touch the
question of the boundary. But, if we start from evolutionary theory and its
implications and try to reconstruct Kant and salvage what we can in that
context, then something like Lorenz's view is more plausible. We remain
‘metaphysical sceptics' after all, to the extent that these evolutionary
considerations force us to a ‘transcendental realism” as envisaged by
Lorenz and Clark.
Skagestad objects, not to Campbell’s realism, but to the claim that
realism is an empirical question on a par with biological evolution
or visual perception (Skagestad 1978, pp. 6171f). These latter involve
“matching” data and creating “fits” in the context of the “phenomenal
world,” whereas the question of realism involves a “fit” between our
(“phenomenal”) theories and the (*noumenal”) world. This is a variation
on the objection that an epistemology of the other is an epistemology in
name only. Skagested sees Campbell as attempting to forgean uneasy
alliance between epistemological naturalism and ontological realism (Ska-
gestad 1981, p. 77). Campbell, Skagestad argues, although in the tradition
of epistemological naturalism, departs from the mainstream of that tradi-
tion by his insistence on the objectivity of truth as correspondence. The
mainstream, as exemplified by James, e.g,, sees truth as a relation between
experiences (cf. also Simmel 1982). This mainstream tradition leads to a
form of epistemological relativism (which Campbell calls “ontological
nihilism”) to the effect that
Once we have ascertained the economic, sociological, or psychological causes of the
origin of a belief, as well as the junctions through which it is maintained, there is no
further question to be asked about truth or validity; we can at most ask how well the
belief fulfills its particular function within its social and cultural context. (Skagestad
1981,p.30)
Campbell, wishing to reconcile naturalism with realism, argues, as we have
seen, for a “hypothetical realism, which while admitting the fallibility of
our current theories and the consequent probability that they are not the
final truth about the matter, nevertheless does not abandon the idea that
there is a final truth to the matter to which our current theories are
groping approximations. As Skagestad sees it, the hypothetical realist
transfers the triadic model of theory-data-environment from the study of
rats in a maze to the human condition. By focusing on “knowledge of
others” Campbell tries to avoid the Jamesian charge that the concept of
truth is vacuous. As Skagestad notes, this will not silence the pragmatist or
instrumentalist who insists on first person epistemology and will not
accept a third person variety (Skagestad 1981, p. 85).
Skagestad goes on to consider the content of the thesis, its explanatory
power and its testability. What, then, does hypothetical realism say? As
Skagestad sees it, hypothetical realism it the view that there is a real world
which scientific theories approximately “fil. As science 'progresses”, it
converges on a ‘true description’ of that world.
Given the inadequacy of the ‘rat-maze’ analogy, what is the content of
this claim? Skagestad argues that since hypothetical realism is the claim
that current theories only approximately fit, the claim is not logically
vacuous. The real world is not ideantical to the world as described by
science at time £ The slack allows for the ‘advancement’ of science and
thus, in turn, supports the contention of the hypothetical realist. Science
progresses, or at least changes, because there is a real world which is not
exactly captured by our current theories.
The notion of “approximate fit,” captures the sense, discussed earlier, in
which this is a “transcendental” as well as “hypothetical” realism. As such,
it plausibly controverts certain forms of instrumentalism or pragmatism. It
does not allay all forms of pragmatism, however, since some form of
ontological relativism is still possible. The extrapolation from historical
consensuses to ultimate consensus through convergence requires addition-
al assumptions (as Skagestad admits, p. 88). That our theories change and
may even converge locally, in itself, does not entail the Campbellian
conclusion (via Peirce and Popper) that an ultimate consensus as the limit
goal of inquiry is legitimate.
Following Peirce, with caveats, Skagestad argues that realism explains
the fact of error elimination and the drift towards consensus (Skagestad
1981, pp. 87f). The caveats are that stability of belief is, in itself, no
guarantee of truth and that the drift towards consensus may be due of the
“invisible hand” of common influences or shared biases (cf. Brewer and
Collins 1981a, p. 2). These cautions notwithstanding, realism, according to
Skagestad, explains that
448 MICHAEL BRADIE
(1) “well established beliefs in science tend to remain stable,” and

(2) “investigators independently addressing the same question tend
ultimately to come up with the same answer.” (Skagestad 1981,
pp. 891)
The strong version of convergent realism held by Campbell, Popper,
Peirce and Skagestad can, indeed, explain these facts, but weaker forms
of
realism, not commited to convergence can do so as well. One might point
out that well established beliefs in science do tend to remain stable . . . as
long as methods and practices remain more or less stable, but otherwise
not. [ think this is true, although Tm not prepared to defend it here at
length. In order to do so, one would have to define more carefully
what is
entailed by saying that a practice or method is stable. The same, naturally
,
holds for stability of beliefs. Both problems seem quite difficult. Similar
remarks hold for the robustness of answers addressed by ‘independent’
investigators. To what extent does the stability of community and
practices
tend to rule out investigators who do not come up with the “same answer”
when addressing the “same question”? Question: When are two putative
questions or two putative answers the “same”?
On the testability issue, Skagestad argues that if there is no general
trend toward convergence, then realism (at least, the convergent variety)
is
false (Skagestad 1981, p. 91). Realism, as Skagestad understands
it, is
compatible with specific failures of stability and specific drifts away
from
convergence. Some lines of inquiry may converge or stabilize while others
may not. The convergence or not of a given line of inquiry is a matter
of
the accumulation of inductive evidence. It follows that neither realism
(nor
its rivals) is experimentally disprovable, although empirical consideration
can be brought to bear on the question of rationally choosing between
realism and its rivals. If all lines of inquiry failed to converge or
stabilize,
this would constitute a meta-induction against realism of the
kind that
Putnam argued for in Meaning and the Moral Sciences (Skagest
ad 1981,
p. 92).
Skagestad suggests that one testable difference between realism and
what he considers to be its chief rival, phenomenalism, may be grounded
in the intersection between the history of science and the history
of
philosophy. Phenomenalists will not predict stability and converge
nce
although they can, after the fact, accommodate it. The prediction then,
Which constitutes the testable difference between realism and phenome
n-
alism is alleged to be
realistically inspired scientists [should] take the lead over phenomenal

istically inspired
ones ín promoting novel theories and accepting novel discoveries. (Skagestad 1981, p.
93)
But, how are we to check this? If we scour the pages of the history of
science, do we find phenomenalistically inspired scientists less daring and

more conservative than realistically inspired ones, on balance? [ do not
know, but 1 suspect that one would have a hard time deciding. We might
find ourselves inclined to dismiss the phenomenalistic tendencies of
scientists who were daring as being “out of character.” Even so, suppose
we found that phenomenalistically inclined scientists were more conserva-
tive than their realistically inspired rivals? How does this show that it is
more rational (as opposed to more likely to produce innovative results) to
be a realist rather than a phenomenalist? Realism, in some forms, may
very well be false even though thinking like a realist is more productive
than thinking like a phenomenalist. That is, unless phenomenalism com-
mits one to view that science should never change. Then, a more appro-
priate test would be the following: if phenomenalism were true, then no
changes in science should occur. Changes do occur. Therefore, phenomen-
alism is false. The problem, even if this argument were sound (which 1
suspect it is not), is what kind of realism does it establish as true? Realism
comes in both convergent and non-convergent versions. Neither test
establishes the one rather than the other.
Shimony (1981) attempts to “dialectically” establish, within a frame-
work of naturalized epistemology, certain key claims of Campbell’s
hypothetical realism which Campbell merely assumes to be true, viz. an
ontology of independent real objects, the correspondence theory of truth,
and a causal theory of perception.
In particular, Shimony argues that science, in general, and evolutionary
theory, in particular, is epistemologically significant only if theories are
construed realistically. But, then some further “independent argument
based on “analytic” considerations must be given to support the realist
interpretation of theories. Shimony's argument is as follows. If evolu-
tionary biology is merely a predictive instrument, then no natural history
of the development of cognitive capacities is forthcoming since the truth
= empirical adequacy) of evolutionary biology does not entail the real
historical development of relevant cognitive structures. The theoretical
evidence, the “truth” of evolutionary biology and physics, would not, then,
support a naturalistic interpretation of the “a priori” elements of our
conceptual world view. Therefore, the epistemological significance of
evolutionary theory for establishing the status of those a priori elements
would be nil. So, to be epistemologically significant, evolutionary theories
and scientific theories, in general, have to be construed realistically. But,
to interpret scientific theories realistically without any further argument
means, in effect, saying that the categories of understanding employed by
organisms are products of natural evolutionary processes because our
theories, which are themselves, in part, products of those same categories,
entail they are. Thus, we cannot use evolutionary theory to establish
epistemological claims including realism because our use of evolutionary

Bradie Assesing-1-50

Uploaded by

Copyright:

Available Formats

Bradie Assesing-1-50

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bradie Assesing-1-50

Uploaded by

Copyright:

Available Formats

ResearchGate

See discussions, stats, and author profiles

Assessing Evolutionary Epistemology

Mr. Drummond, I hope [ haven't said anything to offend

Jerome Lawrence, Inherit the Wind

This paper is a work in progress. At this point [ am not completely sure

Biology and Philosophy 1 (1986) 401—459.

what is called evolutionary epistemology in order to determine what, if

II. THREE DISTINCTIONS

pictures of reality are to be understood as the a posteriori evolutionary

(1) questions about the validation and limitations of human knowl-

On the biological side, the issues include

(1) the recognition that biological systems are knowledge systems,

I should remark in passing that Plotkin's formulation of the problem is

Sir Karl Popper is a well known advocate of both programs in evolu-

In a 1977 paper, he sees himself in agreement with Richards (1977) in

. we shall deal with the role of biology-analogous selection phenomena in the

evolution of biological substrate EEM. evolution of brain

that “Epistemology needs Darwinism” (Ruse 1986). Ruse thinks that a

2. T want to briefly address the question of the application of the evolu-

knowledge and the ontogenesis of knowledge should be amenable to the

3. To close this section, we should take note of yet a third distinction.

III. EXPLORING THE METAPHOR

tionary model of conceptual development along biological lines. The

1. The Making of the Analogy

What links the historical development of intellectual disciplines to populational

Toulmin (1967) and Popper (1972) disagree. In his 1967 paper,

A related point involves Popper's use of the branching tree (phyloge-

we might expect to find epistemic analogues for “organism,” “species,”

2. Critique of the Models

(i) Common Problems

(97 STT ‘161 de -anjono 01 ajgen Antondun vossmasip jO ¡XoIIO o,

-femor [BIL0ISIY se UONN[0A? 0) 19afgns

pue dusered jO JEYI SI [apoWI 1Aq YL Suipnppur suonouny aanrudoo sOMEA

uNj UN SIUIUNWIUIOD yHUaros ay -Nuwos 9y7 pue sisyuaros ¡enprarpur JO

vopearumuioj sainpaoo1d Junsa y, S861 INH

(a8papmouy E-puom sipzo 1addog 1eym)

-191ur pue spaau ay) se aajona 01 190fqns (STE-9)

and Lewontin (1982) considers some alleged analogous difficulties in

(ii) Blind Variation

tion in favor of “methodological Darwinism,” the view that methods are

spontaneously and at random with respect to the needs of organisms,

(iii) Selective Retention

knowledge are different from those which occur in biological evolution

(iv) The Problem of Progress

(v) Miscellaneous Objections

(vi) Isit Darwinian?

Thagard (1980, p. 188) also thinks that Toulmin's coupling of variation

the Origins is really a collection of five distinct theories. In addition, no

3. What Is the Metaphor?

Ruse, in commenting on a different aspect of Popper's program, his

. the language of the biologists and that in vogue among psychologists or

which is enforced in the course of embryo-genesis, Such language applies to

argues that “unfolding” is as much a metaphor for embryology and “trial

(Lewontin1982. pp. 159—163). On this view, the metaphor of “adaptation”

. an environment causally prior to, and ontologically independent of organisms

IV. WHAT IS EVOLUTIONARY EPISTEMOLOGY SUPPOSED TO DO FOR US?

2. The Locus of the Epistemological Problem

(3) Darvinism is false.