160

CRITICAL STUDY

TOULMIN'S EVOLUTIONARY EPISTEMOLOGY

B Y LARBY BEISKMAN

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

Human Understanding, Volume I. By STEPHEN TOULMIN. (Oxford : Claren-
don Press, 1972. Pp. xii+520. Price £4.75.)
The first volume of Stephen Toulmin's Human Understanding is itself a
work of mammoth proportions, and there are two further volumes yet to
appear. The present volume, which contains the General Introduction and
Part I (on " The Collective Use and Evolution of Concepts "), draws heavily
on many of Toulmin's earlier works : from his The Philosophy of Science he
derives his fundamentally instrumentalistic attitudes toward scientific
theory ; from The Uses of Argument, his general distrust of formal logic ;
from Foresight and Understanding, his notion of " ideals of natural order ";
and from The Discovery of Time, his Heraclitean metaphysics and an in-
sistence on an historical perspective. These are served up with liberal help-
ings of Darwinian metaphor—although Toulmin claims more than meta-
phorical status for his metaphors (p. 134)—to produce an evolutionary
epistemology which, Toulmin hopes, will enable us to recognize the historical
variability or relativity of all our knowledge without thereby sinking into
a self-defeating historical relativism. Thus, Toulmin's aim is to seek a
" middle-way " between what he labels as the " absolutist " conceptions of
Frege and the " relativist " conceptions of Collingwood or Kuhn, a tertium
quid which will enable us to avoid the pitfalls of both.
I shall begin, therefore, by attempting to show how Toulmin sees his
problem as arising out of the views of Frege and Collingwood ; I shall then
proceed to offer what seems to me to be a much clearer formulation of the
problem which Toulmin is trying to solve and to attempt to explain why
Toulmin is prevented from formulating his problem in this way ; I then
shall outline Toulmin's evolutionary solution to this problem and offer some
criticisms of it to the effect that it is in grave danger of landing us right
back into a relativism ; finally, I hope to offer some hints as to how relativism
might be overcome : how we can, in fact, improve our understanding of
human understanding.
Before proceeding, a word of explanation : there is a great temptation
to see problems like that of relativism (or, for that matter, the problem of
induction) as silly problems. That is, relativism seems so obviously a crazy,
self-defeating position that there is no reason for anyone but a philosopher
to take it seriously—philosophy being, by presupposition, the " repository
of silly problems " (to paraphrase William James' pithy saying). Why then
does relativism seem so obviously crazy ? Simply because relativism is a
product of our culture, of our anthropology, sociology, philosophy, etc.
Since the main plank of relativism is the view that the intellectual products
of all cultures must, if we are to be objective, be accorded equal intellectual
status, it follows that the anti-relativist position is the intellectual equal of
 TOTJLMIN'S EVOLTTTIONABY EPISTEMOLOGY 161
relativism. So no relativist can coherently argue for the intellectual superior-
ity of relativism. It seems, therefore, that although relativism cannot be
defended, the fact that we seem so far unable to offer some defensible non-
relativist epistemology indicates a major weakness in our understanding of
human knowledge. It is precisely for this reason that what seems like a
silly problem is, in fact, a crucial and serious one ; what is disturbing is not
the problem in itself but rather our seeming inability to solve it.
Back now to Toulmin : his analysis begins with the recognition that
different men, in different cultures, and in different epochs have paid " intel-
lectual allegiance " to diverse and various sets of concepts. How, he asks,
are we to respond to this obvious fact of the historical relativity or vari-

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

ability of " conceptual authority " ? On the one hand, we have the Platon-
istic " absolutist " approach of thinkers like Frege ; on the other hand, the
relativism of Collingwood ; the first approach seeks to deny history, the
second to bow to it (p. 52). Thus, according to the " absolutist " position,
there is an eternal, fixed, changeless realm of concepts and our aim must
be to strip away from our own concepts all the accidental accretions of
history so as to approximate to these concepts in their " pure " form. Thus,
how men have actually used their concepts becomes irrelevant to an evalua-
tion of their " intellectual worth " ; history is denied in that " any actual,
historical set of conceptions has a legitimate intellectual claim on us, only
to the extent that it approximates to " the ideal, " pure ", form of the
concepts involved (p. 56). Accordingly, Toulmin argues, the " absolutist "
avoids the problem of historical relativism " only at the price of replacing it
by a problem of historical relevance " (p. 59). For even " supposing that we
could find an elegant and consistent way of presenting (e.g.) the standards
of intellectual judgment used in natural science as a timeless, ideal system ",
still " we must show how the resulting ahistorical standards bear on the
actual practice of scientists working in all the varied fields and milieux over
which " these standards claim authority (p. 61).J
For Toulmin, the crucial problem of rationality is the problem of the
rationality of conceptual change. Thus, any abstract and timeless ideals or
criteria of rationality (as in the " absolutist " approach) must be shown to
apply " not merely to judgments within a given family of concepts, but also
to comparisons between different sets of concepts or conceptual systems "
(p. 63), and, Toulmin maintains, these latter types of judgments soon drive
us " beyond the range of a purely formal analysis " (p. 64), characteristic of
the " absolutist " approach, into more historical considerations. It is pre-
cisely here that Collingwood's " relativism " enters upon the scene. For
Collingwood fully recognizes the historical variability or relativity of con-
ceptual systems ; he does not seek to deny this relativity as either " acciden-
tal " or " irrelevant " ; rather, he goes to the opposite extreme of denying
the existence altogether of rational standards for evaluating the transition
from one conceptual scheme (or set of " absolute presuppositions ") to an-
other. Thus, according to Collingwood's Essay on Metaphysics, knowledge
is organized in " presuppositional systems " at the top of which stand the

N o t i c e here Toulmin's shift from ideal concepts to ideal standards. This shift is not
at all insignificant: the reason is t h a t one may want to deny the existence of anything
like fixed ideal concepts to which we successively approximate and yet uphold the
existence, or at least the necessity, of fixed ideal standards for comparing alternative
theories. B y sliding from one to the other Toulmin manages not only to confuse the
issues here, but also to think (mistakenly) t h a t a criticism of the fixed concept view
entails a rejection of the fixed standards view.
162 LABKY BKISKMAK

absolute presuppositions, which are the basic intellectual commitments of

an historical epoch. Questions lower down in this system get their meaning
and can be rationally discussed only relative to the procedures supplied by
the absolute presuppositions, so that when we come to ask questions about
how these are rationally changed or discussed we are stymied. In other
words, Collingwood's analysis avoids the problem of historical relevance only
at the price of replacing it by the problem of historical relativism ; by recog-
nizing that each historical epoch or each culture has its own rational stan-
dards and absolute presuppositions we seem to have to " concede final
authority within any milieu to the particular intellectual standards current
in it, while denying those standards any relevance or authority outside their

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

original context " (p. 67).
Thus, Toulmin's problem is effectively to find some way of recognizing
the historical relativity of our knowledge (which, he claims, the " absolu-
tist " tries to ignore) while at the same time avoiding the call of relativism
which seems to beckon when we take this relativity seriously (as Collingwood
does). In order to achieve this goal, Toulmin begins by trying to see what
common assumptions link the seemingly opposed views of the " absolutist "
and the " relativist ". The utility of such a strategy is fairly clear : if one
can identify some common ground between the " absolutist " and the
" relativist ", then by denying or rejecting this common ground we may be
able to break out of the absolutist-relativist conundrum and hence solve
the problem at hand. Accordingly, Toulmin argues that what unites both
Frege and Collingwood is the view that our knowledge consists of nice, neat,
logically coherent propositional or presuppositional systems and the identi-
fication of the " rational " with the " logical ". Since, according to Toulmin
at any rate, logic " is concerned simply with the inner articulation of intel-
lectual systems whose basic concepts are not currently in doubt " (p. 84),
it follows that the shift from one such system to another, incompatible with
the first, cannot be a matter of logic, and therefore cannot be rational.
Thus, by rejecting the equation of " the rational " with " the logical " and
by introducing the view that our knowledge consists of rather loose concep-
tual " populations " (rather than tight conceptual " systems ") subject to
a Darwinian process of innovation and selection, Toulmin hopes to break
through the dilemma and achieve an evolutionary epistemology of " rela-
tivity without relativism ".
At this juncture it is, I think, worth while pointing out the rather con-
fused nature of the problem as Toulmin sees it. Firstly, if Toulmin is arguing
that we must relinquish the view that being logical is the whole of being
rational, then no one would, I assume, disagree—for it is well-known that
many madmen are perfectly logical and yet highly irrational. On the other
hand, Toulmin might be arguing for the view that being logical has nothing
to do with being rational. But this latter view is patently false—for the
very possibility of being rational vanishes if we abandon logic. If, for ex-
ample, we ceased to rule out the admissibility of contradictions, or if we
ceased drawing consequences from our theories so that we could test them,
or compare them with the consequences derivable from other theories, then
it is hard to see how we could remain rational at all. Yet, there is good
evidence to think that Toulmin, in fact, wants to defend this second position
—for given his incredibly narrow view of logic as " concerned simply with
the inner articulation of intellectual systems whose basic concepts are not
currently in doubt " (p. 84), Toulmin can only see logic as an organon of
proof, wherein, given true (or at least currently accepted) premises, we trans-
 TOULMIN'S EVOLUTIONARY EPISTEMOLOGY 163
mit truth (or at least acceptability) to validly derivable conclusions, and
hence as having nothing to do with the problem of conceptual change. But
it is obvious that any logic strong enough to transmit truth (or accept-
ability) will simultaneously re-transmit falsity (or rejection) and can, there-
fore, serve equally as an organon of criticism.2 It is hard, therefore, to see
how one can rationally go about changing one's views (or even one's concepts)
without using logic ; without, that is, entering into critical discussion.
In fact, there is even stronger evidence for believing that Toulmin wants
to deny the thesis that logic has anything to do with rationality. Remember
that in his The Philosophy of Science, Toulmin argued the Rylean instrumen-
talist thesis that scientific theories are mere " rules of inference " which

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

serve to " license " inferences from one set of observation statements to
another.3 If we now identify " logic " with " rules of inference " it follows
that any change of scientific theory necessarily entails a corresponding
change of logic ! Since, for Toulmin, questions of rationality are precisely
questions about theoretical or conceptual change, we seem to be forced into
assuming that Toulmin does actually want to argue that logic has nothing
to do with rationality. It is only against this background that one can
possibly make sense of Toufmin's claim that " it was, in fact, always a mis-
take to identify rationality and logicality—to suppose, that is, that the
rational ambitions of any historically developing intellectual activity can be
understood entirely in terms of the prepositional or conceptual systems in
which its intellectual content may be expressed at one or another time "
(p. 84).
As we shall see, this inability of Toulmin's to distinguish clearly between
the intellectual content of a science and the (logical or methodological) rules
of procedure at work within that science actually prevents him from clearly
formulating where his problem lies. But for the moment, let me turn to
the other culprit which Toulmin sees as at the basis of the absolutist-relativist
conundrum : namely, the " cult of systematicity ", the view that our know-
ledge forms a coherent, logically tight system. Now I take it that few would
deny that, at any given time, our knowledge fails to form a tight logical
system, but is this any reason for giving up logical coherence as an ideal or
regulative standard ? Clearly not : if we give up this ideal we simultaneously
relinquish both our ability to consider incoherence and contradiction be-
tween various parts of our knowledge as a defect and the aim of seeking
truth. Yet, Toulmin actually seems to want to reject " systematicity " even
as an ideal. Why ? Basically because of his view that if the component
elements of a science did form a logically coherent system " there will then
be no way of modifying them piecemeal or one at a time : the only chance
of radical change will then lie in rejecting the entire " system " as a whole,
and starting afresh " (p. 129). This is the kind of muddled non sequitur
which frequents Toulmin's book all too often : for given that we are dealing
with an empirical system, it is obvious that we can use such a system to
interact with the world, we can use it in praxis, or in Popper's terminology,
we can try to test it. And, as Duhem has pointed out, failures in such cases
2
The idea t h a t logic is simultaneously an organon of proof and of criticism is well
known to any reader of Popper. See, for example, his Objective Knowledge : an Evolu-
tionary Approach (Oxford, 1972), p p . 139-40.
3
For a criticism of this Rylean view, and its relation to Ryle's attempt to defend a
common-sense ontology against physics, see my " Are Philosophers Averse to Science ? :
Discussion " in D. O. Edge and J . N . Wolfe (eds.), Meaning and Control (London, 1973),
pp. 18-27.
164 LABBY BEISKMAN

need only, from the logical point of view, force us to modify one of our
assumptions, not all of them. Once we see how piecemeal change is possible
even with a logically coherent system, and once we recognize that such
changes will only be fruitful if we insist that they help to restore or even
increase the overall coherence and " systematicity " of our knowledge, then
there is no longer any reason to give up logical coherence and " systematic-
ity " as an ideal. In fact, some of the most startling advances in scientific
knowledge have resulted precisely from increases in the coherence and
" systematicity " of that knowledge—e.g., Newton's unification of terrestial
and celestial mechanics, Maxwell's unification of electrical, magnetic, and
optical theory, etc.

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

It thus seems to me that Toulmin locates his problem in a disastrous
way-—to think that we can defend rationality by giving up logic and the
goal of systematic knowledge is just plain crazy. Yet there is a serious
problem which Toulmin is, I believe, striving to articulate, but which he is
prevented from articulating in a clear manner by his instrumentalist views
of science. What, then, is this problem ? Let us begin by drawing a three-
fold distinction between the intellectual content or theories of a discipline
at any given time, the methodological rules or rational standards at work
within that discipline, and the aims or goals of that discipline. Then al-
though (contra Toulmin, p. 84) no one would, I take it, be prone to identify
" being rational " with " paying allegiance to current theory ", many would
want to argue that rationality arises only within the context of agreed rules
and standards. That is, assume we are confronted with alternatives Tx and
T2 at the theoretical level; then a rational choice between or comparison
of Tx and T2 can only be carried out with reference to some agreed standards
of choice or comparison. Thus, if we ask how rational improvement is
possible at the level of theory, our answer will appeal to certain standards
or rules which do not change through the theory-change. The " absolutist "
at this level will, presumably, claim that some fixed set of rational standards
exists against which all such theoretical changes can be judged ; the " rela-
tivist " will, presumably, claim that all standards are of equal value and it
is only an historical " accident " that certain standards are held at certain
times. But what if we want to avoid relativism and yet recognize that our
rational standards can themselves improve ? What, that is, can we do if
we want to recognize the possibility that our standards as to what ought to
count as an improvement in our knowledge can themselves improve ? If
we want to recognize that our rules for improving our knowledge can im-
prove ?4 Here, we seem to have to appeal to fixed aims or goals : that is,
we can compare or evaluate competing rules or standards Rx and R2 by
attempting to see whether following Rx or R2 promises to get us closer to
our aims or goals (this being the standard against which we compare Rx and
R2). The " absolutist " at this level will, presumably, claim that some such
fixed aims or goals exist; the " relativist " that the adoption of such aims
are simply historical " accidents " to be given no particular rational weight.
Finally, what if we want to recognize the possibility of improving even our
aims ? How is a rational evaluation or discussion of rival aims possible ?
Do we here appeal to even higher-level fixed standards and/or aims ? Are
there rules for improving our aims ? We seem to be stymied : it would

*I should here make clear the relation between (methodological) rules and standards :
basically, a rule is a standard formulated as a maxim for action. Thus, take the Popper-
ian standard " Non-ad-ftoc theories are preferable to ad hoc theories " ; as a rule this
becomes : " Try to develop theories of decreasing degrees of ad-hocness ".
 TOULMIN'S EVOLUTIONARY EPISTEMOLOGY 165
appear that we cannot acknowledge the historical variability or relativity of
all of our knowledge (including all of our aims, rules, and standards) without
falling into a relativism; yet it is precisely this which Toulmin hopes to do.
How, then, does he hope to do so ?
Before answering this question, however, a diversion : although Toulmin
takes the first 130 pages of his book to introduce his problem, and although
my formulation of it is far from perfect, he does not, I believe, get anywhere
near as clear a formulation of it as that above. The reason is that Toulmin's
instrumentalism does not allow him to make clear the three-fold distinction
between theories or content, rules or standards, and aims or goals. As an
instrumentalist, Toulmin can only see the content of a discipline as consisting

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

of procedures and methods (namely, explanatory procedures and methods of
representation—see pp. 155-73) ; hence he is misled into thinking that many
who have argued for an " absolutism " at the procedural (i.e., methodological)
level were, in fact, arguing for an " absolutism " at the theoretical level, at
the level of content. Similarly, Toulmin's instrumentalism forces him to see
debates about aims or goals as debates about strategies for producing success-
ful explanatory procedures, rather than as debates about (say) which of two
possible rival metaphysical blueprints for science better reflects what we
hope Nature is like.5 In other words, Toulmin not only collapses the distinc-
tion between methods and content, he also collapses the distinction between
aims or goals and methods. It is, then, no wonder that he fails to formulate
clearly the problem at hand as arising as a result of the content-standards-aims
hierarchy of a discipline and the seeming necessity of rationally evaluating
each lower level in the hierarchy in terms of the next highest level.
To return, now, to Toulmin's attempted solution to the problem : the
key to its solution, he proposes, lies in our taking seriously a Darwinian
analysis of the growth of intellectual disciplines. That is, Toulmin argues
that we must give up the old idea of fixed, timeless principles of rationality
and instead take a more " ecological " or " situation-specific " view of
rationality. According to this view, the " intellectual ecology " of a discipline
at any given time is determined by the " disciplinary ideals " or " explan-
atory goals " current within that discipline. These ideals determine both
the problem-situation within which any worker in the field finds himself and
the selection-criteria for judging the acceptability of novel variants or in-
novations designed to solve these problems. Thus a " disciplinary ideal "
plays the same role in the evolution of intellectual disciplines as an " eco-
logical niche " plays in organic evolution—such ideals determine the demands
of the situation (in that problems = ideals minus current capabilities (p. 152))
and the " survival-value " of any novel proposal (in that such innovations
are to be judged in terms of whether they help get us closer to our goals
(p. 236) ). In other words, a Darwinian process of " innovation-and-selection"
takes place within the context of a disciplinary ideal. Thus, in so far as
Toulmin accounts for rational growth or improvement at the theoretical and
methodological levels, he does so by assuming the relative stability or
" fixedness " of disciplinary ideals or goals. But Toulmin wants to go
further : not only do our concepts and selection-criteria change, so too do
6
In choosing one of two rival metaphysical blueprints we simultaneously choose an
aim : namely, the aim of developing better and better articulations or precisifications
of this blueprint. The idea of aim-articulation and comparison as being crucial to the
rationality of science is Nick Maxwell's—see his " Critique of Popper's Views on Scien-
tific Method ", Philosophy of Science, Vol. 39, No. 2, p p . 131-52, and also his forthcoming
book, The nationality of Science.
166 LARRY BRISKMAN

our disciplinary ideals. When confronted with such changes we are, accord-
ing to Toulmin, at " the frontiers of rationality " (p. 240) ; but, he insists,
such changes can be " rational " (pp. 236-42) and " objective " (pp. 242-53).
What needs to be done, then, is to show just how such changes can be
rational and objective. For unless Toulmin can succeed in this he is forced
directly back into a relativism. The reasons are not hard to find : remember-
ing how seriously Toulmin takes his Darwinian model, he is committed to
the view that there is " no absolute scale of " adaptedness " " (p. 333)—
that is, what is " adapted " in one " niche " may not be in another (see also
p. 443). But it would then appear that the standards of being " better
adapted " are " niche-specific ". Recalling that the ideals of a discipline

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

determine its " intellectual ecology ", we seem to be able to talk about
rational improvement only within an " intellectual niche ", only given a set
of disciplinary ideals. So unless Toulmin can give us a convincing account
of how rival disciplinary ideals can be rationally compared and evaluated,
his theory simply retreats to the relativism which he so desperately seeks
to avoid ; unless he can give us some guide-lines as to how we can rationally
prefer one such goal or ideal over another he lands into the same relativism
of which he accuses Kuhn.
Given how crucial this issue is to Toulmin's whole theory, it is rather
surprising how little attention is actually paid to it. Firstly, we are supplied
with little in the way of actual, clear examples of such disciplinary ideals
(the only fully articulated one I can find appears on p. 124). Secondly,
Toulmin does not make very clear exactly how these ideals determine un-
ambiguously the selection-criteria of a discipline (this is claimed on p. 242
and elsewhere) ; nor does he explicitly acknowledge, as he ought given his
commitment to " relativity", the possibility of debate over selection-
criteria even within the framework of such an ideal. Thirdly, we are not
told why fairly specific disciplinary ideals are needed in order to generate
problems. Why, for instance, can't rather vague ideals, like the search for
truth, determine problems % For example, if two successful theories, such
as Newton's dynamics and Maxwellian electro-dynamics, are in logical con-
flict then we would seem to have a problem even if we only agreed on the
rather general aim that physics seeks truth. Why is it absolutely necessary
to the generation of problems to have more specific disciplinary ideals ?
Unfortunately, Toulmin leaves us rather in the dark here. Most damning,
however, is not what Toulmin doesn't say about disciplinary ideals and their
rational development, but what he actually does say. Thus, he argues that
in a debate over disciplinary ideals the various parties appeal to arguments
" involving the comparison of alternative intellectual strategies, in the light
of historical experience and precedents " (p. 237); they search " for the most
exact and revealing historical precedents " (p. 238) ; their arguments employ
" considerations of an essentially historical kind : using the theoretical
experience of early physicists as a precedent, in estimating the most promis-
ing lines for future theoretical development " (p. 239). Finally, he asserts
that " any profound redirection in the strategy of a discipline thus has to
be justified by appeal, not to previously established patterns of argument,
but to the overall experience of men in the entire history of the rational
enterprise concerned " (p. 241).
What are we to make of all this ? What is Toulmin after here ? My
conjecture is that he hopes to avoid historical relativism by appealing to
history, and not just to history, but to the whole of history.6 But clearly
•For evidence t h a t this is just what Toulmin has in mind, see p . 500.
 TOULMIN'S EVOLUTIONARY EPISTEMOLOGY 167

this won't do, for at least three reasons : firstly, it is incredibly unclear as
to just what " the overall experience of men in the entire history " of a
discipline actually is and so equally unclear as to how appeals to it are to
be made. Secondly, it is a bit hard to see how an historical investigation of
precedents can help us when our problem is precisely to decide whether or
not, or how, to break with precedent (i.e., with past disciplinary ideals).
Thirdly, it simply is not the case that scientists argue over the aims of their
disciplines in this historical manner. (Although this is probably the weakest
of the three arguments philosophically, for Toulmin it ought to be devastat-
ing ! After all, Toulmin's main complaint against the " absolutist " approach
was its lack of historical relevance, which he sees as sounding its death-

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

knell.) Thus, take the case of the debate in the 1930's (which continues
today) over whether or not the new Quantum Theory (QT) ought to be
accepted. One of the aims of classical physics was to develop deterministic
theories, while the new QT was indeterministic. Hence a debate developed
as to the desirability of upholding this classical aim ; was it, after all, an
intelligibility requirement on physical theory that it be deterministic ? (Not,
can we see an element of indeterminism in previous theories ?) Einstein, in
fact, was constantly misunderstood by the defenders of QT, who thought
that his main objection to QT was its indeterministic element. But this
was not Einstein's objection at all : his refusal to accept QT was based upon
his realism, rather than his determinism.7 That is, Einstein was convinced
that physics should not relinquish its traditional aim of attempting to
describe an independently existing reality, and he thought he had a powerful
argument to the effect that not only was QT incompatible with this general
aim, but also with a very precise articulation of this aim : namely, " the
principle of contiguity ", the idea that given two objects far apart in space,
an external influence on one has no direct influence on the other.8 For
Einstein, " unless one makes this kind of assumption about the independence
of the existence . . . of objects which are far apart from one another in
space . . . physical thinking in the familiar sense would not be possible ",
for " if this axiom were to be completely abolished, the idea of the existence
of (quasi-) enclosed systems, and thereby the postulation of laws which can
be checked empirically in the accepted sense, would become impossible ".9
In other words, Einstein did not appeal to precedent per se, but rather to
an argument to the effect that QT is unintelligible as a " complete " descrip-
tion of physical reality.10 Any reader of his dazzlingly clear " Quantum
Mechanics and Reality " can judge this for himself.
To be fair to Toulmin, however, he does tell us how such decisions over
rival disciplinary ideals can be " objective ", in that " the ultimate verdict
on them remains an objective . . . matter " since " the ways in which Nature
'This is made crystal-clear in three letters from Pauli to Born, which appear in
The Bom-Einstein Letters (London, 1971) as letters 112, 115 and 116.
8
In other words, " action-at-a-distance " is impossible. For Einstein, any universe
in which " action-at-a-distance " occurred would be ultimately mysterious, and hence
unintelligible. For we could always ask : " W h a t happens in the space between the
two bodies so as to transmit the signal from one to the other ? ", and in an " action-
at-a-distance " universe the answer can only be : " Nothing ".
•These quotations are from Einstein's " Quantum Mechanics and Reality ", which
is reprinted in The Bom-Einstein Letters, op. eit., pp. 168-73.
"Because, t h a t is, QT conflicts with the " principle of contiguity " and so implies
" action-at-a-distance " (see note 8). Hence, QT needs to be supplemented in order
to get rid of this lacuna, and so cannot claim to give both a " complete " and an in-
telligible description of reality.
168 LARRY BRISKMAN

will actually respond to our attempts at understanding her is something that

goes beyond all human tastes, and all human power " (p. 245). But notice
that it is the verdict which is " objective ", not the choice. In fact, for Toul-
min we can only decide in retrospect which of two rival disciplinary ideals
was the better judged " in light of their actual practical sequels " (p. 244).
In other words, like Lakatos, Toulmin has to appeal to a kind of post hoc
rationality, in terms of the subsequent empirical success of adopting one ideal
over another. But this seems to imply that a rational choice is not possible
at the time the decision has to be made. If this is all Toulmin has to offer us,
then it is hard to see how his own prolonged discussion gets us any further
along than Lakatos takes us.

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

What, then, has gone wrong ? In part, I think, the problem lies in the
fact that Toulmin fails to raise the really crucial question : namely, how is
it possible for our ideals themselves to become problematic ? By defining a
problem-situation in terms of the gap between our ideals and what we have
actually achieved so far, Toulmin debars himself from even asking this
question. Yet, if we are to discover how we can rationally discuss or evaluate
our disciplinary ideals, we must first know the ways in which they can be
brought within the sphere of the problematic. There are, I suggest, at least
three ways in which an ideal or an aim can become problematic : if it can
be shown to conflict with other aims and especially more fundamental
aims; if it can be shown to be unrealizable in principle; and if we are making
little head-way in actually trying to realize it. Only the last of these in-
volves empirical considerations (of success) ; the first two involve non-
empirical or a priori considerations. In fact, the empiricist view of science,
which tries to deny a priori considerations a role in scientific rationality,
actually prevents us from getting clear about how our explanatory or disci-
plinary ideals can become problematic, and hence prevents us from even
beginning to construct an adequate theory of that rationality. In so far as
Toulmin (and Lakatos, for that matter) recognize only the empirical " prob-
lematicity " of our aims, they are unable to offer a satisfactory account of
how those aims are rationally discussable.11
Where, then, are we ? The answer has to be, I think, that we are driven
back to a form of " absolutism " ; we are forced to admit that any theory
of scientific rationality which tries to " escape from the a priori " (p. 495)—
as Toulmin's does—is doomed from the start. Clearly there are problems
with an " absolutist " approach—an approach which assumes, out of a kind
of necessity, that " quasi-fixed " rational standards, and " quasi-fixed "
aims against which these standards are rational ones, exist. But, I suggest,
many of these problems can be overcome if we refuse to see such aims and
standards as existing naturalistically as " facts " (as the Platonist does), but
rather see them as the outcome of our decisions based on argument, as existing
as " reasonable conventions ".12 Clearly, a reasonable convention remains a
reasonable convention even if, in fact, no one employs it. Equally, a con-
vention which is, in fact, employed does not thereby become either a fact or
reasonable. To demand, as an adequacy condition of a reasonable con-
vention, that it should in fact be employed—as Toulmin seems to do when
he demands historical relevance of the " absolutist "—is simply to commit
the " naturalistic fallacy ", to miss the point of the " fact/value " distinction.
u
T h e main ideas in this paragraph are Nick Maxwell's.
12
It is for this reason t h a t I call them " quasi-fixed ", because they are fixed only
as the outcome of our adopting certain " reasoned conventions ", rather than in some
Platonio heaven.
 TOULMIN'S EVOLUTIONABY EPISTBMOLOGY 169
Thus, a reasonable convention may sometimes be irrelevant to actual his-
torical events, but this does not entail that it is ever irrelevant to an evalua-
tion of these events. Similarly, that such decisions can be based on arguments
undercuts Toulmin's other main objection to an " absolutism " : namely,
that such decisions are bound to be arbitrary (see p. 496). Take, for example,
Popper's proposal that empirical falsifiability should be taken as the hall-
mark of scientific status. The fact that, perhaps, certain theories considered
during the history of science have been unfalsifiable does not shake the
reasonableness of Popper's proposals qua proposed convention, only qua
proposed historical fact. But Popper has always made it clear that he is
proposing the adoption of a convention, not making an historical assertion.13

Downloaded from http://pq.oxfordjournals.org/ at University of Strathclyde on April 11, 2015

What, then, is the argument behind Popper's proposal ? Basically, it is like
this : " unless we assume the existence of some pre-established harmony
between Mind and Nature, it is reasonable to suppose that most of the
conjectures we come up with will be false ; to assume otherwise is to assume
that we start from some privileged position epistemologically. But given
this, and given that we aim at truth, it is only reasonable to frame our con-
jectures in such a way that we have some hope of learning both that they
are false and where they are false. So the reasonable way to proceed is always
to try to improve our formulations of our conjectures so that they become
increasingly falsifiable. In this way, we can at least hope to learn about
Nature as best we can." Now it seems clear to me that (a) we needn't look
at the history of science to evaluate this argument; (b) it is an argument;
and (c) to deny the conclusion one must either deny one of the premises or
else show some fault in the reasoning. Until one can do this, Popper's pro-
posal remains that of a " reasonable convention " even if no scientist ever
put it into practice. In other words, Popper's proposal only looks arbitrary
if looked at " quasi-naturalistically " ; once we see it clearly as a " proposed
convention based on argument " its arbitrariness does, I assert, disappear.
Ultimately, then, the main failure of Toulmin's whole approach is its
intolerably historicist emphasis : at rock-bottom only history can be the
judge. It is, then, no wonder that Toulmin can only avoid historical relativism
by taking as the " true standpoint of rationality " the whole of man's accum-
ulated experience in all cultures and in all historical periods (p. 500). In
fact, the very vagueness of Toulmin's whole account makes it tremendously
difficult to see how he actually solves the problems which he claims to solve ;
and for a 500-page goliath there are precious few definite theses to be found.
Perhaps most depressing, however, is the totally conservative " what is, is
rational " tone of the entire discussion. Give me a misguided reformer
(like Bridgman) rather than a reasonable conservative any time. And the
real problem with Toulmin's conservatism is that it isn't even a well-
reasoned one.14

University of Edinburgh

IS
This comes out most clearly in The Logic of Scientific Discovery, sections 9, 10
and 11.
14
I should like to t h a n k Michael Krausz, Nancy Sherman and J a n e Winter for an
opportunity to discuss aspects of this review during the Spring Term 1973 at Bryn
Mawr College. Most of all, I want to thank my friend, Nick Maxwell, for sustained
discussion, criticisms, suggestions, etc. I would be surprised if any page does not bear,
in some way or other, the mark of his influence. Naturally, however, I imagine that
there is much here t o which he will not want to subscribe.