The Critical Rationalist Vol. 01 No. 04 ISSN: 1393-3809 31-Dec-1996
6 Epilogue: Artificial Life
5.1.5 Misconception 3: Adaptation
(96) As with the idea that Darwinism may be tautological, the assertion that it is essentially metaphysical has been more or less independently suggested (and criticised) by a number of different authors.
(97) The criticism of the so-called adaptationist programme by Gould and Lewontin is an example of this kind of argument (Gould & Lewontin 1979, p. 589). Indeed, Lewontin has explicitly claimed that "the adaptationist programme makes of adaptation a metaphysical postulate, not only incapable of refutation, but necessarily confirmed by every observation" (Lewontin 1977, as quoted by Maynard Smith 1978, p. 38).
(98) However, the claim put forward by Gould and
Lewontin is not strictly that Darwinism per se
is metaphysical. Indeed, they accept the empirical
reality of natural selection; their argument is about
which evolutionary phenomena are explicable in terms of
selection. The adaptationist programme, which they
criticise, presumes that all evolutionary
phenomena (or organismic attributes) are a direct
outcome of selection.[12]
Conversely, it seems clear that Gould and Lewontin
accept that what I have called adaptative complexity,
where it exists, does demand an explanation in terms of
selection; and, while such explanations may,
individually, be almost impossible to test in practise,
they are testable in principle, and are not therefore
metaphysical. Thus, in terms of the problem of
adaptative complexity ( 
), Gould and
Lewontin appear to accept that 
is not
metaphysical, and, indeed, that it is the best theory
currently available.
(99) Concerns of a similar sort to those raised by Gould and Lewontin have been independently discussed by various other workers. For example, they were prominent in the discussions at the Wistar symposium on "Mathematical Challenges to the neo-Darwinian Interpretation of Evolution" (Moorhead & Kaplan 1967). Although Popper was not present at this symposium, he was repeatedly cited (directly or indirectly) as the source for such concerns. Thus, for example, Medawar makes the following comment in introducing the symposium:
Then there are philosophical or methodological objections to evolutionary theory. They have been very well voiced by Professor Karl Popper--that the current neo-Darwinian Theory has the methodological defect of explaining too much. It is too difficult to imagine or envisage an evolutionary episode which could not be explained by the formulae of neo-Darwinism.
(100) Unfortunately, there are no detailed citations
to original sources, and I am not aware of Popper ever
having published exactly this criticism of
Darwinism. In any case, my response to this argument
is essentially as already discussed in the case of
Gould and Lewontin: while is, undoubtedly,
difficult to test in respect of the evolution of
specific complex adaptations, it is still testable in
principle (i.e. is not metaphysical) and is the best
theory currently available.
(101) However, Popper has published a slightly different argument for regarding Darwinism as metaphysical--or, more precisely, as a metaphysical research programme (Popper 1974a, Section 37). This arose (at least partly) because, as already noted in section 5.1.4, Popper's earliest considerations of the status of Darwinian theory suffered from a form of misconception 2, and he felt that the theory was therefore "almost tautological"; yet he also felt that, despite this, the theory had considerable explanatory power. Popper seems to have thought that this apparent contradiction might be resolved by regarding Darwinism as a metaphysical research programme. While I think his interpretation of Darwinism as tautologous was mistaken, I actually agree that, in a certain limited sense, it can usefully be regarded as metaphysical.
(102) Firstly, as discussed in (McMullin 1992a), I consider that it is not unreasonable to describe the abstract form of Darwinian theory, presented in that essay, as a metaphysical research programme in Popper's sense. It is not a scientific theory until the primitive entities (especially D-actors) are given some specific empirical interpretation. It must be emphasized that to view this admission (of the metaphysical nature of the abstract theory) as a criticism of any particular interpretation of the theory (such as Organismic Darwinism) would be to indulge again in a form of the tautology misconception 1. This, for example, is the only way in which I can understand one of Peters' purported criticisms of Organismic Darwinism (Peters 1976, p. 4), which is apparently based on its being a specific interpretation of the axiomatic Darwinism of Williams (1970).
(103) Having said that, it must be recognised that the abstract or axiomatic form of Darwinism is metaphysical in a deeper or more profound sense than the conventional abstract theories underlying all science. The general kind of problem which abstract Darwinism seeks to solve is the growth of knowledge; and its mechanism of solution entails a refusal to make predictions--it "works" (in the face of the impossibility of a logic of induction) precisely by declining to predict the growth of knowledge. This is a quite unique kind of abstract theory. It follows that, even when the abstract theory is interpreted in specific empirical terms (such as in the form of Organismic Darwinism) it can never predict the future growth of knowledge. We must say that, as long as such a particular interpretation of Darwinism is viewed as an historical theory of the past growth of knowledge it is perfectly testable (in terms of "retrodictions") and scientific; but if it is mistaken for a "universal law" of the growth of knowledge, then, since it cannot predict such growth, it must be treated as metaphysical. The (metaphysical) position adopted here is, of course, that no "universal law" of the growth of knowledge exists.
(104) It can be seen that this argument for viewing even Organismic Darwinism (as opposed to Darwinism in the abstract) as metaphysical hinges on its (in)ability to predict the future growth of knowledge, or adaptive complexity; thus it is closely related to what I have labeled tautology misconception 3. This is brought out clearly by considering Popper's most substantive presentation of this viewpoint:
...assume that we find life on Mars consisting of exactly three species of bacteria with a genetic outfit similar to that of three terrestrial species. Is [organismic] Darwinism refuted? By no means. We shall say that these three species were the only forms among the many mutants which were sufficiently well adjusted to survive. And we shall say the same if there is only one species (or none). Thus Darwinism does not really predict the evolution of variety. It therefore cannot really explain it. At best, it can predict the evolution of variety under "favourable conditions". But it is hardly possible to describe in general terms what favourable conditions are--except that, in their presence, a variety of forms will emerge.And yet I believe I have taken the theory almost at its best--almost in its most testable form. One might say that it "almost predicts" a great variety of forms of life. In other fields, its predictive or explanatory power is still more disappointing. Take "adaptation". At first sight natural selection appears to explain it, and in a way it does; but hardly in a scientific way. To say that a species now living is adapted to its environment is, in fact, almost tautological. Indeed we use the terms "adaptation" and "selection" in such a way that we can say that, if a species were not adapted, it would have been eliminated by natural selection. Similarly, if a species has been eliminated it must have been ill adapted to the conditions. Adaptation or fitness is defined by modern evolutionists as survival value, and can be measured by actual success in survival: there is hardly any possibility of testing a theory as feeble as this.
(105) As already mentioned, Popper had earlier (Popper 1961) recognised the danger of misconception 3, and the consequent need to keep adaptation and fitness (selection) clearly separated; but in this later consideration of the problem he seems to have decided that such separation cannot be achieved. Viewed as a putative theory of the growth of adaptive complexity (i.e. of the evolution of a "rich variety" of more or less "well adapted" forms), Darwinism then becomes irrefutable (metaphysical), for, no matter how little the variety or adaptation we observe, it could still result from Darwinian processes.
(106) Clearly, I agree with the essence of Popper's argument; the difference is that instead of discarding Darwinism, I discard the idea that adaptation and fitness be defined in terms of one another. Granted, Darwinism cannot then "predict" the growth of adaptation or adaptive complexity. But, once adaptation is interpreted in terms of knowledge this becomes precisely consistent with Popper's general evolutionary epistemology, and is seen as the strongest kind of theory we can expect. And, as a theory of the historical growth of adaptive complexity in the biological world, it is perfectly scientific.
(107) I have expended some effort in considering
Popper's criticism of the scientific status of
Darwinism because I think it relates to some difficult
and important issues, which are relevant objectives of
this article as a whole. However, it must be added that
Popper himself subsequently modified his views
significantly (Popper 1978).
In particular, Popper recognised and corrected
the error implicit in tautology misconception 2, and
accepted that (Organismic) Darwinism can be so
formulated that it is definitely not tautologous.
Unfortunately, he then goes on to say that, in such a
form, it is literally false (has been refuted). At
first sight this is an even worse accusation than the
original assertion that the theory was (almost)
tautologous and/or metaphysical. However, Popper's
reformulation is the strong one that all
aspects of the phylogenetic tree are the outcome of
cumulative natural selection; such a strong claim,
which is essentially equivalent to the adaptationist
programme criticised by Gould and Lewontin, is, indeed,
false. Popper does not explicitly consider the lesser
(but still non-tautologous) claim that all instances of
the growth of adaptive complexity (i.e. my
) are the outcome of cumulative natural
selection (i.e. my ), and certainly does not
argue that this formulation has been refuted;
so there is, in fact, no conflict with the views I have
expressed.
(108) In conclusion, let me reiterate that I consider
, viewed as an historical theory of the
growth of adaptive, organismic, complexity, to qualify
as a good scientific theory-- despite the
fact that actually testing it in specific cases is
enormously difficult. More specifically,
qualifies as scientific according to Popper's
criteria. I emphasize this last point because, even
though Popper might be called a "naive"
falsificationist with respect to the logic of
(scientific) theories, he has always been a critical
falsificationist with respect to the
methodology of actually carrying out
scientific research. This point has, apparently, been
commonly misunderstood or misrepresented (Magee
1973, pp. 23-24; Popper
1974b, pp. 981-984). The
relevance of the distinction here is that it can be a
perfectly consistent Popperian position to assert that
Darwinism is scientific by virtue of its (logical)
falsifiability, even if such falsifiability is
methodologically almost impossible to exploit
(i.e. tests which could falsify the theory may be
perfectly conceivable, yet wholly impractical).
6 Epilogue: Artificial Life
5.1.5 Misconception 3: Adaptation
The Critical Rationalist Vol. 01 No. 04 ISSN: 1393-3809 31-Dec-1996
Copyright © 1996 All Rights Reserved.
TCR Issue Timestamp: Tue Dec 31 17:37:08 GMT 1996