What Darwin Got Wrong

Jerry Fodor & Massimo Piattelli-Palmarini (2010)

Brace yourself for a more extensive summary than usual.

Boisterous diatribe to the contrary, ‘science’ is far from free of taboos. Nietzsche remains right about that. Examples: questioning the potential merits of astrology, or links between ‘race’ (however ill-defined) and particular traits. First off, kudos to J&M for having the cajones to put this work out, under such a provocative title. It is hard, even for established intellectuals, to criticise the theory of Darwinian evolution and escape ostracism. And as it turned out, J&M did not escape their un/fair (?) share of ostracization.

J&M think that the Theory of Natural Selection (‘TNS’) is flawed. They come at it with a dual-barreled attack, and the two barrels can be quite cleanly divided. The barrel that takes up most of their narrative concerns whether TNS has what it takes to be a proper theory. They don’t put it in these foregoing terms, though: if the authors are aware just how heavily the shots fired through this first barrel weigh on their views on what-it-takes-to-be-a-proper-theory, then they sadly don’t show it.

Put as curtly as I can allow myself to get away with, their issue numero uno rests with what they regard as a lethal indeterminacy in TNS’ model of causal operation. To wit, TNS’ inability to distinguish actual fitness-enhancing traits from those that are merely free-riding tag-along ‘correlates’ that accompany phenotypic variation. Think of Gould-Lewontin’s arches and spandrels. Phenotypes embody (J&M seem to think of it closer to ‘constitute’, à la J. S. Mill, but I think here they are merely following what biologists have already been doing for ages) traits, and some traits may and others may not enhance fitness. TNS is helpless when it comes to predicting which is which, and evolutionary biologists do so only by committing the heinous sin of circularity, by conveniently pouncing on chosen traits ex post. The authors do not shy away from taking this line of reasoning all the way to its extreme: rather than a theory, TNS is much more akin to Natural History, just one damned thing after another.

Philosophically, the most creative move J&M pull is to establish an analogy between TNS on the one hand, and Skinner’s now deeply unpopular theory of learning by Operant Conditioning on the other. I am disappointed to see so few (none that I am aware of, in fact) critics levelling their sights on J&M from this particular flank. I think the reason is, partly, that virtually nobody in the field of behavioural science has skin in the game; not to mention that probably few biologists are sufficiently well versed in the inns and outs of Skinner’s theory. Anyway, the story goes that the structures of Skinner’s theory and TNS map neatly on to one another, and that if Skinner’s theory is wrong on ‘principled grounds’, which everyone who matters agrees it is, then TNS must be wrong for structurally similar reasons. And it must be said that the authors do a persuasive job of drawing the analogy between the two theories – applause warranted, Philosophy of Science at her best! What they do not do, unfortunately, is carry out the task of demonstrating where and how exactly Skinner’s theory falls flat with any degree of commensurate rigor. I guess I’ll have to read Chomsky to bottom that out, because what J&M had to say about it was less than convincing. What is clear is that the issue has to do with the individuation (of stimuli and responses), and how a habit-forming mechanism would be able to pick between material aspects of each. Was it the triangular shape or the yellow colour that accounts for the yellow triangle acting as a stimuli? Was it running (moving?) to the right, east, or anti-clockwise that made up the learned response? Don’t ask me why we can’t trust cognitive systems to disentangle these on their own …

Far more interesting to me is J&M’s second barrel. There exist serious endogenous forces that shunt natural selection in particular directions. The phenotypic possibility space is highly constrained: by physics (especially thermodynamics), by what-can-and-what-cannot-be-coded, by the fact that ‘traits’ come in packages, and by ontogeny (the developmental history of an organism). Once in an evolutionary branch, inhabiting a particular environment, there are only so many places for evolution to go even before selection pressures come into play. But here’s a question for J&M: to what extent to we need to treat these considerations as anything more than passive constraints on the ‘pseudo-random phenotypic variation generating mechanism’? Still, there’s a wellspring of profundity here that deeply resonates with me: it may well turn out to be the case that our Universe is awash with upright-walking bipedal hominids for reasons that don’t exactly align with the Theory of Natural Selection as currently understood.

When all is said and done, I believe J&M’s actual contribution is – unwittingly – to Ontology. How are traits individuated? If we were to play it safe, and have TNS do away with ‘traits’ altogether and stick to phenotypic variation as the natural kind object, would that make a difference? Some phenotypic variations would be fitness-enhancing within certain ecological contexts, others would not, and the former would get selected. Who then cares about ‘superfluous’ traits? Surely, this would buy us explanatory power, at least? Perhaps it is too much to ask more than this from a theory of evolution. The number of ‘variables’ at play during natural selection, or the nature of the link between phenotypes and fitness (e.g., superveniency – see Alexander Rosenberg’s thinking on that) may preclude predictive power. But it seems that, by J&M’s bar, many so-called ‘theories’ would face severe demotion. Ultimately, this book perhaps says more about the problematic ontological standing of science and ‘lawful determinism’ than about Darwinian evolution per se.

Leave a Reply

Please leave these two fields as-is:

Protected by Invisible Defender. Showed 403 to 135,827 bad guys.