Foldit

Now you can play a game to help scientists fold proteins. When I first saw this being linked I was sceptical that players were really folding proteins — I suspected that the game was just Rosetta@home with a small game tacked on that had nothing to do with the actual protein folding but would keep the computer user occupied while the Rosetta@home ran.

And it seems that Foldit players aren’t helping to fold proteins in the direct sense either:

Can humans really help computers fold proteins?

We’re collecting data to find out if humans’ pattern-recognition and puzzle-solving abilities make them more efficient than existing computer programs at pattern-folding tasks. If this turns out to be true, we can then teach human strategies to computers and fold proteins faster than ever!

So they intend to collect data from the game to help them spot heuristics humans use that can then be implemented in computer programs. They aren’t going to take the results of human protein folding as the ‘correct’ structures.

Intuitions about life

I’m finding Peter van Inwagen’s Material Beings annoying so far. It annoys me because it evinces an unhealthy respect for common intuitions, sometimes to the extent of declaring conflicts of said intuitions with facts to be philosophical problems. I tend to be only too ready to jettison the intuitions in such cases, but some philosophers go on to derive some outrageous conclusion or other by way of solving the apparent problem.

I don’t intend to attack van Inwagen’s argument for his outrageous conclusion that “there are no tables or chairs or any other visible objects except living organisms”. I’m only 1/3 through the book so the argument hasn’t been fully laid out yet. But the book got on my nerves early on with some heavy intuition-reliance on several probably minor issues. And I’m blogging this now because I’ve just read a particularly irritating passage that relies on some rather naive biology. Again, this reliance might turn out to be incidental to his main argument; I don’t know. But it’s certainly not incidental in the context of van Inwagen’s methodology in this book.

The offending passage comes in the context of van Inwagen’s explanation of how ‘lives’, the spatial-temporal processes that are biological organisms, are fundamentally different from non-living events:

…there is an interesting and important feature of lives that is not shared by waves. Consider two waves… which are moving in opposite directions and which pass through each other. A still photograph taken at the moment the waves coincide spatially will show what seems to be one wave whose amplitude is the sum of the amplitudes of the two coincident waves. I think we must say… that both the waves exist at the moment of superposition and that each is at that moment constituted by the activities of the same water molecules. We may describe this possibility — the possibility of two waves’ being simultaneously constituted by the activities of the same objects — by saying that a wave is not a jealous event. Lives, however, are jealous. It cannot be that the activities of the xs constitute at one and the same time two lives. Lives are, in fact, so jealous that only in certain special cases can two lives overlap… The only clear case, in fact, is the case in which one of the lives is subordinate to the other, as the life of one of my cells is subordinate to my life. (…the only possible case… I think, would be the case in which the activity of the ys constitutes the life of a cell and the activity of the xs constitutes the life of a multicellular organism. I doubt whether there could possibly be xs and ys such that the activity of the xs constitutes a life, the ys are properly among the xs, and the activity of the ys constitutes the life of, say, a hamster.)

Here’s an example of non-jealous lives that is not a case of small units of life being ‘subordinate’ to a larger unit: the bacteria in human guts. I claim that they are not subordinate, because although they are useful to humans, they are also ‘free’ to serve their own [reproductive] interests; it is arguable that they are not really ‘our’ cells — they don’t share their host’s genetic material, for one. It would be surprising if, among the microorganisms in our gut, there aren’t any that are mere free-riders who don’t add to our [evolutionary] fitness. Yet, I claim, that the processes constituting these bacteria are part of the overall process constituting a human life. Given that the bacteria outnumber cells containing human DNA in a given human body, and given the importance of the bacteria to normal human digestive functions, it does not seem appropriate to exclude these bacteria from the overall process of a human life. (If you need one, and even more spectacular example of dependence on bacteria is that of the Siboglinid tube worms, which are completely dependent on bacteria for their nutrition.) To put it in van Inwagen’s vocabulary, if we let ys = gut bacteria, then it seems like his ‘hamster’ example has been found, except we have bacteria instead of a hamster. Isn’t it correct to say, in some sense, that “the activities of commensal gut bacteria constitute at one and the same time the lives of humans and bacteria”?

There is nothing to stop the indignant metaphysician from just insisting that the objects constituting the life of a human are the tiny minority of cells that contain human DNA, and everything else is other lives. But I would gently suggest that the biologically messy notion of a “life”, at least going by van Inwagen’s choice of functional individuation, may not be the best starting point from which to build a complex metaphysical thesis. Is there perhaps a better way that he could have defined a “life”? Well, there isn’t one that’s obvious to me. If he chooses to define it my similarity of genetic material, he runs into the problem of having to say that a pair of genetically identical twins constitutes the “same life”.

I did not mention the long, elaborate analogy he had used to sketch his functional definition of life. This right after saying that “it is the business of biology to answer this question [of what a life is]“. I just don’t see the analogy as enlightening everything. It seems instead to expose the weaknesses of his reliance on an assumed clear notion of what a life is.

Another strike against van Inwagen’s claim that a life is a “reasonably well-individuated event” comes from an example that he uses to demonstrate what are not reasonably well-individuated events. The example is that of a flame that spreads into several spatially separated fires. These, he claims, are not well-individuated. But it’s difficult to see how, if we consider only their level of individualization, a kind of unicellular organism that reproduces by simple binary division could be said to have members that are any better individuated. Since the binary division requirement implies that the ‘event’ that begins with the very first cell of this kind of organism is spatially and temporally continuous with all the descendants of that cell, what’s stopping us from calling saying that every cell thus reproduced is a part of a larger mega-organism? How are the individual cells any better individuated than the flame which splits into seven flames?

Now, maybe that is a clear biological definition of a life to be had after all. But on current knowledge, I wouldn’t base my entire metaphysical project on anything as uncertain as the individual nature of lives. I suspect there is significant potential of us altering our ontological terms as we learn more biology, so our current usage of the word “life” need not necessarily reflect any fundamental fact about actual lives, if such even exist.

Some JEB commentaries on group versus kin selection

Ever since I read The Selfish Gene, I’ve been wondering what exactly group selection could bring to the table of theoretical population genetics that kin selection could not. Every case of group selection observed in nature that I read about turned out to involve groups of individuals that have a higher-than-average relatedness within their group (compared to the rest of the population). So why are the group selectionists still going strong?

A recent exchange of commentaries in the Journal of Evolutionary Biology, all free access, seemed like they could go some way to answering my question. They started out with West, Griffin and Gardner putting forth a new classification system for social behaviours, particularly those that are broadly regarded as ‘altruistic’. They devote one section of their paper towards ironing out the ‘semantic confusion’ created by the group selection literature. First, they distinguish the ‘old’ and ‘new’ group selection theories. The old one, associated with Wynne-Edwards, treated the group as the only level of selection — behaviour that was ‘for the good of the group’ is selected for, and groups were selected at the expense of other groups — selection was interdemic. The ‘new’ group selection, in contrast, involved intrademic selection: it considered small groups of individuals that had mutual interactions, and showed that cooperative behaviour within those small groups could be favoured. West et al put it this way:

the new group selection approach looks at the evolution of individual characters in a group structured population, whereas the old group selection approach looks at the evolution of group characters

West et al point out that the ‘old’ group selection works under only extremely restrictive conditions. The ‘new’ group selection, on the other hand, can be interpreted as kin selection — group-selectionist and kin-selectionist descriptions of it are mathematically identical. The implicit suggestion is that there is no point in keeping the group-selectionist description: why not leave everything to kin selection and individual selection?

Unsurprisingly, prominent group selectionist David Sloan Wilson objects to the above account of group selection. Strangely, his reply involves a lengthy recap of the history of group selectionist ideas. Less strangely, he uses the catch-phrase ‘pluralism’ all the time.

Wilson doesn’t try to deny that every plausible case of group selection discovered so far can be reinterpreted as kin selection. He maintains, however, that group selection has brought new insights to population genetics, citing the examples of population viscosity and human cooperation. In particular, he claims that group selectionist perspectives were important in originating modern lines of research on those two issues.

In their reply to Wilson’s commentary, West et al concede that group selection made a pioneering key contribution to models of population viscosity. Their comeback, however, is that later approaches using kin selection models were able to provide analytical proofs that Wilson’s group selection approach could not. On the issue of whether the effect of local competition can be overcome by dispersing in small groups (‘buds’), they point out that kin selection models have managed to provide a solution with just “a few lines of algebra”, whereas group selectionists had long lamented that the problem was too mathematically complicated.

In response to Wilson’s charge that group selectionist perspectives provided new insights into human cooperation, West et al have a similar reply: briefly, that group selectionist approaches have not been able to derive analytical solutions where kin selectionist approaches were able to. They further charge that on this issue, “the group selection approach has failed to clarify the underlying selective forces, and has led to confusion”. They believe that kin selection better isolates these ‘underlying selective forces’. For example, kin selection has demonstrated that “punishment or strong reciprocity are not alternative evolutionary explanations for cooperation, as had been implied, but merely specific mechanisms for providing direct or indirect fitness benefits to cooperation” — these ‘fitness benefits’, rather than the punishment or strong reciprocity in themselves, are the ‘underlying forces’ of selection.

As I’d hinted at earlier, I was uncomfortable with Wilson’s heavy citation of history in his defence of group selection. For even if group selection has historically been a pioneering approach to many issues in evolutionary biology, it doesn’t follow that it’s still worth keeping around as an alternative approach. To show the latter, you really have to bring up examples of issues in which group selection has produced results that kin selection has not been able to. And it seems that West et al have responded admirably to Wilson’s examples by countering with kin selectionist approaches to them that have not just matched but bettered group selectionist approaches.

It probably reflects my physicist-bias that I also find West et al’s argument that group selection should be dropped for reasons of theoretical unification persuasive. It seems that some element of above-average relatedness between cooperators is always present in successful group selection scenarios. The natural move is then to isolate this element as ‘underlying’ group selection, which is exactly what kin selection does. And even on an instrumental level, kin selection seems to beat group selection. Sure, there was a time when group selection had its nose ahead of kin selection on certain issues. But many abandoned scientific theories also at some point or another were the ‘leading’ approaches in their field. We should embrace pluralism only if it’s going to be useful to us now.

WEST, S.A., GRIFFIN, A.S., GARDNER, A. (2007). Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. Journal of Evolutionary Biology, 20(2), 415-432. DOI: 10.1111/j.1420-9101.2006.01258.x

WILSON, D.S. (2007). Social semantics: toward a genuine pluralism in the study of social behaviour. Journal of Evolutionary Biology DOI: 10.1111/j.1420-9101.2007.01396.x

WEST, S.A., GRIFFIN, A.S., GARDNER, A. (2007). Social semantics: how useful has group selection been?. Journal of Evolutionary Biology DOI: 10.1111/j.1420-9101.2007.01458.x

Fodor’s anti-Darwinism

The structure of Fodor’s argument against Darwinism (a paraphrase of the summary in his paper):

1. Explaining the distribution of a phenotypic trait in a population requires determining which trait[s] were selected for, not just which traits happened to be selected. In particular, it requires determining which of two coextensive traits was the one that was selected for.
2. To determine which of two coextensive traits was the one selected for, we have to consider counterfactuals about which of them would be selected for in possible worlds where they are not coextensive.
3. Such counterfactuals can be answered only if there (a) there is an agent that does the selecting, or (b) there are laws of selection.
4. There is no agent that does the selecting in Darwinism.
5. Due to contextual sensitivity, there are probably no laws of selection.

Conclusion: Darwinism cannot explain the distribution of phenotypic traits in a population.

As I see it, 4. is the only premise that is relatively unproblematic. 1., I think, indicates how Fodor misunderstands Darwinism/natural selection. It is simply untrue that “explaining the distribution of a phenotypic trait in a population would require a notion of ‘selection for’ a trait.” This had already been pointed out to Fodor by Coyne and Kitcher, but evidently he didn’t accept their point. It is quite possible to explain the distribution of traits in a population by how certain genes/individuals/groups (pick your favourite level of selection) were simply [plain vanilla] selected. I doubt that biologists who use the phrase ‘selection for X’ really mean ‘selection for’ in the way Fodor construes it — when pressed, they would concede that it really is just ‘selection of’. ‘Selection for’ is more like a useful mode of thinking that is often wrong but produces enough fruitful predictions that it continues to be used.

Perhaps Fodor wasn’t thinking of the sophisticated explanations of the Modern Synthesis of evolutionary biology — perhaps his ‘Darwinism’ is vastly different from that. In his Darwinism, it would seem that distributions of traits can be explained only by their being selected for. It is, in short, indistinguishable from an extreme version of adaptationism. Which would explain very nicely how he claims to have discovered the fatal flaw of Darwinism from a consideration of adaptationist evolutionary psychology (EP). It would also explain why he thinks evo-devo is evidence against ‘Darwinism’. His argument as laid out above may be OK against EP, but not against any non-adaptationist version of ‘Darwinism’. And I doubt (although, as I’m neither a biologist nor philosopher of biology, I may be wrong on this) it’s conventional usage to equate Darwinism with strictly adaptationist natural selection. Indeed, the SEP article on Darwinism (written by James Lennox) does not list adaptationism or anything resembling it as one of its principles. All we get there, in fact, is exactly the kind of ‘story’ that Fodor derided Coyne and Kitcher for providing:

The theory can be set out as a series of causal elements that, working together, will produce the needed transformations.

1. Species are comprised of individuals that vary ever so slightly from each other with respect to their many traits.
2. Species have a tendency to increase in size over generations at an exponential rate.
3. This tendency, given limited resources, disease, predation, and so on, creates a constant condition of struggle for survival among the members of a species.
4. Some individuals will have variations that give them a slight advantage in this struggle, variations that allow more efficient or better access to resources, greater resistance to disease, greater success at avoiding predation, and so on.
5. These individuals will tend to survive better and leave more offspring.
6. Offspring tend to inherit the variations of their parents.
7. Therefore favorable variations will tend to be passed on more frequently than others, a tendency Darwin labeled ‘Natural Selection’.
8. Over time, especially in a slowly changing environment, this process will cause the character of species to change.
9. Given a long enough period of time, the descendant populations of an ancestor species will differ enough to be classified as different species, a process capable of indefinite iteration. There are, in addition, forces that encourage divergence among descendant populations, and the elimination of intermediate varieties.

Note that there is nothing in this account that says that the every distribution of every phenotypic trait can be explained by natural selection, or even that most such distributions can. And because this account does not require traits to be selected for, we don’t have to care about answering counterfactuals about possible worlds where polar bears are white but do not match their environment. Hence we don’t need to have laws of selection. In fact, Peter Godfrey-Smith’s reply to Fodor was that there aren’t laws of selection, but not every scientific theory needs to have laws, so it’s not a problem that natural selection doesn’t. Fodor’s reply to him reveals what I think is an appalling lack of understanding of the empirical status of natural selection:

I’m puzzled as to what Godfrey-Smith takes the substance of the theory of natural selection to be. ‘ [W]e can think of the textbook as describing abstract processes, describing actual-world mechanisms, and also bringing the two together… ’ (p. 35). So, is adaptationism merely the thesis that speciation is the product of [some or other] mechanisms and abstract processes? Cf. ‘ we had to destroy it in order to defend it ’ . With such friends, Darwin doesn’ t need enemies.

I have no idea why he feels that “merely” is appropriately used there — why he thinks Godfrey-Smith’s description of the theory of natural selection is demeaning. Any set of “mechanisms and abstract processes” that gives us the quantity and quality of predictions about empirical observations that we have from natural selection should not be brushed aside with a blithe “merely”.

And what of Fodor’s claim, near the end, that although explanations via natural selection aren’t nomic, they can be ‘salvaged’ as ‘historical narratives’? Although I agree that they probably aren’t nomic, I would have to disagree that they are necessarily post hoc. The theory of natural selection is not solely concerned with explaining the past. It also does an admirable job of predicting future observations, as admirably (dare I say) as the likes of quantum mechanics. It’s not just that when such explanations work “they provide plausible historical narratives”. It’s a lot more than that.

Turns out Fodor has a response to the ‘but it works’ line as well (nestled in his reply to Dennett):

‘But it works’. That’s not obvious; in fact, it’s what is in dispute. To be sure, the theorist can often distinguish between confounded variables; but it doesn’t follow that the theory can since the theorist has much more than the theory to go on: In particular, he has access to all sorts of intuitions about the relative plausibility of one or other natural-history scenarios. (That’s why selectionist explanations, like historical explanations, aren’t always just-so stories.) It’s often a delicate matter to distinguish what the theorist knows from what the theory tells him; but it ’ s essential to do so if one is to determine what the data do or don’t say about the confirmation of theory.

Too bad he doesn’t say anything about how the theory works beyond providing plausible historical narratives. People aren’t just working on ‘intuitions’ about the mere plausibility of natural scenarios. They have hard, quantitative evidence, gleaned from paleontology and genetics.

I’ll cap off with an amusing observation: In their replies, both Sober and Dennett take Fodor’s argument as a reductio that should lead Fodor to abandon one or more of his premises. Fodor, however, embraces what Dennett calls the ‘absurd conclusion’. Dennett expends nearly a full page mocking him for that, which only invites an even more childish reply from Fodor:

Dennett fears that scientists won’t take me or my sort of arguments seriously. I might not get invited to the Biology Department ’ s Spring Picnic, which would put a terrible hole in my social schedule. But I am dauntless; if it be so, then so be it. I have spent about fifty years palling around with scientists, and here is what I have discovered: They are a lot like us. That is, they are often precipitous and confused and not reliable as to the significance of their theories and discoveries. It is therefore desirable to distinguish between two quite different methodological principles, one of which I cleave to but the other of which I treat with circumspection. Namely: ‘take the science seriously’ and ‘take the scientist seriously’ . Often enough, in my experience, doing the one precludes doing the other.

Well, I hardly think that scientists have the last say on everything, but in this matter it really is a case of Fodor not taking the science seriously either — if he did he would not condemn all the correct predictions the theory of natural selection has made about genetics experiments (say) as ad hoc “historical narratives”.

Hah

When Fodor’s bizarre argument for the incoherence of Darwin’s theory of natural selection first appeared, I remarked that he seemed to be conflating extreme adaptationism (a la some quarters of evolutionary psychology, usually marked out by the capitalized label Evolutionary Psychology) with natural selection. Now he’s published a more comprehensive screed in Mind & Language, where he begins by narrating how he came to realise the flaws of adaptationism in general when making arguments against Evolutionary Psychology. I haven’t read through the whole thing yet. I must confess to having read the replies to his screed from Dan Dennett, Peter Godfrey-Smith and Elliott Sober before even dipping my toes into Fodor’s paper. Interestingly, they each have different accounts of what went wrong in Fodor’s argument, and some accounts aren’t consistent with others. Dennett’s reply was in his usual caustic style, which I semi-guiltily enjoy reading even if it doesn’t make his arguments better. My favourite sentence from his commentary: “[T]he explanation of heterozygote superiority in the case of sickle cell anemia is not a matter of people competing with each other in a great malaria tournament.”

I’ll blog more details of their arguments and Fodor’s after I finish reading Fodor’s paper and his reply to them.

Fodor, Again

The latest exchange between Fodor and a large chunk of his detractors, regarding his claim that natural selection provides incoherent explanations for the traits it claims to explain. Unfortunately, both sides seem to be talking past each other again.

I wasn’t sympathetic to Fodor’s stance the last time I read the back-and-forths on this issue. And I still have the same problem with his latest: Why does he think that natural selection must distinguish between the hypothesis that polar bears were selected for being white and the hypothesis that polar bears were selected for matching their environment? And even if that is an important question, doesn’t natural selection answer that? Natural selection says that a variant trait that enhances reproductive success will be selected for. Whiteness enhances reproductive success in a white environment, but probably not in most others. Camouflage enhances reproductive success in all environments. So camouflage (environment-matching) was the trait selected for.

In general, I think all scientific theories are incapable of distinguishing between some historical hypotheses. Why exactly is it so important to distinguish between those that Fodor picks? We are interested in explaining the polar bear’s current appearance, which is both white and matches its environment. If whiteness was selected for, then that explains the matching of the environment too, since the two traits are coextensive. If the matching was selected for, then that explains the whiteness too. Whichever way it is, natural selection explains it. It doesn’t explain everything, though. If we can’t find evidence that the trait in question is reproductively advantageous, then we would not insist that natural selection explains it. Only extreme adaptationists would do that, and natural selection can do just fine without extreme adaptationism.

But, but, the revolution is coming! We’re all too blind to see it, that’s all. Fodor concludes:

I am, to be sure, in danger of having insufficient ‘acquaintance with the biological theory that [I aspire] to replace’; but I’m prepared to risk it. A blunder is a blunder for all that, and it doesn’t take an ornithologist to tell a hawk from a handsaw. Tom Kuhn remarks that you can often guess when a scientific paradigm is ripe for a revolution: it’s when people from outside start to stick their noses in.

Fodor’s Strange Picture of Natural Selection

The flurry of replies to Jerry Fodor’s screed against adaptationism in the LRB prompted me to go back and re-read the article more carefully. Fodor’s ultimate target seems to be evolutionary psychology, which he takes to be founded on adaptationism. So he launches a none-too-original attack on adaptationism, but can’t resist inserting hyperbole like

…an appreciable number of perfectly reasonable biologists are coming to think that the theory of natural selection can no longer be taken for granted. This is, so far, mostly straws in the wind; but it’s not out of the question that a scientific revolution – no less than a major revision of evolutionary theory – is in the offing.

Unsurprisingly, Jerry Coyne and Philip Kitcher have an acid reply to this assessment of natural selection. I largely agree with their defence of natural selection, and share their puzzlement over the import of the so-called conceptual problems Fodor locates in natural selection.

Take the following ‘problem’ Fodor raises:

The crucial test is whether one’s pet theory can distinguish between selection for trait A and selection for trait B when A and B are coextensive: were polar bears selected for being white or for matching their environment? Search me; and search any kind of adaptationism I’ve heard of.

I don’t see why it’s important whether polar bears were selected for being white or for matching their environment. And my efforts to understand why Fodor should think this is important aren’t helped by Fodor’s mystifying reply to Coyne and Kitcher along the lines that adaptationism cannot tell us “whether purple polar bears would have survived in the ecology that supports ours”. Purple polar bears neither are white nor match the environment in their ecology. So surely adaptationism can tell us that purple polar bears would most likely have not survived in the Arctic today.

But, Fodor says, for the purposes of determining if natural selection is conceptually coherent, we don’t care about whether we can tell if polar bears were selected for being white or for matching their environment. Instead,

The problem is that it makes no sense at all to speak of the aspect of a causal history that selection focuses on; to say (as it might be) that selection focused on the whiteness of the polar bear rather than its match to the surround. Selection doesn’t focus: it just happens.

But it seems to me that Coyne and Kitcher are making exactly that point. That selection “just happens”. The second and third paragraphs are devoted entirely to explaining how selection can just happen without any concept of ‘for’:

The concept of ‘selecting for’ characteristics is largely a philosopher’s invention, one put to hefty work by philosophers of mind and language in particular as they strive to understand how psychological states can have content. Fodor knows all this, but he seems to know nothing about the way the notion of natural selection has been used in evolutionary explanations for the past 148 years.

Darwin would have seen the history of the polar bears along the following lines: some ancestors had different versions of the hereditary material that caused them to be paler than their fellows; this difference caused them to be less visible to their prey in their Arctic environment, and thus to have an edge when it came to hunting; that edge made them more successful in leaving descendants who inherited the fortunate variation. After Mendel, Thomas Morgan, Watson and Crick, we can do better: the ancestral bears had some difference in their DNA (perhaps a mutation or a gene rearrangement); that difference led to a difference in the type or expression of proteins affecting the biochemistry of hair follicles; that difference led to paler fur and a better match to the surroundings, producing greater prowess in hunting and increased reproductive success. Nobody has to decide if there was selection ‘for’ the modified DNA, or ‘for’ the protein differences, or ‘for’ the different organisation of the cells, or ‘for’ the whiteness, or ‘for’ the camouflage.

I don’t know how right Coyne and Kitcher are that “the concept of ‘selecting for’ characteristics is largely a philosopher’s invention”, but I certainly don’t understand why Fodor thinks it’s so central to natural selection. I think biologists can perfectly well talk about traits being preferentially propagated without invoking a process of ‘selecting for’. I’m not convinced that biologists have to speak of traits being ‘selected for’, as opposed to merely using the phrase as a convenient shorthand for ‘being preferentially propagated under blah blah blah conditions’.

This notion that a ‘selecting for’ concept is central to natural selection could explain why Fodor seems to think that his criticisms of evo-psych-type adaptationism undermine natural selection. In fact, it seems to me that Fodor conflates adaptationism a la David Buss with natural selection. The whole excursion into spandrels and evo-devo are relatively old (and valid) criticisms of extreme adaptationists who take every phenotypic trait as an adaptation to the environment. But, as Coyne and Kitcher point out, evo-devo is consistent with natural selection, as is, in general, the idea that certain phenotypic traits are historical artefacts of evolution that need not be adaptive. However, like Fodor, Coyne and Kitcher seem to freely interchange the uses of ‘adaptationism’ and ‘natural selection’. The latter is perfectly consistent with Fodor’s insistence that we consider non-adaptive causes of phenotypes, but extreme adaptationism is not. Fodor’s piece would probably be a rather standard critique of adaptationism. What raised such a flurry in the ‘dovecote’ (as Fodor cutely puts it) is his grand conclusion that natural selection is likely to be overthrown as a theory of evolution.

Fodor, in his reply, claims to recognise that his critics tend to be of the opinion that “Fodor is, of course, right about EP; but he’s wrong about natural selection at large”. Well, I do think he’s right about EP, but more broadly, I think he’s right about adaptationism. I just don’t construe natural selection as requiring adaptationism. Fodor doesn’t formally define what he means by adaptationism in that article. The closest he comes to it is the following (and that only after using the term about ten times beforehand):

Adaptationism is a species of what one might call ‘environmentalism’ in biology. [...] The basic idea is that where you find phenotypic structure, you can generally find corresponding structure in the environment that caused it.

Well, I’m not sure that natural selection need commit to the ‘generally’ part of that statement. I think even Darwin must have understood quite well that “pigs lack wings because there’s no place on pigs to put them”. To paint natural selection as claiming that ‘most’ phenotypic traits correspond to some structure in the environment seems to be setting up a straw man. Yes, extreme adaptationism has severe flaws. But natural selection lives on.

See also Brian Leiter’s post on this kerfuffle.

The Evolution of Fauna in Madagascar

Thanks to an early exposure to Gerald Durrell’s writings and Douglas Adams’ Last Chance to See, I’ve always been fascinated by the wildlife of Madagascar. So this recent paper by Dewar and Richard, “Evolution in the hypervariable environment of Madagascar”, caught my eye. Their central thesis is that many of the peculiarities of fauna on Madagascar is due to the unusually high intraannual variability of rainfall in Madagascar. The interannual variability of the rainfall of these stations, on the other hand, was not significantly different from that of continental African stations that had similar mean annual rainfalls.

The authors argue that this environmental variability accounts for both the unusually extreme iteroparity and unusually extreme semelparity of certain Malagasy mammal groups. The carnivorous mammals are extremely iteroparous, as are some species of lemurs. At the other extreme, the tenrecs are extremely semelparous, as (again) are some species of lemurs. Comparing these mammal groups with their relatives in the same taxa, the authors conclude that such extremes of semelparity and/or iteroparity are unusual for their respective taxa. The adoption of either of these two extreme strategies can be interpreted as a response to extreme environmental variability. Iteroparity is seen as the ultimate risk-spreading strategy, optimal when adult mortality is low but infant mortality is high — having too many offspring in one go could easily backfire in such conditions, since the whole batch could perish due to externally imposed travails, so parents should spread their bets over as many years as possible. Semelparity is seen as the optimal strategy when adult mortality is high but infant mortality is low, for in such conditions the parents should attempt to have as many offspring as quickly possible before perishing in their probably short lives. For the lemurs and tenrecs, there is evidence supporting the relative rates of adult and infant mortality that are supposed to correlate with iteroparity or semelparity. No such evidence, however, was mentioned for the carnivores.

And that’s about it by way of evidence for the hypothesis that the reproductive strategies of Malagasy mammals are due to the high intraannual variability of rainfall in Madagascar. I must say that while I find it plausible, I’m not entirely convinced. One of my concerns is that I don’t see why intraannual variability in rainfall would have so much more impact (at least with regards to reproductive strategies) than interannual variability in rainfall. Presumably there are places in continental Africa with considerably more interannual variability in rainfall than Madagascar has. Prima facie, it doesn’t seem like intraannual variability should be any more damaging or risk-inducing than interannual variability. So why would extremes of reproductive strategies appear particularly commonly in Madagascar and not in those parts of Africa with higher interannual variability of rainfall? A possible answer is that since these other parts are continental, their inhabitants can combat unpredictability by migration, an option not open to Malagasy mammals. Malagasy mammals have to adapt to unpredictability on their home turf, as it were. But is it the combination of isolation and unpredictability that is the explanation, or just the isolation? It would be interesting to study other islands of similar size and isolation to see if Madagascar’s fauna is any more unusual than that of those islands. If the trend of unusual fauna holds across several such locales, then perhaps it is a simple ‘island’ effect rather than a climate effect. Of course, deciding which islands (if any) are comparable in size and isolation to Madagascar is not a straightforward affair. Another interesting thing to do would be to study the life cycles of Malagasy birds. The current evidence, for mammals only, has a narrow enough base to make me nervous about confirmation bias.

The Costs of Massive Modularity: a paraphrase

A more philosophical way to express the concerns I raised about evolutionary psychology’s massive modularity hypothesis:

Evolutionary psychologists argue that modules must exist because for a given problem, it is an evolutionary advantage to have a specialised function in the brain that addresses the problem according to its specific characteristics, and not just as any kind of problem out of the millions that could arise in an organism’s lifetime. Thus, to detect cheaters, it is faster and hence more ecologically rational to do so via a cheater detection module rather than a general reasoning mechanism. However, even if we grant them that for any particular problem it is better, for the purposes of solving that particular problem, to have a brain programme that specialises in solving that type of problem alone, it does not follow that, in order to solve the collection of problems an organism experiences, it is better to have a collection of brain programmes each tailored to solve a different type of problem in the array of problems. For, as I hypothesised, there could be emergent effects in having a collection of functionally separate brain programmes, effects that could themselves create new problems.

The Costs of Massive Modularity

When sociobiology passed on its mantle to evolutionary psychology, it seemed that the emphasis shifted away from rigorous mathematical theories in population genetics to plausibility arguments for psychological mechanisms that do not incorporate selection as thoroughly as the theoretical biologists did. The clean-cut equations of Maynard Smith, Williams, Hamilton, Trivers and Price et al are taken as part of the argument for why evolutionary psychology is plausible. However, the more specific hypotheses constituting evolutionary psychology are not supported by similarly rigorous evolutionary reasoning. The standard Tooby-Cosmides argument for massive modularity in our mental architecture considers all the advantages modularity has over domain-generality, but fails to discuss any possible additional energy costs modularity might incur, or whether the details of genetics and molecular biology could possibly conspire such that it is more difficult to acquire many specific adaptational mental modules rather than one monolithic general reasoning device.

I have not found any literature discussing the energy disadvantages a modular brain might have compared to a domain-general brain. I think it is plausible that there are such energy disadvantages, simply because with massive modularity, most specialised mechanisms are not operating most of the time, coming into action only in certain environmental conditions. Therefore, in some sense, these mechanisms are lying dormant most of the time. A domain-general brain, on the other hand, would not have such specialised mechanisms lying dormant. In short, at any given point in time, a domain-general brain has less redundancy — it might not be using its full range of computational resources, but it doesn’t have a massive suite of information processing programmes that are just standing around twiddling their thumbs, as such. Since our brain is not a solid state flash drive, just maintaining these programmes in a dormant state consumes energy, so a massively modular brain would have to channel energy towards the upkeep of these programmes even when they are not working, which most of them aren’t most of the time. The domain-general brain would seem to be free of this burden. So goes my plausiblity argument for why massive modularity could be more costly than domain generality. I have tried Googling for discussions of the energy costs of massive modularity, but have come up blank. Robert Richards, too, says he doesn’t know of any arguments on this score. I suspect that sociobiologists, if they, rather than psychologists, were the ones spearheading the evolutionary psychology movement, would not tolerate having an evolutionary theory that does not take into account the costs as well as the benefits of a particular adaptation.

I also could not find any research on whether it might have been ‘easier’ (more probable) for organisms to acquire massive modularity (or to acquire domain general information processing mechanisms). My intuition is that there must be some bias, in the landscape of genotypes, towards a particular class of mental structures. In other words, I would be very surprised if, in the topology of genotypes, the measure of genotypes with massively modular mental architectures was equal to the measure of genotypes with domain general mental architectures. Now, of course, natural selection would, if we assume equal energy costs for both alternatives and accept the evolutionary psychologists’ arguments for the evolutionary advantages of massive modularity, tend to favour the genotypes with massive modularity, so measure alone, without consideration of fitness functions, cannot tell us everything. But I do not think it is entirely implausible that massive modularity could have a negligible measure compared to domain generality, and if this were so, it would be plausible that we may attain a local fitness maximum on the set of genotypes coding for domain generality and remain there, simply because the set of genotypes coding for massive modularity was, metaphorically speaking, located in an obscure, tiny and relatively inaccessible part of the landscape of possible genotypes, such that even millions of years of recombination and mutations have failed to transport humans to that location.

Such considerations I find more important than the kind of things the psychologists go on about. I do think that evolutionary psychology could do with an injection of good old fashioned mathematical population genetics. Fuzzy arguments make my head spin.

Addendum: The Sperber article I linked to above discusses the problem of allocating energy amongst modules. The idea is that we do not need all modules to be active all the time, so it makes sense to have an energy allocation algorithm whereby energy is allocated preferentially to the modules that would have the largest cognitive benefits. This brings up another possible way in which massive modularity could be more energy-consuming — it demands that the brain have a informational relevance monitoring system and energy consumption prediction system to allocate energy between modules, and these systems naturally consume energy themselves.