Lucus a non lucendo

A major problem in popularization of evolutionary psychology (pop ev psych? blech…) is a really inadequate understanding of evolution and genetics/heredity. It is silly, and unproductive, to approach evolutionary biology as a series of just so stories explaining how trait X developed so species Y could have more progeny. In the aggregate, evolutionary biology predicts that fecundity will drive genetic diversity (i.e. genes that make you reproductively successful will have a higher frequency in a population), but there is no reason that *every * trait locus must affect fecundity. If I have dark hair (and I do) am I more or less successful than a blonde (depends).

The interesting thing here to me, aside from the generally poor state of scientific education, is a sort of catch-22 implicit in the discussion. As a whole, I want people to read popularizations of science, because I hope it will get them interested and involved in one of the great enterprises of our civilization. However, the terrible oversimplifications that are essentially required in all but the best of these books make the whole thing seem counterproductive.

“Part of the issue here is that any form of psychology makes a poor sociology”

Would you say more about that (or point me to someone who does)? It sounds interesting.

Perhaps the thinking is marred by the whole “survival of the fittest” idea. There was an interesting article a few years back in Phys Rev which noted in passing that the proper model of evolution isn’t best understood as “survival of the fittest” but rather “elimination of the least fit.” While the two ideas sound similar, they lead to different results. The latter, in particular, doesn’t support the “just-so” stories of evolutionary psychology nearly as well.

I’m thinking of population genetics models which use selection coefficients. If you are working on a one-gene-one-trait model, these selection coefficients have a natural interpretation as the “adaptiveness” of the trait in question. If the model has context-dependence, then this natural interpretation can’t be made and it’s very difficult to say anythign coherent about individual traits.

Basically, the citation is to Sobers & Lewontin “Artifact, Cause and Genic Selection”.

Someone who really wants to get Popper spinning in his grave should marry ev psych and punctuated equilibrium.

One common chestnut in p.e. is that some genes spread through a population simply because they’re present by chance in an isolated group that happens to end up with some other reproductive advantage, or in the survivors of some (unrelated to that gene) major weeding-out event. Another is that genes that confer fitness during punctuations may spread through a population despite a neutral or even negative effect during normal periods. (On the micro level, the former has been inordinately useful to students of congenital disease, and that latter has been invoked to explain high blood pressure in african-americans and anomalous body mass among pacific islanders.)

If you invoked a few well-placed catastrophes and migrations (some of them even documented) you could use ev psych to explain pretty much anything, no matter how apparently counter to reproductive success under sensible conditions.

dsquared,

It seems to me that the kind of broad-based critique of sociobiology that you are attempting requires a stronger result than the interpretational worry of Sober and Lewontin. In fact, I can’t quite see what makes their discussion more than a red herring in this case.

Whether or not any explanations offered by sociobiologists (and here I don’t mean only human sociobiologists – presumably a population-genetic critique would apply equally to animals) are good ones, they do not depend on spelling models out in genic terms, even if this is routinely done. What you would need to demonstrate is that one cannot obtain the results of optimisation models (say, ESS approaches) when you take the details of inheritance into account. However, this is not the case. See, for instance, Eshel, 1996, J. Math. Biol., 34: 485-510; Hammerstein, 1996, J. Math. Biol., 34: 511-532; Eshel, Feldman and Bergman, 1998, J. Theor. Biol. 191: 391-396.

There is a lot of shoddy sociobiology out there, but the core enterprise does not seem to be theoretically unsound. Sticking to empirical objections is therefore more effective.

Freud was far, far more careful in his claims than the EP folks Daniel discusses here, none of whom are likely, I think it’s safe to say, to be considered one of the major intellectual figures of their era.

Kevin:

1. I’m not responding to argumentum ad hominem about Lewontin. His paper with Sobers is widely cited in the literature on selection, and not just in the context of sociobiology. Lewontin is a genuine authority in the field. Everyone else is prepared to address his arguments rather than trying to claim he’s “tainted” by having argued against what he considers to be wrong science, so you can bloody well do so too.

2. The brain, like the hand, is a multi-purpose organ. General features of the hand (like having five fingers) are genetically coded, but specific actions of the hand (like twirling a baton) aren’t. There are specific genes which determine whether you’re going to have legs or not, but no specific genes which determine whether you’re going to be able to walk (this is well-established, btw; walking is a learned behaviour and babies who don’t have human adults around to copy, don’t learn).

Brains have many more general functions and fewer specific functions than physical organs, and so more of what we do with them is learned and less is determined by their physical structure. For example, there is a module in the brain which helps babies recognise faces. But it’s not very good. Any pattern of three dots roughly in a triangle will set it off. In order to distinguish faces from triangular dot patterns, babies have to learn.

3. Evolutionary psychology will not turn out to be a useful tool as long as it keeps saying things which are definitely untrue about the anatomy of the brain. Steve Rose has been banging on about this for years, to no obvious effect.

Anno-nymous

1. Unless traits are modular, it is very difficult to see how they can be acted upon through Darwinian selection.

2. Learned behaviours are also relevant to fecundity, and learned behaviours can be inherited through what one might provocatively call “culture” and less provacatively call “childrearing”. This evolution is Lamarckian in its operation.

3. And evolutionary psychology needs the selection to be Darwinian if it is ever to make any interesting predictions. If the dominant force determining human behaviour is the Lamarckian selection of learned behaviours through culture, then the talk of “fecundity” is a red herring; all one is saying is that people try to do things which have had good results in the past.

You need the modularity of traits in order to get the argument through that important human behaviours are the result of “human nature” in the sense of being behavioural traits which have been selected for in the past and which are not subject to any quick method of change.

JP:

1. The sociobiological explanations do depend on some specific features of the models. In particular, they depend on the coding of individual traits being small relative to the whole genotype, so that it is meaningful to talk of selection pressure on individual traits. See my response to Anno-nymous above; the sociobiologists need modular traits for their research programme to be worth pursuing.

Sober & Lewontin’s critique matters because it shows that, in a model which takes account of the realities of inheritance selection coefficients don’t have a causal role. This seriously damages the sociobiologists’ case because they need behavioural traits to be both a) causally active and b) units of selection.

Your own cites seem pretty tangential to me. The fact that optimisation works in poopulation genetics models is hardly surprising. But the point at issue is whether it makes sense to apply these models to behavioural traits as units of selection. It doesn’t.

To make it clearer, we could say that men who wear Old Spice get laid and men who don’t don’t. So wearing Old Spice aftershave is the equilibrium of any model which takes into account both evolution and preference in toiletries. One can agree with this without being committed to any particular view on whether aftershave preference is genetically coded.

I wasn’t suggesting that Lewontin isn’t a genuine authority, only that he’s one guy and his prejudices are well known. It doesn’t mean he’s wrong, but it does mean a good argument needs more support than just him. An argument that relied solely on, say, Noam Chomsky, would probably meet the same fate.

I also think you’re overstating what the ev psych folks say, although I’ll grant that they often give you plenty of reason to do it. It’s not so much that baton twirling is genetically coded, but that our ability to twirl batons might be the result of some other features that are genetically coded. This seems a reasonable argument. It’s also going to lead to lots of crap getting thrown out, but that’s the way it goes.

I don’t have a problem with arguing against the crap. May the fittest argument win. But I don’t think the entire field is necessarily crap simply because some individual results are wrong.

In any case, I’d also argue that ev psych has had a prophylactic effect. Hopefully there are damn few people anymore who argue that human behavior is largely unrelated to biology. That wasn’t the case 25 years ago.

If you were to have a political argument, using Chomsky as your source, then clearly that would be an issue. In his academic field, if I understand correctly, his ideas are very well respected (although by no means unquestioned), and a respectable argument about the development of language could be made using his work as a main resource. I think the overall question, “is ev psych bunk?” is a bit oversimplified, because naturally there are good practitioners and bad ones (I’m certainly no expert). My personal opinion is that breathless reports of “the gene for brain size” or “the gene for intelligence” are dreadful oversimplifications that are worse than useless (see my now ancient in blog years running argument with “Godless Capitalist” for more on my opinions in that realm!) The reason that these ideas get oversimplified and distorted is because the overall conception of what evolution is (survival of the fittest) fails at a basic level when applied to our lives. This conception itself is a gross oversimplification, leading the useless and counterproductive Darwin wars, arguing over things that not only arent in doubt, but are basically giant strawmen. My totally citationless feeling on the subject is that the most popular ev psych books, i.e. the ones that get the most press, are the bad ones, like the Thornhill book or the Bell Curve. Frankly, a good, honest book detailing the science of human heredity’s role in behavior would sell in the single digits.

It’s easy to find examples of bogus speculation based on evolutionary reasoning; depression and rape being but two. If you look at the local bookstore one could disparage any one of our fields.

The increasing recognition of kin selection, which has been criticized but to my mind not discredited, links evolutionary psychology to sociology. I don’t think we should put our disciplines in pigeon holes. Paradigm shifts often occur when seemingly disparate fields merge.

“The only things which are even a candidate for evolutionary explanation are the big drives; eating, reproduction, childcare, status, suppression of same-species aggression.” Huh? Why?

Is there something in our evolution which makes it impossible for us to have wired-in behaviors as complex as those of the social insects? Could you mention it?

(Ants, too, for that matter. If, as you say, evolutionary psychology descends from sociobiology, perhaps it’s not entirely a coincidence that it doesn’t descend from people whose evolutionary and neurobiological worldview rests on psychotherapy and pigeon conditioning experiments…)

It doesn’t follow that our brains have as much detail hardwired in as bees do. But it does seem to make it somewhat plausible, and to falsify arguments which are so general that they claim to show that no evolved organism could.

(As a wild guess — not being a psychologist, but a molecular biologist turned computer programmer — it seems plausible that our brains could have about as many bits of preprogramming as bees do. Then since our brains have a whole lot more space for learned stuff, the preprogramming would tend to be less conspicuous. But I doubt that stops rather complex behaviors from being hardwired in. Similar scaling arguments would apply to birds, and I’ve seen some pretty impressive bird nests which I doubt were learned behavior, and I’ve read about planetarium experiments showing that babies of some species of birds are wired to watch the rotation of the stars from the nest at night in order to figure out what’s a reliable guide to north.)

regarding “lucus”, I think the etymological intuition, if the not the analysis, was correct – “lucus” was originally a clearing in a forest with ritual uses; the sense then transferred to a ritual woody site.

Shai — the fact that we, unlike the !Kung, do not consciously instruct our infants in the arts of sitting and standing, but allow them to acquire the knack of it on their own, does not mean they are not “learning” it. [I have no clue as to the truth of that anecdote about the !Kung but am taking it on your word, that it’s true.] Behavior can be “learned” by observing it in one’s elders or peers without any direct tutelage. Indeed most behavior is — hence the adage “Do as I say, not as I do”, a vain effort to prevent undesirable behaviors from being learned through imitation.

My impression (gleaned, I think, largely from Lewontin) is that the types of explanations offered in EP often fall quite easily to arguments from developmental biology (DB). If I’m remembering correctly, Lewontin condemns casual use of “gene for [some trait]” language without knowledge of the developmental processes that produce said trait. From the DB perspective, there is always a question of the effects of the environment on the expression of any given trait, and these effects can derive not just from the inputs to a developing organism (levels of nutrients, etc.) but also to the simple thermodynamic “noisiness” of the physiochemical surroundings. The argument against many EP theories is then that, if given quantitative expression (when this is even possible), they are too mechanistic to fit the (empirically well confirmed) framework that describes a developing organism. I recommend Lewontin’s _The Triple Helix_ which lays out the specifics behind what I’ve been trying to say, and which is pretty much polemic-free.

I don’t think I’ve ever come across a response by EP folks to these criticisms. Has anyone seen one?

paul orwin:
Gould states explictly that, while he was raised by Marxist parents, he himself was not Marxist. Is there some reason not to take him at his word on this? (Lewontin, of course, is a different matter.)

On blind children learning to walk, my reference is a book I picked up about a month ago called “How Babies Learn” or something – don’t have it to hand. Basically, you have to look after a blind baby and encourage it; unsurprisingly nobody’s really gone through the experiment of abandoning blind babies and seeing if they learn to walk unassisted (actually people have, and the evidence from Thai orphanages is that they don’t). Children with mental handicaps do tend to learn to walk more slowly; Down’s syndrome children walk about a year later than non-Down’s.

Anno-nymous: Thanks very much, and I’m using the italicised quotation format below for clairty rather than to be combative.

We don’t have to say that there is a gene for walking, only that people who learn to walk (presumably) fare better in life than those who don’t. Therefore, we can expect that a gene-complex will develop that leads to the behavior of walking, given the relatively common environment of being raised around others who walk.

Yes. But this gene-complex is just “having legs and a brain”.

This should go without saying — after all, most animals which walk really are genetically programmed to do so, and if humans have lost this trait then it’s only because the environment rendered it unnecessary.

Human beings do walk instinctively; they (apparently) scamper, bent over and with their knuckles hanging down. They just don’t walk in the normal way you see people walking by instinct.

I would agree that the Old Spice case is not a lucendo in the particular constructed example, but in any real world case it would be. This is my whole problem with evolutionary psychology; the principle “People do things that will help them get laid” doesn’t actually help you pick out any specific behaviours. If it did, we’d all get laid a lot more.

Kevin: Most of your comments seem like pious hopes for a better evolutionary psychology (plus you’re digging yourself in deeper on the silly ad hominem argument about Lewontin). There will be no such thing. Evolutionary psychology is not a good faith research program; as Paul Orwin points out, the only real advances in this field have been made by serious researchers working outside the “EP” community. It’s a pedigree-creating exercise for sociobiology, which is in turn a cargo-cult sociology.

And your statement that:

Hopefully there are damn few people anymore who argue that human behavior is largely unrelated to biology. That wasn’t the case 25 years ago.

is just weird as a summary of the behavioral sciences in 1980. Nobody since Freud has argued that human behaviour is “unrelated to biology”, or at least nobody who is not a fictional construct. Lots of people have argued that biological imperatives provide few useful explanations of specific human behaviours, and they’ve argued this because they’re right.

Bill: There’s certainly room in our brains for a lot of “wired-in” behaviours, but as a matter of fact we don’t have all that many (or at least, not compared to the number of learned behaviours we have). The majority of our behaviours are learned rather than wired-in, so analysis of behaviour is more likely to be productive if it looks at how that behaviour might be learned than if it tries to tell adaptationist stories about how it might have come to be wired in.

Shai: There are a number of items on that list which are regarded by developmental psychologists as simply false. Basically everything on the “innate abilities” list is more or less settled to be learned behaviour. (interesting link.

Of the “Innate Social Behaviours”, many of them are developmental stages identified by Piaget. Note also that whoever wrote the list couldn’t resist sneaking “Variation in intelligence (leading to inequalities)” onto it; the conflation of extremely arguable biology with extremely unlikely economic theory is classic Pinker.

Developmental biology also supports plenty of examples where neural plasticity does NOT work: for example, language aquisition (especially phoneme recognition, where you have a brief window of opportunity to train.)

The example of “adapting” visual cortex to braille recognition is interesting, but I wonder if the possibility of normal (perhaps minor) use of tactile recognition is also handled at least in part by the visual cortex. If so, then the plasticity is not as dramatic, not a “water into wine” conversion.

“Nobody since Freud has argued that human behavior is ‘unrelated to biology’ or at least nobody who is not a fictional construct”- such as Sartre or Lacan?

Well, I’m not near as learned as many here, but I’ll offer my $.02 worth. I seems to me that genetically retained behavioral selection applies primarily to instincts. Instincts are neural programs that set off a fixed chain of behaviors given the relevant trigger stimulus. If I recall rightly from ancient reading of Konrad Lorenz, he said that, as neural complexity increases, fixed chains tend to break down and their component behaviors tend to gain an independent drive toward expression. The exampled he cited was house cats. Now it seems to me that as neural complexity increases, animals would tend to acquire an increasing capacity for behavioral learning relatively unchained from fixed instincts based on the implicit grasp of behavioral rules rather than causally fixed processes. But then would not this capacity tend to co-evolve with instinctual endowments, such that those instinctual endowments become inter-nested with requirements for rule-governed behavioral learning, such that the operations of instincts come to depend upon such learning? My speculation would not be at all clear as to the specific selection pressures that would trigger such a line of development, but its specific adaptive advantages would seem obvious.

As for the evolution of “mind”, by which we actually mean the evolution of the brain as an organ with complex interactions between physiological and mental functionnings, it would seem to me that the role of specific, genetically fixed selections would tend to decrease with increases in size/capacity; in other words, that a nonspecific increase in superfluous neural capacity would effectively find its “uses” without being pre-programmed. With respect to human evolution, the growth in brain size is a contingent outcome of bipedalism, which apparently is now being pushed back to 3 to 4 million years ago. The resulting narrowing of the birth canal required neotenic or premature birth to allow the infant’s head to get through without being crushed, which means that embryological rates of growth continue for some time after birth. This, in turn, allowed for further neotenic mutations to extend the term of such embryological growth, leading to increased brain size. Presumably, the increased dependency of the infant and thus the increased dependency of nursing and tending mothers created selection pressures for a further intensification of sociality. (The peculiar evolution of human sexuality, with its year-round estrus, may have evolved so as to increase attachment to the dependent mother-infant dual unit, resulting in the paradoxical configuration that human beings are at once excessively horny and excessively prone to attachment.)

Have y’all heard of the woman who teaches parrots to talk? No, not imitate sounds, but use such sounds contextually and indicatively, very much as efforts to teach great apes sign language have yielded some significant results. The woman’s legit, a professor of psychology I think in the U.of Cal. system, and she’s been at it for about 30 years, with a few generations of parrots. Now parrots are birds and they don’t call ’em “bird-brains” for nothing, though parrots as a genus have the largest relative brain sizes among birds. But still, they lack mammalian brain structures. I would think such results suggest something of the “plasticity” of neural matter. And as for phonemic pattern recognition capacities, there is a case at the Yerkes Lab of a bonobo, who was raised while its mother was undergoing language training experiments, who shows considerable evidence of such speech recognition.

Surely it is way too Aristotelian of me to think this, but, if the terms of supposed explanations of human behavior do not accord with anything that we experience of the complexion of human behavior, I would think that prima facie grounds for finding them dubious.

Perhaps rather than ev psych, relying on speculative hypotheses about what happened way long ago, there should develop a scientific research program for human ethology, taking human behavior as it actually is and considering it in terms of the interpenetration of neuro-biology and culture.

That part needed irony tags.

At least some of the ideas behind EP are true: The evolution of thinking and language helped survivability and were preserved traits. Likewise with sexuality, the emotional trait of getting angry when hurt and the very existance of memory and short-term memory. A whole bunch of very fundamental things. Probably including an inherited hatred of the smell of feces.

So it makes sense to see how far these ideas can be pushed, how much can be proved somehow.

In practical terms, evolutionary psychology will become less important as neurological knowledge increases.

Traits that expensive will not be preserved unless they confer some reproductive advantage. The ability of African blood cells to repel malaria by forming a sickle shape comes to mind as a biological example. Having legs is another.

So expensive behaviors that are preserved across cultures and time are good candidates for having a genetic origin. Possible candidates include pair-bonding, our protective attitude toward babies, being attracted to fertile, healthy potential mates, particular facial expressions, an attraction to living near bodies of water, hatred of insects, fear of heights, fear of death. You can make your own list.

Genetic inheritence will be less useful in explaining behaviors with (apparent) negative reproductive/survival value, such as suicide, homosexuality, neuroses and psychoses. So it makes sense to look for alternative materialist explanation for these behaviors. Many have been offered, none have been proved.

These behaviors with (apparently) negative survival value may be the result of imperfections inherent in the genetic system. Mutations, transcription errors and so forth might be involved. Alternatively, one might look for a reproductive/survival value for these apparently negative behaviors. It’s at least an intelligent question, as these behaviors are expensive and preserved across time and culture.

Proving things is much harder, of course.

Pots, kettles, glass houses…

Sure, some evo-psych (possibly the majority) is bad science.

But anyone who casts stones at evo-psych (on grounds of scientific methodology) while admiring (take your pick) Freud, Structuralist anthropology, Marxist sociology (or for that matter most other sociology), simply has no shame.

“To say, for example, that female preference for a wealthy mate is an adaptation implies that:

1. Female preference for a wealthy mate is genetically encoded, i.e., implemented in the realm of nucleic acids and proteins.”

Not exactly. First of all, the selectable trait doesn’t need to be quite that specific. It’s far more likely that a more generalized behavioral tendancy — like “be attracted to high-status males” — is what’s adaptive. And it will continue to be adaptive insofar as wealth is correlated with high status.

Secondly, there is no need for the specific trait to be encoded in a specific set of genes. Any combination of genes and environment that leads to the behavioral tendancy is enough to make it selectable. All that matters is that it is heritable in some fashion, no matter how indirect.

“2. There are at least two alleles for “female mate wealth preference”, corresponding to more and less (or presence and absence) of this trait.”

I have no idea where this comes from. It makes no difference if the genetic influence is due to a single allele, multiple alleles, a single locus, multiple loci, regulatory elements, or any combination thereof. Perhaps you mean that there must be some variation within the population in order to be selectable. That’s true (or at least it would have to have been true in the past), but that’s a far cry from requiring the trait to be monogenic and diallelic.

Daniel writes:

And the fact that the simple existence of a sex drive is next to useless in explaining any specific human sexual behaviour has been known since very shortly after Freud. It’s the same with all our broad drives.

I would say that Ev Psy explains a lot more than the simple existence of a sex drive. It explains, for example, why men tend to be attracted to younger women but women tend to be attracted to older men. It explains why men moreso than women tend to be concerned with phyical attractiveness in a potential mate. It explains why there is a strong taboo against incest in basically every culture you come across. It explains why polygyny is much more common than polyandry (which is next to nonexistent). And so on.

I think it explains quite a bit, just so long as we realize that it explains tendencies and not specifics. And that the thing about tendencies is that they’re statistical, and that one can always find some exceptions.

One place where self-styled evolutionary psychologists really get themselves into trouble is in trying to explain abnormal psychology. The truly horrid “sad hominid” explanation given by Daniel above is one such example. I think people fall into this trap because they’re used to thinking in terms of adaptiveness, and so they come up with adaptive explanations for things which clearly are nonadaptive. It’s much better to view abnormal psychology as, well, abnormal, and to realize that the human mind is far from perfect, so that enough negative stimuli can throw it pretty far off kilter.