Sociobiology 4: individuals as groups, and a summary

This is the fourth of a series: Part 1, Part 2, Part 3

Wilson and Wilson (W&W) then continue on to employ some recent work on individuals as groups, and the "major transitions" literature.

Ed Wilson is well known for his idea of "superorganisms", in which eusocial hives or colonies of insects (his speciality) are treated evolutionarily as single organisms, resolving a problem Darwin had with these species' evolution. In other words, an ant colony is a fitness bearer. Here, W&W appeal to Lynn Margulis' theory of the eukaryotic cell as a symbiosis between prokaryotic cells to undergird the idea that selection occurs on groups (that is, a group of cells that includes the outer cell and the internal symbionts). This leads directly into the idea that major transitions in evolution occur when "selection within groups is suppressed, enabling selection among groups to dominate the final vector of evolutionary change". Phenomena like chromosomes, meisosis, gamete sequenstration and so on stabilise organisms and prevent them from becoming "a mere group of evolving elements". They quote Okasha:

For by ‘frameshifting’ our perspective downwards, it becomes apparent that individual organisms are cooperative groups, so are the product of group selection!

This "frameshifting", or perspectival focus, is crucial to their view, but I wonder if it tends to undercut their point. If we can simply reconceptualise individuals as groups, then we can reconceptualise groups as individuals. Hence we can apply individual selection models to whatver level we need to, supposing that the degree of integration warrants it. And, I suspect, while trait groups warrant it (and this is happily subsumed under kin selection theory), higher level groups like species will not. But if the point is that we always need some group level, I think nobody would deny that apart from the really gene-centric types, whose views are now not in the ascendent.

Following Maynard Smith, W&W treat eusociality as a major transition also, so ant colonies occur when selection operates between colonies rather than within them. But, like the origins of individuality in eukaryotes, the fitness of the hive is rather like the fitness of the individual, and in both cases can be reduced to the fitness of the fertilised ovum of the individual (including the queen of the hive). Whether or not that is a form of group selection depends very much on what one counts as the fitness bearer, and this is the point here. W&W say this is a form of group selection. Hamiltonian approaches see it as a form of individual selection.

I wonder a lot at this point if the entire debate is a matter of what gets defined as a fitness bearer. Differing frames have been employed under individualist selection accounts for such within-soma processes as immunological selection (clonal selection theory) and cancer, where cancerous genomes are seen as outcompeting ordinary somatic genomes. These selection processes treat the somatic corpus as a population, and the fact that (usually) cancerous cell lines and antibodies don't pass through somatic generations is irrelevant to the fact that these are selection processes. Assume for the moment that these can be framed under individual selection, only the individuals in question are B cells and cancer lineages. Then why the move to claim that all selection is group selection? Yes, organisms that have poor antibody repertoires and aggressive cancer lineages are selected as groups. But at the level of individual organisms, they are selected as individuals. And this doesn't seem to me to be problematic for individual selection. It becomes a kind of gestalt switch - you can switch between seeing the group as an individual and vice versa, depending on the level of description.

So while a multi-level selection approach is opposed to the simple individual selectionism of the 1960s, I don't see that it has to be true that we must think of selection at any level as being "group" selection, and most certainly not the kind of group selection that Wynne Edwards and earlier thinkers held was in play. So I don't think that this, in itself, helps sociobiology, or hinders it.

Let's return for a minute to Darwin's selection of (moral) groups. One of the major differences between selection of moral tribes and selection of a genetic moral "instinct" is that there is a bottleneck in the latter case (the fertilised zygote) that allows filtering of the genetic trait, while in the case of tribes, moral institutions are not so sequestered, and hence are not selectable as such. When one tribe takes over another, there is something akin to mixing inheritance, or blending inheritance. Individuals have memories, and unless the moral tribe completely annihilates the immoral tribe (which wouldn't be all that moral), the vanquished will tend to incorporate their old value system and moral instincts into the new conglomerate. If, and this is something Darwin would not have distinguished at the time, moral instincts are biological rather than cultural, selection for cooperative behaviour is just individual selection. If it is culturally mediated, through memes or somesuch, then the selection on the memes is also a form of individual selection (on each token of the meme alleles). In neither case does it resolve to group selection.

So the appeal to eusociality as a form of superorganism is telling - it means that we have a kind of individual (the queen and her progeny) that gets selected for that doesn't match the prototype we hold of what counts as an individual. But that is a fact of description, not of there being some novel kind of selection. If groups and individuals are the same thing at some level in a selection process, we should spend our time trying to individuate the fitness bearers and leave the old debates alone.

In a comment about the differences between Popper's view of science and Kuhn's, Stephen Toulmin once noted (1970) that the debate between gradualists and catastrophists in the 19th century changed to the point where they agreed on everything substantive, and only used different words. I suspect a similar thing is true in this case.

Transitions and Human interactions

Maynard Smith's and Szathmáry's "major transitions" literature has been widely discussed (Griesemer 1999, Maynard Smith and Szathmáry 1995, Szathmáry and Maynard Smith 1997). It presumes that there are transitions of a sui generis kind between certain "grades" of evolution - the evolution of replicators, of multicellularity, of neural systems, and of course of human language and thought. W&W appeal to this as a justification for sociobiology. Selection on linguistic and cultural groups is thought to be an outcome of multi-level selection on this "novel" form of inheritance.

I have a problem with this. It seems to me that transmission of evolutionary novelties, no matter what they are, from pollen collecting to vision in the ultraviolet or whatever, are just as amenable to ordinary (that is, individual) selection accounts as that of language. The only reason why we think human symbolic communication is a "major transition" is because we privilege it, as human beings ourselves. In historical terms, this would be the Whig Interpretation; in biology it is yet another form of the Great Chain of Being. As Darwin himself said in his Notebooks:

It is absurd to talk of one animal being higher than another. We consider those, when the intellectual faculties (cerebral structure) most developed, as highest. A bee doubtless would when the instincts were . . .

This tendency to privilege human evolution is a kind of Whiggism. It leads people to argue that there is some abstract property - information - that is transmitted in a peculiar manner only for human culture. But we know now that culture is not restricted to human beings, nor language (in a very undeveloped form, relative to human language), nor tool using nor any of the old identifiers of human uniqueness. Even morality is now seen as more general than human society.

And here is where I will leave Wilson and Wilson and argue for my own view of sociobiology, or at least, social biology. All this argument about levels of selection seems to me to be quite besides the point. We apply what models of selection need to be applied when we have good reason for thinking that selection has occurred, and a multi-level approach is, I think, no longer controversial. But how do we know what the explicanda are? That does not rely on selection per se, but on something else: phylogeny.

Phylogenetic sociobiology

Attempts to ground sociobiology in selection models and arguments about levels of selection are futile. That's harsh, but I think it's true. I do not think that W&W have indeed offered a new theoretical framework for doing sociobiology; they have rehashed the existing framework and continued old debates. Now these debates are not, in themselves, irrelevant or uninteresting, but they miss the point, in my view.

Ethology and its successors relied on arguments by analogy. Do gazelles have a "stamping ground"? Then so must we. Do chickens have a "pecking order"? So must we, and so on. But Gazelles, and even more chickens, are phylogenetically a long way from humans, so how well does the analogy hold up? In more generic terms, must we, as a social species, exhibit the same dynamics that other social species do? I don't think so - evolution is particularlistic, and what holds for one need not hold for others. In fact, the phylogenetic distance between two species lessens the analogous weight of inferences between them.

Each species has what phylogenetic taxonomists refer to as "autapomorphies", unique derived traits that make that taxon what it is. We might argue, contrary to the sociobiological analogies from ants, bees, birds and beasts, that humans are in fact their own kind. That would be wrong, of course, because we are related to other organisms. But, to murder Tolstoy, all species of a taxon are alike, but each species is itself in its own way. So if we want to do sociobiology, and evolutionary principles suggest that we should, we have to attend to both the shared traits of a taxon and its unique traits.

And this is what researchers like Frans de Waal are doing. Instead of drawing questionable inferences from distal taxa, they use our nearest neighbours to both suggest the shared traits and highlight the differences between humans and our nearest relatives. Morality falls out of such studies. The question why apes and humans share some traits of fairness and yet differ in the expression of those traits will illuminate more about human sociobiology than all the analogies one can find. And at the same time, it will guard against anthropomorphising those other organisms. Symbolic thinking also falls out of those studies. Apes can solve problems very like ourselves. They exhibit displacement, some of them, the ability to refer to objects not present. They make peace when there is conflict, and make conflict when there are resources to be fought over. In sum, humans turn out to be a lot less exceptional than we traditionally have held.

The only way to nonsubjectively identify the unique and shared traits of the human species is to analyse homologous traits in our nearest relatives, chimps, bonobos, gorillas, other apes, and to a lesser degree primates in general. Add to this experiments designed to work out what pre-cultural skills human children have, like this study, and we can start to do real sociobiology rather than trying to shoehorn human behaviour into analogies from distantly related taxa.

I think of this as a way of setting up the problems that selectionist accounts might (and I stress might) explain, rather than beginning with selectionist explanations for poorly individuated and classified behavioural traits. And many traits will be not selectively favoured in humans. As a commentator noted in the first post of this series, what counts as the EvPsychology Environment of Evolutionary Adaptation or EEA may, in fact, go back to well before we even evolved as primates. The human skeleton, for example, contains homologies with most mammals, such as two hind limbs, that are not, in themselves human adaptations. Likewise, we might expect that many of our "innate" behaviours are pre-hominid, and even pre-primate. Aversions to stimuli that are possibly dangerous might be included here. Social behaviour? Well, who knows? Better that we identify what those behaviours are, and explain cultural behaviours in terms of inherited traits, than that we have to give a selective account of them first.

Take religion: I have argued before that religion comes out of our dominance hierarchical nature, something we do share with our ape cousins, but which we might expect has evolved in different ways (after all, if bonobos differ from apes, and they are only some three to four million years separated, then we who have evolved for six million years distinct from that lineage might be expected to evolve in our own way. If religion explained as dominance behaviour is correct, then the "natural" class here is dominance behaviour, not religion, and it is homologous to chimp and bonobo dominance hierarchies. Do we need to explain why there are dominance hierarchies in humans specially? No, because it is a synapomorphy, or shared derived trait, of all apes, and possibly all primates. What we need to explain is why it has evolved the way we (hopefully) can identify as unique among humans. And that, not the existence of social dominance, is the explicandum for human social behaviour.

This is something I have been worrying at for some time - it seems to me that "tree thinking", or more generally genealogical approaches, is better as a starting point for considering any evolution than adaptation - adaptive stories rely on identifying what it is that a species has as typical traits. In effect, I think that the sociobiology of the past has had it backwards. That is what I mean by "adaptationism" - the presumption that the first inferential step is to explain why a trait is adaptive, rather than first attempting to identify why a trait is there in terms of taxonomic inheritance.

Your mileage may vary.

References

Griesemer, James. 1999. Materials for the Study of Evolutionary Transition. Biology and Philosophy 14 (1):127-142.

Maynard Smith, John, and Eörs Szathmáry. 1995. The major transitions in evolution. Oxford, New York and Heidelberg: WH Freeman/Spektrum.

Szathmáry, Eörs, and John Maynard Smith. 1997. From Replicators to Reproducers: the First Major Transitions Leading to Life. Journal of Theoretical Biology 187 (4):555-571.

Toulmin, S. 1970. Does the distinction between normal and revolutionary science hold water? In Criticism and the Growth of Knowledge, edited by I. Lakatos and I. Musgrave. Cambridge UK: Cambridge University Press.