About Rationally Speaking
Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.
Friday, June 29, 2012
Enjoying natural selection on multiple levels
Meanwhile, Richard Dawkins was picking another fight.
Normally, this would not be an occasion worthy of comment. The best way to distinguish between Professor Dawkins’ waking and sleeping states is probably on the basis of how contentious he is at a given time. Nevertheless, I’m compelled to say something for two reasons. First, this particular fight happens to be taking place right in my proverbial (and professional) wheelhouse; second, I’ve just finished my annual re-reading of Michael Crichton’s Jurassic Park duology.
That last bit requires some explanation, I know. As I mentioned in my last post, Crichton spent most of his later career playing the role of anti-establishment gadfly. For The Lost World, his sequel to Jurassic Park, he set his sights against the theory of natural selection. Indeed, the centerpiece of the book—almost literally, coming precisely halfway through the page count—is a chapter entitled “Problems of Evolution,” wherein Crichton asked the following about the evolution of human intelligence:
“... where does natural selection act? Does it act on the body … on the developmental sequence … on social behavior … Or does it act everywhere all at once—on bodies, on development, and on social behavior?”
This is an issue known to philosophers of biology as the “levels-of-selection problem” (I’ll abbreviate it as LOS hereafter). Biologists don’t have a clear answer to Crichton’s question, and so he took it to be the case that the theory of natural selection is deeply flawed. But Crichton missed an important point: that LOS is a philosophical rather than biological problem.
Professor Dawkins’ newest fight is about LOS; as it happens, so too is my PhD thesis.* I’m therefore going to take this opportunity to summarize the debate and to show that we do have decent answers to Crichton’s question—so long as we ask those questions in the right context.
Let’s start with a big question. What exactly does the theory of natural selection say, and why is LOS a problem?
There’s a reason that Thomas Henry Huxley (AKA “Darwin’s Bulldog”) responded to the publication of The Origin of Species with the exclamation, “How stupid not to have thought of that!” The reason is that the theory is, very broadly speaking, incredibly simple. It says that individuals that are better equipped for survival in their environments will leave more offspring in a population than worse-equipped individuals, and so each subsequent generation will see a greater proportion of beneficial traits in the population—and this spread of traits is evolution.
More specifically, evolution requires three things: first, that there is variability between the individuals in a population; second, that the variations can be inherited by individuals in the population’s next generation; third, that there is a consistent reason that individuals with one kind of variation leave more offspring than individuals with a different kind of variation. We can call these the requirements of variability, heritability, and differential reproduction. We can imagine these requirements combining into a sort of informal formula for evolution:
Variability(x) &; Heritability(x) &; Differential Reproduction(x) ➝ Evolution(y)
In this formula, x is a variable that stands for the individual and y is a variable that stands for the population. Bottom line: if a variable trait benefits most individuals x that have that variation, then you’ll eventually (over multiple generations) see that variation spread to most of a later population y.
Benefit therefore plays an important role in evolution. But cui bono?
To hear Darwin tell it, the correct answer to Crichton’s question is “the body” (as Crichton put it), since whole organisms benefit from their advantageous variations. To wit: tigers have stripes because any individual tiger with stripes—as opposed to one with, say, spots—will (all else being equal) blend in better with its grassy surroundings, thus giving it a better chance of ambushing its prey and surviving long enough to leave lots of tiger babies. This works as an explanation of most organisms’ traits, but there was one trait that always defied Darwin’s explanation: altruistic behavior.
By definition, altruistic behavior—that is, being helpful—benefits organisms other than the one that behaves altruistically. I spend a lot of time helping to care for my infant nephew; that’s time that I could be spending, say, going out on dates that might eventually secure me a child of my own. Helping to care for my nephew benefits him, but it hurts me (evolutionarily speaking). The impulse to care for nieces and nephews should therefore be rare among humans. It’s not. Why not? Cui bono?
If you know of Dawkins at all, then the odds are slightly better than even that you know something about his response to the altruism problem. The so-called “selfish gene” theory, technically known as gene selection, is an elaboration of work done by W.D. Hamilton and G.C. Williams on a phenomenon known as “kin selection.” Kin selection is predicated on the idea that the impulse I feel to care for my nephew is stronger than the impulse I feel to care for (say) my neighbor’s nephew. I know that my nephew is my sister’s son, and that my sister and I were born of the same parents; I therefore know that he carries 50% of my sister’s genetic alleles, and that there’s a 50% chance that any one of my sister’s alleles is one that I also carry. For any one of my nephew’s alleles, then, there’s a 25% chance that I also carry that allele. If I care for my nephew, then my genes have a one in four chance of helping themselves; if I care for my neighbor’s nephew, the odds are much, much lower. Gene selectionists therefore argue that genes are the individuals who benefit in the process of natural selection. Hence Dawkins’ famous claim that organisms are “gigantic lumbering robots” for carrying genes around: I have an impulse to care for my nephew because it helps (some of) my genes, even though it hurts me as a whole.
In 2010, E.O. Wilson and two collaborators wrote an article in Nature attacking the viability of kin selection. We won’t get into the details of their mathematical argument; the bottom line is that things rarely work out so neatly as “my nephew has half of my sister’s genetic alleles and she has half of mine,” and the complexities ultimately call into question the idea that gene selection can explain altruistic behavior. In his newest book and a recent New York Times “Stone” column (interestingly, a philosophy blog!), Wilson proposes an alternative that he calls “multi-level selection.” His account is so called because Wilson believes that nature sometimes selects genes, sometimes selects organisms, and sometimes selects groups—and that the latter option is the one that explains altruism. It was this claim that prompted Dawkins’ scathing review of Wilson’s book, linked in the first paragraph. Undermining the very foundation of Dawkins’ account of selection probably had something to do with it, too.
Group selection says that I may not benefit from caring for my nephew, but my family does, and if my family prospers then so too will I. Altruistic behavior evolves in organisms that develop strong social bonds because benefit ultimately comes back to the altruistic organism through those social bonds. Groups maintained by social ties will therefore be “individuals” that nature can select.
To summarize: LOS is a debate over what gets plugged into x in our evolution formula. Darwin said x=organisms; Dawkins says x=genes; Wilson says x=social groups. (And Crichton says that the spread of possible answers makes the formula wrong, which is about as valid an argument as saying that the equation “2x=y” is wrong because there are just so many numbers that can fit into each variable.)
Between Dawkins and Wilson, each side of the LOS debate gets things right and each side of the debate gets things wrong. We’ll see how by asking another big question: why is there a theory of natural selection?
Darwin once wrote that “all observation must be for or against some view if it is to be of any service.” Recognizing what natural selection stands against should therefore be useful towards recognizing what natural selection stands for. Remember creationism? How about intelligent design? Natural selection provides a (vastly superior) alternative to these theories in explaining why organisms seem so well-adapted to their ways of living; it is an answer to what we might call the design question.
To date, gene selection has provided an astoundingly successful answer to the design question. In addition to explaining altruism, it also explains things like why tigers have stripes: after all, traits like stripes are coded (at least in part) by genes, and so selection of tigers with stripes is also selection of the genes for stripes in tigers. Gene selectionists like Dawkins argue that their account works better than organism-level selection because it explains more, and this seems very largely correct.
Group selection, by contrast, doesn’t do as well at answering the design question, because plugging “social groups” into x raises a number of problems. Let’s assume that my family does benefit from my altruism. Cui bono? Maybe my nephew learns the value of altruism and behaves similarly towards his nephew. My nephew therefore has the altruistic impulse; what sense does it make to say that my family has it? The family is altruistic only if it has individual organisms that are altruistic. Doesn’t that mean that it’s really the organisms that benefit? Groups aren’t coherent individuals in the same sense as organisms or genes, and so it’s much more difficult to say that they have beneficial traits. Without beneficial traits, there’s no evolution formula. Again, Dawkins is very largely correct on this count.
Still, it does seem that group selection is at least theoretically coherent. It may work; what does it work for?
As it turns out, the design question is not the only one that the theory of natural selection is meant to answer. Organisms are well-adapted to their environments, yes, but there’s also the fact that organisms in different species are well-adapted to a very wide variety of different environments. Why? This is what we’ll call the diversity question.
As it turns out, group selection may provide a much more direct answer to the diversity question than does gene selection. Understanding why requires that we understand extinction.
Diversity requires extinction. Tigers are well-adapted to jungle hunting and lions are well-adapted to savannah hunting, but why aren’t there big cats that hunt in both the jungle and the savannah? Natural selection predicts that this hypothetical intermediary group once existed, but went extinct; since the intermediary no longer links the tiger and lion groups, those groups are now distinct species.
Extinction is not random. Certain traits—amount of genetic variability, population size, etc.—have a very strong influence on which species go extinct. But take notice: these are traits of groups, not organisms or genes. In fact, gene selection has a difficult time explaining diversity: it can explain why there are feline jungle hunters and feline savannah hunters, but not so much why there aren’t any remaining feline part-timers that hunt in both. On this point, Wilson is correct: a multilevel account of selection can draw on gene selection to answer the design question and group selection to answer the diversity question, and so benefits from the strengths of both accounts.
This has been an admittedly broad sketch of an overview of a deeply nuanced topic, and I’m sure that more than a few readers inclined towards evolutionary biology are now engaged in the harmful act of banging their heads against a wall. (Some of them may be on my dissertation committee.) Sorry: it’s tough to capture nuance in a blog post.
Nevertheless, I hope all readers can take this much away: the question of where natural selection works is one that depends on a philosophical question. What is an individual in biology? We’re used to thinking in terms of organisms—if we don’t count humans as individuals, then what would we count?—but we have compelling reasons to think of entities at other levels of organization as individuals, too. Are my genetic alleles individuals with “interests” distinct from mine? Why should I be able to call the group Homo sapiens a biological individual, but deny that individuality to the group “Finkelman family”? It’s difficult to imagine what sort of empirical data could decide these questions, and this is where the philosopher makes his living.
So: fight on, Professor Dawkins. Fight on, Professor Wilson. It’s a hell of a view from here in the wheelhouse.
* If I make fewer attempts at humor here than I normally do, it’s only because this is obviously the most important problem that anyone can possibly work on, under any circumstances, and it must be treated with all due severity.