Tuesday, May 04, 2010


I suspect that not even evolutionary biologists pay enough attention to the fact that in any evolutionary process, selection is a fundamentally negative process. Natural selection in evolution does not select for, it selects against.

The usual narrative of evolution is that some heritable variation (such as a mutation) gives an individual a reproductive advantage. The individual's descendants inherit that advantageous variation, and because it is indeed advantageous, the variation comes to dominate the gene pool of the population. Not a bad story, and it gets the gist across, but it's missing an important component of the story: what precisely do we mean by "reproductive advantage"? The straightforward intuitive meaning isn't bad, but it misses important subtleties, especially when we want to intelligently direct evolutionary processes, such as social evolution.

A more accurate and direct narrative of evolution is that there is no selection-for, there is only selection-against. Some individuals have — for various reasons — a reproductive disadvantage; they fail to create descendants to inherit their individual variations. A specific variation can confer a reproductive advantage only indirectly, by changing the environment so that lacking the specific variation directly entails a disadvantage. A variation dominates the population's gene pool only when its competitors are selected against.

The selection-against narrative makes the creationist objection that, "If human beings are descended from monkeys, why are there still monkeys?" more obviously fallacious, A naive or superficial grasp of the selection-for narrative makes this objection plausible, but the selection-against narrative makes the answer more obvious: because nature did not select against monkeys nor did nature select against humans; nature selected against that which we do not see.

Another somewhat less naive example is Jerry Fodor and Massimo Piattelli-Palmarini's book, What Darwin Got Wrong (reviewed by Ned Block and Philip Kitcher as well as PZ Myers). I don't want to excuse the authors — they're professional intellectuals, and they shouldn't allow themselves to be misled by superficialities — but it seems clear to me that a narrative of negative selection would make their thesis more obviously incorrect.

Fodor and Piattelli-Palmarini argue that
Darwinists say that evolution is explained by the selection of phenotypic traits by environmental filters. But the effects of endogenous structure can wreak havoc with this theory. Consider the following case: traits t1 and t2 are endogenously linked in such a way that if a creature has one, it has both. Now the core of natural selection is the claim that phenotypic traits are selected for their adaptivity, that is, for their effect on fitness. But it is perfectly possible that one of two linked traits is adaptive but the other isn't; having one of them affects fitness but having the other one doesn't. So one is selected for and the other "free-rides" on it.
The follow this critique up with a deeper philosophical critique. As Block and Kitcher describe
... Fodor and Piattelli-Palmarini think that problems about selection-for are omnipresent... [b]ecause they envisage a vast space of properties and expect proponents of natural selection to discriminate among all the rivals. Not only is there a property of being-a-melanic-moth, there is also a property of being-a-melanic-moth-and-smaller-than-Manhattan. These properties are not only correlated in the world’s actual moth populations, they are correlated universally. Maybe it is impossible, even with the most rarefied genomic technology, to build a moth bigger than Manhattan. If so, the correlation between these properties could not be broken. How then could there be a sense in which one of the properties—being-a-melanic-moth—rather than the other—being-a-melanic-moth-and-smaller-than-Manhattan—caused the increased reproductive success?
Block and Kitcher would have us simply shrug off this object (nobody really cares whether nature is selecting for being-a-melanic-moth or being-a-melanic-moth-and-smaller-than-Manhattan), but even a definite error can be an opportunity to learn.

Block and Kitcher accept the narrative of selection-for, but they seem more flexible about its interpretation.
Natural selection, soberly presented, is about differential success in leaving descendants. If a variant trait (say, a long neck or reduced forelimbs) causes its bearer to have a greater number of offspring, and if the variant is heritable, then the proportion of organisms with the variant trait will increase in subsequent generations. To say that there is “selection for” a trait is thus to make a causal claim: having the trait causes greater reproductive success.
But a narrative of selection-against makes Fodor and Piattelli-Palmarini argument more obviously invalid. We don't have to distinguish between selecting for being-a-melanic-moth and being-a-melanic-moth-and-smaller-than-Manhattan, because nature is not selecting for anything. Rather, nature in this case is selecting against getting eaten by a bird (before leaving descendants). It happens to be the case that being-a-melanic-moth (or perhaps being-a-melanic-moth-and-smaller-than-Manhattan) exempted individuals from this selection-against.

Fodor and Piattelli-Palmarini philosophical critique just doesn't apply to a selection-against narrative. A selection-for narrative makes what is being selected for ambiguous; it's easy to see, however, in a selection-against narrative that nature is selecting against being-eaten-by-a-bird and being-eaten-by-a-bird-and-smaller-than-Manhattan (and being-eaten-by-a-bird-and- anything else). There's no ambiguity we have to shrug off.

If you want to understand why different outcomes occur in different evolutionary scenarios, look for differences in the negative selection pressures.

Consider three modes of negative selection in a population with a fixed food source. There's never enough food for all the individuals in the population. Some individuals will starve (before they reproduce): nature will select against those individuals and their heritable characteristics. Under ordinary circumstances, without any relevant variation, this selection is random: some individuals are just "unlucky" and don't happen to find enough food. But random selection is still selection, and over time it will from time to time happen to be the case by pure chance that all or most of the individuals with some variation will happen to starve by accident. Thus the competing variation(s) will dominate the population even though there is no causal reason why the dominant variation is "better than" the eliminated variation. Selection-for creates a mystery; selection-against provides an explanation: sometimes shit just happens; given enough time, shit will just happen.

In the second case, consider a variation that appears to provide a "reproductive advantage": individuals with this variation are faster, smarter, stronger or whatnot, and more likely to find food... specifically food that, but for their "advantage", other individuals would have eaten. This variation will (probably) dominate the population, but it will do so not because it's advantageous, but because the competing variations have become disadvantageous: individuals without the variation are disproportionally selected against.

Consider the third case: a variation that gives individuals access — even a little — to an alternative food source. In this case, the individuals have a reproductive advantage — if they fail by chance to eat the primary food source, they might not starve and be selected against — but this advantage does not create an immediate corresponding disadvantage in the rest of the population: all the food that was available to individuals without the variation is still available. The outcome in this situation is unclear precisely because we don't know what negative selection-against will actually operate: we have to look more deeply into the situation to identify how nature will select against individuals.

The selection-for narrative does not really distinguish between these situations. Why does some particular trait dominate the population? Because it was selected for. Why was it selected for? Using a selected-for narrative encourages the adaptationist fallacy: we don't know that just because a trait dominates the population it therefore confers a reproductive advantage. We can simply invert the question: the trait was selected for because it wasn't selected against, which encourages the question: why wasn't it selected against. But that's still not quite the correct question. We want to ask: what was selected against, and why? The selection-for narrative leads us down a winding mental path with many pitfalls; the selection-against narrative leads us by the nose to the correct question.

Clarity and precision are important in their own right, but there's a more compelling reason to stress the negative nature of selection in evolutionary processes. Human societies also evolve — dialectical/historical materialism is evolution — and although the low-level mechanisms obviously differ greatly (people do not transmit ideas and beliefs to their children through their DNA) the abstract mechanisms of heritable variation and negative natural selection still apply. If we are to "intelligently" affect our social evolution, we must understand how evolution actually works, and the negative character of selection becomes critically important.

UTA: I don't want to minimize the importance and subtlety of forces other than selection: e.g. mechanisms of heritability and variation as well as accidents large and small. My point is that to understand how selection operates in an evolutionary system, you have to look at what is selected against, not for.