Thinking about how to define and measure fitness is a challenge in evolutionary biology – most definitions tend to cause “survival of the fittest” to collapse back into a truism; if the fittest are those that survive, then survival of the fittest is true merely by definition rather than being a result of empirical science.
That’s an interesting issue I’d like to explore more, but not one I have anything to say about now. Instead, let’s look at a sentence found in discussion of the roles of constraint and selection in evolution: “Selection takes the … offerings available by natural variation, and leaves the few variants that are more fit, perhaps zebras that are larger and therefore better able to fend off lions, despite being less maneuverable.” (Philosophy of Biology, Edited by Rosenberg & McShea, pp. 74-75)
It occurs to me as I read this that this manner of talking about selection is probably not justified by our best understanding of the process. What selection actually offers in the zebra case is the existence (survival) of the progeny of a particular zebra that managed to survive in the past through fending off a lion (and achieving numerous other evasions of death and the failure to reproduce). Selection doesn’t ‘care’ about (feature these things causally) future propensity or ability to fend off lions, or past propensity or predicted ability per se, but only what actually happened contingently in the past. If a particularly weedy zebra managed to escape a pack of lions and go on to reproduce while objectively stronger members of his herd were either taken down or failed to reproduce for some other reason, then the weedy genes will be continued, and that surely is “selection” (whether there is an objective distinction between this and “drift” is an interesting question; presumably there must be…). Of course, selection will tend overall to favour the strong over the weak in situations involving zebras vs lions, but the strength of the strong is arguably an arbitrary classification on our part which is not directly relevant when accounting for survival; it may intuitively strike us as the most relevant feature of the situation, but that is no guarantee of truth.
To rephrase my concern here, it seems that it is tempting to see the survival of a strong zebra as an archetypal instance of “selection” and hence to see “strength” as a trait selected for, whereas as I understand it, whichever zebra (or nematode worm) ends up surviving in whatever circumstance is equally an instance of selection. Part of the worry is that it is difficult to single out factors which were causally relevant in any organism’s (and hence set of genetic information’s) survival. The fact that the zebra had legs, for instance, was presumably quite useful for it. The other zebras had legs too, so that might make legs unlikely to be the favoured candidate for a selected trait in this instance; but perhaps the zebra in question was just in the right place at the right time and also happened to be strong enough to fight off a particular lion at a particular instant; whether we could go on to make the inference that “the ability to fight off lions” has been selected for is rather unclear. In the absence of controlled and replicated trials, we only have our best guess.
Selection is a short-hand for historical environmental processes which have led to the non-continuance of some lineages but not others.
I’m still working through all of this, so my views are likely to change.