We’ve all laughed at the naive “turtles all the way down” story. What follows is all Eric Anderson’s “randomness all the way down” story:
[WD writes:] I’m saying the non-random survival . . .
It is pretty much randomness all the way down. How did the particular particle interact with the copying mechanism to cause a mutation? How did that particular mutation end up interacting in the organism to produce an effect? What result did that have in that particular organism, as opposed to another? How did that particular mutation get spread in the population? What environmental factor happened to come along after the mutation that resulted in it making a difference? Which organism happened to be on a high rock when the flood came, or under protection when the hail fell, or hidden from sight when the predator arrived? And on and on. Everything that goes on within a lineage to get an organism to where it is today; everything that went on in the predator’s lineage to get them where they are today; all the vagaries and hazards of nature. It is essentially randomness all the way down.
Natural selection is not any kind of force. It is simply an after the fact label attached to the results of processes that are seldom understood, rarely identified, and that (as a practical matter) are essentially random. Natural selection doesn’t impart any “non-random” directionality to evolution. It is simply a label attached to the outcome, and attaching a label to the result of what is essentially a random process does not make the process non-random.
[WD writes:] “It’s not simply that survivors survive, but those individuals that are best adapted to their environment survive, and so over time lineages become better adapted.”
And how, pray tell, do we know that a particular organism was “best adapted to its environment”? Because it survived.
Look, if someone wants to use the two words “natural selection” as a shorthand way to avoid having to say: “Organisms are more likely to survive if they happen to be in a lineage that happens to have conferred a (generally unidentified) mutation that (in some typically unknown way) happened to provide a characteristic that happened to be helpful in the particular environment in which they happened to be living at the time, as compared to other organisms that were less lucky.” — If someone wants to use “natural selection” as ashorthand expression so that they don’t have to say all that, then fine.
The problem arises, as it does so frequently, when natural selection is put forward as an explanation for an organism’s survival. In that case it almost always falls back on survivability as the (often unspoken) definition. In that case it is a useless tautology. Worse, it gives people the false impression that some kind of “scientific” explanation has been proffered, when it is really just a confession of ignorance about the real underlying processes.
Think of it this way:
If we can identify, with particularity, what actually caused an organism to survive — the specific trait, particular molecular machines, identifiable proteins and DNA sequences, the particular environmental factors, predation, weather, flood, drought, and so on — if we can identify precisely what caused the “differential survival” in the population, then we can talk about the real, physical, concrete, underlying, specific cause just fine, thank you very much, without ever invoking the label of “natural selection”.
It is only when we don’t know what the actual forces and causes were at work that “natural selection” need be invoked. Unfortunately, in that case, it functions as little more than an observation that those that survived, survived.