This is the promised response to Jim Manzi’s response to Jerry Coyne. It does not rebut any claims that he made in an article he referenced in The American Scene, since I have not had time to read the article. This rebuttal is solely dependent on his writing on The Daily Dish.
First, Manzi expounds his primary contention in the opening:
Accepting evolution, therefore, requires neither the denial of a Creator nor the loss of the idea of ultimate purpose. It resolves neither issue for us one way or the other.
Of course, as near as I can tell, Coyne never argued that evolution requires the denial of a creator in toto. He has specifically admitted that it is compatible with a deistic notion of some impersonal creator. On this, I agree. But I also agree with Coyne that evolution requires to abandon the idea of divine purpose, and I don’t think Manzi has refuted that. I also find Manzi’s justifications for a creator to be weak and unconvincing.
First, there is his version of the first cause argument:
…being able to describe a set of physical rules that explain scientifically how particles can interact to create so much of nature (what I called a scientific result of “stupendous beauty and power”) does not address the problem of ultimate origins – i.e., where do these rules and the most fundamental particles come from? Note that I didn’t assert that one must accept even as general an idea as Uncaused Cause, never mind God, merely that he must either accept there was a first cause, or just live with the problem of infinite regress.
Manzi’s unstated premise here is that some causal chain is always necessary to explain something. But we already have physical theories that entail the reality of acausal events. Namely, quantum mechanics gives us a picture of physical systems that are not deterministic; particles are described with probability amplitudes. It is also worth noting that pure quantum systems evolve in a time-reversible fashion, unlike our everyday macroscopic experience where thermodynamics (specifically, the second law) provides us with the “arrow of time”.
Given this, we don’t have to live with the problem of infinite regress. There always exists the possibility that there exist physical systems that are, unlike in our common experience, non-deterministic and/or time-reversible. This concept can be made more intuitive when the modern cosmologist suggests to us that asking what came “before” the big bang is like asking what is north or the north-pole, since what we perceive as time is a contingent property.
Next, Manzi makes the following curious claim:
But of course, implicit in Coyne’s response is the point that the problem of first cause existed before Charles Darwin was born – many centuries before, in fact, as per my post. Once again, my point in this regard was not that the only resolution to this philosophical question is that one must believe in God, only that evolution through natural selection does not resolve it.
But Darwinian natural selection doesn’t have to resolve all problems posed by the universe which could be taken to necessitate a divine explanation. As far as I can tell, Coyne claimed that it only resolved one: whether biology gives evidence of such a divine plan. There exists the logical possibility, of course, that there is a divine plan and biology is not part of it. But that doesn’t seem to be under consideration.
But moving beyond that, we get to the real meat of Manzi’s argument, which is probably the most unconvincing part and what I had found objectionable before. He disputes the claim that evolution precludes a divine plan by appealing to evolutionary computation:
This is precisely the argument that I was trying to address in the post – the argument that the structure of the evolutionary process is inherently undirected. Coyne simply repeats it without reference to my laborious (in Coyne’s words, “tedious”) attempt to demonstrate that it is invalid. Every one of the issues that Coyne raises in this paragraph – roughly speaking, competition between genes, a complex and changing fitness landscape, and parallel development of alternative lineages – was addressed directly in my post. Coyne asks “What kind of goal-driven process is that?” As per my post, I can produce a goal-driven process that has these characteristics with a $2,000 laptop and a Java compiler.
But there is a slight technical detail of these algorithms that Manzi is either overlooking or neglecting to mention: evaluation functions. When we use an evolutionary algorithm to solve a problem like the Hamiltonian cycle problem, we design an evaluation function that makes populations of solutions more likely to reproduce if they travel through either more nodes the most nodes the fewest times (you could design a much less naive heuristic than this, but it suffices for an example). Natural evolution is nothing like this: the “goals” are distinctly local, and the local conditions are contingent. Given this, the analogy fails.
All in all, Jim Manzi has failed to provide an effective rebuttal to Coyne’s original claim. It’s really a textbook example of why analogies are generally considered a weak form of argumentation.