Pat Gunn (dachte) wrote,
Pat Gunn

Thoughts on Evolution and Provability

Semi-recently, I came across a pointer to this essay, which attempts to tackle the counterfactuals that would disprove evolution, to show that it is a theory in the scientific sense rather than post-hoc. As much a fan I am of people thinking about this, I think they got some of the details wrong. Note that I am not a evolutionary scientist, so as usual, I am not perfectly representing academic consensus; take my reasoning with several large grains of salt.

  • Fossils in the wrong place would at least undermine a lot of gathered evidence, although other evidence for evolution is good enough that it might instead cause us to doubt the process of fossilisation. Or it'd introduce some tension between the two conclusions and inspire us to do more research.
  • Adaptations in one species good only for a second species - I think this is possibly off; evolution is a messy process, and provided the adaptation were not grossly harmful, the described "disproof" might happen just as much as river currents sometimes flow backwards for a bit. Selective pressures make the most sense being viewed over long periods of time, and they only win out in the long run in a statistical way.
  • I agree on the "lack of genetic variation" point, mostly. It'd at least remove a lot of the power of evolutionary mechanisms if mutations had to occur in a static population for any change
  • Adaptations that could not have evolved stepwise - I think this is a bit off, in that the mechanisms of gene replication, and the kinds of errors possible, mean that it is possible for large changes in genotype to happen in one mutation event, and at least theoretically that could result in new features appearing en toto (provided the embryonic environment remains suitable). It's all essentially just bits of information, and any mutation is possible at a given probability. Note that I don't know enough about how probable larger mutations are to be confident that this is a good argument.
  • Two observations that seem to be a snide stab at group selection. Both of these fail to convince; memes and multi-level selection seem pretty logistically solid to me
  • Disjoint morphological/fossil evidence and genetic evidence on the family tree of organisms - This would indeed be a deeply serious problem for evolution; the closest we see is convergent evolution, but the factors behind weaker form of "ontogeny recapitulates phylogeny", specifically the mutation-fragility of the early stages of the developmental process making mutations in genes expressed later in development of the organism less likely to result in unviability, presumably keep convergence from extending so far as to make only-distantly-related organisms frequently identical.

  • Testing functions in Perl

    (Nothing particularly profound or my-idea-centric here, and I was tempted to post it to my personal blog instead, but it's worth trying to learn…

  • Abstract strategies for abstraction

    There are a few purposes of abstraction in programming; one of them is to construct a uniform API that is independent of the backend that can work…

  • Statistical Software Components

    A few months ago I mentioned my big library of useful generic C/Perl functions (libpgunn). There are plenty of other general-purpose libraries out…

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded