Improbable Destinies, Jonathan Losos
The question Losos asks, and tries to answer, is this: can we predict evolution? Are certain things inevitable in development — birds, humans, antibiotic resistance, etc, etc? He writes engagingly about field work, experiments, thought experiments, the various theories and people who have supported them… I definitely want to do more reading on this.
Am I convinced? Well, I’m not sure Losos is convinced that evolution can be predicted in detail; he presents some good evidence that suggests that you can predict the sorts of changes in gene function that will be beneficial in a certain environment, but that you can’t predict exactly how those changes will come about. Sometimes one gene might be altered, sometimes another. The phenotype is predictable (unsurprisingly: look for what would benefit the species in breeding successfully) but the genotype is not, unless it’s a fairly simple case of one particular molecular switch needing to be flipped on or off. There is a great deal of contingency in the process of evolution: Gould was (at least to some extent) correct in suggesting that we can’t “rewind the tape of life” and then see things proceed in exactly the same way.
As with determinism in any sense, I generally believe that if all factors were known, we would also know the result. I’m just not sure we can know those factors (and I dislike and squirm away from applying it to human ethics — our actions may be caused by previous events, but we don’t experience the process that way, so it’s irrelevant in how to be moral) — especially given events on a quantum level.
It seems to me that the mutations that occur are stochastic and environmental changes are not easily predictable, so you have a situation where the right mutations for a particular evolutionary path would have to occur at the appropriate time. For that to look macro-deterministic over the whole of planetary history seems unlikely to me. Particularly, the earlier you go the more likely it is that a random fluctuation in either the environmental pressures or the mutations that occur will have a radical impact on future evolutionary paths. It would be really interesting to get quantitative about it, maybe with some hybrid Bayesian/stochastic/game theory approach? Exciting stuff!
That’s the general evidence, yes, e.g. from the Cit+ E. coli lineage.
The reason I think of it as deterministic is that things usually have causes. But that doesn’t mean evolution would proceed in the same way if started again from scratch, if things weren’t controlled to precisely the same parameters as the first time, down to the movements of subatomic particles. If everything, truly everything was the same, I don’t see how there could be a different result. Whether we would ever be able to comprehend all the factors in retrospect or make accurate predictions for the future, I very very much doubt. And control the factors? Ha.
Even if you could specify the initial conditions exactly, quantum processes would make things veer away from the exact same path over time anyway – and DNA mutations are essentially quantum mechanical in origin, so I think even if you could build a huge series of “early Earths” with the same starting state, evolution would diverge over time in each.
Well yeah, I’m aware of that, that does fall under the heading of “truly everything” and “precisely the same parameters”.