Tuesday, March 03, 2015

What is wrong with standard evolutionary synthesis

Preamble:
With this explanation, I will use terms that may normally be associated with informatics. This is strictly in use for shorthand in the same way that "running the program" is expressing the genes that are associated with DNA that may or may not have mutated. It should be in no way taken to mean that it is scientifically accepted to use the arguments made of a Turing machine or computer science to come to conclusions in regards to Dna mechanisms. 
Also, I am not big on citations. I may mention people like Popper, Wickramasinge, Lennox, etc. but not based on their authority on any subject or another. My thoughts are to be considered based on first principles, and perhaps more abstract like in a mathematical argument rather than a debate scored by a selected audience.

Looking at the synthesis in historical terms, it has evolved from Darwinism, where the main thrust is the origin of species. Life on earth, when exposed to a different environment will adapt to that environment over generations. This was documented and well observed in the Galapagos. Extrapolating from the changes in observed Galapagos species from mainland ones, to the length of time of the age of the earth, it is quite logical to extend that process to all known species. Naively, the process was seen to extrapolate back past the Last common ancestor, and Darwinism included abiogenesis in a kind of continuum from no life to the variety we see today. Now, in maths as in science, interpolation is always a safer bet than extrapolation, thus two species which are closely related in time, genetics and space can confidently have intermediates placed in there, and have a rough idea of smaller changes over time leading from one to another. Extrapolation beyond the last common ancestor is a completely different kettle of fish, because we have no evidence to go on at all, but at any rate, in a philosophical sense, belief in evolution is highly correlated to a belief in abiogenesis.
This leads on to the hypothesis of "how" adaptation happens. In the case where a beneficial mutation reflects a particular environmental "selection", I will use the shortcut that the environment "programs" that particular mutation. As said in the first paragraph, this is just a shortcut, in no way am I ascribing informatics to the process. In this sense it just means that selective pressures are colluding to give differential advantage to individuals and/or groups and/or populations that end up with that mutation. For the individual that first gets this mutation, this is like winning at the roulette wheel. Blind watchmaker analogies and the use of the word random imply that there need not, and probably is not any direct or indirect, partial or even trace causal interaction between the environment and the beneficial mutation. It is like the roulette wheels are assumed to be perfectly balanced, the programming is strictly done, and only done by selective pressures, which involves more successful reproduction and survival with than without.
Clearly the "goal" of a beneficial mutation is one which better enables survival in a (perhaps changed) environment, and of course the environment will have the final say, but Darwinially speaking, a more (even trivially more) efficient programming technique than that envisaged, by a "blind watchmaker" will *always* win over the completely causally unconnected technique. Thus just as "fitter" organisms will win over the less fit even at the cost of extinctions of some of the less fit species, so too fitter "programming systems" will win over "blind watchmaker" programming systems.

This is perhaps like saying that roulette wheels are trivially non-level, but professional gamblers that measure the level in secret, and know exactly how to win will consistently end up in front while everybody else will lose to the house.

The "environment" is replete with symbiotic organisms that are both part of the environment, react to environmental cues, cause trivially non-random mutations in other organisms (eg through causing stress, horizontal gene transfer, activity of viruses, transport of mutagens, etc.) and thus can allow organisms to "beat the house" of the blind watchmaker. Species that do not or cannot have these "programmer helpers" will have much less chance to thrive under new environmental stresses. 

This is the crux of my argument that natural selection on random mutations *cannot* explain adaptation that is as efficient as observed in, for example resistance to pesticides, because more efficient adaptations win against naive randomness and brute force selection any day.

4 comments:

Chris Fellows said...

I have been slow to comment since I seem to essentially agree with everything you are saying except your conclusion, and need to re-read it again more carefully.

Marco Parigi said...

The argument rests on the statistical significance of beneficial mutations, and the burden of proof. In situations where a player is winning at Roulette in a statistically significant way, the burden of proof can shift back to the player to prove that he is making selections based on random turn actions of the roulette wheel. Freakonomics touched on it with teachers cheating on their students behalf. The "simplest" explanation can be that it is not random, even if the mechanism is not known, or perhaps even unknowable. What may appear to be trivially non-random may be a system which hides amongst complexity.

Chris Fellows said...

Your post has handily recrystallised some thoughts for me.

Clearly the "goal" of a beneficial mutation is one which better enables survival in a (perhaps changed) environment, and of course the environment will have the final say, but Darwinially speaking, a more (even trivially more) efficient programming technique than that envisaged, by a "blind watchmaker" will *always* win over the completely causally unconnected technique. Thus just as "fitter" organisms will win over the less fit even at the cost of extinctions of some of the less fit species, so too fitter "programming systems" will win over "blind watchmaker" programming systems.

I agree completely: if there is a better way, organisms would have found it. I think they have, but it is too obvious for us to see it. It is important to remember that ‘beneficial mutations’ are a side product of the main game, which is reproductive success. So a more efficient ‘programming’ technique cannot interfere with the short-term goal of accurately reproducing what already exists. I think, if we look at how organisms actually behave, sexual reproduction is obviously a wildly successful effort to improve upon mere ‘natural selection on random variations’ as a programming technique.

The environment is changeable, so an organism cannot know in advance what mutations might be required. What it has to do to maximise its reproductive advantage is to produce a large variety in offspring that are at the same time meticulously error-checked to try to ensure that they are at least as viable as their parents. Sexual reproduction provides, first of all, a way of mixing and matching known solutions to the problem to give unique offspring. Second, it gives the option of ‘pushing the envelope’ by subjecting gametes to mutational stresses, knowing that the ones that would give non-viable offspring are likely going to be unviable themselves. Third, it provides a way for the species to get more information about the environment than is available to the ‘blind watchmaker’; the ‘sighted matchmaker’ can evaluate potential mates based on environmental cues, and this is how, at least in higher organisms, your ‘fitter programming technique’ manifests itself. I am going to use the terminology ‘higher’ here because I think in this case it is actually justified.

The "environment" is replete with symbiotic organisms that are both part of the environment, react to environmental cues, cause trivially non-random mutations in other organisms (eg through causing stress, horizontal gene transfer, activity of viruses, transport of mutagens, etc.) and thus can allow organisms to "beat the house" of the blind watchmaker. Species that do not or cannot have these "programmer helpers" will have much less chance to thrive under new environmental stresses.

I think if you look at where these things have an impact, you will find they are mostly in the ‘lower’ organisms. By themselves, these ‘programmer helpers’ are solutions which are cruder than sexual reproduction, and for the finely balanced systems required for sexual reproduction to run smoothly, an organisms will seek to exclude them as much as possible. I agree that more efficient adaptations win against naive randomness and brute force selection, and suggest that sex is the way to achieve these more efficient adaptations. The fact that so much of evolutionary change is driven by the mysterious process of 'sexual selection' suggests that where organisms are putting their effort, that is where the key to how they might be beating the house at roulette.

Marco Parigi said...

Fair enough. Perhaps the insistence in the synthesis is virtually all about crowding out Intelligent Design. I feel that science is all the poorer because biologists are not having conversations like these; perhaps even making bold new hypotheses, ignoring that some of them may be stalking horse for ID.
I think that sex is great, and your ideas in that regard are pretty similar to the thoughts rattling in my head. I don't think it rules out other symbiotic things happening in parallel, though I do see your point. Remember that the human biogenome, with all the organisms we rely on, have more cells and more dna than our genome. I cannot see them not being part of the overall adaptation scheme, perhaps only at at least a level lower than Mendelian traits.