num1cubfn wrote:Hey, Cal or Susu or anyone else who is awesome, can you answer something for me?
I´ve read this a while ago and didn´t have the time for the response it takes. But sure, I´ll answer it.
num1cubfn wrote:Why do people say random mutation? We now that there are many things that can directly cause mutation, which suggests rather strongly that in fact, there is nothing random about it, it may be the result of an extremely complex set of variables, and it may even be a set of variables about which we haven't discovered the effects of, but random seems completely
wrong when used to describe what causes mutations.
Wouldn't "Pseudo-random" be more appropriate? Or even "Undefined-cause mutation"? Random is just complete bollocks. I don't even know that there even IS such a thing as "random". I know at the quanum mechanical level things are said to be, but those may well just be the result of us visualizing a multiple-dimensional process in a 4-dimensional frame. But I digress. Could I please get Cali or Susu to address this for me? I could be wrong, but I'd like to know in what way if I am.
You are wrong. You are in good company though, Einstein got it wrong, Newton got it wrong... Random means non-deterministic and the question there is whether there are things that are unpredictable, not only practically but absolutely. You could be omniscient and still couldn´t predict them. And it turns out that this question does have an answer and it´s: Yes, there are such things. Now, there was the idea that there could be a local hidden variables theory in physics (there would be such variables, but we couldn´t know them).
For quite some time physicists thought that there was no experiment to distinguish LHV theories from non-deterministic ones. Then physicist named Bell found out, that the two did differ in a particular way: For a certain experiment LHVs predicted that a particular inequality would be met. In the mid 20th century it became possible to perform this experiment and refined versions have been performed ever since. And the inequality failed to show up - i.e. the universe actually is non-deterministic.
The direction in which a radionucleid emits gamma radiation is in fact random. When it decays is random as well. And the same holds for all interactions of atomic particles. The law of large numbers in probability theory leads to this becoming less aparent at larger scales. Certain macro-scale properties are means of quantum-scale properties and the more material there is the closer to the expected value the mean gets. Now, mutations are random in this sense - you´ve got radiation in random directions, which may or may not hit a strand of DNA or RNA at a particular place. You´ve got copying mechanisms which work on a molecular scale.
Now, mutations don´t "average out" in this way. You get two copies of a gene (once from each parent) at best and whether that one mutated or one in another egg or sperm of your parents does matter. So this is non-determinism that doesn´t go away. Population resampling is another process that is non-deterministic. And so are the processes into which it can be split (drift, selection and migration). Now, being random in this sense is a
strenght of evolution, not a weakness.
Here´s an analogous example: Radioactive decay. You´ve got a gram of
235Uranium. After 7.038*10
8[/sup}a about 0.5g will have decayed. In fact, the standard deviation is 9.88*10[sup]-12g, about 10pg. Now imagine a simple deterministic process: all
235Uranium nuclei take 1.4 billion years to decay precisely. We would have our gram of it, but of course we couldn´t say how old any of the nuclei are. So how much of it will be left after one half life? We can´t say at all.
So being random leads to
more predictability than being non-random. If evolution was non-random, fine tuning arguments would actually make sense - throw a grain of sand into a clockwork and it´ll stop working. Throw a grain of sand into cooking water and it won´t make a lot of difference. That´s because the macroscopic behaviour of a pot of cooking water comes about from the random motion of lots of water molecules and thus is robust to disturbances (which occur from the randomness as well).
Now, evolution didn´t stop when a 10km diameter rock hit earth at the KT boundary, it´s not a clockwork, it´s a boiling pot... I´d go so far as to say: In a non-random universe, evolution wouldn´t happen. Now would the 2nd law of thermodynamics hold. And I strongly doubt that science would work. Random has to obey rules which non-random doesn´t. Draw a card from a well shuffled deck with 52 cards and there´s a 1/52 chance to get an ace of spades. Reshuffle and do it again a lot of times and the frequency with which you draw an ace of spades will approach 1/52. Do this with stacked decks and it could be
any number.
Now, some people in evolutionary biology use random in a somewhat ideosyncratic manner - as statistically independent from selection coefficients, or from fitness or another measure. The problem with that is that the same people will call mutations random (they aren´t - the expected number of times a particular allele will be lost through mutation is N*E(f)*µ and E(f) depends on selection coefficients, or fitness) and call drift random (it isn´t either - selection is the expected change in allele frequencies E(f)-f'. This however leaves the random variable f-E(f) which of course still depends on selection coefficients or fitness). So for that one, there´s nothing in evolution that is random (but it has nothing to do with what causes mutations).
by 
