Tuesday, May 27, 2008

Evolution of Random: Random of Evolution

I attended the 2008 Western Evolutionary Biologists (WEB) conference on Sunday, sponsored by NERE. The winner of this year's Western Evolutionary Biologist of the Year (WEBY) was Kevin Padian of UC-Berkeley, in recognition of his many accomplishments in research and outreach in evolutionary biology. Kevin delivered a fun WEBY address, that mentioned his recent research briefly, and also discussed how evolutionists might better communicate with the public. In terms of bettering the abysmal anti-evolution statistics in the USA, Kevin thinks we shouldn't really focus much on the fundamentalists (who in my words are brain washed beyond rescue). Rather we should focus on a huge middle ground of people who may not understand the vast weight of evidence for evolution, and therefore have doubts, owing in large part to the anti-evolution propaganda machine that exists here. One of the topics Kevin brought up, which struck a chord in me because I just mentioned this in my Macroevolution class recently, is the difference between colloquial usage of words, and the meaning of those words in the scientific community.

Language and words, just like species and genes, evolve. Languages have been used to build phylogenetic trees to test hypotheses of human migration, and the meanings and relative usage of words often shift, sometimes radically, over time. An interesting example of this is what I've seen as a rather recent upsurge in colloquial use of the word "random". In fact, this "evolution of random" may have implications for how people understand the "random of evolution", the meaning of the word random in the context of natural selection.


Evolution of Random

In his best California-surfer-valley-girl accent, Padian illustrated current colloquial usage of the word random, "I, like, met this random guy in the mall the other day who...." (I actually can't remember the exact line Kevin used, but it was something like that). Kevin also defined current usage of random as identifying a non-sequitur, something that doesn't follow. I've certainly noticed this type of usage, and it seems to have surged tremendously in the past 5 years or so (perhaps because that is when I moved to California?) I've also thought of the colloquial use of random as meaning "unexpected", or even "unknown". For example, a "random" guy at the mall means to me that the speaker doesn't know the person. It also means that the reason the person was there is unimportant to the story, or unknown to the speaker. These colloquial meanings of random are quite different from statistical meanings, which is how I often think of its use in evolution.

Random of evolution
When I use the word random, I use the statistical meaning. Statistical randomness might mean that one possible outcome usually has an equal chance of being observed compared to another defined possibility. This is quite different than the colloquial usage that I mentioned above. A "random guy at the mall", according to the meaning often implied, does not include someone known to the speaker. So the speaker's dad is not included in the set of possibilities, which might be considered a decidedly non-random subset of all possible people, based on the statistical definition I presented above.

Kevin also made an interesting analogy, illustrating yet another subtlety in the statistical meaning of random. House fires may be modeled as random events. Given no prior information about one's house, there is a given probability that one's house will burn down. But if we have more information, we find out that most fires are caused by smoking in bed and arson. So, if a person is not a smoker, and has no pyromanical enemies, the chance of their house burning down is much lower than "average". Here, we are defining a random event (house fire) as stochastic, happening "by chance", yet recognizing that specific factors can alter the probability in specific instances. As Padian mentioned, if John Doe has an angry ex-spouse looking to torch the house lost in a divorce, the probability of Doe's house burning down goes up.

Given these sometimes subtly different usages of the word random, it's no wonder that sometimes people misunderstand natural selection, which has a "random" element. For natural selection to act, heritable variation must be introduced into a population. Evolutionists conceive of mutations as random events, and just as a certain prior knowledge of house fire risk can guide our estimation of probabilities, so can our prior knowledge of mutation. For example, flies exposed to UV light show a much increased rate of mutation. We might expect whole genome duplication events to be rarer than some other forms of mutation. But we don't mean that a mutation is a non-sequitur, or that a mutation is (necessarily) unknown to us or unspecified. And this is just a part of the entire process of natural selection!

Kevin's point in this is that we must be very careful about language. We must consider who is our audience, and what is their exposure and typical usage of different words. We must define clearly what we mean, and how our usage might differ from theirs. None of this is easy. As I grade exams, I see that understanding natural selection can be difficult for people; a difficulty probably caused by one random misunderstanding after another.

8 comments:

mikeg said...

todd, pz sent me here... just read all of your posts... global common descent! that's what i'm talkin' about... please keep us informed

Hank Fox said...

Great post! Thank you; gave me some new ideas.

One of my peeves is how people use the word "theory." I can't guess how many times I've had to tell someone there's a colloquial and a technical definition of the word, and that they're VERY different in meaning.

The bottom line in technical vs. colloquial meanings is that each one is fine for its purpose ... you just have to be careful to use the word that applies to the specific subject under discussion.

Fundies deliberately misuse them, though, and seem to know it but not care. If you try to make the point about some words having different meanings, their selective hearing kicks in and you can't get the point across to them. They actually have a vested interest in NOT understanding. What a nasty way to react to the world, huh?

Anyway, thanks. I like your blog, and I hope you write more.

Unknown said...

Very interesting post. I agree that confusion about the meaning of 'random' or 'chance' are often at the root of misunderstandings of evolutionary causality.

One of the very best discussions of this subject I've read is a paper by Gunther Eble, "On the Dual Nature of Chance in Evolutionary Biology and Paleobiology," Paleobiology 25 (1999), 75-87.

lylebot said...

Good points. I think there's another colloquial definition of "random" that leads to confusion: something to the effect of "all mixed up" or "no discernible pattern". For example, there's a sequence in The Da Vinci Code (yes, I read it... unfortunately) in which the protagonists need to figure out a password that's supposed to be an n-digit random number. They think they know what it might be, but what they think it is doesn't "look random"---i.e., it has sequences of repeated digits. So they "randomize" it in a particular non-random way. I'm sure I got some of the details wrong, but the basic idea is there: something that had some easily-identifiable patterns was considered "non-random", and changing it to something without easily-identifiable patterns made it "random".

(In fact, in neither case was it random: the original sequence was the first k Fibonacci numbers; the "randomized" one was the first k Fibonacci numbers in reverse order, I beleve.)

zilch said...

I, too, enjoyed this post. I must take issue, however, with your last point:

As I grade exams, I see that understanding natural selection can be difficult for people; a difficulty probably caused by one random misunderstanding after another.

Some of these misunderstandings are probably random, true; but others are Intelligently Designed.

Todd Oakley said...

Hank: I agree, misuse of "theory" is absolutely the best example of this!

Lylebot: Yes, I think your "non-patterned" definition of random is another example. Although pretty similar to "unexpected" in my post, there is a subtle and important difference. It's as if people think that "111111" couldn't occur randomly, because it's such a clean pattern. Probably people rarely buy lottery tix with "11111" because they think it's too unlikely to come up by chance, compared to 19283, or whatever their # is.
I too read Da Vinci Code, it was quite entertaining, actually. After reading it, I got inspired to write a conspiracy historical fiction novel about Darwin. I checked to see if it'd been done, and it has! It's pretty good, called The Darwin Conspiracy. Starts in the Galapagos, and goes back and forth between now and Darwin's time....

zilch: Non-random, and (un)intelligently designed indeed!!

odrareg said...

Now, will you just tell us what is random in evolution?

Random applies to muttion?

Random mutation means the mutation will occur only we don't know when?


And when it occurs we don't know what is mutated?



So, it is all statistical randomness that is being talked about in random mutation in the theory of evolution?


Again: there is the statistical probability that in the long terms or in the longest time mutation will occur?

Then when it occurs there is the statistical probability that such or such particular trait to be mutated will mutate.

Wherefore, from statistical probability we know when a mutation will occur in a population?

And when it does occur, the mutation, we just have to choose or we must have chosen a particular trait for the mutation to occur on and we wait until that particular trait suscedptibkle to mutation in the courwe of time while we wait, when all the factors are conducive so that probabiity becomes certainty, then that trait will mutate?


Anyway, you are not going to publish ths message, so what is the use, unless owing to statistical probability there is .00001 percent of its being published, just that I wait or try again and again and again and again until by statistical probability even against all probability it just so happens that in time "due' this message gets published....?


Anyway,suppose you tell us now what is random in evolution, in not more than a hundred words.



gerry

odrareg said...

[ This is the same post from me above, I am just correcting the poor spelling with some words. ]


Now, will you just tell us what is random in evolution?


Random applies to mutation?

Random mutation means the mutation will occur only we don't know when?

And when it occurs we don't know what is mutated?


So, it is all statistical randomness that is being talked about in random mutation in the theory of evolution?


Again: there is the statistical probability that in the long terms or in the longest time mutation will occur?

Then when it occurs there is the statistical probability that such or such particular trait to be mutated will mutate.

Wherefore, from statistical probability we know when a mutation will occur in a population?

And when it does occur, the mutation, we just have to choose or we must have chosen a particular trait for the mutation to occur on and we wait until that particular trait susceptible to mutation in the course of time while we wait, when all the factors are conducive so that probability becomes certainty, then that trait will mutate?


Anyway, you are not going to publish this message, so what is the use, unless owing to statistical probability there is .00001 percent of its being published, just that I wait or try again and again and again and again until by statistical probability even against all probability it just so happens that in time "due' this message gets published....?


Anyway,suppose you tell us now what is random in evolution, in not more than a hundred words.




gerry