The Talk.Origins Archive Post of the Month: September 2000

Subject:    Re: Good mutations happen and increase info
Newsgroups: talk.origins
Date:       September 6, 2000
Message-ID: wilkins-4CA209.10212407092000@news.unimelb.edu.au

In article <8p4ron$s7a$2@news.duke.edu>,
mturner@snipthis.acpub.duke.edu (mel turner) wrote:
> In article <39B51FEE.58E3983C@bellsouth.net>, scott@home.com wrote...
> >In <39B51FEE.58E3983C@bellsouth.net>, Mike Owen <m_owen@bellsouth.net> writes:
> >>scott@home.com wrote:
> >>[snip]
> >>
> >>> There is no reason to think that, over long time spans,
> >>> these rare positive mutations will accumulate. Therefore,
> >>> there is no reason to expect any positive overall net
> >>> change.
>
> Recall "selection", the thing you've suggested was a tautological
> truism? Are you now arguing it's false? If it's really a thing that's
> necessarily true, then those positive mutations would have to
> accumulate.
>
> >>> You will have a difficult time indeed attempting to convince
> >>> your t.o. evolutionary colleagues that evolution is inevitably
> >>> in a positive direction.
> >>>
> >>> There is no reason to think that the greater number of
> >>> negative mutations would not, over long time spans,
> >>> completely negate the occasional rare positive mutation.

> >>Wow. I have never seen anyone demonstrate so succinctly their
> >>complete and total misunderstanding of evolution.
> >>
> >You should have left the original context.

> >>What exactly is to prevent positive mutations from "accumulating"?
> >>If you throw a penny in a jar every day, they accumulate. Negative
> >>mutations, BY DEFINITION, hinder reproductive success compared to
> >>the general population, and therefore are differentially selected
> >>against.
> >>
> >This might be an accurate analogy if throwing
> >pennies into a jar was anything like tossing
> >a positive mutation into a population. As it
> >is however, while your pennies may accumulate
> >this is nothing like how nature handles mutations.
>
> Right. In nature, the 'pennies' are also happily reproducing more of
> themselves.
>
> >Since "Negative mutations, BY DEFINITION, hinder
> >reproductive success," does this mean that your
> >"Natural Selection" is nothing more than a tautology?
>
> No, but those definitions of "negative mutations" & "positive
> mutations" are, to some extent.

> Again, if "Natural Selection" is nothing more than a tautology, then
> it is a statement that is obviously, necessarily true, right? And so
> it would seem your earlier claim that "there is no reason to think
> that, over long time spans, these rare positive mutations will
> accumulate" must therefore be obviously, necessarily false. That
> inescapable, necessarily-true process of natural selection is a reason
> that positive mutations must accumulate.
>
> The claim that NS is tautologous amounts to an argument that "positive
> mutations" could be defined as ones that must accumulate in the
> population, right?
>
> Or was there a flaw in your suggestion that NS is "just" a tautology?
>
> let's see...
>
> "Differential reproductive success in a given environment that is due
> to heritable differences among individuals causes adaptive
> evolutionary changes* in the population."
>
> [*and a lack of unadaptive change, as in cases of stabilizing
> selection]
>
> Yep, clearly tautologous. ;-)
>
> http://x69.deja.com/getdoc.xp?AN=622752063
> http://x69.deja.com/threadmsg_ct.xp?AN=588205794
>
> link to some past discussions of the tautology thing
>
> also, there's
> http://www.talkorigins.org/faqs/evolphil/tautology.html
>
> cheers

I've been thinking a bit more about this recently, in the light of correspondence with a feedback contributor. A large part of the problem is how to make it clear that just because something is a formal consequence of definitions and axioms, it does not make the result either nonscientific or useless as a hypothesis.

So, let's take a simple example of a tautology. By definition, 1+1=2. Nothing could be more tautologous than that. We could infer that the use of arithmetic is unscientific, because it is not falsifiable. Right?

Wrong. The point about any actual situation is that it need not be a 1+1=2 case. There is no necessity that 1+1=2 for, say, the mixing of two equal volumes of a gas. If you mix 1 volume of hydrogen and 1 volume of oxygen (plus a little heat), do you get 2 volumes of gas? No, you get something else.

This "could have been otherwise" feature is what makes the use of logical tautologies interesting. If we learn that 2 added single volumes of gas do add up to 2 volumes, then we learn that the two gases aren't reacting with each other. If we learn that they do not, we learn that they are reacting with each other.

The thing about selection is that if the conditions are met, then the outcome is a logical necessity, but the conditions need not be met. If we predict that a given outcome will occur, and it does, then selection explains it (because explanation is the logical inference of conclusions from empirical premises). If it fails to predict the outcome, then either we haven't done the logic right (which is a learning experience) or we should look for other causes (which is a learning experience).

So, is selection falsifiable? Strictly speaking nothing is falsifiable, at least not as a modus tollens (see http://www.bu.edu/wcp/Papers/Logi/LogiDagl.htm), because unless a law or hypothesis is intended to be an exceptionless universal statement, exceptions do not make it invalid. No law of science is exceptionless as of now, because we do not have a valid Theory of Everything, and even if we did, exceptional laws would still be allowable in restricted fields.

However, we need, for something to be scientific, for there to be good reasons to abandon a hypothesis if needs be. What reasons for abandoning selection would there be, even if it is a tautology. Note that I am allowing for the purpose of argument that selection is a tautology. I do not think it is (just) that.

Why abandon selection, then? We would abandon it if it failed to describe enough of the cases that are of interest in biology. For example, if apparent adaptation instances were not, even plausibly with a just-so story, explicable by selection, then we would start to not use adaptive selection as an explanation. But even more than that, we would expect scientists to stop using selective explanations if there were no good uses being made of selection in detailed cases. Scientists are the ultimate cynics - if some idea isn't doing any work, they won't hitch their wagons to it.

So, selection can be a tautology and still be abandonable, if not actually defeasible. As it happens, selection is not a tautology in another sense - this is the actual case by case sense. It is not a tautology to say that the frequency of melanic moths increased during industrial pollution due to selection. There is no way one can claim that this is a theorem of logic or definition. On this account, selection is a schematic of an explanation, and the explanation of each specific case is detailed, empirical, and disconfirmable if not falsifiable as such.

An implication of this is that selection is an effect, not a mechanism, and that is a conclusion I am happy to adopt. The causal mechanism of the increase in industrial melanism in Kettlewell's moths is bird predation. The effect is selection, and so adaptation. One can dress that up with terms like proximal and distal mechanisms, and such, but selection is an explanatory schema, not a physical mechanism.

Another less dramatic way to say this is that selection is a dynamic (one described by the Fundamental Theorem of Selection) that is sometimes observed of living populations due to the interplay of the ecological interactions of members of that population with parts of its environment and with other members.

-- 
John Wilkins, Head, Graphic Production, Hall Institute
<http://www.users.bigpond.com/thewilkins/darwiniana.html>
Otto: Apes don't read philosophy.
Wanda: Yes they do, Otto, they just don't understand it.

Subject:    Re: Fun with probability (was Cambrian Explosion)
Newsgroups: talk.origins
Date:       September 30, 2000
Message-ID: 8r5sjt$onn$1@bioinfo.med.utoronto.ca

DNA and RNA strands have distinct ends and directionality. There's a convention for determining how to label the ends and how to convey directionality. Let's just say that one end is called the 5' (five prime) end and the other is 3' (three prime).

When you write down the sequence of a single-stranded polynucleotide it's written in the 5' -> 3' direction unless you specifically say otherwise. All the standard tables of the genetic code are written this way and most of the tables specifically label the 5', middle, and 3' nucleotides.

Thus the codons for glutamate are 5'-GAG-3' and 5'-GAA-3' but you can just write GAG and GAA because everyone knows that you are following the convention. (You could also write the codons as 3'-GAG-5' and 3'-AAG-5' but in that case you would have to label the ends to specify that you weren't following the normal convention.)

The codons are present in mRNA. Messemger RNA is produced by transcribing a gene. Only one of the two strands of DNA is copied during transcription and this strand serves as the template for the new mRNA. The template strand of the gene and the mRNA are complementary and they form a double-stranded DNA-RNA hybrid during transcription. Now, here's the complicated part - the two complementary stands are orientated in opposite directions. One is 5'->3' and the complement is 3'->5'. This is a property of the double-helix and the base-pairs.

Because the strand are antiparallel the GAG and GAA codons in mRNA are complementary to nucleotides in the template strand of the gene like so,

If you were writing the sequence of the template strand it would be AAATTCCTCTTCCTC because the convention states that you write from the 5' end to the 3' end.

Genes are composed of double-stranded DNA. I've shown you the sequence of the template strand above. The other strand is called the coding strand or the non-template strand. Here's the sequence of the gene.

Note that the coding strand of the gene has the same sequence and the same orientation as the mRNA except that there are T's instead of U's. That's why it's called the coding strand - because it has the standard codons in the right orientation. So, when you talk about codons in DNA you refer to the nucleotide sequence of the coding strand written in the standard 5' -> 3' direction. (There are other reasons for the convention of defining the direction of a gene using this "top" strand.) The glutamate codons in DNA are GAG and GAA.

Jack claimed that the glutamate codons in DNA were CTC and CTT. This was not in accord with the standard convention for writing these sequences. It was wrong for two reasons; (a) he didn't write them in the correct orientation (b) he used the wrong strand of DNA. At first I didn't even realize how he came up with those sequences.

Is selection falsifiable?

Co-Post of the Month: September 2000

DNA Directionality and Codon Conventions

Co-Post of the Month: September 2000