Genetic "def" is not a definition at all

Stewart Schultz schultz at unixg.ubc.ca
Mon Nov 4 18:29:11 EST 1991


In article <1991Nov3.135246.1856 at ac.dal.ca> arlin at ac.dal.ca writes:

...
>:-)I'm glad to hear
>:-)that all the world's evolutionary biologists have admitted that
>:-)non-allelic changes take place.  I had no idea that my arguments would
>:-)be so convincing, or that the net reached so many people so quickly!  
>:-)I guess we can expect that the allelic "definition" of evolution will
>:-)be stricken from the 1992 textbooks, a much-needed clarification. 

Personally, I hope it's _not_ stricken.  It's a superb pedagogical 
introduction to the process of evolution, and insights gained therefrom 
have tremendous generality.  I see nothing philosophically objectionable 
about the fact that biological terms are used differently in different 
pedagogic or disciplinary contexts.

>If things keep going as well, in a few days all the world's evolutionary
>biologists will have conceded that they weren't really trying to
>*define* or describe evolution with their genetic statements, but that
>all along they were really trying to set up a mechanistic framework
>for evolution, a *causal* theory telling us why evolution occurs.

No, we're doing both at the same time, hence the possibility of 
confusion and the need to distinguish the two.  Evolution, by some 
previous historical definition, was hypothesized to be "caused by 
genetic changes in a population."   We can incorporate genetic change 
into a new definition of "evolution" if we believe the hypothesis to be 
well supported (this doesn't mean we won't change it in the unlikely 
event that the hypothesis is eventually disproven).  However, a genetic 
novelty can cause a wide range of new biotic and abiotic phenomena, and 
it's a personal preference how far along the causal chain one is willing 
to travel before the phenomenon is no longer "evolution."  Everyone may 
agree that the hypothesis is well-supported, but disagree where to draw 
the line.  For the purposes of "organic evolution" most biologists will 
stop at an organism's phenotype.  Others, as we have seen in recent 
posts, take a much broader view.  Your mileage may vary, independently 
of the biological facts; this is one aspect of what I mean by semantics.

Another aspect is that one might agree that genetic changes are 
overwhelmingly important, and still prefer to continue to label 
"evolution" those historically-defined cases that don't conform to the 
new definition.  A desire to hold onto some historical or cultural or 
personal concept of evolution might, for all I know, prompt one to 
assert that a migration barrier is "evolution" because it causes 
reproductive isolation, even though one disputes none of the facts.

>Here
>are three brief arguments why the genetic "definition" is really a
>mechanism:
>
>1.  Isolation is a distinct phenomenon from genetic change _per se_, so I
>argued in previous postings that it should be included in the "definition."
>However, Mr. Schultz has argued forcefully that one does not have to make
>reference to isolation in sympatric speciation, given that the phenotype
>leading to isolation is caused by the genetic change.
...
>The problem is not that genes can't cause phenotypes that cause
>isolation: ITS THAT WE WEREN'T SUPPOSED TO BE TALKING ABOUT CAUSATION!

If we want our definitions to reflect our knowledge about mechanism, 
then causation is precisely what we should be talking about.

But even if we agree about the relevant causal factors, we can still 
disagree whether "reproductive isolation" is necessarily an evolutionary 
process.  I say it is only if it's genetically-based, in which case it 
will be included automatically in my genetic definition; you 
apparently say that "isolation" regardless of its basis must be 
_explicitly_ included.  We seem to agree on the biology, hence my 
feeling that this is a purely semantic question.  You appear to be 
assuming some other, yet unspecified, definition of evolution.

>However, if we start out with this genetic  "definition" of
>evolution, we will inevitably have to start explaining that the
>genetic changes cause other phenomena that are a necessary part of
>evolution.  The reason is simple: the genetic "definition" is not a
>definition at all, but a mechanistic hypothesis.

It started out as a mechanistic hypothesis.  The hypothesis is now so 
well-supported that it's become part of the definition.  In other words, 
most evolutionary biologists are unwilling to consider non-genetically 
based changes in a population as evolution, regardless of how important 
they are, even to the evolutionary process.  You can't get much closer 
to an _a priori_ definition than this.  Is your objection to this 
prejudice based on any biological knowledge whatsoever?  Or is it a 
personal semantic preference?

>2. As I have stated before, a definition of X is a set of statements
>that describe X (listing its salient properties), allowing us to
>identify X and distinguish it from other non-X things with which we
>might otherwise confuse it.  The statements of the genetic
>"definition" of evolution refer to changes in the genetic composition
>of a population, and should allow us to identify evolution and
>non-evolution.

I see the semantic debate has escalated into the definition of 
"definition."  If the genetic definition satisfies your criteria, then 
why the quotes around "definition"?  Why not accept it as a real, bona 
fide definition?

>Some paleontologists spend their entire lives studying evolution
>without ever seeing a gene.  What they see are spatial distributions
>(in rocks representing temporal sequences) of the fossilized remains
>of organisms.  From this they infer (by a type of inference based on
>geologic principles, rather than any evolutionary theory) the
>existence of species of organisms that changed in their properties
>over time.  These sequential transformations are evolution, as we
>would all agree, but they are non-evolution by the genetic
>"definition."

Not necessarily.  The weight of biological knowledge overwhelmingly 
favors the hypothesis that these transformations are genetically-based.  
One can define evolution as genetically-based change (from genotype to 
phenotype) and then decide whether any specific case conforms to it on 
the weight of evidence that the change is genetic.  The weight of 
evidence is of course a continuum, and any two people might disagree 
where to draw the line, but in this case there's certainly a consensus.

>Of course, if we came out in the open and admitted that
>the genetic "definition" is not a definition at all, but the bare
>bones of a mechanistic theory, we could make statements such as "we
>believe that the sequential transformations, a kind of evolution seen
>in the fossil record, is *caused* by genetic processes."

Why would we need to say this?  We could alternatively say that the 
transformations "certainly have a genetic basis, and therefore are 
evolution by our agreed definition."  I can't see any objection to this 
as having anything to do with biology.

Perhaps your position is that, if we equate evolution with genetic cause 
then we can't strictly call these transformations "evolution" because 
they're phenotypic, and thus we're committed to saying they're "caused 
by evolution."  Is this what this discussion is all about?  This 
objection can be simply dealt with, by any of the many ways to 
incorporate genetic mechanism into a definition without confining 
"evolution" solely to genetic changes per se, but including some of 
their effects.

My previous point was that, even if someone, for the purposes of
discussion, prefers to restrict his/her concept of evolutionary
phenomena to genetic changes per se, I see no difficulty 
regardless of the biological context.  In some contexts it will be 
somewhat awkward, as we'll have to consider some phenomena not as 
evolution but as caused by it, but I can live with that.  In fact, I can 
live with just about any concept of evolution (though I may disagree 
about its biological importance), as long as we know at the outset what 
we're talking about.

...
>What we are doing is obvious: we all know that both genotypes and
>phenotypes undergo evolutionary changes, and if anyone tries to
>"define" evolution in terms of genotypes alone, it is because they
>think that the genotypic changes "cause" the phenotypic changes, and
>that therefore the phenotypic transformations can be reduced to the
>effects of the genotypic transformations.

Certainly.  And you apparently are saying this is bad because it implies 
that phenotypic changes aren't evolution after all, but rather are 
caused by evolution.  Would you say the field of evolutionary biology is 
crippled because it regards this distinction too picayune to bother 
worrying about?  Would you be satisfied if we agreed to allow the 
concept to extend up the causal chain to phenotype?

...
>There is no genetic "definition" of evolution, only a halfway attempt
>at mechanistic reduction of the phenomena of evolution to the effects
>of genetic changes.

A makes an assertion about rates of evolution.  B points out that its 
truth depends on what is meant by evolution.  A responds with an honest, 
careful, genetic definition for the purposes of the assertion.  What's 
the point of telling us we're all fooling ourselves if we think this is 
really, in some arcane philosophical sense, a definition?  Especially 
after you've stated above that it _is_ a definition by your criteria.  
How does any of this further our understanding of the original 
assertion?

Incorporating a concept of mechanism into a definition of a process does 
nothing _de facto_ to turn the definition into a hypothesis.  Our 
concept of the phenomenon has simply shifted to those cases that conform 
to the mechanism.  The more widespread the mechanism, the less we lose 
by the change.  In the case of evolution, a genetic mechanism is so 
widespread that I see no problem with the conceptual change.  It's 
inevitable and desirable for our language to evolve as it incorporates 
knowledge about the mechanisms of processes.

>In conclusion:
>Those of you who read my original posting know how I feel about this
>attempt to reduce evolution to one or a few simple genetic mechanisms.

We can take a multifarious historical concept of evolution (one that is 
probably still held by many), restrict it to those cases that are 
genetically-based, slice out only that segment that starts at the 
genotype and ends at the phenotype, and call that "organic evolution."  
Yes, this is a kind of reductionism, but not all the way down to a few 
simple genetic mechanisms, hence the inclusive term "genetic 
composition" or "genetically-based."  You apparently prefer not to 
perform any reduction at all, and think of a much wider set of biotic 
phenomena as evolution.  Fine, we'll keep that in mind in the future.

>There is no genetic "definition" of evolution, only a halfway attempt
>at mechanistic reduction of the phenomena of evolution to the effects
>of genetic changes. But this is getting ahead of things: first I have
>to undergo the inevitable flamings about how I'm trying to turn the
>discussion on its head by saying there is no genetic definition of
>evolution, though I'm sure this has been obvious to some all along.

You haven't been flamed for this because you haven't said it until now.  
What I heard you saying is that a definition is inadequate if based on 
"allele-replacements" as the only genetic mechanism, or fails to 
explicitly use the term "isolation."  Now you seem to be objecting to 
any genetic definition whatsoever, or are unwilling to see it as
a definition.

>:-) I only make semantic arguments. The reason, of course, is that my
>opponents are so smart they only make semantic mistakes.

The problem I think is that your opponents have different interests, and 
are having trouble understanding why all the brouhaha about the 
inconsequential fact that words' meanings vary by context, especially 
when the forum is a molecular biology newsgroup, and especially when
we all seem to agree on the biology.

--
Stewart Schultz
Botany Department
University of British Columbia
schultz at unixg.ubc.ca



More information about the Mol-evol mailing list