Burden of proof and the limits of variation debate

Gordon Simons (simons@stat.unc.edu)
Fri, 28 Jul 1995 10:27:31 -0400 (EDT)

[Reflectorites: After some lively discussions between Glenn and myself
concerning "limits of variation" - about which side has the burden of
proof to establish their position, and logic -, he wrote:

"Excellent analysis of the issue. I think you are correct that the
imprecision IS the problem. That explains why the drumstick to lizard legs
example is thrown away so easily. That is major morphological change but
because they haven't defined the limit they can claim that it doesn't violate
their case.

I think you should post something on this. It is an important point. I
promise not to go "shooting my mouth off:" :-)"

I did present this analysis once - rather concisely - apparently unclearly.
So I apologize to those who find this fuller presentation of my analysis
redundant.

I do think it best, and certainly more interesting, to present the analysis
within its historical context. I am a bit embarrassed by my verbosity.]

On Wednesday, David Tyler wrote (near the end of a long message):

"Evidence for speciation does not constitute proof for large-scale
evolution, as the same evidence fits happily within the YEC and PC
thinking of variation within a kind."

then added:

"This is why [I] picked up the issue of limits to variation: the
neoDarwinians must justify this assumption if their evolutionary
theory is to be considered science."

The issue is "limits to variation" and whose task it is to justify their
position - those (the neoDarwinians) who are arguing against a limit, or
those who are arguing for a limit.

To this issue, Glenn replied:

"Why should they have to be the only ones to prove their point? Why are
anti-evolutionists exempt from having to justify their position? What you
are asking is the impossible. Since logically it is very difficult to
prove a negative (it can't be done), it seems that it would be more
fruitful for the anti-evolutionist to prove the positive i.e. that there
is a limit to variation."

then added:

"No one can prove that there is NO limit because that violates the laws of
logic. But you as a believer in limits, should be able to prove the
extent of variation."

So the question is: Who has the burden of proof to establish a "limit to
variation", those who believe in a limit or those who don't. It is NOT,
"Is there is a limit to variation?"

To the burden question, Glenn said, in effect, "unfair"; Tyler is asking me
to prove a negative, and this is extremely difficult, or, as he put it, "it
can't be done."

This is where I came in. I first asked myself whether establishing "no
limit" is "proving a negative". Then I looked at some clearer cases from
physics, the "exclusion principle" and the "parity principle" (one which
held up and one which did not). Both are claims of a limit acting in
physics. Suppose a physicist wants to argue against either of these
limits. Is he really being forced to "prove a negative?" Well, sort of -
a negative of what is already a negative (a claim of limit). But let's
allow that the physicist must establish a negative. How difficult is his
task? Well, not so difficult - from the standpoint of logic. All he has
to do is to find a single counterexample. This was done for the "parity
principle", and not for the "exclusion principle".

So then I asked myself, "How does this LOGICAL analysis hold up in the
case of a claimed "limit to variation?" From a logical standpoint, all
one needs to do is to find a single counterexample.

But then the real issue becomes clear. If those who argue for a limit of
variation were to provide a clearly stated limit (as in the physics
examples) - such as "no variation can take one beyond the boundary of a
given species" -, then finding a single counterexample does reject the
claimed limit, and the evolutionist can claim his Nobel prize :-). But,
alas, there is no clearly stated limit; this is the real issue: No single
counterexample can destroy a claimed limit that isn't adequately
described.

So indeed the burden is great on the evolutionist. He must establish
counterexample after counterexample, and face the argument, after each
such presentation, that he has not found a genuine departure from a limit
to variation; the limit keeps shifting, but it (supposely) is still "a
limit." Its like a marksmen trying to hit a target that is not in place.
He can shoot where it seems to be, and miss every time.

So, to conclude, the unfairness of placing the burden of proof on the
evolutionist - to show that there is "no limit to variation" - is not one
of logic (having to prove a negative), but one of imprecision.

Finally, in fairness to David Tyler, I shall give him the last word
(which he wrote on Thursday, following my more concise presentation):

"I am sympathetic with the comment. We lack the tools to describe
biological variability. How do we recognise limits? Stein-Cadenbach
says that one family of animals is "deeply separated" from others - but
can this be quantified? I would suggest that we are struggling within an
paradigm which has been defined and developed by neoDarwinians: because
they have not seen this as an area of research interest, they have not
developed tools to permit the serious study of limits to variation. This
is a challenge to biologists who do recognise natural groupings in animals
and plants: new tools need to be developed to quantify the observations."

Gordie