[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: courses



There seems to be many ways you can take Occam's razor, or parsimony. And there is often confusion between psychological notions of simplicity, and scientific parsimony. A little Karl Popper may not go astray here...

I don't cladistics is based on strict simplicity; it does not assume evolution always takes the simplest path to a "destination". If we look a it another way, the popperian way, cladistics explains regularity by maximizing explanation via genetic inheritance.

If character acquisition is assumed to be random (back to this in a minute), we could say that it is highly improbable that two organisms will have the same (or very similar) character sequence. Now this regularity needs explanation, and genetic inheritance from a common ancestor is one such explanation. Cladistic parsimony is an attempt to maximize genetic explanation. It is not really the psychological or evolutionary simplicity it is sometimes characterized as.

Thus genetic inheritance causally explains the highly improbable matching sequences, making it highly falsifiable - as it predicts highly improbable things. The more characters it explains this way - by genetic inheritance rather than chance - the greater it's breadth and falsifiability, or empirical content.

Now, of course is the real problem. We all know that character acquisition is not random, which reduces the improbability of matching character sequences. Biomechanical and environmental factors reduce the improbability of matching character sequences, and so reduce the explanatory power genetic inheritance and by implication cladistics.

So it seems to me that cladistics is the method for discovering the most powerful hypothesis for genetic explanation of regularities in variation, and biomechanical and environmental approaches are hypothesis of regularities of selection.


On Sunday, December 8, 2002, at 01:04 AM, philidor wrote:

Jaime remarked:
It goes beyond this, I think. Minimal evolution requires the idea that
to get to any one product, the course will have as few reversals,
duplicate scenarios, revisions, etc. in the genome. If there is one ...
well the idea is that there isn't. If nature is easy, or minimal in
evolutionary expression, then reversals shouldn't happen. And like I said,
the idea is that is doesn't happen. This means a minimum of support to
every two taxa, not a multitude ... simplicity in all modes.


Flukes happen. If they don't, then the situation is not random.
Flukes with a higher probability will, admittedly, happen more often than
flukes with a lower probability over a very long period. For example, if
you flip a coin 50,000 sets of 20 times each, you will probably see more
sequences of 10 heads in a row than 20.
But take a single set of 20 flips and you have a reasonable shot at seeing
one of the many outcomes you'd call weird.
Evolution is even worse than a coin flip for prediction. In a coin flip
there are only two exclusive outcomes. In evolution, with all the different
things that could be working more or less simultaneously, you can see many
actions and interactions.
That evolution will invariably be easy or simple in its effects seems to me
a blythe assertion.


John Conway, Palaeoartist

"Do I contradict myself? Very well, then I contradict myself, I am large, I contain multitudes." - Walt Whitman

Systematic ramblings: http://homepage.mac.com/john_conway/
Palaeoart: http://homepage.mac.com/john_conway/_palaeoart.html
Skeletals: http://homepage.mac.com/john_conway/_skeletals/