Hear, hear!
It's just a pity the book isn't actually freely available as *.pdf. There's just a measly 8-page brochure...
Quote:
Originally Posted by HydrogenBond
The data we collect for evolution is discontinuous data. Many of the early life forms appeared so long ago there is no practical way to get every data point. Here is the analogy. Say you had a child. At the day they were born and every day until they reached 18 years old you made an entry in a journal. That would be the entire data set. What we have is analogous to asking someone to randomly pick maybe 500 out of the 7000+ pages and using only that data, explain why the child is like he is at 18 years old.
Obviously the child progressed or evolved, but if certain critical pages are missing in the analysis, the conclusions could be out of touch with reality. For example, say at 13 years and 200 days, he was in a car wreck with his mother. This one event could alter his path for many years. If it was not included in the data sampling, one may assume it is genetic.
|
Not that I entirely disagree with you but, usually, with the help of statistic methods, one does not need all data points to construct an informed model, and there is some economy of effort in not looking for all data points, obviously.
Your life journal analogy would be more apt if the parent would reffer previous pertinent important events in a succinct way as if to make each journal entry more or less self-sufficient. So selecting about 1000 pages would make good odds for finding a reference to the car crash
That's what happens with evolution, descent with modification, as populations' genetic pools change with time but carry with them the important events of their path as morphological or biochemical limitations.
The
Tiktaalik case as presented in the brochure linked to at the end of the article linked above is a good example of the predictive abilities of current evolutionary models for basal tetrapod(omorph) evolution. True, it's just models, but it seems now they're pretty robust.
