Still more Comments on Ewert's Dependency Graph

In what way was his approach flawed?

I’d have to re-read a lot to describe it well, but you could read it for yourself here. Briefly, there are other reasons why phylogenetic data might appear to fit Ewert’s model.

Well of course, since all that means is that he draws a new box for every set of genes with the same distribution. He can produce an exact fit to any distribution of data. A hypothesis that’s necessarily true can’t be falsified.

2 Likes

I think TSZ had an extended discussion of this too.

Is it? What’s the hypothesis, and how is it testable?

Something about phylogenetic data fitting a dependency graph better than a nested hierarchy? His approach was flawed, but he had a hypothesis to be falsified, making it actual science (a first for ID).

If I remember his alternative hypothesis was a tree vs a dependency graph. His comparison was gene families.