[seanmcdirmid]

You'd think that an empirical experiment with new programmers could prove the point, or at least provide conclusive evidence.

[tome]

Personally I wouldn't think that. I think that such things are far too complicated to be amenable to scientific study.

[seanmcdirmid]

I would normally agree, but specific usability claims are being made continuously about the benefits of purity, while many programmers think of Haskell has being anything but.