I think I read this paper when it was published, and a few others they wrote on the topic - if it's the one I'm thinking of, the notion seemed plausible enough on the surface (I'm terrified of spiders, to an irrational degree), but their argument was actually borderline ridiculous in places. Didn't they argue that this was built into us in our almost ancient past, circa the dinosaurs? And yet (and I'm working from memory here), I recall practically on page 1 that they reported that some captive primates failed to exhibit the fear response. I worried terribly about the complete lack of learning in their model!

The dinosaur idea was credited to Carl Sagan, and probably not too serious.

The model is about a built in propensity to fear snakes, but it doesn't have to be active at birth - most people learn to fear snakes as they see their parent shouting and running around during a picnic when they see those funny moving sticks. So the learning aspect is likely strong, since it is needed to connect the built-in (and perhaps non-learning) snake detector module with the fear reaction.

Phobias are fairly amenable to removal though training and extinction. But this removal seems to be about the introduction of a new inhibitory connection (from frontal executive functions?) than the removal of the old stimulus-fear association.

Sorry, I'm just now getting back to respond. Your comments are well-taken! I'm probably mis-remembering the paper, since it left a bad impression on me. I recalled them actually arguing for the ancient evolutionary beginning to the module, and not just off-handedly referencing Sagan on it. I'm deferring to you since you've read it in the last 6 months, unlike me! I guess I just remember thinking that they didn't adequately show at all that there was such a disposition, given that learning actually does play so large a role in most fears of those things. (And given that, unless I'm wrong again, which is possible, there are species of monkey that actually *eat* snakes.)

I went into the article thinking the idea was quite clever, but never found it worth pursuing your larger question of where the module might be located simply because I didn't think they had adequately demonstrated any such module existing.

Thanks a bunch for posting this; it's really put the whole argument back on my mind!