(no subject)

From a comment by Richard H. Serlin on a post in Mark Thoma's blog, Economist's View:

Internally consistent is far from enough if the assumptions you start with are ridiculously unrealistic and you make little or no attempt to consider that in your conclusions from the model to reality.

He was talking about economic modeling in particular, but it's obviously applicable in situations involving mental models of all kinds.

Some of those referenced in my earlier post likely suffer from faulty assumptions, despite having internally consistent models.