Busy hands syndrome: Unfortunately, I think one of the main criteria for a successful theory in linguistics is what I call 'Busy Hands". The theory has to be able to give people something to do, and write papers about. It has to contain some mechanisms that are non-trivial (and interact) that take time to learn. Then people are able to apply these mechanisms in a wide array of cases to linguistic phenomena. All this is independent of how good the theory is or whether it corresponds in some sense to reality or how elegant it is or how plausible it is from various points of view. To give a simple example, I think lots of cartography is like this. Just to be absolutely clear, I say this with no negative disposition toward cartography, which I am generally impressed by. One has a set of complex interacting principles that take some investment to learn, then one can apply those principles to a linguistically wide variety of phenomena. And any particular application takes some ingenuity and some insight, and more importantly takes up time, and paper and keeps the hands busy. The output of the work is rewarded by conference presentations, paper is in journals, citations, etc. But one can also argue that lots of that work is more or less descriptive, in the same sense as formulating phrase structure rules was largely descriptive. Another example like this is the early checking theory of minimalist syntax. Once again, it had a certain complexity and could be applied to a wide range of cases, which gave people something to do. So I guess the lesson is that one needs to be vigilant. Are you happy with a theory just because it provides you with something more or less mechanical to do?