Disclaimer: I haven't really been following the argument.
My thought: Does anyone think it's all a matter of human capital or all a matter of signalling? Most people I come across think signalling is very real and human capital investment is very real. This tendency to think of things in terms of "signalling vs. human capital" (or for that matter "rational voter vs. irrational voter") is a little strange to me. These theories have currency because there's evidence of both, and the maintain that currency usually because we have a pretty good sense of the balance between the two - i.e., when signalling is more and less important and when human capital is more and less important, or how people are irrational and how they are irrational and what sorts of trade-offs there are between the two.
Approaching this by saying something like "the human capital theory is wrong" or "the signalling model is right" seems like a funny way of approaching it. Of course the signalling model is right. Of course the human capital model is not wrong. This is all pretty interesting stuff without the contrived sense of conflict between the two.
It's like raging battles between nature and nurture. Meh.