I'm still plugging away at this Prototype Theory stuff. One interesting idea is that it even affects our views of causation. Apparently we view the sort of billiard ball model of causation (ie A causes B to C) as a 'better' version of causality than, say, change resulting from the confluence of a number of factors. I spose that might in part explain why we are drawn to simplistic explanations for events. Excuse my penions geekery but an obvious case in my mind is the idea that the abolition of dividend tax credits 'caused' the closure of final salary schemes.
In similar territory I've just come across a good example of how the way we categorise things depends a great deal of conceptual models we already have in place. The example is to think of the category 'bachelor'. I assume we all have a clear view of what this category means - a man who is not married. It also seems fairly clear that as a category you are either in it or not - it's not a gradiated category. But what about a boy lost in a jungle who grows up into an adult on his own - is he a bachelor? What about the pope? What about a gay bloke in long-term relationship?
In fact whilst initially the designation 'bachelor' seems like a very straightforward yes/no bit of categorisation it is actually built on other assumptions, for example about marriage.