Peter, there are loose, colloquial similarites between evolution by natural selection and search, but one has to be careful about reading too much into them. There are also more formal connections.
A better way of thinking about your pruning analogy is to imagine a search involving multiple agents, at various locations in a search space, where only a fixed number of agents is allowed. The death of an agent before reproduction is a kind of node pruning (without lookahead). This may eventually result in higher concentrations of agents around "fruitful" nodes. (But again, this is only an analogy.)
I suggest reading some of the monographs of John H. Holland for more detailed explanations of genetic search. His writing is both accurate and accessible. Of course, most of the examples he gives involve teleological* search, but the principles are the same for undirected search.
-Neil
* (in edit) By teleological I mean with an end goal in mind, not with a directedness in the details of the process. That is, artificial selection where the same gross measure of fitness is used throughout the process.
[This message has been edited by NeilUnreal, 05-28-2003]