There are (at least) two common reasons why strange and mangled URLs may show up as crawl errors in Webmaster Tools.
The first possibility is that someone has copied your pages (or some other pages that link to yours) and mangled the links in the process. This happens more often than you might think; see e.g. the sixth question in this Google Webmaster blog post.
The other possibility is that Googlebot itself is trying to follow what it thinks are JavaScript links and making a mess of it. You can usually tell these two cases apart by visiting the referring page (which should exist and be accessible, if Google managed to crawl it to begin with) and looking for the name of the target page in its source.
Either way, there are basically two things you can do: either just ignore the links, or come up with some rewrite rules to try and map the broken URLs into working ones. If you can see an obvious pattern in the URLs, and are familiar with regexps, I'd recommend the latter approach — it'll clean up your crawl error list and maybe even give you a small and rather cheesy, but real, PageRank boost.
A third option, if you find that someone's been copying your content without permission, is to try and get them delisted. You could even send a complaint (and/or a formal takedown request) to their hosting provider, if you believe it justified. Of course, given that they are apparently linking back to your site, you might not necessarily find that worth the effort.
