If you're OK with wikitext (the syntax MediaWiki uses), then the easiest thing to do is to grab raw pages with the URL parameter "action=raw", e.g. < http://en.wikipedia.org/w/index.php?title=Plutonium&action=raw >, which will return the raw source code of the page. This probably isn't what you want, though.
There's an API over at < http://en.wikipedia.org/w/api.php >, but I think you'd do better to simply scrape the printable version, which is accessible by adding a parameter "printable=yes" to a page's URL, e.g. < http://en.wikipedia.org/w/index.php?title=Plutonium&printable=yes >.
The printable version is still HTML, but it removes or otherwise makes nicer most of the problems you'd have. I think that with a few regular expressions (< http://enwp.org/Regex >), it would be simple to remove tables and remove most images (you'd want to convert math images to the TeX markup used as their alt text, for example, rather than removing them). You'd have to fix certain kinds of representational formatting, particularly superscript and subscript, as it's often quite relevant e.g. in physics/chemistry-related articles. This conversion process does involve losing some of the content (some Wikipedia content can't be adequately represented without a markup language supporting images, tables, etc.), but you'd get most of it without much work.