Have you considered using the dictionary tools?
Wikipedia (subsets) can be found as DVD-image downloads of essentially pure HTML and as xml query subsets.
http://en.wikipedia.org/wiki/Wikipedia_database
It should not be impossible to convert one of the downloads to a standard dictionary format or to XDXF and then use the conversion tools to produce a Wikipedia "dictionary". Of course, how to bridge that divide is an exercise for the student.
Theoretically it could instead be converted to a searchable epub or FB2 ebook with embedded hyperlinks but I doubt the resulting multi-GB file would even open.