In general, "formalising mathematical vernacular" is a good source for work relevant to your question.  The field generally started in earnest with de Bruijn's Automath project.  

Harvey Friedman has done some nice work on leveraging his theory of explicit definitions for set theory (cf. [The Logical Strength of Mathematical Statements I][1], 1976) to give a language for doing mathamtics in a way that can be easily translated into proofs over ZFC.

I'm afraid I don't have a good reference for this: he gave an invited talk to the ASL 2006 conference in Nijmegen.



  [1]: http://www.math.ohio-state.edu/~friedman/pdf/histLogStr1.pdf