A clarification on the requirement of "entry level": I would like to work with entry level proofs so as to concentrate on the linguistic means used in proofs while still being able to understand the proofs. I am mostly interested in the expressions used to denote premise->conclusion substructures.

I will be looking for a way to represent the proofs as a symbolic structures which would enable further processing. Formulae and symbols will be treated along with the "natural language bits". Ideally an identical proof step expressed as a formula and expressed in words should yield the same output structure. The long term goal is the interpretation of arbitrary proofs in expressions of Artemov's Justification Logic (see e.g. logic.uconn.edu/readings/Milnikel1.pdf). Checking the correctness of a proof is not a goal.