Fast, deep-linguistic statistical minimalist dependency parsing

We present and evaluate an implemented statistical minimal parsing strategy exploiting DG charateristics to permit fast, robust, deeplinguistic analysis of unrestricted text, and compare its probability model to (Collins, 1999) and an adaptation, (Dubey and Keller, 2003). We show that DG allows for the expression of the majority of English LDDs in a context-free way and offers simple yet powerful statistical models.