Tracking Down the Origins of Ambiguity in Context-Free Grammars

Context-free grammars are widely used but still hindered by ambiguity. This stresses the need for detailed detection methods that point out the sources of ambiguity in a grammar. In this paper we show how the approximative Noncanonical Unambiguity Test by Schmitz can be extended to conservatively identify production rules that do not contribute to the ambiguity of a grammar. Furthermore we can identify tree patterns that will never occur in derivations of ambiguous strings. We prove the correctness of our approach and consider its practical applicability.