Robust Chart Parsing with Mildly Inconsistent Feature Structures
Carl M. Vogel, Robin Cooper · Arrow@dit (Dublin Institute of Technology) · 1994
We introduce the formal underpinnings of our theory of non#4;classical feature structures#2; The resulting expanded universe of feature structures has direct implications for robust parsing for linguistic theories founded upon feature theory#2; We present an implementa#4; tion of a robust chart parser for Head#4;driven Phrase Structure Grammar #7;HPSG#8;#2; The problem of relaxed uni#5;cation is in limiting it so that arbitrarily nongrammatical in#4; puts needn\tt be accepted#2; In the worst case excessively #11;relaxed\t parsers cannot provide meaningful interpretations to process#2; However parsers which can guarantee minimally nongrammatical #7;inconsistent#8; interpretations provide an important tool for grammar de#4; velopment and online robust processing of natural language#2; Our parser prefers maximally consistent interpretations and accommodates inconsistencies by discovering the minimal set of constraints that need to be relaxed for a parse to go through without employing backtracking or post processing#2; The system is di#12;erent from other related relaxational techniques for uni#5;cation grammars which require advanced naming of features whose constraints are allowed to be relaxed#2; Yet it is compatible with those approaches in that there is a well de#5;ned location for preferences on sources of inconsistency to be named as well as for resolution of inconsistent information#2; We use a simple approach to the problem of unknown words and suggest generalization of that for coping with missing and extra elements in an input