Meta-Rules as a Basis for Processing III-Formed Input

If natural language processing systems are ever to achieve natural, cooperative behavior, they must be able to process input that is i l l-formed lexically, syntactically, semantically, or pragmatically. Systems must be able to partially understand, or at least give specific, appropriate error messages, when input does not correspond to their model of language and of context. We propose meta-rules and a control structure under which they are invoked as a framework for processing ill-formed input. The left-hand side of a meta-rule diagnoses a problem as a violated rule of normal processing. The right-hand side relaxes the violated rule and states how processing may be resumed, if at all. Examples discussed in the paper include violated grammatical tests, omitted articles, homonyms, spell ing/typographical errors, unknown words, violated select ion restrictions, personification, and metonymy. An implementation of a meta-rule processor within the framework of an augmented transition network parser is also described.