Interactions between linguistic constraints: Procedural vs. declarative approaches

Grain Size

“…These mappings can be de ned for large sub-classes of lexical entries. For example, the SSNounVerb mapping de ned below applies to verb subcategorizing for a subject which is mapped to an agent in the semantic zone (see Emele et al, 1992). This de nition would be inherited by all appropriate entries (intransitive v erbs where the subject is the agent): by selecting the appropriate mapping for classes of entries, it is possible to hide the mapping from the acquirer since these mappings would be de ned in a lexical class, not in an instance.…”

“…is expressed in the terms of the debate on e ability, a v ersion more attuned to the environment of computational microtheories is (28ii.). (28) i. Hypothesis of Practical E ability: Each s e n tence can be translated into another natural language on the basis of a lexicon with equally limited polysemy. ii.…”

Tricolor DAGs for machine translation

“…1 Attempts to design a pure interlingual MT system were therefore abandoned, 2 and the notion of "semantic transfer" [24,22] came into focus as a practical solution to the problem of handling the language mismatching. The constraint-based formalism [2] seemed promising as a formal definition of transfer, but pure constraints are too rigid to be precisely imposed on target-language sentences. Some researchers(e.g., Russell [14]) introduced the concept of defeasible reasoning in order to formalize what is missing from a pure constraint-based approach, 1 For example, Yasuhara [26] reported there was an overlap of only 10% between his group's English and Japanese concept dictionaries, which covered 0.2 million concepts.…”

The nondirectional representation of Systemic Functional Grammars and semantics as typed feature structures

“…The current state of the art in natural language processing calls for components Callable of sophisticated deep semantic analysis, modular representation of resources, and re-usability of those resources across different NLP applications. Furthermore, it has been demonstrated that the sheer diversity of interactions between distinct kinds of linguistic information is guaranteed to defeat any staged approach to generation/understandiug that successively maps between levels of representation [1]. One strategy for addressing these problems is to stratify resources so that inter-stratal mappings are simplified.…”