Any model of natural language comprehension must have a lexicon containing the vocabulary of the language. In this chapter we do not propose to construct a lexicon, but rather to discuss some of the principles and problems, to give some examples, and to leave it to the interested reader to construct his or her own lexicon. As we will see, entries in a lexicon can be quite complex and may include rules and procedures for word formation, lexical selection, insertion, and so forth. The complexity of a model lexicon and its size will depend on the approach and on the type of project one undertakes. Computer models for natural language processing (NLP) usually have a lexical processor which is an autonomous component. That is, words of the input sentences are looked up in the dictionary, and the information associated with each word, including multiple meanings of ambiguous words, are extracted for further processing and possible disambiguation at the subsequent stages of syntactic and semantic analyses.
KeywordsNatural Language Processing Direct Object Lexical Processing Indirect Object Count Noun
Unable to display preview. Download preview PDF.