Introducing Linguistics Theoretical and Applied Approaches

This chapter brings together aspects of computation (as computer software) and linguistics (as provided by the study of natural language). The chapter first shows how natural language is often ambiguous, and the underlying structure is not immediately visible. Therefore, computer software that deals with natural language as input must cope with this inherent ambiguity. Readers come to realize that, if language were not ambiguous, we could reliably prepare computer algorithms that resemble the computer programs which many of us are familiar with: a sequence of actions that produces an outcome (e.g., a sequence of words forming a sentence and compute the meaning of a sentence). But because of ambiguity, we must use a different computational paradigm. Ambiguity comes in the form of word level ambiguity, syntactic ambiguity, and semantic ambiguity. Ambiguity causes a rapid increase in the number of possible interpretations of a natural language sentence. Different methods for avoiding the problems of ambiguity are detailed, including machine learning. The chapter ends with some examples of how computational linguistics may be applied.