On Lexical Ambiguity

In his 2000 book entitled "Meaning in language", Cruse argues that truly ambiguous words have multiple senses that exhibit the phenomenon that he calls antagonism, i.e. we can only focus our attention on one of the ambiguous senses. Those words that do not show antagonism may still have multiple discrete readings that can be detected via various tests. He also points out that the relatedness of senses or readings is continuous in nature. This continuum includes "clear cases" of homonymy, as well as various forms of polysemy. Lexical disambiguation (the assignment of a discrete reading to each word in the sentence) is usually considered as a necessary and well-encapsulated linguistic task. I argue that this approach is an oversimplification of a very complex situation to a degree where it is bound to fail and it does fail in practice.