METHODS & DESIGNS Lexical ambiguity of words used in English text
Bruce K. Britton · 1978
Lexical ambiguity is generally assumed to be widespread in natural languages but quantitative estimates have not been available. At least 32% of the words used in English text were found to be ambiguous in studies reported here. This figure is probably very conservative. Temporary word definitions established for special purposes occurred in 30% of a sample of texts. This paper is concerned with lexical ambiguity, ambiguity that occurs when a word has more than one meaning. For example, in The soldiers liked the port, the word port can refer to a wine or a harbor. In the literature on language processing. it is widely assumed that natural languages are ambiguous (Garrett, 1970). Substantial amounts of theoretical and experi mental work have been devoted to models of word and sentence processing to account for the resolution of lexical ambiguity. Apparently, no quantitative estimate of the amount of lexical ambiguity has been attempted. Of particular interest is the frequency of ambiguity among words most commonly used in discourse. Rare words are less interesting. If a word is used rarely, it contributes little to the ambiguity encountered by language processors. About 41%of the words appearing in Kucera and Francis' (1967) word frequency count occurred only once in the million-word sample. All of these words combined constitute only 2% of the words used in English text. If lexical ambiguity is extensive among words used in English text discourse, as has been generally assumed, then further investigation of processing mechanisms for resolving ambiguity may be warranted. If there is rela tively little lexical ambiguity among the words com monly used in English text, then research attention might better be directed elsewhere. In this paper, lower bounds are determined for the amount of lexical am biguity of words used in English text, two special types of ambiguity are described and quantified, and a refined method for quantifying the ambiguity of indi vidual lexical units is presented. STUDY 1