Web-Inspired Sentence Complexity-Index

摘要

The intelligibility of speech is conditioned by its semantic complexity. Context is a precious cue in the recognition of words by the auditory system, as proven by the decreased recognition rates of semantically unpredictable sentences. The linguistic features of a test corpus directly impacts on the word recognition scores reported by subjects. An assessment method of the semantic complexity of short sentences is proposed on the base of counting search-results across the World Wide Web. Predictions are successfully compared with the empirical method proposed by Boothroyd et al., 1980, showing that the tool proposed could be used for the design of corpuses with controlled word predictability.