The Effects of Contextual Priming on the Intelligibility of Semantically Ambiguous Degraded Speech

Abstract

A significant problem in the area of speech perception is that noisy listening environments often make it difficult to understand what is being said. Furthermore, speech overwhelmingly contains ambiguous words that carry multiple meanings, which can make speech comprehension even more difficult. Previous research has found that spoken sentences containing ambiguous words (e.g. “the woman was told that the mint was used for making coins”) are harder to understand in noise than matched sentences without such words; we call this phenomenon the “ambiguity effect”. The current study examined individuals’ understanding of speech in noisy environments when this speech contains ambiguous words, and how context can influence this understanding. By manipulating the context in which sentences are presented, I examined whether listeners’ interpretation of the sentence can be biased towards a particular meaning, thereby affecting intelligibility. Participants listened to noisy sentences, each of which was preceded by a priming word intended to provide a particular context to the sentence. Two main predictions follow from this study. First, I predict that listeners will be able to understand less from sentences that contain ambiguous words, compared to those that do not. Furthermore, I predict that the priming words will be of greater benefit (particularly related primes) to listeners in understanding sentences with ambiguous words, rather than sentences without ambiguous words. Preliminary findings will be discussed in the presentation. This work will contribute to the current literature concerning how we use semantic information to understand speech in challenging listening environments.