The Phenomenon of Speech Perception

1098 Words3 Pages

Introduction
A major issue in the study of spoken language is how humans are able to successfully perceive speech in spite of its variability. For instance, speakers can differ in speech rate, dialect, and even in the rate of the syllables within the words of speech (Newman & Sawusch, 1996). Words in speech often become distorted as with coarticulation, a phenomenon in which speakers overlap words in normal continuous speech (Dilley & Pitt, 2010). In some cases, the overlapping of adjacent words can be so severe that words appear to vanish (e.g. “Do you want to?” becomes “Djawana?”).
Although there are no explicit signals in spoken language that indicate where boundaries should be placed in the same way spaces are used to separate words in written language, listeners are still able to both segment and recognize words in normal continuous speech (Darwin, 2012). Studies on speech perception show that there are several cues listeners use to segment and subsequently, perceive words. Currently, researchers recognize two categories of cues used by listeners to locate boundaries in speech: knowledge-driven processes and acoustic speech signals (Dilleys, Mattys, & Vinke, 2010).
Knowledge-Driven Processes
Phonotactics, syntactic expectations, and context encompass the knowledge-driven processes used by listeners to segment words in speech. Phonotactics “refer to the constraints on the ordering of segments within and between the words of a language” (Mattys & Jusczyk, 1999). People know what order sounds can come in and use this to help them segment speech. Using phonotactics, an English listener, for instance, would recognize that “the sequence, [br], is generally located at the beginning of a word, whereas the sequence, [nt], is typical...

... middle of paper ...

...German, 2011).
Summary
Speech perception continues to be an ambiguous phenomenon in the study of spoken language. Several studies have been conducted in the struggle to understand how humans are able to segment and recognize words in normal continuous speech, which is in itself, prone to changes in speaking rate and numerous distortions depending on the speaker. Thus far, researchers recognize that knowledge-driven processes and acoustic signals are both prominent cues used to segment words in speech. However, the use of these cues, particularly acoustic cues, in segmentation has been mainly tested in adults while studies in children remain widely unexplored. This experiment will examine the effect of acoustic properties, pitch and speech rate, on word segmentation in four to five-year-old children in the hopes of further understanding how humans perceive speech.

Open Document