Speech Perception

The term ‘speech perception’ is used in a variety of contexts. Critical terminological distinctions should be made. To begin, speech perception is not equivalent to language comprehension; rather, the perception of the speech signal constitutes only one subroutine of comprehension. The latter is comprised of a set of linguistic computations that can be initiated by auditory input (speech), visual input (text or sign), or somatosensory input (Braille). ‘Speech perception’ thus refers to the set of operations that transform an auditory signal into representations of a form that makes contact with internally stored information – that is, the stored words in a listener’s mental lexicon. There are at least three empirical approaches that are typically grouped under the term. Because they are different in terms of the structure of the input signal, the perceptual subroutines implicated, and the endpoint of the processes, it is important to be aware of these distinctions. (a) A large body of research on speech perception addresses effects at the level of individual speech sounds, that is, sublexical or prelexical units of speech. Subjects are typically exposed to vowels or syllables and execute particular behavioral tasks, such as discrimination or identification. For example, subjects listen to consonant–vowel (CV) syllables constructed from an acoustic continuum – for example, a series exemplifying a / ra/-/la/ tongue-shape contrast or a /bi/-/pi/ voicing contrast or a /tu/-/ku/ place-of-articulation contrast – and asked on presentation of a single token to identify the category. Such designs focus on sublexical properties of speech and examine questions concerning, for example, the nature of categorical perception in speech (e.g., Liberman, 1996), the phonemic inventory of speakers/listeners of different languages (e.g., Harnsberger, 2000), perceptual magnet effects (e.g., Kuhl et al., 2008), changes associated with first (e.g., Eimas, Siquelan, Jusczyk, & Vigorito, 1971) and second language learning (e.g., Flege & Hillenbrand, 1986), phonotactic constraints (e.g., Dupoux, Kakehi, Hirose, Pallier, & Mehler, 1999; Kabak & Idsardi, 2007), the role of distinctive features (e.g., Kingston, 2003), and other issues that can be experimentally evaluated at the prelexical level of analysis. This research strategy has been highly productive in the behavioral literature, and experiments of this type are now also prominent in the cognitive neurosciences, both in the context of imaging (fMRI, PET) and electrophysiological approaches (EEG, MEG, ECog). For example, using fMRI, several teams have examined regionally specific hemodynamic effects when subjects perform judgments on categorically varying stimuli (Blumstein, Myers, & Rissman, 2005; Liebenthal, Binder, Spitzer, Possing, & Medler, 2005; Raizada & Poldrack, 2007). These studies aim to identify areas responding differentially to signals belonging to different categories or that