Distinguishing universal and language-dependent levels of speech perception: Evidence from Japanese listeners' perception of English “l” and “r”

Abstract

Native speakers of Japanese may be unable to correctly identify the phonemes /l/ and /r/ in spoken English. Nevertheless, in perceiving English utterances, they, like native speakers of English, respond to the different acoustic patterns which convey /l/ and /r/ as if they are sensitive to differences in the vocal tract movements that convey /l/ and /r/. Support for this conclusion is provided by a study in which native speakers of Japanese and native speakers of English labelled stimuli along a synthetic /da/-/ga/ continuum when the stimuli were preceded by natural tokens of /s/ or /∫/, /al/ or /ar/. Each pair of precursors had contrasting effects on the location of the category boundary between /da/ and /ga/, and neither the direction nor the extent of contrast depended on native language experience. Significantly, /al/ gave rise to more /ga/ percepts than /ar/ for Japanese and English speakers alike, regardless of their ability to identify /al/ and /ar/, as such. Interpretation of these results rests on previous observations that the contrasting perceptual effects of /al/ vs. /ar/ and /s/ vs. /∫/ find parallels in the acoustic structure of natural utterances of /al-da/, /ar-da/ etc., due to coarticulation of the vocal tract movements that convey the preceding consonant and those that convey the following /da/ or /ga/. Apparently, native speakers of Japanese can be sensitive to the acoustic consequences of coarticulating /l/ or /r/ with /d/ or /g/ while being unable to categorize /l/ and /r/ as different phonemes. Preceding a language-specific level of perception where speech sounds are represented in accordance with the constraints of a given phonological system, there may exist a universally-shared level where the representation of speech sounds more closely corresponds to the articulatory gestures that give rise to the speech signal.