The discrimination of an /r/-/l/ speech continuum presented in an isolated monosyllabic context has been studied by a number of researchers. However, when one considers normal listening environments, it is unusual for people to hear syllables in isolation. In the present research, we investigated whether the characteristics of the discrimination performance for a /ra/-/la/ continuum presented under the isolated condition are also observed when the continuum is presented under nonisolated conditions, or more ordinary listening conditions. Two nonisolated conditions were employed: 1) the continuum was presented within a sentence, and 2) the continuum was preceded and followed by pure tones. Experiments revealed that the discrimination performance under condition 2) was similar to that under the isolated condition; however, the performance under condition 1) was different from those under the other conditions. The research suggests that the characteristics of the discrimination performance under nonisolated conditions are not necessarily identical to those under the isolated condition.

No abstract

The main hypothesis behind the perceptual assimilation model (PAM) is that listeners perceive non-native speech sounds based on articulatory similarities between the non-native speech sounds and native speech sounds. When non-native and native sounds are substantially similar, nonnative sounds are perceived as equivalent to native ones. Former research shows that the schwa vowel is perceptually assimilated to Japanese /a/ almost exclusively. In this study, we investigate acoustical cues that Japanese listeners rely on to assess the articulatory similarities between the English schwa and the Japanese vowels during the perceptual assimilation. Traditionally, the first two formants (F1 and F2) are considered to be effective for assimilation judgements; however, former investigations imply that these dimensions may be insufficient. In this study, we compared the schwa and the five Japanese vowels in the dimensions of vowel openness and backness using additional information, i.e., the third formant (F3) and fundamental frequencies (F0). The results of the analyses suggest that Japanese listeners use F1, F2, F3 and F0 information to assess articulatory similarities between the schwa and the Japanese vowels during perceptual assimilation.

No abstract

No abstract