Pitch accent in spoken-word recognition in Japanese

Abstract

Three experiments addressed the question of whether pitch-accent information may be exploited in the process of recognizing spoken words in Tokyo Japanese. In a two-choice classification task, listeners judged from which of two words, differing in accentual structure, isolated syllables had been extracted ͑e.g., ka from baka HL or gaka LH͒; most judgments were correct, and listeners' decisions were correlated with the fundamental frequency characteristics of the syllables. In a gating experiment, listeners heard initial fragments of words and guessed what the words were; their guesses overwhelmingly had the same initial accent structure as the gated word even when only the beginning CV of the stimulus ͑e.g., na-from nagasa HLL or nagashi LHH͒ was presented. In addition, listeners were more confident in guesses with the same initial accent structure as the stimulus than in guesses with different accent. In a lexical decision experiment, responses to spoken words ͑e.g., ame HL͒ were speeded by previous presentation of the same word ͑e.g., ame HL͒ but not by previous presentation of a word differing only in accent ͑e.g., ame LH͒. Together these findings provide strong evidence that accentual information constrains the activation and selection of candidates for spoken-word recognition.

Perception of suprasegmental structure in a non-native dialect

“…For each of the extracted syllables, the total syllable duration and the duration of the vowel were measured, plus, across the voiced portions of the signal only, "ve f measures (minimum f ; maximum f ; f range; mean f ; standard deviation of f ) and two amplitude measures (mean rms amplitude; standard deviation of rms amplitude). All these acoustic data can be found in Cutler & Otake (1999).…”

“…For the original subject group, Cutler & Otake (1999) reported that subjects were signi"cantly more likely to decide that a syllable was H when it had high minimum f , high maximum f , high mean f , and high mean amplitude, and were signi"cantly less likely to decide that a syllable was H when it had a large f range or when it had a large f standard deviation. These four positive and two negative correlations also appeared, across the full set of 96 tokens, in the responses of each of the two new subject groups.…”

“…The materials were those used by Cutler & Otake (1999;Experiment 1): 32 bisyllabic Japanese words, all with the segmental structure CVCV (where V was always short), and each containing the mora/syllable ka. Half of the words had initial accent (HL), half did not (LH); these patterns were checked against the Tokyo reference data given by Sugito (1995).…”

“…A further joint analysis compared these two subject groups with the original subject group tested by Cutler & Otake (1999). This original group was tested at Soka, in Saitama Prefecture, and consisted exclusively of listeners from that area, i.e., native speakers of a variety with Tokyo-Japanese-style pitch accent (the horizontal-striped area in Fig.…”

Audio-Visual Prosody: Perception, Detection, and Synthesis of Prominence

“…In Japanese [39], it was found that pitch movement can help in the selection of word candidates. In Swedish, syllables in words are contrasted through lexical stress.…”