关键词: 言语知觉, 听觉理论, 动觉理论, 频谱对比效应, 启动效应

Abstract: A long-standing debate in the field of speech perception concerns whether specialized processing mechanisms are necessary to perceive speech sounds. The motor theory argues that speech perception is a special process and non-speech sounds don’t affect the perception of speech sounds. The auditory theory suggests that speech perception can be understood in terms of general auditory process, which is shared with the perception of non-speech sounds. The findings from English subjects indicate that the processing of non-speech sounds affects the perception of speech sounds. Few studies have been administered in Chinese. The present study administered two experiments to examine whether the processing of non-speech sounds could affect the perception of speech segments in Chinese listeners.
In experiment 1, speech sounds were a continuum of synthesized consonant category ranging from /ba/ to /da/. Non-speech sounds were two sine wave tones, with frequency equal to the onset frequency of F2 of /ba/ and /da/, respectively. Following the two tones, the /ba/-/da/ series were presented with a 50ms ISI. Undergraduate participants were asked to identify the speech sounds. The results found that non-speech tones influenced identification of speech targets: when the frequency of tone was equal to F2 onset frequency of /ba/, participants were more likely to identify consonant series as /da/; when the frequency of tone was equal to F2 onset frequency of /da/, participants had more /ba/ responses, especially for ambiguous intermediate stimuli. The responses of participants showed a kind of spectral contrast effect.
In experiment 2, speech sounds were two synthesized vowels—/i/ and /a/. Non-speech sounds were two tones and two tone complexes. The frequency of two tones was equal to the F2 frequency of /i/ and /a/ respectively, and the complexes were composed of two tones at the frequencies of the first two formants of the vowel /i/ and /a/ respectively. The ISI between non-speech primes and speech targets was 25ms. Participants were asked to label vowel with /a/ or /i/ quickly. The results found that in the tone-prime condition the identification of vowel /a/ was significantly faster with the frequency-matched prime than with the frequency-clashed prime; in the complex-prime condition, the identification of vowels /a/ and /i/ both showed priming effect.
The results of experiments 1 and 2 confirmed that the processing of non-speech primes affected the perception of speech sounds (including consonants and vowels) in Chinese listeners. It provided cross-language evidence for the auditory theory of speech perception, and indicated that speech perception experiences a process of pre-lexical spectral analysis and the process is shared with the perception of non-speech sounds.

Key words: speech perception, the auditory theory, the motor theory, spectral contrast effect, priming effect