Studies on speech perception by primates

通过反应时间任务，研究了日本猕猴、黑猩猩和人类的听觉敏感性、“响度”和对语音的感知。使用纯音（62.5 - 32k Hz，-10 - 90 dB SPL）测量听觉灵敏度和“响度”。根据反应时间数据，获得等响度轮廓。人类呈 U 形曲线，对 500-4k Hz 音调最敏感。黑猩猩表现出W形曲线，与旧世界和新世界猴子的曲线相似。奇曼奇人对 1k 和 8k Hz 音调敏感。然而，在2k-4k Hz时观察到听觉敏感性下降。然后研究元音的辨别。使用自然和合成的日语和法语元音。当元音对的成员彼此相似时，预计反应时间会更长。通过 MDSCAL 和聚类分析对反应时间数据进行分析。黑猩猩和人类之间的主要区别在于，黑猩猩区分 /i/ 与 /u/ 和 /e/ f … More rom /o/ 的反应时间更长。因此，这表明黑猩猩需要更长的反应时间来根据第二共振峰频率的差异来区分元音。人类和黑猩猩之间的物种差异可能部分地由W形黑猩猩的听觉敏感功能来解释。对辅音（主要是塞音）元音（/a/）音节的感知进行了研究。具有不同发声时间值和第二和第三共振峰瞬态频率的合成塞音也用于研究发声（/ga/-/ka/）和发音位置（/ba/-/da/）特征的分类感知。获得以下结果。尽管区分浊音和清音塞音很容易，但根据发音位置区分塞音对于黑猩猩来说却很困难。黑猩猩表现出对发音和塞音辅音位置特征的明确辨别。每一个境界的位置都和人类相差不大。较少的

Using a reaction time task, auditory sensitivity, 'loudness' and perception of speech sounds by Japanese macaques, chimpanzees and humans were studied. Auditory sensitivity and 'loudness' were measured using pure tones (62.5 - 32k Hz, -10 - 90 dB SPL). Based on reaction time data, equal loudness contours were obtained. Humans showed U-shaped curves and were most sensitive to 500-4k Hz tones. Chimpanzees showed W-shaped curves, which were similar to those of Old- and New-world monkeys. Chimanzees were sensitive to 1k and 8k Hz tones. However, a decrease in auditory sensitivity was observed at 2k-4k Hz.Then the discrimination of vowels was investigated. Natural and synthesized Japanese and French vowels were used. Longer reaction times were expected when members of a vowel pair were similar to each other. Reaction time data were analyzed by MDSCAL and cluster analyses. The main difference between chimpanzees and humans was that the reaction times for discriminating /i/ from /u/ and /e/ f … More rom /o/ were longer for chimpanzees. Thus, it is suggested that chimpanzees require longer reaction times for discrimination of vowels based on differenses in the frequency of the second formant. The species difference between humans and chimpanzees may be partially explained by the W-shaped chimpanzees' auditory sensitivity functions.Perception of consonant(mainly stops)+vowel(/a/) syllables was studied. Synthesized stop consonants with different values of voice onset time and 2nd and 3rd formant transient frequencies were also used for studying categorical perception of voicing (/ga/-/ka/) and place of articulation (/ba/-/da/) features. Following results were obtained. Although discrimination between voiced and unvoiced stops was easy, discrimination of stops based on the place of articulation was difficult for chimpanzees. Chimpanzees showed categorical discrimination of voicing and place features of stop consonants. The position of each boundary was not so different from that of humans. Less