This paper presents an examination of the variable lexical labels used by listeners to identify 16 social affective expressions in Japanese language, in audiovisual presentations. A free-labeling task allows an open approach to variability in the perception of social affects, that is constrained by pre-defined force-choice paradigms. 27 L1 Japanese listeners participated in the experiment. Subjects were asked to write down one word (noun or adjective) that best describes the intended expression they perceived from the speaker in each stimulus. Results cluster into coherent groups-relative to the expressions intended by the speakers. One Japanese-specific social affect, kyoshuku forms one cluster by itself among the 8 main clusters. This result emphasizes its specificity in Japanese culture: this expression was not singularized the same way by L1 French listeners from the same situation. The results also indicate the importance of a separation between assertive and dubitative speech acts in the meaning carried by prosody.

No abstract

In this paper, we investigate how French listeners with various levels of knowledge in the Japanese language, as well as Japanese native speakers, recognize the social affective meanings of utterances expressed during face-to-face interactions. A lexically neutral sentence consisting of 3 morae uttered with 8 different social affects by 6 native Japanese speakers during a conversation was used as stimulus for this experiment. Listeners had to recognize the expressions among the 8 possibilities. The perceptual results of the three groups of French listeners (levels 0 to 2) and of the group of native listeners are compared, to see if there may be different perceptual behaviors due to L1 or L2. Results show that L1 listeners' recognition rate was the most accurate, followed by French groups of level 2, 1 and 0. It was also found that the expression of surprise does not need any particular training, while the Japanese arrogant expression was quite difficult to be understood by French listeners. Further investigations of the visual information are suggested.

No abstract

This study investigates whether people can infer a speaker's social affective expressions in a language that is not their native language, by examining the variable lexical labels used to identify them. A freelabeling paradigm shows evidence of culturalspecificity in the perceptual behavior of listeners. Subjects were 16 US English listeners, asked to name the affective expressions carried by a set of stimuli. The results showed that their perceptual behavior was quite coherent relative to the expressions intended by the speakers. Also, despite the fact that "kyoshuku" is not a conventionalized expression in US culture, listeners' terms to describe it were in accordance with the intended expression. The Japanese prosodic expression of seduction was described as "happy" or "joy" by native listeners whereas "flirtatious" was a term used by US English listeners; a so-called "seductive" expression may be more conventionalized in western cultures than in Japan.