Decoding speech prosody in five languages

Recognizing Emotions in a Foreign Language

“…The results of this study demonstrated above-chance identification of all emotions in all languages, although English listeners performed significantly better in their native language (in-group advantage). The notion that vocal emotion recognition is influenced by linguistic similarity (Scherer et al 2001) was not strongly indicated by Thompson and Balkwill's (2006) data as English listeners demonstrated comparable accuracy to identify emotions in a related language such as German (67.5% correct) and in an unrelated language such as Tagalog (72.2% correct). However, potential shortcomings of this study were that very few tokens were presented to listeners in each language condition (n = 16) and the vocal exemplars were not extensively piloted to establish their emotional validity to a group of native listeners of each language prior to cross-cultural presentation (stimuli were included based on the impressions of two native speakers of each language).…”

“…While less abundant, supporting evidence of cross-cultural agreement in how basic emotions are recognized from a speaker's vocal expressions has been reported (Albas et al 1976;Scherer et al 2001; Thompson and Balkwill 2006;Van Bezooijen et al 1983). During speech communication, listeners attend to changes in pitch, loudness, rhythm, and voice quality (emotional prosody) to form an impression about the speaker's emotion state in conjunction with linguistic decoding (Wilson and Wharton 2006).…”

“…In fact, studies of both the face and voice commonly report differences in the overall degree of cross-cultural agreement witnessed across study groups and/or for specific emotion types (Beaupré and Hess 2005;Ekman et al 1987;Matsumoto 1993;Thompson and Balkwill 2006). Typically, these data reveal an ''in-group advantage'' for identifying emotions posed by members of the same rather than of a different culture (Elfenbein and Ambady 2002), underscoring that social factors play a critical role at stages of regulating and displaying emotions (i.e., nonverbal display rules, Ekman et al 1987;Mesquita and Frijda 1992).…”

“…Given the additional challenge of constructing stimuli which ''isolate'' the effects of emotional prosody in each language of interest (Pell 2006), it is perhaps unsurprising that cross-cultural research on the voice is not as rich as that on the face. These methodological considerations likely explain why many researchers in the vocal literature have adopted ''unbalanced'' experimental designs which involve repetition of items produced by a single cultural group to a number of different listener groups (Scherer et al 2001), or presentation of vocal expressions produced by speakers of several different languages to listeners from a single language/culture (Thompson and Balkwill 2006). While these research designs may not allow for the precise separation of effects due to the culture of the speaker versus the listener, they are highly valuable in elucidating how different languages and linguistic ability influence vocal emotion recognition in a manner which ensures rigorous control of both emotional and linguistic characteristics of the stimuli.…”

“…In a recent undertaking which adopted such an approach, Thompson and Balkwill (2006) studied 25 English-speaking listeners who identified the emotion conveyed by ''semantically-neutral'' sentences produced by speakers of English, German, Chinese, Japanese and Tagalog. In each of the five language conditions, listeners were presented four tokens representing each of four basic emotions (joy, sadness, anger, fear).…”