The research group on interaction and communication at the University of Giessen

The research group on interaction and communication was established in 1973 by Prof. Kiaus R. Scherer, who holds the Chair of Social Psychology in the Department of Psychology at the Justus-Liebig-University, Giessen. Activities of the group also include the teaching of psychology students in the areas of social psychology, the psychology of communication, and related fields. Our research was originally motivated by social psychology research into communication, and especially into nonverbal behavior. Thus, our research activities cover not only speech communication, but also other modalities of nonverbal behavior such as gestural and facial behavior. Because of special research interests, much of our research has focused on multichannel studies of nonverbal behavior, i.e., vocal, gestural, and facial behavior as well as psycho-physiological reactions. With regard to speech communication, our interests concentrate on nonverbal vocal bchavi" with particular reference to the following twt; pects: (1) the identification of the objective acoustic parameters of speech that are indicators of emotional and attitude-related states of speakers, and of personality traits of speakers, and (2) the attribution of these acaustic parameters to speaker states and traits by listeners in order to assess the functional relevance of vocal behavior for communication processes on the whole (see Scherer, 1978; Scherer and Bergmann, 1984). As an illustration of our research activities, three current research projects are outlined below. The "stress study" ("Indicators of affective arousal in communication", Wallbott and Scherer, 1985a; Scherer et al., in press, a) aims at a multichannel evaluation of person-specific responses to experimentally induced emotional arousal. Subjects were selected using two-dimensional questionnaire approach to coping behavior, and were exposed to cognitive and emotional stressors. In addition to obtaining vocal utterances during and immediately after exposure to stressful material, a variety of physiological measures as well as measures of facial behavior were obtained in addition to an assessment of subiects' feelings via questionnaire. The vocal indicators of stress that were analyzed were mean fundamental frequency (FO), FO-variability and FO-floor (calculated by means of the final FO-values in declarative utterances, see Ladd, 1983). In addition, the articulation parameters studied were the energy spectra and formant structures which were found to serve as personand situation-specific indicators of arousal (Tolkmitt and Scherer, 1984). Both aggregated FO-parameters and features of the F~) time cour.~e such as intonation avid the i~,, ~g,.: e~ ,~'~-J (projt:~t "Pragmatic ~ea~ings of intonation contours") are important for the affective impression of the voice. In tests on two different approaches with regard to the question of covariance between vocal cues and affective meaning, a signal-masking experiment (content-filtering, random splicing, reverscd " " ~ ' SHC..~,.,.j was conduc ted in order to isolate particular vocal cues and to determine whether judges were still able to infer the underlying affective meaning (Ladd et al., 1983; Ladd et al., 1984a; Scherer et al., in press, b). A very promising approach to the study of vocal cues in experimental settings is the use of digital resynthesis. Based on the same natural sentences, several selected parameters, such as intonation contour type, FO range, and perturbations were manipulated in isolation, indepcndent-