Pause reports for spontaneous dialogic speech

Abstract

Native speakers of English (N = 40) reported occurrence and estimated duration of pauses for English and German spontaneous dialogue. Occurrence was underestimated significantly more for the German than for the English. Duration was underestimated significantly more for the English than for the German, and for positions where no actual pauses occurred (false positives) than for positions where there were actual pauses. The results for the English sample only are compared with similar estimates based on ethnomethodological transcriptional procedures (N = 7). Both methods are found to yield similarly nonveridical reported occurrence and estimated duration.Recent research (Carpenter & O'Connell, 1988;Chiappetta, Monti, & O'Connell, 1987;Duez, 1985;Gardner, McMillan, Townsend-Handscomb, Barrett-Bates, & O'Connell, 1990;Monti, O'Connell, & Chiappetta, 1990;Stuckenberg & O'Connell, 1988) has challenged the veridicality of reported occurrence and estimated duration of pauses in oral discourse. In general, both overestimates and underestimates of duration occur, and both false positives (report of pauses where no actual pauses occur) and false negatives (failure to report actual pauses) are frequent.There remain, however, numerous questions about these phenomena. The research methods have complicated this uncertainty in that Duez (1985) has used a method other than the remainder of the research listed, and the Monti et al. (1990) research was concerned not with actual oral discourse as stimulus, but rather with the surmising of pauses that would be appropriate were a printed text actually to be spoken. Several native languages have been studied, but, to date, no spontaneous dialogic speech has been investigated (except by Duez with the alternate method and only for French). The research of Gardner et al. (1990) did indeed investigate dialogic speech, but the dialogue was according to dramatic convention rather than spontaneous.Another recent research development has involved the questioning of psycholinguistic and ethnomethodological transcriptional methods for oral discourse. Deese (1984) has found all such methods, whether used for research purposes or for official or legal records, • 'to be inaccurate to varying degrees" (p. 21). More specifically, for the purpose of recording pause occurrence and duration, these methods have been found to be inadequate (O'Connell & Kowal, 1990a, 1990b. To date, there has been no direct comparison of an experimental method and a transcriptional method of identifying pauses.The present research extends the evidence regarding reported occurrence and estimated duration of pauses to spontaneous dialogue. It is hypothesized that native speakers of American English will be both more accurate 223and more efficient in reporting occurrence of pauses in English than in reporting pauses in German. It is also hypothesized that there will be a difference in the estimated duration across the two languages. The limitations of extant evidence necessitate that this hypothesis be nondirectional. Fin...

Estratégias de finalizações frasais e a análise automática da entoação: um estudo piloto

“…Em relação à duração da pausa, pesquisas indicaram que a partir de 200 ms, a pausa é mais saliente (DUEZ, 1985;FLETCHER, 1987;FRIEDMAN;OCONNELL, 1991;OLIVEIRA, 2002;SILVA, 2017). Além disso, Boomer e Dittmann (1962), como já vimos, optaram por diferenciar pausas de hesitação, que ocorriam com maior duração de, pelo menos, 50 ms, e pausas delimitadoras frasais, que se manifestavam já a partir dos 200 ms. Para este artigo, optamos pela reprodução desse princípio, valendo-nos de pausas com duração de 200 ms como marcadores prosódicos liminares de fronteiras interfrasais e pausas como duração maior ou igual a 500 ms como marcadores intencionais de segmentação frasal.…”