Predicting pauses in L1 and L2 speech: the effects of utterance boundaries and word frequency

Abstract

This paper compares the distribution of silent and filled pauses in first (L1) and second language (L2) speech. The occurrence of pauses of 52 L2 and 18 L1 Dutch speakers was evaluated with respect to utterance boundaries and word frequency. We found that L2 speakers paused more often than L1 speakers within utterances; but not between utterances. Similarly, only within utterances, L2 pauses were longer than L1 pauses. Regarding word frequency, both L1 and L2 speakers are more likely to pause before lower frequency words as compared to higher frequency words. These findings imply that L1 and L2 speakers' production processes may be similar in that (1) pauses at utterance boundaries are used for conceptual planning mostly and (2) lexical retrieval difficulties are comparable for L1 and L2 speakers. These findings furthermore imply that when using fluency for L2 testing, pause locations must be taken into account.

Counting ‘uhm’s: How tracking the distribution of native and non-native disfluencies influences online language comprehension

“…These typically fully crossed the factor Fluency with the particular target manipulation at hand (e.g., disfluency was equally likely to occur before known vs. unknown objects; Arnold et al, 2007), which also deviates considerably from the natural disfluency distribution in native speech (i.e., biased towards more complex referents). Finally, although disfluencies occurring before high-frequency referents are unlikely (compared to occurring before low-frequency referents), they are not rare (De Jong, 2016). Based on these arguments, we propose that the present findings should be taken as a case in point: Listeners are demonstrated to be capable of implicitly tracking different disfluency distributions.…”

“…Of course, the 'typical' and 'atypical' distributions used in the present study represent rather extreme examples of disfluency distributions; they do not reflect the much more balanced disfluency distribution observed in spontaneously produced (native and non-native) speech (De Jong, 2016). Nevertheless, we think that the present findings obtained with these arguably unnatural distributions are highly relevant and carry implications for our understanding of the processing of more natural fluent and disfluent speech, based on three arguments.…”

“…Listeners' so-called disfluency bias for relatively more complex referents has been attributed (at least in part) to prior exposure to the non-arbitrary distribution of disfluencies in native speech (Bosker et al, 2014a). Given that disfluencies tend to occur more often before complex lexical items than before relatively simple words (Arnold & Tanenhaus, 2011;De Jong, 2016;Hartsuiker & Notebaert, 2010), listeners can cleverly use disfluencies as a predictive cue in comprehension. We refer to this account as the associative account of the disfluency bias: listeners anticipate more complex referents upon encountering a disfluency due to association with the non-arbitrary disfluency distribution in spontaneous speech.…”

“…The authors explained the absence of a disfluency bias for non-native speech in terms of the atypical distribution of non-native disfluencies. Non-native speakers typically produce fewer low-frequency words, resorting to more commonly used lexical items (De Jong, 2016), and more disfluencies, such as silent and filled pauses (Bosker, Quené, Sanders, & De Jong, 2014b), negatively affecting non-native fluency perception (Bosker, Pinget, Quené, Sanders, & De Jong, 2013). As such, non-native disfluencies occur more often before high-frequency words than they do in native speech.…”

La vocal de relleno en español y en ruso: caracterización acústica e implicaciones teóricas

“…De Leuuw 2007;Clark y Fox Tree 2002), su distribución en la cadena hablada (Shriberg 1994; O'Connell y Kowal 2005) y su frecuencia de aparición, no solo en la lengua materna sino también en la interlengua de los aprendices de L2 (Dewaele 1996;Cenoz 2000;Wu 2008;García Amaya 2015;de Jong 2016), o bien han analizado los fillers con vistas a la caracterización idiolectal de los hablantes para fines forenses (Künzel 1997;Brander 2014, Cicres 2014) o para precisar su posible impacto en el rendimiento de los sistemas de reconocimiento automático del habla (Shriberg 2001;Vasilescu, Rosset y Adda-Decker 2010).…”