Abstract
Fluency assessment is part of many official language tests (e.g., TOEFL iBT) which evaluate non-native speakers’ language proficiency. Operationalizing and validating the notion of fluency involves disentangling the different factors that influence fluency judgments. In a previous study the authors found a primary role in L2 fluency perception for pause characteristics of speech. Therefore, the present experiment was designed to zoom in on the contribution of silent pauses to fluency perception. Native and non-native speech fragments from turns in simulated discussions were recorded. The number and duration of silent pauses in these fragments was manipulated. The manipulations resulted in three conditions: NoPauses (pauses >250ms excised); ShortPauses (pauses >250ms received an altered duration between 250-500ms); LongPauses (pauses >250ms received an altered duration between 750-1000ms). These manipulated native and non-native speech fragments were rated on oral fluency by untrained raters using a Latin Square design. Preliminary results (using Linear Mixed Models) demonstrated that non-native speech was rated as significantly less fluent than the native speech. In both native and non-native speech, the NoPauses condition was rated significantly more fluent than the other conditions. Also, a significant difference was established between ShortPauses and LongPauses. Despite the clear difference in fluency level of native vs. non-native speakers, it is concluded that both the number and duration of silent pauses have an equally strong effect on fluency perception in native and non-native speech. These results suggest that, at least with respect to pauses, the notion of fluency is constant across native and non-native speech.
