Horizon Hall, #4225
June 09, 2023, 09:30 AM to 12:00 PM
Listeners use a variety of cues in the speech signal to aid them in identifying nouns. For instance, English speakers use the phonological distinction between a and an to facilitate processing of following nouns (Nozari & Mirman, 2016; Gambi et al., 2018). Listeners’ use of cues is also modulated by the identity of the talker: listeners are less likely to use cues in nonnative talkers’ speech (Romero-Rivas et al, 2015; Schiller et al., 2020). Using visual-world eye-tracking, the current study explored native listeners’ use of morphophonological cues on the Arabic definite article in native- and foreign-accented speech.
The Arabic definite article /ʔal-/ provides at least three morphophonological cues to the identity of a following noun. First, the coda /l/ assimilates to following coronal consonants but not to noncoronal consonants (/ʔaddulfin/ “the dolphin” vs. s/ʔalbab/ “the door”). This assimilation carries two additional sub-phonemic cues depending on the coronal onset: coarticulation associated with emphatics and longer pre-voicing associated voiced stops.
In two experiments, participants saw picture-pairs accompanied by auditory instructions to click on one of them. In Informative trials, the two pictures’ names differed in their initial consonants (/ʔaˤsˤsˤaruχ/ “the rocket” vs. /ʔassullam/ “the ladder”). In Uninformative trials, initial consonants were the same (/ʔaˤsˤsˤaruχ/ “the rocket” vs. /ʔaˤsˤsˤaqr/ “the falcon”). In Experiment 1, participants listened to native-accented Arabic and in Experiment 2, they listened to foreign-accented Arabic. If listeners use the available cues, they should look at the target image earlier and/or longer in informative than uninformative trials. If foreign-accented speech disrupts language processing, cue use will be more evident in Experiment 1 than in Experiment 2.
Fixation latency and proportion looks-to-target were measured and analyzed. In Experiment 1, mixed effects models showed shorter latencies and higher accuracy in informative than uninformative trials in the Emphasis condition and shorter latencies in the Coronal condition. In Experiment 2, models revealed shorter latencies in the Emphasis condition and higher accuracy in the Coronal condition. No statistically significant effects were found in the Voicing condition for either experiment.
These results suggest that native listeners use some of the morphophonological cues on the Arabic definite article to facilitate noun processing. However, the cue use depended on the condition and the identity of the talker. Moreover, unlike previous findings that listeners do not rely on phonological cues in foreign-accented speech, the current results show that foreign-accented speech reduces phonological cue use but does not completely block it.