| Sumario: | This study addresses deconstructing L2 English comprehensibility and accentedness. The topic has interested scholars as comprehensibility and accentedness have been found partly independent phenomena, and modern L2 teaching aims for comprehensible rather than native-like speech. It is hence important to establish the constituents of comprehensibility and focus on those when teaching and assessing students. Previous research on the topic has found suprasegmental features of speech to be more strongly associated with comprehensibility than segmental features. The present study seeks verification for these results by using novel L1 speaker groups. This is important because some speech features may be more prevalent in some accents than in others, causing stronger effects on comprehensibility. In this study, speech samples were elicited from 40 L1 Finnish and L1 Finland-Swedish teenagers with CEFR B1–B2 English proficiency. The speech samples were rated for comprehensibility and accentedness by English-speaking listeners (n = 34) and analysed for twelve speech features, (e.g., speed of delivery, pausing, and segmental accuracy). Then, the speech data measurements were tested for connections with the ratings of comprehensibility and accentedness. The results strengthen our understanding that speech and articulation rate and mean length of run are important constituents of L2 English comprehensibility and accentedness and that segmental accuracy is more strongly linked with accentedness than comprehensibility. In addition, the study suggests that the link between pausing and L2 comprehensibility may be stronger than previously assumed.
|
|---|