What is second language pronunciation proficiency? An empirical study

Abstract

The current study set out to examine which segmental and suprasegmental factors discriminate different levels of global second language (L2) pronunciation proficiency. First, a total of 40 extemporaneous speech samples were elicited from Japanese learners of English with diverse experience/proficiency levels. Subsequently, experienced raters holistically assessed the global pronunciation qualities of the samples using the rubrics in IETLS Pronunciation Scale (Low to High). Finally, the dataset was submitted to a comprehensive set of segmental and suprasegmental measures. The results revealed that the raters attended to, in particular, the ratio of segmental errors with high communicative value (determined via the functional load principle) to distinguish between Low- and Mid-level L2 pronunciation proficiency. Other specific measures—segmental errors with low communicative value, the schwa vowel insertion in complex syllables, and the absence of word stress—played a significant role in the raters’ decision to assign high ratings to identify High-level L2 pronunciation proficiency.