Assessing the Pedagogical Potential of Google Translate's Speech Capabilities: Focus on French Pronunciation

As the capabilities of web-based machine translation develop, online translators such as Google Translate (GT) have attracted computer-assisted language learning (CALL) researchers’ attention for their potential to aid second/foreign language (L2) instruction. Using its built-in text-to-speech (TTS) and automatic speech recognition (ASR) features, GT can be used for L2 pronunciation practice. The aim of this study (part of a larger project investigating L2 learners’ use of speech technologies in homework settings) is to examine the impact of self-regulated pronunciation practice using GT's TTS and ASR features on the development of French liaison (the re-syllabification of latent consonants when they appear in consonant-plus-vowel contexts across words, e.g., /z/ in tes amis [te.za.mi] “your friends”). Participants were 20 adult beginner learners of French studying at an English-speaking university in Canada. Their phonological development (i.e., awareness, perception, and production) was assessed before (pretest) and after (immediate and delayed posttests) the completion of a semi-autonomous, GT-based pronunciation practice. The results of the analysis of variance (ANOVA, the statistical method used) indicate that the proposed treatment led to a statistically significant improvement in liaison production between the pretest and the delayed posttest, while phonological awareness and perception remained unaffected, probably due to a ceiling effect.