The literature reports a number of limitations that affect the teaching and learning of foreign languages, including a lack of teacher preparedness (Hu, 2005) and insufficient time for practice (Life, 2011). To mitigate these challenges, we proposed a self-directed learning (SDL) environment assisted by a combination of text-to-speech synthesis (TTS) and automatic speech recognition (ASR) technologies, as found in Microsoft Translator (MT), to examine whether this translation tool and its built-in speech features can promote the acquisition in pronunciation of English regular past tense -ed in a self-directed manner.

This study followed a pretest/posttest research design in which participants received autonomous but teacher-assisted TTS- and ASR-based treatment to learn about the pronunciation of English past -ed allomorphy: this suffix can be pronounced as play/d/, visit/ɪd/ and walk/t/, depending on the preceding phonological environment. We compared 29 participants’ performance in past -ed allomorphy by assessing their phonological development in terms of phonological awareness, phonemic discrimination, and oral production, as per Celce-Murcia et al.’s (2010) framework for pronunciation instruction. The t-test results showed that there were significant improvements in participants’ phonological awareness and oral production of English past -ed allomorphy. For the phonemic discrimination tests, the results were inconclusive: the participants only improved in recognizing the /t/ allomorph. These findings highlight the affordances of MT and its speech capabilities regarding its pedagogical use for improving second language learners’ pronunciation.