Analysis of EFL speech production data according to theoretical models : English plosives by Native Spanish speakers

Abstract

Within language acquisition, speaking in a non-native language accurately is extensively recognized to be the hardest skill confronted by learners. It seems an arduous task as it demands the correct employment of second language phonology, a challenge that usually results in foreign accents. In the light of this problem, the present paper examines the realisation of English plosives by Native Spanish (NS) speakers in an attempt to understand the causes that lead learners to errors. In order to carry out the analysis, L2 (second language) perception models are reviewed as well as a list of the major factors that usually provoke foreign accent. The following investigation is based on a corpus of task-based conversational speech in which 3 pairs of Native Spanish (NS) learners were selected. Data were analysed with PRAAT, a tool that allows to see the spectrograms of speech together with the sentences and the sound. For this analysis, approximately 3 minutes of speech per person were scrutinized in which certain specific words were selected for the examination of stop consonants. Findings demonstrate different types of realisations of English plosives by NS speakers that reveal an evident foreign accent. The study has focused on Flege’s Speech Learning Model (SLM) so as to prove its suitability for foreign accent speech. However, results of this investigation show that using Flege’s model to understand the causes of foreign accent is not enough. Therefore, it would seem more efficient to examine a wider set of factors that SLM does not include in its predictions.