ABSTRACT
The Acoustic Phonetic Correlates of Stops in L2 English - is devoted to the acoustic- analysis of some 500 [p, b, t, d, k, g] produced by 67 non-native speakers English and the impact of their pronunciation on intelligibility. It has been known since Lisker and Abramson (1964) that if the Voice Onset Time (VOT) of [p] and [b] is ≤ of 25 ms, if that of [t] and [d] is ≤ 34 ms, or if that of [k] and [g] is ≤ 42 ms, poor or mediocre intelligibility ensues. The acoustic measurements indicate that, though [p, d, g] do not exist in several of the native languages under consideration, all speakers produced the phonemic contrast between [t] vs. [d] and [k] vs. [g] intelligibly. Pronunciation problems surfaced only for Arabic and Somali speakers because their attempted [p]s mask their [b]s acoustically. Both L1s lack [p]. Given the extremely high RFL of 98% between [p] and [b], poor intelligibility is unavoidable. Korean and Mandarin on the other hand produce perfectly intelligible [p]s and [b]s because they transfer the long VOT of [p]s from their L1s into English. Spanish and Slavic speakers produce a shorter VOT for [p], but it is not confused with [b] because their [p]s contain no voicing during the Close-Hold-Release phases of articulation. The measurements also reveal that Slavic speakers have a propensity to produce negative VOTs for [b, d, g], that is, their vocal folds begin vibrating even before closure takes place while these voiced segments are being articulated. However, their negative VOTs have no adverse impact on intelligibility.
