Abstract:

The importance of improving perceptual fluency in the early stages of second language (L2) listening is well recognized. Recently developed synthetic speech has the potential to be utilized in improving L2 perceptual fluency, as it can generate various L2 sounds. However, it is unclear whether the use of synthetic speech is effective for L2 listening, particularly for perceptual learning beyond single words, such as formulaic language. This study used the mere-repeated-exposure paradigm to examine changes in perceptual learning among 15 Japanese university students, who were English learners, using eight types of artificial intelligence generated synthetic speech: American and British English, with two types of male and female voices each. The mere-repeated-exposure paradigm suggests that the smoothness of information processing resulting from repetitive contact leads to an increase in favorability. In this study, repeated presentation of the stimulus object was the independent variable, and favorability, or preference, was the dependent variable. The stimulus object included two groups of formulaic sequences (48 high- and 48 low-familiarity) in two conditions: voice changing each time and voice remaining the same (24 formulaic sequences each). Participants were randomly presented with all formulaic sequences twice and asked to rate their favorability. The experiment was repeated with the same formulaic sequences 1–2 months after the initial test. PsychoPy® was utilized for both experiments, and Pavlovia was utilized for the second experiment. A Wilcoxon signed-rank test was conducted, and the results indicated that favorability of familiar formulaic sequences that were repeatedly presented using the same voice was significantly higher (Z = -2.045, p = .041). These results suggest that listening to synthetic speech enhances L2 perceptual fluency with repeated exposure as long as the perceptual information is consistent, thereby promoting perceptual learning. A discussion of presentation methods using synthetic speech that could enhance L2 listening is provided.

Keywords:

Synthetic speech, second language listening, perceptual fluency, formulaic language, mere-repeated-exposure paradigm, favorability.