Naïve listeners’ perceptual learning of Chinese tones: The influence of L1 background in prosody and the effect of an auditory-image (AI) training paradigm


Summary, in English

The present perceptual learning study was conducted in an attempt to investigate to which extent naïve listeners’ L1 prosodic background affected their learning performance of a lexical tone language and whether an auditory-image (AI) training paradigm curtailed the online-learning process of lexical tones. Thirty-three Mandarin-naïve Swedish-native listeners (NS listeners) and 33 Mandarin-naïve English-native listeners (NE listeners) participated in a behaviour experiment designed to answer the question of whether native listeners of a pitch accent language (NS listeners) had an advantage over native listeners of a non-tonal non-pitch-accent language (NE listeners) in learning to perceive tones in Mandarin Chinese (MC). The results obtained – by on-line measuring the two groups’ general mean values of response accuracy (RA) and response time (RT) for the whole experiment and by comparing the groups’ block-to-block improvements in terms of mean values of RA and RT in both all and matched trials – indicated that (a) both NS and NE listeners made significant improvements across blocks in identifying tones in MC; (b) NS listeners were quicker learners of tones in MC compared with NE listeners; (c) NS listeners outperformed NE listeners in both matched (significantly) and mismatched (non-significantly) trials; (d) NS listeners’ intrinsic pre-activation of suffixes associated with initial stem tone to some extent impeded the online-acquiring of some tone combinations in MC. Furthermore, the AI training paradigm is more effective compared to traditional auditory training paradigm in terms of percentage gain (in correct identification) and training time. The results suggest that the short-term laboratory perceptual learning of lexical tones at a naïve level is determined by factors such as linguistic experience, training paradigm and procedure, and influenced by factors such as musical aptitude, psychoacoustic ground for and physiological bias toward perception of certain pitch patterns.
Different possible explanations were discussed regarding why initial falling tone combinations with high initial stem tone in MC were processed slower and less accurately compared with initial falling tone combinations with low initial stem tone for NS listeners. It remains an open question whether NS listeners’ intrinsic use of stem tone in predicting the incoming suffixes is transferable in perceiving other non-native tones aside from MC, until more studies are carried out concerning NS listeners’ perceptual learning on other tonal or pitch accent languages at same phonological level.


  • Languages and Literatures


  • Tone perception
  • Tone acquisition
  • Mandarin 声调感知


  • Mikael Roll