Phonetic characteristics of Mandarin-English code-switching in second language (L2) English learners

This study examined how phonetic characteristics differ across three speech conditions: Mandarin-only, English-only, and code-switched speech (where speakers switch between languages). To explore this, the study looked at three elements: sentences, phrases, and keywords within these conditions. Principal component analysis (PCA) was used to identify which components can best differentiate between these speech conditions. The results showed that the first three main components-overall pitch, vocal variability (jitter, shimmer, and harmonics-to-noise ratio or HNR), and pitch variability at the sentence level-explained 78.8% of the variation between the speech conditions. The findings revealed that Mandarin-only speech and code-switched speech involved more flexible and dynamic pitch modulation compared to English-only speech. This difference was likely due to the cognitive and phonological demands of Mandarin as a tonal language, as well as the added complexity of switching between two languages in code-switched speech. Additionally, vocal variability differed significantly across speech conditions, suggesting that differences in phonation type between Mandarin and English also played a crucial role in shaping their phonetic characteristics. Code-switched speech showed the most instability in vocal variability, Mandarin-only speech showed moderate instability, while English-only speech demonstrated the highest stability.