Prosodic patterns of code-switched speech

Code-switching (CS), the use of features of more than one language in an utterance, has been thoroughly researched through the syntax lens but little is known about the relationship between syntax and prosody. The main prosodic elements that have been investigated recently are fundamental frequency (f0), voice onset time (VOT), speech rate, and vowel duration, with some studies finding a phonological cue indicating a succeeding CS. However, it is still unclear what this prosodic cue is, and how listeners use it to predict an upcoming switch. This study aims to fill this gap in the literature by investigating the prosodic patterns of Bulgarian-English CS speech. Spoken data from fifteen Bulgarian-English bilinguals extracted from a pre-existing Bulgarian-English CS corpus was examined. Syntactically similar CS and unilingual Bulgarian utterances were selected for comparison, with the CS utterances containing either an insertion or an alternation. The sentences were analysed with Praat for f0, speech rate, and VOT. The findings show a lower f0 before the CS and a higher f0 before the unilingual equivalent. The speech rate before the CS is slower, and the VOT appears slightly shorter in CS utterances. These findings confirm the presence of prosodic cues in CS speech, and advance our understanding of the role of prosodic cues in the processing of CS speech.