There was a story on NPR’s Morning Edition this morning about a song that is making at least some people in Iraq happier than they might otherwise be. A somewhat older Iraqi woman was asked about her opinion (in English) of the song. Of particular interest to readers of this blog is the woman’s pronunciation of the word “street”.
Here’s the first instance of “street” (actually, “the street”), and here’s a waveform and spectrogram for the more visually-oriented (sorry, I don’t have much time or technical expertise to label it):
These sound (and look) something like [sɪtɪɾi:t] to me (putting aside the heavy aspiration at the end of the first instance). What I found interesting were the vocalic elements interrupting the segments of the cluster. The one between the [s] and [t] is certainly longer and more prominent, as indicated by the lack of superscripting in my transcription (and it may be somewhat lower and/or more back than suggested by [ɪ] in the first instance). The one between the [t] and the [ɾ] is the same svarabhakti previously discussed here and here.
So now I’m curious about the analysis of these vocalic elements. What I know from Lisa Davidson‘s work leads me to at the very least doubt that the longer vocalic element between [s] and [t] is a straightforward case of phonological epenthesis. But why is it longer and more prominent than the vocalic element between [t] and [ɾ]?