Voiceless TH, the sound at the start of words such as 'thin' and 'three', can be something of a shibboleth: some people worry about pronouncing it as [θ], even though there is little evidence that pronouncing it as [t] (as is common in places such as Brunei, Malaysia and Singapore) has much impact on intelligibility. Indeed, plenty of people in Ireland use [t], and many people in England use [f]; so how important is use of [θ] for voiceless TH?
In China, people tend to use [s] for voiceless TH. Does that impact on intelligibility? Do listeners ever mistake 'thing' for 'sing' as a result?
Recently, while in China, I presented my analysis of thirteen five-minute conversations with students in Yangzhou. (For more information about my presentation, see here.) In those thirteen conversations, I found eighteen instances where I did not understand the speakers, and just one case involved voiceless TH. Let us consider that token in some detail. In the following transcript, Int is the Interviewer (me) while F1 is the female speaker from Yangzhou:
Int : er lots of people want to become teachers do they?
F1 : ermi think many girls want to
Int : mm
I initially heard this as 'as many girls want to', so it seems that the use of [s] at the start of 'think' contributed to the misunderstanding. However, let's consider it a bit more.
Here is a spectrogram of 'i think many' from this utterance:
In fact, one can see that 'i think' is pronounced as [aɪs], and there is little evidence of a separate syllable for 'think'. In other words, the use of [s] for voiceless TH was just one factor, and if 'think' had been pronounced as [sɪŋk], it is likely that the misunderstanding would not have occurred.
My conclusion, therefore, is that variant pronouncing of voiceless TH has little impact on intelligibility.
