Friday, 3 October 2014

consonant cluster reduction #2

Paul Carley inspired me to write another blog entry on consonant clusters (the previous one is to be found here). How does one pronounce the plural form of month - /mʌnθs, mʌntθs, mʌnts, mʌnʔs, mʌns/ - or what?
Listen to this extract from the morning news of the 2nd of October on BBC 4 (Paul, our truffel-finder, spotted it; the speaker is the International Development Correspondent to the BBC, Mark Doyle):
If the disease [Ebola] continues at current rates a million people across the West African region could die within months."
 Here is the whole sentence plus repetitions of months:


How does Mr Doyle pronounce months?Which of the above-mentioned variants does he use. The choice is actually between /mʌnts, mʌnʔs, mʌns/.
Let's look at the waveform first:

/mʌ/ ~ 100 ms
/n/ ~ 74 ms
pause ~ 58 ms
1st aperiodic low volume signal section ~ 39 ms
2nd aperiodic higher volume signal section ~ 160 ms.
The pause section seems to represent the hold stage of an epenthetic /t/ in /mʌnts/.

Could the speaker also have said /mʌns/? In that case the waveform would have looked something like this

and sounded like this:

Not much of a difference, is it?


  1. There's a well known difficulty pronouncing -ns-, or nasal+s generally: the velum has to be closed for [s] in order for sufficient pressure to build up, so there's a slight delay between nasal vibrations and sibilant hiss. Is what you're seeing in the first waveform such a delay, or a [t] occlusion?

    How silent is the environment here (a radio studio should be silent, but TV news is usually read in a large studio with several other people present, cooled cameras, cooled computers etc). That possible occlusion in the first waveform has plenty of low-level activity - is that background noise, intruding noise added by the soundfile's adventures since initial recording (compression?), or is it feeble friction for the [θ]? The answer to any of these questions so far could take you in different directions for different explanations.

    A possible scenario. The velum closes early before the "absolute end" of /n/ to seal the nasopharyngeal port for /s/, and the glottis opens early before the end of /n/ to permit the massive airflow for /s/; /θ/ is elided so there's no activity for it (your title already assumes it's cluster reduction); the /n/ occlusion is still in place, pressure builds up for /s/; the tongue blade shape is changed from /n/ to /s/ and the channel aimed at the central incisors, by now there's sufficient airflow for /s/ turbulance. What you've said is /ns/, but it might be perceived as [nts].

    1. Sidney,
      as you describe in extenso, it could be one or the other due to a host of factors. I know I'm walking on slippery ground.

    2. There's more to it of course. This is one of the old riddles about sound change - why do we hear [p] in -ms- or [t] in -ns-, and why do infants sometimes re-interpret them as /mps/ etc., sometimes not?