BACK IPS Chapter 8

 
This is a brief tutorial on drawing parametric diagrams for voicing and soft palate position. We will use the English word printed as an example.  Click on the Forward and Back links to navigate through the tutorial.  With thanks to Dynamic Drive for the sequential content script.
First, we need to transcribe the word phonetically.  It would not be a good idea to try to draw a diagram like this using an orthographic representation of the word.  (Why not?).  Then we need to divide the word into segments. Remember that diphthongs and affricates should be treated as single segments and not as sequences. We happen to have neither sound type in our example word which contains seven segments, so we have seven columns plus a space at the beginning and at the end of the word.
We will deal with the soft palate parameter first.  By convention, we assume that the word is pronounced in isolation and that before and after the utterance the speaker is breathing in quietly through the nose, so the soft palate is lowered.  Look at the diagram above,
Next we need to identify all the obstruent sounds in the word.  We have three in our word [p], [t] and [d]. For these the soft palate must be up throughout their duration.
The next step is to identify any nasal sounds in the word.  We have one of these: [n].  The soft palate must be lowered so that air can escape through the nasal cavity.  But we also need to look at what happens after the nasal.  Here we have an obstruent in our word.  Notice that the soft palate moves up during the nasal and not during the obstruent.
Notice how the soft palate begins to lower during the first vowel in anticipation of the following nasal.  This means that the vowel is nasalised.  Compare this with what happens for the second vowel.  This is not next to a nasal so the soft palate remains in the up position.
Now we turn to the voicing parameter. Here again before the utterance and after the utterance we assume that the speaker's glottis is open and the vocal folds are not vibrating.
Our word has two voiceless consonants: [p] and [t].  There are also two segments which are usually voiced, but in our word they have lost most or all of their vocal fold vibration. Lets find out how they got that way.
The second sound of the word is a postalveolar approximant.  Like all English approximants, it is voiced in most environments.  However, when it is immediately preceded by a voiceless plosive,  which in turn is in a position where it would be aspirated,  the approximant is wholly or partially devoiced.  We have chosen to show it with no vocal fold vibration at all, but the final part of the segment may be voiced for some speakers.
The final voiced alveolar plosive is also devoiced because it is not surrounded by voiced sounds.  Voiced (lenis) obstruents are devoiced in many varieties of English if they are adjacent to a voiceless sound or to silence.  Notice that the nasal consonant is not devoiced, even though it is next to a voiceless sound.  Nasals are sonorants, not obstruents.

End of tutorial