Repeat this several times, intro-
specting about what it feels like. Contrast this dorso-velar slop [k] with a dorso-palatal stop [c]. Note how the body of the tongue is thrust well forward for [c], so that contact can be made with the highest part of the hard palate. For [k], however, the body of the tongue, though clearly further back than for [c]. doesn't feel particularly strongly drawn back.
Once you are satisfied you can feel the midvelar (mid-soft palate) contact for ]k] hold the tongue in that |k]-position and, while taking care not to shift the tongue either forward or back, open up a very small channel between the tongue and the soft palate. If you propel an cgressivc air-stream (i.e. blow) through that narrow dorso-velar channel you should hear the sound of the vo/ce/fs; W-
Make a prolonged |x x x x x . . .]. then do it again, switching on vmcf, but making no other change: |x x x y y y . . .], where [y] is the symbol for a voiced dorso-velar fricative.
Starting from the voiced velar fricative |y] develop a velar approximant [mj. Say a prolonged |y y y . . .], noting that it is truly fricative, i.e. there is a fricative hiss-noise superimposed upon the smooth sound of voice. Now while saying prolonged [y y y . . •], very slowly, and very slightly, open up the articulatory channel, just to the point where the hiss-noise of turbulent airflow ceases: [y y y m mj. This is the velar approximant , [u/J.
Now that you are clear about dorso-ve/ar (or simply ve/ar, as we often call it) articulation you can note that there is also a velar nasal [nj. This is the nasal that occurs at the end of /w/ig [lArj] in English. Note that [rj] never occurs at the beginning of a syllable in English. However, if you isolate |rj] you will find it quite easy to put a vowel after it, and say [rja] [ni], etc.
In English, and many other languages (it is very noticeable in French, for example), [k] and [g] are purely velar only before such vowels as [a] and |u], as in English car, gw?/Y/, and mo/, goo. Before front, or palatal, vowels Like [i] as in &ey, gea$e , however, the articulation of [k] |g] is shifted forward a little. If you articulate these words silently, then isolate the [k]/(g] of each and compare the ones that occur before [i] with the others, the forward shift of tongue-position before [i] will be obvious. But note that even
98
Articulation: Locations
Articulation: Locations
99
though the tongue is somewhat advanced in key and geese, it does not go nearly as far forward as the palatal position of [c] and [jj. The English velar stops are somewhat 'fronted' or 'palatalized' before |i]: but they do not become palatals.
The next, and the last, articulatory location in the oral area is (/o/"jo-«M//o/", or simply HW/f/r. We investigate sounds made at this location in Experiment 64.
64 Make a |kj-closure and then, silently, or almost silently, make a prolonged series of faint [k|-type sounds [k^ k^ k'' .. .|, etc. while slowly sliding the tongue back and down as far as you can. You will end up making a stop at the very furthest back part of the soft palate. The extreme hack of (he tongue is in contact with (he uvula and the extreme hack of the velum (soft palate). If you let a little pressure huild up behind this extreme back closure, then release the closure, you will hear a mv(/w .wo/j |qj. If you repeat this experiment-a series of faint stops of the |k'' k^ k''.. .| type steadily moving back from the t'e/w to the mv;/M/- positions-you will observe that the sound of the little burst of noise occurring on the release of each stop goes down in pitch by about an octave over the whole range.
Another observation you may make is that (he release of velar |k] is relatively 'clean', while tha( of |q| is more 'sloppy'. This is because the convex tongue-surface can break away from the whole contact-area of (he concave velar surface almost instantaneously but separation from the more flexible and irregular surface of the extreme back of the velum, including the uvula, is less instantaneous, less clean-cut.