Posted: Apr 09, 2012 9:17 pm
by Zwaarddijk


Not only are the sounds we make affected by where in the mouth the closure is, there's a number of manners of articulation that are relevant. While introducing the nasal-oral passages as a tube, I already presented the stop. The stop is a full closure, behind which some pressure is allowed to build up and then released. The stop part of it is actually fairly silent, what gives away what sound it is in that case generally is an effect on the surrounding sounds' frequency distributions just prior to and after the stop. Depending on where along this tube it occurs, it'll affect the neighbouring sounds in subtly different ways, which we parse as different actual sounds.

There's a number of things we can do with stops. We can voice them - this normally distinguishes /b/ and /p/, /d/ and /t/, /g/ and /k/ (although the situation in English is a bit more complex than just that, and in some languages, the distinction actually isn't so much voicing as it's voicing onset). To voice them, there's a need to keep the air stream going through the vocal cords, which normally happens even when they're not voiced, but in order to keep the phonation going, the required air stream is a bit stronger than in the unvoiced case. This leads to the result that in languages that distinguish length on stops, it's not unusual for long g not to be present. The pressure builds up too fast, as the space for it to spread out in is rather small. Length is of course a thing some languages distinguish as well. Estonian has a remarkable length distinction system, in that it distinguishes three lengths on its stops. (Orthographically, short consonants are marked by what corresponds to English voiced stops, middle ones by a single voiceless stop, and long ones by doubled voiceless stops).

If you divert the air into the nasal cavity instead, you get a nasal stop. English distinguishes three of them, corresponding to each pair of stops: /m/ (labial), /n/ (alveolar), /ŋ/ (velar). Some languages have nasal release of the stop pressure: initially, it is just like a stop, then some of the pressure is diverted through the nose. Generally, nasals are voiced, but this is not always the case. In many varieties of Swedish, the suffix -ism gets its final m devoiced. I have seen claims of languages that distinguish voicing on nasal stops, but I have not seen them verified - I suspect there's faulty analysis going on behind those conclusions.

Another important group of sounds that almost all languages on the earth have (the only exceptions being almost all Australian aboriginal languages) are fricatives. In a fricative, there is only a slight obstruction, but sufficient enough to produce turbulence in the airflow. English has the following fricatives: /f/, /v/, /θ/, /ð/, /s/ , /z/, /ʃ/, /ʒ/, /h/. At all stages, there's a voiced-voiceless distinction, except at h. The symbols used here are the IPA symbols encoding fricatives (unvoiced and voiced) at the labiodental, dental, alveolar, postalveolar and glottal. /h/ can be a bit exceptional as far as this goes, in English it apparently often is but fonation - e.g. the vocal cords causing a vibration and some turbulence and no other thing whatsoever.

If you slow down the release of a stop, the release will sound a bit like a fricative, and this gives one rather common kind of stop as well: the affricate. English has two (distinguished by voicing), /t͡ʃ/ and /d͡ʒ/. German also has /p͡f/. Other possible manners of release exist.

There's yet another step obtainable by just reducing the amount of obstruction so that no turbulence appears, viz. the approximant. The most basic approximants in English are some varieties of /r/, (consonantal) y, transliterated in IPA as /j/, and most varieties of w. (Most dialects don't even have non-approximant variations of w).

There appear some varieties on these as well. Lateral approximants are what you get if there's closure at the top, but a passage on the side of the tongue. This gives l-like sounds. Some languages (notably Welsh, but also any number of languages in other parts of the world) have lateral fricatives. Stops can have lateral releases as well, but these are unusual in the languages of Europe.

Another kind of articulation is the trill, a vibration between the active and passive articulator. These can't happen everywhere in the mouth - e.g. labiodental ones are afaict unattested. Bilabial ones, alveolar/postalveolar/etc ones, and uvular ones at the very least are possible. Most English-speaking people would recognize at least the uvular and coronal trills as variations of r, and probably associate them with some foreign accent (French, East European, ...)

A tap is basically a single iteration of the vibration a trill consists of. In English, intervocalic d and t are often rendered as taps.

Next installment will be secondary articulations and some stuff on non-pulmonic airstreams, and I hope, an introduction to the ideas of phones vs. phonemes.