B07 Sound patterns in Human Language: phonation

1: Basic anatomy of the larynx

The larynx is a complex structure consisting of several cartilages (that's gristle) arranged to form a big ring, and two loose flaps, called the vocal folds which span the ring. At one end of the each fold is a small pyramid-shaped cartilage called the arytenoid cartilage. The other ends are pretty much fixed and held together. The arytenoids can be moved from side to side, or swiveled, so that they pull the vocal folds open or closed, and optionally tighten them up. Pulling the folds together is called adduction, while pulling them apart is abduction. During breathing, airflow should be unimpeded, so the folds are abducted. During voiced sounds, the folds are adducted and air blows through them, causing them to vibrate (just like a little Bronx cheer). You can see pictures of the folds during voicing and during breathing by clicking below, though it may take a moment to load.


The glottis is the vocal folds and the space between them. It is thus the most important part of the larynx for speech production.


View photos of the glottis in action!


2: States of the glottis

A: Glottal closure

We met abduction (open, as during breathing) and adduction (together, as during voicing). When the folds are maximally adducted, no air can escape, and we have glottal closure. Some sounds which involve complete glottal closure are
  1. Glottalic ejectives /p',t',k'/
  2. Glottal stop //
  3. Syllable final voiceless stops, especially /t/

B: Whisper

Earlier in the course we pretended that voicing was like a switch. Either a sound was voiced or it was not. Now it is time to attend to detail. Regular voicing involves a close approximation of the vocal folds, and regular vibration along their complete length (see the photos). We can also whisper. Try singing in a whisper. Can you sing a high note and a low note? The reason there is no pitch is that the vocal folds are not vibrating. Rather, they are slightly adducted, leaving enough of a constriction to cause some frication. Whispering is like adding a /h/ sound to all articulations. Although most languages have some form of /h/, I know of none which use whispering as a distinctive form of phonation.

C: Murmur (breathy voice)

It is possible to combine a whisper with voicing. This is called breathy voice or murmur. This is achieved by bringing the vocal folds together along part of their length (producing voicing) and spreading them for the rest (producing the whisper). Sounds tough, but we do it automatically in the oooohhhh! and aaahhhhhhh! conventionally used at fireworks displays on the 4th of July. Produce an exaggerated oooohhhh!, holding it for as long as you can. You notice that you run out of air rather quickly. This is because much more air flows through the slightly spread glottis than through a glottis in its normal voicing state. Most speakers will produce a murmured sound when /h/ occurs between two vowels, as in aha! or ahoy!.

Any sound which can be voiced, can be produced with murmur. The IPA symbol is two dots underneath the non-murmured symbol. Murmured vowels form minimal pairs with non-murmured ones in some languages, such as Gujarati (India). (see Ladefoged, p. 141)

Just as an otherwise voiced sound can be produced with murmur, so an otherwise whispered (/h/-like) sound can be produced with murmur. This is used in some Indian languages, such as Hindi, to produce a distinctive form of "aspiration" on voiced stops. This allows Hindi to have a four-way "voicing" distinction (remember, English had a 2-way distinction, Thai had a 3-way distinction):
Hindi Voiceless
unaspirated
Voiceless
aspirated
Voiced Breathy voiced
Bilabial /pal/ /pal/ /bal/ / bal/
Dental /tal/ /tal/ /dal/ / dal/
etc. see Ladefoged, p. 145 for glosses and further examples.

D: Creaky voice

In creaky voice, the vocal folds are especially tight, so that they vibrate rather irregularly. Individual flaps are further apart than in normal voicing and they are quite irregular. In extreme cases, you can clearly hear individual taps of the folds. Try imitating a creaky door. Creak is an idiosyncratic feature of some speakers voices, and is often found at the end of an utterance. It is thus not distinctive in English, but it is distinctive in Hausa (Nigeria), where the glide /j/ is either creaky or not creaky (the term usually used is laryngealized). Stops too can be distinctively laryngealized, which means that the part of the vowel immediately following stop release is creaky voiced (Hausa is again an example). In the IPA, creaky or laryngealized sounds are marked by putting a tilde (~) underneath them.
This page has been developed by Fred Cummins, Department of Linguistics, 2016 Sheridan Road, to whom all questions should be referred.
Copyright (1998) Northwestern University

Goto top of page
Goto index