Airstream mechanism

Pulmonic initiation

Initiation by means of the lungs (actually the diaphragm and ribs) is called pulmonic initiation. The vast majority of sounds used in human languages are pulmonic egressives. In most languages, including all the languages of Europe (excluding the Caucasus), all phonemes are pulmonic egressives.

The only attested use of a phonemic pulmonic ingressive is a lateral fricative in Damin, a ritual language formerly used by speakers of Lardil in Australia. This can be written with the extended version of the International Phonetic Alphabet as . !Xóõ has ingression as a phonetic detail in one series of its clicks, which are ingressive voiceless nasals with delayed aspiration, . Peter Ladefoged considers these to be among the most difficult sounds in the world. Other languages, for example in Taiwan, have been claimed to have pulmonic ingressives, but these claims have either proven to be spurious or to be occasional phonetic detail.

In interjections, but not in normal words, pulmonic ingressive vowels or words occur on all continents. This is commonly done for back-channeling (as with in Ewe) or affirmation (as with in Swedish). In English, an audible intake of breath, , or an indrawn consonant such as or is used in a conversation to indicate that someone is about to speak or is preparing to continue speaking. In some languages, such as Finnish and Amharic, entire phrases may be uttered with an ingressive airstream. (See ingressive sound.)

Glottalic initiation

It is possible to initiate airflow in the upper vocal tract by means of the vocal cords or glottis. This is known as glottalic initiation.

For egressive glottalic initiation, one lowers the glottis (as if to sing a low note), closes it as for a glottal stop, and then raises it, building up pressure in the oral cavity and upper trachea. Glottalic egressives are called ejectives. The glottis must be fully closed to form glottalic egressives, or the air column would flow backwards over it; it is therefore impossible to pronounce voiced ejectives. Ejective allophones of voiceless stops occur in many varieties of English at the ends of intonation units.

For ingressive glottalic initiation, the sequence of actions performed in glottalic pressure initiation is reversed: one raises the glottis (as if to sing a high note), closes it, and then lowers it to create suction in the upper trachea and oral cavity. Glottalic ingressives are called implosives, although they may involve zero airflow rather than actual inflow. Because the air column would flow forwards over the descending glottis, it is not necessary to fully close it, and implosives may be voiced; indeed, voiceless implosives are exceedingly rare.

It is usual for implosives to be voiced. Instead of keeping the glottis tightly closed, it is tensed but left slightly open to allow a thin stream of air through. Unlike pulmonic voiced sounds, in which a stream of air passes through a usually-fixed glottis, in voiced implosives a mobile glottis passes over a nearly motionless air column to cause vibration of the vocal cords. Phonations that are more open than modal voice, such as breathy voice, are not conducive to glottalic sounds because in these the glottis is held relatively open, allowing air to readily flow through and preventing a significant pressure difference from building up behind the articulator.

Because the oral cavity is so much smaller than the lungs, vowels and approximants cannot be pronounced with glottalic initiation. So-called glottalized vowels and other sonorants use the more common pulmonic egressive airstream mechanism.

There is no clear divide between pulmonic and glottalic sounds. Some languages may have consonants which are intermediate. For example, glottalized consonants in London English, such as the t in rat , may be weakly ejective. Similarly, fully voiced stops in languages such as Thai, Zulu, and Maidu are weakly implosive. This ambiguity does not occur with the next airstream mechanism, lingual, which is clearly distinct from pulmonic sounds.

Lingual (velaric) initiation

The third form of initiation in human language is lingual or velaric initiation, where a sound is produced by a closure at two places of articulation, and the airstream is formed by movement of the body of the tongue. Lingual stops are more commonly known as clicks, and are almost universally ingressive. The word lingual is derived from Latin lingua, which means tongue.

To produce a lingual ingressive airstream, first close the vocal tract at two places: at the back of the tongue, as in a velar or uvular stop, and simultaneously with the front of the tongue or the lips, as in a coronal or bilabial stop. These holds may be voiceless, voiced, or nasalized. Then lower the body of the tongue to rarefy the air above it. The closure at the front of the tongue is opened first, as the click "release"; then the closure at the back is released for the pulmonic or glottalic click "accompaniment" or "efflux". This may be aspirated, affricated, or even ejective. Even when not ejective, it is not uncommon for the glottis to be closed as well, for a triply articulated consonant, and this third closure is released last to produce a glottalized click. Clicks are found in very few languages, notably the Khoisan languages of southern Africa and some nearby tongues such as Zulu. They are more often found in extra-linguistic contexts, such as the "tsk tsk" sound many Westerners use to express regret or pity (a dental click), or the clucking noise used by many equestrians to urge on their horses (a lateral click).

Lingual egressive initiation is performed by reversing the sequence of a lingual ingressive: the front and back of the tongue (or lips and back of the tongue) seal off the vocal cavity, and the cheeks and middle of the tongue move inward and upward to increase oral pressure. The only attested use of a lingual egressive is a bilabial egressive "spurt" in Damin. Transcribing this also requires the use of the Extended IPA, .

Since the air pocket used to initiate lingual consonants is so small, it is not thought to be possible to produce lingual fricatives, vowels, or other sounds which require continuous airflow.

Clicks may be voiced, but they are more easily nasalized. This may be because the vocal cavity behind the rearmost closure, behind which the air passing through the glottis for voicing must be contained, is so small that clicks cannot be voiced for long. Allowing the airstream to pass through the nose enables a longer production.

Nasal clicks involve a combination of lingual and pulmonic mechanisms. The velum is lowered so as to direct pulmonic airflow through the nasal cavity during the lingual initiation. This nasal airflow may itself be egressive or ingressive, independently of the lingual initiation of the click. Nasal clicks may be voiced, but are very commonly unvoiced and even aspirated, which is rare for purely pulmonic nasals.

Airstream contours

In some treatments, complex clicks are posited to have airstream contours, in which the airstream changes between the front (click) and rear (non-click) release. There are two attested types: Linguo-pulmonic consonants, where the rear release is a uvular obstruent such as or ; and linguo-glottalic consonants, where the rear release is an ejective such as or .

Not only are simultaneous (rather than contour) implosive clicks possible, i.e. velar (e.g. ), uvular (), and de facto front-closed palatal (), but velar implosive clicks are easier to produce than modally voiced clicks. However, they are not attested in any language.

Percussive consonants

Consonants may be pronounced without any airstream mechanism. These are a subset of percussive consonants, where sound is generated by one speech organ striking another. Purely percussive consonants – that is, those without any airstream mechanism – are not phonemic in any known language, though the extensions to the IPA for disordered speech provide symbols for a bilabial percussive (a smack of the lips), a bidental percussive (a gnashing or clashing of the teeth), and a sublaminal (sublingual) lower-alveolar percussive (a tongue slap). The only percussive confirmed to be used in normal speech is a tongue slap that appears allophonically in the release of alveolar clicks in the Sandawe language of Tanzania. (See Alveolar click § Percussive release and Palatal click § Percussive release.) There is also a speech register in Pirahã in which the tongue is released from a lateral flap to make a sublaminal strike of the lower lip . (See labiolingual consonant.)

Tap and flap consonants in general are described as percussive sounds, but in an articulatory sense; the airstream mechanism is still pulmonic because the sound is generated by a (percussive) interruption of a pulmonic airflow. The abrupt releases of other pulmonic consonants have also been described as percussive, for example the audible velar release often heard in a velar nasal before a vowel, .