Tenuis alveolar linguo-pulmonic stop | |
---|---|
ǃ͡q |
Voiced dental linguo-pulmonic stop | |
---|---|
ǀ͡ɢ |
Aspirated lateral linguo-pulmonic stop | |
---|---|
ǁ͡qʰ |
Breathy-voiced palatal linguo-pulmonic stop | |
---|---|
ǂ͡ɢʱ |
Voiceless labial linguo-pulmonic affricate | |
---|---|
ʘ͡χ |
Voiced retroflex linguo-pulmonic affricate | |
---|---|
‼͡ʁ |
Pulmonic-contour clicks, also called sequential linguo-pulmonic consonants, are consonants that transition from a click to an ordinary pulmonic sound, or more precisely, have an audible delay between the front and rear release of the click. All click types (alveolar ǃ, dental ǀ, lateral ǁ, palatal ǂ, retroflex ‼, and labial ʘ) have linguo-pulmonic variants, which occur as both stops and affricates, and are attested in four phonations: tenuis, voiced, aspirated, and murmured (breathy voiced). At least a voiceless linguo-pulmonic affricate is attested from all Khoisan languages of southern Africa (the Khoe, Tuu, and Kx'a language families), as well as (reportedly) from the Bantu language Yeyi from the same area, but they are unattested elsewhere.
Traditionally, contour clicks were believed to be uvular in their rear articulation, whereas non-contour clicks were thought to be velar. However, it is now known that all clicks are uvular, at least in the languages that have been investigated, and that the articulation of these clicks is more complex than that of others but no different in location. Miller (2011) analyzes them as contours (that is, as a transition from one kind of sound to another within a single consonant), whereas Nakagawa (2006) analyzes them as sequences of a click followed by a uvular consonant (that is, as consonant clusters). The benefit of a cluster analysis is that it greatly reduces the consonant inventory of the language. Taa, for example, has 164 known consonants, including 111 (and potentially 115) clicks, an extraordinary number considering that the largest inventory of any language without clicks, that of Ubykh, is 80 (84 consonants including loanwords). With a cluster analysis, the number of clicks in Taa is reduced to 43, and the total number of consonants to 87, only slightly surpassing Ubykh for the most consonants in the world. There are, however, some disadvantages to a cluster analysis: Although the click series and non-click series can often be made to align, in some languages there are consonants in these purported clusters that never occur alone, something that never happens with other kinds of consonants. Also, all other languages in the world that allow obstruent clusters (as English does with s and t in steep, and as these click clusters would be) also allow clusters with sonorants (as English does with r in treat). However, no Khoisan language allows a cluster of any consonant, click or otherwise, with sonorants like l, r, y or w. Miller concludes that the remarkably large numbers of consonants in these languages is real, a consequence of the greater number of permutations of clicks, where there are two places of articulation that can be independently manipulated.