This article needs additional citations for verification. (January 2010) (Learn how and when to remove this template message)
|Manners of articulation|
|IPA chart non-pulmonic consonants|
|This table contains phonetic symbols, which may not display correctly in some browsers. [Help]|
|help • help (audio) • image • template|
In phonetics, ejective consonants are usually voiceless consonants that are pronounced with a glottalic egressive airstream. In the phonology of a particular language, ejectives may contrast with aspirated, voiced and tenuis consonants. Some languages have glottalized sonorants with creaky voice that pattern with ejectives phonologically, and other languages have ejectives that pattern with implosives, which has led to phonologists positing a phonological class of glottalic consonants, which includes ejectives.
In producing an ejective, the stylohyoid muscle and digastric muscle contract, causing the hyoid bone and the connected glottis to raise, and the forward articulation (at the velum in the case of [kʼ]) is held, raising air pressure greatly in the mouth so when the oral articulators separate, there is a dramatic burst of air. The Adam's apple may be seen moving when the sound is pronounced. In the languages in which they are more obvious, ejectives are often described as sounding like “spat” consonants, but ejectives are often quite weak. In some contexts and in some languages, they are easy to mistake for tenuis or even voiced stops. These weakly ejective articulations are sometimes called intermediates in older American linguistic literature and are notated with different phonetic symbols: ⟨C!⟩ = strongly ejective, ⟨Cʼ⟩ = weakly ejective. Strong and weak ejectives have not been found to be contrastive in any natural language.
In strict, technical terms, ejectives are glottalic egressive consonants. The most common ejective is [kʼ] even if it is more difficult to produce than other ejectives like [tʼ] or [pʼ] because the auditory distinction between [kʼ] and [k] is greater than with other ejectives and voiceless consonants of the same place of articulation. In proportion to the frequency of uvular consonants, [qʼ] is even more common, as would be expected from the very small oral cavity used to pronounce a voiceless uvular stop. [pʼ], on the other hand, is quite rare. That is the opposite pattern to what is found in the implosive consonants, in which the bilabial is common and the velar is rare.
Ejective fricatives are rare for presumably the same reason: with the air escaping from the mouth while the pressure is being raised, like inflating a leaky bicycle tire, it is harder to distinguish the resulting sound as salient as a [kʼ].
Ejectives occur in about 20% of the world's languages. Ejectives that phonemically contrast with pulmonic consonants occur in about 15% of languages around the world. They are extremely common in northwestern North America and frequently occur throughout the western parts of both North and South America. They are also common in East Africa and Southern Africa. In Eurasia, the Caucasus forms an island of ejective languages. Elsewhere, they are rare.
Language families that distinguish ejective consonants include:
According to the glottalic theory, the Proto-Indo-European language had a series of ejectives (or, in some versions, implosives), but no extant Indo-European language has retained them[a]. Ejectives are found today in Ossetian only because of influence of the nearby Northeast Caucasian and/or Kartvelian language families.
It had once been predicted that ejectives and implosives would not be found in the same language but both have been found phonemically at several points of articulation in Nilo-Saharan languages (Gumuz, Me'en, and T'wampa), Mayan language (Yucatec), Salishan (Lushootseed), and the Oto-Manguean Mazahua. Nguni languages, such as Zulu have an implosive b alongside a series of allophonically ejective stops. Dahalo of Kenya, has ejectives, implosives, and click consonants.
Almost all ejective consonants in the world's languages are stops or affricates, and all ejective consonants are obstruents. [kʼ] is the most common ejective, and [qʼ] is common among languages with uvulars, [tʼ] less so, and [pʼ] is uncommon. Among affricates, [tsʼ], [tʃʼ], [tɬʼ] are all quite common, and [kxʼ] and [ʈʂʼ] are not unusual ([kxʼ] is particularly common among the Khoisan languages, where it is the ejective equivalent of /k/).
A few languages have ejective fricatives. In some dialects of Hausa, the standard affricate [tsʼ] is a fricative [sʼ]; Ubykh (Northwest Caucasian, now extinct) had an ejective lateral fricative [ɬʼ]; and the related Kabardian also has ejective labiodental and alveolopalatal fricatives, [fʼ], [ʃʼ], and [ɬʼ]. Tlingit is an extreme case, with ejective alveolar, lateral, velar, and uvular fricatives, [sʼ], [ɬʼ], [xʼ], [xʷʼ], [χʼ], [χʷʼ]; it may be the only language with the last type. Upper Necaxa Totonac is unusual and perhaps unique in that it has ejective fricatives (alveolar, lateral, and postalveolar [sʼ], [ʃʼ], [ɬʼ]) but lacks any ejective stop or affricate (Beck 2006). Other languages with ejective fricatives are Yuchi, which some sources analyze as having [ɸʼ], [sʼ], [ʃʼ], and [ɬʼ] (but not the analysis of the Wikipedia article), Keres dialects, with [sʼ], [ʂʼ] and [ɕʼ], and Lakota, with [sʼ], [ʃʼ], and [xʼ] . Amharic is interpreted by many as having an ejective fricative [sʼ], at least historically, but it has been also analyzed as now being a sociolinguistic variant (Takkele Taddese 1992).
Because the complete closing of the glottis required to form an ejective makes voicing impossible, the allophonic voicing of ejective phonemes causes them to lose their glottalization; this occurs in Blin (modal voice) and Kabardian (creaky voice). A similar historical sound change also occurred in Veinakh and Lezgic in the Caucasus, it and has been postulated by the glottalic theory for Indo-European. Some Khoisan languages have voiced ejective stops and voiced ejective clicks; however, they actually contain mixed voicing, and the ejective release is voiceless.
Ejective trills are rare, if they exist as distinct sounds at all. An ejective [rʼ] would necessarily be voiceless, but the vibration of the trill, combined with a lack of the intense voiceless airflow of [r̥], gives an impression like that of voicing. Similarly, ejective nasals such as [mʼ, nʼ, ŋʼ] (also necessarily voiceless) are possible. (An apostrophe is commonly seen with r, l and nasals, but that is Americanist phonetic notation for a glottalized consonant and does not indicate an ejective.)
Other ejective sonorants are not known to occur. When sonorants are transcribed with an apostrophe in the literature as if they were ejective, they actually involve a different airstream mechanism: they are glottalized consonants and vowels whose glottalization interrupts an otherwise normal voiced pulmonic airstream, somewhat like English uh-uh (either vocalic or nasal) pronounced as a single sound.
In the International Phonetic Alphabet, ejectives are indicated with a "modifier letter apostrophe" ⟨ʼ⟩, as in this article. A reversed apostrophe is sometimes used to represent light aspiration, as in Armenian linguistics ⟨p‘ t‘ k‘⟩; this usage is obsolete in the IPA. In other transcription traditions, the apostrophe represents palatalization: ⟨pʼ⟩ = IPA ⟨pʲ⟩. In some Americanist traditions, an apostrophe indicates weak ejection and an exclamation mark strong ejection: ⟨k̓ , k!⟩. In the IPA, the distinction might be written ⟨kʼ, kʼʼ⟩, but it seems that no language distinguishes degrees of ejection. Transcriptions of the Caucasian languages often utilize combining dots above or below a letter to indicate an ejective.
In alphabets using the Latin script, an IPA-like apostrophe for ejective consonants is common. However, there are other conventions. In Hausa, the hooked letter ƙ is used for /kʼ/. In Zulu and Xhosa, whose ejection is variable between speakers, plain consonant letters are used: p t k ts tsh kr for /pʼ tʼ kʼ tsʼ tʃʼ kxʼ/. In some conventions for Haida and Hadza, double letters are used: tt kk qq ttl tts for /tʼ kʼ qʼ tɬʼ tsʼ/ (Haida) and zz jj dl gg for /tsʼ tʃʼ cʼ kxʼ/ (Hadza).