Flora Yu


Chapter 7  Glottalic and Laryngealized Consonants

From7.4 7.10


7.4  Voiceless laryngealized segments

1.          Usually, voiceless laryngealized segments are ejectives.

2.          Three languages have voiceless laryngealized segments: Korean, Ashuslay and Siona.


/p¼/, /t¼/, /k¼/, /t¼ /, /s¼/


  /p¼/, /t¼/, /k¼¼/, /t¼ /




  /s¼/ voiceless laryngealized fricatives

3.          These segments, in Korean, are usually called “fortis” or “tense” obstruents. They are produced with a narrow glottal aperture.


7.5  Implosives and voiced laryngealized plosives

1.          Implosives are the only kind of glottalic ingressive segments.

2.          Except the two segments /p / and /t / in Igbo116, all the implosives are voiced.

3.          In UPSID, 32 languages are implosives. 10 languages are voiced laryngealized plosive. Total is 13.2% in UPSID.

4.          These two are always mentioned together. We use /?b/ to represent both /b/ and /b/. Greenberg used the term “injective” to cover both implosives and voiced plosives.

5.          In UPSID, there are two languages have /?d/ alone; one languages has /?d/ alone. This shows the hierarchical relationship of injectives is:

(1)      The presence of /?b/ implies the presence of /?d/ or no other implosives.

(2)    The presence of /?d/ implies the presence of /?b/ or no other implosives.

(3)    The presence of /?j/ implies the presence of both /?b/ and /?d/.

(4)    The presence of /?g/ implies the presence of /?b/, /?d/, and /?j/.

6.          Most common places of these two segments are bilabial and dental/alveolar. And velar is disfavored for voiced implosives and voiced plosives, e.g., Doayo128and Yulu216.


7.          The number of injectives vs. the languages

The number of injectives42

Languages or examples



/?b/: kpelle, Igbo, Zulu, lakkia, K’ekchi


/?d/: Berta, Kullo

/?d/: Somali


Usually are bilabial and dental/alveolar; velar is disfavored, e.g. Doayo.




/?b/, /?d/, /?j/: Kadugli, Angas, Ngizim


/?b/, /?d/, /?g/: Harmer more special


/?b/, /?d/, /?j/, /?g/: Nyangi, Swahili, Maasai


/?b/, /?d/, /?j/, /?G/uvular:IK unsual

The nine languages with more than two injectives are from Africa. Therefore, an important area trend is that the languages using palatal and velar places for segment are Africa language.

8.          About / j /

(1)    Greenberg: Languages often have / j / in place of /?j /.

(2)    But Hamer265in UPSID does not support his suggestion.

(3)    In UPSID, most languages28 of 33,with a series of injectives but no palatal series, lack / j /. Most languages8 of13which have / j / lack injectives.

(4)    No language in the survey has both /?j / and / j /, which might be evidence for the suppletion of /?j / by / j /.

9.          Retraction of dental/ alveolar implosives

Greenberg and Haudricourt: An implosive, which corresponds to a non-implosive dental/ alveolar, is often retroflexed or articulated further back than the non-implosives. E.g., Somali258, Tama210and Yulu216. However, the data in UPSID are not sufficient to support this.

10.      Voiceless implosives

Overwhelmingly, implosives are voiced. But there are some exceptions, e.g., Igbo116: /p / and /t /

K’ekchi714: uvular ejective /q’/ has an allophone /q /.Pinkerton 1980


7.6  Languages with both ejective stops and implosives

1.          Thirteen languages in UPSID contain both some ejectives and implosives. The large number suggests that these two classes of segments have a tendency to occur together in a language.

2.          Although there is the articulation preference for ejectives and implosives, it is still possible that they occur at the same place of articulation. E.g.,

/p’/ & /?b/: Zulu, Koma, Maidu, Otomi

/t’/ & /?d/: Koma, Kullo, Maidu, Otomi, Mazahua, Southern nambiquara.

/k’/ & /g/: Hamer


7.7  Laryngealized sonorants

1.          Greenberg: Laryngealized sonorants are as counterparts in some way to the glottalic and laryngealized consonants.

2.          All segments of laryngealized sonorantsnasals, liquids, approximants and vowelsare voiced.

3.          In general, languages which have laryngealized sonorants must have glottalic stops. The exception: Tiddim Chin513))

4.          In laryngealized sonorants, laryngealized nasals voiced approximants are more frequent than laryngealized liquids.Table 7.9, P.116

5.          In most cases13 of 17 languages, /m¼0/ and /n0/ occur together, just as /?b/ and /?d/ occur together.

6.          The laryngealized approximants /j0/ and /w0/ usually occur together12 of 16 languages.

7.          There are 5 languages with laryngealized trill, tap or flap /r0/. Wapishana822has a laryngealized voiced retroflex fricative /z0/.


7.8  Diachronic implications

1.          The diachronic hypothesis from Greenberg:

(1)    At least one source of injectives might be a sound shift from voiced plain to voiced implosive stops.

(2)    Loss or addition of implosives should follow the place of articulation preference hierarchy.

(3)    Implosive at the same place of articulation as voiced plosive should be rather rare.

2.          Refutation:

(1)    In Swahili124, it appears that the source of implosives may be from voiced plosives. But it is still not clear because Swahili has merged voiced plosives and implosivesGuthrie 1967-70.

(2)    In Proto-Bantu language, Stewart1972has found implosives, in addition to voiced plosives.

3.          The failure to confirm Greenberg’s prediction does not completely refute the diachronic hypothesis, since the languages that do not conform to the prediction may also have been a shift in another stop series, or the original voiced plosive series may have split into plosive and implosive set.

7.9  Phonetic explanation for the structure of glottalic systems

1.          For ejectives, velar place of articulation is more common than a dental/ alveolar one, and both places are preferred to a bilabial one.

2.          For implosives, bilabial and dental/ alveolar places of articulation are equally common, and both are preferred to velar.

3.          Javkin1977explained implosive and ejective distributions by Boyle’s Law, correcting a misinterpretation by Greeberg.

(1)    Greenberg: Back articulations confer an advantage in compressing air in the supraglottal chamber, and front articulations confer an advantage in rarefying air.

(2)    Javkin: It takes the same effort to produce either compression or rarefaction in a chamber of a given size. What matters is the proportional change in the size of the chamber.P.118

4.          Why does implosive do not share the preference for back articulation? The voicing of implosives.

(1)    The absolute volume expanded by oral cavity must be greater than the volume of pulmonic air, or there will be no rarefaction, hence no implosion.

(2)    A chamber created by a back oral closure may not permit expansion by the required absolute volume, whereas one further forward may allow greater absolute expansion through adjustments of tongue position and oral cavity walls.

(3)    Lindau1982: The achievement of an even or rising amplitude of voicing throughout the closure may be a major part of the “target” in production of an implosives.

(4)    The further forward the oral closure is formed, the less the intraoral air pressure increase for a given volume of transglottal airflow.

5.          Why are the patterns for voiced laryngealized plosives are similar to those for true implosives?

7.10  Summary of generalization

The recapitulation of the most important observations about glottalic and laryngealized consonants.P.120