In his Akkadian grammar (specifically the appendix on phonology), Huehnergard lists the following Proto-Semitic consonants:

consonant table

Most of this looks familiar to me. However, * caught me by surprise; I'm not used to an emphatic velar fricative being included. (Semitic is very much not my specialty, though, so I'm not very well-read on it; it's entirely possible that it's a mainstream thing that I just haven't come across before.)

Is this part of the mainstream consensus? And what is the evidence for it?

2 Answers 2


One page further (p. 587), Huehnergard gives as one of the changes from Proto-Semitic to Old Babylonian:

Common Semitic * and * merged to (Huehnergard 2003):
     *ḫamisum > ḫamšum ‘five’; *saḫānum > šaḫānum ‘to be warm’;
     *x̣apārum > ḫepērum ‘to dig’; *rax̣āṣ́um > raḫāṣ́um ‘to wash’.

The reference is to the author's 'Akkadian and West Semitic *'; pp. 102–119 in Leonid Kogan, ed., Studia Semitica; Orientalia: Papers of the Oriental Institute, 3 (Alexander Militarev volume).

The background is that Akkadian lost (preserved in West Semitic, whence the traditional / distinction), and that corresponds to in West Semitic, while there are exceptions to this (where corresponds to in West Semitic). The numbers are as follows (p. 112):

  • Akkadian Ø corresponds to West Semitic * in about 60 examples; there are in addition some 50 examples of Akkadian Ø with a > e, i.e., in which the earlier presence of either Proto-Semitic * or Proto-Semitic ʕ is likely.

  • Akkadian corresponds to West Semitic * in about 50 examples.

  • Akkadian corresponds to West Semitic in about 90 examples; there are also nearly 200 additional Akkadian roots with for which no West Semitic cognates are known.

The traditional solution is that Proto-Semitic stays unaffected, whereas Proto-Semitic * is dropped in East Semitic. However, the second group, where West Semitic * corresponds to East Semitic , is too large for this. In default of a phonological explanation that can predict which instances of * become Ø and which become , we must posit a third consonant for this second group (Huehnergard compares this to the situation of ð merging with d in Aramaic but z in Hebrew; cf. Aramaic dkr vs. Hebrew zkr 'remember').

Because this hypothetical consonant merged with and it should be a fricative articulated near the velum or pharynx. Most consonants in Proto-Semitic come in a triadic opposition (voiced/voiceless/emphatic), the exceptions being bilabials and velar and pharyngeal fricatives (none of which have emphatics). Because a emphatic pharyngeal fricative is 'unexpected on articulatory grounds' (p. 115), an emphatic velar fricative is proposed, i.e., IPA [x'].

We then get (p. 116):

  • Proto-Semitic is unaffected in West Semitic but dropped in East Semitic.
  • Proto-Semitic is unaffected throughout.
  • Proto-Semitic becomes in West Semitic and in East Semitic.

Is this the mainstream consensus? I can't really comment since I don't work in phonology, but I can say that (1) you are right to be surprised (2003 is pretty recent) and (2) Huehnergard is usually right.

  • The other possibility is that at least some of the words with ḥ in West Semitic and ḫ in East Semitic are in fact loanwords in the latter.
    – fdb
    Nov 15, 2021 at 16:39

Today´s consensus is that the emphatic sounds were realized as glottalized (as, e.g., in Akkadian), not as pharyngalized (cf., e.g., Arabic). Therefore, one for a root, no more (in Akkadian, known as the Geers´ Law). The problem with Huehnergard´s theory (as I see it) is that it reconstructs/creates quite a lot of roots with two emphatic radicals in his reconstruction of, say, Protoakkadian. OK, perhaps we can postulate that it was a retention from another stage (between Protosemitic and Protoakkadian?; H. self was playing with many additional stages in his reconstruction of the history of Aramaic), but it violates the language economy. I do not see an elegant way out of this "Huehnergards´ paradox". Perhaps we should expect another secondary articulation for this velar fricative?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.