Why does Hebrew transcribe Akkadian š inconsistently?

Question

Biblical Hebrew consistently uses the letter ס (s) to transcribe names with the Akkadian consonant š. For example, Esarhaddon for Aššur-aḥa-iddina, Esther from Ištar, Sargon from Šarru-ukīn (all Akkadian transcriptions copied from Wikipedia). Etymologically, Akkadian š and Hebrew š almost always correspond (š and ṯ, although Akkadian š can also correspond to Hebrew ś, but not to s to my knowledge).

The only exception I can think of is Aššur which is transcribed that way (with š). Since the Aramaic reflex is Attur, the word seems to have reached both languages as a descendant (through *Aṯṯur) and not a borrowing.

On the other hand, Shalmaneser comes from Šulmanu-ašarid, and yet preserves the š at the beginning of the word, while the second š is transcribed as s.

My question is: Why does Hebrew not transcribe Akkadian š with Hebrew ש (š) instead of ס (s)? Does this indicate that Akkadian š was actually pronounced s (maybe only in Assyria)? And why was the š preserved in the name of Shalmaneser (only one of the two times)?

Answer 1

While Akkadian š is generally cognate with Hebrew š or ś, there's good reason to believe its pronunciation was quite different! The reason it's transcribed as š is mostly historical—Akkadian was first deciphered by comparison to other Semitic languages, so when a certain phoneme seemed to correspond regularly to Hebrew š, they named it š.

But there's evidence that, at least in some dialects, this š was pronounced as /s/. When the Hittites used Akkadian cuneiform to write their own language, they used the š signs for /s/ and left the s signs unused. (We know the Hittite phoneme was pronounced /s/, not /ʃ/, because of transcriptions of Hittite names into Egyptian: Ḫattušiliš becomes Egyptian xtsl, Muršiliš becomes Egyptian mrsl, Šuppiluliumaš becomes Egyptian spll. Egyptian had both /s/ and /ʃ/, so the consistent use of one transcription over the other is meaningful.)

EDIT: Further evidence comes from the tablet Tell el Amarna 1921 1154, which transcribes some Egyptian words in Akkadian cuneiform; Egyptian s is definitely rendered with Akkadian š, and Egyptian š may be rendered with Akkadian s (this half of the equation is less clear).

Huehnergard and Woods (in the Cambridge Encyclopedia of the World's Ancient Languages) explain it thus. Akkadian originally inherited four sibilants from Proto-Semitic: š s z ṣ /s t͡s d͡z t͡sʼ/. Later, s became a simple fricative, but this development happened separately in different dialects: š s /ʃ s/ in Babylonian but š s /s ʃ/ in Assyrian. Huehnergard had previously suggested that [s] and [ʃ] were realizations of a single phoneme, which would explain Šalmanesar, but the dialect-variation theory seems to have supplanted this.

Answer 2

Yes, some people think Akkadian š was pronounced [s].

For the sibilants, traditionally /š/ has been held to be postalveolar [ʃ], and /s/, /z/, /ṣ/ analyzed as fricatives; but attested assimilations in Akkadian suggest otherwise. For example, when the possessive suffix -šu is added to the root awat ('word'), it is written awassu ('his word')

(https://en.wikipedia.org/wiki/Akkadian#Consonants)

There is also relevant discussion in the Wikipedia article on Proto-Semitic:

The "maximal affricate" position additionally posits that *s *z were actually affricates [t͡s d͡z] while *š was actually a simple fricative [s] [...] According to Kogan, the affricate interpretation of Akkadian s z ṣ is generally accepted

(https://en.wikipedia.org/wiki/Proto-Semitic_language#Fricatives)

I am not familar with any of the literature, so I can't expand beyond what Wikipedia says.