[whythismatters]

Valid point, while cognatic kinship can be clearly established through afro-asiatic linguistics, that doesn't say anything about actual pronunciation. My comment wasn't meant as an absolute statement, but merely reflects the scientific consensus on the matter, i.e. how an assyriologist would read those characters in a transliterated text out loud. With cuneiform languages we're talking millennia, so of course there has been considerable variety (space) and change (time) in pronunciation, some of it reflected in writing, cf. Assyrian vs Babylonian dialects. I could imagine that ṣ pronounced as "ts" may have been inferred/validated for specific timeframes through variant "spelling", or maybe indirectly via e.g. Hittite or Old Persian, but that's just a wild guess.

[schoen]

Thanks for the explanation! Yeah, I guess it's a similar or at least analogous situation for Proto-Indo-European, where there might be conventional ways to pronounce reconstructed forms, and some evidence about what is more and less likely, but people don't really know. There are even some reconstructed consonants where they actively have no definite consensus at all:

https://en.wikipedia.org/wiki/Laryngeal_theory#Pronunciation

I guess it's quite an imperfect analogy, because we have Akkadian writing (by people who actually spoke the language) and we have no equivalent at all for PIE.