How should the Sumerian ergative marker be read?

Question

In Foxvog's Sumerian grammar, he assumes that the ergative marker -e was pronounced as -e, even after the possessives -(a)ni and -bi. For example, he transcribes "her king" in the ergative as lugal-(a)ni-e > lugal-a-né, and "its king" in the ergative as lugal-bi-e > lugal-bé.

However, the cuneiform signs né and bé are more commonly read as ni and bi. And distinct, commonly-used signs for ne and be certainly existed in the syllabary. The fact that the Sumerians chose to use the NI and BI signs instead of NE and BE seems like a reasonably solid indication, to me, that the actual pronunciation was /ni/ and /bi/. And this is backed up by looking at how it combines with other possessives: lugal-ŋu-e "my king" > lugal-ŋu₁₀, with no /e/ in sight.

So, why does Foxvog read lugal-a-né and lugal-bé instead of lugal-a-ni and lugal-bi? I assume he knows what he's talking about and has a reason for writing it this way, but his grammar doesn't really present any rationale for this decision.

Answer 1

Foxvog has a note on this on page 31:

Related to this phenomenon is the matter discussed by Thomsen in §107. As the Table of Syllabic Sign Values (p. 157) shows, the NI and BI signs can also be read né and bé; the choice of values is entirely up to the reader of a text. Though it is not universal practice, in this grammar the presence of a presumed case marker -e after the possessive suffixes -ni and -bi will consistently be indicated by the transliterations -né and -bé. Thus {lugal+(a)ni+Ø} 'his king (subject)' should be transliterated as lugal-a-ni, but {lugal+(a)ni+e} 'by his king' (ergative) as lugal-a-né.

Chasing the pointer back to Thomsen's grammar:

The postposition /-e/ usually disappears after the vowel of the possessive suffixes: /-ani-e/ > -a-ni. If there is a reason to assume the presence of the postposition the possessive suffix is often transliterated as -a-né (or -bé). However, we cannot know whether a pronunciation [ane] actually was opposed to normal [ani] and the transliteration -a-né or -bé is thus merely an aid for the translation. In the present study the suffixes have always been written -a-ni and -bi also in cases where a loc. term. or ergative element probably is present.

Foxvog doesn't offer any further justification for this choice, so I chalk it up to personal authorial style.