Research session 5 · Proto-cuneiform: the archaic texts of Uruk
Officials' names: the foreign-workforce explanation finds no support
Split by the header each name sits under, archaic names outside the worker lists, including 61 under title headers, carry the Sumerian name elements no more than the workers' names (odds ratios 1.27–1.40), far below every Sumerian sample (4.5 and up).
No group of archaic names, by context, approaches the Sumerian name profile: the title-header names (61) reach an odds ratio of 1.38, with an upper bound of 2.34 against 4.5 and up for every Sumerian sample.
The foreign-workforce explanation is not supported; the live alternatives are non-Sumerian name-givers and a labelling notation, which an element-reuse test against Early Dynastic names can separate.
What this changes: Central question: one more piece of evidence against H-SUMERIAN, of suggestive weight; the remaining contest for the names is between H-UNKNOWN or H-OTHER (non-Sumerian name-givers) and H-NOTATION (labels).
Compare how archaic names reuse their elements (type-token growth, share of elements in two or more distinct names, recurring element pairs) with Early Dynastic Sumerian and Semitic names at matched sample sizes. Linguistic names reuse a small vocabulary productively; ad hoc labels should not.
Next steps live on their questions, where anyone can take one up.
What was done
Design. Proto-cuneiform officials are normally named by title, and the editors' name marking is confined to 41 personnel tablets [@englund1998texts] [@englund2009smell]. So instead of a separate officials corpus, the marked names were split by the header each one sits under, fixed before scoring. Worker-category headers (female, male, child, age marks) give 245 names. Other headers (work units, commodity recipients) give 269. Headers carrying a title or office sign (SANGA, EN, PA, GAL, NAMESZDA and others) give 61.
Result. The odds ratios are 1.40 (0.88–1.77) for workers, 1.27 (0.93–1.65) for other headers and 1.38 (0.60–2.34) for title headers. Early Dynastic Sumerian names give 4.5–5.8, and Early Dynastic dependents 4.7.
::: figure 1
Reading. The foreign-workforce explanation predicts that names outside the worker categories should look more Sumerian. They don't. Two readings of the archaic names remain: whoever gave these names was not drawing on the Sumerian name stock, or the entries are labels rather than spelled names (Damerow's notation view) [@damerow2006origins]. The test is limited because the groups overlap in archive and the 'titled' names are people in officials' accounts, not demonstrably the officials themselves.
Element odds ratio (Englund's list without EN) per context group, with a tablet-cluster bootstrap; compared with the Early Dynastic benchmark
Uruk III administrative non-name entries · Early Dynastic Sumerian benchmark from RUN0DFVQR5WN96F and RUN0DG2CWRH4M9P
No context group approaches the Sumerian range. The upper 95% bounds are 1.65–2.34 for the groups of useful size, against 4.5 and up for Sumerian names. The foreign-workforce explanation finds no support among the names available, but the groups come from the same personnel tablets and do not include a clean sample of officials.