📄 Gideon Goldenberg - Ariel Shisha-Halevy (eds.): Egyptian, Semitic and General Grammar

Summary

This is not an NLP paper. It is a short review by Alessandro Bausi, published in Aethiopica 13, of a memorial volume for Hans Jakob Polotsky (1905–1991), who founded an influential school of Egyptian, Semitic and general linguistics in Jerusalem. The volume was edited by Gideon Goldenberg and Ariel Shisha-Halevy and grew out of a 2001 workshop held on the tenth anniversary of Polotsky's death. It contains 25 studies on Egyptian, Coptic, Akkadian, Arabic, Hebrew, syntax and general linguistics.

The reviewer briefly lists the contents and then looks closely at the two Ethiosemitic papers:

• Rainer Voigt on Southern Tigrinya dialects: their phonology and personal pronouns.
• Goldenberg on the first attempts to write the Gurage language Kəstanəñña.

Bausi praises the volume for continuing Polotsky's rigorous, source-based approach. For a computational library, its value lies in the methodological ideas it contains about dialect variation, grammaticalization and the effect of writing on language. It offers no resources or metrics.

Key points

• The book was published in Jerusalem by the Israel Academy of Sciences and Humanities in 2009. It has xix + 501 pages, costs US$75 and contains 25 contributions.
• Topics range across Egyptian classifier script, Coptic morphophonology, Akkadian nexus focusing, Arabic dialects, Maltese prepositions, Hebrew clefts and passive participles, and the "emotive genitive."
• Voigt sets out a basic split in Tigrinya:
  • Northern dialects (Aksum/Adwa to Asmara) are the basis of the literary standard.
  • Southern (southeastern Tigray) dialects spirantize velars and labiovelars, with morphonological consequences.
• Voigt traces the Southern pronoun nəssu (compare Amharic ərsu) to a Ge'ez noun meaning "belly." He places this within a wider pattern of pronouns built from nouns, and he suggests Amharic may have grown out of a Southern Tigrinya dialect.
• Goldenberg analyses Kəstanəñña (also called Aymälläl or Soddo) texts collected in Wärqu Täsfa's 1995 Amharic book. He asks what writing does to a language that was previously only spoken.

Interesting ideas

• I1. Pronouns can grow out of ordinary nouns such as "belly," "head" or "soul," so a function word may carry a traceable lexical history worth recording.
• I2. A dialect boundary can be described by a few regular sound changes, each with predictable morphological side effects, rather than by a long list of word differences.
• I3. Writing down an unwritten language forces explicit choices about phonology and morphology, and those choices then reshape how the language is perceived.
• I4. Polotsky's school insisted on examining primary sources first-hand before theorizing.
• I5. The editors openly flag one contribution in their own volume as controversial, which keeps the collection honest.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

• I1 (world): The ontology could link grammaticalized items like English "-self" forms to their noun sources, so explanations of reflexives can cite where the form came from.
• I2 (tree): UD dialect conversion could be written as a small set of regular, explainable transformation rules, each with its knock-on effects stated.
• I3 (tokens): Tokenizer and lemma conventions should be documented as analytical decisions, not neutral facts, because they shape every later layer.
• I4 (learn): Silver data and repair rules should be regularly audited against raw sentences, not judged only by aggregate scores from witness parsers.
• I5 (learn): Evaluation reports could explicitly mark weak or disputed components alongside the strong results, much as negative controls are kept visible.