📄 Gäbrä Iyyäsus Kifle, ታሪኽ ትውልዲ ልሳን ግእዝ Tariḵ tǝwlǝddi lǝssan Gǝʿǝz [History of the origin of the Gǝʿǝz language – in Tigrinya]

Summary

This is a short book review, not an NLP paper. Olga Kapeliuk assesses a Tigrinya-language history of Gǝʿǝz by Abba Gäbrä Iyyäsus Kifle, a monk and teacher. He studied Gǝʿǝz in the traditional church schools for about twenty years, then at the Hebrew University in Jerusalem. He wrote the book because he was frustrated that Gǝʿǝz was studied in foreign universities but not in its homeland.

He finished an Amharic version in the early 1970s, but the Derg-era censor refused to approve it. The author blames this on his not praising the regime. Kapeliuk suggests a more likely reason: officials preferred to place Ethiopian languages in an African, mainly Cushitic, frame rather than a Semitic one. After Eritrean independence the author moved to Asmara and translated the book into Tigrinya. It is probably the first history of Gǝʿǝz in that language.

The book places Gǝʿǝz in the Semitic family using comparative tables. It also traces the Ethiopic script and ends with teaching material. The review praises its scope but questions some of its linguistic claims and its treatment of Amharic.

Key points

• Publication details: Asmara, Asär Press, 2001 (1993 E.C.), 165 pages. The review appears in Aethiopica 6 (2003), pp. 263–265.

• Semitic survey (pp. 16–102): The book surveys Aramaic, Hebrew, Arabic, Akkadian, Ugaritic and Epigraphic South Arabian. It gives lexical comparison tables in Ethiopic script and a pronoun/imperative table on p. 99. It also documents how the Ethiopic script developed from the ESA alphabet.

• Descent of Tigre and Tigrinya: The book treats both as direct descendants of Gǝʿǝz.

• Neighbouring languages (pp. 103–112): The chapter covers Argobba, Gafat, Gurage and Harari but leaves out Amharic. The author justifies this by Amharic's many non-Semitic words and cites Aešcoly.

• Teaching material:

  • a Gǝʿǝz–Tigrinya word list (pp. 114–131);
  • thematic vocabularies;
  • a basic grammar;
  • 18 conversation exercises on everyday secular topics, to be translated in both directions.

  The exercises reflect the author's view that Gǝʿǝz could be revived as Hebrew was.

Interesting ideas

• I1. The author coins a suffix -it for language names and calls it a feminine marker. Kapeliuk shows it actually comes from a participle ending plus -t, matches Amharic and Hebrew forms, and should not be confused with endings that sound the same.
• I2. Kapeliuk separates the author's stated cause of the censorship from the more probable cause she infers from the political context.
• I3. The comparative tables set related languages side by side and explicitly mark both shared and diverging items.
• I4. The conversation exercises are meant to be translated both ways, Gǝʿǝz→Tigrinya and Tigrinya→Gǝʿǝz, which tests whether meaning survives in each direction.
• I5. Kapeliuk notices that Amharic is missing from the chapter on neighbouring languages and asks why, treating the gap itself as evidence.

For Mova

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

• I1 (tokens): Tag identical English surface endings (e.g. -s, -ing) by their real morphological source rather than their shape, and record which analysis was chosen.
• I2 (frames): Keep a speaker's attributed explanation and the narrator's or critic's alternative explanation as distinct, sourced claims, rather than merging them into one "cause" slot.
• I3 (world): Cross-lexicon alignments in the compiled knowledge, such as WordNet–FrameNet–Wikidata links, could likewise store explicit divergence entries, not only matches.
• I4 (back): Evaluate the text → graph → text compressor with paired round-trips in both directions, and flag cases where one direction loses roles or entities.
• I5 (learn): Evaluation reports should check which categories a training set or rule set silently leaves out, not only how it performs on what it covers.