📄 December 13–14, 2007 in Mainz: Workshop on “Language contact in Ethiopia: Examples from Cushitic, Omotic and Semitic languages”

Summary

This item is a short conference report, not a research paper, so it contains no data, experiments or measured results. It describes a two-day international workshop held in Mainz on 13–14 December 2007 about how contact between Ethiopian languages changes them. The languages came from three families: Cushitic, Omotic and Semitic. Joachim Crass and Ronny Meyer of the collaborative research center SFB 295 organized it. More than 20 researchers from six countries attended.

The opening talks set up the idea of convergence zones, with special attention to the Gurage/Highland East Cushitic (HEC) area. Later talks looked at:

• whether that area's features also appear in Omotic languages,
• lexical borrowing,
• verbal derivation across families,
• contact effects in single Ethiopian languages.

The report concludes that contact phenomena matter a great deal for describing and understanding individual Ethiopian languages. Its value is as a map of who was working on which language-contact questions at that time.

Key points

• The workshop took place on 13–14 December 2007 in Mainz and was organized by Joachim Crass and Ronny Meyer (SFB 295 "Cultural and Linguistic contacts"). Its more than 20 participants represented six countries: Germany, Ethiopia, Switzerland, Italy, Norway and the Netherlands.
• The central framing was the Gurage/Highland East Cushitic (HEC) convergence area. Walter Bisang and Joachim Crass gave the opening remarks. Silvia Zaugg-Coretti and Christian Rapold then traced HEC-area features in Omotic.
• The Ethiosemitic talks covered:
  • the origins of Amharic (Girma A. Demeke),
  • Argobba (Andreas Wetter),
  • Chaha vs. Gumär (Sascha Völlmin),
  • quotative-verb constructions in Gurage (Ronny Meyer).
• The Cushitic and Omotic talks covered:
  • Libido morpho-syntax (Crass),
  • Kambaata lexicalization (Yvonne Treis),
  • Oromo grammaticalization (Debela Goshu),
  • Lowland East Cushitic feature spread (Ongaye Oda),
  • Yemsa–Oromo contact (Silvia Zaugg-Coretti),
  • copula and focus marker in East Ometo (Binyam Sisay),
  • causatives in Shekkacco and Kafa (Tolemariam Fufa).
• The extracted text lists the source as Aethiopica 11 (2008), pp. 303–304, not 2012 as stated in the request. The extracted text also includes the end of an unrelated German-language conference report by Ulrich Braukämper.

Interesting ideas

• I1. Languages that are not related can still come to share grammar when spoken in the same area. This means some patterns are best explained by geographic contact rather than by family tree.
• I2. The workshop kept borrowing words apart from borrowing grammar (derivation, causatives, focus marking) and studied them as separate kinds of change.
• I3. Meyer treated quotative-verb constructions as a family of related structures built on one "say" verb.
• I4. Sisay's talk examined how a copula relates to a focus marker, showing that one grammatical element can play both structural and information-highlighting roles.
• I5. The program was arranged first by contact area and then by language family, which lets one phenomenon be compared from two angles.

For Mova

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

• I1 (tree): When converting between UD dialects or treebank styles, treat shared conventions as area-like influences, and check which differences come from annotation contact rather than from the language itself.
• I2 (tokens): Handle loanwords and foreign morphology as their own category in the lemmatizer, rather than forcing them into native inflection rules, and keep a record of why each word was assigned its lemma.
• I3 (frames): Group English reported-speech and "say"-based patterns under one speech-act frame with several surface variants, so quotations of every shape produce the same explainable who-said-what record.
• I4 (frames): Mark English cleft and copular-focus constructions ("it was X who…") as focus signals in pragmatics, instead of reading them only as plain identity statements.
• I5 (learn): Split evaluation sets both by text genre and by source treebank, so that errors caused by shared conventions can be separated from errors caused by the genre itself.