📄 Verbless Relative Clauses in Gǝʿǝz and their Equivalents in Amharic and Tigrinya

Summary

Kapeliuk examines a less-studied corner of Ethio-Semitic syntax: relative clauses in Gǝʿǝz that have no verb, where the relative pronoun is linked to a nominal predicate through a zero copula. She compares them with their renderings in Amharic and Tigrinya. Her data are mostly Gǝʿǝz Bible passages set against modern translations, plus some contemporary Amharic and Tigrinya prose. She argues that these clauses are not one-member fragments. Each is a complete nominal sentence whose subject is the relative pronoun standing in for the head noun.

She sorts them into relational, equational and qualifying types, according to whether the predicate is a prepositional phrase or adverb, a noun, or an adjective. She also traces how some of them became lexicalized and how they overlap with the genitive marker. In the modern languages, the missing predication is filled in by a relative verb of existence or a relative copula.

The study helps explain why Ethio-Semitic leans so heavily on relativization. The author tentatively links this tendency to a Cushitic substrate, probably Agaw.

Key points

• Frequency counts: 30 pages of Haddis Alämayyähu's Fǝqǝr ǝskä mäqabǝr contain 368 relative verbs, about 12.27 per page. The Picture of Dorian Gray has about 4.8 per page.
• Growth over time: Citing Cotterell (1973), Amharic relative verbs rose from roughly 30 to almost 100 per thousand words between 1880 and the 1970s.
• Three types of verbless relatives:
  • relational (preposition or adverb predicate);
  • equational (noun predicate);
  • qualifying (adjective or passive participle predicate).
• Modern equivalents: Amharic and Tigrinya use the existential verb for relational clauses (Amh. y-allä, Ta. z-ällo). They use the relative copula for equational and qualifying clauses (Amh. yä-honä, Ta. zǝ-ḵonä).
• Lexicalization: Relativized prepositional phrases in Gǝʿǝz produced new lexical items, such as zä-bä-ʾaman ("true/truly") and zä-lä-ʿaläm ("eternal"). The modern languages stopped coining such forms, and only Amharic keeps the genitive use of the relative marker. (The PDF header gives Aethiopica 12, 2009, pp. 143–154.)

Interesting ideas

• I1. A reduced modifier like "animals under the tree" is analysed as a full predication with a silent copula, with the relative element as its subject.
• I2. The relational / equational / qualifying split classifies a predicate by its type: location or time, identity, or property.
• I3. Gǝʿǝz relativizes a prepositional phrase partly to signal that it modifies a noun, not the verb, which removes an attachment ambiguity.
• I4. The author doubts that the pleonastic "which is + adjective" pattern is emphatic, because some writers use it so routinely that it carries no extra force.
• I5. Some relative-plus-preposition sequences froze into single adjectives or adverbs, so their meaning is no longer built compositionally.

For Mova

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

• I1 (frames): Mova could turn noun-attached PPs and reduced relatives into explicit state frames (located-at, is-a, has-property) with the head noun as the subject, rather than leaving them as bare modifiers.
• I2 (world): This gives a ready schema for storing facts from copular and verbless descriptions as three distinct state kinds in the world of the text.
• I3 (tree): Paraphrasing a PP as "which is + PP" versus attaching it to the verb could serve as a consistency probe when witness parsers disagree on PP attachment.
• I4 (frames): Pragmatics rules should not treat a marked construction as emphasis or implicature by default; they should first check how often the author uses it.
• I5 (tokens): A lexicon of frozen multiword relatives or phrases, tagged as one adjective or adverb unit, would keep downstream frames from over-analysing fixed expressions.