Skip to content

Should a Ge'ez compound name written as two words read as one? (መንግስቱ ኃይለ ማርያም gives middle ኃይለ, family ማርያም) #1201

Open

Description

@derek73

Many Ethiopian and Eritrean names are compounds of a bound word and a holy name: ገብረ (servant of), ኃይለ (power of), ወልደ (son of), ተክለ (plant of), ሀብተ (gift of), ወለተ (daughter of), joined to ማርያም, ሥላሴ, ሕይወት, መድህን, ጊዮርጊስ, ሚካኤል, እግዚአብሔር. In Ge'ez they are usually written as two words; in Latin script usually joined (Gebremariam) or abbreviated (G/Mariam). Written as two words, the compound splits into two names (measured on master ccaf6ad):

Input given middle family
መንግስቱ ኃይለ ማርያም መንግስቱ ኃይለ ማርያም
ኃይሌ ገብረ ሥላሴ ኃይሌ ገብረ ሥላሴ
አበበ ገብረ ሕይወት አበበ ገብረ ሕይወት
ገብረ ማርያም ተስፋዬ ገብረ ማርያም ተስፋዬ
ሣህለ ሥላሴ ሣህለ '' ሥላሴ
አበበ ገብረመድህን አበበ '' ገብረመድህን
አበበ ገ/መድህን አበበ '' ገ/መድህን

Measured (2026-11-08):

  • Wikidata labels: of 445 Ethiopians with an Amharic label (Wikidata, citizenship Q115), 26 hold a compound written as two words, and all 26 split. In 14 of them the English label writes the compound as one word (ለማ ገብረ ሕይወት / Lemma Gebrehiwot, ተዘራ ኃይለ ሚካኤል / Tezera Hailemichael).
  • Other corpora: JRC-Names holds 35 Ge'ez spellings; 2 are this shape (the first two rows above). The committed corpora hold no Ge'ez beyond CLDR's 31 fictional renderings.

This is the shape of P12 (a bound given name and the word that completes it): Arabic-script عبد and the closed list of names that complete it. P5 (bound given names) covers the first position.

Proposal, in rule shape

When: a Ge'ez bound word (ገብረ, ኃይለ/ሀይለ, ወልደ, ተክለ, ሀብተ, ወለተ, ፍቅረ, ሣህለ, ኪዳነ, ዘርአ) stands before a word its reading lists as completing it.

Then: the two words are one name unit. Past the first name word, P12 applies (completed_by); at the first name word, P5 applies, with its reserve.

"መንግስቱ ኃይለ ማርያም"       →  family="ኃይለ ማርያም"
"አበበ ገብረ ሕይወት ከበደ"      →  middle="ገብረ ሕይወት"
"ገብረ ማርያም ተስፋዬ"         →  given="ገብረ ማርያም"
"ኃይሌ ከበደ"                →  given="ኃይሌ"  · boundary

Because: Ge'ez spelling settles which form is bound. The bound words end in the construct vowel (ኃይለ, ገብረ), while the given names written alone are spelled differently (ኃይሌ Haile, ገብሬ Gebre). So the bound form is certain from its spelling, where a Latin Haile or Gebre is also a whole given name (Gebre Kebede Tesfaye gives given Gebre, rightly). As P12 decided for Latin abdul, the Latin spellings would list no completions.

Questions

  • Is completed_by closed enough here? The holy names are a short list, but ገብረ also takes saints' names (ገብረ ክርስቶስ, ገብረ ጻድቅ).
  • A two-word name that is all compound (ሣህለ ሥላሴ, ተክለ ሃይማኖት): with no word to spare, P5 keeps the halves apart. Is that the reading wanted?
  • Should Tigrinya (ti.toml) take the same words? Eritrean names use the same compounds, but Wikidata's Eritreans hold only 11 Ge'ez labels.

Metadata

Metadata

Assignees

Labels

non-latin-scriptNames written in a script other than Latin (rules/W)vocabularyFixed by a word-list entry: per-word vocabulary and language packs

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions