Logo Lanfrica

unimorph/amh

Domaine:

natural language processing

Type de record:

dataset
Créateur:
uni
Hôte:
Amharic # Amharic (amh) ## Contents - `amh`: inflection tables for Amharic (ISO 639-3 `amh`) - `amh.args`: a UniMorph 4.0 compatible verion of `amh` - `amh.gloss`: English glosses for the lemmas in `amh` - `README.md`: this file ## Data All data were generated using the Amharic morphological generator in HornMorpho, described in Gasser (2013). Decisions about lemma frequency and subcategorization were based on Biniam Gebremichael's Amharic word list. Common orthographic variants of lemmas and common variant plural forms of nouns are included. In these cases, the variants are distinguished with the features `LGSPEC1` and `LGSPEC2`. Predictable orthographic variants are not included. These include alternate representations of labialized consonants (ቆ/ቈ, ቁ/ቍ, ፏ/ፍዋ, etc.), the two ways of writing the vowel ɨ/i before y (e.g., ጅያ/ጂያ), and the common replacement of ው with ዉ. The total number of lemmas is 2,465: `V` 670, `N` 1599, `ADJ` 196. ### Verbs Although the formation of the passive-reflexive, causative, reciprocal, and frequentative from bare roots is relatively productive, many roots fail to appear in the basic, underived form, in which case the morphologically passive-reflexive or causative form is not semantically or syntactically passive-reflexive or causative (for example, "passive-reflexive" ተቀመጠ 'sit', "causative" አለቀሰ 'cry'). For this reason, each of these derived forms of a root is treated as a separate lemma, and only those appearing in the word list have been included. For example, based on the root 'break', there are the following lemmas in the repository: ሰበረ 'break', ተሰበረ 'be broken', ሰባበረ 'break repeatedly'. Other possible forms derived from this root are not included. Verb lemmas appear in all of the major tense-aspect categories and the infinitive: * perfective (UM: `PFV`) * finite (UM: `IPFV`) and non-finite imperfective (UM: `NFIN;IPFV`) * gerundive (UM: `V.CVB`) * present perfect (UM: `PRF`) * jussive-imperative (UM: `IMP`) * infinitive (UM: …

Languages