---
title: "A Part-of-Speech Tagset for Morphosyntactic Tagging of Amharic"
description: "The paper tackles a basic first step for computational work on Amharic: settling on a part-of-speech… 12 references, cited by 2 papers here."
authors: ["Magdalena Krzyżanowska"]
year: 2018
venue: "Aethiopica 20"
source: "https://journals.sub.uni-hamburg.de/aethiopica/article/view/1142"
doi: "10.15460/aethiopica.20.1.1142"
licence: "CC BY-SA (journal licence per DOAJ; Aethiopica)"
licence_url: "https://creativecommons.org/licenses/by-sa/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/krzyzanowska-2018-part-speech-tagset-morphosyntactic-tagging/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/krzyzanowska-2018-part-speech-tagset-morphosyntactic-tagging/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2018](https://md.research.elamur.ai/free/en/papers/2018/page/7/) / A Part-of-Speech Tagset for Morphosyntactic Tagging of Amharic

# A Part-of-Speech Tagset for Morphosyntactic Tagging of Amharic

**Authors:** [Magdalena Krzyżanowska](https://md.research.elamur.ai/free/en/authors/magdalena-krzyzanowska/) · **Year:** 2018 · **Venue:** [Aethiopica 20](https://md.research.elamur.ai/free/en/venues/aethiopica/) ([Aethiopica 2018](https://md.research.elamur.ai/free/en/venues/aethiopica/2018/)) · **Publisher:** [Hamburg University Press](https://md.research.elamur.ai/free/en/publishers/hamburg-university-press/) · **Licence:** [CC BY-SA (journal licence per DOAJ; Aethiopica)](https://creativecommons.org/licenses/by-sa/4.0/)

**Links:** [Publisher's page](https://doi.org/10.15460/aethiopica.20.1.1142) · [PDF](https://journals.sub.uni-hamburg.de/aethiopica/article/download/1142/1064) · [DOI 10.15460/aethiopica.20.1.1142](https://doi.org/10.15460/aethiopica.20.1.1142) · [OpenAlex](https://openalex.org/W2795310470)

**Citations:** 3 citations (OpenAlex, 2026-10-14) · [2](https://md.research.elamur.ai/free/en/papers/krzyzanowska-2018-part-speech-tagset-morphosyntactic-tagging/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
The paper tackles a basic first step for computational work on Amharic: settling on a part-of-speech inventory that human annotators can apply consistently. It was written within the TraCES project, which had already built digital annotation tools and a tagset for Gǝʿǝz. That Gǝʿǝz tagset does not fit Amharic's very different, heavily agglutinative morphosyntax, so the author designs a new scheme. It aims to be detailed enough to be useful for corpus search yet simple enough for manual tagging. The proposal has three parts:
- a transliteration convention that keeps every Ethiopic grapheme distinct;
- a segmentation policy that splits wordforms into tagged morphological pieces, each linked to a lemma;
- 47 tags across 12 POS classes, with grammatical features such as person, number, gender and politeness.

Annotation is planned in GeTa, with search in ANNIS. The author explains the difficult decisions: grammaticalized items, compounds, clitic homophones, and hard-to-classify small words. The scheme is meant as the basis for annotator guidelines and an Amharic corpus usable for both synchronic and diachronic linguistics.

## Key points
- The final tagset has 47 tags in 12 POS groups. Nominals, pronouns, verbs, numerals and quantifiers, adpositions, conjunctions, adverbs, particles and interjections each get their own group, as does a catch-all class called "Bound Grammatical Morphemes" for affixes and clitics.
- An Amharic wordform can hold up to six morphemes, so annotators tokenize it into pieces before tagging. Each piece is linked to a lemma, and Kane's Amharic–English dictionary is suggested as the lexicon base.
- Transliteration keeps homophonous letters apart (ሀ/ሐ/ኀ become h/ḥ/ḫ, and ሠ/ሰ become ś/s), so the corpus can also support studies of spelling practice. The first- and fourth-order vowels are written ä and a to match actual pronunciation.
- The scheme generally ranks form over function, with grammaticalization as the main exception. For example, items of nominal origin that cannot be pluralized and carry no gender are tagged as postpositions, not nouns.
- The paper names two earlier efforts: the news corpus of the Ethiopian Languages Research Center (Addis Ababa University) and the HornMorpho morphological analyser. Compounds are marked "MW", a label borrowed from the British National Corpus.

## Interesting ideas
- Default to form-based tags, but switch to function when there is clear morphosyntactic evidence of a category change, such as losing number and gender inflection. This yields a principled and checkable rule rather than a case-by-case judgment.
- Each constituent of a compound or reduplication is tagged separately at the morphological level, and the multiword status is recorded at a higher level via an MW label. This keeps the low layer simple without losing the unit.
- Homophonous markers are kept apart by function: the same surface suffix gets a derivational "adverbializer" tag in some contexts and is tagged as the article or accusative marker elsewhere.
- Hard-to-segment items are deliberately left unsplit, even at the cost of some inconsistency, because annotation throughput and agreement matter more than theoretical purity there.
- The orthography is preserved faithfully during normalization (distinct graphemes kept despite sharing a sound), so the same corpus can serve unforeseen research questions.

## For Mova

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

- **I1** (tokens): The tagger could use an explicit evidence test, such as whether a word accepts plural or other inflection, to decide when a noun-like word has become a preposition, and log that test as the explanation.
- **I2** (tree): Multiword expressions could stay as separate tokens in the UD tree and be linked by a fixed/compound-style relation plus an MW annotation, so later frames can treat them as one lexical unit.
- **I3** (tokens): English homographs like "-ly", "that" or "'s" could carry separate function tags decided by context, rather than one tag per surface form, which makes downstream reasoning clearer.
- **I4** (learn): When building silver data, low-agreement cases could be collapsed into a coarser, explicitly flagged unit instead of forcing witnesses to split them, reducing noise in training.
- **I5** (tokens): The tokenizer could keep the original spelling, casing and variant forms alongside the normalized lemma, so normalization never destroys information that explanations or generation might need.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Universität Hamburg](https://md.research.elamur.ai/free/en/universities/universitat-hamburg/) — University, Hamburg, Germany · 81 works · 5,549 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works

🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works · [🔤 Part-of-speech tagging](https://md.research.elamur.ai/free/en/topics/part-of-speech-tagging/) · 126 works

## Cite

Cite as: Magdalena Krzyżanowska. 2018. A Part-of-Speech Tagset for Morphosyntactic Tagging of Amharic. Aethiopica 20.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{krzyzanowska2018part,
  title = {{A Part-of-Speech Tagset for Morphosyntactic Tagging of Amharic}},
  author = {Magdalena Krzyżanowska},
  booktitle = {Aethiopica 20},
  year = {2018},
  doi = {10.15460/aethiopica.20.1.1142},
  url = {https://doi.org/10.15460/aethiopica.20.1.1142}
}
```

- [← previous The Role of Diacritics in Increasing the Difficulty of Arabic Lexical Recognition Tests](https://md.research.elamur.ai/free/en/papers/hamed-2018-role-diacritics-increasing-difficulty-arabic/)
- [next → A Study on Dialog Act Recognition using Character-Level Tokenization](https://md.research.elamur.ai/free/en/papers/ribeiro-2018-study-dialog-act-recognition-using/)
- [↑ back to papers of 2018](https://md.research.elamur.ai/free/en/papers/2018/page/7/)

## References

12 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Cotterell, F. P. 1964. ‘Amharic word classes’, Journal of Ethiopian Studies, 2/1 (1964), 33–48. [find](https://scholar.google.com/scholar?q=Cotterell%2C+F%2E+P%2E+1964%2E+%E2%80%98Amharic+word+classes%E2%80%99%2C+Journal+of+Ethiopian+Studies%2C+2%2F1+%281964%29%2C+33%E2%80%9348%2E)
2. Gambäck, B. 2012. ‘Tagging and Verifying an Amharic News Corpus’, in G. De Pauw, G.M. de Schryver, M. L. Forcada, K. Sarasola, F. M. Tyers, and P. W. Wagacha, eds, Proceedings of the Workshop on Language Technology for Normalisation of LessResourced Languages SaLTMiL 8–AfLaT 2012. May 22, 2012, Istanbul Lütfi Kır-dar Convention & Exhibition Centre, Istanbul, Turkey (Istanbul: European Language Resources Association, 2012), 79–84. [find](https://scholar.google.com/scholar?q=Gamb%C3%A4ck%2C+B%2E+2012%2E+%E2%80%98Tagging+and+Verifying+an+Amharic+News+Corpus%E2%80%99%2C+in+G%2E+De+Pauw%2C+G%2EM%2E+de+Schryver%2C+M%2E+L%2E+Forcada%2C+K%2E+Sarasola%2C+F%2E+M%2E+Tyers%2C+and+P%2E+W%2E+Wagacha%2C+eds%2C+Proceedings+of+the+Workshop+on+Langu)
3. Aethiopica 20 (2017) Miscellaneous Gambäck, B., F. Olsson, Atelach Alemu Argaw, and L. Asker 2009. ‘Methods for Amharic PartofSpeech Tagging’, in G. De Pauw, G.M. de Schryver, and L. Levin, eds, EACL 2009. Proceedings of the EACL 2009 Workshop on Language Technologies for African Languages, AfLaT 2009. 31 March 2009, Megaron Athens International Conference Centre, Athens, Greece (Athens: The Association for Computational Linguistics, 2009), 104–111. [find](https://scholar.google.com/scholar?q=Aethiopica+20+%282017%29+Miscellaneous+Gamb%C3%A4ck%2C+B%2E%2C+F%2E+Olsson%2C+Atelach+Alemu+Argaw%2C+and+L%2E+Asker+2009%2E+%E2%80%98Methods+for+Amharic+PartofSpeech+Tagging%E2%80%99%2C+in+G%2E+De+Pauw%2C+G%2EM%2E+de+Schryver%2C+and+L%2E+Levin%2C+eds%2C+EACL+)
4. Gankin, E. B. 1969. Amharsko–Russkij Slovar. አማርኛና፡ መስካብኛ፡ መዝገበ፡ ቃላት። (Amharic–Russian Dictionary. Amarǝññanna mäskobǝñña mäzgäbä qalat) (Moskva: Izdatel’stvo Sovetskaja Entsiklopedija, 1969). [find](https://scholar.google.com/scholar?q=Gankin%2C+E%2E+B%2E+1969%2E+Amharsko%E2%80%93Russkij+Slovar%2E+%E1%8A%A0%E1%88%9B%E1%88%AD%E1%8A%9B%E1%8A%93%E1%8D%A1+%E1%88%98%E1%88%B5%E1%8A%AB%E1%89%A5%E1%8A%9B%E1%8D%A1+%E1%88%98%E1%8B%9D%E1%8C%88%E1%89%A0%E1%8D%A1+%E1%89%83%E1%88%8B%E1%89%B5%E1%8D%A2+%28Amharic%E2%80%93Russian+Dictionary%2E+Amar%C7%9D%C3%B1%C3%B1anna+m%C3%A4skob%C7%9D%C3%B1%C3%B1a+m%C3%A4zg%C3%A4b%C3%A4+qalat%29+%28Moskva%3A+Izdatel%E2%80%99stvo+Sovetskaja+Entsiklopedija%2C+1969%29%2E)
5. Gasser, M. 2011. ‘HornMorpho: a system for morphological processing of Amharic, Oromo, and Tigrinya’, in Organizing Committee, Proceedings of Conference on Human Language Technology for Development. 3–5 May 2011, Bibliotheca Alexandrina, Alexandria, Egypt (online publication, 2011), 94–99, [http://www.cle.org.pk/research/](http://www.cle.org.pk/research/) rep/HLTD2011.pdf, accessed 5 September 2017. — 2012. ‘HornMorpho 2,5 User’s Guide’ (online publication, 2012), http://homes.soic.indiana.edu/gasser/L3/horn2.5.pdf, accessed 5 September 2017. [↗](http://www.cle.org.pk/research/)
6. Girma Awgichew Demeke and Mesfin Getachew 2006. ‘Manual Annotation of Amharic News Items with PartofSpeech Tags and its Challenges’, ELRC Working Papers, 2/1 (2006), 1–17. [find](https://scholar.google.com/scholar?q=Girma+Awgichew+Demeke+and+Mesfin+Getachew+2006%2E+%E2%80%98Manual+Annotation+of+Amharic+News+Items+with+PartofSpeech+Tags+and+its+Challenges%E2%80%99%2C+ELRC+Working+Papers%2C+2%2F1+%282006%29%2C+1%E2%80%9317%2E)
7. Hartmann, J. 1980. Amharische Grammatik, Äthiopistische Forschungen, 3 (Wiesbaden: Franz Steiner Verlag GmbH, 1980). [find](https://scholar.google.com/scholar?q=Hartmann%2C+J%2E+1980%2E+Amharische+Grammatik%2C+%C3%84thiopistische+Forschungen%2C+3+%28Wiesbaden%3A+Franz+Steiner+Verlag+GmbH%2C+1980%29%2E)
8. Hummel, S. and W. Dickhut 2016. ‘A part of speech tag set for Ancient Ethiopic’, in A. Bausi with assistance from E. Sokolinski, eds, 150 Years after Dillmann’s Lexicon: Perspectives and Challenges of Gǝʿǝz Studies, Supplement to Aethiopica, 5 (Wiesbaden: Harrassowitz Verlag, 2016), 17–30. [find](https://scholar.google.com/scholar?q=Hummel%2C+S%2E+and+W%2E+Dickhut+2016%2E+%E2%80%98A+part+of+speech+tag+set+for+Ancient+Ethiopic%E2%80%99%2C+in+A%2E+Bausi+with+assistance+from+E%2E+Sokolinski%2C+eds%2C+150+Years+after+Dillmann%E2%80%99s+Lexicon%3A+Perspectives+and+Challenges+of)
9. Kane, T. L. 1990a. Amharic–English Dictionary, I: ሀ–ነ (Wiesbaden: Otto Harrassowitz, 1990). — 1990b. Amharic–English Dictionary, II: ኘ–ፐ (Wiesbaden: Otto Harrassowitz, 1990). [find](https://scholar.google.com/scholar?q=Kane%2C+T%2E+L%2E+1990a%2E+Amharic%E2%80%93English+Dictionary%2C+I%3A+%E1%88%80%E2%80%93%E1%8A%90+%28Wiesbaden%3A+Otto+Harrassowitz%2C+1990%29%2E+%E2%80%94+1990b%2E+Amharic%E2%80%93English+Dictionary%2C+II%3A+%E1%8A%98%E2%80%93%E1%8D%90+%28Wiesbaden%3A+Otto+Harrassowitz%2C+1990%29%2E)
10. Leslau, W. 1995. Reference Grammar of Amharic (Wiesbaden: Harrassowitz Verlag, 1995). [find](https://scholar.google.com/scholar?q=Leslau%2C+W%2E+1995%2E+Reference+Grammar+of+Amharic+%28Wiesbaden%3A+Harrassowitz+Verlag%2C+1995%29%2E)
11. Vertan, C. 2016. ‘Bringing Gǝʾǝz into the digital era: computational tools for processing Classical Ethiopic’, in A. Bausi with assistance from E. Sokolinski, eds, 150 Years after Dillmann’s Lexicon: Perspectives and Challenges of Gǝʿǝz Studies, Supplement to Aethiopica, 5 (Wiesbaden: Harrassowitz Verlag, 2016), 31–42. [find](https://scholar.google.com/scholar?q=Vertan%2C+C%2E+2016%2E+%E2%80%98Bringing+G%C7%9D%CA%BE%C7%9Dz+into+the+digital+era%3A+computational+tools+for+processing+Classical+Ethiopic%E2%80%99%2C+in+A%2E+Bausi+with+assistance+from+E%2E+Sokolinski%2C+eds%2C+150+Years+after+Dillmann%E2%80%99s+Lexicon%3A+)
12. Wintner, S. 2014. ‘Morphological Processing of Semitic Languages’, in I. Zitouni, ed., Natural Language Processing of Semitic Languages, Theory and Applications of Natural Language Processing (Heidelberg–New York–Dordrecht–London: Springer, 2014), 43– 66. Summary [find](https://scholar.google.com/scholar?q=Wintner%2C+S%2E+2014%2E+%E2%80%98Morphological+Processing+of+Semitic+Languages%E2%80%99%2C+in+I%2E+Zitouni%2C+ed%2E%2C+Natural+Language+Processing+of+Semitic+Languages%2C+Theory+and+Applications+of+Natural+Language+Processing+%28Heidelb)

*Updated 2 November 2026*