Old Norse as a Problem of Automatic PoS-Tagging

About the author:

Raul D. Karimov, Postgraduate Student, Department of Romance and Germanic Languages and Cross-Cultural Communication, Chelyabinsk State University; raoul.karimov@hotmail.com; ORCID: 0000-0003-0313-0309

Abstract:

This article dwells upon automatic PoS-tagging of Old Norse by computational means, including machine learning. It analyzes the available language material in diachrony from the standpoint of how language evolution might have affected the quality of automatic PoS-tagging. This article further describes the phonetic traits that have assumingly led to any classification errors.

The research material is an Old Norwegian educational text titled Konungs skuggsjá, or “King’s Mirror”, vectorized by the moving average method and then used to train an Ada-Boosted random forest model. The resulting classification accuracy is about 97%. However, being non-contextual, this vectorization method enables no complete differentiation of morphologically similar parts of speech: verbs, nouns, adjectives, and adverbs. This becomes evident when digging into the identified high-weight classification features, each being a vectoral dimension corresponding to a specific alphabet character; another indicative factor comprises Morfessor-identified high-rank morphs, analyzing which reveals the morphogrammatic units that cause the most classification errors.

Historical consideration of these morphs shows that their collision is due to them being inherited from Proto-Germanic (PG) while undergoing rhotacism, or conversion from PG /z/ to ON /r/. However, the same process effectively prevents the collision of rhotacized finite verbal forms with the genitive case that inherits the PG suffix -s.

The key finding is that such morphological collision being unavoidable, character-based vectorization might not suffice when using a small training set or when trying to classify not only by parts of speech, but also by specific forms in the paradigm.

References:

1. Arapov M. A., Hertz, M. M. 1973. Mathematical Methods in Historical Linguistics. Moscow: Nauka. [In Russian]

2. Nikolayeva N. A. 2003. “Thematization of the present tense of strong verbs in celtic and germanic languages”. Cand. Sci. (Philol.) diss. Moscow: Moscow State University. [In Russian]

3. Bandle O. 2002. The Nordic languages: an international handbook of the history of the North Germanic languages. Edited by O. Bandle, K. Braunmüller, E. H. Jahr, A. Karker, H. P. Naumann, U. Telemann, L. Elmevik, and G. Wildmark. Berlin: De Gruyter Mouton.

4. Gade K. E. 1986. “Homosexuality and rape of males in Old Norse Law and literature”. Scandinavian Studies, vol. 58, no 2, pp. 124-141.

5. Hagland J. R. 1978. “A note on Old Norwegian vowel harmony”. Nordic Journal of Linguistics, vol. 1, pp. 141-147.

6. Haugen O. E. 1994. Norrøne tekster i utval. Oslo: Ad Notam Gyldendal. [In Norwegian]

7. Haugen O. E. 1995. Grunnbok i norrønt språk. Oslo: Ad Notam Gyldendal. [In Norwegian]

8. Jahr E. H., Lorentz O. (eds.). 1993. Historisk språkvitenskap. Oslo: Novus. [In Norwegian]

9. Karttunen L. 2000. “Applications of Finite-State Transducers In Natural Language Processing”. Proceedings of the 5^th International Conference “Implementation and Application of Automata”, CIAA 2000 (24-25 July), pp. 34-46.

10. Kytö M. 1996. Manual to the Diachronic Part of the Helsinki Corpus of English Texts. Helsinki: University of Helsinki.

11. Loftsson H., Kramarczyk I., Helgadóttir S., Rögnvaldsson E. I. 2009. “Improving the PoS tagging accuracy of Icelandic text”. Proceedings of the 17^th Nordic Conference of Computational Linguistics (NODALIDA 2009), pp. 103-110. Odense, Denmark: Northern European Association for Language Technology (NEALT).

12. Medieval Nordic Text Archive. Accessed 11 May 2019. http://www.clarino.uib.no/menota

13. Silva A. P., Silva A., Rodrigues I. 2015. “An approach to the POS tagging problem using genetic algorithms”. In: Computational Intelligence, pp. 3-17. Berlin: Springer.

14. Smit P., Virpioja S., Grönroos S.A., Kurimo M. 2014. “Morfessor 2.0: toolkit for statistical morphological segmentation”. Proceedings of the Demonstrations at the 14^th Conference of the European Chapter of the Association for Computational Linguistics, pp. 21-24. Stroudsburg, PA, USA: Association for Computational Linguistics.

15. Spurkland T. 1989. Innføring i norrønt språk. Oslo: Universitetsforlaget. [In Norwegian]

16. Stroh-Wollin U. 2016. “The emergence of definiteness marking in Scandinavian — new answers to old questions”. Arkiv för nordisk filologi. 2016, no 131, pp. 129-169.

17. Takala P. 2016. “Word embeddings for morphologically rich languages”. European Symposium on Artificial Neural Networks (27-29 April, Bruges, Belgium), pр. 177-182.

18. Tharwat A. 2018. AdaBoost Classifier: An Overview. Frankfurt: Frankfurt University of Applied Sciences.

19. Vrieland S. D. 2004. Old English and Old Norse. An Introduction to West and North Germanic. Copenhagen: University of Copenhagen.

20. Witten H. I. 2011. Data Mining: Practical Machine Learning Tools and Techniques. Burlington, Massachusets: Morgan Kaufmann Publishers Inc.