Elamur Research⌕
☰
home / papers / 2020 / Real-Valued Logics for Typological Universals: Framework and Application

Real-Valued Logics for Typological Universals: Framework and Application

Summary

Computational typology usually treats features from databases such as WALS or URIEL as Boolean and mostly tests simple two-variable implications. Dönicke, Yu and Kuhn propose a more flexible framework. It handles propositional formulae of any complexity and gives each one a graded truth score computed from corpus data.

Languages and properties form a matrix whose cells take values between 0 and 1. In the authors' setup, these values are relative frequencies of dependency constructions extracted from the UD v2.5 treebanks. Logical connectives are evaluated with fuzzy logic, which makes no independence assumption about related properties. Per-language scores are then averaged with weights based on the phylogenetic tree, so that heavily represented families such as Indo-European do not dominate the result.

The authors test the framework by re-checking known universals and by systematically scoring pairwise word-order implications. The top-ranked results recover classic generalizations, including some that earlier database-driven models had missed. The framework also proposes untested candidate universals. Its main value is that typological hypotheses can be tested on real corpora with little manual work, and its data, logic and weighting components are all interchangeable.

Key points

  • Data: UD v2.5 covers 90 languages in 20 families and 39 subfamilies. Features are single-link (head plus one dependent) or double-link (head plus two dependents) construction frequencies.
  • Logic: The framework uses Zadeh's fuzzy logic. Negation is 1 − x, conjunction is min, disjunction is max, and implication and equivalence are built from these three. The authors chose it over probabilistic logic, whose product rule assumes independence.
  • Weighting: Each family-tree node passes its weight down in equal shares to its children. All data is kept, with no undersampling, and early experiments showed better agreement with known universals than an unweighted average.
  • Pairwise experiment: The authors tested 20 × 20 implications over 10 relations. The top result, acl-HEAD ⇒ nmod-HEAD (score 0.97), matches Universals Archive #176 (Hawkins 1983), yet neither Bjerva et al. (2019b) nor Daumé and Campbell (2007) had found it. Part of Greenberg's universal 18 (UA #57) was also recovered with a score of 0.92.
  • New candidates and code: The search suggests new candidates, such as a postverbal auxiliary implying postpositions and prenominal genitives, articles and numerals. Code is available at github.com/tidoe/typology-coling.

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

Fields: Linguistics & languages

Cited by

2 papers of our library cite it (of the 140673 whose bibliographies we read); 1 of them have a page here, 1 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2020.coling-main.353

@inproceedings{donicke2020real,
  title = {{Real-Valued Logics for Typological Universals: Framework and Application}},
  author = {Dönicke, Tillmann and Yu, Xiang and Kuhn, Jonas},
  booktitle = {Proceedings of the 28th International Conference on Computational Linguistics},
  year = {2020},
  doi = {10.18653/v1/2020.coling-main.353},
  url = {https://doi.org/10.18653/v1/2020.coling-main.353}
}

References

34 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Merrie Bergmann. 2008. An introduction to many-valued and fuzzy logic: semantics, algebras, and derivation systems. Cambridge University Press. find
  2. Balthasar Bickel. 2008. A refined sampling procedure for genealogical control. STUF - Language Typology and Universals, 61:221–233. find
  3. Balthasar Bickel. 2011. Absolute and statistical universals. In Patrick Colm Hogan, editor, The Cambridge Encyclopedia of the Language Sciences, pages 77–79. Cambridge University Press, Cambridge. find
  4. Johannes Bjerva, Yova Kementchedjhieva, Ryan Cotterell, and Isabelle Augenstein. 2019a. A probabilistic generative model of linguistic typology. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1… find
  5. Johannes Bjerva, Yova Kementchedjhieva, Ryan Cotterell, and Isabelle Augenstein. 2019b. Uncovering probabilistic implications in typological knowledge bases. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3924–3930, Florence, Italy, July. Associatio… find
  6. Johannes Bjerva, Robert Östling, Maria Han Veiga, Jörg Tiedemann, and Isabelle Augenstein. 2019c. What do language representations really represent? Computational Linguistics, 45(2):381–389. find
  7. Hal Daumé, III and Lyle Campbell. 2007. A Bayesian model for discovering typological implications. In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics, pages 65–72, Prague, Czech Republic, June. Association for Computational Linguistics. find
  8. Tillmann Dönicke. 2020. Evaluation of complex typological universals with language vectors and real-valued logics. Master’s thesis, University of Stuttgart. find
  9. Matthew S. Dryer and Martin Haspelmath, editors. 2013. WALS Online. Max Planck Institute for Evolutionary Anthropology, Leipzig. find
  10. Matthew S. Dryer. 1989. Large linguistic areas and language sampling. Studies in Language, 13:257–292. find
  11. Matthew S. Dryer. 2003. Adverbial subordinators and word order asymmetries. In John A. Hawkins and Anna Siewierska, editors, Performance principles of word order, EUROTYP Working Papers, pages 50–67. European Science Foundation. find
  12. Matthew S. Dryer. 2018. On the order of demonstrative, numeral, adjective, and noun. Language, 94(4):798–833. find
  13. Didier Dubois and Henri Prade. 1993. Fuzzy sets and probability: misunderstandings, bridges and gaps. In [Proceedings 1993] Second IEEE International Conference on Fuzzy Systems, volume 2, pages 1059–1068. find
  14. Michael Dunn, Simon J. Greenhill, Stephen C. Levinson, and Russell D. Gray. 2011. Evolved structure of language shows lineage-specific trends in word-order universals. Nature, 473(7345):79–82. find
  15. Kim Gerdes, Sylvain Kahane, and Xinying Chen. 2019. Rediscovering Greenberg’s word order universals in UD. In Proceedings of the Third Workshop on Universal Dependencies (UDW, SyntaxFest 2019), pages 124–131, Paris, France, 26 August. Association for Computational Linguistics. find
  16. Joseph H. Greenberg. 1963. Some universals of grammar with particular reference to the order of meaningful elements. Universals of language, 2:73–113. find
  17. Michael Hahn, Dan Jurafsky, and Richard Futrell. 2020. Universals of word order reflect optimization of grammars for efficient communication. Proceedings of the National Academy of Sciences, 117(5):2347–2353. find
  18. John A. Hawkins. 1983. Word Order Universals. Academic Press, New York. find
  19. Gerhard Jäger. 2018. Global-scale phylogenetic linguistic inference from lexical resources. Scientific data, 5(1):1–16. find
  20. Stephen C. Levinson and Russell D. Gray. 2012. Tools from evolutionary biology shed new light on the diversification of languages. Trends in cognitive sciences, 16(3):167–173. find
  21. Patrick Littell, David R. Mortensen, Ke Lin, Katherine Kairis, Carlisle Turner, and Lori Levin. 2017. URIEL and lang2vec: Representing languages as typological, geographical, and phylogenetic vectors. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational… find our card
  22. Haitao Liu. 2010. Dependency direction as a means of word-order typology: A method based on dependency treebanks. Lingua, 120(6):1567–1578. find
  23. Bethany C. Lochbihler. 2017. Typology across treebanks. Master’s thesis, University of Stuttgart. find
  24. Xia Lu. 2013. Exploring word order universals: a probabilistic graphical model approach. In 51st Annual Meeting of the Association for Computational Linguistics Proceedings of the Student Research Workshop, pages 150– 157, Sofia, Bulgaria, August. Association for Computational Linguistics. find
  25. Jan Łukasiewicz and Alfred Tarski. 1930. Untersuchungen über den Aussagenkalkül. Comptes rendus des séances de la Société des Sciences et des Lettres de Varsovie, 23, cl. iii:39–50. English translation by J. H. Woodger, “Investigations into the Sentential Calculus.” In: Alfred Tarski, Logic, Se… find
  26. Chaitanya Malaviya, Graham Neubig, and Patrick Littell. 2017. Learning language representations for typology prediction. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2529–2535, Copenhagen, Denmark, September. Association for Computational Linguisti… find
  27. Eduardo Mizraji. 1992. Vector logics: The matrix-vector representation of logical calculus. Fuzzy Sets and Systems, 50(2):179 – 185. find
  28. Joakim Nivre, Marie-Catherine de Marneffe, Filip Ginter, Yoav Goldberg, Jan Hajič, Christopher D. Manning, Ryan McDonald, Slav Petrov, Sampo Pyysalo, Natalia Silveira, Reut Tsarfaty, and Daniel Zeman. 2016. Universal dependencies v1: A multilingual treebank collection. In Proceedings of the Tenth I… find our card
  29. Arturo Oncevay, Barry Haddow, and Alexandra Birch. 2019. Towards a multi-view language representation: A shared space of discrete and continuous language features. Phonology, 27(28):84–51. find
  30. Frans Plank and Elena Filimonova. 2000. The Universals Archive: A brief introduction for prospective users. STUF-Language Typology and Universals, 53(1):109–123. find
  31. Nicholas J. J. Smith. 2012. Many-valued logics. The Routledge companion to philosophy of language, pages 636–651. find
  32. Jae Jung Song. 2018. Language samples and sampling methods. In Linguistic Typology, chapter 5, pages 77–117. Oxford University Press. find
  33. Lotfi A. Zadeh. 1965. Fuzzy sets. Information and control, 8(3):338–353. find
  34. Daniel Zeman, Joakim Nivre, Mitchell Abrams, Noëmi Aepli, Željko Agić, Lars Ahrenberg, Gabrielė Aleksan-dravičiūtė, Lene Antonsen, Katya Aplonova, Maria Jesus Aranzabe, Gashaw Arutie, Masayuki Asahara, Luma Ateyah, Mohammed Attia, Aitziber Atutxa, Liesbeth Augustinus, Elena Badmaeva, Miguel B… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/donicke-2020-real-valued-logics-typological-universals/