Friday, July 17, 2015

The 18 Languages of the 13th International Linguistics Olympiad (2015, Blagoevgrad)

The International Linguistics Olympiad treasures linguistic diversity, this is why we do not require all participants to compete English or another large language, but instead we give all participating countries the opportunity to compete in their own language. We cannot, unfortunately,  accommodate all 7,000+ of the planet. This year, the 44 teams from 29 countries will be provided the problem set in 18 languages - a most impressive task on behalf of the Problem Committee and Jury, as I'm sure you will soon understand. 

The IOL problem committee (the creators of the problem set for the international contest) request that all of the different national contests that are participating in the international contest submit what language they will compete in, well before the contest. Typically this is the same as the language of their national contest. Our policy is that the national contests should be held in the language of the majority of the population or the language of education.  The Problem Committee have a special method for creating the basic problem as free from any one language as possible and later employs translators. The identity of the translators are only known to the Problem Committee, the board and the team leaders are not involved.

This year the participants will be competing in the following languages:
  • Estonian (Finnic, Uralic)
  • English (American & British version separate) (Germanic, Indo-European)
  • Swedish (Germanic, Indo-European)
  • Spanish (Romance, Indo-European)
  • Hungarian (Hungarian, Uralic)
  • French (Romance, Indo-European)
  • Bulgarian (Slavic, Indo-European)
  • Romanian (Romance, Indo-European)
  • Russian (Slavic, Indo-European)
  • Mandarin Chinese (Sinitic, Sino-Tibetan)
  • Dutch (Germanic, Indo-European)
  • Slovene (Slavic, Indo-European)
  • Polish (Slavic, Indo-European)
  • Turkish (Oghuz, Turkic)
  • Japanese (Japonic)
  • Czech (Slavic, Indo-European)
  • Latvian (Baltic, Indo-European)
  • Ukrainian (Slavic, Indo-European)
Most of these languages are Indo-European and they reflect where there is a strong tradition of these contest, i.e. Slavic countries. I'm a typologist (me = Hedvig), i.e. I work with systematic cross-linguistic comparison of the world-s languages, and just as a side note I'd like to remark that this sample is not at all representative of the languages of the world in terms of regions and families, Indo-European is only the fifth largest language family of the world. It is however quite representative of language with large speaker populations. For more about these kinds of things I can recommend exploring Ethnologue statistics and this post.

As I stated previously, the Problem Committee employs a special method to make the core of the problem as unbiased for any languages as possible.  If you want to read more about how that works, read this: Derzhanski, Ivan (2013) Multilingual Editing of Linguistic Problems, In Proceedings of the Fourth Workshop on Teaching NLP and CL August 2013 Sofia, Bulgaria Association for Computational Linguistics 27–34  Ivan Derzhanski is one of the founders of the international contest, a constant member of the problem committee & jury and co-chair of the board of the IOL.

If a participant from one country wishes to compete in different language from their team mates in the individual round, this is alright as long as the language is already represented anyway. Say for example that a Swedish contestant wants to compete in Turkish, this is perfectly alright. A contestant who would like to compete in Jiddisch or Finnish for this year will unfortunately not be able to do so, even if it is their mother tongue and/or an official language of their country.

Participants cannot get the problem set in more than one language, and for the team contest each team can only have the problem set in one language. Having the problem in more than one language is usually an advantage, this is why it is not allowed.

Remember, the contestants do not only receive the problem set in different languages, they also submit (in handwriting, mind you) their solutions in their respective different languages. The jury is not divided into different groups according to languages, but by problems. This means that one or several jury members grade all solutions for one problem, in all languages (no doubt with the help of those with more expertise in specific languages, but still). This is pretty impressive, to say the least, and something that we in the IOL are very, very proud of. If you meet a member of the Problem Committee, be sure to thank them and be very friendly to them. While the participants are all out on the excursions etc they are slaving away correcting solutions to several very complicated and hard problems in 18 languages, submitted in handwriting.


It is now 3 days left until this years contest starts. All the best of luck to everyone.

POST-EDIT: Just to be perfectly clear, the IOL is not a contest in competence of language - i.e. we are not assessing how well the contestants know different languages. We are testing their skill at figuring out linguistic puzzles, and we appreciate their linguistic diversity. To see samples of the problems they are facing, click here.