2016Q3 Reports: Linguistics Olympiads

Linguistics Olympiads originated in the Soviet Union in the 1960s. By the 1980s they were being offered in other countries. The International Linguistics Olympiad was founded in 2003. The countries competing in the 14th IOL in 2016 in Mysore, India (http://iol14.plo-in.org/) were: Australia, Bangladesh, Brazil, Bulgaria, Canada, China, Czech Republic, Estonia, Hungary, India, Ireland, Isle of Man, Kazakhstan, Japan, Latvia, Nepal, Netherlands, Pakistan, Poland, Korea, Romania, Russia, Slovenia, Spain, Sri Lanka, Sweden, Taiwan, Turkey, Ukrain, United Kingdom, and United States. The contest was hosted by Microsoft and Infosys.

Results of the 2016 IOL will be posted at http://www.ioling.org.

Next year, the IOL will be in Ireland (http://www.iol2017.ie/), hosted by the ADAPT center at Dublin City University.

NAACL has supported the US and Canadian contest (NACLO) with a donation of $5,000 per year for several years. We encourage ACL to support the IOL.

Internationally, about 15,000 students per year participate in Linguistics Olympiads in their countries, and global participation increases by about 10% per year. (This number includes only those students competing in national Olympiads specifically dedicated to linguistics and computational linguistics; if we include Olympiads that test logic and linguistics together such as the Slovenian contest in mathematical logic, this number rises to about 40,000 students.) Currently, 31 national competitions are accredited by the International Linguistics Olympiad, and about 2-3 new national competitions come into existence each year.

While these national contests are in many ways as diverse as the countries that host them, there are strong unifying principles. Each national Olympiad takes an empirical rather than prescriptive approach to the language sciences, and given that linguistics and computational linguistics are rarely encountered in secondary education, each Olympiad strives for a contest in which prior knowledge is not required for success. Contestants are challenged with new phenomena in unfamiliar data, rather than being tested on languages and concepts they already know, and come to learn that languages are something that you can discover through analysis of real language data, rather than something prescribed by teachers and in books.

The IOL is the culminating event that motivates bringing linguistics to 40,000 students per year in 31 national contests worldwide. Although less than 150 of these students attend the IOL, it provides a focus for admiration of linguistic talent and a target for setting goals in academic achievement in a fun way outside of their normal high school subjects.

The IOL problems are puzzles with solutions that snap into place and can be reached by a thread of analytic reasoning. (IOL problems can be viewed on the IOL web site, ioling.org.) Every problem is carefully reviewed and tested for solvability, but they are not meant to be solved easily. From the IOL jury's perspective, the ideal high score would be around 85%, with average scores under 40%.

IOL problems exercise skills needed for any academic field. Students are challenged to find an initial hook or clue, formulate an approach, recognize patterns, form hypotheses, support or reject hypotheses based on data, restructure the solution space many times, and finally write a winning explanation of what they have discovered in the process. Also, as emphasized by Slovenia's national logic contest, it is important to bring rigorous logical and deductive reasoning to all students, even those who are not studying mathematics.

In addition to exercising analytic skills, each IOL problem rewards the solver with a linguistic discovery about sounds, sound change, morphemes, lexical semantics, number systems, poetic meter and rhyme, or another area of linguistic structure.