Twenty-two years of linguistic problems. The archive.
Every International Linguistics Olympiad since 2003 — from the founding contest in Borovets, Bulgaria to the most recent edition in Taipei, Taiwan — has produced a packet of original problems drawn from the world’s languages. This page indexes them by year, by problem type, by language family, and by the path a new student should walk to learn them.
Twenty-two contests, twenty-two cities.
Each card below points to one IOL edition. The official problem packets, marking schemes, and best-solution awards for every year live on the by_year index at ioling.org — click any card to read the source.
No prior knowledge of linguistics is required — what IOL tests is the structure of reasoning itself, applied to the wild diversity of human language.Competition Regulations · ioling.org
Five categories. Each contest packet covers all five.
Every IOL packet aims to present one problem from each of the five core linguistic categories. The list below explains what each category tests, what languages tend to appear in it, and what kind of reasoning a contestant needs to bring.
Phonology
The contestant receives a corpus of words or short phrases in an unfamiliar language, often transcribed in IPA or a close approximation. The task is to deduce the language’s sound system: which vowels and consonants distinguish meaning, what alternations occur, and what governs them.
Common phenomena tested include vowel harmony (Turkic, Uralic, and some African languages), tone systems (Sinitic, Bantu, Mesoamerican), consonant gradation (Finnic), and stress patterns (Indo-European, Polynesian). The contestant rarely needs to know phonetic theory — the corpus contains everything required.
Morphology
Morphology problems test how words are constructed. The contestant is given inflected forms of words from an unfamiliar language and must reconstruct the language’s affix system, paradigm structure, or stem alternations.
These problems often feature agglutinative languages (Turkic, Finno-Ugric, Bantu) where many morphemes stack onto a single stem with predictable rules, or fusional languages (Slavic, Romance) where forms compress several grammatical categories into single morphemes. Some of the most rewarding IOL problems are from polysynthetic languages (Mohawk, Inuit varieties).
Syntax
Syntax problems test how meaning is built from word combinations. Given a corpus of complete sentences with translations, the contestant must figure out the language’s word-order rules, agreement patterns, case marking, and any unusual alignment systems.
Particularly fruitful for IOL are languages with non-Indo-European alignment: ergative-absolutive (Basque, Inuit, many Australian languages), split intransitive (Choctaw, Lakota), or direct-inverse (Algonquian). The contestant must hold multiple competing hypotheses in mind and test them against the corpus.
Semantics & scripts
Semantics problems test conceptual organisation: how a language carves up kinship terms, spatial deixis, colour terms, or evidentiality. Script-decipherment problems present an unfamiliar writing system with partial glosses and ask the contestant to read previously unseen text.
Kinship problems are an IOL specialty because they require the contestant to think carefully about relative perspective: a Mohawk speaker’s word for “uncle” may distinguish maternal from paternal, may distinguish older from younger, and may obligatorily mark the speaker’s gender. Script problems range from logographic (Maya, Egyptian) to abugida (Devanagari, Ethiopic) to syllabaries (Vai, Cherokee).
Numeral systems
Numeral problems are short, dense, and deeply rewarding. The contestant gets a small table of numbers written in an unfamiliar language and must figure out the counting base — often not base 10 — and the arithmetic structure (additive, multiplicative, or compositional).
Famous IOL numeral problems include vigesimal systems (base 20: Mayan languages, Welsh archaic), body-part counting systems (Oksapmin uses 27 body parts as a counting sequence), and mixed bases (French 80 = “quatre-vingts”, or four-twenties). Numerals are typically the shortest problem on each packet but disproportionately difficult relative to length.
The world’s language diversity as a contest curriculum.
IOL problem-setters deliberately draw from typologically diverse languages, ensuring that no single family dominates the corpus. The eight broad families below appear most often in the archive — though hundreds of individual languages have featured across twenty-two editions.
Indo-European
The world’s largest family. Common contributors: Romance, Slavic, Indic, Iranian, Celtic, Albanian, Armenian, Greek. Often used for inflection paradigms.
Uralic & Turkic
Agglutinative powerhouses. Finnish, Estonian, Hungarian, Turkish, Kazakh, Uyghur. Showcased in morphology and vowel-harmony problems.
Niger-Congo & Afro-Asiatic
Tone systems, noun classes, root-and-template morphology. Swahili, Yoruba, Hausa, Arabic, Amharic, Berber. Famous for noun-class concord.
Austronesian
Voice systems, reduplication, evidentiality. Malagasy, Tagalog, Maori, Fijian, Hawaiian, Indonesian. Hosts famously head-marking syntax.
Algonquian & Iroquoian
Polysynthesis, direct-inverse alignment, complex kinship. Mohawk, Cree, Ojibwe, Algonquin. Famous for verb-as-sentence morphology.
Mayan & Andean
Ergativity, vigesimal numerals, classifiers. K’iche’, Yucatec, Tzeltal, Quechua, Aymara. Classic source for script and numeral problems.
Sino-Tibetan
Lexical tone, classifiers, isolation. Mandarin, Tibetan, Burmese, Hmong. Often features in tone and classifier problems.
Australian & Papuan
Ergative split systems, kinship complexity, switch-reference. Warlpiri, Dyirbal, Wamesa, Yimas. Often hosts the team-round corpus.
A three-tier progression from first problem to selection-ready.
The archive is large, and starting from year one can be intimidating. Below is the progression we recommend in the club walkthroughs — tested with grade 7-12 students worldwide and based on what most national olympiads expect.
Start with the official sample problems
Start at ioling.org/problems/samples/. These are intentionally warm-up problems, written for students new to linguistic reasoning. Most can be solved in thirty to fifty minutes.
- ARead the problem twice before writing anything
- BUnderline every gloss to map data first
- CAim for full solutions, not partial guesses
- DCheck answers; understand why each choice works
Work the national-olympiad archives
Move on to NACLO, UKLO, OZCLO, NOL, and other national olympiads’ past papers. These bridge the gap between sample problems and full IOL difficulty, and they typically publish complete answer keys and commentaries.
- ATime yourself — aim for one hour per problem
- BCover all five categories within each fortnight
- CWrite out your reasoning, not just final answers
- DCompare your method against the official commentary
Solve recent IOL papers cold
Sit a full IOL paper in six hours, single sitting, paper and pen only. The most recent editions (Taipei 2025, Brasilia 2024, Bansko 2023) are calibrated to current IOL difficulty and are best paired with the solvers’ choice walkthroughs.
- ASimulate full contest conditions, no distractions
- BScore against the official rubric for self-feedback
- CDiscuss problems with peers post-test
- DRead solution slides for the categories you lost most
The full archive lives on ioling.org.
This page is a community index, not a copy. The IOL Board publishes every contest packet, marking scheme, and best-solution award on its official site — with PDFs in multiple languages, post-contest commentaries, and the solvers’ choice walkthrough series. Click any link to read the source. Where we cite a problem in a club walkthrough, the link returns to the IOL Board’s PDF, not a copy.