The UK Linguistics Olympiad doesn't test what your child knows about French, or Mandarin, or any other named language. It tests whether they recognise a handful of recurring ideas — how words are built, how sentences are ordered, how words agree with each other — and can apply them to a language they've never seen before. Here's the toolkit.
If you've read our full guide to UKLO, you already know the headline: it's free, it's run through school, and it needs no prior language knowledge. But "no prior knowledge" doesn't mean nothing helps. It means the thing that helps isn't facts about a specific language — it's familiarity with the shapes that linguistic puzzles keep taking. A child who has met these shapes before spends their exam time solving the puzzle, not discovering the concept from scratch under time pressure.
We built a set of flashcard decks that trace these ideas across Foundation-tier UKLO material, and a pattern falls out immediately: the same handful of concepts turn up again and again, dressed in different invented languages each time. This article is a tour of that toolkit, organised by category.
Grammar's building blocks: word classes and morphemes
Every puzzle starts with words, so the first skill is naming what kind of word you're looking at. Nouns name things, verbs name actions, adjectives describe nouns, and pronouns stand in for nouns — the classic Year 7 quartet, and the single most recurring idea in Foundation-tier material by a wide margin. A puzzle's word list rarely labels these for you; you infer them from the English meanings, and once you've spotted the verb, the rest of the sentence tends to fall into place around it.
Below the level of the whole word sits the morpheme — the smallest piece of a word that carries its own meaning. "Unkindness" has three: un- (not), kind, and -ness (a state). This matters because invented puzzle languages build words the same way: a root plus bolt-on pieces for plural, tense, or possession. Spot the pieces, and you have the grammar.
Those bolt-on pieces have names depending on where they sit. A prefix goes before the root, a suffix after — and the labels describe position, not meaning, which trips more children up than you'd expect. Less commonly, a language slots a piece into the middle of a word — that's an infix — and a handful of languages (the Celtic family among them) change the word's very first sound depending on what came before it, called initial mutation. All three — prefix, suffix, mutation — are just different places grammar can hide, and a puzzle will tell you which one you're dealing with by showing you where two related forms differ.
How sentences are ordered
English is what linguists call SVO: subject, verb, object — "the cat (S) chased (V) the mouse (O)." It feels so natural that it's easy to assume every language works this way. Most don't. Japanese and Turkish are SOV (verb last); Welsh and Irish are VSO (verb first); a handful of languages are even OSV. This three-letter labelling system — find the subject, the verb, and the object, then note the order they come in — is, alongside word-building, the single most recurring idea across Foundation-tier years.
The trap is assuming the first word of a sentence must be the subject, or translating word by word in English order. The fix is mechanical: find the word that matches the English verb, note what sits on either side of it, and let that — not habit — tell you the order.
Some languages don't rely on word order at all. Instead, a noun's ending shows its job in the sentence — this is case — which frees the word order up to move around without losing clarity. Latin has six cases; English keeps a trace of the idea only in its pronouns ("he" as subject, "him" as object). When a puzzle's word order looks unusually loose, the fix is nearly always a case ending doing the work that order does in English.
Agreement: words that copy each other
Agreement is one word copying another's grammar. English keeps only a scrap of it — "she runs" but "they run" — but many languages make far more of the idea. A verb might copy its subject's person (I/you/he) and number (singular/plural); an adjective might copy its noun's number too, something English adjectives never do ("red house", "red houses" — "red" never changes). Some languages go further and sort every noun into a grammatical gender class — masculine, feminine, sometimes more — that the verb or adjective then has to match, for no reason connected to biological sex.
The reasoning move here is always the same: when two forms of one word differ, and the difference lines up with a change in a nearby word, agreement is very likely the explanation. Sort the data by which form goes with which partner, and the classes reveal themselves.
Cognates, loanwords, and the trap of lookalikes
English "night" and German "Nacht" descend from the same ancient word — they're cognates, and if a puzzle language is a relative of one your child knows, a lookalike word is often a free clue. But a lookalike can also be a false friend: a word that resembles an English one without sharing its meaning. The right habit is to treat every lookalike as a hypothesis to test against the data, never as a fact to bank on.
A related idea is the loanword — a word borrowed wholesale from another language, which often keeps the spelling or sound habits of where it came from and so breaks the borrowing language's own rules. When a puzzle's rule fits nearly every word except two or three, the exceptions are often loanwords, not proof the rule is wrong.
Reading unfamiliar scripts
Not every UKLO puzzle uses our alphabet. An alphabet gives one sign per sound; a syllabary (Japanese kana, for instance) gives one sign per whole syllable; a logographic script (Chinese characters) gives one sign per word or meaning, not per sound at all — so you can't "sound it out," you have to match it. Knowing which type you're facing tells you what to count and how to compare word lengths.
Two more terms round this out. Transliteration is rewriting a word letter by letter into a different alphabet — turning Cyrillic Москва into "Moskva" — and it is not the same as translation, which changes the meaning, not just the spelling. And the IPA (International Phonetic Alphabet) gives one symbol per sound, in any language, which matters because English spelling is a famously unreliable guide to how a word actually sounds — "thought" has seven letters but only three speech sounds.
Why this toolkit, and not a vocabulary list
Notice what's absent from this list: no French verb tables, no Mandarin tone rules, nothing that belongs to one named language. That absence is deliberate, and it's the entire reason UKLO can't be crammed the way a normal subject test can — a theme we return to in how to revise for UKLO with spaced repetition. What can be built, well before the day of the competition, is fluency with the concepts themselves, so your child spends the exam applying ideas they already recognise rather than discovering them cold.
Two companion pieces go deeper on how this plays out. UKLO patterns that keep coming up looks at exactly how often each of these ideas resurfaces across years, with the numbers behind it. UKLO common mistakes and misconceptions looks at the reasoning habits that go wrong even when a child knows the concept — because knowing what a case ending is and remembering to check for one under exam pressure are two different skills.
FAQ
Do I need to teach my child linguistics before UKLO?
No — UKLO puzzles are solvable from the data given, with no outside knowledge required. But knowing the handful of ideas on this page means your child recognises the shape of a problem faster, instead of spending exam minutes reinventing each idea from scratch.
Is this the same as learning grammar for GCSE?
It overlaps but isn't the same. GCSE English grammar is mostly about English. UKLO concepts are about how ANY language could work — the same word class or agreement idea, but applied to unfamiliar data. A child strong in one often finds the other easier, but neither is a substitute for the other.
Which of these concepts matters most?
Word order (SOV/SVO/VSO/OSV) and word-building (roots, prefixes, suffixes) are each present in around 16 of the 17 Foundation-tier years we reviewed — by far the most common. If you only have time for two ideas, start there.