Nonsense Words: Why Teachers Use Them

Ask a five-year-old to read 'dreft' and you learn more in two seconds than a page of real words would tell you. Real words can be recognised from memory — a child who has seen 'cat' a hundred times is not necessarily decoding it. A made-up word cannot be memorised, so it tests the only thing that matters at that stage: can this child turn letters into sounds and blend them together?

Why a fake word is the better test

Early reading runs on two engines: decoding, which works a word out from its letters, and sight recognition, which retrieves a familiar shape. Both are useful, but only decoding generalises to words the child has never met. Because a pseudoword is unfamiliar by definition, it isolates decoding completely. That is why nonsense words show up in formal reading assessments rather than being a teacher's quirk — England's Year 1 phonics screening check mixes pseudowords in with real ones, and the widely used American nonsense-word fluency measures do the same with simple consonant-vowel-consonant items.

It also explains the little monsters. Screening materials often print an alien next to each pseudoword so children understand the word is not supposed to mean anything. Without that cue, a conscientious reader will try to convert 'dreft' into 'drift' or 'draft' and get marked wrong for being clever.

What a good pseudoword looks like

A usable nonsense word follows English spelling rules exactly and breaks only the dictionary. It should be pronounceable on sight, contain the specific pattern you want to test, and not accidentally be a real word — or a near-miss of a rude one. Difficulty comes from structure, not length: simple consonant-vowel-consonant first, then blends, then the long-vowel spellings.

• vop
• nid
• tam
• chex
• streb
• plimp
• dreft
• snalk
• quib
• thusp
• blone
• frayle

The first half tests single letters and simple blends. The last few carry digraphs and long-vowel patterns, which is exactly where most readers stall.

A five-minute routine at home

Generate five or six nonsense words. Ask the child to read each one aloud and invent a silly meaning for it — the definition step matters, because it turns the exercise into a game and proves they are comfortable with the word being meaningless. Mark nothing. Just note which patterns caused a hesitation, then use real words containing that same pattern in the next few days of reading.

One trap to avoid: doing this with a child old enough to know they are being measured and already anxious about reading. For them, hide the same work inside hangman or a spelling relay, where sounding out letter strings is the mechanic rather than the assessment.

What the errors actually tell you

Three patterns are worth recognising. Reading 'dreft' as 'drift' means the child is pattern-matching to real words — common in strong readers, a sign of comprehension rather than failure, but it hides decoding gaps. Dropping a consonant out of a blend, 'sreb' for 'streb', points at blending rather than letter knowledge. And guessing wildly from the first letter alone means decoding has not started yet, and it is worth returning to single letter-sound work before anything longer.