A game teaches English only when English is load bearing. If the same game can be played in the home language, it will be, usually within two minutes and without anybody deciding to, because a child playing a game is trying to win rather than trying to practise.
So the test for any game is one question: take the English out and see whether the game still works. If it does, the English was decoration, and no amount of enthusiasm from an adult will keep it in.
Eight games, and the rule underneath all of them
- language is the tool, not the topic. A game where English is needed to win teaches; a game about English does not.
- the game must be winnable. A child who cannot win disengages within minutes, and the language goes with them.
- simon says teaches listening and imperatives, and it demands comprehension without demanding speech.
- hide and describe teaches prepositions. Hiding an object and describing where it is produces under, behind and next to naturally.
- snap teaches vocabulary recall under mild time pressure, which is exactly the skill flashcards fail to build.
- board games force turn taking, and with it the small social phrases: your turn, I win, that is not fair.
- keep the rules in English, even when explaining takes longer, because the explanation is the richest language in the session.
- stop while it is still fun, which is the same rule as reading and for the same reason.
Make a word the mechanism
The games that work are the ones where saying something is how a move happens. Guessing games, where the only way to find out is to ask. Instruction games, where one person can see the picture and the other has to build it from what they are told. Hiding games, where warmer and colder are the entire information channel.
In each of those, a child who switches language has not cheated, they have stopped playing, because the move does not exist in the other language within the rules you set up. That is a much stronger constraint than any instruction to speak English, and it does not require the adult to police anything.
Be the referee, never the marker
The adult's job during a game is the rules and nothing else. The moment you correct the English, the game becomes a lesson, and the child now knows the real activity was the test and the game was the wrapping.
That does not mean errors go unaddressed. It means they are addressed the way a fluent speaker addresses them, by using the correct version back in your own next turn, without comment and without pausing. The child hears it, in a sentence that was part of the game, and a proportion of it goes in.
Arguing about the rules is worth protecting for the same reason. A dispute about whether a move was legal is the highest-value English in the whole session: it is unscripted, the child is motivated, and nobody is thinking about the language.
Where it stops: games where English is the decoration
The stall is a vocabulary game that is a test with a card in front of it. Bingo with pictures, matching pairs, snap with words on. The child produces the words to satisfy an adult, and nothing about winning depends on the language, so what is being practised is compliance.
These are popular because they can be bought, they look like teaching, and they generate visible output. They also fail the test above completely: every one of them works exactly as well with the words in the home language.
Nothing corrects it because the session looks successful. Words were said, a game was played, everybody enjoyed it. Whether any of it was English rather than a naming exercise is not visible from inside the twenty minutes.
An ordinary pack of cards
A deck of playing cards is the most useful thing on this page, because a dozen games come out of it and every one of them can be given a rule that requires speech. Call the card before you play it. Ask for what you need rather than taking it. Say why the other player cannot do that.
It replaced the educational game with English printed on the box, which a child identifies as school equipment on sight and which usually contains one mechanic wrapped around a word list.
Its limitation is that the vocabulary is narrow: numbers, colours, a few verbs and the language of turns and rules. That happens to be the most useful narrow vocabulary a beginner could have, and it is still narrow, so it is a chassis rather than a syllabus.
What it looks like when it goes wrong
The recognisable failure is a game that drifts into the home language halfway through and finishes there, with nobody having noticed the switch. That is the design failing, not the child.
The second is an adult who wins. A game a child cannot win stops being worth playing very quickly, and losing on purpose badly is transparent, so the fix is a handicap built into the rules rather than a performance.
Why the reward has to be the game rather than the language is teaching a child without turning it into homework, and the phrases that make a turn-taking game possible at all are what to start with, and in what order.
What a game cannot supply is anything beyond the room it is played in, since the vocabulary of rules and turns is small and closes quickly, and the only material that reliably brings in words nobody in the house would have said is a book: picture books, and what to do when they get stuck.