A language game is a concept developed by Ludwig Wittgenstein at the beginning of his book Philosophical Investigations. A language game is a simple language, combined with a context that shows what to do with the language. One example he gives is a language for building, containing two words, 'slab' and 'brick'. When A says 'slab' to B, B finds a slab and gives it to A; likewise, when A says 'brick' to B, B finds a brick and gives it to A. Language games are also known as play languages or ludlings.

In another sense, language games are not technically artificial languages so much as heuristics for altering language, like a code. They are used primarily by groups attempting to conceal their conversations from others. Most common examples are: Pig Latin, which is used all over the globe; the Gibberish family, prevalent mainly in the United States and Hungary, and Verlan in France. Each of these language games involves a usually simple standard transformation to speech, thus encoding it. The languages can be easily mentally encoded and decoded by a skilled speaker at the rate of normal speech.

A common difficulty with language games is that they are usually passed down orally. While written translations can be made, they are often imperfect, and thus spelling can vary widely. Some factions argue that words in these spoken tongues should simply be written the way they are pronounced, while others insist that the purity of language demands that the transformation remain visible when the words are imparted to paper. Contrary to what proponents of either side may tell you, there is no one definitive written lexicon for language games, but it is rather a matter of dialect.

See also word play, word game.