The critical period hypothesis holds that there is a window early in life during which language is acquired readily and after which full acquisition becomes difficult or impossible. It matters well beyond linguistics, because it bears on when deaf children should be given access to sign language, how second languages should be taught, and what is owed to children deprived of language early.
Age effects on language learning are real and large. Adults who begin a second language rarely reach native-like pronunciation, and the younger a learner starts, the better the eventual outcome tends to be. In first-language acquisition the evidence is starker: deaf children who receive no accessible language in early childhood show lasting deficits in grammar that later exposure does not fully repair, and this finding is consistent across studies of late signers.
Eric Lenneberg set out the hypothesis in 1967, proposing a period ending around puberty, tied to the brain losing plasticity as it matures and lateralises. The underlying idea came from ethology, where critical periods for imprinting and for the development of vision are well documented, and from clinical observation that children recover from brain injury affecting language far better than adults do.
Is it a cliff or a slope? A genuine critical period implies a relatively sharp offset. Much of the data instead shows steady decline across the lifespan with no obvious discontinuity, which is better described as a sensitive period. A large study using online data from hundreds of thousands of speakers suggested the ability to learn grammar to a native standard declines sharply only in the late teens, later than Lenneberg proposed, and that finding is itself contested on methodological grounds.
Is it maturation, or circumstance? Older learners differ from children in more than brain plasticity: they have an established first language interfering, less immersion, more self-consciousness, and different motivation. Disentangling biological maturation from these is genuinely hard, and critics argue the age effect may be substantially social rather than neural.
Which parts of language? The evidence is much stronger for phonology and for grammar than for vocabulary, which adults often learn faster than children. Any single window covering all of language is difficult to sustain.
The deprivation cases prove less than they appear. Children found after extreme isolation, of whom Genie is the best known, are cited on both sides. They were also starved, abused, and deprived of ordinary social contact, so their later difficulties cannot be attributed to language timing alone, and the sample is tiny.
Whatever the mechanism, the practical implication is not in serious doubt: delay is costly. This is the basis of newborn hearing screening and of the argument that deaf infants should be given sign language immediately rather than waiting on decisions about implants, since a child who waits may lose access to a full first language during the period when acquisition is easiest.
