The Cartoon That Proved Children Don't Just Imitate Grammar

In 1958 Jean Berko, a 26-year-old Harvard graduate student in social relations, drew a small blue bird with two thin legs on a card. She labeled it WUG. To pre-school children she showed the card and said, "This is a wug. Now there is another one. There are two of them. There are two ____." Almost every child filled in the blank with wugs.

The finding looks obvious. The point is that the children had never heard the word wug before. They could not have memorised its plural. To produce wugs they had to apply, on the spot, an unconscious rule that says English plurals end in /-z/ after a voiced consonant, /-s/ after a voiceless one, and /-ɪz/ after a sibilant. Berko ran the same trick on past tenses (ricking → ricked), possessives (the wug's hat), and derivations. Children three and four years old produced the right form for words they were inventing live.

The study, published as The Child's Learning of English Morphology, reframed how linguists thought about acquisition. Skinner's behaviorist account had children learning language by imitating heard sentences and being reinforced. Wug-test results were direct evidence against that: there is no possible adult sentence the child could have copied. Whatever was happening in the kids' heads was rule-based and productive, the same kind of pattern that Noam Chomsky was at the same time arguing must underlie any natural language.

Berko Gleason kept the field, and the test, alive for the next sixty years. Researchers still hand children pictures labeled with nonce words — gutch, bik, spow — to probe how the morphology of dozens of languages develops. The original blue wug, redrawn from memory in marker, has become the unofficial mascot of psycholinguistics.