Lila Gleitman while driving with her year-old daughter, advised her daughter to "hold on tight" at a tricky turn. The toddler responded: "Isn't that 'tightly?" It was a turning point in the young scholar's career. She had already switched from English literature to classics, in which she quickly became bored of her teacher's digressions on Athenian society (she wanted to get back to grammar). Realising that her two-year-old already had an understanding of language made Gleitman want to get into her child's head and those of other children.

Gleitman, who died on August 8th at the age of 91, turned children's learning of language into a research career that helped define psycholinguistics, a field that hardly existed before. Her early interest coincided with the emergence in the 1950s of Noam Chomsky, a frequent visitor to the University of Pennsylvania when she began taking courses there. Until then linguistics largely involved concentrating on what people said, shying away from what they might be thinking

The two scholars' work and that of others instead considered the mental systems that might produce the sentences you hear, which are shaped by abstract rules that speakers may not even know that they know An early piece of Gleitman's research, for example, investigated small children's "telegraphic" speech, in which many words are left out: a toddler might say "throw ball" rather than "throw me the ball". This seems to imply that the child's knowledge is primitive. But she found that children nonetheless comply with instructions better when their parents use adult-style English than when they mimic their offspring. She and her colleagues concluded that youngsters know more than they can say.

As the learning process goes on, children deploy some remarkable strategies. They often seem to correctly guess what a word means after hearing it just once.

The physical environment is an obvious spur (as when they hear "dog" and see one at the same time). But how would a child guess the meaning of the verb in "I believed that he lost his keys"? Gleitman noticed that the sentence structure is identical to those with other verbs that mean similar things (i.e., refer to states of mind): saw, remembered, imagined, forgot, worried and doubted. This intuitive aid helps children learn astonishingly quickly, a process she called "syntactic bootstrapping"

Her work also had implications for the debate over whether a person's native language strongly influences how they think or even what they can think. She was convinced that all languages shared fundamental traits, forged by the nature of the human mind itself, the effects of using a particular one on cognition were modest and fleeting. The notion that speaking a different language entails a profoundly different way of thinking was romantic and tempting, but she would not buy it.