News Release

NYU Researchers Find Infants Understand How To Put Words Together Into Simple Language-Like Sentences

Peer-Reviewed Publication

New York University

Cognitive Scientist Gary Marcus Says Infants Can Recognize Simple Language-Like Rules

Findings May Have Implications For Diagnosis And Treatment Of Language Disorders



A seven-month-old infant in Marcus' experiment looks surprised as she listens to a sentence with an unfamiliar structure.

A team of researchers led by New York University psychologist Gary Marcus has discovered that seven-month-old infants have a previously undiscovered ability for learning about the world and attacking the problem of language acquisition. Marcus and his colleagues found that infants can recognize and generalize simple language-like rules.

Marcus said, "Over the past forty years, scientists who study language have found that the mind puts together sentence components -- nouns, verbs, and adjectives -- as if they were Xs, Ys, and Zs in an equation. We have now found that even seven-month-olds can do the same. Of course, babies of this age don't literally know nouns, verbs, and adjectives. But our results show that they understand the algebra of how to put words together into sentences.

"Abstract rules form the core of everything from computer programs to grammars. Our results show that babies' minds are built to look for such rules -- even without being told. Far from being passive, babies are active learners, constantly trying to understand the world around them, always analyzing the world and looking for regularities.

"Our results may have important implications for the diagnosis and treatment of language disorders. If we can spot children at risk for language problems while they are still infants, we may be able to keep problems from developing."

In a "Perspective" article that will also run in the January 1st Science, MIT cognitive scientist Steven Pinker wrote, "Marcus et al have now shown that infants as young as seven months can abstract simple rules from language-like sounds....Their demonstration suggests that the ability to recognize abstract patterns of stimuli .... is a basic ability of the human mind."

Marcus and his colleagues had each infant listen for two minutes to a set of 16 different 3-word "sentences" from an artificial language. For each infant, all the sentences within the set followed a uniform structure. For example some the infants heard 'sentences' that followed an A-B-A structure ("Li ti li," ), while other infants heard sentences that follow an A-B-B structure ("Li ti ti") or an A-A-B structure (Li li ti.)

Each infant was then presented with a new set of 3-word sentences. Half of the sentences followed the same structure as the ones in the initial set, such as A-B-B. Half followed a different structure (e.g., A-B-A for infants who initially heard A-B-B sentences).

The researchers found that more than 90 percent of the infants were able to recognize the sentence structure that they had heard before. The babies listened longer and paid more attention to the sentences with unfamiliar structure, as measured by how long the infants looked at a light that blinked alongside the speakers as the sentences were presented.

Various controls ensured that the children did not simply like the sound of some sequences more than others, or recognize partial chunks of sentences that they had memorized.

Gary F. Marcus is an associate professor in NYU's psychology department. He received his BA in 1989 from Hampshire College and his Ph.D.in 1993 from MIT's Department of Brain and Cognitive Sciences . His research on language acquisition and computational models has been published in journals such as Cognition, Cognitive Psychology, Journal of Child Language and the Monographs of the Society for Research in Child Development. His book, The Algebraic Mind, is being published by MIT Press. He is on the editorial boards of Cognition and Developmental Psychology. In 1996, he received the Robert L. Fantz award for new investigators in cognitive development.

###



Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.