A small plastic witch and her broom are being moved around a table in the UConn Child Labs by graduate student Kazuko Hiramatsu. Also at the table is Lulu, a colorful velvet hand puppet, who often makes mistakes when she talks. Lulu, a fish from the moon, is being operated by Diane Lillo-Martin, a professor of linguistics.

"The witch sweeping the floor is," says Lulu. Pre-schooler Joanne giggles at the error, and tells Lulu the right way to say it: "The witch is sweeping the floor."

To pre-schooler Joanne, it's a game. To Lillo-Martin and Hiramatsu, it's important research. They are learning how children acquire language, one of the most fundamental human qualities.

Language acquisition is one of several areas studied by researchers in the University's nationally ranked linguistics department.

"We look at the principles that govern the structure of languages," says Lillo-Martin, "and in particular, the ways in which languages are the same in structure and the ways in which they differ."

Despite the superficial differences between languages, there are striking and important underlying similarities. "Our work is mainly devoted to discovering and explaining these commonalities," says Lillo-Martin, who heads the department. Researchers look for generalizations that describe the patterns that are found in languages, focusing on syntax (sentence structure), phonology (sound structure) and language acquisition.

Human language is a complex, rule-governed system and yet children acquire their native languages rapidly, in the space of a few years - even before they can tie their shoes.

Lillo-Martin, whose specialty is language acquisition, says when people hear three- or four-year-olds talking, they tend to focus on the errors that the child makes, Lillo-Martin says. But the errors are minimal compared to the child's successes, especially when one considers the complexity of language, she adds.

One error children often make that intrigues Lillo-Martin is how they ask questions. "Although children acquire language effortlessly, there are certain types of mistakes they make, but outgrow." The majority of pre-schoolers in the study at the Child Labs, for instance, make mistakes when they ask a question that has a negative in it. But when a puppet asks the question incorrectly the children are able to judge it as being wrong.

The linguistics department is a leading center for theoretical research in syntax and phonology and for experimental research on child language acquisition. The National Research Council's most recent study of graduate programs in linguistics gave the department its highest ranking: "extremely effective."

"Our graduate students get great jobs," Lillo-Martin says. Since 1991, 32 students have completed the Ph.D. in linguistics. Of those, 16 went directly into tenure-track positions, seven took post-doctoral or lectureship positions, followed by tenure-track positions. Another five have post-doctoral or other temporary academic positions. "We're very proud of that. They have done well," Lillo-Martin says.

This fall the department will offer a new joint major in linguistics and psychology for undergraduates.

In another research project with assistant professor William Snyder, Lillo-Martin and several graduate students are conducting a cross-language study of two-year-olds who speak Russian, Spanish, English and Japanese.

Lillo-Martin also is conducting research on the impact of delayed exposure to American Sign Language on the process of signed and written language acquisition among deaf children who have hearing parents.

Sherry Fisher