The ability to speak and understand language is probably the most intricate skill that people possess. It is certainly our most uniquely human ability. This project investigates how such an important skill is acquired and continues to develop throughout our lives. The resulting research findings will have applications in three areas: (1) the design of computer systems that can recognize and produce language, (2) the teaching of language skills, including reading, writing, and foreign language learning, and (3) the remediation of language disorders such as aphasia and dyslexia.
The skills of language production and comprehension together comprise the language processing system, a complex system within the brain that turns thoughts into grammatical sequences of words and then articulates those words, and also understands the speech of others. The central claim of this project is that the language processing system is constantly changing. It adapts quickly to recent experience, while continuing to reflect the accumulated experience of a lifetime of speaking, listening, reading, and writing. Such adaptation, or learning, which usually occurs without any awareness, enables us to speak and comprehend effortlessly in the face of ever-changing situations. This perspective leads to two key research questions: (1) What learning mechanisms enable the language processing system to adapt swiftly and effectively to experience? (2) How can the study of learning tell us more about how the language processing system works?
The current project addresses these questions by investigating the comprehension and production of individual words as well as words combined into sentences. The research is multi-disciplinary, integrating the work of psychologists, linguists, and computer scientists. It involves experiments that measure how the abilities to speak and understand change with experience. For example, one set of experiments will determine how quickly children and adults can recognize a spoken word, as a function of how recently they have heard or said the same word before. More generally, the project's experiments involve measurement of the time it takes people to say or to recognize words, the kinds of errors that commonly occur during speech, and the patterns of eye movements and brain electrical activity that accompany word and sentence comprehension. The research is carried out on normal adults and children, and also on individuals with brain damage affecting their abilities to learn skills. The experimental work is combined with the development of computer models of the language processor that are capable of learning. These models, some of which will consist of artificial neural networks, will serve two functions. First, the models will instantiate theories of how people process language, and second, the models will offer insights into how machines can be designed to interact with people through language and speech.
