Betty Tuller - US grants

This proposal is aimed at understanding the formation of articulatory patterns during speech production, using a strictly operational approach in which theoretical tools and experimental data are developed together. The empirical approach is guided by synergetics, the theory of spontaneous structure formation in open, nonequilibrium systems (Haken, 1975). In synergetics, behavioral patterns produced by multidegree of freedom systems can be characterized by (low-dimensional) collective variables, the so-called order parameters. We aim to discover these order parameters and their dynamics in the speech production system by studying the stability and loss of stability of spatiotemporal speech patterns using experimental paradigms that allow monitoring of trajectories of multiple articulators. Order parameters have two characteristics to be investigated here. First, they should remain stable or exhibit only weak dependence on "control parameters" that potentially deform the system. Second, they should exhibit loss of stability at bifurcation points, where the behavior of the system changes abruptly. Based on preliminary work, Section I of the present proposal tests the hypothesis that the continuous relative phasing among observed articulatory components is an essential macroscopic variable that remains stable across transformations in style of speaking, stress, rate, and loudness. Continuous relative phase is a measure derived from individual articulator motions (e.g., tongue, jaws, lips, glottis) and vocal tract variables (e.g., lip aperture, tongue construction). Test of convergence on this candidate order parameter are carried out in Section II, which examine the mechanisms underlying voluntary and involuntary switching among collective states. In addition to their theoretical importance, the experiments proposed are intended to answer the question of where, when and how the speech motor system changes its parameters in simple utterances. In short, a program of research is proposed that seeks to identify the lawful regularities in speech motor patterns that arise in a self- organized fashion, and that structure sound for a listener. Such a characterization of normal speech production is a necessary precursor to the definition and categorization of disordered stated (so-called articulation disorders).

Every year people immigrate to the United States who do not understand or speak English. While learning a new language is not too difficult for children and adolescents, it is a much harder feat for adults. This is in part because adult speakers cannot usually hear meaningful sound distinctions in English that are not part of their native language. For example, in English "bit" and "bead" have different meanings but in Spanish the two English words may sound identical. It is often extremely difficult for native Spanish speakers to learn to hear the difference between these two words. Another common example is Japanese speakers' difficulty in perceiving the difference between "r" and "l." It works both ways, so to speak, typical American English speakers cannot tell the difference between, for example, the Hindi "d" (as in dal, which is the word for lentils) and the American English "d", which is produced with the tongue slightly further back in the mouth. Typical courses in English for Speakers of Other Languages (ESOL) concentrate on vocabulary, grammar, and pronunciation, but much is as yet unknown about how adults can learn to perceive new speech sounds.

With support from The National Science Foundation, Dr. Betty Tuller (a cognitive scientist) and Dr. Viktor Jirsa (a theoretical physicist) will advance our understanding of 1) how learning of a meaningful speech sound occurs over time, 2) exactly what is learned, 3) the impact of a newly learned distinction on a similar, but not identical, native speech sound, 4) transfer of learning to a novel speech contrast, and 5) variables that may facilitate learning. The collaboration between a cognitive scientist and a theoretical physicist has the potential to be exceptionally fruitful, as behavioral research and theoretical modeling, grounded in the dynamics of the problem, can inform and guide each other. This multidisciplinary approach has proven to lead to new insights in many realms of human behavior. The proposed experiments have broad implications for understanding the nature of learning itself (what is learned and what processes are involved in learning) as well as for the specific area of second language learning. There is a critical need for language-learning research that can improve second language teaching. Perhaps less obvious are the implications for computer-assisted language learning and speech remediation. Understanding how adult's perceptions are altered by linguistic experience will likely shed light on how meaningful speech perception can be regained after brain damage (e.g. in aphasia) or learned in developmental language disorders. In particular, the work may steer strategies of remediation to become more individually based.