Ray Dougherty

Ray C. Dougherty, Ph.D. is an Associate Professor of Linguistics at New York University. Professor Dougherty completed the MIT Ph.D. program in Electrical Engineering studying the mathematical communication theories of Shannon, Turing, and von Neumann. He transferred to the Linguistics Department finishing a Ph.D. under Noam Chomsky on coordinate structures and recursion in language (1969). Digital Signal Processing (1982) used coding theories to design micro-circuitry for speech recognition and synthesis. Natural Language Computing (1994) offered a Chomsky grammar in Lisp and Prolog. Working with Konrad Lorenz (1976, 1980) led to an interest in formalizing animal sensori-motor capacities. Recent work with Stephen Wolfram on Cellular Automata (see A New Kind of Science) forms the basis for a rather complete 3D computational graphics model of the operation and evolution of the modern human inner ear, and provides tools to compare the cochlea of lizards (disk), birds (cylinder), and mammals (triple concentric helical spiral). Since the cochlea-brain connection is an analog-digital converter, and sound complexities have clear Fourier/Poincare representations in mathematics, one can offer precise descriptions of how human sound processing capacities evolved over hundreds of millions of years, and account for anomalies like the mammalian platypus whose inner ear resembles that of the bird kingdom. Studying the computational universe of conceivable sound complexities and geometrically possible inner ears, one can show persuasively that Gould’s ‘saltations’ occurred at least twice in the evolution of the modern human ear from ancient creatures, one converting a cylinder (bird) cochlea to a helix (mammal), and also in the appearance of three parallel rows of hairs in the mammal cochlea. Most linguists focus on the evolution of the vocal tract, our group focusses mainly on the cochlea. The mouth serves dozens of different functions, the cochlea since its ancient birth, only detects vibrations and converts them to neural pulses. The computations required to analyze the cochlea parallel closely those required to design a sampling music synthesizer, and many students have worked on projects involving sampling synthesizers.