Tone deafness: failures of musical anticipation AI\D self-reference

« Tone deafness » is a term that tends to be applied indiscriminately to a constellation of musical processing, perceptual, and production deficits. It has been estimated that 3 - 10% of the population are tone deaf (Cox, 1947; Joyner, 1968). Tone deafness can result from organic trauma (in which case the term "amusia" is often applied) or from some as-yet-unknown combination of genetic, neurochemical and environmental (i.e., learning) factors. Although the medical term "amusia" was first applied over a century ago (Edgren, 1895) the various forms of the syndrome have not been systematically classified. Preliminary research suggests that at least some of those individuals who are labeled as tone deaf lack the cognitive structures necessary to anticipate musical tonality and harmony; or lack internal self-referencing tonal schema within which to understand, process, and remember musical material. In this paper, I propose a taxonomic system for classifying the various forms of tone deafness, as a precis to new empirical research. First, I propose that tone deafness can be grouped according to four different deficits: production deficits, perceptual deficits, memory deficits, and deficits in symbolic manipulation (either music reading or writing). Among the medically documented deficits is a condition that parallels "pure word deafness" in which the subject can perceive sound, but is unable to recognize any musical or melodic qualities. Within each of these deficit families, I propose a number of specific deficits with varying causes, and describe tests that can determine the nature of a given individual's deficits. The present work, in addition to providing a formal theoretical structure within which to think about musical processing, also has relevance for artificial intelligence researchers attempting to model human cognition and music processing. Specifically, the study of individuals with tone deafness presents us with a window into the neural mechanisms of musical processing, revealing evidence for which sub-processes might be modular, and which might be interlinked