We present the detailed analysis of a patient, GL, who presents auditory atonalia as a consequence of brain damage. GL was found to be unable to use tonal knowledge in the interpretation of melodic closure, in discrimination as well as in preference of melodies. This breakdown in the tonal representation of melodic patterns occurred in the presence of accurate encoding of melodic contour and, to some extent, of interval sizes. It also occurred in isolation from disturbances in the processing of temporal information. Thus, the pattern argues against a general deficit in melodic organisation processes but, rather, argues for a specific loss of access to tonal knowledge. These data are discussed with regard to the independence and organisation of tonal knowledge within the music processing system and as a function of the current classification schema underlying musical disorders.
INTRODUCTIONTonal knowledge refers to the system of mental representations that specify the set of pitches that can be used and the rules governing their combination in pieces written in our Western musical idiom. This implicit knowledge is considered to be the product of a shared experience that is acquired very early in the ontogenic development without explicit tutoring. Tonal knowledge is essential for organising most music, by helping memory, by allowing the building of expectancies, and by shaping singing performance. This range of mental processes in which tonal knowledge plays a central role is even larger in musicians, who exploit it in reading, in performing, and in composing musical sequences.Accordingly, impairments of this knowledge may constitute a severe handicap for understanding and responding to the musical environment. Requests for repMts should be addressed to Isabelle Peretz. Dcpartment de psychologit. Univerdtc dt Montreal, C.P. 6128, succ. A, Montrcal (at.), H3C 317, Canada. This research was supported by a research grant and fellowship from the National science and Engineering R d Council of Canada. I wish to thank Christiane Cyr-StaEord for facilitating access to GL. List qagnon and Sylvie Htbert for helping to construct test materials and for collecting normal data, and Karalyn Patterson as well as John Sloboda for insightful editorial suggestions. I am especially grateful to GL for his preciouS collaboration.