Absolute pitch: perception, coding, and controversies Daniel J. Levitin and Susan E. Rogers Department of Psychology and Center for Interdisciplinary Research in Music, Media and Technology, McGill University, Montreal, Quebec, Canada Recent findings in cognitive neuroscience and cognitive psychology are converging to shed light on the nature of processing, categorization and memory for pitch in humans and animals. Although most people are unable to name or place pitch values in consistent, well-defined categories, as they do for color, stable long-term memory for pitch has been shown in certain animal species, in infants, and in both adult musicians and non- musicians. ‘Absolute pitch’, the rare ability to label pitches without external reference, appears to require acquisition early in life, and involves specialized brain mechanisms, now partially identified. Research on pitch coding strategies informs wider theories in cognitive science of semantic memory, and the nature of percep- tual categories. Introduction The study of the memory codes used for musical pitch is one of the oldest in experimental psychology, beginning with Helmholtz, Fechner, Stumpf and Wundt. Indeed, the Gestalt psychology movement was launched with the following question: how is it that a melody composed of specific musical pitches retains its identity despite transposition and when none of the original pitches are present? A related question is that of why some people are able to label all the notes of the musical scale as effortlessly as most of us label colors – a phenomenon known as absolute pitch (AP). Recent findings from neuroimaging, psychophysics, developmental psychology and cognitive science are converging to create a critical mass of knowledge on which to build new theories and experiments. Research on absolute pitch has grown exponentially over the past 120 years. Defined as the ability either to identify the chroma (pitch class) of a tone presented in isolation or to produce a specified pitch without external reference [1–3], AP occurs in 1 in 10 000 people [2]. People with AP presumably possess an internal template to map musical tones to linguistic labels. Sometimes regarded as a mark of musicianship, AP is in fact largely irrelevant to most musical tasks. Being unable to turn it off, many possessors of AP perform dramatically poorer at judging whether a melody and its transposed counterpart are the same, a task that non-AP musicians accomplish with ease [4,5]. Comparisons are often made between color labelling in most humans and pitch labelling in AP possessors [2], because AP possessors categorize and label pitches quite effortlessly and automatically. However, the human visual system is constructed in such a way as to allow discrete categories to emerge readily: information about color is separated into three (or sometimes four) streams by cones in the retina [6], and remains separated up to the cortex. By contrast, information from the cochlea and peripheral auditory system is much more continuous and lacks the one-to-one mapping between excitation patterns and percepts [7]. Accordingly, pitch and color perception are phenomenologically different: colors are experienced as belonging to categories; pitches are experienced (by most of us) as continuous. The fact that some individuals place pitches into categories requires an explanation of what is different about these individuals in their development or neural architecture. Absolute pitch versus relative pitch AP should not be confused with relative pitch (RP), an ability all trained musicians learn that allows them to identify or produce musical ‘intervals’, or relations between pitches (see Box 1). A trained musician presented with the tones A and C will identify the musical interval as a ‘minor third’ without being able to name either component tone. Told that the first note was an A, she will use her learned knowledge of musical scales to report that the second note was a C. Interestingly, if we tell the musician that the first note was G, she will report that the second note was a Bb – the note a minor third above G – and not know that she had been deceived. The absolute labels for notes are typically not attended to by RP possessors except when performing music from a score. Indeed, most musical training in the classical, jazz and popular traditions emphasizes the ability to play musical patterns and scales equally well in all keys. In general, RP information is used to recognize and produce melodies, but some residual AP information might also be coded (see Box 2). This issue relates to a longstanding debate in the animal literature about the extent to which learning and internal representations are relational or absolute, and whether the function of memory is to preserve the details of experience or rather, to form abstractions and preserve the gist of experience [8]. Corresponding author: Levitin, D.J. (daniel.levitin@mcgill.ca). Available online 8 December 2004 Review TRENDS in Cognitive Sciences Vol.9 No.1 January 2005 www.sciencedirect.com 1364-6613/$ - see front matter Q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2004.11.007