Musical Forces: Motion, Metaphor, and Meaning in Music

MUSICAL FORCES

This book is a publication of

INDIANA UNIVERSITY PRESS

601 North Morton Street

Bloomington, IN 47404–3797 USA

iupress.indiana.edu

Telephone orders 800–842–6796

Fax orders 812–855–7931

© 2012 by Steve Larson

All rights reserved

No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage and retrieval system, without permission in writing from the publisher. The Association of American University Presses’ Resolution on Permissions constitutes the only exception to this prohibition.

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ANSI Z39.48–1992.

Manufactured in the

United States of America

Library of Congress

Cataloging-in-Publication Data

Larson, Steve, [date]

Musical forces : motion, metaphor, and meaning in music / Steve Larson ; foreword by Robert S. Hatten.

p. cm. – (Musical meaning and interpretation)

Includes bibliographical references and index.

ISBN 978-0-253-35682-6 (cloth : alk. paper) – ISBN 978-0-253-00549-6 (e-book) 1. Music – Physiological aspects. 2. Music – Physiological effect. 3. Music – Psychological aspects. I. Title.

ML3820.L36 2012

781’.1 – dc23

2011024022

1 2 3 4 5    17 16 15 14 13 12

IN MEMORY OF ROBERT TROTTER

CONTENTS

FOREWORD

ROBERT S. HATTEN, SERIES EDITOR

Musical Forces is the culmination of over 25 years of speculation, research, and empirical inquiry into the ways we experience motion, and hence meaning, in music. Inspired by the work of Rudolf Arnheim on visual perception and Douglas Hofstadter on analogy, Steve Larson develops a theory of musical forces that affect our perception of both melody and rhythm, by analogy to our embodied (and cultural) understanding of physical forces.

Written for a wide audience, the book reflects Steve’s engagement with several often-overlapping scholarly and pedagogical communities. Cognitive scientists will find extensive empirical testing of Steve’s hypotheses concerning our metaphorical understanding of musical forces in terms of our experiencing of corresponding physical forces, along with a generous sampling of basic musical examples. Students will find a careful sequencing of concepts and applications, with frequent summaries and cross-referencing, and a helpful glossary. Experienced musicians will find more sophisticated music examples demonstrating Steve’s sensitivity to tonal musical styles, ranging from classical to jazz. Speculative music theorists will appreciate Steve’s careful balancing of musical intuitions and hard science, and his encyclopedic knowledge of the relevant scientific, music theoretical, and music historical literature on movement, metaphor, and meaning. And future researchers will find a suggestive list of recommendations for further exploration within the paradigm Steve has so artfully constructed.

In the fall of 2010 Steve learned he had a brain tumor, and although he had completed the entire manuscript, he did not live to see this book appear. I am thus especially grateful to music editor Jane Behnken, her able assistant Sarah Wyatt Swanson, June Silay, and the production staff at Indiana University Press for their diligent work preparing the manuscript for publication. Jodi Jolley compiled the index. As series editor, I responded to the copyeditor’s queries after consulting on the phone with Steve as much as was then possible, and in late summer I gave the entire manuscript a close proofreading in his stead. Not only were Steve’s ideas fully formed, he also expressed them with his inimitable voice, wit, and charm—all of which you will find preserved here.

Those of us who were privileged to have known Steve will remember his courage and serenity throughout his last months. He was a dear friend and colleague to many, and we will miss him terribly. This book may serve as a fitting memorial to his visionary ideas, his love of music, and his passion for sharing both.

PREFACE

This book describes a theory of musical forces and some of the evidence for that theory. Thus it is aimed primarily at professional music theorists. I hope, however, that it may interest other types of readers as well – including cognitive scientists (especially those with interests in psychology and expectation, the phenomenology and aesthetics of music, computer science and artificial intelligence, and cognitive linguistics), musicians from other sub-disciplines of music (especially history, ethnomusicology, music education, jazz, and popular music), and general readers with an interest in music.

Because it is written for such different audiences, members of each will inevitably find some portions too difficult or maybe even painfully simple. I hope the overview in chapter 1 will help readers find those portions of greatest interest (and to identify passages they may wish to skip). I have also included a glossary at the end of the book, partly as a way to summarize and review its basic concepts but also to help those readers who find its terms unfamiliar (or who find that my use of those terms differs from what they are used to or would have expected).

The theory of musical forces concerns the experience of certain listeners: those listeners of tonal music who have internalized the regularities of common-practice tonal music to a degree that allows them to experience the expectations generated by that music (the music of Bach to Brahms and much jazz and popular music). When this book talks about we or us, it is talking about members of that musical culture.

This book assumes that its readers are familiar with musical notation and have some grasp of the fundamentals of harmony. (Readers who have not learned harmony by taking piano lessons or an undergraduate course on this topic may wish to consult a book on music fundamentals such as Victor Zuckerkandl’s Sense of Music [1959].) While I hope that this volume will be of interest to professional music theorists, much of what I say does not require a prior understanding of the technical vocabulary of advanced music theory.

ACKNOWLEDGMENTS

First things first: thank you, Sonja Rasmussen, for your love and support – I have incorporated a number of your valuable substantive suggestions, and your encouragement has really helped me to put this all into a final form.

My interests in motion, metaphor, and meaning go back to my years as a student at the University of Oregon, where my 1981 master’s thesis, written under the supervision of Robert Trotter and Peter Bergquist, explored the role of metaphor in musical meaning by offering my own metaphorical descriptions of pieces by the group Oregon and then asking how music analysis could illuminate the types of motion described.

At the same time Marion Guck was doing her seminal work on metaphor at the University of Michigan. And when, just a few years later, I encountered her work – and her encouragement to pursue my own – I decided to dig deeper into the topic. Marion’s work continues to be an inspiration and model for my own.

This book began to take shape more than twenty-five years ago, as a term paper for a course at the University of Michigan. That course, called Psychology of Comparative Arts, was taught by Rudolf Arnheim. During his office hours we had many conversations about how his ideas on Gestalt psychology and the perception of art might be applied to music. My correspondence and conversations with him after completing that course continued to inspire my work. In fact, this book may be regarded as an elaboration of his 1986 article Perceptual Dynamics in Musical Expression.

At the same time I had the great fortune to work as Graduate Research Assistant to Robert Hatten, exploring issues of music and meaning in twentieth-century opera. He and many of my other teachers (especially Richmond Browne, Robert Hurwitz, and William Rothstein) may notice ideas they shared with me long ago. I hope they will forgive my inability at this point to say where some of those ideas came from.

Although there are very few specific citations of the work of Alexandra Pierce, her approach to musical meaning through movement has long been an inspiration to my thinking in and thinking about music.

The Center for Research on Concepts and Cognition (CRCC) at Indiana University, under the direction of Douglas Hofstadter, supported research that started turning ideas about musical forces into conference presentations, published articles, and computer models. (I was a Visiting Faculty Research Associate at the Center from 1992 to 1993 and again from 2002 to 2003.) My colleagues at CRCC (especially Doug Hofstadter, Gary McGraw, and James Marshall) offered thoughtful ideas on music and considerable help with computer programming. Doug read several chapters of this book and offered many helpful suggestions. He has also provided support in myriad other ways.

I am deeply grateful to Hallgjerd Aksnes, who invited me to serve as Visiting Faculty Research Associate for her project Music, Motion, and Emotion: Theoretical and Psychological Implications of Musical Embodiment at the Institutt for Musikkvitenskap of the University of Oslo in 2010. That appointment provided time and support for the final stages of the book and gave me the pleasure of working with her and members of her research project (Anita Høyvik, Ingvild Amundsen, Mari Haugen, and Ragnhild Solberg), whose feedback enriched this book and helped shape a planned sequel (on musical forces and music analysis).

I have had numerous opportunities to share these ideas with students in the courses I have taught. These ideas have shaped, and have been shaped by, my teaching of courses on musicianship skills and music analysis (in fact, I first came up with the patterns described in chapter 6 while teaching aural skills in the late seventies). Temple University provided a large number of such courses. But some of the most fruitful of these courses were taught at the University of Oregon. One course, a seminar titled Music and Meaning, which I team-taught with my colleague Mark Johnson, led to our article, Something in the Way She Moves – Metaphors of Musical Motion (Johnson and Larson 2003), which appears here as chapter 3 (and as a chapter in Johnson 2007). Part of that article also appears in chapter 2 (in a section on conceptual metaphor theory). Thank you, Mark, for allowing me to include our article in this book. Mark and I are indebted to Arnie Cox for his help with that article; we also learned a great deal about the metaphorical understanding of musical experience by working with him on his doctoral dissertation. We also thank our colleague Scott Pratt for his helpful comments on an earlier draft of that article. Mark also read several chapters of this book and offered suggestions that improved them all. Students in another seminar taught at the University of Oregon in 2004, called Musical Forces, read a nearly complete draft of this book. The students in that course (Darin Hoskisson, Kalin Kirilov, Kaori Noland, Jennifer Russell, Keith Salley, LeeAnn Sterling, and Jamie Webster) all provided thoughtful feedback.

Another colleague at the University of Oregon, Leigh VanHandel, coauthored an article with me called Measuring Musical Forces, parts of which appear throughout this book. Her skills as experimental psychologist, computer programmer, and writer were very helpful indeed. Thanks Leigh!

Thanks also to Fred Lerdahl, who heard one of my first public presentations of these ideas in 1993 and has offered suggestions and encouragement on this work in various stages since then. At that same 1993 presentation, Carol Krumhansl and Jamshed Bharucha noted the similarity between my computer’s algorithm for the interaction of musical forces and the general formula for multiple regression, and encouraged me to learn about this method of statistical analysis.

A grant from the University of Oregon provided time in the summer of 1999 to help finish computer models and articles based on them. Robin High and Joe St. Sauver of Statistical Consulting at the University of Oregon helped with the statistical analyses.

Bill Lake supplied me with a copy of his dissertation and offered help and advice. Erick Carballo helped translate Lake’s transcriptions of his participant responses from musical notation into integers that the computer could read.

My computer models were written in MacScheme. Scheme is a dialect of LISP invented by Guy L. Steele Jr. and Gerald Jay Sussman. Scheme for the Macintosh was developed at the Massachusetts Institute of Technology (© 1984). MacScheme was published by Lightship Software, Inc. (© 1992).

Thanks also to David Butler, Richard Cohn, Anne Danielsen, Steven Demorest, Gina Fatone, Cynthia Folio, Thomas Goolsby, Robert Hatten, Robert Hurwitz, Bunny Laden, Patrick McCreless, Anne Dhu McLucas, John Rahn, Sonja Rasmussen, Lee Rothfarb, William Rothstein, Caitlin Snyder, Steven Strunk, and Keith Waters for reading chapter drafts (or drafts of articles that became chapters in this book).

I also appreciate the encouragement that so many friends and colleagues have provided for this project. To give just one example, when I mentioned that I suspected there might be some relationship between the pattern map of Example 6.9 and the motives that make up hidden repetitions, Janet Schmalfeldt suggested I write the journal article (Larson 1997–98a) that became the starting point for some of the repertoire-based types of evidence presented in part 2.

Because parts of this book were originally published as free-standing articles or book chapters, I also extend my thanks to the patient and helpful suggestions of several editors – Thomas Demske and Ramon Satyendra (Journal of Music Theory), Mary Wennerstrom (Journal of Music Theory Pedagogy), Raymond Gibbs (Metaphor and Symbol), Anthony Gritten and Elaine King (Music and Gesture), Lola Cuddy and Mari Riess Jones (Music Perception), and Taylor Greer (Theory and Practice) – as well as other, anonymous readers for those journals and books.

The book publication process gave me encouragement and helpful suggestions from a number of readers (most of them anonymous). Of these readers, the comments of David Temperley were particularly valuable. I also thank the staff at Indiana University Press, including copyeditor Rita Bernhard.

Finally, I thank my parents, not only for their love and support in other ways but also for providing, in their cottage on the Oregon beach, many opportunities to write and think in an inspiring and idyllic location.

MUSICAL FORCES

1

Introduction

Sometimes we ask "How does that melody go?" At times we might say that a melody "moves by steps and leaps." Or we might talk about melodies "ascending and descending." In fact, it is hard to think of words for describing physical motion that have not been applied to musical motion. So this book’s first question might be, "Why do we talk about music as if it actually moved?"

Sometimes music can make us laugh, cry, or want to dance. At times it might induce us to see or remember colorful shapes and images, to recall painful or pleasant episodes, or to experience all kinds of different feelings. In fact, music and emotion have long been described as intimately linked. So this book’s second question (to rephrase our first question only slightly and to shift from motion to emotion) might be, "Why does music actually move us?"

This book gives partial answers to both questions by showing some ways in which physical motion influences our experience of classically tonal music (music of the Bach-to-Brahms era as well as much jazz and popular music) and contributes to its meaning. (As noted in the preface, the theory of musical forces concerns the experience of certain listeners: those listeners of tonal music who have internalized the regularities of common-practice tonal music to a degree that allows them to experience the expectations generated by that music. When this book talks about we or us, it is talking about members of that musical culture.)

The central argument of this book is that our experience of physical motion shapes our experience of musical motion in specific and quantifiable ways – so that we not only speak about music as if it were shaped by musical analogs of physical gravity, magnetism, and inertia, but we also actually experience it in terms of musical forces. Part 1 (chapters 2–7) describes this theory of musical forces, and part 2 (chapters 8–13) summarizes the substantial (and growing) body of evidence that supports it.

Of course, writers have long discussed music in terms of the metaphors of motion and forces.¹ But the theory of musical forces presented in this book takes five further steps. Although none of these is entirely unprecedented, their combination is novel. First, the book identifies and rigorously defines three melodic forces (melodic gravity is the tendency of notes above a reference platform to descend; melodic magnetism is the tendency of unstable notes to move to the closest stable pitch, a tendency that grows stronger as the goal pitch is closer; and musical inertia is the tendency of pitches or durations, or both, to continue in the pattern perceived) and two rhythmic forces (rhythmic gravity and metric magnetism; these are discussed in chapter 6, where the rhythmic aspects of musical inertia are also explored in greater depth).² Second, the book embraces the metaphorical status of those musical forces as central to, explanatory for, and constitutive of both our discourse about music and our experience of music. In other words, although these forces may feel to us as though they are inherent in the sounds themselves, they are actually a creation of our minds, where they are attributed (consciously or unconsciously) to the music – shaping our thinking about music and our thinking in music – so that they become a part of what may be called musical meaning. Third, the book explicitly grounds the operation of those musical forces in the theories of Heinrich Schenker. Fourth, it shows that the musical forces provide necessary and sufficient conditions for giving a detailed and quantifiable account of a variety of musical behaviors (for example, the regularities in the distributions of musical patterns in compositions and improvisations, and the responses of participants in psychological experiments). Fifth, it finds converging evidence for the cognitive reality of musical forces from a variety of practical and experimental sources.

In order to clarify the claims of this book, it may also be helpful to say what I am not arguing. I do not claim that the account given here completely explains the roles of musical forces in our experience of music. I do not claim that musical forces completely explain musical experience. I do not claim that every melodic motion results from giving in to musical forces. I do not claim that musical forces have the same universality or natural status that physical forces do. I do not claim that such forces are an important part of every culture’s music. I do not claim that gravity, magnetism, and inertia are the only forces that shape melodic expectations. I do not claim that musical forces and musical motion are the only metaphors that inform music discourse and musical experience. And I do not claim that the ideas of pattern, analogy, and metaphor offered in this book give a complete account of human meaning-making.

A THEORY OF EXPRESSIVE MEANING

I suspect, however, that readers will want to think about how the idea of musical forces might contribute to a larger theory of expressive meaning – one that explains more thoroughly how and why music moves us the way it does. Therefore, in this introduction, I sketch the larger theory of expressive meaning in music that guides my thinking about musical forces. I do this in order to lay my cards on the table, as well as to contextualize the arguments of this book. But I do not claim that mine is the only possible theory of meaning that would encompass the idea of musical forces.³ I do not suggest that this book provides a complete description of that theory of expressive meaning. Nor do I believe that this book provides a complete defense of the value or the truth of that larger theory. Although the claims of this book may be seen as supporting that larger theory, they should be understood as claims about musical forces rather than claims about such a larger theory of expressive meaning in music. In other words, although this book makes claims about musical forces that could be understood as part of a larger theory of expressive meaning in music, I am not going to defend that larger theory – but I do want readers to see how the idea of musical forces might fit into such a larger theory. The theory of expressive meaning in music toward which my earlier writings point makes the following claim:

It is useful to regard part of what we call expressive meaning in music as an emergent property of metaphorical musical forces.

The remainder of this chapter clarifies what I mean by that statement – elaborating on the terms expressive meaning, emergent property, and metaphor – and it offers an overview of how some of these ideas inform the remainder of the book. Along the way it will be useful to respond to some of the objections that have been raised to this theory of expressive meaning.

Much has been written about emotion and meaning in music. In a book with that title, Leonard Meyer (1956) wrote that music excites emotions when it departs from what we expect. Others (e.g., Coker 1972; Cooke 1959; Davies 1994; Ferguson 1960; Kivy 1989; and Robinson 1997) have written about whether music can express emotions (or anything, for that matter), about whether music can express or simply be expressive of emotions, and about how musical meaning (if it exists) might differ from other types of meaning.

The theory of expressive meaning simply assumes that music has expressive meaning – that quality we experience in music that allows it to suggest (for example) feelings, actions, or motion. This quality may not translate well into words, nor relate clearly to the emotions felt by the creators (the composers, performers, and improvisers) of the music, but it seems to be one reason why we derive so much pleasure from listening to music. In other words, I use this term in the same way that Robert Hatten (1994, 2004) does and in the same way that Rudolf Arnheim (1974) uses the term expression.

Like Meyer, I also find an intimate connection between meaning and expectation. My theory further argues, however, that expressive meaning in music arises from the specific ways in which music moves when it denies – or confirms – our expectations. And I argue that we understand those specific ways of moving, in part, by using our knowledge of analogous physical motions – and the forces that shape those physical motions.

To illustrate some of the ways in which music can move us, and to help clarify the theory of expressive meaning in music, Examples 1.1 and 1.2 offer two types of musical examples. Example 1.1 gives the music for Dido’s Lament from Purcell’s Dido and Aeneas – an example of a lamento bass. Example 1.2 (p. 16) gives a few excerpts that include what I call the hallelujah figure).

EXAMPLE 1.1. Dido’s Lament from Purcell’s Dido and Aeneas.

TWO LAMENTO BASSES

In Dido’s Lament, Dido sings of how she would like to be remembered, just before she takes her own life. The music seems to capture the meaning of its text very well. It is based on a repeated bass line, the so-called lamento bass, which descends chromatically from tonic to dominant, in a minor key, and in a slow triple meter. A similar repeated bass line occurs in the Crucifixus from Bach’s B-minor Mass – which tells of Christ’s death. The lamento bass has a long history of association with texts expressing sadness and death (Williams 1997).

Why have these (and other) lamento basses appealed to their composers as an excellent way to set such texts? One answer to this question might be caricatured as rote learning of conventions. According to this argument, we first hear Dido’s Lament, then notice that the words are sad, and finally come to associate that sadness with that piece and its bass line. Later, when we hear another lamento bass, we think something like Oh, this sounds like ‘Dido’s Lament’ . . . so it must be sad, too. Although I do not deny that learning such associations is possible, I am convinced that this view of musical meaning leaves out something important.

CONVENTION AND CULTURE

One problem with assuming that all musical meaning relies solely on rote learning of conventions is that it seems to assume that the musical material is entirely arbitrary – that it is purely conventional. However, if the musical material were really arbitrary, then any association between material and meaning would be possible.

One reaction to this argument concerning associations between material and meaning is to believe that musical association is culturally determined – symbolic – and that is why just any association is not possible.⁴ This reaction is emblematic of a line of thinking that is counterproductive to the project of this book, so I examine it here at some length.

Note, first of all, that such a reaction makes an important point: culture does play a central role in helping to shape the associations we make between material and meaning. In fact, although the specifics of their claims have proven controversial (or at least thought, by some, in need of qualification), ethnomusicologists have noted relationships between culture, meaning, and musical material that seem to go beyond rote learning of conventions. Judith Becker (1981) has shown interesting connections between Javanese conceptions of time and the organization of their gamelan music. Steven Feld (1981, 1988) has discussed relations between the music of the Kaluli people and their stories about native soundscapes, as well as their habits of conversation. Alan Lomax ([1968] 2000) has suggested that, in folk songs from all around the world, correlations between attitudes about sex or authority seem to be reflected in aspects of a culture’s music (such as its vocal tension or textural organization). If these authors are right (and even if their claims need qualification), then culture does play an important role in shaping musical material and meaning.

At first glance, this reaction (association is culturally determined – symbolic – and that is why just any association is not possible) also seems to agree with at least part of the idea I am arguing for here: that it is not possible to make just any association between musical material and musical meaning. Nevertheless, this reaction implies that culture is the sole determinant of such associations ("association is culturally determined – symbolic – and that is why just any association is not possible), and it describes those associations as symbolic," a term from semiotics implying that the association is arbitrary (more about this term in a moment).

THE SINGLE-MECHANISM FALLACY

Let us begin with three responses to the idea that associations between musical material and musical meaning could be determined solely by culture. These three responses concern single-mechanism explanations, the complexity of culture, and the logic of this reaction.

First, I am skeptical that anything as complicated as associations between musical material and musical meaning could be determined solely by any single mechanism – even one as complex as culture. Nevertheless, it is an interesting aspect of our minds (which seem to have a drive toward simple explanations) that they tend to seek causality in terms of a single mechanism. This idea (that our minds prefer to attribute meanings to single mechanisms even though mechanisms, including mental processes, tend to be multiple) is a recurrent theme in this book. For example, chapter 11 (Evidence from Music-Theoretical Misunderstandings) will re-examine single-mechanism explanations by noting that some prior writings about musical forces make an understandable but mistaken assumption: that only one force (whether physical or musical) operates at a time.

Second (or perhaps this just puts the same argument in a different way), if we understand culture in such a way that it is reasonable to regard it as the sole determinant of such associations, then culture must itself be a very complicated entity; that is, in this view, culture must surely include things like embodiment, body image, and body schema – factors that, as this book argues, reflect nonarbitrary relations between material and meaning.

Third, the logic of the reaction (and that is why) is not sound. Even if culture were the sole determinant of associations between musical material and musical meaning, it does not automatically follow that culture would therefore forge such associations arbitrarily. The reaction seems to suggest that if culture were the sole determinant of such associations, then the nature of the material would necessarily be irrelevant – that the nature of the material would not help determine such associations. The reaction implies that for any association that one culture might make between material and meaning, another culture could make apparently opposite associations – and that nothing about the musical material itself would make this improbable.

LEARNING, STATISTICS, AND INTERNAL REPRESENTATION

The reaction we are discussing raises an important practical question: If associations are arbitrary, then how do we learn them? In other words, if we assume that the association of the lamento bass with sadness (especially the sadness of death) is solely culturally determined – symbolic, then we might next ask, How did that association come about, both for individual listeners and for our musical culture as a whole?

The belief that members of a culture can learn anything simply through repeated exposure is widespread. In an important and recent book, David Huron (2006) notes that auditory learning is dominated by statistical exposure (72) and that listeners appear to be sensitive to the frequencies of occurrence of different auditory events (73). Huron’s work provides ample evidence that musical expectations are shaped by, and tend to reflect, the frequencies with which musical events occur.

Once again, if one were prone to what I am calling the single-mechanism fallacy, one might assume that we expect what we expect solely because it happens more often – and then dismiss any investigation into what those expectations mean. I do not level this criticism at Huron’s work. I just caution that the combination of statistical learning and the single-mechanism fallacy might discourage the kind of thinking essential to the goals of this book.

Huron’s work with Paul von Hippel on post-skip reversal nicely illustrates the point I want to make here. Von Hippel and Huron (2000) studied a sample of traditional European folk songs, Chinese folk songs, South African folk songs, and Native American songs. They found that the majority of larger intervals were followed by a change in directions (what they called a post-skip reversal).

The results seem consistent with the age-old compositional advice that leaps (especially large leaps) should be followed by steps in the opposite direction. Their further analysis showed, however, that the results could best be explained not in terms of post-skip reversal but rather as regression to the mean. In other words, statistically, large upward leaps that took melodies to the upper part of a melodic range tended to descend because they were thus in the upper part of the melodic range; large upward leaps from low pitches that took melodies into the middle part of a melodic range do not show a consistent tendency to be followed by reversals in direction.

But the punch line is that von Hippel’s subsequent experimental studies showed that (despite the statistical facts) the expectations of trained listeners are better explained in terms of a preference for post-skip reversal (rather than an expectation for regression to the mean). In other words, experienced listeners seem to have learned the wrong lesson from their statistical learning. My point here is not to argue with the notion that we are sensitive to frequencies of occurrence of musical events (there is plenty of evidence that we are). Rather, I suggest that meaning and expectation involve more than memorizing the frequency with which events have occurred (and while Huron is interested in what we expect and how we came to expect it, this book asks about the relationship between those expectations and our sense of meaning). As Huron notes, how minds represent music has repercussions for what listeners remember, what listeners judge to be similar, and other musically important functions (2006, 73).

Let us return, then, to the question of how we learn to associate the lamento bass with sadness – and the role of mental representation and meaning in that learning. Each lamento bass is at least a little bit different from other lamento basses, and each may be embedded in a different musical context. If presented with fifty-seven different passages containing lamento basses, do listeners have to learn – separately – that each one implies an affect of sadness? How do they learn that association without using the words lamento bass and sadness, and without someone else pointing out how each new passage shares those features (including its slow tempo, its descending contour, its triple meter, its chromatic motion, the tonic and dominant end points it connects) that we think of as making up a lamento bass – especially when each new passage will likely contain features not necessarily contained in the other passages, some of which may be thought of as essential to a lamento bass, some of which may be thought of as specific to the passage but not essential to describing it as a lamento bass, and some of which may also appear in passages having a contradictory affect? If presented with a fifty-eighth passage of music that also has a (different) lamento bass, how would the listener know that it also was a passage of music implying a sad affect?

Eventually the listener would have to develop the concept of a lamento bass as an abstract concept having a set of potentially shared features (some of which may not be present in every example). Without the ability to create such abstract categories, listeners would not recognize the fifty-ninth passage of music with those same features as a lamento bass, and would thus (according to the quoted reaction) be unable to attribute the associated affect to it.

Notice that the problem does not go away if we simply move to a lower level. One might claim that listeners are not primarily learning to associate the lamento bass with sadness but that they are learning to associate sadness with the lower-level features that make up the lamento bass (such as slow tempo, descending contour, chromatic motion). This relocates the problem but does not solve it. We would still need to explain how listeners recognize those lower-level features as categories (with their own still-lower-level features).

Furthermore, a single musical culture’s associations between material and affect seem to form rich and nuanced webs of meaning. Members of any one musical culture share numerous such associations. We know enough about cognition to see that it is much easier for a group of people to learn the same or similar associations if they are nonarbitrary. Given the large number of associations built up by a single musical culture, it seems highly improbable that all those associations could be arbitrary.

The point I am getting at here is that we have (or develop) the ability to create such abstract categories. That ability is necessary in order to forge the associations that the quoted reaction attributes solely to the black box of culture. Listeners within a culture cannot learn to associate a specific affect with a specific musical relationship in any deep or consistent way until they can recognize that specific musical relationship as a representative of an abstract category – a category with its own feature