You are on page 1of 306

NORTHWESTERN UNIVERSITY

A Theory of Analogy for Musical Sense-Making and Categorization:


Understanding Musical Jabberwocky

A DISSERTATION

SUBMITTED TO THE GRADUATE SCHOOL


IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

for the degree

DOCTOR OF PHILOSOPHY

Field of Music

in the Program in Music Theory and Cognition

By

Janet Eileen Bourne

EVANSTON, ILLINOIS

August 2015
2

© Copyright by Janet Bourne 2015

All Rights Reserved


3

ABSTRACT

A Theory of Analogy for Musical Sense-Making and Categorization:

Understanding Musical Jabberwocky

Janet Bourne

This dissertation argues that listeners use analogy (or relational similarity) to make sense of

unexpected or unusual musical events and categorize musical patterns by connecting music to

similar patterns from past experiences. Using Structure-Mapping Theory as a basis, I present an

interdisciplinary framework for analogical listening: retrieval (past experiences), mapping (associating

element of one structure with a corresponding element in a different structure), and evaluation

(inferences based on the event mapped over in context). Humans make analogies from music to

music or music to another domain (between music and language, and so on). I use the framework to

help explain how music theorists perceive musical irony and categorize thematic variations. First, I

use the framework to analyze Beethoven string quartets that scholars consider ironic (op. 95/iv, op.

131/V, op. 130/I) since perceiving musical irony relies partly on a music and language analogy.

Using empirical studies on ironic language along with formal function, sonata theory, and schema

theory, I argue why some theorists hear these movements as ironic. Second, this dissertation argues

that analogy is used to categorize musical themes; themes are relational categories and not perceptual

categories. To demonstrate that a theme is a relational category, I use the analogy framework to

analyze theme and variation movements by Beethoven (op. 109/III), Mozart (K 377/II), and Daube

(from Musical Dilettante). This project responds partly to an “ideal listener” assumption; I consider

how contexts shape perception—particularly, music theorists connecting relations of patterns and

inferences.
4

Acknowledgements

I would not have been able to write this dissertation without the help of many people.

First, I want to thank my advisor, Richard Ashley, for everything that he has done for me.

Ric has inspired me, supported me and believed in me my entire time at Northwestern. In his music

cognition class my first year, it felt as if I had come home. He has an amazing encyclopedic

knowledge of the field that has helped guide me over the years to interesting sources that I would

not have found otherwise. His ideas have shaped this project, always pushing me to think critically

about the ideas here. I want to thank him for his patience, encouragement and care for my well-

being over the years. He saw me as not just a graduate student, but a person, and I am infinitely

thankful for that.

My committee members Robert Gjerdingen, Vasili Byros and Dedre Gentner have not only

shaped this project, but my development as a scholar.

I am grateful to Bob for introducing me to schema theory. Ever since the galant class my

first year, schema theory has completely re-wired the way I think about music (and for the better).

He has not only supported me, but been an incredible source of academic knowledge and practical

advice. Not unlike a student of partimenti, I have looked to him as a model for creating engaging

conference presentations and writing beautiful paragraphs. I thank him for the copious amount of

time he has spent working with me. I have learned much about the field, music, radio dramas, and

musicals.

I first starting writing about irony because of Vasili’s Beethoven seminar my first year. Since

then, he has been a fountain of knowledge on music-theoretic matters. He has challenged me to be

more analytically minded in my scholarship, steering me back towards music analysis when I have
5

needed it most. I thank him for generously spending time with me and giving me honest advice on

professional matters. I am grateful for everything he has done for me.

I am indebted to Dedre for introducing me to the wonders of analogy. Because of her

analogy and similarity class, I felt like I could finally understand my own thinking. Since then, I have

grown more and more obsessed with analogy. I am so grateful to her for introducing me to these

ideas which play such an important role in this dissertation. I am also grateful for the time she has

spent working with me considering she is an invaluable resource on experimental design.

At Northwestern University, I have had a wonderful support group. I want to thank all the

members of the Music Theory and Cognition Program (past and present) who have helped shape

my ideas: Ji-Chul Kim, Jung-Nyo Kim, Ben Anderson, Ben Duane, Karen Chan Barrett, Matt

Gilmore, Melissa Murphy, James Symons, Cora Palfy, Olga Sanchez-Kisielewska, Kristina Knowles,

Bruno Alcade, Rosa Abrahams, Stephen Hudson, and Miriam Piilonen. I have been grateful not only

for their encouragement, but also for the chance to get to know them and hear their perspectives. I

also want to thank NU’s Graduate Christian Fellowship and the wonderful friends and support I

have found there over these past few years. With prayers and love, they have sat through mock job

talks and mock teaching demonstrations as well as talked with me about ideas in the dissertation. I

want to especially thank the Tuesday night small group (Lauren Sturdy, Jonathan Syrigos, Vlad

Serban, Sylvia Wright, Megan Mascarenhas, Caitlin Makatura, Bryan Van Scoy, and Vienet Romero)

for letting my prayer request always be “my dissertation.” I also want to thank everyone in my

interdisciplinary writing group: Faith Kares, Andrew Owens, Aaron Greenberg, Alex Lindgren-

Gibson, Kate Dugan, and Jaimie Morse. In particular, I want to thank our fearless leader Elizabeth

Lenaghan, who brought us all together. It has been wonderful this year to learn about non-

governmental organizations in the Philippines, Dark Shadows, early modern vitalism, prayer in
6

Catholicism, British nonelites in India, and legality of medical examinations. Without their gracious

comments and feedback, the writing in this dissertation would have been considerably worse.

Outside of Northwestern, my family and friends have been supportive of me through the

years. My parents, Jim and Leota Bourne, have always encouraged me and believed in me. I would

not be here today if it weren’t for them. I am so grateful for all the sacrifices they have made for me

over the years. I hope I have made them proud. I thank my brothers, Jeremy and Davy, who have

supported me and endured hearing me talk about the dissertation. I also want to thank Carrie Crow,

who helped my writing through advice and reading suggestions. Finally, I want to thank the people

at Faith EPC for their prayers and thoughts.

Finally, I want to thank my husband, Allen. With infinite patience, he has supported me

throughout this entire process. And, without him, I am not sure where I would be. He celebrated

with me when I was happy, and held me when I cried. He always believed in me, even when I didn’t

believe in myself. For that, I want to thank him. This dissertation is as much his as it is mine. I could

never imagine finding someone I love more dearly, and I thank God every day that He brought

Allen in my life. Although no words express my love for him, I want to include the following quote

for him.

CAPT. MALCOLM REYNOLDS: But it ain't all buttons and charts, little albatross. You
know what the first rule of flying is? Well, I suppose you do, since you already know what
I'm about to say.
RIVER TAM: I do. But I like to hear you say it.
CAPT. MALCOLM REYNOLDS: Love. You can learn all the math in the 'Verse, but you
take a boat in the air that you don't love, she'll shake you off just as sure as the turning of the
worlds. Love keeps her in the air when she oughta fall down, tells you she's hurtin' 'fore she
keens. Makes her a home.
RIVER TAM: Storm's getting worse.
CAPT. MALCOLM REYNOLDS: We'll pass through it soon enough.

I close this acknowledgements section the way J.S. Bach closed many of his pieces: SDG.
7

Table of Contents
CHAPTER 1. “Somehow it seems to fill my head with ideas—only I don’t exactly
know what they are!” Who is Listening and what do They Hear?” ………………. 21

 Analogy ………………………………………………..…………..………….. 23

 Research Questions and Hypotheses …………..…………..………………….. 26

 Interdisciplinary Audience: Music Theory and Music Cognition ………………. 32

 Methods and Evidence ….…………..…………..…………..…………………. 34

 Organization …………..…………..…………..…………..…………………… 36

CHAPTER 2. “We have learned to think about the relatedness of things”: Analogy,
Music Theory, and Music Cognition ..…………..…………..…………..……………… 39

 Definitions of Analogy ……………..…………..…………..…………..…………. 41

 Analogical Processes: Retrieval, Mapping, and Evaluation …………..……………. 44

 Mapping: Gentner’s Structure-Mapping Theory …………..…………..………….. 46

 Structural Alignment …………..…………..…………..………………….. 47

 Two Forms of Comparison: Projective Analogy and Mutual Alignment


51
Analogy …………..…………..…………..…………..…………..………..

 Retrieval: The “Analogical Paradox” …………..…………..…………..………….. 52

 Evaluation …………..…………..…………..…………..…………..…………….. 53

 Holyoak and Thagard’s Multiconstraint Theory …………..…………..………….. 54

 Analogy and Similarity …………..…………..…………..…………..……………. 56

 Analogy and Relational Categories …………..…………..…………..……………. 58


8

 Approaches to Analogy in Music …………..…………..…………..……………… 63

 Music Theory …………..…………..…………..…………..……………... 63

 Music Cognition …………..…………..…………..…………..………….. 64

 Similarity and Categorization in Music Theory …………..…………..……………. 65

 Meyer’s Conformant Relationships …………..…………..……………….. 66

 Music-Theoretic Patterns …………..…………..…………..…………….. 66

 Thematic Variations …………..…………..…………..…………… 67

 Meyer and Gjerdingen’s Schemata …………..…………..………… 72

 Hanninen’s Associative Sets …………..…………..………………. 76

 Similarity and Categorization in Music Cognition …………..…………..………….. 78

 Pollard-Gott (1983) …………..…………..…………..…………..……….. 78

 Bigand (1990) …………..…………..…………..…………..……………… 79

 Lamont and Dibben (2001) …………..………...…………..……………… 80

 Ziv and Eitan (2007) …………..…………..…………..…………………... 81

 Eitan and Granot (2009) …………..…………..…………..………………. 82

 A Common Theme …………..…………..…………..…………..……….. 82

 Conclusion …………..…………..…………..…………..…………..…………….. 83

CHAPTER 3. “A Knowledge about a thing is a knowledge of its relations”: An


Analogy Framework for Music Analysis …………..…………..…………….………….. 85

 Musical Event …………..…………..…………..…………..…………..………….. 86

 The Analogy Framework …………..…………..…………..…………..………….. 86


9

 Step 1: Retrieval …………..…………..…………..…………..…………………… 87

 Musical Use/Exposure …………..…………..…………..………………... 90

 Exposure to Specific Musical Pieces …………..…………..………. 92

 Conventional Musical Patterns …………..…………..……………. 93

 Comparing Listeners …………..…………..…………..…………..………. 94

 Step 2: Mapping …………..…………..…………..…………..…………………… 94

 Musical Relations …………..…………..…………..…………..………….. 95

 Two Levels of Musical Relations: Relations within a Musical Event


and Relations between Musical Events …………..……………….. 99

 Relations within a Musical Event …………..…………..………….. 100

 Relations between Musical Events …………..…………..………… 117

 Two Forms of Analogical Comparison: Unidirectional Listening and


Bidirectional Listening …………..…………..…………..………………… 122

 Four Forms of Analogical Distance in Music: Intra-opus, Inter-Opus,


Intertextual, and Extra-musical …………..…………..…………..………... 123

 Intra-opus Analogy …………..…………..…………..…………… 124

 Inter-opus Analogy …………..…………..…………..…………… 127

 Intertextual …………..…………..…………..…………..………. 129

 Extra-musical Analogy …………..…………..…………..………… 130

 Step 3: Evaluation …………..…………..…………..…………..…………………. 131

 Inferences …………..…………..…………..…………..…………………. 131

 Context …………..…………..…………..…………..…………..……….. 132

 Goals …………..…………..…………..…………..…………..………….. 134


10

 Retrieval, Mapping, and Evaluation: Pattern Mapping Taxonomy …………..…… 135

 Pattern that Belongs …………..…………..…………..…………………. 137

 Pattern that Does Not Belong …………..…………..…………..……….. 138

 Pattern that Belongs, but Manipulated …………..…………..…………… 140

 Benefits of Analogy Framework …………..…………..…………..……………… 145

 Conclusion: The Analogy Framework and Research Questions …………………… 149

CHAPTER 4. “No Interpretation could Palliate this Error of a Genius”: Using


Analogy to Perceive Irony in Beethoven’s String Quartets …………..………………. 151

 What is Irony? …………..…………..…………..…………..…………………….. 153

 Perceiving Irony in Language …………..…………..…………..…………………. 154

 Music and Language Analogy …………..…………..…………..…………………. 156

 Perceiving Irony in Music using Analogy …………..…………..………………….. 158

 Retrieval: Violation of Expectation …………..…………..………………... 161

 Mapping: Flouting the Gricean Maxims …………..…………..…………… 164

 Mapping: Extra-musical Analogy …………..…………..………….. 166

 Mapping: Relations between Musical Events …………..………….. 166

 Case Studies …………..…………..…………..…………..…………..…………… 166

 Op. 95/iv (1810) …………..…………..…………..…………..………….. 168

 Op. 131/V (1826) …………..…………..…………..…………..………… 181

 Op. 130/I (1825) …………..…………..…………..…………..………….. 187

 Analyzing Listeners …………..…………..…………..…………..……………… 201


11

 My Own Coda …………..…………..…………..…………..…………..………… 205

CHAPTER 5. “Appears Clothed in a New Manner”: Analogy and Hearing


Thematic Variations …………..…………..…………..…………..…………..…………. 207

 A Variations Vignette …………..…………..…………..…………..……………… 208

 Thematic Variations as Relational Categories: Melodies Clothed in a New Manner .. 210

 Analogy Framework and Analyzing Variations…………..…………..…………… 216

 Mapping: Intra-opus Analogy …………..…………..…………..…………. 217

 Mapping: Relations within a Musical Event …………..…………..……… 217

 Mapping: Comparison and Bidirectional Listening …………..……………. 218

 Variations from the Eighteenth to Nineteenth-Century: From Embellishing to


Reinterpreting …………..…………..…………..…………..…………..………….. 221

 Relational Categories: Thematic Categories and Galant Schemata ………………… 225

 Beethoven’s Piano Sonata, Op. 109/III (1820) …………..…………..……………. 227

 Mozart’s Violin Sonata, K. 377/ii (1781) …………..…………..…………..……… 244

 Daube’s Variation from The Musical Dilettante (1773) …………..…………..……… 257

 Conclusion …………..…………..…………..…………..…………..……………. 267

CHAPTER 6. “You’re sure to get somewhere, if you only walk long enough”:
Future Research …..………….…..………….…..………….…..………….…..……… 269

 Dissertation Summary …..………….…..………….…..………….…..…………. 269

 Future Research …..………….…..………….…..………….…..………….…..… 271

 An Empirical Investigation into Analogy for Categorizing Themes …..…… 271


12

 Stimuli …..………….…..………….…..………….…..…………. 272

 Experimental Design, Hypotheses and Procedure …..………….… 272

 Predictions …..………….…..………….…..………….…..……… 276

 Follow-Up Studies …..………….…..………….…..………….….. 277

 Connection between Schemata and Topics: Relational Categories,


Perceptual Categories, and Construction Grammar …..………….…..…… 277

 How Modern Listeners Hear Common Practice Music through Film Music. 278

 Analogy as a Window to Musical Communication …..………….…..……………. 280

 Conclusion …..………….…..………….…..………….…..………….…..……… 281

WORKS CITED …..………….…..………….…..………….…..………….…..……… 282


13

List of Figures

Figure 1.1.: Diagram of the solar system and atom analogy (graphic from Corral and Jones

2012, 1434) 25

Figure 1.2: Mozart, 12 Variations on “Ah, vous dirai-je maman,” K. 265/300e, mm. 1-8,

25-32 28

Figure 1.3: The basics of the analogy framework 35

Figure 2.1: Disciplines and research questions 40

Figure 2.2: Other disciplines and research questions 41

Figure 2.3: Major components of analogical reasoning (graphic from Holyoak 2012, 236) 44

Figure 2.4: Circles steadily decreasing in size (graphic from Gentner and Smith 2012, 131) 49

Figure 2.5: Causal scenes containing a cross-mapping. The woman in the top scene is

receiving food, while the woman in the bottom scene is giving food (graphic form

Markman and Gentner 1993, 436) 50

Figure 2.6: Schematic of two forms of comparison in analogy (graphic from Jee et al.

2010, 3) 52

Figure 2.7: Visual representation of similarity space (graphic from Gentner and Smith

2013, 10) 57

Figure 2.8: Schematic of the category “visit” (graphic from Goldwater, Markman, and

Stilwell 2011, 360) 60

Figure 2.9: Contextual subtypes list, re-constructed from Hanninen (2012, 37-38) 77

Figure 3.1: The analogy framework 87

Figure 3.2: Sections and subsections of the retrieval step 89


14

Figure 3.3: Analogy as Structure-Mapping (graphic from Gentner and Smith 2012, 132) 95

Figure 3.4: Bach, Art of Fugue, Contrapuntus VII, mm. 2-4 98

Figure 3.5: Visual representation of two levels of musical relations: relations between

musical events and relations within a musical event 99

Figure 3.6: Graphic representation of relations within a musical event 104

Figure 3.7: Higher-order relation thematic category 105

Figure 3.8: Revised graphic representation of relations within a musical event 105

Figure 3.9: Proportional rhythms 106

Figure 3.10: Showing rhythm as a relation within a musical event; Clementi, Sonatina Op.

36, no. 1, m. 5 and m. 6 respectively 107

Figure 3.11: (a) Ordinal duration vector key, (b) Clementi rhythm—length approach 108

Figure 3.12: Clementi rhythm—proportional approach 109

Figure 3.13: Dittersdorf, Symphony in C, K. 1, first movement, mm. 1-4 110

Figure 3.14: Example of Aprile from Aprile, Soleggio, MS fol. 40v, Larghetto, mm. 1-4

(graphic from Gjerdingen 2007a, 123) 111

Figure 3.15: Graphical representation of scale degree relation 111

Figure 3.16: Scale degrees 112

Figure 3.17: Clementi, Sonatina in C Major, Op. 36, No. 1, first movement, mm. 1-19 113

Figure 3.18: Dittersdorf, Symphony in C, K. 1, first movement, mm. 1-4, and outline of

contour relation 114

Figure 3.19: Mozart, Sonata in C Major, K. 545, first movement, mm. 1-4, and outline of

contour relation 114

Figure 3.20: Contour 115


15

Figure 3.21: Dittersdorf, Symphony in C, K. 1, first movement, m. 1-4 117

Figure 3.22: Harmony 117

Figure 3.23: Correlation (graphic from Hatten 1994, 37) 118

Figure 3.24: Major vs. minor onto tragic vs. nontragic (graphic from Hatten 1994, 38) 119

Figure 3.25: Mapping of minor vs. major onto tragic vs. nontragic 119

Figure 3.26: Mozart, K. 332, first movement, mm. 1-26 120

Figure 3.27: Mapping of singing style vs. horn fifths onto indoors vs. outdoors 121

Figure 3.28: (a) Beethoven, Piano Sonata in F Minor, Op. 2, No. 1, first movement, mm.

1-8, (b) Beethoven, String Quartet in F, Op. 135, third movement, mm. 3-10

(graphics from Caplin 1998, 10) 122

Figure 3.29: Visual representation of analogical distance in music 124

Figure 3.30: Mozart, “Adagio” from Piano Concerto in A major, K. 488, second

movement, mm. 1-12 126

Figure 3.31: Prinner schema (graphic from Gjerdingen 2007a, 455) 128

Figure 3.32: (a) Wodiczka, Op. 1, No. 3, first movement, adagio, mm. 1-3 (graphic from

Gjerdingen 2007a, 46), (b) L’Abbé, Op. 8, No. 1, first movement, mm. 10-14

(graphic from Gjerdingen 2007a, 55) 129

Figure 3.33: (a) Mozart, “Miserocordias Domini,” mm. 23-26, (b) Beethoven, Symphony

No. 9, fourth movement, mm. 92-99 130

Figure 3.34: Beethoven, Piano Sonata No. 26 in Eb major, Op. 81a, first movement,

“Lebewohl,” mm. 1-5 133

Figure 3.35: Visual Representation of the Pattern Mapping Taxonomy 137

Figure 3.36: Prinner schema (graphic from Gjerdingen 2007a, 455) 138
16

Figure 3.37: Stravinsky, op. 7, etude no. 1, mm. 1-6 139

Figure 3.38: Beginning of “Batman’s Theme,” Elfman, Batman (1989) 139

Figure 3.39: Visual representation of types of pattern manipulation 141

Figure 3.40: Beethoven, Piano Sonata No. 26 in Eb major, Op. 81a, first movement,

“Lebewohl,” mm. 1-5 141

Figure 3.41: Quotation of Wagner in Debussy’s “Golliwog’s Cakewalk” (graphic from

Everett 2009, 32) 142

Figure 3.42: Beginning of Idée Fixe in Berlioz, Symphonie Fantastique, fifth movement 143

Figure 4.1: Grice’s maxims, a component of the linguist H.P. Grice’s Cooperative

Principle 156

Figure 4.2: Linguistic slots that map onto musical slots (graphic from London 1996, 52) 157

Figure 4.3: The Analogy Framework 159

Figure 4.4: The Retrieval Step 160

Figure 4.5: The Mapping Step 160

Figure 4.6: The generic layout of Sonata Form (graphic from Hepokoski and Darcy 2006,

17) 164

Figure 4.7: Beethoven, Op. 95 (Serioso), fourth movement, mm. 126-153 169

Figure 4.8: Beethoven, Op. 132, finale, mm. 290-309 170

Figure 4.9: The analogy framework and Op. 95, relation maxim, form 172

Figure 4.10: The analogy framework and Op. 95, relation maxim, genre 174

Figure 4.11: Graph of Sonata-Rondo form and placement of EEC and ESCs (graphic

from Hepokoski and Darcy 2006, 428) 177

Figure 4.12: The Indugio from Gjerdingen (2007a, 464) 178


17

Figure 4.13: Beethoven, Op. 95 (Serioso), fourth movement, mm. 118-20 179

Figure 4.14: The analogy framework and Op. 95, quality maxim 180

Figure 4.15: Beethoven, Op. 131, fifth movement, mm. 13-48 182

Figure 4.16: Beethoven, Op. 131, fifth movement, mm. 41-48 183

Figure 4.17: The analogy framework and Op. 131 185

Figure 4.18: Beethoven, Op. 130, first movement, mm. 13-28 189

Figure 4.19: The Meyer schema from Gjerdingen (2007a); the Aprile variant has a closing

dyad of 2-1 instead of 4-3 in the soprano 191

Figure 4.20: Example of the Aprile variant (with a not as common 5-1 close in the bass

instead of 7-1) from Aprile, Solfeggio, MS fol. 40v, Larghetto, mm. 1-4 (graphic

from Gjerdingen 2007a, 123) 191

Figure 4.21: The Aprile schema in Beethoven, Op. 130, first movement, mm. 13-19 192

Figure 4.22: Beethoven, Op. 130, first movement, mm. 47-58 193

Figure 4.23: Beethoven, Op. 130, first movement, mm. 104-132 196

Figure 4.24: The analogy framework and Op. 130 199

Figure 4.25: The analogy framework and Op. 95, a modern listener 202

Figure 4.26: The analogy framework and Op. 130, an eighteenth-/early-nineteenth-century

listener or a modern theorist 203

Figure 4.27: The analogy framework and Op. 130, a modern listener 204

Figure 5.1: Autograph of Beethoven’s 1793 letter to Eleonore von Breuning (Beethoven

1793) 209

Figure 5.2: Beethoven, Symphony No. 5, opening 214


18

Figure 5.3: Visual representation of similarity space (graphic from Gentner and Smith

2013, 10) 215

Figure 5.4: The Analogy Framework 216

Figure 5.5: A thematic category as a relational category (or schema) that accounts for

theme and its variations 218

Figure 5.6: Major components of analogical reasoning (graphic from Holyoak 2012, 236) 225

Figure 5.7: Locatelli, Op. 2, various opening basses (graphic from Gjerdingen 2007a, 14) 226

Figure 5.8: Beethoven, Op. 109, third movement, theme, mm. 1-16 230

Figure 5.9: Op. 109 thematic category, mm. 1-4 231

Figure 5.10: Key for notating “lengths approach” in rhythm 232

Figure 5.11: Venn diagram rom Beethoven Op. 109 234

Figure 5.12: Beethoven, Op. 109, third movement, var. 6, mm. 1-4, score and box

notation 235

Figure 5.13: Beethoven, Op. 109, third movement, theme, mm. 1-4, score and box

notation 236

Figure 5.14: Beethoven, Op. 109, third movement, var. 1, mm. 1-4, score and box

notation 237

Figure 5.15: Beethoven, Op. 109, third movement, var. 3, mm. 1-5, score and box

notation 238

Figure 5.16: Beethoven, Op. 109, comparing theme and var. 1 239

Figure 5.17: Beethoven, Op. 109, third movement, var. 4, mm. 1-5, score and box

notation 241
19

Figure 5.18: The order of variations in op. 109. The shaded yellow sections are relations

shared with the thematic category. The shaded red sections are ambiguous

whether the relations are shared with the thematic category. 243

Figure 5.19: Mozart, K. 377, second movement, theme, mm. 1-16 245

Figure 5.20: Mozart, K. 377, second movement, theme, mm. 17-32 246

Figure 5.21: Mozart, K. 377, second movement, theme, mm. 1-4, grouping structure 246

Figure 5.22: Mozart, K. 377, thematic category, box notation 247

Figure 5.23: Venn diagram for Mozart K. 377 248

Figure 5.24: Mozart, K. 377, theme, box notation 249

Figure 5.25: Mozart, K. 377, second movement, var. 3, mm. 1-4, box notation and score 250

Figure 5.26: Mozart, K. 377, second movement, var. 5, mm. 1-4, score and box notation 251

Figure 5.27: Mozart, K. 377, second movement, var. 6, mm. 1-4, score and box notation 252

Figure 5.28: Mozart, K. 377, second movement, var. 2, mm. 1-4, box notation and score 253

Figure 5.29: Mozart, K. 377, comparing var. 3 and var. 4 254

Figure 5.30: The order of variations in K. 377. The shaded yellow sections are relations

shared with the thematic category. 256

Figure 5.31: Variation of the first chord without pasting tones (graphic from Daube [1773]

1992) 258

Figure 5.32: Daube’s theme (graphic from Daube [1773] 1992, 139-140) 259

Figure 5.33: Daube’s theme, mm. 1-10 (graphic from Daube [1773] 1992, 139) 260

Figure 5.34: Do-Re-Mi schema (graphic from Gjerdingen 2007a, 650) 261

Figure 5.35: Prinner schema (graphic from Gjerdingen 2007a, 646) 261

Figure 5.36: Cudworth’s cadence galante (graphic from Gjerdingen 2007a, 147) 262
20

Figure 5.37: Daube, thematic category, box notation 263

Figure 5.38: Daube, theme, mm. 1-5 (graphic from Daube [1773] 1992, 139) 264

Figure 5.39: Daube, potential thematic category after hearing only the theme 264

Figure 5.40: Daube, var. 1, mm. 1-10 (graphic from Daube [1773] 1992, 140) 265

Figure 5.41: Daube, var. 2, mm. 1-8 (graphic from Daube [1773] 1992, 141) 265

Figure 5.42: Daube, beginnings of (a) var. 3, (b) var. 4 (c) var. 5, (d) var. 6, (e) var. 7

(graphic from Daube [1773] 1992, 141-142) 266

Figure 6.1: Adjective pairs for excerpt descriptions (graphic from Lamont and Dibben

2001, 253) 274

Figure 6.2: Schematic of experimental design 275


21

CHAPTER 1
“Somehow it seems to fill my head with ideas—only I don’t exactly
know what they are!” Who is Listening and what do They Hear?

‘Twas brillig, and the slothy toves


Did gyre and gimble in the wabe;
All mimsy were the borogoves,
And the momre raths outgrabe
—Lewis Carroll, Alice in Wonderland and Through the Looking-Glass and
What Alice Found There

After reading the (in)famous Jabberwocky poem, Alice of Alice in Wonderland said:

“It seems very pretty,” she said when she had finished it [the Jabberwocky poem], “but it’s
rather hard to understand!” (You see she didn’t like to confess, even to herself, that she
couldn’t make it out at all.) “Somehow it seems to fill my head with ideas – only I don’t
exactly know what they are! However, somebody killed something: that’s clear at any rate – ”.
(Carroll 1865, 165–66, italics in original)

The poem is filled with “Jabberwocky” sentences, which are sentences with “proper” syntax, but

nonsense words. Understanding music is analogous to understanding the Jabberwocky poem. At

times, music is hard to comprehend, yet somehow it fills our heads with ideas. How can a reader

infer meaning from a nonsensical poem such as the Jabberwocky? How can a listener infer meaning

from music? To paraphrase Alice, “somebody heard something, at any rate.”

As a listener parses incoming information into categories (musical events), he or she learns

associations and expectations. A listener’s exposure to repeating musical patterns may be enough to

create a category in his or her mind (a mental representation). Important to categorization is the

ability to relate musical patterns to each other, either by relating a current pattern to one in long-

term memory or to another pattern recently heard. Yet, music is composed of multiple dimensions.
22

For instance, a musical melody can take shape either as a waltz or a march, but still be the same

melody. On the other side of the coin, two different waltzes can sound similar to each other even

though they involve different pitches and rhythmic content. How could theorists understand, from a

cognitive perspective, these relationships in music? One idea may be that theorists—and listeners in

general—use analogy, a cognitive process, to understand musical relationships.

Similarities of relationships between one musical pattern to other musical patterns—or

analogies—is crucial to musical communication, but it has received less attention as a way to

understand musical meaning. Instead, theorists favor analogies between music and other domains

(e.g. “this melody is like falling snow”) to understand music as communicative (e.g. Hatten 2004;

Zbikowski 2002), some making analogies between music and drama (e.g. Maus 1988; Levinson

2004) or music and narrative (e.g. Almén 2008). Approaches to musical communication have been

varied, from the collection by David Hargreaves, Raymond MacDonald and Dorothy Miell (2005)

which approaches the topic from a variety of disciplines, to Danuta Mirka and Kofi Agawu’s (2008)

investigation limited to eighteenth-century music. Yet, Northup Frye (1957) wrote that “poetry can

only be made out of other poems; novels out of other novels” (97); by extension, music can only be

made out of other musics, and listeners make sense of music through relating musical patterns.

When music theorists do analyze musical patterns, for musical communication or other

reasons, they often discuss clear-cut examples that are not distorted. Some examples are

prototypical, while some are atypical or distorted. These “distorted” examples are members of that

pattern, simply “exploited.” Leonard Meyer (1980) describes Beethoven exploiting musical

conventions: “Beethoven overturned no fundamental syntactic rules. Rather, he was an

incomparable strategist who exploited limits—the rules, forms, and conventions that he inherited

from predecessors such as Haydn and Mozart, Handel and Bach—in richly inventive and strikingly
23

original ways. In so doing, Beethoven extended the means of the Classic Style” (Meyer 1980, 190,

italics in original). These distorted patterns are when patterns “go wrong”—they are on the fringe of

what a pattern should be, but still that pattern. Still, a listener is “supposed to” recognize these

patterns, despite distortion and appearance in unusual contexts. Analogy provides a framework for

relating musical patterns to others in memory for both unconventional and conventional patterns. It

also offers a means for theorists to discuss how these categories, even atypical exemplars, may be

perceptually salient depending upon a listener’s use and exposure to similar previous music. Once

the listener has identified a pattern, he or she can interpret it within a piece’s context. In the pages to

come, I explain how analogy helps answer the following question: When listening, what do listeners

do to make sense of a sequence of musical events, especially when “things go wrong”?

Analogy

Analogical reasoning1 is a powerful, domain-general process—a pillar of psychological

thought, even claimed to be the “core of cognition” (Hofstadter 2001), and the fundamental process

by which people understand their surroundings (e.g. Gentner, Holyoak, and Kokinov 2001; Holyoak

and Thagard 1995). An analogy involves “recognizing a common relational system between two

situations [or events] and generating further inferences guided by these commonalities” (Gentner

and Smith 2013, 2–3). In an analogy, one domain or situation—called the base/source—is more

familiar, better understood or more concrete, while the other domain or situation—called the

target—is less familiar, less understood and more abstract (Gentner and Smith 2012, 130). By “better

understood,” Holyoak (2012) explains that, “the reasoner has prior knowledge about functional

relations within the source analog—beliefs that certain aspects of the sources have causal,

1 I use terms analogy, analogical processing, and analogical reasoning interchangeably.


24

explanatory, or logical connections to other aspects” (234, italics in original). Humans retrieve, or

remember, relational information concerning a source and transfer that information to understand a

target.

According to Jerry Fodor (1983), this transfer from source to target underlies creation of

new scientific theories.

It really does look as though there have been frequent examples in the history of science
where the structure of theories in a new subject area has been borrowed from, or at least
suggested by, theories in situ in some quite different domain: what’s known about the flow of
water gets borrowed to model the flow of electricity; what’s known about the structure of
the solar system gets borrowed to model the structure of the atom; what’s known about the
behavior of the markets gets borrowed to model the process of natural selection, which in
turns gets borrowed to model the shaping of operant responses. And so forth. The point
about all this is that “analogical reasoning” would seem to be isotropy in the purest form: a
process which depends precisely upon the transfer of information among cognitive domains
previously assumed to be mutually irrelevant. (107)

Through analogy, humans perceive one situation through the lens of another; both situations share

the same structure, but not the same appearance. Humans in Western society conceptualize

electricity (target)—an invisible concept—by thinking in terms of water (base/source)—a more

concrete concept. Using analogy, people infer electricity “behaves” like water; they expect electricity

to “flow” and have a “current.” It is not because water and electricity look similar on the surface,

but because they share similar relations. The Oxford English Dictionary defines “relation” as: “an

attribute denoting or concept expressing a connection, correspondence, or contrast between

different things; a particular way in which one thing or idea is connected or associated with another

or others” (“Relation, N.” 2015). Shared relations (how objects connect to each other), rather than

perceptual features (how objects look on the surface), determine whether two situations are an

analogy. An “analog” is a “situation or domain involved in analogical mapping, either the base

(source) or target” (Gentner and Smith 2012, 130). Gentner and Smith (2012) argue that, for

analogy, “what is crucial is similarity in relational structure” (2–3, italics in original).


25

This dissertation’s psychological model of analogy is cognitive psychologist Dedre Gentner’s

Structure-Mapping Theory (Gentner 1983; Gentner 2003). In Structure-Mapping Theory, people

employ analogies through a “comparison process” that aligns two situations (generally, a base and a

target) to reveal common relational structure (Gentner 2003, 107). Humans reason analogically in

three steps: retrieval, mapping, and evaluation. Consider the analogy “atom is like the solar system.”

In the first step, retrieval, people retrieve their knowledge about solar systems and atoms. In the

second step, mapping, they use comparison to align this knowledge so to highlight and map over

relational structure; in this case, one possible relation is smaller objects revolve around a larger one (figure

1.1).

Figure 1.1.: Diagram of the solar system and atom analogy (graphic from Corral and Jones 2012, 1434)

Finally, during the third step of evaluation, listeners evaluate inferences from this mapping: such as,

“if smaller objects revolve around a larger one faster in solar systems, then they probably revolve

faster in atoms as well.” Analogies pervade everyday life, including science, mathematics, problem

solving, creative cognition, legal reasoning, and teaching (Holyoak 2012, 346).

If analogy shapes understanding in so many areas of everyday life, then it may shape

understanding in music as well. If analogy is the “core of cognition” (Hofstadter 2001), then it

follows logically that it is also the “core of music cognition.” This dissertation builds on the idea that

listeners make sense of music partly through analogy, a cognitive process that connects music to past
26

experiences. I argue that analogy can illuminate the cognitive processes behind music theorists’

understandings of music, as they relate music to familiar patterns or musical events. Structure-

Mapping Theory and its analogical steps (retrieval, mapping, and evaluation) will be shown to play a

part in musical thinking and reasoning.

Research Questions and Hypotheses

The psychologist Kevin Dunbar strives to understand how laboratory scientists make

discoveries, and the role analogy plays in this process. Dunbar (1995) used an “in vivo”

methodology to study analogies made “in the wild” by real world scientists in laboratories.2 He

began studying analogy “in the wild” after observing that subjects in controlled psychology

experiments failed to make the analogies he expected. He observed that, instead, “subjects in many

psychology experiments tend to focus on superficial features [surface, not relational, features] when

using analogy, whereas people in non-experimental contexts, such as politicians and scientists [“in

vivo” analogy], frequently use deeper more structural features” (Dunbar 2001, 313). Participants

often used perceptual features (what something looks like on the surface), rather than relations

(connections between objects). He labelled this phenomenon the “analogical paradox” (Dunbar

2001, 313): the fact that participants easily use analogies in naturalistic settings, but often fail to use

them in psychology experiments (at least, without hints or reminders).3 Thus began a part of

Dunbar’s research agenda: contrasting analogy in experimental settings with analogy in naturalistic

settings.

2 According to his analyses, analogical thinking is an important component of scientific reasoning, including hypothesis
generation, experimental design, explanations, and data interpretation (Dunbar 2001, 315).
3 Participants in Gick and Holyoak’s (1980) experiments solved a problem using analogy; yet, only after experimenters

reminded them of an analogous example that could be used to solve the problem.
27

If analogy is difficult for participants in experimental settings, then musical thinking rooted

in analogy should also be difficult for participants in experimental settings. Participants in music

cognition experiments may follow a similar pattern to that of participants in analogy experiments.

This observation motivates my first research question.

 Research Question 1: Research has established that listeners in experimental settings


do not categorize musical themes the way music theorists expect them to. Why is
that the case?

Music theorists assume that listeners abstract away perceptual features (e.g. dynamics, tempo,

etc.) to categorize musical themes based on structural features (e.g. pitch, rhythm, meter, etc.) (e.g.

Meyer 1973; Zbikowski 2002). In conflict with these assumptions, results from experiments on

musical similarity (e.g. Lamont and Dibben 2001) and categorization (e.g. Pollard-Gott 1983;

Koniari, Predazzer, and Mélen 2001; McAdams et al. 2004; Ziv and Eitan 2007) indicate participants

often categorize musical excerpts based on perceptual features instead. These findings have

perplexed many music cognition scholars, especially those—like music theorists—expecting

participants to categorize such musical events based on structure. After all, listeners in naturalistic

settings use structural features to categorize themes in various contexts, from a movie-goer

recognizing variations of a protagonist’s theme to a concert-goer noticing repeated sections of a

sonata. Why do participants not categorize themes the way music theorists think they should? To

categorize musical themes based on structural features, I argue that listeners in naturalistic settings

use analogy to map structurally important notes from one musical passages onto parallel notes in

another. In figure 1.2, a listener maps melodic notes between the theme in Mozart’s “Ah, vous dirai-

je maman” and the first variation to help him or her hear these phrases as the same theme. Yet, it is

not just the melody line that facilitates this mapping, but the bass as well, which stays the same up to

a mid-point of the variation.


28

Figure 1.2: Mozart, 12 Variations on “Ah, vous dirai-je maman,” K. 265/300e, mm. 1-8, 25-32

My research question, then, echoes Dunbar’s when he encountered the “analogical paradox”: why

do humans make analogies “in the wild” but not in experimental settings? I create the following

hypothesis for research question 1.

 Hypothesis for Research Question 1: The discrepancy of listeners not categorizing


musical themes in experimental settings is a consequence of listeners needing to use
analogy, or analogical processing, to categorize musical themes based on structural,
as opposed to perceptual, features.

I argue that listeners use analogy to categorize themes based on structural parameters since

these structural parameters—pitch, meter/rhythm, harmony—are often relational. Research

question 1 asks why listeners fail to use these relations to categorize musical themes in experimental

settings. Participants in analogy experiments struggle to make analogies since they often overlook

shared relations between two situations. In a similar way, participants in music cognition

experiments on categorizing musical themes overlook shared relations between two musical

passages. Yet, participants in analogy experiments notice relations—and so make an analogy—if

relations are highlighted and made more salient. The outcomes of several analogy experiments (e.g.

Gentner and Markman 1997) indicate that the act of comparison highlights relations, prompting

participants to notice them. If participants compare two objects or situations, then common

relations are more prominent. If participants in music cognition experiments compare musical
29

passages before categorizing, then they may also notice shared relations between musical passages,

prompting them to use analogy for categorizing musical themes based on structural—pitch,

meter/rhythm, harmony—over perceptual features (see chapter 2 for a review of similarity).

To introduce my second research question, I revisit Dunbar’s (1995; 2001) research on

analogies “in the wild”—specifically, when scientists use local over distant analogies and vice versa.

Depending on analogy’s purpose, distance between target and sources domains vary.

[The] distance between the source and the target may be large or small. For example, a
designer trying to develop door handles for the auto industry may make an analogy to other
door handles in the auto industry (within-domain, or local, analogies) or may make an
analogy to telephones or oysters in developing design (between-domain, or distant,
analogies)…. Local analogies involve greater superficial similarity between the source and the
target, as compared with the lesser amounts of superficial similarity involves in distant
analogies. (Christensen and Schunn 2007, 29)

Both local and distant analogies depend on relations; yet, local analogies also share perceptual

similarity. For example, furnace : coal :: woodstove : wood have closer semantic relations than furnace : coal

:: stomach : food, which have semantically distant relations (Vendetti, Wu, and Holyoak 2014, 929).

According to Vendetti, Wu, and Holyoak (2014), “whereas near analogies can be solved by matching

identical relations (e.g., the furnace burns coal as a woodstove burns wood), distant analogies require

evaluating or generating a more abstract relation that bridges the domains (e.g., the furnace burns

coal as the stomach “burns’ food)” (929).

While studying scientists in developmental biology and pathogen laboratories, Dunbar

(2001) found that the goal or reason for making the analogy influenced distance between source and

target (316). When scientists were trying to fix experimental problems, they often used local

analogies, which are analogies from highly similar domains (e.g. between HIV and HIV) or a

common superordinate category (e.g. between one virus and another virus) (Dunbar 2001, 316).

When scientists were reasoning about unexpected findings, though, not only was analogy the first
30

cognitive tool they used (Dunbar 2001, 317), but they often used distant analogies. Dunbar (2001)

writes:

Finding that more distant analogies are used following a set of unexpected findings is
interesting for two main reasons. First, analogy use changes with the current goal; a series of
unexpected findings act as a trigger for the wider search for sources and provide a new
mechanisms for expanding the search process. (318)

In addition to discussing musical analogy in experimental settings, I will explore musical analogy in

naturalistic settings: what musical relationships that rely on analogical thinking do music theorists

assume are important? What prompts music theorists to make distant or local analogies? As a first

step at “naturalistic” musical analogy, I analyze when music theorists (as listeners) use distant

analogies to make sense of music.

Local analogies in music are between two musical events (e.g. one musical theme to another

musical theme) while distant analogies are between music and a different domain (e.g. analogy

between music and language, music and narrative, and so on; henceforth called “extra-musical

analogies”). Local analogies use musical relations while distant—or extra-musical—analogies use

more abstract relations. Similar to Dunbar’s finding in scientists, listeners may use extra-musical

analogies for “unexpected findings” in music. When music “behaves” in a drastically unexpected

fashion and not much in the musical domain can explain this behavior, listeners may rely more on

extra-musical analogies to make sense of what they heard. This observation motivates my second

research question.

 Research Question 2: Music theorists perceive some unusually behaving musical


excerpts as ironic. Why do theorists use irony to understand these “unexpected”
findings in music analysis?

The Oxford English Dictionary defines “irony” as: “The expression of one’s meaning by using

language that normally signifies the opposite, typically for humorous or emphatic effect….A state of
31

affairs or an event that seems deliberately contrary to what was or might be expected; an outcome

cruelly, humorously, or strangely at adds with assumptions or expectations” (“Irony, N.” 2015). On

the face of it, irony seems an odd leap to make when interpreting music, especially since music lacks

denotative semantics propositional truth value (Cross 2005). Yet, several music theorists (e.g. Hatten

1994; Balter 2009; Zemach and Balter 2007) interpret certain Beethoven string quartet movements

(op. 95/iv, op. 130/I, and op. 131/V) as ironic, or something close to irony. When faced with these

examples, why do music theorists use “irony” to understand these “unexpected” moments in music?

Dunbar’s scientists made distant analogies when encountering unexpected findings; music

theorists—and certain listeners as well—may also make distant analogies when encountering

unexpected findings.

 Hypothesis for Research Question 2: Listeners use distant analogies to make sense
of “unexpected findings” in music—for example, a listener uses an analogy between
music and language to perceive irony in music as a way to make sense of certain
“unexpected findings” in music.

Research question 2 asks why listeners (and music theorists) use irony to make sense of

“musical behaviors” in certain Beethoven string quartets. I hypothesize that when music “behaves”

in an unexpected fashion—and not much in the musical domain can explain it—listeners, like

Dunbar’s scientists, could rely on distant or extra-musical analogies to make sense of these

“unexpected findings” in music. An “unexpected finding” triggers a wider search for sources to help

a person understand what the music could be communicating.4 As a case study, I analyze how some

listeners and theorists perceive irony in music. I hypothesize that perception of irony stems from an

analogy between music and language. Even though this music behaves unexpectedly, its behavior

4 Yet, the opposite is not necessarily true: when music behaves in an “expected fashion,” listeners are just as likely to use
extra-musical analogies. In my project, however, the interest is more on how listeners use extra-musical analogy as a tool
to make sense of unexpected instances in music, more so then on how listeners may use extra-musical analogies even
when the music is expected, as a way of description or interpretation.
32

imitates that of ironic language. Therefore, listeners/theorists may infer that musical passages

suggest irony if they are analogous to linguistics passages that also suggest irony. Not all

“unexpected findings” in music are ironic; yet, all instances of ironic music violate expectations, and

so are “unexpected findings.”

On the surface, these two questions seem unrelated. Yet, both ask how listeners use analogy

to understand music. I explore analogies that listeners make between musical themes as well as

analogies that listeners make between music and another domain (specifically, language). I attempt

to do for music theorists what Dunbar did for psychologists and scientists. There are several reasons

I chose to include both questions within the scope of one dissertation. First, a systematic theory of

analogy in music is novel and more examples provide more evidence for analogy in musical thinking.

Second, analogies between musical themes and analogies between music and another domain

demonstrates that it can happen on different levels in music, encouraging the reader to see various

potential for analogy as a method for analyzing music. Finally, this breadth allows me consider

analogy in experimental settings (more in line with music cognition) as well as naturalistic settings

(more in line with music theory). In the “big picture,” I hope to understand who the listener is and

why, from a cognitive perspective, they find certain interpretations of music or musical

communications intuitive and perceive musical patterns the way they do. As part of this, I seek to

understand why listeners make certain choices, categorizations and inferences while listening—and

the role analogy plays in these processes.

Interdisciplinary Audience: Music Theory and Music Cognition

I intend this dissertation to speak to two groups of scholars: music theorists and music

psychologists. I have two goals related to my audience and interdisciplinarity. As a first goal, I hope

this project continues to unite the disciplines or music theory and music cognition; in fact, I think
33

my research questions demand such a cross-disciplinary marriage. In my first research question, I

explain why results in music cognition do not follow music theory’s predictions. For music

cognition, then, I further research on this domain-general process and how it relates to music and

analysis. In my second research question, I use music cognition to partly explain music theorists’

intuitions about certain pieces. For music theory, then, I introduce the cognitive process of analogy

as an analytical tool to partly describe cognitive reasons behind music theorists’ intuitions on musical

relationships.

As a second goal, related to my music theory audience, I promote a type of music analysis

Temperley (1999) labels “descriptive.” He divides music theory into two schools, “suggestive” and

“descriptive.” The objective of “suggestive” theory and analysis is “to find and present new ways of

hearing pieces, not to describe the way people hear pieces already” (Temperley 1999, 70). Carl

Schachter and John Rahn’s approaches to analysis represent different sides of this “suggestive” type.

For Schachter (1976), analysis means finding the hidden structures of music: “of course the deeper

levels of structure, by definition, are not as readily accessible to direct perception as are events of the

foreground…. If they were, there would be no point to our analyzing music” (285–286). For Rahn

(1980), analyzing music “is to find a good way to hear it and to communicate that way of hearing it

to other people” (1). In “suggestive” analysis, the role of the theorist is to find imperceptible musical

structures or propose new readings or hearings of a piece.

In “descriptive” music analysis, on the other hand, the role of the theorist is to explain

listeners’ knowledge and to explain why they hear music a certain way. This type of music analysis:

Attempts to describe listeners’ unconscious mental representations of music… What is


sought is a general theory of some aspect of musical perception, a theory that describes
listeners’ general knowledge of music and the principles whereby they infer certain structures
from certain musical inputs…. Explaining why it is that a certain piece or certain musical
features in a piece, bring about a certain experience in the listener. (Temperley 1999, 68,
italics in original)
34

Fred Lerdahl, Ray Jackendoff and Leonard Meyer’s approaches to analysis represent this

“descriptive” type. For example, Lerdahl and Jackendoff (1983) insist that: “We take the goal of a

theory of music to be a formal description of the musical intuitions of a listener who is experienced in a musical

idiom” (1, italics in original). Meyer (1973), talking about music criticism, writes:

The primary goal of criticism is explanation for its own sake. Because music fascinates,
excites, and moves us, we want to explain, if only imperfectly, in what ways the events within
a particular composition are related to one another and how such relationships shape
musical experience. (17, italics in original)

This project—and my reasons for analyzing music—falls in the “descriptive” camp. By analyzing

music from a cognitive perspective, I analyze music in a way I think listeners may hear it. Therefore,

I promote “descriptive” music analysis. This stance helps me speak to my audience of both music

theorists and music cognition scholars.

Methods and Evidence

To address my research questions, I combine traditional music-theoretic tools with concepts

and methodologies from cognitive science and linguistics. To transfer analogy from cognitive

science to music analysis, I present an interdisciplinary framework for analogy during musical

listening (henceforth, the “analogy framework”) based on Gentner’s (1983) Structure-Mapping

Theory and the three analogy steps—1) retrieval, 2) mapping, and 3) evaluation. A theorist can use

the analogy framework (figure 1.3) in analysis by considering how a listener performs these steps in

regards to an analogy and (a) piece(s) of music.


35

Figure 1.3: The basics of the analogy framework

The retrieval step accounts for “who” is listening, or information a listener has from experience with

past music or another domain. In the mapping step, listeners associate each element of one structure

with a corresponding element in a different structure, noticing shared relations between two

domains. Listeners then evaluate; they evaluate inferences, recognize musical context, and consider

their own goals as well as composer and performer goals. Listeners make various inferences when

making analogies while listening, including affect (emotion), formal function (Caplin 1998), learned

associations, expectations (or implications) among others. Through the evaluation step, theorists can

help tie together meaning (or inferences) and musical structure.

I combine the analogy framework with more traditional music-theoretic methods of score

analysis, including thematic analysis (Meyer 1973), schema theory (Gjerdingen 2007a), sonata theory

(Hepokoski and Darcy 2006), formal function (Caplin 1998), and, to a lesser extent, associative sets

(Hanninen 2012). My use of score analysis for the first research question differs from my score

analysis for the second. For research question 1, I use score analysis to understand listener

perceptions in experimental settings. Even though I do not analyze stimuli from these experiments,

my analyses of different thematic variation movements are a direct response to results from these
36

experiments. For research question 2, I use score analysis to understand listener perceptions in

naturalistic settings.

In addition to the analogy framework and music-theoretic score analysis, I often quote

claims music theorists, historical figures and others make on their hearing or interpretation of a

musical passage or piece. I incorporate more of these quotations in my text than expected (or,

perhaps, even normal). Yet, my research investigates “the listener” and their modes of hearing. I

consider these quotations additional evidence for the arguments I make: that someone hears this

passage or piece in “that way.” For research question 2, for instance, several listeners (e.g. Hatten

1994; Longyear 1970; Balter 2009) hear Beethoven’s final movement of Op. 95 as ironic.

Organization

Before continuing, I summarize each dissertation chapter.

Chapter 2 reviews scholarship on analogy in psychology, music theory and music cognition

as well as categorization in music theory and music cognition. First, I review how scholars in

psychology conceptualize, define and study analogy. Cognitive psychologists and cognitive scientists

argue that analogy also influences how humans perceive similarity and categorize concepts. Next, I

discuss how music theory and music cognition incorporate analogy (either explicitly or implicitly).

Unlike cognitive psychology, most music-related literature do not place significance on analogy,

though they do for similarity and categorization. I review similarity and categorization in music

theory, before discussing music-theoretic surface-level patterns relevant to this project. Finally, I

review similarity and categorization in music cognition, an area dominated by experiences on

categorizing musical themes.

Chapter 3 introduces the analogy framework and how theorists can use it to analyze music.

The first step is retrieval. As part of this step, I discuss listener cultural context. The second step is
37

mapping. First, I define two levels of musical relations: relations within a musical event and relations

between musical events. I discuss constraints for these relations. Next, I define two forms of

analogical comparison: bidirectional listening and unidirectional listening. Then, I introduce four

levels of analogy in music: intra-opus (local analogy; within a same piece), inter-opus (local analogy;

between pieces), intertextual (local analogy; between specific pieces) and extra-musical (distant

analogy; between music and another domain). Finally, I introduce a third step of evaluation and how

it is used in music: inferences, context, and goals. A pattern mapping taxonomy demonstrates how

these mappings can be executed in music for different effects. I close with benefits of the analogy

framework.

Chapter 4 considers how listeners perceive irony in music. This chapter addresses research

question 2: how music theorists as listeners use distant analogies to make sense of something

“unexpected” in music. In this case, the distant analogy is between music and language, and how

listeners use this analogy to perceive irony in Beethoven string quartets op. 95/iv, op. 131/V, and

op. 130/I.

Chapter 5 considers how listeners use analogy to perceive thematic variations as one

category. This chapter addresses research question 1: how listeners may need to use analogy to

categorize musical themes, a process inhibited by the design of music cognition categorization

experiments. I use the analogy framework to argue when listeners would use intra-opus analogy to

hear relations within a musical event. I discuss differences in how composers vary a theme from the

eighteenth- to the nineteenth-century. Then, I analyze the following theme and variation forms:

Beethoven’s Piano Sonata, Op. 109/III, Mozart’s Violin Soanta, K. 377/ii, and Daube’s pedagogical

example of theme and variations in his The Musical Dilettante.


38

Chapter 6 concludes the dissertation and describes future research. I discuss the design of an

experiment in categorizing thematic variations. I hypothesize that if experimentalists use methods

from analogy experiments, participants are more likely to categorize themes based on structural

notes as opposed to surface features. I also discuss how the analogy framework could be used for

two other score analysis projects: 1) connecting schemata and topics, and 2) analyzing how exposure

to film music impacts how modern listeners hear common practice music. In conclusion, I discuss a

possible model for musical communication based on analogy.


39

CHAPTER 2
“We have learned to think about the relatedness of things”: Analogy,
Music Theory, and Music Cognition

Whether or not we talk of discovery or of invention, analogy is inevitable in human thought,


because we come to new things in science with what equipment we have, which is how we
have learned to think, and above all how we have learned to think about the relatedness
of things. We cannot, coming into something new, deal with it except on the basis of
the familiar and the old-fashioned.
—J. Robert Oppenheimer, “Analogy in Science,” an address to the Sixty-Third
Annual Meeting of the American Psychological Association, 1955

[Characters FRY, LEELA, and BENDER are in a spaceship]


FRY: Usually on the show, they came up with a complicated plan, and then explained
it with a simple analogy.
LEELA: Hmmm…. If we can re-route engine power through the primary weapons and
configure them to Mellvar’s frequency that should overload his electro-quantum structure.
BENDER: Like putting too much air in a balloon!
FRY: Of course! It’s all so simple!
—“Where No Fan Has Gone Before,” animated television show Futurama

Both J. Robert Oppenheimer and the writers for the television show Futurama (1999-2013)

draw attention—albeit, in different ways—to humans’ ability to think about the relatedness of things

and how they use something familiar to understand something new or complex. In the preceding

chapter, I proposed two research questions and corresponding hypotheses, which suggest the

cognitive process of analogy as a way to answer these questions. The aim of this dissertation, to

understand how listeners use analogy to make sense of and categorize sequences of music, is

inherently interdisciplinary. To answer these research questions, I draw from literature addressing

analogy, similarity and categorization in at least three disciplines: cognitive psychology, music theory,

and music cognition (figure 2.1).


40

Music
Cognition

Research
Questions

Music Cognitive
Theory Psychology

Figure 2.1: Disciplines and research questions

I first review research in cognitive psychology and its consensus on a definition of analogy and how

humans use the cognitive process of analogy, not only for analogical reasoning but also similarity

and categorization. Then, I review literature on analogy, similarity and categorization in music

theory, including certain music-theoretic categories or patterns. I close by reviewing literature on

similarity and categorization n music cognition experiments, particularly experiments that ask

listeners to categorize musical themes. In chapters to come, I contextualize psychology literature in

music to create a theoretical framework for analogy in listening (chapter 3) and draw from this other

literature in analyses that address research question/hypothesis 1 (chapter 5) and research

question/hypothesis 2 (chapter 4). In the fourth and fifth chapters, I supplement this literature with

other, though minimal, literature from (psycho)linguistics and musicology (figure 2.2).
41

Music
Cognition

Cognitive
Musicology
Psychology

Research
Questions

Linguistics Music Theory

Figure 2.2: Other disciplines and research questions

This chapter, then, reviews the core literature for this interdisciplinary project, guiding a reader

through ink spilled in primarily cognitive psychology, music theory and music cognition.

Definitions of Analogy

As it exemplifies a common understanding of analogy, I begin this definitions section with

an example of analogy from Keith Holyoak and Paul Thagard’s Mental Leaps:

Consider the following discussion between a mother and her four-year-old son, Neil, who
was considering the deep issue of what a bird might use for a chair. Neil suggested,
reasonably enough it would seem, that a tree could be a bird’s chair. A bird might sit on a
tree branch. His mother said that was so and added that a bird could sit on its nest as well,
which is also its house. The conversation went on to other topics. But several minutes later,
the child had second thoughts about what a tree is to a bird: “The tree is not the bird’s chair
– it’s the bird’s backyard!” In this conversation Neil makes a mental leap, exploring
connections between two very different domains. He is trying to understand the relatively
unfamiliar world of creatures of the air in terms of the familiar patterns of everyday human
households. This small example conveys what we mean by analogy, or analogical thinking. The
child’s everyday world is the source analog: a known domain that the child already understands
in terms of familiar patterns, such as people sitting on chairs and houses that open onto
backyards. The bird’s world is the target analog – a relatively unfamiliar domain that the child
is trying to understand. … In a loose sense, there is indeed a some sort of logic – call it
analogic – that constrains the way the child uses analogy to try to understand the target
42

domain by seeing it in terms of the source domain. (Holyoak and Thagard 1995, 2, italics in
original)

Analogy involves a mental leap, using a well-understood source5 as a basis for making inferences

about a less understood target6.

Analogy seems intuitive and easy to define; yet, it appears in different forms depending on

its purpose. It can take shape as a literary turn of phrase (“Juliet is like the sun”), but also as a

problem-solving technique in a lab (“Since it worked with this virus, it should work with this one as

well”). It can be a humorous quote, such as the following by Benjamin Franklin: “A countryman

between two lawyers is like a fish between two cats” (quoted in Gentner and Smith 2012, 135).

Analogy can persuade someone to feel a certain emotion; to “convince you to adapt an emotional

attitude” (Thagard and Shelley 2001, 344). As an example, in 1995, the people of Quebec voted

whether or not to separate from Canada. A side opposed to Quebec’s separation used a persuasive

analogy with a negative emotional content: “It’s like parents getting a divorce, and maybe the parent

you don’t like getting custody” (quoted in Thagard and Shelley 2001, 345). Or, humans can use

analogy to convince others to make certain decisions (Bassok 2001). For example, consider this

exchange in The Social Network (Fincher 2010), a movie about Mark Zuckerberg and the creation of

Facebook.

SEAN PARKER: And that’s where you’re headed, a billion dollar valuation. Unless you take
bad advice, in which case you may as well have come up with a chain of very successful
yogurt shops. When you go fishing you can catch a lot of fish, or you can catch a big fish.
You ever walk into a guy’s den and see a picture of him standing next to fourteen trout?
CHRISTY: No, he’s holding a three-thousand point marlin.
SEAN PARKER: Yup!
MARK ZUCKERBERG: That’s a good analogy.
5 Definition of source: “analog from which inferences and explanatory structure are drawn; typically, the more familiar or
concrete domain: for example, in the analogy ‘An electric circuit is like a plumbing system,’ the base is a plumbing system”
(Gentner and Smith 2012, 130, italics in original).
6 Definition of target: “analog one is drawing inferences about; typically the less familiar or more abstract domain: for

example, in the analogy ‘an electric circuit is like a plumbing system,’ the target is electric circuit” (Gentner and Smith 2012,
130).
43

EDUARDO SAVERIN: Okay, but we all know that marlins don’t really weigh three-
thousand pounds, right?

Sean Parker convinces Mark Zuckerberg to “go big or go home” by advising him, using analogy, to

catch a “big fish” instead of “a lot of little fish” (Sean Parker succeeds in persuading Mark

Zuckerberg, even though Eduardo Savein identifies flaws in the analogy). Despite an abundance of

ways analogy can be realized, I commit to a single overarching definition here from cognitive

psychology that accounts for different ways analogy can be used.

This project’s core model of analogy is Dedre Gentner’s (1983, 2003) Structure-Mapping

Theory. According to Structure-Mapping Theory, a definition of analogy is:

A type of similarity in which two examples share [the] same system of relations. The concrete
features of the examples may be similar or dissimilar; analogical comparisons are concerned
with whether the features relate to one another in the same way in each case. (Jee et al. 2013,
177)

Analogy, then, is more than just figurative language or an artistic mode of expression, it also helps

humans interpret mundane facets of everyday life, often without them realizing it. Samuel Day and

Gentner (2007) claim that:

Analogical processes may also be involved in the far more routine task of organizing and
interpreting our daily experiences. Intuitively, it seems that analogies with our prior
experience could contribute to our fluency in processing current situations, even when such
analogies are not overtly noticed or intentionally pursued. (39)

Analogy, a cognitive process that helps humans understand their everyday experiences, depends on

comparing two situations or analogs (where one is a base and another a target) that share similar

structures of relations. It helps humans generate inferences on these as well as influences similarity

perception and even categorization.


44

Analogical Processes: Retrieval, Mapping, and Evaluation

When a person reasons analogically, he or she performs the following steps, as described by

Gentner and Smith (2012) (for a visual representation, see figure 2.3):

 Retrieval: Given some current topic in working memory, a person may be reminded of a
prior analogous situation…
 Mapping: Given two cases present in working memory (either through analogical retrieval
or simply through encountering two cases together), mapping involves a process of
aligning the representations and projecting inferences from one analog to the other.
 Evaluation: Once an analogical mapping has been done, the analogy and its inferences are
judged. (131)

As seen in figure 2.3 below, Holyoak (2012) adds a learned “schema” as a product of these steps:

In the aftermath of analogical reasoning about a pair of cases, some form of relational
generalization may take place yielding a more abstract schema for a category of situations…
of which the source and target are both instances. (235)

Figure 2.3: Major components of analogical reasoning (graphic from Holyoak 2012, 236)
45

When defining the “mapping” step above, Gentner and Smith (2012) use the following phrases:

“two cases present in working memory,” “aligning representations” and “projecting inferences from

one analog to another.” Before continuing, I explain what each of these phrases mean in turn.

Gentner and Smith (2012) mention that a person considers “two cases present in working

memory,” either through analogical retrieval or two cases encountered together. On one hand, to

have two cases present, humans can remember a better understood situation from long-term

memory, what Gentner and Smith (2012) call “analogical retrieval.” In the analogy electricity is like

water, for example, “water” is a better understood domain from long-term memory used to

understand “electricity,” whereby a person projects one domain onto another. On the other hand,

humans can also compare and notice relational similarities between two cases encountered together.

For example, one might compare one picture of a woman giving food to a squirrel to a second

picture right next to it of a man giving food to a woman. After comparing, one notices both pictures

share a relation of giving. A person did not retrieve either of these pictures from long-term memory;

instead, he or she encountered the pictures together and it prompted him or her to make the

analogy.

Also, in the mapping stage, Gentner and Smith (2012) describe “a process of aligning the

representations.” Although I clarify what “aligning” entails at a later point, this statement assumes

there are typical conceptual cognition representations as well as interactions of representations and

processes (e.g. Palmer 1978). Gentner and Markman (1997) write:

We must have a representational system that is sufficiently explicit about relational structure
to express the causal dependencies that match across the domains. We need a
representational scheme capable of expressing not only objects but also the relationships and
bindings that hold between them, including higher order relations such as causal relations.
(46)
46

In order to have this concept of analogy as “aligning representations,” I presume cognitive

representations exist that encode relationships between objects.

Finally, I further explain “projecting inferences” as part of a mapping stage. After a person

maps over necessary relational structure, he or she generates inferences by filling in “holes” or

“vacancies” in the target’s representation. According to Gentner and Smith (2012):

Once the base and target have been aligned and their common relational structure found, if
there are additional parts of the relational pattern in the base that are not present in the
target, then this missing pattern will be brought over as a candidate inference….Thus, one
way to think about inference generation is as a process of relational pattern completion. The
requirement that candidate inferences be connected to the common relational pattern
effectively filters which inferences will be considered. (132)

As an example of inference generation, Gentner and Smith (2012) use a bathtub analogy:

The amount of water in a bathtub is determined by the rates of water flowing into the tub
and water flowing out through the drain. As long as the inflow of water into the tub exceeds
the outflow the bathtub will continue to fill. Likewise, the amount of carbon dioxide (CO2)
in the atmosphere is determined by the rates of CO2 emissions and CO2 removal. (130)

A person aligns the known fact that “the amount of water entering and leaving the tub determines

the total amount of water in the tub” with the known fact that “the amount of CO2 entering and

leaving the atmosphere determines the total amount of CO2 in the atmosphere” (Gentner and Smith

2012, 132). Then, a person draws new inferences. One such inference might be “the amount of CO2

in the atmosphere will decrease if CO2 removal exceeds CO2 emissions,” since water in a tub will

decrease if the amount draining exceeds water entering (Gentner and Smith 2012, 133). A person

notices a “hole” in the representation of “amount of CO2 in the atmosphere” and uses their

knowledge of “water in the tub” to predict would happen if CO2 removal exceeds its emissions.

Mapping: Gentner’s Structure-Mapping Theory

Mapping, or a “systematic set of correspondences” between the elements of a source and a

target analog, is the heart of analogical thinking (Holyoak and Thagard 1995, 4). Out of all these
47

steps, psychologists of analogical research have studied “mapping” the most since “the mapping

process is viewed as pivotal” (Spellman and Holyoak 1996, 308). In the mapping step, listeners

associate an element of one structure with a corresponding element in a different structure; for

music, this generally means placing relationships between musical elements (including note-to-note,

phrase-to-phrase, etc.) in dialogue with each other. Mapping is also “the stage at which knowledge

about the source is carried over to the target” (Blanchette and Dunbar 2002, 672).

Structural Alignment

Structure-Mapping Theory entails a “comparison process” which “involves finding an

alignment between the base and target representations that reveals common relational structure”

(Gentner 2003, 107). Mapping begins with structural alignment (Goldstone and Medin 1994; Gentner

1983; Gentner and Markman 1997): “Identifying correspondences between two analogs, based on

their common relations” (Gentner and Smith 2012, 130). Humans normally focus on objects instead

of relations between those objects, especially if the two analogs do not share surface similarity.

When representations are structurally aligned, however, it prompts humans comparing the situations

to notice relations instead of objects (Markman and Gentner 1993).

In order for structural alignment to occur, the representations must be structurally consistent

due to tacit constraints: one-to-one mapping, parallel connectivity, and the systematicity principle. The first, one-

to-one mapping or one-to-one correspondence (Gentner and Markman 1997; Gentner and Smith 2013),

requires that each of the representation’s elements match at most one element to another in the

representation. On element cannot map onto two different elements in the other representation.

The second, parallel connectivity (Gentner and Markman 1997; Gentner and Smith 2013), ensures that

if two predicates (or relations) correspond to each other, then their arguments also correspond to
48

each other. For example, let us return to our atom and solar system analogy from the previous

chapter. If the relation is REVOLVES, as in smaller objects revolving around a bigger one, and the relation

matches across both representations, the arguments will also match: electrons will map onto planets

(both are the things that REVOLVE) and the nucleus will map onto sun (both are the things that

are being REVOLVED around). Therefore, the alignments must be structurally consistent (Gentner

and Markman 1997, 47). The systematicity principle (Gentner 1983; Clement and Gentner 1991;

Gentner 2003) biases humans to map relations that participate in a common system of relations—

“sets of common relations connected by higher-order relations that can themselves be mapped”

(Clement and Gentner 1991, 92). Thus, humans prefer to map “lower-order matches” (e.g. events)

that are connected through “higher-order constraining relations” (e.g. causal relations) (Gentner

2010, 754). Gentner (2010) writes: “The systematicity principle stems from a tacit preference for

coherence and predictive power. Thus, when a given analogy affords more than one consistent

interpretation, people prefer the more systematic interpretation, all else being equal” (Gentner 2010,

754). The more systematic interpretation is preferred, the only exceptions being when factual or

relevance considerations overrule it in some way.

Consider figure 2.4: two analogous scenes that can be aligned. The elements can be put in a

one-to-one correspondence with each other where the relational structure is consistent in both

scenes. In this example, the relational structure could be labeled steadily decreasing size, as Circle

1>Circle 2>Circle 3 (Gentner and Smith 2012, 131). A perceptual/object match (one that shares

surface similarity) for the pointed-to circle in (a) is the middle circle in (b); it is the same size and

looks the same. A relational match (one that shares relational similarity) for the pointed-to circle in

(a) is the smallest circle on the far right in (b); it is the smallest circle, or the final evolution of the

relation steadily decreasing size.


49

Figure 2.4: Circles steadily decreasing in size (graphic from Gentner and Smith 2012, 131)

Comparison highlights common relational structure and accentuates structural

commonalities necessary for structural alignment (Loewenstein, Thompson, and Gentner 2003;

Gentner et al. 2009; Catrambone and Holyoak 1989; Christie and Gentner 2010; Gentner and

Markman 1997; Gentner and Namy 1999; Gick and Holyoak 1983; Markman and Gentner 1993).

Thus, comparison is a “powerful learning mechanism when it occurs; however, it often fails to occur in

situations where it could be effective” (Gentner et al. 2009, 1353, italics in original).

Cross-mappings are often used in experimental stimuli to illustrate the importance of

comparison in analogical processing. A cross-mapping is “a comparison in which two analogous

scenarios contains similar or identical objects that play different relational roles in the two scenarios”

(Gentner and Markman 1997, 47). The 3 in 1:3::3:9 is an example of cross-mapping (Gentner and

Markman 1997). Although 3 is present in both 1:3 and 3:9, it plays different relational roles in each

situation.

Markman and Gentner (1993) used cross-mapping to demonstrate how comparison

highlights relational structures. For stimuli, they used scenes where the same object played different

roles in each scene (cross-mapping; figure 2.5).


50

Figure 2.5: Causal scenes containing a cross-mapping. The woman in the top scene is receiving food, while the

woman in the bottom scene is giving food (graphic form Markman and Gentner 1993, 436)

In the top scene of figure 2.5, a man gives food to a woman. In the bottom scene of figure 2.5, a

woman gives food to a squirrel. In these scenes, the act of “giving from one agent to another” is a

relation. The woman is cross-mapped since she appears in both scenes, but plays different roles. In

the top scene, she is given to, while, in the bottom scene, she is the giver. The experimenter, after

pointing to the woman in the top scene, asked participants which object in the bottom scene “went

with” the first object. If the participant chose the woman in the bottom scene, then he or she chose

the same object without considering its relational role, or an object mapping (Markman and Gentner

1993, 438). If the participant chose the squirrel in the bottom scene, then he or she chose the object
51

with the same relational role whether it was the same object or not, or a relational mapping

(Markman and Gentner 1993, 438). If participants just answered the “went with” question, then they

tended to make object mappings. However, if participants compared the two scenes beforehand,

then they were far more likely to make relational mappings.

Markman and Gentner (1993) conclude that comparison causes people to align scenes and

makes common relations more salient. Hence, comparison altered how participants viewed the

scenes. If they just looked at the scenes, they picked the object match. If they compared the scenes

first, it changed what they attended to in the scene; relations became more noticeable, and so

participants chose the relational match instead. Thus, one group of participants engaged in the

activity of comparison, while the other did not. Humans tend to notice surface features or objects

first. Yet, they attend to relational structure if they participate in certain actions, such as comparison.

Two Forms of Comparison: Projective Analogy and Mutual Alignment Analogy

There are two forms of comparison in analogical reasoning: projective analogy and mutual

alignment analogy. Projective analogy is someone relating something familiar to a novel instance,

while mutual alignment analogy is someone comparing two novel situations to each other (figure

2.6):

In projective analogy, the learning results chiefly from inference projection. A well-understood
situation (the base) is aligned with a less understood situation (the target), and inferences are
mapped from the base to the target…. Projective analogy plays an important role in learning
and instruction… But although projective analogy is important in learning, it cannot explain
how the process gets started…. This brings us to a second kind of analogical learning,
analogical encoding or mutual alignment… Analogical encoding occurs when two analogous
situations are present simultaneously and are compared to one another. Here the key process
is not directional projection of information (though inferences can occur), but aligning,
rerepresenting, and abstracting commonalities. If inferences are drawn they may be
bidirectional, with both examples serving as bases as well as targets. (Gentner and Kurtz
2005, 255, italics in original)
52

Figure 2.6: Schematic of two forms of comparison in analogy (graphic from Jee et al. 2010, 3)

Mutual alignment analogies are often from the same domain or topic (Jee et al. 2010, 8) as well as

often both partially understood. Thus, humans make analogies either from retrieving a base from

long-term memory or from comparing two situations together in working memory.

Retrieval: The “Analogical Paradox”

“Retrieval” means transferring a prior situation in long term memory to a current one. I use

retrieval to discuss a listener’s past experience with music: musical conventions, cultural information,

and so on. However, retrieval also subsumes a listener’s experience with other domains. These other

domains may be activated if the listener makes an analogy between it and music. In the literature on

analogy in cognitive science, the most relevant information (especially in regards to my first research

question) is the “analogical paradox” (e.g. Dunbar 2001).

Humans often fail to retrieve source analogs from long term memory, even if they are useful

in the new context (Gick and Holyoak 1980; Gick and Holyoak 1983; Keane 1987; Novick and

Holyoak 1991; Ross 1987; Gentner et al. 2009). Dunbar (2001) calls this phenomenon the
53

“analogical paradox,” although it has appeared under other names including “inert knowledge

phenomenon,” “the failure to appropriate retrieval” (Gentner et al. 2009, 1344), “the retrieval gap”

(Holyoak 2012, 224), and “transfer failure” (Blanchette and Dunbar 2002, 673). When humans are

reminded of prior situations, it is often due to surface similarities—such as characters and settings—

instead of relational similarities (Brooks, Norman, and Allen 1991; Catrambone 2002; Gentner,

Ratterman, and Forbus 1993). It is not that people lack the right knowledge, but instead are unable

to retrieve the “right knowledge at the right time” (Loewenstein, Thompson, and Gentner 2003,

120).

Evaluation

After someone matches a base and target domain, the knowledge he or she transfers from

base to target may be proposed as candidate inferences (Gentner and Markman 1997, 51). Gentner

and Smith (2013) indicate that inferences are: (a) highly selective (people do not bring over all they

know about the base to the target), and (b) not necessarily true (6). Gentner and Markman (1997)

provide the following example:

Imagine you have a friend with a sarcastic sense of humor that makes her difficult to get
along with but a helpful temperament that wins her the loyalty of her friends. If you met a
new person and discovered that he had a sarcastic sense of humor, then based on his
similarity to your other friend, you would probably be more willing to suppose that he is
difficult to get along with than to infer that he has a helpful temperament that wins him loyal
friends. (51)

A person is selective in their inferences, yet the inference may not necessarily be true. In the case

described above, the new person met may not be difficult to get along with after all.

After structural alignment and inferences projected from base to target, a person evaluates

the analogy and its inferences as part of the larger evaluation step. According to Gentner and Smith

(2012), at least three factors impact how humans evaluate analogical inferences: 1) factual
54

correctness, 2) goal relevance, and 3) new knowledge (133). First, people generally reject inferences

that are clearly false. Factual correctness relates to “adaptability: how easy it is to modify a fact from

the base to fit the target” (Gentner and Smith 2012, 133). Second, inferences that relate to the

reasoner’s current goals are more important in evaluating the analogy than inferences that do not

(e.g. Clement and Gentner 1991). Spellman and Holyoak (1996) claim that:

An analogy may be drawn to help achieve a variety of different goals, such as solving a
problem that has arisen in the target domain, predicting what is likely to happen if various
alternative actions are taken, or generating an explanation of why the target domain behaves
as it does. (308)

Spellman and Holyoak (1996) demonstrate that people who have two possible mappings available to

them for a given analogy select the mapping whose inferences are most applicable to their goals.

Third, reasoners accept inferences that yield more new knowledge than ones that do not: “The idea

is that inferences that potentially yield a significant gain in new knowledge may be desirable (even if

somewhat risky), especially when brainstorming or dealing with unfamiliar domains” (Gentner and

Smith 2012, 133).

Holyoak and Thagard’s Multiconstraint Theory

Structure-Mapping Theory is not the only theory of analogy. Holyoak and Thagard (1995)

put forth a Multiconstraint Theory where analogy is understood as “three interrelated types of

constraints” (15). The first is that “analogy is guided to some extent by direct similarity of the

elements involved” (Holyoak and Thagard 1995, 5, italics in original). Second, there is a “pressure to

identify consistent structural parallels between the roles in the source and target domain” (Holyoak

and Thagard 1995, 5, italics in original). Finally, the third is that “the exploration of the analogy is

guided by the person’s goals in using it, which provide the purpose for considering the analogy at all”

(Holyoak and Thagard 1995, 6, italics in original). These three aspects—similarity, structural and
55

purpose—are not rules, but guidelines to creating and understanding analogy (Holyoak and Thagard

1995, 6). Holyoak and Thagard (1995) argue that:

Ideally, the three constraints will all work together to suggest a single interpretation of how
the source is applicable to the target. But in less-than-perfect analogies the constraints can be
at odds with each other, for example, when structural considerations support
correspondences that are incompatible with object similarities or with the purposes of the
analogist. (37)

To develop their multiconstraint approach, Holyoak and Thagard (1989) combine assumptions of

structural consistency with a pragmatics focus (e.g. attainment of goal states).

There is much agreement between Holyoak and Thagard’s Multiconstraint Theory and

Structure-Mapping Theory. According to the Multiconstraint Theory, for instance, “people implicitly

favor mappings that maximize structural parallelism (in agreement with Gentner’s, 1983, structure-

mapping theory)” (Holyoak 2012, 239). Holyoak and Thagard (1995) also insist that:

The multiconstraint theory and structure mapping theory are in general agreement that the
core of analogical thinking involves finding mappings between relational structures, and that
the use of analogy depends on sensitivity to structure, semantic similarity of concepts, and
the purpose of analogy. However, the theories formulate the constraints in different ways.
The multiconstraint theory….integrates all three types of constraints within a process of
parallel constraint satisfaction. (268)

When relevant, I adopt positive elements that Multiconstraint Theory offers. In my analogy

framework for listening to music, for instance, I consider listener and composer goals, which I

adopted from the Multiconstraint Theory. Yet, Structure-Mapping Theory is the dominant

analogical model used here. This is because Structure-Mapping Theory includes an act of structural

alignment, often prompted though comparison. The Multiconstraint Theory lacks this aspect, at

least in systematic detail. Structure-Mapping Theory also places greater importance on comparison. I

consider this action important to understanding how people listen to music. In addition, Structure-

Mapping Theory includes a projection of inferences. Finally, I interpret Structure-Mapping Theory

as a more active-oriented approach to analogy than the Multiconstraint Theory. Structure-Mapping


56

Theory focuses on actions, such as aligning representations, projecting inferences, and so on.

Constraints, on the other hand, seem more passive. They “happen to” a person as opposed to a

person “acting toward something.” This connects to my goal of creating an action-oriented

framework for analogy in listening. Overall, Holyoak and Thagard (1995) call Multiconstraint

Theory a “very general theory” (15). Structure-Mapping Theory, on the other hand, is systematic and

specific; overall, a cleaner model more easily transferred to the listening experience.

Analogy and Similarity

According to Keane and Costello (2001, 287), mechanisms such as structural alignment and

analogy are behind multiple phenomena, including similarity. The Oxford English Dictionary defines

“similarity,” as “likeness, resemblance” and “points of resemblance” (“Similarity, N.” 2015).

Similarity, in the words of William James ([1890] 1950), could also be defined as a “sense of

Sameness” which is “the very keel and backbone of our thinking” (459). Though similarity is a

central theoretical construct in psychology, scholars have accused similarity of being too flexible

(Goodman 1972; Murphy and Medin 1985). Although, similarity loses some flexibility when

grounded in analogy. The philosopher Nelson Goodman (1972) claims that similarity needs a frame

of reference; similarity of A to B is meaningless unless a person knows “in what respects” A is

similar to B. In Medin, Goldstone and Gentner (1993), similarity is often considered flexible but

“respects” as described by Goodman (1972) are systematically fixed through a similarity comparison

process (255).

Gentner and her colleagues recognize three different “classes” of similarity in “similarity

space”: analogy (or relational similarity), literal similarity and mere-appearance (or perceptual

similarity). Perceptual similarity is present when two objects look similar on the surface; they share

similar attributes. For example, a sun and an orange-yellow ball share similar attributes: they are both
57

yellow-orange and round. Relational similarity is present when objects share relations even though

they do not look the same on the surface. For example, a nest and a house are both homes (for birds

and humans respectively): they both share a relation of something living inside. Yet, they do not look

similar on the surface; they do not have shared attributes. Instead of strict categories, these different

“classes” rest on a continuum (Gentner and Medina 1998, 266) where the y axis is “relations shared”

and the x axis is “attributes shared” (figure 2.7).

Figure 2.7: Visual representation of similarity space (graphic from Gentner and Smith 2013, 10)

Consider the often-used analogy between an atom and the solar system (see chapter 1). In analogy,

then: “there is substantial relational overlap with very little object similarity: objects correspond not

because of inherent similarity but by virtue of playing like roles in the relational structure” (Gentner

and Medina 1998, 266). When object-similarity (or perceptual similarity) begins to increase, the

comparison becomes literal similarity. Consider comparing one solar system with another. Literal

similarity then “involves both relational and object [perceptual] commonalities [or similarity]”

(Gentner and Medina 1998, 266). Mere-appearance (or perceptual similarity) contrasts with analogy,

as when comparing a planet with a round ball (Gentner and Medina 1998, 266). The planet and ball
58

do not share relations, though they look alike on the surface since they both have a round shape. A

planet revolves around the sun, and a ball does not revolve around anything. Therefore, they share

perceptual features, but not relations.

Gentner and Markman (1997) argue “similarity is like analogy,” since humans use the

alignment process in both (48). Participants judge stimulus pairs with aligned relational similarity as

more similar than stimulus pairs with primarily perceptual similarity (Goldstone, Medin, and

Gentner 1991; Gentner, Ratterman, and Forbus 1993). This implies that humans perceive a

difference between relational similarity and perceptual similarity, a distinction that Gentner and

Markman (1997) find emerges at different points in processing (53). Studies of relational

comparisons find that participants base similarity on local as opposed to relational matches when

they need to respond within 1000 ms (Goldstone and Medin 1994). When they have a longer

response deadline, participants base similarity on relational matches. (Goldstone and Medin 1994).

Analogy and Relational Categories

In the same way that relational similarity differs from perceptual similarity, a category

defined by relations (a relational category) differs from a category defined by perceptual features (a

perceptual or entity category). Barr and Caplan (1987) first made this a distinction between categories,

although they used “intrinsic features” for “perceptual features” and “extrinsic features” for

“relations.” Features are intrinsic when it is true of an entity in isolation (“has wings” is an intrinsic

feature of a bird). A feature is extrinsic when it represents a relationship between two or more

entities. “Used to work with” could be an extrinsic feature of a hammer in the relation between

hammer and worker (Barr and Caplan 1987). Barr and Caplan’s (1987) extrinsic and intrinsic features

connect to Gentner and Kurtz’s (2005) relational and entity (perceptual-based) categories. Entity

categories (or perceptual categories) are categories where the members have “highly overlapping
59

intrinsic features and feature correlations” (Gentner and Kurtz 2005, 152). Relational categories, in

contrast, are ones where common relational structure determines category membership (Gentner

and Kurtz 2005, 151). Relational categories can be based on relations that:

May include common function (e.g., both are edible), mechanical causal relations (e.g. both are
strong so they can bend things), biological causal relations (e.g., both need water to grow), role
relations (e.g. both grow on trees), and progeneration (e.g., both have babies). Relational categories
can also be based on perceptual relations such as symmetric in form, mathematical relations
such as prime, or logical relations such as deductively sound. It is these relational systems that
provide the theory-like aspects of concepts and categories. (Gentner 2005, 216)

Relational categories are often contrasted with perceptual one, even though they are not always so

distinct in our perceptual world. An object can be a member of multiple categories simultaneously.

For instance, a cow could be a member of the perceptual category “cow,” but could also be a

member of the relational category “barrier” if it is between a person trying to escape from an angry

bull. Whether or not a person attends to the object as a member of “cow” or “barrier,” most likely

depends on their attention and needs in the situation. Therefore, perceptual features often describe

how an object looks on the surface and are intrinsic to the object (e.g. the color red, a round shape,

etc.). On the other hand, relations describe how an object connects to another object (e.g. to get

from point A to point B).

Markman et al. (2011) recognizes three types of relational categories: role-governed

categories, schema-governed categories, and thematic categories (361). Role-governed categories are

determined by the role that category plays in a relation (Goldwater, Markman, and Stilwell 2011).

Guest is a role-governed category (figure 2.8) since its definition depends on a relation between it and

something else. A guest is a guest when an agent visits a home or establishment that is not their own.

In the same way, host is a role-governed category (figure 2.8) since its definition depends on an agent

whose home is being visited by another agent. Neither of these categories depend on what these

agents look like—they could be men, women, or animals (as in fairy tales or children’s stories).
60

Thematic categories are organized around the goals of an organizer (e.g. Barsalou 1983). For

example, a category diet foods could have a wide range of members since their membership depends

on a goal on the part of an agent. Finally, schema-governed categories are determined by a relational

system (Goldwater, Markman, and Stilwell 2011). For instance, consider x visits y (figure 2.8). Due to

its relational system, visit is a schema-governed category as there is one agent in relation to another

agent; the relation of going to a house or establishment that belongs to one agent, but not the other

(Goldwater, Markman, and Stilwell 2011, 360). In this example, role-governed categories guest fits in

x, while host fits in y. In general, category membership can depend on relationships, features or both.

Figure 2.8: Schematic of the category “visit” (graphic from Goldwater, Markman, and Stilwell 2011, 360)

Goldwater, Markman and Stilwell (2011) differentiate these three types of relational categories—

role-governed, schema-governed, and thematic categories—from feature-based or perceptual

categories, which have received the most attention in the categorization literature (see Murphy

2002). Goldwater, Markman and Stilwell (2011) define feature-based or perceptual categories,

writing they “are represented as collections of features describing category members, e.g., birds are

animals with wings and a beak. These features are primarily about the category members themselves

and not about relations among category members and other entities” (361). In their discussion of
61

feature-based categories vs. role-based categories [relational categories], Goldwater, Markman and

Stilwell write: “While feature-based categories and role-governed categories are distinct… their

representations can become connected because whenever a role is filled, it is filled with a member of

a feature-based category. For example, the roles of guest and host are most typically played by people”

(361).

As comparison helps structural alignment in analogy, comparison also helps humans

categorize based on relational information. Gentner and Namy (1999) examined how children

acquired categories. Early in development, children think of “like kinds” in terms of perceptual,

object-level commonalities, such as “shape bias” (e.g. Baldwin 1992; Imai, Gentner, and Uchida

1994). Experimenters showed children an orange and called it a “blicket.” When asked whether a

banana or a ball was also a “blicket,” children chose the ball a significant number of times. Children

were basing categorization on “perceptual” categorization (the roundness of the shape) as opposed

to “taxonomic” categorization (fruit). When Gentner and Namy (1999) gave children two examples

of the “blicket” category, children chose the taxonomic choice significantly over the perceptual

match. Gentner and Namy (1999) argue that, “the process of structural alignment may act as a

bridge from an initial perceptually-based category to a later more sophisticated understanding of the

category” (506). The comparison between the exemplars highlighted the relational commonalities

(the taxonomic category) and promoted structural alignment.

Comparison facilitates categorization of relational categories not just for children, but also

adults (Elio and Anderson 1984; Medin and Ross 1989; Jee et al. 2013). For instance, Jee et al. (2013)

used comparison to help participants learn between categories in a geosciences context. Participants

saw images of faults in rocks that were highly similar except for a crucial difference, helping them
62

categorize the image as category of “image-with-fault” or “image-not-with-fault.” They found that

“comparing alignable contrasting images facilitates fault classification” (Jee et al. 2013, 180).

In addition to alignment, comparison influences future transfer through relational schema

abstraction, or the use of an abstract relational category. Comparison was positively associated with

participants abstracting a relational schema (Gentner et al. 2009, 1372); essentially, an abstracted

relational category. After comparing two analogs, participants create a relational schema that can be

used as a memory item to transfer relationally similar instances from autobiographical memory to

future situations (Gick and Holyoak 1983; Gentner et al. 2009). Gick and Holyoak (1983)

demonstrated that comparison creates a schema that leads to future transfers, while Gentner et al.

(2009) demonstrated that there could be a backwards effect of using a schema to retrieve from past

memories. According to Gentner et al. (2009), “analogical abstraction [of a relational schema] aids in

relational retrieval from past memories and also applies to future instances in relational transfer”

(1353).

Gentner et al. (2009) ran five experiments which supported the following: “(a) that

analogical abstraction [creation of relational schemas] at recall time promotes relational retrieval of

prior exemplars from long-term memory and (b) (in concert with prior findings) that analogical

abstraction at learning time promotes relational transfer to future exemplars” (1363, italics in original).

Gentner and colleagues (2009) realized relational schemata have two properties: 1) diminished surface

competition, and 2) stronger relational matches. The first property, diminished surface competition, means that

relational schemas have fewer surface features. If used as a memory probe, it will retrieve few

surface matches. The second, stronger relational matches, means that relational information is

“weighted” heavier in memory (Gentner et al. 2009, 1363).


63

Approaches to Analogy in Music

Since I have reviewed analogy in cognitive psychology, I will now review analogy in music

literature.

Music Theory

In music theory, Steve Larson, Lawrence Zbikowski and Robert Hatten explicitly discuss

analogy in their research. Larson, especially, uses analogy as a basis or his book Musical Forces. He

writes:

I define an “analogy”…. As a mapping that calls attention to similarities between two


different things – regardless of whether those things belong to the same or to different
domains… When the mapping is between things in different domains, I call that analogy a
“metaphor.” (Larson 2012, 36)

Unlike Gentner’s definition, however, Larson (2012) does not restrict analogy to relational similarity,

but considers any type of similarity. Larson (2012) describes various potential analogies in music:

hearing similarities between passages in one piece, between two different pieces, and hearing music

as similar between two domains (36–50). Zbikowski (2013), in his definition of analogy, draws

attention to relations as a necessary component:

In the case of similarity, both attributes and relations are shared: a pencil and a pen are
similar to each other both in appearance and in function, although the kind of marks each
makes on a writing surface (permanent or impermanent; of relatively consistent coloration or
subject to gradation) are different. In the case of analogy, only relations need be shared: a
finger is analogous to a pen in that it is an approximately cylindrical structure that can be
used to trace characters on a writing surface; unlike a pen or pencil, however, the finger
leaves no discernible marks on the writing surface and its ‘cylinder’ is firmly attached to the
larger structure of the hand. (12)

Zbikowski (2013) mentions relations; yet, features he calls relations—such as “cylindrical

structure”—sound more akin to perceptual features than relations. When Hatten (1994) defines
64

analogy, he mentions relationships, and even types of relationships in music, such as part-for-whole,

and so on.

An analogy is a relationship arising from a comparison of relationships: A is to B as X is to Y,


where “as” implies a figural meaning illuminating the nature of one rational pair in terms of
another. A correlation is a more literal mapping of meaning (literal for a given style)
coordinated by the analogous markedness values of the two pairs of oppositions. It is in this
sense that A can “mean” X. “A is to B as X is to Y” does not imply that A means X, but that
it relates to B in some way (similarity, contiguity, part-for-whole, etc.) comparable to the way
X relates to Y. Often, that “way” requires the creativity of metaphor to interpret. Analogy
can best be understood as one of the motivations underlying the mapping of a correlation.
(Hatten 1994, 38)

Although Hatten (1994) does not allude to cognitive literature on analogy, Structure-Mapping

Theory could be a cognitive underpinning to Hatten’s theory of relating music to outside world and

its significations. In summary, these theorists engage with “analogy” to a varying degree and for

different purposes. Larson uses “analogy” to analyze how listeners relate musical concepts to

physical concepts, Zbikowski uses it to understand how listeners hear emotions in music, and

Hatten uses it to motivate correlations in his theory.

Music Cognition

In music cognition, Kielian-Gilbert (1990) as well as Eitan and Granot (2007) connect

analogy to a cognitive or empirical study of music. Kielian-Gilbert (1990) argues that humans relate

different phrases and sections of a piece to each other—or “similarity of musical roles”—using

analogy (69). In Eitan and Granot’s (2007) experiments, participants heard musical phrases as more

similar to each other when they shared analogous changes in intensity. For example, a musical

phrase with a crescendo (gradual increase in loudness) was judged as more similar to a phrase with an

accelerando (gradual increase in speed) than a decelerando (gradual decrease in speed). These results

suggest that listeners judge similarity not just using musical dimensions (e.g., timbre, contour) but
65

also relational structure across dimensions (e.g. intensity change or “isomorphic intensity contours”)

(Eitan and Granot 2007, 64). Eitan and Granot (2007) claim that theorists, especially those who

study atonal music, may use such intensity changes as a motivic device in analysis (64).

Similarity and Categorization in Music Theory

Since this dissertation examines how analogy connects to similarity and categorization in

music, I review literature not just on analogy, but also on similarity and categorization in music

theory and cognition (a more copious literature anyway). In music theory, Hanninen (2004) observes

that, “association and categories’ central place in analytical practice may well explain why theorists

have paid them scant attention: plain stones are rarely noticed” (147). In his book Conceptualizing

Music, Zbikowski (2002) devotes an entire chapter on categorization, writing specifically about a

Wagner example:

To understand this music – to make sense of the sonic texture Wagner weaves – requires
being able to assimilate these various musical phrases into a single cognitive construct and
then recall that construct, often after an hour or more of Wagnerian effusion. Understanding
Wagner – or most music, for that matter – requires being able to think in terms of categories
of musical events. (24)

Before these current theorists, Meyer (1967), using the term “pattern” as opposed to category, writes

in Music, The Arts, and Ideas: “Understanding music is not merely a matter of perceiving separate

sounds. It involves relating sounds to one another in such a way that they form patterns (musical

events)” (46). In the following sections, I discuss how music theorists have analyzed similarity and

categorization in music. In particular, I discuss Meyer’s conformant relationships as well as review

music-theoretic patterns or categories relevant to this dissertation.


66

Meyer’s Conformant Relationships

Meyer (1973) describes conformant relationships in music as: musical relationships in which

one “identifiable, discrete musical event is related to another such event by similarity” (44). Whether

a listener perceives a musical relationship as conformant or not depends upon the degree of

similarity between them (Meyer 1973, 46). Ockelford (2009) claims that the perception of similarity

is at the heart of understanding musical structure. Meyer (1973) argues that:

The more all the parameters are duplicated in model and variant, the stronger the
conformant relationships. This is specially the case with the primary pattern-forming
parameters of pitch, duration, and harmony. For instance, a motive can be changed in
register, dynamics, tempo, and instrumentation and still be recognizably the same. (Meyer
1973, 46)

For Meyer (1973), two musical themes are similar based on pitch and pitch-class relationships as well

as secondarily rhythmic-metric structure. On the one hand, melody, rhythm-meter and harmony are

primary parameters since they create musical syntax (Meyer 1989, 14). On the other hand, secondary

parameters7 are musical dimensions that do not give rise to musical syntax; for instance, dynamics,

tempo, articulation, timbre, sonority, and texture (14). Secondary parameters have been called

“surface” features of music while primary parameters have been called “deep” features (Lamont and

Dibben 2001, 247). Thus, large differences in secondary parameters should not change how a

musical theme is “categorized.”

Music-Theoretic Patterns

Since my research question 1 asks how listeners categorize musical themes, I review

literature on themes and thematic variations. In addition, I review literature on other music-theoretic

patterns that could also be categorized using analogy: Meyer and Gjerdingen’s schemata and

7To be clear, I do not agree with Meyer’s use of terms such as “primary” and “secondary.” It implies that primary
parameters are more important than secondary parameters. I do not agree with this stance.
67

Hanninen’s associative sets. I consider how Hanninen uses musical relations in her associative sets

when creating my analogy framework (chapter 3), while I use schema theory and variations analytical

techniques in the analysis chapters (schemata: chapters 4 and 5; variations: chapter 5).

Thematic Variations

When music theorists discuss similarity, they often discuss musical themes. The theme, or

motive, is any kind of musical event in “a piece or section of a piece that, despite change and

variation, is recognizable as present throughout” (Schoenberg [1934-36] 1995, 169). The intuition of

most music theorists, such as Meyer (Meyer 1973; Meyer 1989), Réti (1951), Schoenberg (1967;

1978), and Zbikowksi (2002), is that themes are perceived as more or less similar due to primary

parameters. If the two musical themes have similar pitch structure, then they would be similar to

each other and possibly in the same musical category. Other features, including instrumentation,

dynamics, tempo, texture and so on, may “create variants within a motivic or thematic category, but

do not define or constitute motives and themes” (Eitan and Granot 2009, 140). Thus, two musical

passages are the same thematic category if they share pitch and meter/rhythm information, whether

or not one is very loud and the other is very soft. Primary parameters that constitute a theme are

often relationship-oriented; therefore, a listener may use analogy to categorize two musical passages

as the same thematic category (see chapter 5).

Analogy, then, helps listeners recognize not just a theme, but also variation on a theme since

variations elaborate upon the basic thematic identity of relations. Ivanovitch (2010) opens his article

on themes and variations by describing this common understanding of variations as elaboration

upon elements of a base theme:

Variation is often seen as a straightforward affair: a self-contained piece called the theme
gives rise, through various elaborations and manipulations, to a series of variations that are
68

“based upon” that theme. According to this view, which appears to honor the linear
temporal sequence in which a variation set unfolds, the theme functions effectively as a
repository of elements, a coordinated collection of structural features (such as harmony,
melody, motives, or – more flexibly – a composite voice-leading structure) from which the
variations will select, emphasizing now one aspect, now another. (Ivanovitch 2010, 1)

In the eighteenth-century, music writers recognized that variation form had three different, but

related connotations, most of which considered variation as elaboration or modification:

First, it was a method of composition, by which a simple melody or bass line could be
elaborated in shorter note values to create a fully-worked out piece. In this sense, all music
could be seen as composing out a simple framework…. Second, it was a technique designed
to add interest to a composition by modifying simpler material previously heard…. Finally,
variation could be elevated to the dominating structural principle of a musical work, resulting
in the well-known form of theme and variations. Here, both pattern and elaboration are
experienced in a temporal series. (Sisman 1993, ix–x).

Variation seems straightforward; yet, cognitively-speaking, how do listeners categorize different

variations as one thematic category?

As Ivanovitch (2010) implies, scholars assume a theme is a list of features that may or may

not be changed over time in variation. According to Nelson (1948), the variation theme “may be

regarded as a complex of separate, through interdependent elements. Conspicuous among these

elements are melody, bass, harmony, structure (meaning the plan of parts and phrases), tempo,

dynamics, rhythm, and instrumental tone color” (7). “Every element” the theme has can change,

either in terms of a “slight modification” (“simple embellishment or decoration”) to “radical

alterations that could “threaten the recognizability of the original element or even destroy it

completely” (Nelson 1948, 7). Instead of being “single changes” or “groups of changes,” Nelson

(1948) argues, variation means “combinations of changes with constants” (8). As part of my analyses

(chapter 5), I study combination of changes to measure musical distance between variations (which

are exemplars of a thematic category).


69

When humans listen to a theme, Ivanovitch (2010) describes the process as listening for

“potential” (5). Ivanovitch (2010) claims that variations relies on listener engagement:

Creative engagement on the part of the listener is always vital in the realm of variation. To
hear an element of the theme where none literally exists on the page, to imbue a scale with
the properties and implications of a theme, indeed, to hear a theme or variation at all – these
are crucial aspects of the environment of variation that take place with the creative
cooperation of the listener; they are not to be found by pointing at notes. Variation is, in
fact, a matter of trust, a contract between composer and listener, who both agree to behave
in the “variation manner”…. Listeners who become “preoccupied” with variation allow the
composer to lead them through a musical world in which relations between two musical
objects are not measured according to some finely calibrated absolute scale, but are governed
by strategic reassurances (which is why the beginnings and endings of variations are so
important), general shapes, and a willingness to exert one’s mental faculties to fill in the
blanks. (28)

Thus, Ivanovitch (2010) hints at elements important to analogy: a contract between composer and

listener (a focus on listener or composer goals) as well as engagement with the process.

Although several musical forms depends on variations of a theme (e.g. sonata form; Rosen

1976), I limit my analyses to the theme and variations form. The Oxford Music Online defines the

theme and variation form as such:

A self-contained theme is repeated and changed in some way with each successive
statement…. As a genre, strophic theme and variations has often had a poor reputation. This
is in part because its form is paratactic (a chain of separable links) and can therefore seem
like a loose assemblage of small pieces without a coherent shape. However, many composers
have grouped individual variations to create larger-scale musical forms and rhetorical
patterns…. Variation form is unusual in furnishing both the most vacuous and some of the
most profound examples of Western instrumental music. (Jones 2014)

Sisman (1993), arguing that variation form has mostly been denigrated by music scholars, insists that

“the ornamental and decorative techniques assumed to prevail in the variations of Haydn and

Mozart are considered “surface” features, failing to penetrate and transform the thematic model like

“deeper” contrapuntal, characteristic, developmental, or transformational techniques” (1–2). In the

eighteenth-century, aesthetic acceptance of the variation form fell away with emergence of the

“organic metaphor in the arts, in the works of Schlegel, Coleridge, and many others” (Sisman 1993,
70

15). Unlike sonata form, variations could not equate with “growing” or “becoming”; individual

variations could not be considered a continuous process of growth, even with the theme serving as a

metaphorical “seed” (Sisman 1993, 15).

The theme and variations form as a whole covers a surprising array of types. To have a

clearer understanding of these different theme and variation form types, two different typologies

emerged in the twentieth-century. The first, by Nelson (1948), groups variation types

chronologically, as seen here:

1. Renaissance and baroque variations on secular songs, dances, and arias:


“In pieces of this kind, the primary aim was the decoration and embellishment of the theme
by means of new figurations and counterpoints” (Nelson 1948, 3–4).

2. Renaissance and baroque variations on plain songs and chorales.

3. The baroque basso ostinato variation.

4. The ornamental variation of the eighteenth and nineteenth centuries:


“Successor to renaissance and baroque variations on secular songs, dances, and arias. Both
types aim at the figural decoration of the theme, as indicated above, and the distinction
between them is therefore primarily stylistic…. The stylistic divergence between the two
classes is too wide to be ignored, and in this study ornamental variation will be used in the
restrictive sense of a specific eighteenth- and nineteenth-century decorative type” (Nelson
1948, 5, italics in original).

5. The nineteenth-century character variation:


“Whereas previous variations end to preserve the expression of the theme throughout a
series, the separate members of the character variation frequently alter the expression, or
“character,” of the theme profoundly” (Nelson 1948, 5).

6. The nineteenth-century basso ostinato variation.

7. The free variation of the late nineteenth and early twentieth centuries.

The second, by Sisman (1993; 2014), groups variation types by musical elements, though the list is

still chronological:

1. Ostinato variations:
71

“Built upon a short pattern of notes, usually in the bass register, which functions as an
ostinato or ground bass, this type includes continuous variations of late 16th- and 17th-
century dance frameworks” (Sisman 2014).

2. Constant-melody or cantus firmus variations:


“The former is the broader category of which the latter is a historical instance. A melody,
usually widely known, appears intact or with only slight embellishments in every variation,
moving from voice to voice in the texture” (Sisman 2014).

3. Constant-harmony variations:
“This broad category includes many variation sets of the 16th, 17th and 18th centuries in which
the harmonic progression takes precedence in retentive power over the melody. The more
sectional harmonic-metric schemes of Italian and Spanish dance frameworks, such as the
folia and the Romanesca, may be included here, as well as such topical, expressive and
contrapuntal Baroque variation sets” (Sisman 2014).

4. Melodic-outline variations:
“The theme’s melody, or at least the ‘outline’ of its main notes, is recognizable despite
figuration, simplification (unfigured variation) or rhythmic recasting” (Sisman 2014).

5. Formal-outline variations:
“Aspects of the theme’s form and phrase structure are the only features to remain constantly
in this predominantly 19th-century type” (Sisman 2014).

6. Characteristic variations:
“Individual numbers take on the character of different dance pieces, national styles or
programmatic associations” (Sisman 2014).

7. Fantasy variations:
“In this 19th- and 20th-century type, occasionally used as a title, the variations allude to or
develop elements of the theme, especially its melodic motifs, often departing from any clear
structural similarity with it” (Sisman 2014).

8. Serial variations:
“Modification of a serial theme (a 12-note row or some slightly longer or shorter
configuration) in which figuration and accompaniment are derived from the row” (Sisman
2014).

As implied by these taxonomies, the theme and variation form changed over time. Thus, a scholar

could more closely examine shifts from one type to another over time.

I focus, in my analyses, on how composers shift their understanding of variation from the

late eighteenth- to the early-nineteenth-century. In Nelson’s language, this would be the shift from
72

ornamental variation to character variation; or, in Sisman’s language, the shift from melodic-outline

variations to either formal-outline, characteristic or fantasy variations. Variations are a profitable area

for studying analogy in music as they depend on a category (theme) that has been altered on the

surface, but still heard as the same due to shared relations. Thereore, a theme could be a relational

category. Also, the genre as a whole prompts listeners to compare variations, meaning that listeners

are more likely to align relational structure.

Meyer and Gjerdingen’s Schemata

Leonard Meyer applied the psychological concept of schemata to music by writing about

patterns that listeners implicitly learn:

Schemata are patterns that, because they are congruent both with human
perceptual/cognitive capacities and with prevalent stylistic (musical and extramusical)
constraints, are memorable, tend to remain stable over time, and are therefore replicated
with particular frequency. (Meyer 1989, 51)

Robert Gjerdingen (2007a) defines “schema” as a “packet of knowledge, be it an abstracted

prototype, a well-learned exemplar, a theory intuited about the nature of things and their meanings,

or just the attunement of a cluster of cortical neurons to some regularity in the environment” (11).

Similarly, Byros (2012) claims that it is a “mentally abstracted prototype of a statistical regularity in a

particular musical style which forms the basis for apprehending future phenomena” (280). Since

musical patterns are style-specific, schemata are mental representations of these patterns as governed

by the grammar conventions of a specific style (Meyer 1973, 27). A “replicated pattern,” such as a

musical schema, is a “culturally and historically determined category of mind” (Byros 2012, 306).

The fact that composers replicate patterns has “direct consequence” on the creation of a mental

category (Byros 2012, 282). Yet, schemata should not be static, but active and engaging.
73

In retrospect, the history of music theory may consider Meyer to be the “father” of schema

theory. He described voice-leading or contrapuntal schemata, solidifying this type as the most

commonly discussed schemata in the field. At an earlier point in his career, Meyer used the word

“habit” as opposed to “schema.” For Meyer (1973), norms (or patterns) are the “rules of the game,”

the way a person perceives music (213). In Explaining Music, Meyer discusses how patterns, or

schemata, are events which may imply a particular continuation, striving for closure or stability. He

calls these “implicative relationships” (Meyer 1973, 110), where an implication may or may not be

realized (actualized). By the time of his 1989 book Style and Music¸ Meyer shifted from using “habit”

to “schema,” briefly flirting with the term archetype in his article “Exploiting Limits” (Meyer 1980)

and Explaining Music. In Style and Music, Meyer corrects his understanding and use of the terms

“schemata” and “archetype”:

I have discussed the nature and function off schemata in Explaining Music, pp. 213-26,
“Exploiting Limits,” and (with Burton S. Rosner) “Melodic Processes.” In these studies I
usually referred to such stable, replicated patterns as archetypes. I prefer the term schema,
however, not only because it is commonly used in cognitive psychology, but because there is
a possible confusion with Jungian psychology, which uses archetype to refer to presumably
innate universals. But as far as I can see, the schemata of, say, tonal music are significantly a
matter of learning; that is, they arise on the levels of style rules, not cognitive universals.
(Meyer 1989, 50 n31, italics in original)

In “Exploiting Limits,” Meyer writes about schemata (though, here he uses “archetype”) as part of

style change (Meyer 1980).8 Constraints change over time and patterns do as well; however, some

patterns seem more “archetypal” within a cultural tradition such as Western tonal music (Meyer [1980]

2000, 194). Schemata then provide a listener with a type of “context” (Gjerdingen 1988, 6) in order

to make sense of a piece.

8 All page numbers will be the reprinting in Meyer’s Collection of essays The Spheres of Music (2000).
74

Robert Gjerdingen extended Meyer’s work in his galant9 music project, beginning with his

1988 Classic Turn of Phrase which investigates a certain schema of Meyer’s from 1700 to 1900

(Gjerdingen 2007a; Gjerdingen 1988). He shows that the schema event is not characterized by any

single feature, but is a “coordinated set of movements” (Gjerdingen 1988, 64). In identification,

Gjerdingen calls for a “reciprocal relationships” between features and schemata (Gjerdingen 1988,

6). Features are cues to the selection of schemata; however, schemata serve to detect features.

Corpus studies give scholars a way to identify a schema in many different compositions in

order to amass statistical evidence for its presence (Gjerdingen 1988, 34). In Music in the Galant Style,

Gjerdingen discusses many different schemata found in this era of music (which he defines as 1720

to 1780). He writes that the goal is to help people:

Hear this music more as Mozart might have heard it, to imagine musical behaviors more
consonant with the premises and goals of those who lived at galant courts, and to seek a
more realistic account of how galant musical craftsman fashioned raw tones into finished art.
(Gjerdingen 2007a, 452).

9 Many scholars often restrict schema theory analyses to voice-leading patterns in eighteenth-century music. Others,
however, have extended this ideology to musical elements outside of voice-leading (rhythm in particular) and to other
musical styles. Benjamin Anderson (2012) adopts schema theory methodology to explore voice-leading/harmonic
progression patterns in the style of popular music, specifically the music of Elton John. His corpus study concludes with
a theory of musical “archetypes:” patterns, stable over time, which persist and transcend style periods (Anderson 2012,
201). Common features, or central tendencies, emerge when style-specific information is removed from 1970s popular
music and galant style patterns (Anderson 2012, 203). Anderson identifies a 4-3-2-1 archetype as a result of comparing
the galant Prinner to an Elton John schema called the “Levon” (Anderson 2012, 205–208). Stefan Love’s (2012)
schematic analyses of Charlie Parker and jazz include both phrase structures (rhythmic elements) as well as “melodic
paths.” He compares the high-speed composition of galant composers to the improvisation of jazz musicians (especially
Charlie Parker). He identifies phrasing schemata, an adaptable template for beginning and endings phrases and a “path”
through the metrical structure of a blues chorus (Love 2012, 3.2). He follows this by looking at “melodic schemata” or
“recurring stepwise paths” which tend to appear in particular “zones” of the twelve-bar blues. (Love 2012, 4.1) These
melodic patterns, though having features like schemata, differ from the “precise” schemata in other corpus studies. For
instance, the main implications of Love’s “descent to 1” schema are to reach scale degree 1 and appear in a particular
“zone” of the form. It, prototypically, begins on 6, but could, theoretically, begin on another note. These melodic
schemata are more akin to gap-fill melodies (Meyer 1989; Meyer 1980) than the changing-note schema and other galant
schemata. They seem to “reach” up to a particular note before descending to fill in the leap. Love’s melodic schemata, in
this way, may be “plan-based” as opposed to “script-based.”
75

Overall, Gjerdingen writes that differences in learning and experience influence how listeners in

various time periods would interpret the same set of pitches and “could lead to somewhat different

meanings in different eras” (Gjerdingen 2010a, 62).

Gjerdingen (2007b; 2010b) connects the acquisition of schemata to the pedagogical style of

eighteenth-century Neapolitan conservatories. Through every exercise the student experiences, she

or he abstracts away features to create a schema. Pedagogical exercises—such as partimenti (see

Sanguinetti 2012)—provided conservatory students with thorough basslines to be realized. The

pedagogical exercises created a “rich store of memories” that a student could later draw on for an

artistic purpose (Gjerdingen 2007b, 115). A musician—through use and experiences—achieves

fluency in these stock musical phrases. A composer who is a “non-native speaker” and does not

have such fluency may use the schemata in a “non-normative manner,” analogous to a person

speaking English as a second language may misuse a common idiom such as “raining cats and dogs”

(Gjerdingen and Bourne 2015).

Other scholars extend the schema concept into other areas of eighteenth and early

nineteenth-century music. Byros (2009a; 2009b; 2012) discusses how eighteenth-century listeners

may use schemata to recognize the key of a piece. For instance, problematic sections in Beethoven’s

Eroica can be explained using schema theory. Byros (2012) writes that his approach is similar to

reader-response theories in literature and he believes his approach to be the “inverse process” of the

Penn School since he ascertains schemata by beginning with reception history as opposed to

corpora, or the “music itself” (286). He does, however, perform a large corpus-study (in addition to

looking at reception history) to pioneer a schema which he calls the le-sol-fi-sol (Byros 2009b; Byros

2009a). The le-sol-fi-sol illustrates why listeners in the nineteenth-century heard beginning measures of

Beethoven’s Eroica in the key of g minor while contemporary theorists hear the same passage Eb
76

major. The le-sol-fi-sol – a four-event schema in two stages – has a “dominant” chord orientation

(revolving around the 5 scale degree) and – in turn – has key defining characteristics (Byros 2012,

292). The strong key-defining profile is evidenced by the “normative closing usage,” consistently

creating a cadence (particularly a half cadence) or a cadential-type function (Byros 2012, 295). As a

way to further extend schema theory, some scholars identify rhythmic schemata. Burns (2010), for

instance, writes about six rhythmic archetypes in African music, while Ito (2013) writes about

hypermetrical schemata

A schema, a base abstraction, depends on relations between scale-degrees, contour and

harmonic functions; thus, a listener could categorize it using analogy. Goldwater, Markman, and

Stilwell (2011) refer to relational categories as “schema-governed categories,” providing further

evidence for musical schemata fitting in this category type (361). In addition, Holyoak (2012) also

label relational categories “schemas” (Holyoak 2012, 235). A contrapuntal musical schema

discussed—as an abstraction of a category of situations—depends on relational information, so a

listener may use the cognitive process of analogy to perceive it.

Hanninen’s Associative Sets

The music theorist Dora Hanninen (2004; 2012) has created a formalistic analytical system

that relies on relations between musical elements; therefore, I briefly review her understanding of

musical relations. Hanninen (2004) develops a theory of association using “associative sets,” which is

a “set of segments interrelated by contextual criteria – a category in the context of music analysis”

(150–151). A “contextual criterion” is a “rationale for segmentation that identifies an association

between two or more segments by repetition, equivalence, or similarity within a specific musical
77

context” (Hanninen 2004, 149). She defines “contextual criteria” further by claiming that it refers to

relational properties:

The properties that contextual criteria record are not merely predictable but relational: a
contextual criterion represents as association between segments as a property of (two or more)
segments…. Predictable properties such as a particular pitch-class set or rhythm become
significant when they associate two or more segments within a context under consideration –
that is, when they function as relational properties. (Hanninen 2012, 33, italics in original)

Contextual criteria are made up of contextual subtypes (C Subtype), which is “a subtype of

contextual criteria that names the musical space in which association occurs” (Hanninen 2012, 482).

Some examples of contextual subtypes include: pitch contour, pitch-class sets, pitch intervals, scale-

degree ordering and rhythm (Hanninen 2012, 36). For instance, the Cpitch <C4, D5, E5> indicates an

association between two or more instances of a pitch ordering of <C4, D5, E5> (Hanninen 2012,

36). The following (figure 2.9) are a selection of contextual subtype related to tonal music:

C Subtype Description and Comments Sample Individual Criterion


Cpitch Pitch set (ordered) Cpitch <C#3, A2>
Ccseg Pitch contour Ccseg <0132>
CSD Ordered set of scale degrees CSD <565>; CSD <543>, CSD <6543>
(uninterpreted)
CSD :a:<543> Ordered set of scale degrees, key CSD :a:<543>
specified (stronger than above)
CSDint Scale degree interval (+,- indicate CSDint <-3, -4, +2, -3> (NB: SDint <D4,
directed intervals) C4> = -2 not -1)
CSDintq Diatonic interval size and quality (+,- CSDintq <-M3, -d4, +m2, -M3>
indicate directed intervals)
CRN Roman numerals (uninterpreted) CRN, a: <I, ii6, V, I>
CHF Harmonic functions (T=tonic; CHF <T, PD, D, T>
P=predominant; D=dominant)
(uninterpreted)

Figure 2.9: Contextual subtypes list, re-constructed from Hanninen (2012, 37-38)
78

As contextual subtypes are relational structure in music, I incorporate some of these subtypes into

my understanding of musical relations (see chapter 3).

Similarity and Categorization in Music Cognition

Since listeners use similarity to extract information from music (Reybrouck 2009),

understanding musical similarity is important not just to music theory, but also music cognition.

McAdams et al. (2004) maintain that:

The perception of similarity between musical materials is a crucial topic in the field of music
psychology because it underlies a large part of the listener’s musical experience, including the
perception of associations between themes or motifs and their variations, the formation of
musical categories, and the sense of familiarity. (207–208)

Special issues of journals dedicated to these topics reflect a growing interest within the field of music

cognition. Music Perception dedicated a 2001 special issue and Musicae Scientiae followed suit with the

2007 and 2009 discussion forums.

A number of experiments have studied how listeners perceive similarity (Lamont and

Dibben 2001) and categorize different musical themes (Deliège 1996; Eitan and Granot 2009;

Pollard-Gott 1983; Ziv and Eitan 2007). Although there are a variety of music-theoretic categories,

as elaborated upon above, most experimenters have chosen “themes” for stimuli in music cognition

experiments. I elaborate upon the main experiments in this area, and note that participants in these

experiments do not categorize themes based on primary parameters as music theorists expect.

Pollard-Gott (1983)

Pollard-Gott (1983) asked musicians (including two experts) and non-musicians to listen to

Franz Liszt’s Sonata in B minor, which employs two main themes (theme A and theme B). To begin,

participants listened and took notes on the first half of the sonata. After this listening phase,
79

participants rated similarity (on an 11-point Likert scale) of pairs of eight target passages (four

varitions on each of the two themes). After similarity data was collected, the experimenters told

participants that passages were variations on two themes. At this point, subjects heard one example

of theme A and one example of theme B. Subjects, then, in a forced-choice task, categorized the

same eight target passages as well as unheard themes from the sonata’s second half as either theme

A, theme B, or neither. There were three conditions: a single-listening condition (subjects came in

for one session as described above), a repeated-listening condition (subjects came in three separate

days within a week, but did the same tasks each time), and an “expert” condition (musicians who

had played the piece and were familiar with it). The experts relied on primary parameters for

similarity and categorization. However, all listeners (musicians and non-musicians) in the single-

listening condition rated similarity between excerpts, and categorized them, based on secondary

parameters, such as dynamics and register. Participants in the repeated-listening condition began by

using secondary parameters to judge similarity, but changed dimensions over the course of the

sessions. By the end, they used more primary parameters. For listeners in this repeated-listening

condition, Pollard-Gott (1983) contended that, “the relationships that listeners

perceived…corresponded with higher order thematic structure after repeated exposure, but not after

a single exposure to the music” (92). Only after an increase in familiarity were listeners able to base

their similarity (and categorization) on primary parameters as opposed to secondary parameters.

Bigand (1990)

In Bigand (1990), experimenters constructed a family of four melodies (a1, b1, c1, d1) with

different rhythmic-melodic contours (a, b, c, d) but the same underlying harmonic structure (1).

They applied the same rhythmic-melodic contours to a different harmonic structure (2) to create
80

Family 2 melodies (a2, b2, c2, d2). The listeners heard family 1 (a1, b1, c1, d1) twice with instruction

that these melodies were related to each other in the same way. Participants were told to listen

carefully because they would hear the same melodies again as well as four other unrelated, but

similar, melodies. After the listening phase, participants heard eight melodies (4 old and 4 new) in

the test phase. Participants indicated whether the melody in the test phase belonged to the family

heard in the listening phase or not. Bigand (1990) calculated the number of errors for each

participant’s answers. In one condition the family was a True Family (a1, b1, c1, d1), while the

second condition was a False Family (a1, b2, c1, d2). The False Family controlled for memorization;

however, this meant that the melodies in the False Family were not actually members of the same

category. Not surprisingly, the subjects made fewer mistakes when a True Family was presented.

Bigand (1990) concluded that a listener was able to “abstract an underlying structure common to

four melodies that have different rhythmic-melodic contours…. A listener can go beyond the level

of superficial organization to reduce the musical surfaces to a minimal underlying structure

represented them”10 (56).

Lamont and Dibben (2001)

Lamont and Dibben (2001) experimented with thematic similarity. They used Beethoven’s

piano sonata, Op. 10, No. 1, first movement for musical stimuli. Similar to Pollard-Gott’s (1983)

10 A confound with this experiment is that listeners heard the exact same stimuli in the testing phase as in the listening
phase. The participants were told that any new melodies heard in the testing phase would not be part of the Family
heard in the listening phase. Instead of abstracting the underlying harmonic structure as Bigand (1990) concluded,
listeners could have used memorization. If they realized during the testing phase that they had not heard the melody
before, they knew to classify the theme as not part of the Family. If they had heard the melody, they knew that it was
part of the Family. Bigand (1990) said memorization is controlled; however, the one control for memorization was
leading the subjects astray with a “False Family,” a “false” category. A more effective control for memorization would
have been presenting all new melodies in the testing phase (four different melodies that had the same underlying
harmonic structure as the Family in the listening phase and the four melodies with the same rhythmic-melodic contour
heard in the listening phase but with a different underlying harmonic structure).
81

Liszt piece, this movement had two different themes (theme A and theme B) with variations. After

hearing the Beethoven piece in its entirety, participants heard thirty-six pairs of themes (each theme

paired with every other theme on the list). All participants rated each pair for their similarity on an

11-point Likert scale. Lamont and Dibben (2001) found that secondary parameters, such as

dynamics, articulation and texture, were the principal dimension of thematic similarity. They had

only a single-listening condition, however, and noted that their results may change with repeated

listening or increased familiarity (Lamont and Dibben 2001, 263).

Ziv and Eitan (2007)

Ziv and Eitan (2007) used the same material as Lament and Dibben (2001) to see if they

could gather different results. Instead of asking about similarity, they decided to take the route of

categorization. They hypothesized that “thematic categorizations based on representations in long-

term memory, would emphasize deeper-level theory based features more than similarity ratings using

short-term memory traces (Ziv and Eitan 2007, 119). They found that features listeners used to

categorize included aspects of texture, melodic contour, rhythm and dynamics, while some “surface

aspects of pitch structure,” such as mode, harmonic color, and common pitch configurations also

may play a part (Ziv and Eitan 2007, 120–121). They found that, “As in L&D’s [Lamont and

Dibben’s] similarity judgments, structural pitch-based structures preeminent in music analysis ---

harmonic, intervallic, or (in Schoenberg) serial – do not seem to play a major role in listener’s

thematic categorizations” (Ziv and Eitan 2007, 121).


82

Eitan and Granot (2009)

Eitan and Granot (2009) manipulated both secondary and primary parameters to see which

made more of an impact when participants categorized themes. They created their own stimuli to

control the manipulations of different melodies. In the first experiment, Eitan and Granot (2009)

manipulated three variables: interval class, pitch contour, and “expression” (a compound of

secondary parameters, including dynamics, texture, articulation, and register). They wanted to know

which (if any) of the manipulated variables would influence listeners in classifying musical themes.

As in previous experiments, Eitan and Granot (2009) showed that “motivic classification can often

be related to similarities and differences in secondary parameters – general auditory features, not

specifically related to music” (163). The variables of the second experiment were: pitch intervals

(instead of interval class), rhythm, and “expression” (the same secondary parameters). Participants

tended to use “expression” in categorization. However, the second experiment demonstrated that

musically training participants used rhythm more than “expression,” while non-musicians continued

to use “expression.”

A Common Theme

A common theme (pun slightly intended) is that listeners more often use secondary

parameters (or perceptual features) to categorize themes/motives than primary parameters (or

relations). This finding contradicts music theorists’ ideas on what listeners should use to categorize

themes/motives. Eitan and Granot (2009) concede that:

Our results seem to problematize [music] theorists’ notion of the hierarchy of musical
dimensions shaping motivic identity: pitch intervals or IC [interval class], dimensions
supposedly central in determining motivic identity, were marginalized, and secondary
parameters, assumed to be marginal, were centralized. Correspondingly, the processing of
Express-based categorization was fastest, while that of pitch-based categorization was
slowest. (163, 165)
83

Ziv and Eitan (2007) similarly claim that:

Studies of motivic-thematic perception suggest that features considered structurally


important by music theorists are often not those selected by listeners (including experienced
musicians) in categorizing musical events or rating the similarities among them. (102)

Lamont and Dibben (2001), in yet another similar statement, also admit:

There is no evidence to suggest that listeners are using thematic or motivic similarities
[primary parameters], even listeners with musical training … listeners prioritized the more
surface features. (263)

To add to the collection, McAdams et al. (2004) acknowledge that listeners prefer to use surface

features to determine musical similarity (231). Thus, there seems to be a “thematic paradox”:

listeners in experimental settings do not use musical dimensions that music theorists would expect to

categorize themes.

Of the different perceptual features, scholars note that a few are more prominent in

categorization than others. In McAdams et al. (2004), participants, according to their verbalizations,

used tempo, regularity of durational patterns, syncopation, register, contour, repetition, directionality

of gestures, continuity/fragmentation, fluidity/choppiness among others (231). For the Beethoven

piece in Lament and Dibben (2001), dimensions of similarity included dynamics, articulation and

texture.

Conclusion

In research question 1, I mentioned experiments, which I discussed here in depth, that show

participants do not often categorize musical themes based on primary parameters. If musical themes

are based on relational structure (or relational categories), then participants in these experiments do

not notice shared relations in the musical stimuli. They are not making the analogy necessary to hear

similarity between musical stimuli. Many articles have illustrated the necessity of comparison in

making analogies (e.g. Markman and Gentner 1993; Medin, Goldstone, and Gentner 1993; Gentner
84

and Namy 1999). To categorize musical themes, listeners may need to compare musical passages to

recognize relational structure, assuming that themes are relational categories. To demonstrate that a

theme is a relational category, I will analyze theme and variations movements “in the wild” as a real

world laboratory (see chapter 5). If themes are relational categories, then “analogy” explains

participant behavior in music categorization experiments. In research question 2, I ask how theorists

use an analogy between music and another domain to make sense of unexpected findings. In this

chapter, I overviewed how humans use analogy to make inferences as well as connect a source and a

target, even ones from different domains. I use this knowledge to make a case for using analogy for

the perception of irony in certain musical movements (chapter 4). To answer both research

questions, though, a theoretical framework for analogy in music listening is needed. In the following

chapter, I create this framework based on literature covered here.


85

CHAPTER 3
“Knowledge about a Thing is Knowledge of its Relations”: An
Analogy Framework for Music Analysis

Knowledge about a thing is knowledge of its relations.


—William James, The Stream of Consciousness

The White Rabbit put on his spectacles. “Where shall I begin, please your Majesty?” he
asked.
“Begin at the beginning,” said the King gravely, “and go on till you come to the end: then
stop.”
—Lewis Carroll, Alice in Wonderland and Through the Looking-Glass and
What Alice Found There

The preceding chapters have been groundwork for the dissertation’s more complicated task,

which is to consider how listeners use the cognitive process of analogy to make sense of sequences

of music. In the published version of his address to the American Psychological Association,

Oppenheimer (1956) wrote: “we come to new things in science with what equipment we have… we

cannot, coming into something new, deal with it except on the basis of the familiar and the old-

fashioned” (129). By replacing “science” with “music,” the quote becomes, “we come to new things

in music with what equipment we have… we cannot, coming into something new, deal with it

except on the basis of the familiar and the old-fashioned.” Analogy lets listeners use what equipment

they have to understand music in the moment. In the first chapter, I outlined research questions and

hypotheses related to musical analogy. In the second chapter, I reviewed scholarly literature related

to analogy in both cognitive psychology and music. Now, I proceed to the core of this dissertation,

which is creating a theoretical framework for analogy in listening by contextualizing analogy from
86

cognitive psychology in music. What is musical retrieval, mapping and evaluation? What is a musical

relation? This theoretical approach, based on Structure-Mapping Theory (Gentner 1983), will be

used to analyze musical “texts” so to answer my research questions. Before explaining the

framework, I first define a “musical event.”

Musical Event

I define a “musical event” as a discrete and identifiable musical unit (e.g. Hanninen 2012,

12). It occurs within a “perceptually significant time span” (Caplin 1998, 9) and its perception as an

individual segment depends on listeners dividing “a continuous spectrum into discrete bins”

(Margulis 2013, 37). This division, or segmentation, depends on a listener assembling the “sound in

groups on the basis of their temporal and/or acoustic properties” (Deliège and Mélen 1997, 391).

Yet, a musical event is not only “raw” musical “properties” (e.g. pitch, duration, etc.), but also

socially and historically constructed (Dibben 2003, 196). I use “musical event” to discuss any

identifiable, discrete, and readily perceptible musical unit in the abstract.

The Analogy Framework

The theoretical framework for analogy during musical listening (or the analogy framework) is

based on research in cognitive psychology on analogy. Figure 3.1 schematizes components of the

analogy framework.
87

Figure 3.1: The Analogy Framework

To create this framework, steps of analogy—retrieval, mapping, and evaluation—are applied to

listening to music. A listener may not realize he or she is completing each analogical step. Most

music listening proceeds without conscious attention. Likewise, the cognitive process of analogy in

music listening often proceeds automatically and without conscious attention. Although, a listener

may be aware of interpretations formed from the evaluation step, he or she may not be aware of

how these interpretations came to be.

Step 1: Retrieval

Analogical retrieval in music involves a listener being reminded of relationally-similar musical

or non-musical analogs, depending on knowledge structures formed through musical exposure. As

Gentner and Smith (2013) note, analogical reasoning does not always means “retrieval” in the

definition of “when a person is reminded of a prior relationally similar case” (4). Though a person

can be reminded of a previous analog, sometimes both analogs are present in working memory

and/or physically, as, for example, in instructional analogies (“electric current is like waterflow”) or

persuasive analogies (“Afghanistan is like Vietnam”) (Gentner and Smith 2013, 4). Seeing that

listeners could either be reminded of a past/familiar analog stored in long-term memory or


88

encounter both analogs together, a scholar using the retrieval step asks either one or two question:

he or she asks two questions if one analog is a reminding and one question if both analogs are

encountered together. If one analog is retrieved from long-term memory, the first question is: “What

relationally similar situation (in music or otherwise) could a listener be reminded of?” Whether an analog is

retrieved from long-term memory or both analogs are encountered together, an analyst using the

retrieval step should always ask the following question: “What knowledge does a listener have about these

two analogs?”

Some music theorists may consider only the context of the musical work itself to answer

these questions. For example, relationally-similar analogs come only from other sections of the work

and listener knowledge consists of knowledge only of that piece. Dan Harrison (2000) affirms these

assumptions when he maintains music theory’s founding principles include: “A musical object can

be abstracted from history for inspection of parts that are unaffected by the passage of time” and “a

system of tonal organization, though it emerge in some historical period, can be regarded as timeless

and, hence, ahistorical” (31). To begin his description of “the musical object,” Matthew Butterfield

(2002) describes the danger of an analytical approach where “a musical work is usually treated as an

entity that is self-identical through time and putatively exists fully independent of the real contexts in

which it is encountered” (327). These quotations illustrate a stance of historical de-contextualization

in music theory. If they de-contextualize a musical work, then these theorists would de-contextualize

a hypothesized listener as well.

In contrast, I argue that music theorists should consider the cultural context of listeners, and

the questions of the retrieval step give them the opportunity to do so. Context for interpreting a

piece should come not just from the confines of a “musical work,” but also pieces, associations,

conventions and other musical experiences familiar to a listener in his or her time and place. The
89

cultural context of a listener—including experiences of this work, but also other pieces or

conventions—impacts potential musical analogs retrieved and their associations. The analogy

framework gives theorists space to consider who is listening (his or her context) and his or her

knowledge structures. It accounts for listener context by incorporating knowledge structures a

listener may have learned through musical use or exposure. Analogies mean making connections

between relations in knowledge structures. An analogy framework provides a systematic way to

consider how listeners connect relational knowledge, and evaluate inferences made from these

connections (see chapter 4 for analyses).

To systematize the retrieval step in analysis, I divide the retrieval step into sections and sub-

sections. The focus of systematization is always on relationships instead of objects themselves. The

primary overarching categories are “listener cultural context” and “perceptual capabilities.”

However, I limit my exploration to “listener cultural context” (figure 3.2) because of my interest in

cultural context.

Figure 3.2: Sections and sub-sections of the retrieval step


90

A seen in figure 3.2, the two main sections are i) other domains, and ii) musical use/exposure. Sometimes

listeners do not retrieve a musical analog, but one from another domain instead. Humans in Western

culture gravitate toward making analogies between music and certain other domains, including

language (e.g. Swain 1997) and bodily gestures (e.g. Hatten 1994; Brower 2000; Larson 2012). Since

this section encompasses any analog outside of music, I do not discuss it further; but, focus instead

on analogs within music. I divide music into two subsections (figure 3.2): i) exposure to specific musical

pieces (relationships in an individual exemplar), and ii) conventional musical patterns (relationships in a

general conventional pattern). The first, exposure to specific musical pieces, is more likely to map onto

intra-opus patterns, while the second is more likely to map onto extra-opus patterns. Under

conventional musical patterns, I list form, voice-leading harmonic patterns, and musical associations (figure

3.2). This list is not exhaustive by any means. However, it reflects areas of listener knowledge that

motivate analogical choices as well as conventional patterns in my analyses. In the following

paragraphs, I discuss musical use/exposure, exposure to specific musical pieces, and conventional musical patterns.

Musical Use/Exposure

A listener learns musical knowledge structures, often unconsciously, through exposure to or

use of music. In Explaining Music, Meyer (1973) writes: “Understanding music, to paraphrase what

Bertrand Russell has said of language, is not a matter of knowing the technical terms of music

theory, but of habits correctly acquired in oneself and rightly presumed in others” (16). Hatten

(1994), in his book Musial Meaning in Beethoven, holds a similar sentiment, “How, then, does a musical

style become understandable for a listener? Much of the general expressive significance of tonal

music did not have to be directly taught; it was enculturated through progressive inferences about its
91

use and contexts” (248). Both Meyer and Hatten argue that implicit and unconscious musical

exposure influences how listeners learn and understand musical styles.

Yet, musical use influences not just knowledge listeners learn, but also “meanings” and

“communication” ascribed to musical knowledge and pieces. In Emotion and Meaning in Music, Meyer

(1956) first mentions “habits” of listening as a way listeners create musical meaning: “the norms and

deviants of a style upon which expectation and consequently meaning are based are to be found in

the habit responses of listeners who have learned to understand these relationships” (61). “Musical

use” is another way to say “historical contextualization,” but from the listeners’ perspective.

Historical contextualization, according to William Sewell (2005), implies, “We cannot know what an

act or an utterances means and what its consequences might be without knowing the semantics, the

technologies, the conventions—in brief, the logics—that characterize the world in which the action

takes place” (10). A few scholars, often inspired by reader-response theory in literature (e.g. Fish

1980; Iser 1978), have begun to examine how experiences of a listener change his or her

understanding of musical communication. In reader-response perspectives, the “texts were as much

a product of their readers as of their writers… All of these scholars realized that the prior knowledge

and unique frames of reference possessed by individual readers and listeners must substantially

affect their experience of a text” (Gjerdingen 2010a, 62). Gjerdingen (2010a) transfers this ideology

to music by arguing that “people living in different eras will likely develop slightly different schemata

owing to their differing experiences” (62). He penned similar thoughts more than ten years earlier,

writing:

“Native listeners” of eighteenth-century court music – whether we mean by that term the
deceased members of those courts or modern listeners who have immersed themselves in
the galant style to the point of acquiring it as a second language – hear in its compositions
discrete chunks that match memories of meaningful gestures and phrases. Other, more
casual listeners will perceive a pleasant flow of tones, gross changes in texture and dynamics,
and those elements of musical syntax that may transcend the period in question. Because
92

their prior musical experiences differ substantially, these two classes of listeners will have
different musical perceptions of the same piece. (Gjerdingen 1996, 380)

The retrieval step analyzes how musical use—or listeners’ historical contextualization—guides

analogical choice in musical perception as well as illustrates how listeners build a musical category or

abstract schema. A listener, when being reminded of a musical analog, could draw from two main

sources of experience: exposure to a specific musical piece (an individual exemplar) or exposure to

conventional musical patterns (a category composed of multiple exemplars).

Exposure to Specific Musical Pieces

A listener could be reminded of, and so retrieve, an individual exemplar of music—a specific

musical piece—and map its relations onto a current analog. According to Edward Cone (1974),

musicians especially rely on a “musical memory,” a store of these individual exemplars of music

ready to be retrieved: “When he [a musician] tries to verbalize the significance of a certain

composition, he often does so by relating it to other musical works” (Cone 1974, 173). If a musical

work alludes to a past work, then listeners connect the past work to the current one.

However, the way someone learns this musical piece may influence the inferences or

understandings he or she has of it. For example, in the episode “The One after Joey and Rachel

Kiss” (Bright 2003) of the TV show Friends, the character Monica (Courtney Cox Arquette) starts

humming Wagner’s “Ride of the Valkyries.” Her husband Chandler (Matthew Perry) responds:

CHANDLER: What are you humming?


MONICA: It’s “Bolero” from 10.
CHANDLER: It’s “Ride of the Valkyries”….from Apocalypse Now!

The fact that Chandler’s exposure to “Ride of the Valkyries” is through the movie Apocalypse Now

(he may not even know it originates in an opera) influences his perception or associations of the
93

piece. If he were to retrieve it to as an analog, his inferences would stem from a war movie instead

of an opera.

Conventional Musical Patterns

A listener could be reminded of, and so retrieve, conventional musical patterns, which are

abstract categories created by exposure to multiple exemplars. Much of musical understanding

depends on knowing musical conventions.

Music comprehension depends to a considerable extent upon the listener’s knowing a


traditional tonal syntax and a set of conventional signs and schemata (even though these may
be grounded in and limited by the nature and capacity of the human ear and mind)…. The
existence of such conventions means that there has been a continuing tradition of musical
representation in which later manners of delineating a particular moral character, a kind of
event, an affective state, or some phenomenon in the physical or mythical world are based
upon and influenced by earlier ones. Just as there are histories of elegiac poetry and of carpe
diem lyrics, of paintings of the Annunciation and of pastoral scenes, so there would seem to
be histories of battle music, pastoral music, love music, lamentation music, and so on.
(Meyer 1973, 69)

Much of the music in this list have both an “artistic” side and a “real world” or “functional” side,

indicating that some conventions may be born more out of necessity then “art.” For instance, there

is music that has been historically played in actual battle, music that has historically been played in

“virtual” battle (opera, moves, etc.), and instrumental music that sounds enough like battle music

that people think “battle” when they hear it. Exposure to a musical style means exposure to its

conventions, implications and associations.

Listeners retrieve from long-term memory—either consciously or unconsciously—

conventions based on relational structure. They then make analogies between a current musical

passage and this convention. In my examples, I often limit my analysis of conventional patterns to

phrasal conventions (such as galant schemata) and musical form conventions (such as sonata form).

In the second chapter, I described phrasal conventions. Here, I discuss why form and genre depend
94

on listener experience. According to Grove Music Online, a genre is defined as “a class, type or

category, sanctioned by convention” (Samson 2007). In particular, genre or form entails a set of

conventional expectations held by the listener (Caplin 1998; Hepokoski and Darcy 2006; Cobly

2005; Cobly 2008). Paul Cobly (2005) defines genre as “not a set of textual features that can be

enumerated; rather, it is an expectation” (41). According to James Hepokoski and Warren Darcy

(2006), “genres exist only insofar as production and reception communities agree to act as if they

really did exist, as sets of rules, assumptions or expectations” (606).

Comparing Listeners

Depending on a theorist’s goals, he or she could use the retrieval step to analyze an

individual listener or a group of listeners. The retrieval step is inherently flexible and can take into

account different listeners, even without altering subsections. For example, one could theoretically

compare an eighteenth-century listener to a twenty-first-century listener. Under the “voice-

leading/harmonic pattern” category, an eighteenth-century listener may have a strong representation

of a galant schema, yet a twenty-first-century listener may have a strong representation of a

conventional popular music harmonic pattern (e.g. I-vi-IV-V). The “item” in the voice-

leading/harmonic pattern category changed with the listener even though the category did not.

Step 2: Mapping

A theoretical framework for analogy in music based on Structure-Mapping Theory describes

how listeners align musical relations, and then project and evaluate inferences based on this

alignment. In the mapping step, listeners recognize relations between objects, then associate each

element in a relation of one structure with a corresponding element in a relation of a different

structure (figure 3.3).


95

Figure 3.3: Analogy as Structure-Mapping (graphic from Gentner and Smith 2012, 132)

When making musical analogies, listeners map relations between musical elements (note-to-note,

phrase-to-phrase, etc.) from a musical analog onto relations in either another musical analog or a

non-musical analog. First, I further define a musical relation (concept expressing a connection between

musical elements). Then, I discuss how listeners compare the source and the target (unidirectional

listening and bidirectional listening) and the distance there is between the two analogs (intra-opus, inter-

opus, intertextual, and extra-musical). Finally, I discuss mapping patterns in general and a taxonomy for

classifying pattern mappings.

Musical Relations

A musical relation is a concept expressing a connection between musical elements. Relations

always have multiple arguments and these arguments can be either objects or other relations. For

example, the following is a notation of relation x bigger than y, where x and y are objects.

BIGGER-THAN(x, y)

If x is a ball 1 and y is ball 2, then ball 1 is bigger than ball 2.

BIGGER-THAN(ball1, ball2)

For music, examples of pitch relations include higher to and lower to, while examples of rhythm

relations include longer to and shorter to. I purposely use “higher to” instead of “higher than” so I can list
96

arguments in the order they appear in the music while simultaneously describing where the relation

is headed (“going higher here”, etc.).

HIGHER-TO(C4, D4)

LOWER-TO(D4, C4)

SHORTER-TO(quarter-note, eighth-note)

LONGER-TO(eighth-note, quarter-note)

When arguments are relations instead of objects, these are called higher-order relations (Gentner

1983). For example, a rhythmic motive is a higher-order relation comprised of relations from every

note to its adjacent note. Consider the relations of this Clementi sonatina rhythmic motive.

RHYTHMIC-MOTIVE[SHORTER-TO(quarter-note, eighth-note), SAME-AS(eighth-note,

eighth-note), LONGER-TO(eighth-note, quarter-note), SAME-AS(quarter-note, quarter-

note)]

According to the systematicity principle (e.g. Clement and Gentner 1991), humans prefer to map

relations connected to higher-order relations. The higher order relation is RHYTHMIC-MOTIVE,

since it is a relation that has arguments that are other relations (e.g. SHORTER-TO(quarter-note,

eighth-note), etc.). Therefore, listeners are likely to map over the rhythmic motive as a higher-order

relation that encompasses the note-to-note relations instead of each note-to-note relation

individually.

The following are several constraints that limit perception of musical relations, and will be

discussed in turn.

i. Musical Parameter
ii. Order
97

iii. Contrast or change-over-time


iv. One-to-one mapping
v. Parallel connectivity
vi. Working Memory
vii. Active orientation through comparison

Musical parameter: relations form within a single musical parameter. For example, do to re

forms a relation (e.g. HIGHER-TO) while eighth-note to quarter note forms a different kind of

musical relation (e.g. LONGER-TO). Yet, listeners fail to recognize a relation between do and an

eighth-note since these objects exist between different musical parameters of pitch and rhythm

respectively. However, relations within a single musical parameter can map onto a relation within

another musical parameter. For example, Eitan and Granot (2007) illustrated that listeners mapped

the relation crescendo [GRADUAL-INCREASE-TO(soft, loud)] onto accelerando [GRADUAL-

INCREASE-TO(slow, fast)].

Order: the order of objects within a single musical parameter changes the relation. For

example, do  re creates a different relation than re  do, even though the same objects (do and re)

are used.

Contrast or Change-over-Time: A musical parameter changes over time. Music’s ability to

unfold in time differentiates it from most art forms, and so a musical relation takes place within a

time span. One exception may be an excuse that a score “represents” music. Even with a notated

score’s “static” nature, it still represents music occurring temporally.

One-to-One Mapping: Each of a representation’s elements match one element at most in

another representation. For example, a rhythmic motive half-notedotted-quartereighth-note

follows the relation gradually getting shorter:

RHYTHMIC-MOTIVE; GRADUALLY-GETTING-SHORTER [SHORTER-TO(half-

note, dotted-quarter), SHORTER-TO (dotted-quarter, eighth)]


98

In diminution, this gradually getting shorter relation is still represented even though the notes are now

quarter-notedotted-eighthsixteenth-note.

RHYTHMIC-MOTIVE; GRADUALLY-GETTING-SHORTER [SHORTER-TO(quarter-

note, dotted-eighth), SHORTER-TO (dotted-eighth, sixteenth)]

In the Art of Fugue excerpt in figure 3.4, Bach transformed the lower voice of the canon using

diminution; but, the duration’s relations are still the same and can be mapped one-to-one from the

upper voice to the lower voice.

Figure 3.4: Bach, Art of Fugue, Contrapuntus VII, mm. 2-4

Parallel Connectivity: if two relations correspond to each other, then the arguments (objects

or relations) also correspond to each other. In the Bach Art of Fugue example above (figure 3.4), the

arguments of the relation SHORTER-TO(half-note, dotted-quarter) in the upper voice correspond

to SHORTER-TO(quarter-note, dotted-eighth) in the lower voice. Upper voice half-note maps onto

lower voice quarter-note while upper voice dotted-quarter maps onto lower voice dotted-eighth.

However, upper voice half-note could not map onto lower voice dotted-eighth.

Working Memory: a listener can only make relations between objects retained in working

memory. Using Miller’s (1956) famous seven-plus-or-minus-two, I speculate that no more than

seven objects could create a relation.


99

Active Orientation through Comparison: a listener actively participates in the music through

creating and building relations while listening. One aspect of active participation is the act of

comparison, where a listener chooses musical objects and relations to compare.

Two Levels of Musical Relations: Relations within a Musical Event and Relations between Musical Events

Music can be relational on at least two levels: 1) a musical event defined by internal relations

within that event, and 2) relations between musical events. I deliberately chose prepositions within and

between to further demonstrate how musical relations can together create a shared system of relations.

As illustrated in figure 3.5, the musical event defined by the relations within can be the same musical

events with relations between.

Figure 3.5: Visual representation of two levels of musical relations: relations between musical events and relations

within a musical event

This hypothesis of relations in music as “within” and “between” draws inspiration from Gentner’s

(2005) discussion of relational categories. Gentner (2005) explains: “By relational categories, I mean

categories whose meaning consists either of (a) relations with other entities, as in predator or gift, or

(b) internal relations among a set of components, as in robbery or central force system” (245, italics in

original). In Gentner’s relational category (a), a category is formed based on its relation with some
100

other entity. For example, a gift is only a gift if an agent gives that object (a relation) to another agent.

It might be a baseball bat, a doll or a puppy. The example gift, an agent or entity has to “do

something” to that object in order it to be a member of that relational category. In a similar way,

relations between musical events entails something has been “done” to the musical event. The relation

does not come from that musical event alone, but only in relation to another entity; generally,

another musical event. In Gentner’s relational category (b), the relation is built into the definition of

the category. A robbery is a robbery when an agent takes something they do not own from another

agent. It is defined by internal relations between a set of components within the category. In a

similar way, internal relation within a musical event can define or create that musical event.

Relations within a Musical Event

Relations within a musical event are music-specific relations that are restricted to—and even

define—a musical event. Since I cannot list all possible relations within musical parameters, I only

list parameters that commonly form relations in music most applicable to relations within a musical

event. To create this list, I build off of Meyer’s (1989) primary and secondary parameters and

Hanninen’s (2012) contextual subtypes. Meyer’s primary and secondary parameters serve as an

“umbrella” for “relational” and “perceptual” features in music respectively, while Hanninen’s (2012)

contextual subtypes (figure 2.9) serve as inspiration for specific musical relations.

In eighteenth- and early nineteenth-century Western art music, composers often use

relations with Meyer’s (1989) primary parameters (e.g. melody, rhythm/meter, and harmony) to

define musical events more so than relations in secondary parameters (e.g. dynamics, tempo,

articulation, timbre, sonority, texture). Primary parameters create musical syntax (Meyer 1989), are

music-specific and are defined by learned and culture-specific higher-order sets of relations: tonality
101

and meter. For instance, Western pitch and functional harmony connect to a tonal hierarchy

(Krumhansl 1990; Agawu 1991), so tonal context organizes perception of individual notes. E4

functions differently in different contexts depending on key, style, and chord played. In F major, a

melody ending on E4 implies something unfinished and a desire to resolve to F4. Yet, a melody

ending on E4 in E major is already resolved. In an E major chord, E4 is consonant, but dissonant in

a B major chord. In a similar way, rhythm also connects to a hierarchy in meter (London 2004). Two

eighth-notes sound complete in 4/4 time signature, but the same two eighth notes sound vaguely

incomplete in a 6/8 time signature. Therefore, an eighth note functions differently depending on its

metrical context. Meyer (1998) further implies that primary parameters are based on functional

relationships: “In order for syntactic [primary parameter] relationships to arise, the elements

comprising a parameter must be related to one another in a functional way – e.g., leading-tone/tonic

or upbeat/downbeat.…. To take the clearest example, in any syntactic tonal system there are large

and smaller intervals” (Meyer 1998, 8). Elements of primary parameters are often used relationally

and can be used to make musical analogies.

In contrast, secondary parameters are “natural” aspects of auditory organization, not music-

specific and do not often connect to any certain higher-order sets of relations (Eitan and Granot

2009), especially the way they are used in eighteenth and early-nineteenth century art music. For

example, loudness and other secondary parameters are dimensions found in sounds not music such

as speech (e.g. Cruttenden 1997). Depending on their use, listeners can sometimes perceive

secondary parameters as relational. For example, a book falling on the floor is loud compared to the

sound of running water, but soft compared to an explosion. In an experiment by Eitan and Granot

(2007), results suggest listeners can perceive tempo and dynamics relationally if the musical material

heavily emphasizes the contrast. Meyer (1998) further implies that secondary parameters are
102

perceptual: “Dynamics may become louder or softer, tempi may be faster or slower, sonorities

thinner or thicker, and so on. But they cannot be segmented into perceptually discrete relationships.

Because they are experienced and conceptualized in terms of amount, rather than in terms of kinds

or classlike relationships (such as “major third” or “antecedent-consequent”)” (8–9). Since

secondary parameters are often not music-specific and accessible without context, then secondary

parameters are often treated as more “intrinsic” and “perceptual” features. As a reminder, perceptual

features describe how an object looks on the surface and are intrinsic to the object (e.g. the color

red, a round shape, etc.). On the other hand, relations describe how an object connects to another

object (e.g. to get from point A to point B) or the internal relations of an object (e.g. it is only X if Y

connects to Z).

When listeners use relations between primary parameters to categorize musical events, these

categories are relational categories. To categorize two musical events as members of one relational

category, a listener can map relations within a musical event. If listeners define a musical event using

non-relational secondary parameters (e.g. events played loudly categorized together), this would be a

perceptual category (e.g. of loudness) since it is characterized by high intrinsic similarity of dynamics.

One musical event is often a member of both a relational category and a perceptual category. For a

non-musical example, a mansion is simultaneously a member of relational category “home”

(assuming someone lives there) and a member of perceptual category “house” (assuming it fits the

perceptual features of “house”). For a musical example, a musical event with a descending melody

from fa to do (primary parameter is pitch and the relation is the organization of pitches in a Prinner

schema; Gjerdingen 2007a) but also a fast tempo (secondary parameter) is a member of both

relational category “Prinner” and perceptual category “fast excerpts.” This approach is not “all or

nothing”; instead, how a listener categorizes a musical event depends on what is salient in the music
103

(what a listener attends to), what is within a listener’s perceptual limitations, and a listener’s

expertise. Meyer (1998) discusses how the distinction between primary (syntactic) and secondary

(statistical) parameters benefits theory:

Tonal dissonance, which is a matter of syntactic function, needs to be distinguished from


acoustic discord, which is a matter of statistical amount. Thus the interval of a minor seventh
in a dominant seventh chord implies a resolution because of our understanding of tonal
syntax. But the implication of a minor seventh in a chord built of two perfect fourths are
enigmatic because they are not constrained by a shared syntax. Unlike dissonance, which is
functional, discord is a statistical – a matter of amounts that result from factors such as the
proximity of simultaneously sounding pitches and their partials, attacks and dynamic level,
the acoustical properties of the instrument and so on. (Meyer 1998, 10)

The same auditory signal can be interpreted differently, as either “tonal dissonance or “acoustic

discord”; therefore, the interpretation could be either a relational category or a perceptual category.

Next, I list specific parameters that often create relations within a musical event, based on

Meyer’s (1989) primary parameters and Hanninen’s (2012) contextual subtypes. Listeners use

relationships between these parameters to form relational categories. The first “level” (labeled with

roman numerals) represents a relation within a parameter most likely perceived. The second “level”

(labeled with letters) represents a relation connected to the first “level” that needs either more

cognitive effort, attention, or domain-specific expertise to be perceived compared to the first

“level.”

i. Rhythm
ii. Scale Degrees
iii. Contour
a. Interval size and quality
iv. Harmonic stability/instability
a. Harmonic Functions (local to context)

I discuss each parameter in more depth to argue why listeners hear these musical parameters

as relational, although listeners sometimes struggle or need expertise to hear these relational

categories. As a disclaimer, listeners could hear any element of music as relational with enough
104

effort. However, some musical dimensions need less effort than others to be heard as relational

since some dimensions are often relational due to Western cultural listening practices.

Since my research question investigates categorizing musical themes, I discuss these musical

relations in the context of categorizing a musical theme (a thematic category). The term “thematic

category” means a “relational category of a musical theme.”

To graphically represent relations within a musical event in my analyses, I use a version of the

relations within box from figure 3.5 (see figure 3.6).

Figure 3.6: Graphic representation of relations within a musical event

The circle represents the relation itself while the diagonal lines attach arguments to that relation

(whether objects or other relations). The graph moves left-to-right, showing time and order of

objects in musical event. Different rows represent different musical parameters that may give rise to

relations and thus define a thematic category: rhythm, scale degree, contour, and harmonic

stability/instability. A thematic category is a higher-order relation composed of these other relations.

Even though not visually represented, the relations within box (figure 3.6) implies a structure like in

figure 3.7.
105

Figure 3.7: Higher-order relation thematic category

In figure 3.7, the () is where relations are categorized parametrically. To make graphics cleaner, I

dispense with individual relations between adjacent notes. Therefore, the relations within box of figure

3.6 is often visually represented like figure 3.8:

Figure 3.8: Revised graphic representation of relations within a musical event

i. RHYTHM

Rhythm cannot be defined easily because it has at least two meanings: as a parameter and as

a local pattern; but a simple definition might be the pattern of time that elapses between note onsets,
106

since onsets, as opposed to note durations, determine rhythms. Scholars often describe rhythm

using inter-onset interval (“IOI”), or the time “between the attack-points of successive events”

(London 2004, 4). Since rhythms are based on relative time, humans conceptualize rhythmic

relationship as ratios between two adjacent time-spans, also known as serial ratio (Jones 1976).

Figure 3.9: Proportional rhythms

Honing (2013) writes that rhythms are notated proportionally, calling this “proportional or relative

representation,” since “it indicates how the durational interval between the notes relate” (374, italics in

original, see figure 3.9). These proportional rhythmic patterns are not always exact in performance

due to expressive timing and tempo, though these aspects give musical performance interest (Ashley

2002). According to Honing (2013), both rhythmic pattern and expressive timing “are available at

the same time, with the categorization functioning as a reference relative to which timing deviations

are perceived” (375). Listeners categorize rhythms as serial ratios even if the ratios are more complex

(Clarke 1987). Clarke (1987) asked participants to listen to 10 short, regularly timed musical items

(five or six notes). Then, he played three test notes of serial ratios, all between 1:1 and 1:2. The

results indicated that humans often hear the test notes as either 1:1 or 1:2 even if the note timings

actually formed a more complex ratio. He argues that listeners interpreted the more complex ratios

due to “expressive information, or perhaps accidental inaccuracy” (Clarke 1987, 30).

Rhythmic patterns are based on proportional ratios, and listeners use these relations to make

musical analogies. Proportional ratios are a common analogical relation. In fact, Hofstadter and

Sander (2013) consider proportional analogies (e.g. west : east :: left : right) an analogy-making
107

“stereotype” (15). This analogy “stereotype” is relevant for perceiving rhythmic patterns. When

comparing two musical events, listeners could use rhythmic proportionality as a relation to perceive

thematic similarity and categorize musical events as members of the same category. In figure 3.10,

for instance, the two Clementi motives both share a long-short-short-long-long rhythmic relation.

Figure 3.10: Showing rhythm as a relation within a musical event; Clementi, Sonatina Op. 36, no. 1, m. 5 and m.

6 respectively

Some common rhythmic relations include: SHORTER-TO(x, y), LONGER-TO(x, y), DOUBLE-

TO(x, y), HALF-TO(x, y) and so on.

To graphically represent rhythm, I use one of two approaches (depending on rhythmic

complexity). First, a 1) length approach, where a series of durations is represented as a series of ordinal

values. Figure 3.11a is a key for the ordinal duration vector, from which I assign the ordinal values.

Then, I note symbols between the ordinal values—+, - or 0—to represent (respectively) whether

shorter-to-longer, longer-to-shorter, or stays the same. In figure 3.11b, I represent the musical event

in figure 3.10 using a length approach. I write 4 (=quarter note) and 2 (=eighth note): the first note

and the last two notes are longer compared to the second and third. This approach is preferred

when a musical passage is rhythmically complex.

(a)
108

(b)

Figure 3.11: (a) Ordinal duration vector key, (b) Clementi rhythm—length approach
109

The second is a 2) proportional approach. This approach is preferred when notes neatly fit

proportionally against each other. In figure 3.12, for example, the first, fourth and fifth notes are

notated twice as long as the second and third notes. Therefore, I use a proportional approach,

writings 1 (=eighth note) and 2 (=quarter note, or two eighth notes).

Figure 3.12: Clementi rhythm—proportional approach

Both lengths and proportions are inherently relations. A note is “long” only in relation to a

following note that is shorter than it. Therefore, in a piece, a “long” note could be an eighth note or

a whole note, depending on its surrounding notes in context.

ii. SCALE DEGREES

A melody traveling from one scale degree to another creates a relationship between scale

degrees. Huron (2006) claims he himself has no “native mental code for melodic intervals,” but a

“scale-degree code”; instead of hearing a perfect fourth, he hears so-do or mi-la, but rarely re-so (117).

He and collaborator Bret Aarden ran an experiment to see if listeners encode intervals as scale-

degree dyads. The results were mixed—some musicians encoded intervals as scale-degree dyads

while others did not (Huron 2006, 119).


110

The schemata of Meyer and Gjerdingen are based on relations between ordered scale

degrees that often create distinct contours (although I discuss contour separately). Consider the

beginning to Dittersdorf, Symphony in C, K. 1, first movement (figure 3.13). This piece begins with

a musical abstraction that Meyer and Rosner (2000) call the “changing-note schema” and Gjerdingen

(2007a) later renames the “Meyer” schema. Meyer and Rosner (2000) describe this schema as a

melody where:

The main structural tones of the pattern consist of the tonic (1), the seventh or leading tone
of the scale (7), the second degree of the scale (2), and then the tonic again…. A variant of
the changing-note process may occur beginning on the third degree of the scale, producing
the succession 3-2, 4-3. (Meyer and Rosner 2000, 166)

In a Meyer, the soprano begins with a 1-7 dyad and closes with a 4-3, as seen in Dittersdorf’s do-ti-fa-

mi opening. The theme retains its identity even transposed to a new key, demonstrating the

importance of relations between notes and not notes themselves (C-B-F-E).

Figure 3.13: Dittersdorf, Symphony in C, K. 1, first movement, mm. 1-4

The relation can hold even when some pitches differ. For instance, one pitch can substitute

for another, assuming that the substitute pitch: 1) retains a similar melodic contour overall, and 2)

retains the same harmony. Evidence for this can be seen in Meyer and Rosner’s (2000) description

of the changing-note schema and its possible variants. One could argue that Dittersdorf’s Meyer (1-

7, 4-3; see figure 3.13) and the 1-7, 2-1 variant (what Gjerdingen (2007a) calls the “Aprile”; see figure

3.14) both share an abstracted scale degree relation.


111

Figure 3.14: Example of Aprile (with a not as common 5-1 close in the bass instead of 7-1) from Aprile, Solfeggio,

MS fol. 40v, Larghetto, mm. 1-4 (graphic from Gjerdingen 2007a, 123).

A listener could perceive the Aprile example (do-ti-re-do) and Dittersdorf (do-ti-fa-mi) example as

members of the same musical category:

SCALE-DEGREE-CATEGORY[DESCENDS-TO(tonic-note, dominant-note),

ASCENDS-TO(dominant-note, dominant-note), DESCENDS-TO(dominant-note, tonic-

note)]

Figure 3.15 illustrates a graphical representation of this relation.

Figure 3.15: Graphical representation of scale degree relation

In a relation between scale degrees, one scale degree builds on the previous one. This parametrical

relationship also has closer ties to contour relations and harmonic functions than others discussed

here.

To represent scale degrees, I use solfége, as in figure 3.16:


112

Figure 3.16: Scale degrees

In scores, I use numbers to mark scale degrees. In graphs, however, I use solfége since numbers are

sometimes used to represent rhythm. The analyses typically show only structural notes instead of all

notes in passage.

iii. CONTOUR
a. INTERVAL SIZE, DIRECTION AND QUALITY

Instead of using pitch and perfect intervals to categorize melodies, listeners prefer to

categorize melodies based on contour similarity (Halpern 1984, 166). Contour is defined as “the

pattern of rises and falls along an auditory or musical dimension” (Schmuckler 2010, 169). Although

contour can be relevant to other auditory dimensions (including loudness and timbre), I limit my

definition of contour here to changes in pitch information. When comparing musical motifs,

contour is the most important attribute (Dowling 1978), though participants also use size of tonal

motion (Eiting 1984, 90). Dowling (1978) concludes that contour, an “abstraction,” of atonal

melodies can be held in memory independently of interval sizes or pitches (346). These empirical

studies demonstrate that listeners are sensitive to melodic shapes (as in the Clementi sonatina of

1798-1799 in figure 3.17), implying that contour—or shape of a melodic abstraction—may be a

simpler relation to perceive than intervals.


113

Figure 3.17: Clementi, Sonatina in C Major, Op. 36, No. 1, first movement, mm. 1-19

Contour may mean all pitches in a melody, as in the Clementi excerpt (figure 3.17), but also

structural notes of a schema or melodic abstraction. Scholars, more often than not, consider contour

a “surface feature,” since contour often means “all notes in a phrase.” However, I restrict contour to

structural notes. Contour here means shape of structural notes or shape of relational category of the theme.

More often than not, relations become generalizations or abstractions. Listeners can perceive a

contour shape of a theme’s mental representation as a relation. For example, a Meyer or a

“changing-note schema,” a historically-grounded term, gets its name due to a melodic contour

resembling a melodic turn sign (figure 3.18).


114

Figure 3.18: Dittersdorf, Symphony in C, K. 1, first movement, mm. 1-4, and outline of contour relation

A contour relation is often represented spatially, maybe more so than other relation types discussed.

Therefore, it is possible to draw the shape of the relation, as in figure 3.18. A theme with a different

contour relation then creates a different shape. For example, mm. 3-4 of Mozart’s Sonata in C

Major, K. 545 (1788), first movement, form a schema called the Prinner (Gjerdingen 2007a), where

the soprano has scale degrees 6-5-4-3.

Figure 3.19: Mozart, Sonata in C Major, K. 545, first movement, mm. 1-4, and outline of contour relation

Instead of a turn shape, as in Dittersdorf, the shape of Mozart’s contour is simply a descent.

For contour, I use spatial or pictorial representation, where a line illustrates direction and

shape of contour.
115

Figure 3.20: Contour

I have several reasons for representing contour pictorially (as opposed to contour segments; e.g.

Straus 2005). First, others have already represented contour this way. For example,

ethnomusicologist Charles Adams (1976) created a typology for categorizing contour types and he

used pictorial representations for his graphics. Second, listeners often evoke spatial reasoning when

listening to contour. For instance, they often hear contour in terms of “up” and “down.” A drawing

of contour, then, may provide an accurate shape for how a musical event is heard.

I have grouped ascending/descending interval size and quality in this section as well, since,

even though listeners rely more on contour, their reliance on contour depends on an understanding

of interval distance. Listeners may not recognize specific intervals (e.g. major sixth), but they

understand some intervals are larger than others (e.g. major sixth as larger than a major second). For

example, this could be represented as:

LARGE-INTERVAL-TO(C4-A4)

SMALL-INTERVAL-TO(C4-D4)

Contour relations are connected both to intervals (in the descending/ascending interval size and

quality) as well as scale degree relations.


116

iv. HARMONIC STABILITY/INSTABILITY


b. HARMONIC FUNCIONS (LOCAL CONTEXT)

Western listeners (both trained and untrained) can perceive relationships of musical tensions

and relaxations in short and long chord sequences (Bigand, Parncutt, and Lerdahl 1996; Bigand and

Parncutt 1999). In Western music, tension-relaxation and stability-instability relations are partly

determined by “harmonic relations that exist among chords” (Bigand, Parncutt, and Lerdahl 1996,

125) as indicated by seminal music theory texts (Lerdahl and Jackendoff 1983; Meyer 1956; Meyer

1973). Bigand, Parncutt, and Lerdahl (1996) give the following example:

In the key of C major, the chord containing the notes G-B-D-F (dominant seventh chord)
creates tension partly because of the tritone relation between the tones B (leading tone) and
F (the seventh of the chord). This tension needs to be resolved to the tones C and E,
respectively, of the C major triad (tonic chord). (Bigand, Parncutt, and Lerdahl 1996, 125)

They found that tonal hierarchy influenced perceived musical tension. Chords can connect to create

a harmonic stability-instability relation. On a finer note, some listeners—perhaps ones with more

experience—could also be sensitive to relations formed from tonic, predominant and dominant

harmonic functions (Caplin 1998, 23).

In the Dittersdorf excerpt in figure 3.21, a relation of stable-unstable, unstable-stable is

formed due to a I-V-V7-I harmonic progression. In regards to the changing-note schema, Meyer

and Rosner (2000) write: “the pattern is always harmonized by a progression that moves from a

tonic chord (I) to dominant harmony (V) and then from the dominant back to the tonic: that is the

progression is always I-V, V-I. Surrogates for these harmonies are possible: for example, vii for V or

vi for I” (166).
117

Figure 3.21: Dittersdorf, Symphony in C, K. 1, first movement, m. 1-4

Chords relate to each other within a musical event and within a tonal context to form a stable-

unstable relation. This perception of harmonic function changes depending on local harmonic

context, which is more salient than an overarching tonal context for an entire piece. Therefore, I am

not arguing for a large scale idea of harmonic function (an entire section functioning as a “V”).

Instead, perception of harmonic function relies on local context.

To represent harmony, I often use Roman Numerals.

Figure 3.22: Harmony

My representation of harmony never goes beyond short prolongations, when a melody may be an

elaboration of a tonic or dominant chord. In some of my analyses (chapter 5), harmony is the least

discussed relation since it remains consistent.

Relations between Musical Events


118

Relations between musical events describe how one musical event relates to another within some

dimension. For example, there is a change of mode relation when a piece begins with a musical event in

major followed by a musical event in minor.

CHANGE-OF-MODE-TO(major, minor)

Binary oppositions are a common way to discuss relations between musical events, especially for distant or

extra-musical analogies. According to Robert Hatten (1994; 2004), major and minor are binary

oppositions, with minor “marked” in relation to major. The change of mode relation from major to

minor can be mapped onto similar relations in cultural meaning. He writes: “an opposition in

musical structure (minor versus major mode) can correlate [figure 3.23] with an opposition in

meaning (tragic versus nontragic, for the Classical style), which provides a systematic motivation for

association that is stronger than association by mere properties or contiguities” (Hatten 2004, 12).

Figure 3.23: Correlation (graphic from Hatten 1994, 37)

Since minor is marked, it has a narrower range of meaning: it consistently conveys the tragic (Hatten

1994, 36). Major, on the other hand, is unmarked, and so has a broader range of meanings: it

corresponds to the nontragic in general (Hatten 1994, 36). Hatten (1994) uses the graphic in figure

3.24 to represent this correlation.


119

Figure 3.24: Major vs. minor onto tragic vs. nontragic (graphic from Hatten 1994, 38)

As illustrated in figure 3.25, the minor vs. major are relations between musical events that create an

opposition Hatten (1994) “maps”—to use analogy terminology—onto tragic vs. nontragic.

Figure 3.25: Mapping of minor vs. major onto tragic vs. nontragic

Another way to create musical oppositions—and so relations between musical events—for non-

musical analogy is through musical topics (e.g. Ratner 1980; Agawu 1991; Monelle 2000; Monelle

2006). Musical “topics” are “a thesaurus of characteristic figures” considered to be “subjects for

musical discourse” (Ratner 1980, 9). Then, a musical topic is a conventionalized musical pattern

understood by a community to have certain referential or extra-musical association. The way a

musical event’s topic relates to another musical event’s topic not only creates relations between musical

events, but also musical oppositions that listeners map onto cultural opposition. Allanbrook’s (1992)

interpretation of Mozart’s beginning of K. 332 (1783) implicitly describes topics in opposition.


120

Figure 3.26: Mozart, K. 332, first movement, mm. 1-26

As seen in figure 3.26, Allanbrook analyzes the first four measures as a “simple singing style” (mm.

1-4), and the next four measures a “parody of learned counterpoint” (mm. 5-8), followed by four

measures of “galant minuet style” (mm. 9-12), which closes the period (Allanbrook 1992, 132). She

writes that the several horn calls in mm. 13-22 “provides an evocative counterstatement to the

opening topics—the out-of-doors answers the salon” (Allanbrook 1992, 133). On the horn call’s

associations, Monelle (2006) writes: “We may observe intuitively that hunting music evoked the

nobility, the outdoors, the forest, adventure, and action” (35). After following the “indoors” singing

style, “outdoors” is a more salient association of the horn calls than “nobility,” “adventure,” and

“action.” Here, singing style vs. the horn fifths maps onto cultural opposition of indoors vs.

outdoors—or even decorum vs. rustic (figure 3.27).


121

Figure 3.27: Mapping of singing style vs. horn fifths onto indoors vs. outdoors

The topics together create a relation between two musical events that can be mapped onto a cultural

opposition. This creates an analogy between a musical analog and a cultural, non-musical analog.

So far, all examples of relations between musical events have been analogies between music and

another domain, in that they have mapped a culturally understood relation signifying meaning (e.g.

nontragic-tragic) onto a musical relation (e.g. major-minor). Yet, relations between musical events make

analogies between music and music as well, such as relations of musical form (e.g. Kielian-Gilbert

1990). Consider Caplin’s (1998) theme type of “sentence,” an eight-measure structure that:

Begins with a two-measure basic idea, which brings in the fundamental melodic material of
the theme….The basic idea is repeated in measures 3-4….As a result of repetition, the basic
idea has been unequivocally “presented” to the listener, and so we can speak of this music
fulfilling presentation function….The strongly ongoing quality created by a presentation
generates demand for a continuation phrase, one that will directly follow, and draw
consequences from, the presentation,” before closing with a cadence. (Caplin 1998, 9–10,
italics in original).

Listeners hear some musical themes as “sentences,” such as the ones in figure 3.28a and 3.28b,

because of relations between multiple musical events.


122

(a)

(b)

Figure 3.28: (a) Beethoven, Piano Sonata in F Minor, Op. 2, No. 1, first movement, mm. 1-8, (b) Beethoven,

String Quartet in F, Op. 135, third movement, mm. 3-10 (graphics from Caplin 1998, 10)

A listener make an analogy between one musical event and another to hear the Beethoven Piano

Sonata, Op. 2, No. 1, first movement (figure 3.28a) and the Beethoven String Quartet, Op. 135,

third movement (figure 3.28b) as similar, since both themes have relations of a “sentence.” A

listener can align these musical themes and map over the “sentence” relations.

Two Forms of Analogical Comparison: Unidirectional Listening and Bidirectional Listening

There are two ways listeners compare musical analogs, each forming a distinct way of

listening to music: unidirectional listening and bidirectional listening. An important component of

Structure-Mapping Theory is comparison. Unidirectional listening is projective analogy, which is


123

comparing a novel analog to a familiar or well-understood one, transferred to music.11 In

unidirectional listening, listeners understand the target in terms of the source; in that they project

inferences from the source onto the target. A listener uses a familiar musical passage heard earlier

(either in the piece or their musical experience) to shape their understanding of incoming musical

passages. For example, listeners use unidirectional listening when using familiar, conventional

patterns to inform perception of a musical event. On the other hand, bidirectional listening is mutual

alignment analogy, which is comparing two novel analogs to each other, transferred to music. In

bidirectional listening, listeners understand both analogs in terms of each other: the analogs inform

each other; both are sources and targets. A listener uses bidirectional listening when recognizing

idiosyncratic or emergent patterns, such as a theme and its variations within a piece. For example,

listeners use a musical pattern just heard to re-shape or re-understand a musical passage heard in the

past. Musical styles encourage different listening comparisons. A Romantic piece balking at

convention (e.g. Meyer 1989) could encourage bidirectional listening (actively comparing themes to

each other), while a galant style piece filled with conventional patterns could encourage

unidirectional listening (actively using conventional patterns to inform musical events).

Four Forms of Analogical Distance in Music: Intra-opus, Inter-opus, Intertextual and Extra-
musical

There are four forms of analogical distance in musical analogy: intra-opus analogy (local

analogy), inter-opus analogy (local analogy), intertextual analogy (local analogy), and extra-musical

analogy (distant analogy).

11Gentner uses the term “projective analogy,” which has a different association then in the music literature. By
“projective” here, I do not mean projecting into the future. Instead, I mean projecting from a familiar source onto a
novel target.
124

Figure 3.29: Visual representation of analogical distance in music

1) Intra-opus analogy
Analogies made within one musical piece.

2) Inter-opus analogy
Analogies made between pieces. Within this level, one could distinguish between composer’s
style, historical style, and transhistorical analogy. For these analogy types, a listener may have
more conscious awareness.
a. Composer’s Style
Analogies made between music all by the same composer.
b. Historical Style
Analogies made between music all in the same style.
c. Transhistorical
Analogies made between any music regardless of composer or time period.

3) Intertextual analogy
Analogies between specific compositions; a specific text is modeled by another text.
Within this level, there could also be composer’s style, historical style, and transhistorical analogy.

4) Extra-musical analogy
Analogies made between music and another domain (for example, analogies between music
and language, analogies between music and narrative, and so on).

In theory, a listener could use the same musical event for three of the four analogies: intra-opus,

inter-opus or intertextual, and extra-musical. With enough intra-opus or intertextual analogies, a

listener can learn or create a new conventional musical pattern. That same listener can then make

inter-opus analogies with this new conventional musical pattern.

Intra-opus Analogy
125

Theoretically, listeners make different assumptions when mapping patterns in intra-opus

analogy than inter-opus, intertextual and extra-musical analogies. For instance, listeners assume these

patterns will not be encountered outside the work. Recognizing a pattern from one composition in

another would be a different instance of analogy-making. In addition, patterns of intra-opus

analogies gain significance through repetition. The more intra-opus analogies are made, the more

information a listener learns about this pattern or category. Also, patterns of intra-opus analogy

often overlap with patterns of inter-opus analogy. For example, a melody is rarely so original that it

has no basis in convention. Yet, an analogy made between a musical passage and a convention

would be inter-opus rather than intra-opus. As intra-opus analogy involves making sense of

emergent patterns over time, listeners use bidirectional listening more for intra-opus analogy than

others.

Larson (2012) uses intra-opus analogy to argue why the “Adagio” from Mozart’s Piano

Concerto in A Major, K. 488, second movement (figure 3.30) violates expectation. The opening

forms an “antecedent-consequent” period.


126

Figure 3.30: Mozart, “Adagio” from Piano Concerto in A major, K. 488, second movement, mm. 1-12

Mm. 1-4 make a “motto” with a “Siciliano rhythm,” ending with a half cadence to create a musical

“question” (Larson 2012, 36). The following consequent—mm. 5-12—begins with the same

“motto,” but ends with an authentic cadence to create a musical “answer” (Larson 2012, 36). He

insists that:

When we describe the music in this way, we make an analogy between measures 1-4 and
measures 5-12…. This analogy maps the motto in measure 1 onto the motto in measure 5,
and the cadence of the question on to the cadence of the answer. Finding these similarities
between measures 1-4 and measures 5-12 makes this analogy possible. And finding these
similarities provides a context for noticing meaningful differences (such as the fact that the
motto begins in measure 1 on C# and in measure 5 on D, that G# in measure 3 precedes
the first cadence and that G in measure 9 precedes the second cadence, and that the first
cadence is incomplete and the second one is complete)…. To hear measures 1-4 and 5-12 as
analogous in this sense relies on hearing measures 1 and 5 as “the same” in some sense
(Larson 2012, 37).
127

Larson (2012) writes that listeners familiar with the style will recognize a typical antecedent, and

expect a typical consequent to follow that returns at the same pitch level (39). Instead of returning to

the same pitch level (C#), it begins the expected consequent on D. Because a listener made an

analogy between m. 1 and m. 5, he or she expects a certain degree of similarity. The consequent is

similar, but it is surprising since it begins on the unexpected pitch level D instead of the expected

C#.

The most common intra-opus analogies identify themes, motives, and variations. A theorist

could use intra-opus analogy to analyze theme and variations form since listeners map a main theme

onto in order to hear them as one category. In addition to being from cognitive perspective, this way

of analyzing theme and variations focuses on a theme’s relations instead of addition or subtraction

of features.

Inter-opus Analogy

In inter-opus analogy, listeners relate a current musical event to one in another piece. When

making an inter-opus analogy, listeners often connect relations of stylistic abstractions, such as a

conventional musical pattern that transcends multiple pieces. For example, a listener makes an

analogy between a current musical event and a Prinner schema, which is a stylistic abstraction where

a soprano descends 6-5-4-3 and a bassline descends 4-3-2-1 (Gjerdingen 2007a, 45–60; see figure

3.31).
128

Figure 3.31: Prinner schema (graphic from Gjerdingen 2007a, 455)

When a listener hears mm. 1-3 of Wodiczka’s Op. 1, No. 3, first movement (1739) (figure 3.32a) and

mm. 10-14 of L’Abbé’s Op. 8, No. 1, first movement (1763) (figure 3.32b), he or she can make an

inter-opus analogy between these closing ripostes, connecting both musical events to each other and

to the conventional Prinner schema.

(a)

(b)
129

Figure 3.32: (a) Wodiczka, Op. 1, No. 3, first movement, adagio, mm. 1-3 (graphic from Gjerdingen 2007a,

46), (b) L’Abbé, Op. 8, No. 1, first movement, mm. 10-14 (graphic from Gjerdingen 2007a, 55)

Since the pattern analogized here (the Prinner) is grounded in the galant style, this is inter-opus

analogy limited to a historical style. Inter-opus analogy can also be limited to a composer’s style, or

analogies made only between music by the same composer. As a result of composer’s style analogy-

making, for instance, a composer-specific schema could emerge. Transhistorical inter-opus analogy, on

the other hand, is not limited to a historical time or place. In this analogy type, a modern listener

could use analogy to project a modern pattern onto an older piece of music.

Intertextual Analogy

Intertextual analogy occurs when listeners make an analogy between specific texts; in short, a

specific text is modeled or borrowed by another text (to which a listener may judge whether or not

he or she thinks the borrowing is intentional). Similar to inter-opus analogy, intertextual analogy

could also be limited to composer’s style (specific pieces by one composer), historical style (specific pieces

within one style) or transhistorical (specific pieces across time and place). Intertextual analogy can

highlight salient differences between two themes. For example, consider the first violin’s theme in

Mozart’s “Misericordias Domini” (1775; figure 3.33a) and Beethoven’s Ode to Joy theme in his

ninth symphony (1824; figure 3.33b):


130

(a)

(b)
Figure 3.33: (a) Mozart, “Miserocordias Domini,” mm. 23-26, (b) Beethoven, Symphony No. 9, fourth

movement, mm. 92-99

Both themes share a similar melodic line of mi-fa-sol-sol-fa-mi-re-do-re-mi—even with the same rhythms.

However, once a listener reaches this point in the Mozart excerpt—especially if he or she uses “Ode

to Joy” as a frame of reference—the mi-re-do-ti eighth notes sound different in a salient and striking

way, since a dotted note is expected instead. If a listener uses Beethoven’s “Ode to Joy” as a model,

then he or she is surprised by Mozart’s melody being different. Since Beethoven’s piece follows

roughly fifty years after Mozart’s piece, this would be transhistorical intertextual analogy. On the other

hand, if a listener’s first exposure is to Mozart (a likely scenario considering the timeline), he or she

may instead be surprised in listening to Beethoven.

Extra-musical Analogy

Theoretically, listeners tend to use extra-musical analogies (analogies between music and

another domain) to make sense of “unexpected findings” in music. Dunbar’s (2001) subjects used

distant analogies to make sense of “unexpected findings”; therefore, listeners may take a similar

approach. In Western culture, a common distant analogy humans make to understand music is

comparing music and language (e.g. London 1996; Swain 1997). This comparison is not a new one

in music scholarship. According to Dahlhaus (1991), “in the discussion of musical aesthetics the idea

of likening musical structures to verbal ones—that is, comparing a period to a sentence, and a

motive to a word—is a commonplace that goes back to the Middle Ages” (93). Even though it
131

existed in the Middle Ages, this analogy between music and language dominated in the eighteenth-

century (Bonds 1991; Mirka and Agawu 2008). In her introduction to Communication in Eighteenth-

Century Music, Danuta Mirka (2008) uses the word “metaphor,” but she means “analogy”:

At that time [the eighteenth-century] theoretical and aesthetic discourses about music were
based upon the metaphor of music as language. Within this metaphor, a composer or
performer was compared to an orator, and a musical piece to an oration subdivided into
parts, periods and sentences. Just as the art of rhetoric has its raison d’etre in persuading the
listener, so the art of composition consisted in arousing his sentiments. The musical
repertory labelled by later generations as the ‘Classical style’ was thus an expression of the
aesthetic stance which conceived of music as communication between composer and
listener. (1)

Using this analogy, listeners assume information about musical communication based on music’s

similarity to language. If a listener does not “get” or “comprehend” an element of a piece, he or she

could default to searching for a potential source elsewhere. One potential source could be a listener

defaulting to a linguistic framework, especially since this is a culturally-ingrained analogy (in the same

way that someone in speech may default to an “up-down” metaphor). If what a listener does not

“get” or “understand” in the music seems to mimic language, and reminds them of language, then it

seems likely that he or she would retrieve this source domain and use it as an analogy.

Step 3: Evaluation

In the evaluation step for musical analogy, a listener performs three simultaneous actions: 1)

judgment of inferences, 2) consideration of musical context, and 3) goal relevance. To put it

succinctly, a listener evaluates inferences in context while considering his or her own goals and goals

of others. These actions helps a listener evaluate the analogy as a whole and anything new he or she

learned from it.

Inferences
132

Inferences are first projected during mapping, but then judged during evaluation. Listeners

make various inferences while listening, including affect (or emotion12), formal function (Caplin

1998), learned associations, expectations (or implications) among others. Meyer (1973) describes

how listeners make inferences about function of a musical event:

Not surprisingly, melodies which perform the same general function in a particular kind of
composition often have common characteristics, a kind of family resemblance. Some
melodies seem typical of the beginning of sonata-form movements; others are members of
the class of closing themes. Some melodies are characteristic fugue subjects; others seem like
themes upon which a set of variations might be built. (206)

In this quote, Meyer (1973) describes how melodies that perform the same function often share

common characteristics. He implies that listeners will infer that a musical event with these

characteristics will have the conventional function. When certain listeners make an inter-opus

analogy to map a horn fifths pattern, they infer this pattern has a cadential formal function since the

pitch and harmonic parameters best match a closing phrasal pattern. They then assume that this

passage does (or should) occur at the end of a phrase. The inference most likely takes place as soon

as a listener identifies the pattern.

Context

Listeners evaluate their mappings and inferences in context. An opera buffa coda has a

different meaning as the ending of Beethoven’s String Quartet Op. 95 then it does as the end as an

opera buffa finale. Margulis (2013) describes the importance of context in repetition: “Even the most

literal forms of repetition, then, are differentiated by the associations of the immediately surrounding

12According to Bigand and Poulin-Charronnat (2006), “Emotional responses are consistent between and within
participants, and that they seem to be immediately triggered by short musical excerpts…. It only shows the importance
of the irrepressible emotional responses to music, which nicely mirrors the automaticity and rapidity of cognitive
processes in music perception” (119). This “irrepressible emotional response” could be an example of an inference a
person may make while listening.
133

context—the words or notes that precede or follow it—as well as the relevant compositional,

authorial, historical and intertextual situation of the utterance” (27–28); ultimately, “context shapes

perceptual, cognitive, and emotional orientation” (30). Different types of context occur when

humans listen to music. For example, context could be the other musical events surrounding the one

in questions or it could be the venue of the performance.

An analogy framework benefits scholars studying musical context since analogy often means

understanding a familiar pattern in different ways. Evaluation, then, prompts listeners to consider

these patterns in context, which may alter inferences or cause listeners to reconcile incongruous

elements. Therefore, analogy helps identify recontextualization, which Margulis (2013) defines as:

“The process whereby a given element is positioned or understood within a different context”

(Margulis 2013, 32). The analogy framework balances “patterns” and “context”: a framework for

recognizing patterns exist, but also acknowledging music perception as a holistic process where

patterns are not heard in isolation. For instance, consider figure 3.34, which is Beethoven’s Piano

Sonata No. 26, Les Adieux.

Figure 3.34: Beethoven, Piano Sonata No. 26 in Eb major, Op. 81a, first movement, “Lebewohl,” mm. 1-5

Meyer (1973) writes: “Characteristic melodic gestures usually occur in appropriate and familiar

context…. In some cases, however, there is a discrepancy between the normal function of a gesture

and its actual use in a composition” (Meyer 1973, 208). Les Adieux has horn fifths as a beginning;
134

therefore, it is an ending as a beginning. A listener maps this opening gesture onto a common horn

fifths pattern. When writing about the horn fifths beginning Beethoven’s “Les Adieux,” Meyer

(1973) observes that:

Use of horn fifths in the first measures of Beethoven’s “Les Adieux” Sonata is unusual in
almost every way. Instead of coming at the end of a fast movement, they are the beginning
of a slow introduction; instead of being accompanimental, they are the main substance; and
instead of reaching emphatic closure on the tonic, they end in a deceptive cadence which is
mobile and on-going. The deviant use of the traditional pattern not only emphasizes the
importance of the motto, but contributes considerably to its peculiar poignancy. (244)

Some listeners, including Meyer (1973), infer that this pattern has a cadential formal function. Yet,

this pattern begins Beethoven’s piano sonata; thus, a listener has to reconcile the ending inference

with the pattern’s actual placement.

Goals

Evaluation incorporates listener goals, or the question of “why are they listening?” Yet,

evaluation also incorporates composer goals. I consider composer goals not through authorial

“intent,” but considering what the listener thinks the composer intends (a phenomenon I call perceived

intent). Meyer (1989) argues that music has intention behind it since a composer makes certain

choices within a constrained musical style (138). He differentiates this type of intention from an

“idiosyncratic” or “personal” one:

The distinction between merely miming and replicating indicates, then, that far from being
irrelevant, comprehending the intentions of composers (collective as well as individual) is
crucial for understanding the choices that results in the compositions on which a history of
music is partly based. But – and this cannot be sufficiently emphasized – it is not some kind of
idiosyncratic, personal intention that is crucial for such a history, but the sort that is implicit in the stylistic
constraints that define the goals of the “game of art” itself. In short, the historically significant
intentions of a football coach, a grand master of chess, or a composer are those that result
from choices made among alternative possibilities permitted by the constraints of both the
style of the activity and the cultural context. (Meyer 1989, 138, italics in original)
135

A listener assumes musical gestures are placed together for a reason; therefore, he or she assumes a

composer has intentions. These perceived intentions play into how a listener evaluates inferences in

context.

Retrieval, Mapping, and Evaluation: Pattern Mapping Taxonomy

Listeners retrieve familiar patterns from memory, map those patterns onto incoming music

they hear, and then evaluate the meaning they infer from this mapping. When we discuss intra-opus,

inter-music and intertextual mapping, we often mean how listeners map musical patterns. Pattern is

defined here as a combination of surface-level notes/rhythms/figurations/etc. based on some

convention or commonly understood association; therefore, it is a musical event that depends on

relations within a musical event. The recognition of patterns—or patterns as musical events—is

essential to both music theory and the listening experience in general (Meyer 1967). As a way to

connect pattern mapping and listener background (retrieval step), I create a pattern mapping

taxonomy to analyze what patterns listeners may map while listening.

This pattern mapping taxonomy can be useful in music analysis for a number of reasons.

First, this taxonomy adds to what theorists already do by analyzing patterns according to whether it

is appropriate or not for the historical context of the piece. Second, it can be used in conjunction

with many analytical methodologies, but not tied to a particular one. It could be used to analyze

schemata, topics, themes, and a myriad of other patterns that theorists think a listener perceives in

music. Not only is it not tied to a particular methodology, but it actively connects different analytical

methods, giving theorists a way to easily discuss multiple pattern types. Third, it gives theorists a way

to discuss both novel and conventional patterns in music, and how the listening experiences asks

humans to constantly reconcile convention and novelty. Fourth, it considers how context shapes
136

pattern perception; patterns do not exist in isolation, instead they exist in context. Fifth, it

distinguishes between different types of patterns that may be activated during listening, showing that

patterns are not static moments on a score, but, instead, emergent and flexible. Finally—and,

perhaps, most importantly—the taxonomy gives theorists a way to discuss patterns in relation to a

listener or group of listeners. All patterns in this taxonomy rely on someone listening, and what that

listener thinks or assumes he or she perceives in the music.

To account for listener experience or familiarity, the taxonomy depends on a notion of

“belonging”: whether or not the pattern was “around” during the piece’s time period—was it, in the

words of Meyer (1989), a viable choice?—or whether or not a pattern’s interpretation has changed

over time. The pattern mapping taxonomy (figure 3.35) includes four main types (and some sub-

types): 1) listener maps patterns that belong. Under this type, there are two sub-types: i) listener

maps patterns that belong and all goes as expected, and ii) listener maps patterns that belong, but it

was deceptive or evaded. 2) Listener maps patterns that do not belong. 3) Listener maps a pattern

that belongs, but its parameters or context are manipulated so that a listener evaluates it different. In

this type, there are three subtypes: i) pattern in “wrong place” or context, ii) pattern incongruously

juxtaposed with another, and iii) incongruous addition or manipulation to pattern itself.
137

Pattern Mapping
Taxonomy

Patterns that Don't Patterns that


Patterns that Belong Belong Belong, but
Manipulated

Expected Wrong Place or


Context

Deceptive or Evaded Incongruous


Juxtaposition

Incongruous
Addition

Figure 3.35: Visual Representation of the Pattern Mapping Taxonomy

Next, I give examples for most of these types of pattern mapping.

Pattern that Belongs

First, I discuss when listeners map over patterns that belong, including expectations for how

it will be realized. I create two subtypes: 1) listeners map patterns that belong and all proceeds as

expected, and 2) listeners map patterns that belong, but it is deceptive or evaded. When Meyer

(1973, 1989) discusses perception of schemata, his discussions often fall within this type of patterns

that belong.
138

Figure 3.36: Prinner schema (Gjerdingen 2007a, 455)

When a listener recognizes a Prinner, he or she hears scale degrees 4-3, then has an expectation that

steps 2-1 will follow. Listeners rarely expect a deceptive cadence; instead, they expect an authentic

cadence, but it goes “terribly wrong.” Yet, listeners hear these patterns as “evaded” or “deceptive”

because they map over a pattern with certain expectations for how a composer will realize it.

Pattern that Does Not Belong

A listener can map over a pattern that does not belong. This pattern is anachronistic or not a

possible stylistic choice for the composer. Consider the following literary example; I, as a twenty-

first century person, read an eighteenth-century novel and think the main character acts like a

computer. Since computers did not exist in the eighteenth-century (at least the way think of them),

this pattern mapping is anachronistic. For a musical example, listeners—especially modern

listeners—may connect film music patterns to classical music, even though film music was

composed later chronologically. In these cases, the film music is the source (the more familiar and

better understood concept) and the classical music is the target (the less familiar and less understood

concept).
139

As an example, let me give a personal anecdote regarding Stravinsky’s op. 7, etude no. 1

(figure 3.37).

Figure 3.37: Stravinsky, op. 7, etude no. 1, mm. 1-6

On my first hearing, I heard a thematic similarity between Stravinsky’s melody and Danny Elfman’s

Batman (1989) theme (figure 3.38). This would be transhistorical intertextual analogy since I made an

analogy between two specific pieces across time periods.

Figure 3.38: Beginning of “Batman’s Theme,” Elfman, Batman (1989)

Both themes have a salient do-re-me-le-sol, emphasizing le and its stepwise resolution to sol. Hearing

Stravinsky through Batman implicitly and unconsciously colored my perception and judgment of the

piece, even if I intellectually knew Stravinsky came before Batman. In addition, it impacted my
140

meaning or interpretation of this etude. I was more likely to use adjectives to describe the Stravinsky

etude that I would also use to describe Batman, since the Stravinsky was my target and Batman my

source in the analogy. Therefore, listeners can map over patterns that do not belong. In this case, the

pattern does not belong because Stravinsky came before the Batman theme, so I “cannot” use the

Batman theme to understand Stravinsky.

Music theory, however, does not often engage in this historically “backwards” listening,

although there are exceptions. In his book on intertextuality, Klein (2005) mentions transhistorical

intertextuality, or opening “the text to all time” (12). Some scholars hint that this “backwards”

listening impacts music theory analysis, especially claims made from analysis. For example, Clarke

(2005), discussing topic theory, argues that topic theorists often do not make a distinction between

the “idealized”/“reconstructed” eighteenth-century listener and a twenty-first century reader (or

listener) (160).Instead, a topic theorist should identify their listener, instead of assuming these two

audiences have same or similar knowledge. Music theorists could use this pattern mapping

taxonomy to analyze “backwards” listening in music: how modern listeners’ present day patterns

influence how they hear older music. This connects to the historical listener literature, which I cover

at a later point.

Pattern that Belongs, but Manipulated

There are also patterns that belong, but the pattern itself is manipulated or changed and so

prompts listeners to assume an adapted meaning or perspective. In figure 3.39, I present a visual

representation of each subtype: subtype 1, patterns in a wrong “place” or context; subtype 2,

patterns incongruously juxtaposed with another, and; subtype 3, incongruous addition or

manipulation to pattern itself. Subtype 1 and 3 are simultaneous mismatches (their mismatch occurs
141

within the same temporal space), while subtype 2 is a sequential mismatch (mismatch occurs

between musical events).

Figure 3.39: Visual representation of types of pattern manipulation

For the first sub-type—musical patterns in a wrong “place” or context—consider—once

again—Beethoven’s horn fifths that begin Les Adieux (figure 3.40).

Figure 3.40: Beethoven, Piano Sonata No. 26 in Eb major, Op. 81a, first movement, “Lebewohl,” mm. 1-5

Caplin (2005) recognizes that the horn fifths as a beginning creates a sense of play. He writes:

Here the topic of horn fifths occurs at the very beginning of the work, yet its particular
characteristics are precisely those more naturally associated with closure: the melody
descends and the harmonies are those used in a deceptive cadence…. It is easier to say ‘this
music is in the wrong place.’ And thus we are prompted to explore the aesthetic effect of
this disturbance and even to consider whether the Lebewohl idea will eventually find its more
appropriate formal position as a cadence. (Caplin 2005, 122)
142

Caplin’s (2005) rhetoric is one of attention. Because it is music in the “wrong place,” listeners explore

the aesthetics behind it. The play with this pattern causes listeners to ask: “why did the composer

make that choice?”

According to Monelle (2006), associations of hunting music, such as the horn fifths, include

a wide range of related concepts such as “nobility, the outdoors, the forest, adventure, and action”

(35). With these descriptors, any nostalgic associations may be surprising. Yet, Rosen (2002)

describes the Lebewohl horn fifths as the movement’s “kernel” and “a symbol in poetry well

established by 1810 of distance, isolation, and memory” (202). In a similar vein, Hatten (2004)

believes these horn fifths “suggest the distance of a landscape or forest as well as its pastoral

character” (56). Beethoven’s use of horn fifths Lebewohl would be marked, since it appears in the

“wrong place”. A part of this nostalgic expression most likely originates from the Lebewohl’s

translation as “farewell.” Yet, the ending horn fifths placed as a beginning augments this “farewell”

expression. A listener uses the horn fifths as a source for a musical analogy; yet, since the inferences

from the source do not fit the context, a listener has to use discourse such as “distance” and

“isolation” in order to reconcile the manipulation.

Composers manipulate a pattern not just through putting it in its “wrong place,” but also by

putting it adjacent to other incongruous features or events, a method Everett (2009) calls

“incongruous juxtaposition” (32), and the second sub-type.


143

Figure 3.41: Quotation of Wagner in Debussy’s “Golliwog’s Cakewalk” (graphic from Everett 2009, 32)

Consider Debussy’s quotation of Wagner’s “desire” leitmotif from Tristan und Isolde in his Golliwog’s

Cakewalk (figure 3.41). Everett (2009) describes this quotation as distorted:

First, he [Debussy] extracts the ascending minor sixth motive and exaggerates the sentiment
(“avec une grande émotion”), and then he juxtaposes this quotation with the grace note
figuration that “mocks” the serious affect of Wagner’s music. By embedding the operatic
leitmotif within the genre of ragtime, Debussy blurs the presumed boundary between
highbrow and lowbrow music. Thus while the affect of “desire” is associated with the
borrowed motif, the changes in musical context brought on by its juxtaposition with the
“mocking” motif, exaggerated expressive indication, and formal content of ragtime negate
the sign-interpretant of “desire” by trivializing it. (32)

Sheinberg (2000) argues that Debussy creates “an aesthetic distance, a double outlook which is,

simultaneously, satirizing and self-satirizing” (144). This, then, is mapping patterns that belong, but

manipulated or distorted to prompt listeners to interpret it differently otherwise. Unlike the previous

example, the manipulation comes not from a pattern in a wrong place (pattern related to context);

instead, it is a pattern placed next to an incongruous other. The listener maps over the correct

leitmotif but the leitmotif quotation on its own is not distorted; it is distorted because of what follows

after the quotation.

A third, and last, subtype of patterns that belong but are manipulated is distortion through

“adding” incongruous musical parameters to the temporal space of the pattern, what I consider

manipulating the pattern itself. Consider the last variation of the idée fixe in the fifth movement of

Berlioz’s Symphonie Fantastique as seen in figure 3.42.

Figure 3.42: Beginning of Idée Fixe in Berlioz, Symphonie Fantastique, fifth movement
144

In this example, the composer alters the pattern through addition. There is a base abstraction of the

theme and then Berlioz “adds” features to this abstraction incongruous to what the pattern

represents. The idée fixe represents the artist’s beloved (Harriet Smithson); yet, features added

prompt listeners to interpret it as grotesque. These features include a shrill Eb clarinet, a metric

change to a 6/8 (which creates a lilting, dancing feel), chromatic grace notes (especially at the head

of the theme), and trills (particularly of the 1 and 4 of the 6/8). Taruskin (2005) calls this fifth

movement variant a “character-transformation” of the idée fixe to “an ignoble dance tune, trivial and

grotesque.”Ritchey (2010) also writes that this variation sounds like a “lilting, mocking caricature”

(181). Listeners map the pattern of the idée fixe and they recognize that these added features do not

match the representation of the pattern. They then have to use language such as “grotesque” or

“caricature” in order to make sense of it.

Listeners may hear similarities between features used in the “ridiculing gesture” of Debussy’s

Golliwog’s Cakewalk and the grotesque caricature of Berlioz’s idée fixe. Both use chromatic grace notes

at the beginning of the gestures and have motives grouped in threes. Therefore, some of the features

that make Debussy’s gesture “ridiculing” may be the same that make the idée fixe a grotesque

caricature.

I finish this type of patterns that belong, but manipulated by mentioning a few pertinent

elements. While the previous type of patterns that do not belong depended on listeners’ awareness

of historical placement, this type depends on listeners’ familiarity with patterns and what they

represent. For the Beethoven, familiarity comes from exposure to a particular pattern: the horn

fifths. For Debussy, familiarity comes from knowledge of a famous musical quotation and what it

represented. Finally, for Berlioz, familiarity comes not just from knowledge of the program, but also

exposure to the idée fixe variations throughout the composition (this is the only example where
145

listeners do not need to be familiar with an outside piece or style). Pattern manipulations allude to

composer choice, a la Leonard Meyer in his Style and Music (Meyer 1989), and it prompts some listeners,

after making the analogy, to ask themselves, “What is the motivation for manipulating the pattern in

this way?”

Benefits of Analogy Framework

Using the analogy framework can benefit both music theory and cognition as it serves two

purposes: 1) models a person’s cognitive process when using analogy with music, predicting

inferences a listener could make from different analogies, and 2) guides music theorists when

analyzing music using cognitively-informed analogy. The analogy framework builds upon and

furthers current scholarship by providing a systematic way to study musical relationships as well as

giving space to who the listener is and why, from a cognitive perspective, they find certain

interpretations of music or musical communications intuitive.

That experiencing music depends on perceiving musical relationships is recognized by both

music cognition and music theory. One benefit of the analogy framework is it focuses on musical

relationships instead of musical objects in isolation. In Music, the Arts and Ideas (1967), Leonard

Meyer writes: “Understanding music is not merely a matter of perceiving separate sounds. It

involves relating sounds to one another in such a way that they form patterns (musical events)” (46).

Not only are relationships important to pattern perception, but also to how listeners infer meaning.

According to Steve Larson (2012), “meaning is something that our minds create when they group

things into patterned relations” (33). Therefore, incorporating analogy in music analysis helps

theorists connect meaning (and inferences) and musical structure. In addition, analogy can help

theorists analyze not just how listeners perceive musical relationships and relate them to each other

(e.g. one musical theme to another musical theme), but also how listeners relate musical relationships
146

to relationships in other domains (e.g. analogy between music and language, music and narrative,

etc.). I build off of these previous theorists and psychologists to analyze how analogy plays a part in

musical thinking and reasoning; how, from a cognitive and systematic perspective, listeners, based

on their experience and familiarity, perceive relationships in music and infer knowledge from this

perception.

Another benefit of the analogy framework in music analysis is it gives scholars a means to

explore how cultural context of the listener influences meaning and how different groups of listeners

could react differently when listening to a piece of music (especially in the retrieval step). Analysts

may accept that the cultural context of a piece is important to ground any interpretation. In fact,

many methodologies provide brilliant examples of ways to understand the norms of a time period.

An analyst, especially one interested in musical meaning or communication, however, may look

around in a quizzical manner and cry, “What about the listener? Where does his or her cultural

context play a part in analysis?” Temperley (1999) addresses the type of listener usually assumed in

“descriptive” music analysis: “[it] usually aims to account for the perceptions of a fairly wide

population of listeners, rather than just those with extensive formal training, although it will

normally confine itself to listeners who have had some exposure to the kind of music being studied”

(68). Descriptive analyses, including those by Lerdahl, Jackendoff, and Meyer, limit their discussions

of listeners to an “idealization”—the “experienced listener.”

We will now elaborate the notion of “the musical intuitions of the experienced listener.” By
this we mean not just his conscious grasp of musical structure; an enculturated listener need
never have studied music. Rather we are referring to the largely unconscious knowledge (the
“musical intuition”) that the listener brings to his hearing—a knowledge that enables him to
organize and make coherent the surface patterns of pitch, attack, duration, intensity, timbre,
and so forth….A listener without sufficient exposure to an idiom will not be able to organize
in any rich way the sounds he perceives. However, once he becomes familiar with the idiom,
the kind of organization that he attributes to a given piece will not be arbitrary but will be
highly constrained in specific ways….The “experienced listener” is meant as an idealization.
Rarely do two people hear a given piece in precisely the same way or with the same degree of
147

richness. Nonetheless, there is normally considerable agreement on what are the most
natural ways to hear a piece. (Lerdahl and Jackendoff 1983, 3)

Instead of just an “idealized listener,” what would it look like to analyze how different listeners with

different experiences perceive musical relationships and patterns, possibly even comparing these

different listeners? The analogy framework could help answer this question.

To be fair, current sub-areas of music theory identify different listeners and “modes of

listening.” In particular, several theorists—especially in the past decade—have attempted to recreate

a “historical listener” or “historical modes of listening”—often regarding eighteenth-century music.

In Music in the Galant Style, a book on galant schemata, Gjerdingen considers a historical mode of

listening by wondering if a modern listener could learn to listen like someone of the eighteenth-

century. He writes:

I suspect that traditions of listening have also been slowly transformed. To recover
something of the older, galant tradition, I attempt an archaeology of utterances from that
distant musical civilization, one whose courtiers share with us relatively few social structures
or modes of thought. As the potsherds from my excavations I present musical phrases—
simple musical behaviors from a different time, now given voice in a different social setting.
Can we hear them as Voltaire, Jefferson, or Mozart heard them? Perhaps that is an
unrealistic question….What can be done is to provide an option for the modern listener, a
method for developing a historically informed mode of listening to galant music… It is a
modern reconstruction of an imagined past. But this conjectured galant mode of listening is
nonetheless intriguing and well supported by the writing and practices of eighteenth-century
musicians. (Gjerdingen 2007a, 18–19, italics in original).

Taking a different approach, both Vasili Byros (2009b; 2009a; 2012) and Danuta Mirka (2009) seek

to recreate an eighteenth-century “historical listener” in their analyses by piecing together

information a historical listener may have known. In his case study of the opening to Beethoven’s

Eroica symphony, Byros (2012) establishes a “correlation between real listeners’ responses,

abstracted from documents in the symphony’s [Eroica’s] reception history, and the replicated

patterns unearthed by my own analysis of several thousand compositions from the long eighteenth

century (1720-1840)” (278). Danuta Mirka (2009), in Metric Manipulations in Haydn and Mozart,
148

recognizes that hearing eighteenth-century music the way it was heard by historical listeners is

unlikely, but we can still try to reconstruct the “historical listener”: “one equipped with the

theoretical knowledge of the time (which can be learned from historical treatises) and with a

cognitive mechanism supposedly not different from that of today’s listeners (which thus can be

studied empirically and modelled theoretically)” (xii). Considering any historical listener takes a

stance that historical listening differs from modern, Western culture listening. According to Byros

(2009b), “In order to demonstrate that historical modes of listening may exist, one must articulate

some difference with the present so as to qualify the situatedness of cognition as “historical” in

some way, while maintaining that differences are somehow mediated all the same, in order to allow

“history” a place in cognition” (236). Even though these theorists recognize a difference between

“historical” and “modern” modes of listening, they rarely directly compare the two groups. In

general, theorists do not compare different types of listeners; instead, holding a perspective—such as

a “historical eighteenth-century listener”—for an entire analysis.

Even though theorists could use the analogy framework to analyze the “historical listener”—

and I do for research question 2—, the analogy framework addresses listeners beyond just a

historical listener. It could address listeners with different backgrounds in the same time period or

even compare different groups of listeners to each other (whether “historical” or not). By this, I do

not mean the analogy framework should or does account for each individual, subjective listener

experience—or even that it can account for every individual way a piece of music is heard. Instead, I

mean that the analogy framework provides a space, a method, and terminology for analyzing different

listener experiences. It could be a framework of prediction—what groups of listeners might hear and

when, depending on attention, background, familiarity and so on. Due to the “retrieval stage” in
149

analogical thinking especially, the analogy framework provides a space for considering listener

background, or cultural context of the listener.

Connected still to “the listener,” the last benefit of the analogy framework is that theorists

can use it to approach music analysis as participatory in nature. In music analysis, many elements of

listening become lost: active and participatory acts become frozen; process-oriented engagement

becomes “static.” Even though analogy is a thought process, analogy-making involves an action on

the part of a listener. To make an analogy, a person often has to make some effort, as demonstrated

by multiple studies where participants do not automatically make analogies (e.g. Gentner and

Markman 1997). Therefore, a listener using analogy to understand music is a participatory process—

a listener participates in the music process through the action of analogy. Analyzing music with

analogy, then, contributes to efforts by Christopher Small (1998) and others to transform “music”

from a noun to a verb—music becomes “an activity, and not an object—i.e. one ‘musics’ rather than

performing or listening to a ‘piece’ of music” (Butterfield 2002, 329).

Finally, a theorist can use this framework combined with other music-theoretic

methodologies. The framework is flexible enough to be used with other methods for analyzing

music. In the following analysis chapters (4 and 5), I demonstrate how the analogy framework can

be used with various music-theoretic methodologies, including sonata theory, schema theory, formal

function, and thematic analysis.

Conclusion: The Analogy Framework and Research Questions

The analogy framework outlined here can be used to answer both research questions from

the first chapter. My first research question asks: how do listeners categorize musical themes?

Analogy is used to categorize and relate musical material to each other. To answer this research
150

question, the analyses of theme and variations (chapter 5) entail intra-opus and inter-opus analogies

as well as mapping relations within a musical event. My second research question asks: how do

listeners perceive irony in music by making an analogy between music and language? Analogy is used

to make sense of “unexpected” events in music—nothing else seems to explain it, unless he or she is

using an extra-musical analogy. To answer this research question, the analyses of Beethoven string

quartets (chapter 4) entails extra-musical analogy and mapping relations between musical events.
151

CHAPTER 4
“No Interpretation could palliate this Error of a Genius”: Using
Analogy to Perceive Irony in Beethoven’s String Quartets

Understanding a sentence in language is much more akin to understanding a theme in music


than one may think. What I mean is that understanding a spoken sentence is closer than one
thinks to what is ordinarily called a musical theme. Why is just this the pattern of variation in
intensity and tempo? One would like to say: “Because I know what it all means.” But what
does it mean? I’d not be able to say. As an ‘explanation’, I could compare it with something
else which has the same rhythm (I mean the same pattern). (One says, “Don’t you see, this is
as if a conclusion were being drawn” or “This is, as it were, a parenthesis”, and so on. How
does one justify such comparisons? – There are very different kinds of justification here.)
—Ludwig Wittgenstein, Philosophical Investigations

“And how do you know that you’re mad?”


“To begin with,” said the Cat, “a dog’s not mad. You grant that?”
“I suppose so,” said Alice.
“Well, then,” the Cat went on, “you see a dog growls when it’s angry, and wags its tail when
it’s pleased. Now I growl when I’m pleased, and wag my tail when I’m angry. Therefore I’m
mad.”
“I call it purring, not growling,” said Alice.
“Call it what you like,” said the Cat.
—Lewis Carroll, Alice in Wonderland and Through the Looking-Glass and
What Alice Found There

Lewis Lockwood, in his 2003 book Beethoven: The Music and the Life, writes about the finale for

Op. 95, a tragic F minor string quartet titled “serioso”:

The coda of the Finale has baffled many a dedicated Beethovenian, as it ends the whole
work with a light-fingered, nimble Allegro. Its special character is enhanced by what
immediately precedes it, namely, a pianississimo close in F major that seemingly promises to
conclude the movement on a quiet note of affirmation. But then the coda breaks out with its
running figures in 2/2 meter, sempre piano, building gradually to two climactic arrivals on the
tonic. (329)
152

In his 1971 study The Beethoven Quartets, Joseph Kerman uses genre as a way to explain the

incongruity of this coda:

Then, as though something silently snapped, a very fast alla breve section emerges in the
major mode, a fantastic evocation of an opera buffa finale in which all the agitation and pathos
and tautness and violence of the quartet seem to fly up and be lost like dust in the sunlight.
(182)

William Kinderman (1995) explicitly labels this “brilliant, exhilarating coda” as “problematic” since it

“blithely ignores the dramatic tensions of the work up to that point” (293). Kinderman (1995) and

Kerman (1971) both take note of a contrast between the opera buffa style and the serious and tragic

style heard previously throughout the quartet. The comic effect is emphasized through the use of

2/2 meter and quick harmonic rhythm, all features which Johann Joachim Quantz detailed as traits

of humorous music in his 1752 Versuch einer Anweisung die Flöte traversiere zu spielen (Wheelock 1992,

38).

Vincent d’Indy, obviously frustrated by the coda, wrote: “one might imagine [that] some

light Rossinian finale had strayed into this atmosphere of sustained beauty, and we think that no

interpretation could palliate this error of a genius” (Quoted in Watson 2010, 192). Yet, an

interpretation is afforded by Hatten, who suggests that if passages of a musical work are

inappropriate to the context of an entire piece – such as the “error” of the Op. 95 coda – than an

“ironic interpretation would be one way to reconcile that inappropriateness as a compositional effect

rather than a flaw” (Hatten 1994, 135). Hatten (1994) interprets this coda – which he argues to be

“neither a miscalculation nor a poorly solved problem” – as ironic, describing it as a “shift in level of

discourse” due to the piece’s sudden and extreme change to a contrasting genre (187-88). Rey

Longyear, in his 1970 article “Beethoven and Romantic Irony,” also hears this coda as ironic, writing

that Beethoven “destroys the illusion of seriousness which has hitherto prevailed with an opera buffa-

like conclusion” (649). Tamar Balter (2009), a student of Hatten, considers this coda to be the
153

“prototypical instance of Romantic irony in music” (158). Is there something systematic about this

passage that leads these authors to use irony as a way to interpret the musical “inappropriateness” of

Op. 95? What motivates listeners to perceive a musical passage as ironic as opposed to sincere?

What is Irony?

Linguistic theories may shed light on the way irony works in music since irony is easier to

study in language than music. Scholars have identified many types of irony (Colebrook 2004;

Muecke 1969). In his 1970 book, Douglas Muecke boils all these types down to two principal kinds:

verbal and situational (25), writing: “the former is the irony or the ironist being ironical; the latter is

the irony of a state of affairs or an event seen as ironic” (49). Irony is not a “static rhetorical tool,”

but is part of a “communicative process” (Hutcheon 1995, 13). Raymond Gibbs (2007), a language

psychologist, defines verbal irony as an utterance in which a “speaker says something that seems to

be the opposite of what they meant” (4). According to the language philosopher Cameron Shelley

(2001), a series of events which “defies the normal way in which situations fit with their repertoire of

concepts” can lead to situational irony (775). While verbal irony has commonly been considered a

linguistic phenomenon, situational irony concerns events governed by fate or a “state of the world”

(Attardo 2000, 794). In Hatten’s (1994) definition, “Irony is a higher-order trope inaugurated by the

contradiction between what is claimed (or observed, or done), and a content that cannot support its

reality (or appropriateness) …. there has to be a potential for reversal in interpreting what is “really

meant” by word or deed” (172).


154

Traditional approaches to irony in music often find their origin more in literature or

philosophy than cognitive science.13 Esti Sheinberg (2000)– taking a semiotic and philosophical

perspective – clarifies that “musical incongruities” are what suggest irony in music (50). Byron

Almén (2008) and Michael Klein (2009) use literary theory as a model, considering irony to be a

narrative archetype in the tradition of Northrup Frye (1957). Almén (2008) writes that irony has

different phases: comic and tragic. He describes Frye’s (1957) model:

The comic phases of irony do not entirely displace the initial hierarchy, which is depicted as
humorously flawed and not beyond redemption. Attention is called to the problematic
quality of that hierarchy, but an alternative possibility is not always given …. The tragic
phases of irony, by contrast, feature narratives of despair and integration, in which the safety
of a stable society gives way to unrelieved disasters of oppression. (Almén 2008, 167)

Balter (2009), in her research as well as collaborations with Eddy Zemach (2007), combines

philosophy and semantics of linguistics to identify multiple types of irony in music.

Perceiving Irony in Language

Several empirical studies have shown that certain conditions need to be present for a person

to perceive irony in language. In her experiments, Joan Lucariello (1994) revealed that

unexpectedness – through the violation of some type of norm or schema – and the evocation of

human frailty were always mandatory for understanding situational irony. Lucariello (1994) provides

the following example of situational irony that exhibits both unexpectedness and human frailty: “A

man died when the weather was sunny and calm the day before a hurricane hit; he was electrocuted

by removing an object off of his roof as a precaution against the storm” (132).

13In discussions of musical irony, the role of the performer is often omitted. Performers have the power to enhance the
effect of irony; however, the musical text may prompt musicians to adopt such an interpretation. In this respect, the
performance can greatly maximize or minimize the effect of irony in a piece.
155

Other experiments on verbal irony by Herbert Colston (2001) support many of Lucariello’s

(1994) observations, but he found other conditions as well. Colston (2001) concluded from his

experimental data that two specific conditions were necessary to comprehend verbal irony. First, the

utterance must allude to violated expectations (Colston 2001, 279). Ironic utterances mention prior

statements, desires, beliefs or social norms that do not come to fruition. Second, the speaker flouts

one or more of H.P. Grice’s (1975) maxims of conversation (Colston 2001, 306). Grice (1975)

defines “flouting” as “BLATANLY fail[ing] to fulfill [the maxim(s)]” (49). These maxims are: 1)

Quantity, make your contributions as more or less informative as required; 2) Quality, do not say

what you believe to be false; 3) Relation, be relevant; and 4) Manner, avoid obscurity of expression,

avoid ambiguity (figure 1). H.P. Grice, a Pragmatics14 scholar, created the maxims as part of his

Cooperative Principle: “Make your conversational contribution such as is required at the stage at

which it occurs, by the accepted purpose or direction of the talk exchange in which you are

engaged” (Grice 1975, 45). A person implicitly follows the Gricean maxims in any “cooperative”

conversation.

Grice’s (1975) Maxims


Make your contribution as informative as required (do not be under-
informative)
Don’t make your contribution more informative than is required (do not be
over-informative)
1 Quantity
Example of flouting the maxim of Quantity:
Person 1: “Where did you go yesterday?”
Person 2: “Somewhere on this earth.”
Do not say what you believe to be false
2 Quality
Do not say what you lack adequate evidence for

14Pragmatics is an area of linguistics which looks at conversational context, and not the words themselves, to determine
an utterance’s meaning. Notable figures in pragmatics include H.P. Grice (1957, 1975), J.L. Austin (1962), as well as Dan
Sperber and Deirdre Wilson (1986). For a broader review of the field, see The Handbook of Pragmatics (Horn and Ward
2004).
156

Example of flouting the maxim of Quality:


(The weather is stormy and bad outside)
Person 1: “Nice weather.”
Be relevant

3 Relation Example of flouting the maxim of Relation:


Person 1: “Do you love me?”
Person 2: “I think we should order pizza tonight.”
Avoid obscurity of expression
Avoid ambiguity
Be orderly

Example of flouting the maxim of Manner:


Mother: “How is your new teacher?”
4 Manner
Child: “She’s mean but fair.”
Mother: “How is that?”
Child: “She’s mean to everybody.”

* “Fair” is re-interpreted later on in the joke, pointing to the ambiguity of using the word
“fair” at all.

Figure 4.1: Grice’s maxims, a component of the linguist H.P. Grice’s Cooperative Principle.

Grice (1975) believed that the only maxim which could create irony when flouted was Quality (53).

Since then, scholars such as Salvatore Attardo (2000) and Galia Hirsch (2011) have concluded that

any maxim which is flouted could create irony. Even though the Gricean maxims are most often

used to analyze everyday speech, scholars have also used it as a tool to analyze written prose (e.g.

Rudanko 2007). Colston (2001) found both of these conditions necessary for a person to perceive

verbal irony in language.

Music and Language Analogy

Why should research in linguistics pertain to music? When listeners make an analogy

between language and music, they use language, a more familiar domain, to understand music, a less
157

familiar domain. To perceive irony in music, listeners may make an analogy between music and

language.

A music and language analogy—entrenched in Western culture—structures how a person

listens. According to Justin London (1996), a music-is-language framework “is so well established in

our musical training and the language we use to describe music, it becomes wholly transparent—or

to many listeners music becomes a subclass of linguistic phenomena” (51, italics in original). George

Lakoff and Mark Turner would call this analogy a “basic conceptual metaphor” since it is a common

“conceptual apparatus” and prevalent within a culture (Lakoff 1989, 51). According to London

(1996), this analogy has fixed aspects between source and target domains (52). He includes a list of

some ways where “slots” in language map onto corresponding “slots” in music:

Figure 4.2: Linguistic slots that map onto musical slots (graphic from London 1996, 52)

London considers this analogy useful in two ways. First, it allows listeners to make inferences:

“Thus, we can, for example, regard contrapuntal alternations between instruments as analogs to

turn-taking in conversation or argument, musical discontinuities as parenthesis, and so forth” (52).

Second, this analogy allows listeners “to evaluate musical gestures as we would linguistic utterances,”

for example “anomalous musical structures are considered as a species of Gricean flouting”

(London 1996, 52). Because of the music and language analogy, listeners may have similar

expectations for music as they do for language, such as a desire for composers to “cooperate” and

follow the Gricean maxims. Beyond higher-order relations of “speakers” mapped onto “composers”
158

and “hearers” mapped onto “listeners/audience,” listeners can make other analogies between music

and language.

Not only do listeners use a culturally-ingrained “music and language analogy” as a reference

for reasoning about music, but listeners, also, map relations in linguistic structure—such as between

sections or words—onto musical structure. That is, musical passages may suggest irony if they are

analogous to linguistic passages that also suggest irony. A musical work is “conceived as a sequence

of events” (Agawu 2009, 7). The contradictory events in Op. 95, namely the opera buffa coda

following the pathos of the quartet, exhibit similarities to contradictory events in situational irony.

Situational irony is caused by an act of fate and has no “speaker.” Beethoven, however, serves as a

“voice” or “speaker” in Op. 95 since he placed these events – motives, phrases, and the like –

together (Cone 1974). Since a composer acts as a “speaker,” music has the potential to flout the

Gricean maxims in a way analogous to verbal irony. When a listener hears the incongruous and

inappropriate coda, he or she desires to understand or “solve” the musical anomaly. By realizing the

similarity between this passage in Op. 95 and instances of irony in language, a listener may make an

inference that the piece is ironic.

Perceiving Irony in Music using Analogy

Drawing from Colston’s (2001) research on verbal irony and Lucariello’s (1994) research on

situational irony, I propose the following cognitively-informed conditions for perceiving ironic

communication in music:

1) Violation of a temporal expectation generated by some norm or violation of a general knowledge structure
in the form of a schema
2) Flouting one or more of the Gricean maxims.

A theory of analogy could provide an explanation for how such violations of expectation and

floutings of Gricean maxims are perceived similarly in music and language. In short, these
159

conditions, in addition to priming a culturally-ingrained “music and language analogy,” give clues to

relations in ironic linguistic structure a theorist could map onto musical structure, prompting him or

her to hear a musical passage as ironic. Each condition relates to a different step of the analogy

framework (figure 4.3).

Figure 4.3: The Analogy Framework

The violation of expectation condition relates to the retrieval step, while the flouting of Gricean

maxims condition relates to the mapping step. To recognize a violation of expectation, listeners

need to have expectations, often learned from previous experience with musical structures. I use the

retrieval step to analyze knowledge of musical norms and structures as well as Gricean maxim

conventions. Listeners retrieve knowledge they have of music as well as knowledge of how Gricean

maxims are used in cooperative conversation (figure 4.4), then use this knowledge in the mapping

step.
160

Figure 4.4: The Retrieval Step

To recognize flouting of Gricean maxims in music, listeners map relations in linguistic structures

that flout the maxims onto similar relations in musical structure. I use the mapping step (figure 4.5)

to analyze not only higher-order relations connected to the culturally-ingrained music and language

analogy (e.g. speakers onto composers, etc.), but also relations connected to flouting of Gricean

maxims.

Figure 4.5: The Mapping Step

Before proceeding to case studies of irony in Beethoven, I will explore these knowledge structures in

more depth: violation of expectation and the Gricean maxims.


161

Retrieval: Violation of Expectation

With the right knowledge, music has the ability to create expectations to be either fulfilled or

denied (Meyer 1956; Narmour 1990; Huron 2006; Margulis 2007). Meyer (1956) believed

expectation to be an important part of the listening experience, arguing:

All tendencies, even those which never reach the level of consciousness, are expectations. For
since a tendency is a kind of chain reaction in which a present stimulus leads through a series of
adjustments to a more or less specified consequent, the consequent is always implied in the
tendency, once the tendency has been brought into play. (25)

Musical expectations come about in a number of ways, including “reflexes, conceptual knowledge,

mechanisms of statistical learning, logic, or hard-wiring” (Margulis 2007, 2005). On one level,

musical expectations can be created through the structural elements of the music (e.g. Bharucha

1987; Krumhansl 1995). On another level, genre or form entail a set of expectations held by the

listener (Caplin 1998; Hepokoski and Darcy 2006; Cobly 2005; Cobly 2008). In fact, Paul Cobly

(2005) begins his definition of genre as “not a set of textual features that can be enumerated; rather,

it is an expectation” (41). According to James Hepokoski and Warren Darcy (2006), “Genres exist

only insofar as production and reception communities agree to act as if they really did exist, as sets

of rules, assumptions or expectations” (606). A sonata form movement inherently raises different

expectations than an aria. Since music can create expectations, composers have been known to play

with, delay, or even violate these expectations.

Musical expectations are not all created equal. Elizabeth Margulis (2007) categorizes

expectation by considering the five basic ways they can vary: origin (Where does the expectation

come from?), nature (What is it like to have the expectation?), time course (How long is the

expectation sustained?), object (What kind of entity is the target of expectation?), and consequence

(What is the effect of this expectation?) (205–6). In her research on irony, Lucariello (1994) classifies

ironic events not just as unexpected, but unexpected in a culturally recognized way. These events
162

presuppose a general knowledge structure as opposed to being idiosyncratic. Therefore, ironic

violation of expectation may vary within Margulis’ (2007) taxonomy, but must also be culturally

understood.

To recognize a violated expectations within a musical style, a listener has similar or shared

musical knowledge with a composer, and then retrieves this knowledge from memory when

necessary; therefore, this condition relates to the retrieval step. This sharing of knowledge between

composer and listener, or the cultural context of a listener, is equally necessary for understanding

ironic communication (Gibbs and Colston 2012, 303–4). For example, if a speaker says, “100 Main

Street burned to the ground,” this is not ironic in and of itself. However, if the hearer knows that

100 Main Street is the address of the firehouse, the sentence is understood to be ironic. Some

scholars have noted the importance of shared knowledge that is retrieved to understand irony and

other compositional effects in music. Everett (2004) writes: “Irony is communicated more often

through implicit rather than explicit signals: in subtle cases, the transliteral message is embedded in

such a way that only the culturally and ideologically “competent” audience comprehends the double

exposure” (5). Here, I define shared knowledge retrieved in the retrieval step as the stylistic

knowledge structures common to Beethoven and experienced late eighteenth-/early nineteenth-

century listeners. To approach violation of expectation in the following case studies, I consider the

galant voice-leading schemata of Gjerdingen (2007a), Caplin’s (1998) theory of formal function, and

Sonata Theory (Hepokoski and Darcy 2006) as part of the shared knowledge between Beethoven

and his learned audience.

I review galant schemata here only in connection to Beethoven, since it was considerably

reviewed in the second chapter. A few scholars have reservations regarding galant schemata in

Beethoven, a composer associated as much, if not more, with the nineteenth-century than the
163

eighteenth-century galant. Regarding possibilities of schemata in Beethoven, however, Gjerdingen

(2007a) writes:

Generations of biographers have depicted him [Beethoven] as so archetypically the German


musician that his deep roots in the Italian galant style have been easy to overlook… But
perhaps a simple example from an early piano sonata can suggest how faithfully he had
absorbed the traditions of the galant schemata, even as he was dilating and dramatizing them
in unexpected ways. (237, italics in original)

Since Beethoven trained with galant methods, his music, including his late style, may be a fruitful

repertoire for schema analysis.

Caplin (1998) defines formal function as the “more definite role that the group [a self-

contained “chunk” of music] plays in the formal organization of the work” (9). Listeners recognize

musical sections as having different formal functions:

By means of specific musical criteria, largely based on harmonic-tonal relations but also
involving processes of grouping structure, melodic directionality and texture. The various
formal functions that I have defined relate to three traditional categories of temporal
expression – beginning, being-in-the-middle and ending. In addition, some framing functions
express the sense of before-the-beginning and after-the-end. (Caplin 2005, 115)

Formal function, then, involves the types of formal categories that a listener would most likely

expect in different temporal points of a piece.

Hepokoski and Darcy’s (2006) Sonata Theory (figure 4.6) is more a dynamic process

concerning tasks and goals than simply a mold or template. The classical sonata form is a musical

plot revolving around the drive towards two perfect authentic cadences (PACs), one at the end of

the exposition (the essential expositional closure [EEC]) and one at the end of the recapitulation (the

essential structural closure [ESC]) (Hepokoski and Darcy 2006, 12). Hepokoski and Darcy (2006)

recognize that the “heart of the theory” is this “recognition and interpretation of

expressive/dramatic trajectories toward generically obligatory cadences” (13). Instead of discussing

“themes” or “key areas,” Hepokoski and Darcy (2006) speak of normative procedures within
164

“action-space” or “zones.” A normative exposition begins with a primary theme (P) which proceeds

to a transitional zone (TR). A secondary theme (S) follows in a new key which is secured by the

EEC. After the development, the recapitulation has a similar sequence of “zones,” except that the

return of S resides in the home key, ultimately driving toward the ESC.

Figure 4.6: The generic layout of Sonata Form (graphic from Hepokoski and Darcy 2006, 17)

These modern approaches to eighteenth- and early nineteenth-century Western art music –

based on extensive surveys of pieces from this time period – create an understanding for what might

have possibly been in the sharked knowledge of Beethoven and his audience.

Mapping: Flouting the Gricean Maxims

To hear flouting of Gricean maxims in music, listeners may expect composers to

“cooperate” and so use analogy to map relations of flouting the maxims between language and

music. The Gricean maxims are often applied to – and even assumed in – natural language. Fewer
165

scholars, however, have considered them in relation to musical communication. London (1996), a

notable exception, writes that even though pieces of music are not conversations, a listener may

expect music to uphold conversational norms:

When we find in them [pieces of music] violations of [Grice’s] cooperative principle, we tend
to assume that these violations are intentional floutings of one (or more) of the principle’s
maxims. Indeed, we often encounter musical descriptions precisely along these lines: these
that are too long or too short are described in terms of overstatement or understatement,
i.e., violations of quantity, melodic and harmonic non sequiturs (for example, a ‘deceptive’
cadence) are violations of relation; ambiguous (especially tonally ambiguous), rhythmically
chaotic, or overly dense musical textures are violations of manner. (59, italics in original)

In addition to London (1996), Charles Nussbaum (2007) also believes music can behave similar to

conversational structure. He comments, for instance, on music’s ability to flout the maxim of

Relevance: “Musical ‘irrelevance,’ whether thematic, harmonic, or formal, is something we all have

encountered. Mozart’s Musical Joke, K. 522 makes a study of it” (Nussbaum 2007, 125).

Since the maxims represent rational behavior, Grice (1975) does not limit their application to

language alone, but believes they can be applied to non-verbal communication as well (47). A person

expects any speaker attempting to communicate to follow the Cooperative Principle and, therefore,

the maxims. In his Origins of Human Communication, Michael Tomasello (2008) writes that “human

communication is thus a fundamentally cooperative enterprise” (6). If music is a type of human

communication, then a listener should expect the composer to be “cooperative” and implicitly

follow the Gricean maxims. Music’s possible communicative implications have been widely

discussed. For instance, Ian Cross (2005) argues that “music’s apparent ambiguity does not debar it

from being considered to be a communicative medium” (32).

Central to communication, however, is intention (Grice 1975; Tomasello 2008). In order for

music to be communicative, a listener may need to perceive it as having intention of some kind. As

previously mentioned in the third chapter, Meyer (1989) argues that music has intention behind it
166

since a composer makes certain choices within a constrained musical style (138). Understanding

music as a form of human communication invites the possibility that listeners expect the composer

to more or less follow the Gricean maxims. If a listener expects a composer to follow the Gricean

maxims, then he or she may be primed to map relations from instances of flouting the maxims in

language to music. Then, a listener might make a particular inference if he or she believes the

composer is flouting one of these maxims. In connection to recognizing instances of flouting the

Gricean maxims, there are other facets of the mapping step to consider, including analogical

distance and level of musical relation.

Mapping: Extra-musical Analogy

Perceiving irony in music involves a distant or extra-musical analogy. The two analogs—

music and language—are from separate domains.

Mapping: Relations between Musical Events

Often, relations involving flouting the Gricean maxims map onto relations between musical

events (as opposed to relations within a musical event). Instances of flouting often involve a

composer’s play with sections of a musical form.

Violation of expectation and flouting the Gricean maxims create relations in language that

could mirror relations in music, prompting a listener to make an analogy to hear these musical

passages as ironic. Listeners then use the analogy framework in order to infer that the piece

communicates irony.

Case Studies
167

To illustrate the analogy framework and how it relates to irony, I analyze three Beethoven

movements drawn from middle to late string quartets. These are movements which others scholars

have also heard as ironic.15 I build upon these scholars’ observations and use their analyses as

evidence for my conditions of ironic perception.

With the wealth of repertoire in this time period, the use of only string quartets here may

seem limited. The choice of string quartets was not accidental, however. Beethoven’s society

maintained a divide between amateurs (Liebhaber) and connoisseurs (Kenner) in music. Eighteenth-

century composers, to exercise caution, often composed music catered toward both demographics.

In a 28 December 1782 letter to his father Leopold, Wolfgang Amadeus Mozart wrote: “Although

here and there connoisseurs alone can obtain satisfaction, the non-connoisseurs will be satisfied,

without knowing why” (quoted in Bonds 2008, 36, italics in original). The string quartet genre

especially was connected to the connoisseur, often having associations with “exclusivity” (Hunter

2012, 57). Leon Botstein (1994) writes, “By 1827 the quartet form (the Beethoven quartets in

particular) had gained augmented status as sophisticated and profoundly communicative pieces

aimed at the truly educated” (90). Engaged listeners could form Beethoven quartet clubs, as

Christian Friedrich Michaelis (‘M.’) described in an 1829 article of the magazine Berliner allgemeine

musikalische Zeitung:

For some time musicians and friends of music have founded numerous quartet clubs
[Quartettvereine], whose primary, or exclusive exercise is the study of Beethoven’s quartets.
It could be called more than a ‘club’ when some of the latest and most difficult masterworks
are gone through fifty or a hundred more times in order fully to enter into the spirit of the
master, and to play him worthily. (quoted in Hunter 2012, 58)

15Op. 130 is the only exception. Even though scholars have not used the term “ironic” to describe this movement,
however, many have alluded to ironic-like behavior.
168

The string quartet presents itself as the ideal genre in which to look at violation of expectations.

Since irony depends so much on shared knowledge for comprehension, it may be a compositional

device more apt to be used and recognized by “insiders,” such as these connoisseurs. Beethoven and

other composers expected to share knowledge with their audience of this particularly

“communicative” genre.

Op. 95/iv (1810)

A mention of op. 95 appeared in a letter from Beethoven to Sir George Smart: “NB. The

Quartett is written for a small circle of connoisseurs and is never to be performed in public”

(Anderson 1961, 2:Letter 664). Beethoven held high opinions for the audience of this quartet. It was

him, as opposed to an editor, who titled the work “Quartett[o] Serioso” (Longyear 1970, 649).
169

Coda

Figure 4.7: Beethoven, Op. 95 (Serioso), fourth movement, mm. 126-153

By this point, a reader knows how scholars usually interpret this swift and light end to a previously

serious quartet. As the turbulent F minor movement comes to an assumed close, a “Picardy third”

creates a sense of release (figure 4.7, m. 132). Yet, as we know, the trickster “Picardy third” dissolves

into our opera buffa or “Rossinian” finale (to borrow d’Indy’s description).

An avid Beethoven listener may cry that op. 132 shares a similar finale. Like the “Serioso”

quartet, op. 132 also shifts from pathos-laden minor to joyful major (figure 4.8). Scholars, however,
170

tend not to consider the ending of op. 132 ironic. Why does the coda of op. 95 strike such an ironic

tone, but the coda of op. 132 does not? In her comparison, Balter (2009) argues that op. 95 does not

evoke irony through a change in affect or mood, but because the coda alludes to the different genre

of opera buffa (164). Yet, why is this difference significant?

Figure 4.8: Beethoven, Op. 132, finale, mm. 290-309

The conditions outlined earlier, which I analyze through the analogy framework, may explain

why a listener perceives op. 95 as ironic, but not op. 132. The op. 95 coda violates expectations since

it is less common for movements to shift to a different affect, topical genre or use new musical

material without any transition. This expectation, to address Margulis’ (2007) category of “origin,”

stems from familiarity with stylistic norms. As a result, the expectation would be part of a shared

knowledge of retrieval between eighteenth-/early nineteenth-century listeners and composers.


171

“The ending of op. 132 violates expectations, as well,” our avid Beethoven listener may

insist, “in its shift to a different affect and use of new musical material.” Both op. 95 and op. 132

satisfy the first condition. A violated expectation, then, is not a sufficient indicator of irony in music.

Zemach and Balter (2007) appear to agree, as they write:

First, music raises expectations: competent listeners project situations they consider right for
the (musical) conditions at hand. Second, listeners compare and contrast the anticipated
situation (a musical event) with the one that does occur in the work. These traits, however,
are necessary, but not sufficient, conditions for musical irony, for contrasting situations and
frustrating expectations have many other uses too: they create tension, drama, or simply add
interest to the work. (183)

The flouting of the Gricean maxims may offer a “missing piece” to complete our irony puzzle. The

flouting of the Gricean maxims offers a condition that works in tandem to expectation. The

unexpected coda of op. 95 provides a paradigmatic example of flouting the Gricean maxims,

specifically Relation and Quantity.

According to the maxim of Relation, whatever statements a speaker adds should be relevant

to the conversation at hand. An abstraction of a relation involved for flouting the maxim of Relation

is:

DOES-NOT-RELATE(x, y)

The coda that Beethoven composed for op. 95 is not relevant to the rest of the quartet in either

affect or genre. In addition, the music of the coda was not previously heard before in the movement.

The fact that Beethoven provides no transition material or preparation for the shift further

emphasizes its irrelevance. To make an analogy between music and language, first a listener retrieves

relevant knowledge of sonata form and formal function conventions as well as knowledge on the

maxim of Relation. In this case, a linguistic flouting of Relation could take be realized as such:

DOES-NOT-RELATE(sentence 1, sentence 2)

In Op. 95, a listener hears piece + coda, where coda does not relate to the rest of the piece.
172

DOES-NOT-RELATE(piece, coda)

A listener could make this abstraction from these instances so to map this relation from language to

music (figure 4.9). In the evaluation step, a listener infers the musical passage portrays irony.

Figure 4.9: The analogy framework and Op. 95, relation maxim, form

On the other hand, while the coda for op. 132 may defy expectations, it still relates to other

parts of the movement. Even the change in affect sounds linked, according to Hatten (2004):

Unlike what I have called the addendum [coda] to the finale of the “Serious” Quartet (Op.
95)… The extensive coda to the finale of Op. 132 is thematically and expressively integrated
as a logical outcome to both the movement and the quartet…. Beginning with the
modulation in the first theme at m. 7, we have glimpsed the potential of this ending. (285)

Kinderman (1995), also uses the word “integrated” to describe the difference: “As in the quartetto

serioso, op. 95, the finale of op. 132 contains a coda in the major mode, but the conclusion is now far

more integrated with the work as a whole than it was in the earlier quartet” (298). Both Hatten and
173

Kinderman focus on the coda’s thematic material and how it connects to previous themes in the

movement. Kerman (1971), however, places importance on transitions:

The Finale of the A-minor, after resuming the pathos of the first movement of the
composition in a dynamic analogous to that of the F-minor quartet, concludes with an
analogous volte-face to the major mode, Presto. Again the first impression is of lightness and
play, as though a great weight of involvement has been lifted. But the shadows are not quite
forgotten, and the contrast in mood, though again staggering, is not so puzzling because
(among other reasons) the new major-mode material is led in by a definite thematic passage.
The concluding section itself is longer and self-sufficient; Beethoven can afford really to look
at it. In this ending, the prior pathos seems genuinely – “serious” – encompassed. The play
seems genuinely earned or achieved. (183–84)

Seemingly small differences, such as no transition materials, unconnected themes, and switching

genres, ultimately create the effect that op. 95’s coda purposely does not relate to the rest of the

quartet, prompting a mapping between it and the maxim of Relation. Op. 95’s coda aligns with one

of David Huron’s (2006) nine devices for musical humor; particularly, “mixed genres” (284).

Beethoven flouts the maxim of Relation by drastically switching from tragedy to opera buffa.

Op. 132, on the other hand, applies the strategies that op. 95 lacks, evoking a sincere coda and cuing

an expressive genre—defined as “category of musical works based on their implementation of a

change-of-state schema” (Hatten 1994, 290)—which leads “from tragic to transcendent” (Hatten

2004, 280). This move from “tragic-to-transcendence” relates within the context; therefore,

Beethoven does not flout the maxim of Relevance. As no similar expressive genre (other than

perhaps one of “irony”?) easily maps onto op. 95, its coda seems irrelevant to its surroundings.

Instead of “from tragic to transcendent,” the coda of op. 95 could be considered “from seriousness

to frivolity” (figure 4.10). In a case study of selected humorous Peter Schickele music, Huron (2006)

shows that Schickele’s “most common tactic [of mixed genres] is to juxtapose “high art” and “low

art” styles. Typically, the “high art” style is established first, followed by interjections of “low art”

materials” (285). Op. 95 would be an example of this “high art” followed by “low art” styles, or
174

“seriousness-to-frivolity.” Therefore, the relation is not just DOES-NOT-RELATE in regards to

structure, but also genres and their associations. This can be mirrored in linguistic examples of

flouting the Gricean maxim of Relation, as well, as seen in figure 4.1:

Person 1: “Do you love me?”


Person 2: “I think we should order pizza tonight.”

The question “do you love me?” is serious, followed by a frivolous response about ordering pizza.

Figure 4.10: The analogy framework and Op. 95, relation maxim, genre

For that reason, op. 95 flouts the maxim of Relation while op. 132 does not, creating a sense of

irony in the former but not the latter.

More often than not, the flouting of one maxim will connect to the flouting of another.

Beethoven takes advantage of Formenlehre to flout the maxim of Quantity in op. 95. This flouting

directly relates to the flouting of Relevance previously discussed. According to the maxim of

Quantity, a contribution to a conversation should be no more or no less informative then required.


175

The “perfect amount of information” varies depending on the form of the piece; in the case of op.

95, the form is a modified seven-part sonata-rondo form (ABA’CB’A’’) or Expanded Type 1 Sonata-

Rondo Mixture (Type 41-exp) (Hepokoski and Darcy 2006, 409). Caplin (1998) writes that eliminating

the A between the C and B’, as seen here in Op. 95, is a common-sonata-rondo deviation, often

used by Mozart (238). This modified sonata-rondo also has an effect on the rotational structure of

the movement compared to other rondo types. Hepokoski and Darcy (2006) write:

In one typical scenario, an episodic or developmental “billowing-out” can occur when the
tonic return of Prf, launching Rotation 2 and the recapitulation, is followed by either a closed
episode or a genuine development (often of Prf or TR material) or by both. This episodic
and/or developmental expansion has frequently been mistaken for “Episode 2 (C)” of a
seven-part sonata-rondo in which the third statement of the refrain (A) is eliminated. In
other words, some analysts have parsed this familiarly Mozartian pattern as ABACB’A,
suggesting that it arises as an “incomplete” ABACAB’A design. Once again, we see the
pitfalls of reducing the Type 4 sonata to a mere string of alphabetic symbols. Type 41-exp, the
expanded Type 1sonata-rodo mixture, may more meaningfully be conceived as a rotational
structure…. The Rotation 2 interpolation, even if it is a new and separate episode, frequently
links up at its end with the end of the original TR, now transposed to the to the tonic key.
The Rotation 2 expansion often leads to a crux-point that slips into correspondence
measures in TR-space. This means that it rejoins an ongoing rotation-in-progress, one that
had begun with the second tonic statement of Prf.” (Hepokoski and Darcy 2006, 410)

As opposed to other types, such as the Type 3 Sonata-Rondo mixture, which is four rotations, the

Type 41-exp has three rotations (Hepokoski and Darcy 2006, 410).

Rotation 2: Prf (TR) [development or episode] … TR ‘S/C  RT

Therefore, the second rotation begins at m. 51 and does not end until the beginning of the third at

m. 98.

There are a number of problematic cadences within this piece in addition to expected ones.

The sonata-rondo-space of op. 95 lacks any PACs, a problematic turn of events since a central part

of Sonata Theory involves a piece’s drive to the EEC and ESC. There is a structurally important

i:HC in m. 19 (with caesura-fill in the first violin lasting for a bit). Then all voices stop on a unison
176

Db. In a way, this dramatic Db, highlighted by the rests and emphasized by the “sol-le” movement

in all voices, sounds almost as if a wrong turn, playing into Beethoven’s possible use of “play” in this

piece. However, he recovers by transforming it into the seventh of a viio7 in m. 22, marking the A

material as a seed for a beginning of a dependent transition. M. 23 marks the beginning of this

dependent TR in Hepokoski and Darcy’s (2006) tradition of a grand antecedent and dissolving grand

consequent. After a shift from dependent transition to an independent one, an imperfect Medial

Caesura (MC) occurs in m. 43. The MC is lacking, if not completely blown through as some kind of

rejection. This feeds into an S/C blend in v (m. 43); though the short S theme, in a way, questions its

own existence due to the weak MC and also the length of the S/C space. Finally, the exposition fails

to produce a satisfactory EEC, with only a v: IAC as opposed to the expected PAC in mm. 47-48.

This leads directly into the return of A’ (beginning in m. 51), though a severely truncated version.

Beethoven flouts the maxim of Quantity16 (being “under-informative”) with his shortage of

cadences, specifically PACs within the Sonata-rondo space. Caplin (1998) discusses the importance

of “cadential goals” in a sonata exposition (196). Similar to Hepokoski and Darcy (2006), he requires

a PAC to close the Secondary Theme key (196). Where a PAC should occur (mm. 47-48) to close

the Secondary Theme and mark the exposition’s end, a listener hears the previously mentioned

subtle imperfect authentic cadence (IAC) instead. This cadence hides under a continuing flurry of

sixteenth notes in a rush to return to Primary Theme-Refrain (Prf) material. Such a coy IAC creates a

“failed” exposition as it falls short in marking a satisfactory EEC. Hepokoski and Darcy (2006) note

that an EEC can be “more weakly secured” with an IAC, though it is rare (167). The definition of a

16 Of course, attempting to justify flouting of Quantity is difficult since music traditionally expands, contracts, repeats,
and so on. What is too much or too little information in music? It may be that it can be considered “too much” or “too
little” only in relation to another part of the piece (as in Op. 95). It may also be “too much” if it goes above and beyond
what is “necessary” to establish form and “too little” if it falls short of establishing it (also, as in Op. 95). Finally, it may
be “too much” if it draws a listener’s attention enough for them to question.
177

“failed” exposition is a “nonclosed expositions,” when there is no proposed EEC-effect (Hepokoski

and Darcy 2006, 177). Therefore, one could argue that this makes a “failed” exposition, though

perhaps not as dramatic as more pertinently failed expositions.

Beethoven’s “under-informative” nature does not end with the EEC, but continues on to

the ESC as well. Certain complications exist for the ESC in sonata-rondos (figure 4.11). As

described by Hepokoski & Darcy (2006):

[Sonata-rondos] present a more complicated conceptual situation: As hybrid forms they can
be viewed from two different perspectives: from that of the sonata and that of the rondo…
the “sonata” aspect will ask for the presence of ESC at the end of the recapitulation’s S
theme…. A simple rondo’s ESC is delayed until the moment of the PAC-closure of its final
thematic statement – which in a sonata-rondo occurs after the recapitulation is completed.
Thus a sonata-rondo presents us with the possibility of two conflicting ESC claims. (428)

Figure 4.11: Graph of Sonata-Rondo form and placement of EEC and ESCs (graphic from Hepokoski and
Darcy 2006, 428).

At the “sonata” ESC moment, Beethoven presents an IAC similar to the exposition, yet eliding with

returning Prf material. The “rondo” ESC moment brings even more surprises, however, since there is

no authentic cadence at all. Beethoven teases his listeners with implications that a PAC, or any
178

authentic cadence for that matter, should occur in this moment. For instance, the Indugio (figure

4.12) is a predominant-expansion schema documented by Gjerdingen (2007a). It has a bass that

prolongs scale degree 4, a soprano that outlines scale degrees 2, 4, and 6 (with an emphasis on 6),

and 6/5 figured bass (or, in the case here, the soprano highlights 6 and an inner voice highlights 2).

This schema commonly “served to hold back an expected big cadence, thereby heightening

anticipation” (436). In mm. 118-20 (figure 4.13), an Indugio implies a cadence that is never fully

realized.

Figure 4.12: The Indugio from Gjerdingen (2007a, 464).


179

Figure 4.13: Beethoven, Op. 95 (Serioso), fourth movement, mm. 118-20

Such a lack of PACs, in both ESC positions, creates a non-resolving recapitulation

(Hepokoski 2001/2002, 152). This could indicate that the sonata-rondo is “failed.” This “failed”

sonata-rondo as a whole sets up reasonable expectations for a discursive coda which would solve the

“unfinished business” about F minor or F major. The expectation for a coda to “finish” the

“unfinished” especially applies to Beethoven, as Burnham (2000) contends, “Critics from our own

century tend to regard the Beethovenian coda as the locus for unfinished business” (121). For

instance, Beethoven’s contemporary Egmont overture also features a non-resolving recapitulation,

placing the burden on the coda to fully realize the F major via a PAC (Hepokoski 2001/2002). A

listener may expect a similar coda from op. 95, but a rather different one appears instead. If the

sonata-rondo space flouts the maxim of Quantity by being under-informative, the coda flouts the

maxim of Quantity by being over-informative. The coda contains an abundance of authentic

cadences in F major, including a number of PACs such as m. 137 (weak), m. 151, and m. 175. The

coda proves to be irrelevant (as previously discussed) by not solving the problems outlined earlier as

expected – for example, with a triumphant F major. Because of the sudden shift in discourse and
180

opera buffa display, the coda sounds like the wrong coda for its context. This makes it irrelevant to the

“failed” sonata-rondo at hand. The lack of cadences (“under-informative”) prior the cadence-

abundant coda (“over-informative”) flouts the maxim of Quantity. Figure 4.14 illustrates the

relations of “under-informative,” that of:

LACKS-INFORMATION(x, y)

Though, it can be flipped to account for “over-informative” as well:

EXCESSIVE-INORMATION(x, y)

A listener maps this relation from language to music.

Figure 4.14: The analogy framework and Op. 95, quantity maxim
181

Even though other factors exist, Beethoven’s play with cadences contributes to the

“irrelevant” nature of his coda. Between the violated expectations and the flouting of the Gricean

maxims, a listener perceives irony when listening to op. 95.

Op. 131/V (1826)

The next case study begins more than a decade later. The scherzo of op. 131, according to

both Longyear (1970) and Zemach and Balter (2007), creates irony through a number of unexpected

modulations and play with tempo. During Beethoven’s lifetime, no public performance took place

of this quartet; however, it was rehearsed and “heard by the usual ‘small circle of connoisseurs’,

including Schubert” (Watson 2010, 264).

A playful melody in E major triggers a jesting nature, appropriate for a movement which

Kerman (1971) believes to be the “most childlike of all Beethoven scherzos” (338). The melody

outlines a tonic triad and provides a strong initiating function. After this opening motif, the melody

feeds into a continuation. Beethoven ends the phrase, though, with a jolting rat-a-tat-a-tat of chords

(figure 4.15, mm. 17-18) instead of an expected cadence. After an extended sequence from G#

major (figure 4.15, mm. 19-20), A major (figure 4.15, mm. 21-22), to B major (figure 4.15, mm. 23-

24), a listener assumes Beethoven “means” to conclude the modulation on the dominant. At the

end, however, the scherzo is unable to hold onto its B major dominant: “When it gets there [the B

major], it fusses inordinately, slows down, expostulates with this key and persuades it (against its

better judgment) to try being G# [minor]” (Kerman 1971, 339).


182

Figure 4.15: Beethoven, Op. 131, fifth movement, mm. 13-48

The opening motif returns (figure 4.16, m. 41), but this time in G# minor. As a listener, it

seems that the performers are not convinced this is the appropriate course of action. For all
183

instruments, Beethoven writes in a ritardando. The players then slow down (figure 4.16, m. 44),

creating the effect that they are confused, before landing on V7 of G# minor. From this V7,

Beethoven creates a deceptive move (figure 4.16, m. 45) to what at first seems to be a VI chord, but

is actually a modulation back to the original E major tonic. The motif repeats, satisfied now with the

current key, and reprises the opening as if nothing out of the ordinary had taken place. In terms of

Caplin’s (1998) formal functions, the choice of stopping and re-beginning in the “correct” key

creates an effect of multiple initiating functions strung together. Beethoven’s first “take” at this

motif in G# minor is interrupted, prompting a repeating initiating function in the more appropriate

E major. This second initiating function, then, feeds into a continuation and cadential function as

expected.

Figure 4.16: Beethoven, Op. 131, fifth movement, mm. 41-48

Zemach and Balter (2007) discuss the irony of this movement as such:

These early modulations to the key of the mediant are ironic, because once it becomes clear
that the E major chord is a correct beginning of the reprise, the modulation to G# minor
looks wrong. Beethoven makes it look as if this (unconventional) modulation to the key of
the mediant was a mistake by writing no modulation back to the home key. (187, bold
added for emphasis)
184

This piece violates expectations by modulating to the mediant instead of the expected dominant. In

addition to this unusual move, the scherzo shifts suddenly back to the tonic without preparation. In

regards to Margulis’ (2007) question of “origin” (205), these expectations come from conventions

for modulation that would have been part of the shared knowledge between experienced eighteenth-

/early-nineteenth-century listeners and composers. A listener may argue, however, that play with

tempo strikes him or her as more surprising than the modulations themselves. The performers of

this piece slow down dramatically, as if perplexed, creating a type of disillusionment. This creates the

effect that someone maybe made a mistake.

In relation to the mapping, Beethoven’s musical structure mimics the relation of flouting the

maxim of Quality: do not say what you do not have the evidence for or what you do not believe to

be true. The unexpected and quite abrupt return to the opening motif in the tonic, coupled with the

performers’ “confusion,” reinforces the effect that Beethoven did not provide the listener with the

“correct” information; to put rather bluntly, he was “wrong.” The abstracted relation here would be:

INCORRECT-IN-CONTEXT(x, y).

Here, issues arise regarding music’s lack of propositional truth: music cannot have “wrong

information” in the same way language can. Analogy, however, bypasses some of these issues. For

analogy, music does not need to “be truthful.” Instead, music needs to behave, or have relations,

that imitate either “correct” or “truthful” relations in language. In the case study here, an abrupt

transition returning to an opening motif creates behavior with a similar relation: the move was a

mistake. Therefore, the music may have a relation of ABRUPT-TRANSITION(phrase-in-iii,

opening-motif-in-I), while language may have a relation of UNTRUE-IN(sentence, context/reality).

As seen in figure 4.17, these can be re-represented as relation INCORRECT-IN-CONTEXT(x, y).


185

Figure 4.17: The analogy framework and Op. 131

A listener maps this relation from language to music and may perceive this moment as ironic as he

or she infers Beethoven as flouting the maxim of Quality.

It may be that a listener prefers to use the term “humorous” instead of “ironic” to describe

op. 131. Humor and irony, while often regarded as different phenomena in the psychology of

linguistics (Hirsch 2011, 531), share many similarities. For instance, violation of expectation is often

considered a cue for both irony and humor (Raskin and Attardo 1994; Hirsch 2011). In this

particular case, op. 131’s humorous nature could be attributed to a Western art music tradition of

“performer bungling.” For this style of humor, pieces are purposely composed in a way that the

composer or performer (or both, in the case of composer-performers) seem to commit mistake after

mistake. Friedrich August Weber, in an 1800 allegmeine musikalische Zeitung article, believed that the
186

“artfully imitated bungling”17 of an incompetent composer symbolized the most sophisticated

humorous music (quoted in Mirka 2009, 300). To imitate this “fake bungling,” though, a composer

needed to be “well-acquainted” with the common ways that composers usually “botched”

compositions (Zenck 2008, 57). Essentially, he or she needed to have the right information in his or

her shared knowledge.

Performer/composer bungling provides a methodology for analyzing humor in op. 131.

Claudia Maurer Zenck (2008), for instance, uses such a perspective in her study of Beethoven’s op.

31, no. 1, first movement:

Still, the suspicion arises that those surprising moments are not intended as simply
humorous, but also meant to represent the inept ‘fumblings’ of the fictive composer-persona
– one that apparently is not particularly ingenious nor adequately schooled in the rules of
composition. (62)

Byros (2013) uses Mozart’s K. 279 as an example of performer bungling, writing that Mozart’s

manipulation of common uses of galant phraseology and sonata form create a “parody of a musical

performance” with a “mindless or overenthused performer” (242). For Janet Levy (1992), humorous

music comes from a performer who acts in an inanimate manner. In Beethoven’s Bagatelle in C

Major, op. 3, no. 5, a virtuosic arpeggio repeats for the third time in a “seemingly mindless” fashion

and gives the “impression of something purely mechanical” (Levy 1992, 226). This prompts Levy

(1992) to ask herself: “Has the music gotten stuck? Or the pianist?” (226). The irony conditions

established earlier, however, also explains this Bagatelle example: Beethoven flouts the maxim of

Quantity; he gives too much information.

The tempo fluctuations of op. 131 could easily be heard as a type of performer bungling as

the string players seem lost and confused. In a similar way, the high amount of repetition in this

17 “Künstlich nachgemachte musiklaische Stümperey” (quoted in Mirka 2009, 300 n12).


187

movement could also be interpreted as a type of performer bungling; perhaps, as part of the

performers’ lost and confused nature. They lost track of where they were in the piece and ended up

repeating more than they originally intended. Performer bungling also falls in line with Lucariello’s

(1994) other requirement for situational irony: it exhibits “evocation of human frailty.” The ability

for performers and composers to make mistakes shows how “fragile” human nature can be. In the

end, this movement is often considered both humorous and ironic, due to unexpected modulations

and tempo fluctuations.

Op. 130/I (1825)

The final case study here analyzes the first movement of op. 130, the third and last of the

commissioned “Galitzin” quartets: op. 127, op. 132, and op. 130 (Watson 2010, 244). Unlike the

other pieces discussed here, previous scholars have not used the word “ironic” to describe this

movement. However, descriptors similar to irony have been used by some, as seen in this Kerman

1971 quote: “Paradox has to proceed from norms: suspiciously normal features jostle with abnormal

ones all through the Quartet in Bb” (307).

The basis for irony in this movement comes from transformations of sonata form principles.

In “normative” sonata form, the exposition, a tight-knit and stable section, should “propose the

initial tonic… move to and cadence in a secondary key” as well as provide a “referential arrangement

or layout of specialized themes and textures against which the events … – development and

recapitulation – are to be measured and understood” (Hepokoski and Darcy 2006, 16). The

development tends to be “more active, restless, or frequent tonal shifts—a sense of comparative

tonal instability” while the “thematic choice and arrangement is of paramount importance and

derives its significance through a comparison with what had happened in the exposition”
188

(Hepokoski and Darcy 2006, 18–19). The recapitulation, as a final step, “resolves the tonal tension

originally generated in the exposition by re-beginning on the tonic” and should bring ultimate

closure (Hepokoski and Darcy 2006, 19). To be a “functioning” sonata, each section should

accomplish its specific goal. Op. 130, on the other hand, seems to turn these roles “upside-down”:

both the exposition and recapitulation appear fragmentary, tonally unstable, and restless, while the

development appears static and serene. Beethoven does not simply create an odd sonata, but one in

which each section accomplishes the “opposite” of its expected goal. Why could we expect

Beethoven to follow sonata form, especially if we take Hepokoski and Darcy’s (2006) eighteenth-

century orientated perspective as a model? Beethoven is often considered closer to Mozart’s world

then Berlioz’s. In his essay “Exploiting Limits: Creation, Archetypes and Style Change,” Meyer

(2000) argues:

In my view, it is not surprising that Beethoven’s use [in op. 131] is more like Mozart’s than
like Berlioz’s—even though the last of Mozart’s works considered here proceeded the C#-
Minor Quartet by forty years while Beethoven’s precedes Berlioz’s work by only four years.
What most affects compositional choices in the internalization of prevalent musical
constraints, and such learning almost always takes place before a composer is twenty. In this
respect, Beethoven and Mozart, who were born fourteen years apart, were near
contemporaries, while Beethoven and Berlioz, who were born thirty-three years apart, were
not. Beethoven’s ingrained Classicism seems evident in the high-level organization. (Meyer
[1980] 2000, 208 n31, italics in original)

Therefore, Beethoven may have lived more in an eighteenth-century world even though he

composed in the nineteenth.

We may commence, for instance, with the fragmentary exposition. It begins adagio ma non

troppo, but is immediately followed by an allegro section. Such a beginning would not be odd; a

listener would simply interpret the adagio as a slow introduction. The allegro section ends abruptly,

however, on a I:HC (figure 4.18, mm. 15-20). This half cadence bridges the allegro to a return of the

adagio in the dominant key of F major. The adagio gently reasserts itself (figure 4.18, mm. 20-24),
189

which could be interpreted as the piece either returning to the introduction (perhaps to restart) or it

could prompt a listener to reconsider the beginning as actually a first Primary Theme instead of a

slow introduction. The allegro section follows suit once more (mm. 20-36), also in F major.

Figure 4.18: Beethoven, Op. 130, first movement, mm. 13-28


190

The adagio and allegro materials contrast each other on many levels. Their textures differ, since the

adagio draws from an aria cantabile topical style, while the allegro “paraphrases” a canzone with

shimmering sixteenth notes (Ratner 1995, 215). They differ the most, however, in tempo. Beethoven

grouping these themes together without transitions and within a short amount of time creates an

incoherent and confusing effect.

While this analysis argues for two Primary Themes, other scholars have different

perspectives. Hepokoski and Darcy (2006) write:

The opening (and recurring) Adagio ma non troppo, preparing a contrasting Allegro, is not
only repeated with the whole exposition but may also be understood simultaneously as both
an introduction – clearly its principal role – and the onset of a deformational P[rimary
Theme]…. [Op. 130] slow introduction; false first start with Allegro P-material, aborted;
return to slow introduction; restart or continuation of the Allegro. The impression is of two
or more attempts to launch a P-theme, only the last of which succeeds. In turn this suggests
either a need to go back to the “reflective” introduction to allow the faster theme to be
gestated more sufficiently or a momentary indecision or reluctance to face the task that is to
follow. (299–300)

Daniel Chua (1995), believing the work to be in “crisis,” follows a similar train of thought when he

writes that the piece contains two introductions, the first of which is “detachable” and the second

“integral” (205). Whether or not one hears the repeated adagio as a return of a slow introduction, a

separate theme or a combination of both, op. 130’s Primary Theme zone groups together material

that contradicts each other in significant ways.

One could argue, perhaps, that contrasting Primary Themes may not deter an exposition

from fulfilling its main goal of proposing the initial tonic and providing tonal stability. Yet, the

exposition of op. 130 fails on this front as well. The allegro material begins with a variant of the

Meyer schema (Gjerdingen 2007a, 111–21) called the Aprile (Gjerdingen 2007a, 122–28). In the

Meyer (figure 4.19), a soprano opens with 1-7 dyad and closes with a 4-3, while the bass opens with
191

a 1-2 and closes with a 7-1. For the Aprile variant (figure 4.20), a composer replaces the normal

soprano close of 4-3 with a 2-1.

Figure 4.19: The Meyer schema from Gjerdingen (2007a); the Aprile variant has a closing dyad of 2-1 instead of
4-3 in the soprano

Figure 4.20: Example of the Aprile variant (with a not as common 5-1 close in the bass instead of 7-1) from
Aprile, Solfeggio, MS fol. 40v, Larghetto, mm. 1-4 (graphic from Gjerdingen 2007a, 123).

The Meyer, and its Aprile variant, functions as a “tonally stable” schemata due to its I-V-V-I

harmonic motion, explaining “why it was a preferred choice for important themes” (Gjerdingen

2007a, 112). Beethoven uses an Aprile for the allegro Primary Theme; however, his use of this

schema is atypical (or deformative). This results in an opening theme that lacks any strong sense of

tonic (figure 4.21).


192

Figure 4.21: The Aprile schema in Beethoven, Op. 130, first movement, mm. 13-19

Meyer (1989) comments on Beethoven’s use of this Aprile, maintaining that the “general parallelism

of the parts… and the harmonic process… conform to the norms of the model [the schema]… But

there are striking differences, and because of these, the presence of the schema is much less

apparent” (229). One of the ways the schema is atypical is its missing 1-2 and 7(or 5)-1 dyads in the

bass. As Meyer mentions, these are not the exact notes; however, it still conforms to the harmonic

norms (I-V-V-I) of the schema. In addition to being an atypical version of the schema in terms of its

note choices, the schema is also masked. Meyer (1989) lists various ways the schema is masked,

including a “competing schema” of rising sequential fourths in addition to a greatly elaborated

melodic line (229), which particular masks the convention since it flies by the 1-7 and 2-1 notes. By

disguising this conventional schema associated with tonal stability, the allegro Primary Theme does

not accomplish its goal of establishing a sense of tonic.


193

The arrival and choice of the Secondary Theme key (Gb major) further emphasizes an

unhinged nature in this exposition. The piece modulates to its secondary key area by a unison

chromatic scale, beginning on F (the dominant of the tonic key) and rising to Db (ultimately acting

as a dominant to Gb major). This also signals a triply-obscured MC, hidden through an expanded

and modulating caesura-fill as well as an obscured acceptance (Richards 2013, 184). Richards (2013)

notes that an “implied home-key V in a unison texture sounds within TR at m. 51 but then

mysteriously rises in chromatic steps, reaching and sustaining Db at m. 53. Is this Db an unassuming

start to S, or is the MC-preparation, with what follows constituting single-voice caesura-fill, the start

to S instead occurring at m. 55?” (184) Meyer (1989) may label the lyrical Secondary Theme (figure

4.22, beginning m. 55) a gap-fill melody as it begins with a large interval leap up and filled in by a

subsequent descent.

Figure 4.22: Beethoven, Op. 130, first movement, mm. 47-58


194

Despite the Secondary Theme’s success at securing an EEC (mm. 89-90), the piece feels off-kilter

due to its odd manner establishing this unusual key. Kerman (1971) describes the modulation as one

might describe a confused character in a play, arguing:

This staccato unison chromatic scale is the most devastating event yet in the composition. For
nothing so far, not even the hinting at chromaticism, has prepared us for so mechanistic a
move to Db. When Db is coolly treated as the 5th of Gb major, the second key for the
movement, the tonal situation appears utterly precarious; normally Beethoven would never
dream of establishing a contrasting key-area, let alone a remote key (bVI), in so dissociated a
fashion. (309)

Daniel Chua (1995) also finds problems with the way Beethoven arrives at the Secondary Theme key

area: “Beethoven does not actually arrive Gb major, because there was never a departure towards it

in the first place; it is merely a contingent assertion” (206, italics in original).

The exposition, then, does not achieve its supposed goals, namely, to “propose the initial

tonic”—create an atmosphere of tonal stability—as well as move to and cadence in a secondary key

smoothly. Ultimately, the exposition should have a rhetorical role of tight-knit and stable tonality.

The recapitulation is not a different story. Beethoven’s recapitulation further emphasizes an unstable

atmosphere through fragmentary material and multiple modulations by fifths, traveling from Bb

major to Eb major, a “skirting Ab major” (Brodbeck and Platoff 1983, 159), and Db major before

finally returning (relieved, no doubt) to the Bb major tonic. Chua (1995) believes that the “tonal

ambiguity staged by the double opening has highly problematic repercussions in the recapitulation”

(205). The instability heard in op. 130 undermines the recapitulation’s usual goal of bringing closure

and providing resolution.

While the exposition and recapitulation highlight an unstable and fragmentary nature, the

development provides a vision of serenity and stability. From the beginning (m. 104), op. 130’s

development maintains a consistent ostinato pattern and piano dynamics with homophonic texture

that would be the envy of any stable Primary Theme (figure 4.23). The first violin re-contextualizes
195

the rising-fourths fanfare of the allegro. This fanfare, which originally sounded so flustered, seems

calm in this context. A separate lyrical melody begins with an ascending octave, perhaps alluding to

the gap-fill of the Secondary Theme. Such texture and dynamics highly contrasts what was heard in

the exposition.
196
197

Figure 4.23: Beethoven, Op. 130, first movement, mm. 104-132

The development sounds notably static compared to the “restlessness” that a listener might

expect from this section. Other scholars note the development’s unusual nature as well. David

Brodbeck and John Platoff (1983), for instance, talk about style expectations in relation to this

movement:

The tables are turned, as it were, since the development is relatively more stable than the
exposition. For this reason it sounds strange, at once dreamy and remote: our inner sense of
style tells us that the development should be more, not less, dynamic than the exposition.
(158)

Kerman (1971) also believes this development would appear strange to an experienced listener,

writing that this development is the:

Most eccentric Beethoven ever wrote, and doubtless the most disruptive contrast he ever
used in a sonata-allegro movement…. In the Bb Quartet the entire development section
exists in a trance, as though somehow another movement has got going without our quite
noticing how. (312)

Even though the word “irony” is not penned, Beethoven’s play with sonata form could imply a

paradoxical or ironic interpretation.

Beethoven violates expectations by turning the sonata movement “inside-out”: breaking

rules inherent to the form. A listener would need to be familiar with sonata form to hear the irony in

this piece (though the contrasting material in the beginning may strike even an inexperienced listener

as odd). Such knowledge of sonata form rules would have been part of the shared knowledge

between composer and eighteenth-/early nineteenth-century listeners that could then be retrieved.

Instead of looking just at “origin” of expectation, however, this movement serves a good example of

other categories in Margulis’ (2007) taxonomy. This movement’s “time course,” especially, differs

from the previous case studies.


198

The time course for op. 130’s expectations lasts over the entire form. Many of the other

ironic movements, on the other hand, target a particular moment that violates expectations. This

creates, more often than not, a particular moment that “feels” ironic. To track the irony of this

movement, a listener would need to pay attention to the sections of sonata form and realize what

expectations are being violated and when. Studies in cognitive science imply that listeners struggle

with understanding form as a whole and listeners instead attend to music on a moment-to-moment

basis (e.g. Levinson 1997; Clarke 1999; Tillmann and Bigand 2004). Clarke (1999) writes that the

perceptual present (brief memory store actively available) is short, and so: “it is not possible to have

any direct apprehension of form, but…a sense of form becomes available only through a

retrospective, and in some sense deliberate, act of (re)construction” (476). Yet, Beethoven’s striking

use of tempo, texture and dynamic contrasts makes points in op. 130’s form easier to detect

compared to others and even draw a listener’s attention while listening in the moment. For example,

PT and ST jump back and forth between adagio and allegro, which is a contrast that does not need

(re)construction to be recognized as unusual compared to most sonata forms. In a similar vein, the

development section’s overwhelming calmness in the middle of the piece would also strike a

knowledgeable listener as odd in the moment. He or she recognizes that at this moment in the form,

the music should be more and more tumultuous instead of a calming presence. Even though this

case study recognizes play with overall form, aspects that make it an ambiguous sonata form draws

listeners’ attention in the moment. At the same time, it is possible listeners need several listens of

this movement and even retrospection or (re)construction to perceive irony.

This movement, by playing with conventions of sonata form, has relations that mimic

flouting the maxim of Manner: avoid ambiguity and obscurity of expression. Beethoven can “utter”

an ambiguous form, similar to the way a person can utter an ambiguous sentence. Consider, for
199

example, the following sign outside of a men’s suit store: “These suits are so cheap, they won’t last

an hour!” This is an ambiguous sentence. Most likely the store means that the “suits are so cheap,

they will sell out within an hour.” However, a person could interpret the sentence as “these suits are

so cheap, that they would fall apart after wearing them for only an hour.” In this piece, the sonata

form mimics relations similar to an ambiguous statement: x, where x affords multiples interpretation

of the same utterance or is ambiguous, in the context of y. Sonata form normally has a rhetorical

trajectory of tight-knit-to-loose-knit-to-tight-knit, but instead here it is loose-knit-to-tight-knit-to-

loose-knit. This relation of RHETORICAL-TRAJECTORY can be re-represented as

AMBIGUOUS-IN(x, y) as seen in figure 4.24. The object needs to be different, but still similar

enough, to call into question whether it behaves in a “normative” fashion.

Figure 4.24: The analogy framework and Op. 130


200

A listener perceives enough similarities to hear this piece as a sonata, yet the rhetoric of its sections

simply do not behave like a sonata should. Such ambiguity may persuade a listener to perceive

Beethoven as flouting the maxim of Manner (avoid ambiguity and obscurity of expression). This

movement could also be mimicking relations that flout the maxim of Quality: do not say what you

do not have evidence for or what you do not believe to be true. Beethoven composes a movement

which does not follow the form. Despite his knowledge of “appropriate” sonata form, Beethoven

composes an “upside-down” sonata instead. Yet, an avid Beethoven listener may cry, “Does not he

actually follow the form well enough through the expected exposition, development and

recapitulation, including all important cadences?” In an “upside-down” sonata, the sections a listener

normally associates with a tight-knit or loose-knit are switched. A listener would expect the

exposition and recapitulation to be “tight-knit” sections, while the development “loose-knit.”

Instead, there is a “loose-knit” exposition and recapitulation with a “tight-knit” development.

Beethoven has turned upside down the expected rhetoric of the sections.

It is not just a matter of “obeying” or “not obeying” rules which makes this form

ambiguous; it seems as if the sonata form sections are deliberately switched around. There have been

many composers which “break” sonata form rules; but such “rule-breaking” does not always follow

with a sense of irony. Beethoven does not completely annihilate the form, but instead simply

modifies it enough so to evoke a sense of irony. He switches the rhetorical associations with the

sections, but still keeps the sections themselves (including the cadences) in check. Brodbeck and

Platoff (1983) note this curious sonata form play as well, claiming that:

In the Bb Quartet the composer has stayed within striking distance of the particular sonata-
form movement with slow introduction. His deviations from traditional procedures are thus
much more telling, for each unexpected event must be understood not only on its own
terms but in terms of the norms from which it diverges. (162)
201

Why draw attention to sonata form? An answer may be found if one looks at reasons

humans use irony to communicate in the first place. Considering the risks of miscommunication,

many psychologists wonder why humans are drawn to using irony. Raymond Gibbs Jr. and Jennifer

O’Brien (1991), when discussing various theories of irony, suggest “that the communicative purpose

of irony is to call attention to some idea or attitude that both speaker and listener can derogate”

(527). Beethoven, by using irony, may have attempted to critique sonata form’s expectations by

calling attention to what those expectations would be for each section.

Analyzing Listeners

The analogy framework, since it includes a retrieval step, may predict when or if a listener

hears a piece as ironic. By considering knowledge a listener retrieves during listening, a theorist could

predict if that listener makes a mapping that points toward irony. In regards to these case studies, I

use the retrieval step to briefly compare differences between an eighteenth-/early-nineteenth century

audience—the historical listeners considered here—and a modern audience. I have argued that a

historical audience, as well as modern music theorists, hear irony in these pieces; yet, how about

other modern listeners? To illustrate, I theoretically compare a historical audience and a modern

audience listening to op. 95 and op. 130.

In op. 95, a modern listener retrieves knowledge to perceive irony in this case study. Even if

he or she is not familiar with Western art music, this hypothetical listener most likely expects

different sections to relate to each other through experience with popular music (figure 4.25).
202

Figure 4.25: The analogy framework and Op. 95, a modern listener

Unless unexpected changes are built into a genre, then a modern audience still has expectations that

musical sections will relate to the rest of the piece. Yet, op. 130 may be a different story. This

movement draws on specific knowledge of sonata form (as opposed to a general understanding of

form).
203

Figure 4.26: The analogy framework and op. 130, an eighteenth-/early-nineteenth-century listener or a modern

theorist

A listener with knowledge of sonata form norms—such as an experienced eighteenth-/early-

nineteenth century listener or a music theorists—maps relations as discussed here, prompting him or

her to hear this piece as communicating irony (figure 4.26). A listener with no knowledge of sonata

form norms, on the other hand, will not retrieve knowledge necessary to make mappings and

perceive irony, at least in the way discussed here (figure 4.27).


204

Figure 4.27: The analogy framework and Op. 130, a modern listener

This listener could retrieve other knowledge to make a mapping and infer op. 130 as ironic. Yet, this

mapping would be based on different knowledge than sonata form transformations discussed here.

If theorists consider what knowledge a listener has, then theorists may be able to account for

different kinds of mapping and create different analyses based on different listener experience.

A number of composers outside of Beethoven’s world have reputations for using irony in

their music, and both the conditions and the analogy framework could be used to analyze this irony

as well. Shostakovich, for one, is notorious for his ironic play (Sheinberg 2000). Everett (2004; 2009)

writes on uses of irony, parody, and satire in music by more modern composers, such as György

Ligeti, Kurt Weill, Peter Maxwell Davies, and Louis Andriessen. Due to its inherent flexibility, these

irony conditions could be applied to other composers across time periods. The audience members

present for op. 95 may find its contrasting coda unexpected; yet, the audience members present for a

modern composition where such contrasts may be built into the genre may think nothing of it.
205

Musical expectations and degrees of flouting the Maxims may change depending upon cultural and

chronological context; however, the conditions would change with it.18 Therefore, these conditions

could be applied to composers outside of Beethoven and the types of shared knowledge discussed

here. As long as the theorist determines the appropriate shared knowledge for the piece and its

audience, what could be retrieved, the irony conditions could be applied to almost any time period.

My Own Coda

On comprehending humor in Beethoven’s music, Claudia Zenck (2008) writes:

How can later generations, with their different musical means and their different mentality
and wit identify the transgressions of norms…. with certainty as intentionally humorous and
not, for example, as daring ideas of a genius or, on the other hand, as sheer mistakes of a
composer possibly not of first rank? (74)

The same question can be asked for irony. The irony conditions and analogy framework may step a

bit closer to revealing a possible answer. The relevance to theorists of shared knowledge that could

be retrieved is not limited to issues only of irony and humor. In musical meaning in general, the

cultural context of both the piece and the listener makes an impact.

I have used the analogy framework to discuss when listeners may infer irony when listening

to music based on an analogy between music and language. Irony is only one possibility a listener

could have for making an analogy between music and language. Yet, to discuss how musical

behavior mimics linguistics behavior, irony is a good starting point for analysis. Returning to Dunbar

(2001), and hypothesis 2, listeners may be making distant analogies such as this one in order to make

sense of “unexpected findings” in music. In this case, the “unexpected findings” prompted theorists

18The possible extension to pieces with text may prove a bit trickier. For instance, a number of scholars have discussed
how Schubert and Schumann use irony as a device in lieder (Brauner 1981). The presence of lyrics may prompt a listener
to interpret irony differently in song than in instrumental music. Through text, listeners and composers have an
opportunity to exploit semantic meaning. Therefore, the conditions and analogy framework outlined here may not be as
effective a tool for analyzing music and text.
206

to use a distant or extra-musical analogy between music and language to make sense of these unusual

musical occurrences. The research here, by appropriating concepts from linguistics and cognitive

science, may illuminate how a listener reconciles inappropriateness in a musical work as evoking a

compositional effect such as irony as opposed to poor writing. If the inappropriateness in music

mirrors other inappropriateness in language, people may, in an effort to create coherence, infer a

musical piece as ironic communication due to an analogy between music and language. To return to

op. 95’s infamous coda, the ending which d’Indy confidently asserted could bear “no

interpretation,” we may simply smile to ourselves when hearing this “problem” in Beethoven. It

does not seem “chance” at all that many listeners’ arrive at similar interpretations.
207

CHAPTER 5
“Appears Clothed in a New Manner”: Analogy and Hearing
Thematic Variations

Variation forestalls the monotony, the triviality, in short, that void through which a melody
simply becomes trite, overused, like a street ditty. But if the basic theme, the main melody,
appears clothed in a new manner, under a delicate transparent cloak, so to speak, thus the soul of
the listener obtains pleasure, in that it can automatically look through the veil, finding the
known in the unknown, and can see it develop without effort. Variation demonstrates
freedom of fantasy in treatment of the subject, excites pleasant astonishment in recognizing
again in new forms the beauty, charm, or sublimity already known, attractively fusing the
new with the old without creating a fantastic mixture of heterogeneous figures…. [Variation]
also concerns the freedom of reflection of the listener who now knows how to grasp hold of
the main subject if he is given an inducement to hold onto it when it appears] in different
environments.
—C.F. Michaelis, “Über die musikalische Wiedeholung under Veränderung,”
Allgemeine musikalische Zeitung, 1803

In order to write variations, the composer need not be a great Melopoet, but should possess
all the more Phraseology. His main task is to invent new styles of playing [Spielarten], to adapt
new forms and new figures to the mode; he must successfully retain in the variations the
same analogy between harmony and harmony, between melody and melody that obtains in
the theme; in short, each character which he assigns to the first measure must be continued
throughout.
—Georg Joseph Vogler, Verberesserungen der Forkel’schen Veränderungen
über “God Save the King,” 1793

In the previous chapter, I addressed research question 2 by using the analogy framework to

describe how a listener may perceive irony in Beethoven string quartets. This chapter addresses

research question 1, which asks why participants in music cognition experiments do not categorize

themes based on pitch, harmony and meter/rhythm. I hypothesize that participants fail to categorize

themes based on these parameters since themes are relational categories. Participants, then, need to

use analogy in order to make the desired categorization. Participants often struggle to make

analogies in experimental settings. The best way to facilitate analogy-making is through the act of
208

comparison. Thus, these music cognition experiments, before the categorization task, should ask

participants to compare musical stimuli. Before leading my own empirical study, I first use theme

and variations movement by Beethoven, Mozart, and Daube as a real world laboratory to test my

theory of analogy.

In this chapter, I use the analogy framework to analyze three theme and variations

movements. I use it to make two arguments (A and B): (A) themes are relational categories. Thus,

listeners use the cognitive process of analogy to categorize themes. It is accepted that composers

varied a theme differently in the eighteenth- compared to the nineteenth-century. (B) The shift from

an ornamental eighteenth-century to a nineteenth-century style reflects a difference in how

composes altered a theme’s relations. In eighteenth-century variations, composers left relations

intact; therefore, these variations neatly fit a relational category (as in the Mozart and Daube

examples). In nineteenth-century variations, on the other hand, composers altered one or more

relations, making it harder to find a relational category that accounts for every variation (as in the

Beethoven example). A Romantic composer may alter one, two, or even more interconnected

relations to create a process-oriented variation that changes a theme at its very core. Due to altered

relations over several variations, individual variations of a same theme are more different from each

other in a Romantic style than those of a Classical style variation. Nineteenth-century variations

overall tend to be further from each other in similarity space than eighteenth-century variations.

Therefore, nineteenth-century variations are more difficult for a listener to perceive or categorize

since composers alter these relations.

A Variations Vignette

Beethoven composed variations differently than other composers, leading Sisman (1993) to

describe him as “the true innovator in variations” (235). It did not begin this way, as his first
209

published work, when he was twelve years old, set of variations WoO 63, was among others that

received “devastating contemporary notice” (Solomon 1977, 28). He later began composing

variations with a more original approach. On November 2, 1793, Beethoven wrote a letter (figure

5.1) to his childhood friend Eleonore von Breuning with a postscript that referred to his variations

on Mozart’s Se Vuol Ballare from Figaro, WoO 40:

The V[ariations] will be somewhat difficult to play, especially the shakes in the Coda…. I
never would have written anything of the kind, but I had already frequently noticed that
there was some one in V[ienna] who generally, when I have been improvising of an evening,
noted down next day many of my peculiarities in composing, and boasted about them. Now
as I foresaw that such things would soon appear in [print], I resolved to be beforehand with
them. And there was another reason for perplexing the pianists here, viz., many of them are
my deadly enemies, so I wished in this way to take vengeance on them, for I knew
beforehand that here and there the Variations would be put before them, and that these
gentlemen would come off badly. (Beethoven 1793, quoted in Shedlock 1972)

Figure 5.1: Autograph of Beethoven’s 1793 letter to Eleonore von Breuning (Beethoven 1793)
210

Beethoven acknowledged a shift in his composing of variations, especially when completing op. 34

and op. 35 in 1802. In a letter to Breitkopf and Haertel in Leipzig on October 18, 1802, Beethoven

mentioned these pieces:

I have composed two sets of Variations, one containing 8, the other 30---both are written in
an entirely new style, each in quite a different way. ….. Do not let this proposal be made to
you in vain, for I assure you that you will not regret taking these two works—each theme is
treated in a totally different manner. I only hear what other people say when I have new
ideas, for I never know it myself; but this time I must myself assure you, that the style of
both works is on a totally new plan of mine. (Beethoven 1802, quoted in Shedlock 1972, no.
37, 41)

Beethoven’s claim was justified by a May 1803 review in the Allgemeine musikalische Zeitung on op. 34:

“The variation are very beautiful and are handled in a particular manner that is also different from

this composer’s earlier variations” (quoted in Senner, Wallace, and Meredith 1999, italics in original)

How did composers treat a “theme” category when composing variations in the eighteenth- and

nineteenth-centuries? How does Beethoven begin composing variations in an “entirely new

manner’? How do listeners recognize or perceive this “new manner” of variations as members of

one category? These historically-oriented questions are similar to ones that current psychologists ask

today when they run music categorization experiments: how do listeners recognize or categorize

variations? Sisman (1993) notes that, “Just as variation ennobles repetition, then, variations is in turn

elevated by the yielding up of its secrets, gradually, so that the listener knows how to hear them,

knows how to find the familiar amid the strange” (237). Variation is about finding the familiar in the

strange, and analogy is a tool that listeners use for this task.

Thematic Variations as Relational Categories: Melodies Clothed in a New Manner

In the first chapter, I hypothesized that: “The discrepancy of listeners not categorizing

musical themes in laboratory settings is a consequence of listeners needing to use analogy, or

analogical processing, to categorize musical themes based on structural, as opposed to perceptual,


211

features.” Since participants do not encode relational structure, they do not categorize based on

thematic structure, but perceptual features (e.g. dynamics, timbre, texture, etc.) instead. Before

embarking on my own empirical enquiries (see chapter 6), I first establish a theoretical framework

rooted in music analysis for using analogy to analyze theme and variations.

Scholars in music studies (e.g. Dahlhaus 1991) and cognition (e.g. Margulis 2013) recognize

that a theme and its variations belong to a single category (not unlike how a lion, tiger, and bear (oh

my!) belong to a single category of “carnivore”), so categorization and its cognitive components

impacts how listeners hear themes and their variations. A variation of a theme does not create a new

category; instead, it is an exemplar of a same theme, a new elaboration of an old category. According

to C.F. Michaelis in “Über die musikalische Wiedeholung under Voränderung,” a variation is a same

melody, just clothed in a new manner: “But if the basic theme, the main melody, appears clothed in a

new manner, under a delicate transparent cloak, so to speak, thus the soul of the listener obtains

pleasure, in that it can automatically look through the veil, finding the known in the unknown and

can see it develop without effort” (Michaelis 1803, col. 200; translated by and quoted in Sisman

1993, 236, italics in Sisman). Scholars agree that thematic variations do not constitute each its own

category, but instead are members of one overarching category.

Those in music studies and cognition also acknowledge that listeners abstract this

overarching category encompassing a theme and its variations. Dahlhaus (1991) labels this

underlying structure of a theme’s melody an “abstract melodic framework.” He maintains that, “as a

rule, the abstract melodic framework provides the melodic substance of the variations, and it is to

some extent made concrete by the fact that it advances out of the background into the foreground,

and is not just contained in the melody but actually constitutes the melody itself” (Dahlhaus 1991,
212

157–158). In music cognition, Margulis (2013) writes that listeners abstract a thematic category after

hearing many exemplars.

Acquaintance with a body of varied instances can cause a listener to abstract a thematic
category that might not literally match any particular statement, but rather involve a set of
characteristics—for example: large leap in the melody, tremolo in the lower register, and
movement from major to minor—such that new passages could be accurately classified as
either belonging to the thematic family or not. (Margulis 2013, 177).

Variations are members of an abstracted category that listeners form. Snyder (2009) proposes that

thematic variants are categorized based on abstract prototypes, or “(generalized representations of

thematic material)” (Snyder 2009, 114). As the first exemplar heard, listeners may assume the theme

is the thematic category. Yet, as the piece continues, a listener may create a slightly different thematic

category.

Although I agree listeners create thematic categories, I disagree with how most scholars treat

characteristics often attributed to these categories. According to music scholars, listeners should not

categorize thematic variations based on surface features—such as figuration—or objects—such as

an individual pitch. C-B-D-C is the same theme as D-C#-E-D, simply transposed, and a waltz

version of C-B-D-C is the same theme as a march version of C-B-D-C. A list of characteristics that

determines themes and motives often includes pitch, interval, harmony and meter and rhythm.19

Some scholars treat these characteristics like perceptual features, as a static checklist of features a

category “has.” I argue composers use these musical parameters relationally instead, and so should

be treated as relations. Instead of surface features or objects, musical relationships determine

membership of a “theme” category. Though, it is often relations within certain musical parameters

19As an exception, Ivanovitch (2010) writes that a theme is not just a list of characteristics: “From an experiential
standpoint, that is, a theme is not a static arrangement of structural elements. Rather, it stands in a complex and
reciprocal relationship to the variations: it bequeaths to theme a set of expectations about how they might proceed, and
yet exists as a mutable collection of possibilities or potentialities to be activated and reshaped by the course of variations
themselves” (Ivanovitch 2010, 3).
213

that determine a thematic category. In his definition of “variations,” Hugo Riemann (1893) writes,

“As a rule, a variation transforms only one or a few elements of the theme: time, rhythm, harmony

or melody” (823). This quotation implies that mostly these parameters define a theme, so relations in

dynamics (e.g. crescendo) would not influence thematic categorization, at least not in late

eighteenth- and early nineteenth-century themes and variation.

Within these musical parameters, though, relations determine whether or not an exemplar

belongs to a thematic category, and not perceptual features. Quotations from numerous music

scholars provide additional evidence that humans use relationships between notes to categorize

themes. Gjerdingen (1992) claims that:

Because musical motives and themes undergo many transformations in a typical


composition, their core identities are often best defined in relative terms. For example, the
famous opening motive from Beethoven’s Fifth Symphony is more than just the exact
pitches G-4-G4-G4-Eb4. When later transformations are taken into account, the motive
emerges as an abstract figure: three relatively short iterations of a single pitch followed by a
leap down to a relatively long tone. (238)

When Schellenberg et al. (2014) discusses a listener’s ability to recognize “happy birthday” played

both very fast on a piccolo and very slow on a tuba, they write: “Your ability to imagine these

previously unheard versions demonstrates that a melody’s identity is based solely on relations between

consecutive tones in terms of pitch and duration” (84, italics in original). In addition, Margulis

(2013) insists that, “the overwhelming majority of people possess relative pitch, and experience the

intervals and relations between the notes as essential” (30, italics in original). Hanninen (2012)

considers a motive to not be just a set of features, but of relationships among segments (117). If

thematic categories depend on relations, then analogy would be the cognitive process used in

categorization.

If listeners categorize themes based on shared relations, instead of shared perceptual

features, then themes are more akin to relational categories then perceptual ones. As a review,
214

perceptual or feature-based categories are in force when category membership depends on a

collection of features. For example, birds have wings, beaks, fly, and so on. They are about the

features themselves in isolation (Goldwater, Markman, and Stilwell 2011). Relational categories exist

where category membership depends on relationships between entities or other categories. “Bridge”

is an object over which a person can get from point A to point B (Goldwater, Markman, and Stilwell

2011). Categorizing “themes” may be more similar to the process of categorizing “bridges” than

“birds.” Through comparison, listeners notice salient relations and, after enough comparisons, learn

a relational schema (Holyoak 2012, 236). How this schema type relates to galant schemata

(Gjerdingen 2007a) will be discussed at a later point. Since comparison means action from a listener,

perceiving variations through analogy involves active listening. That themes are relational categories

is significant when one realizes that music cognition experiments, alluded to in research question 1

and hypothesis 1, use methodology that treat themes as if they are perceptual categories. Vogler

writes: “He [the composer] must successfully retain in the variations the same analogy between

harmony and harmony, between melody and melody that obtains in the theme” (Vogler 1793,

quoted in Sisman 1993, 26).

A musical passage can simultaneously be a member of multiple categories. For instance, the

opening to Beethoven’s fifth symphony (example 5.2) is a member of a “thematic” category

(relational category), but also a member of a “loud musical excerpts” category (perceptual category).

Figure 5.2: Beethoven, Symphony No. 5, opening


215

A listener recognizes the musical passage is a member of both categories. In addition, he or she can

“flip” between categories, changing his or her experience of musical structure. Although the “music

itself” does not change, a listener’s perception of its category does. In listening for motivic

development, a listener’s attention shifts to its membership as a thematic category (relational

category). If the listener jumps a bit in his or her seat, his or her attention shifts to its membership as

a loud excerpt (perceptual category). This idea of attention shifting between perceptual features and

relations directly connects to similarity space as discussed in the second chapter (figure 5.3).

Figure 5.3: Visual representation of similarity space (graphic from Gentner and Smith 2013, 10)

A listener can attend to the same musical passage differently; mayhap attending to its perceptual

qualities before its relations, or vice versa. Neither way of attention is “wrong,” simply a different

way to perceive the music. However, humans more easily attend to perceptual features. Thus, the

challenge is often to prompt humans to notice relational structure.

In this chapter, I use the analogy framework (figure 5.4) to analyze how Beethoven, Mozart

and Daube varied themes since I consider themes to be relational categories. I use these movements

as a real world laboratory to test my hypothesis. Yet, how these composers vary relations within
216

their themes changes depending on their temporal context and mentality of variations. The way

composers alter a theme’s relations influences whether it is variation in an “old” or “new manner,”

and reflects changes in variation form from the late eighteenth- to early nineteenth-century.

Analogy Framework and Analyzing Variations

Following my assumption that themes are relational categories, a theorist can use the analogy

framework to analyze variations of a given theme as members of one category and make claims as to

how they are perceived.

Figure 5.4: The Analogy Framework

For these analyses, I focus on the mapping step rather than retrieval and evaluation. Within the

mapping step, listeners categorizing thematic variations make intra-opus analogies (since mappings

occur within one piece). The relations mapped are relations within a musical event. Finally, categorizing

thematic variations is an example of bidirectional listening (listeners compare two novel analogs and

understand both in terms of each other; both are sources and targets). Regarding the retrieval step,

though, listeners retrieve one exemplar from memory instead of retrieving one from long term

memory.
217

Mapping: Intra-opus Analogy

Listeners use intra-opus analogy to categorize variations since the multiple exemplars

compared to each other come from a single composition. Active comparison is necessary for

categorizing thematic variations, then, since listeners encounter two analogs “together” during the

course of a piece instead of retrieving an analog from long-term memory.

Mapping: Relations within a Musical Event

Thematic categories are defined by relations within a musical event (as opposed to between musical

events), which are relations restricted to a time course of a musical event. A listener, then, uses these

relations to assign musical events to a thematic category. Relations that define thematic categories

are often relationships within musical parameters outlined in the third chapter: rhythm/meter, scale

degree, contour, and harmony.

A thematic category, though, does not rely on a musical relation in isolation. An idea

reinforced by the fact that humans, according to the systematicity principle (Clement and Gentner

1991), often map over sets of relations connected by higher-order relations instead of individual

relations. In fact, a downfall of past thematic analyses is that scholars prefer certain relations over

others. When music theorist Rudolph Réti (1951), for example, discusses thematic processes, he

places more importance on intervallic properties and pitches. By focusing on pitches, Réti overlooks

the importance of meter and rhythm in identifying themes (cf. Eitan and Granot 2009). One benefit

to using the analogy framework, then, is that a theorist can not only consider different relations, but

also how relations interact with each other. Sometimes a relation is altered from one variation to the
218

next; yet, its interaction with other relations could be enough for listeners to still recognize that

musical passage as a thematic category.

I adopt the relations within a musical event box notation (chapter 3), using scale degree, contour,

harmony, and rhythm. For the Mozart analysis, I use a proportional approach for rhythm. For the

Beethoven analysis, I use a lengths approach for rhythm. I use this notation styles to analyze

relations for each variation. Then, I determine how closely a variation’s relations align with the

theme’s relations or another variation’s relations. In this way, I measure musical distance between

the different exemplars of the theme and variations; how similar or different they are from each

other. From comparing the many exemplars, I create an overall thematic category, which is the

relational category (or schema) that encompasses or accounts for the theme and all its variations

(figure 5.5).

Figure 5.5: A thematic category as a relational category (or schema) that accounts for theme and its variations

Meyer (1973) contends that:

The greater the amount of change – in both rate and degree – in one parameter, the smaller
must be the changes in other parameters if patterning is to be perceived. If all parameters are
varied simultaneously and independently of one another, the result is not necessarily a more
complex and interesting pattern, but often none at all, a confused hodgepodge of sounds.
The amount of simultaneous variation possible also depends upon the nature of the patterns
219

themselves: the more patently structured and archetypal one aspect of a pattern (for instance,
its melodic shape) the more other parameters (e.g., rhythm, harmony, etc.) can be varied
without destroying the impression of conformance. (54)

Thus, each exemplar should keep at least one relation from the thematic category to facilitate

perception.20 The clearest exemplar for a listener to perceive would have several of the connecting

relations. However, nineteenth-century composers play with this knowledge, making it harder to

perceive variations.

Mapping: Comparison and Bidirectional Listening

Comparison, according to psychologists, is important for making analogies (e.g. Markman

and Gentner 1993). Music scholars, as well, have noted that hearing variations depends on

comparing musical material. In the words of Ivanovitch (2010):

At root, the basic act of construing variation is a comparative one: the task of a listener is to
relate two stretches of music, to hear one passage “in terms of” another. To hear something
“as a variation” is perforce to be engaged in such an activity of comparison. (3)

In order to hear two musical events as members of one category, a listener compares musical events

to each other, or hears one passage “in terms of” another. Knowing a piece is a theme and variations

form may be more than enough to prompt a listener to make these comparisons. According to

Deliège (2007):

There are other occasions where a comparison process takes place automatically while
listening to music. This is particularly the case for forms that are based on a Theme and
Variation structure, where a basic statement is repeated a certain number of times, but always
worked out in different ways. (10, italics in original)

To hear a theme and variations, a listener does not hear a musical event in isolation, but compares it

to other, often previous, musical events retrieved from memory (though perhaps not from long-

20 This is especially the case since relational categories are mostly deterministic (Jung and Hummel 2014).
220

term memory). Broyles (1987) writes that the heart of variation form “is essentially oriented toward

the past. It is an aesthetic of retrospective….the principal interest of a variation structure grows as

the composer’s memory and imagination begin to exceed that of the listener’s, as the obvious is

exhausted and the composer begins to reveal facets of the theme that the listener had not until then

realized” (Broyles 1987, 89). Recognizing variations means not just remembering past instances, but

also comparing present to past (or present/past to future). Comparison is significant to hearing

thematic variations for at least two reasons.

First, comparison involves an active and participatory stance on the part of a listener.

Ivanovitch (2010) speaks of the “task of the listener,” denoting an action from the listener. He also

uses verbs “engage” and “activity,” showing that a listener is participating in the music. In the

previous paragraph, the Deliège (2007) quote uses words such as “process” and “takes place,”

illustrating a difference between a listener taking and not taking part in comparison. Comparison

involves active listening. One benefit of the analogy framework is that it strives to be a framework for

music analysis that incorporates listening as inherently participatory and action-oriented. A listener

needs to do something, take part, to act for the analogy framework to be effective. A listener needs to

make choices, whether conscious or unconscious. Using the analogy framework, a theorist should

build comparison into the analysis; it is a specific action a listener takes.

Second, comparison is essential to making analogies (e.g. Markman and Gentner 1993;

Gentner and Markman 1997). Humans do not align relational structures naturally. Yet, the act of

comparison often prompts humans to structurally align relations between two different situations.

Similar to these psychology experiments, music scholars have noted the necessity of comparison in

hearing musical variations. In fact, comparison may be the missing link in past experiments that ask

listeners to categorize musical themes. By incorporating comparison into analyses, theorists can use
221

a method of analysis that is inherently participatory/action-oriented on the part of the listener as

well as incorporate an action that has been empirically shown to influence the making of analogies.

Since listeners compare two analogs, they are more likely to engage in bidirectional listening

(mutual alignment analogy). In bidirectional listening, both variations a listener compared inform

and shape perception of the other. A listener learns new information about both analogs.

Bidirectional listening is partly facilitated by the act of comparison. In the case of thematic

variations, bidirectional listening serves several purposes. It influences how listeners perceive

similarity between different versions of a theme as well as solidifies variations as members of one

relational category. Also, it can help listeners reinterpret the theme or a previous variation

retroactively. After exposure to two exemplars (e.g. a theme and its first variation), a listener may

begin creating a relational category. After exposure to more exemplars, however, a listener may have

a different representation of the relational category. This newer relational category could be

retroactively applied to the first variation or even the theme to reinterpret it.

Variations from the Eighteenth to Nineteenth-Century: From Embellishing to


Reinterpreting

I analyze two movements of theme and variation form and an eighteenth-century

pedagogical example:

 Beethoven, Piano Sonata Op. 109/III (1820)

 Mozart, Violin Sonata in F Major, K 377/ii (1781)

 Daube, theme and variations example in the Musical Dilettante: A Treatise on Composition (1773)
222

Each of these pieces illustrate a different facet of theme and variations form. Johann Friedrich

Daube’s composition from 1773 taught others to compose an eighteenth-century variation on a

theme. It also showcases galant schemata in as well as bidirectional listening’s use in perceiving

thematic variations. Mozart’s violin sonata illustrates a standard theme and variations of the late

eighteenth-century, one of ornamentation. Beethoven’s piano sonata illustrates a “new type of

variation,” the nineteenth-century’s shift from variation as simple ornamentation or embellishment

to variation as reinterpretation.

Jeffrey Swinkin (2012) argues that composers of the nineteenth-century varied themes in

fundamentally different ways compared to the eighteenth. Classical mid-to-late eighteenth-century

variations reflect characteristics of embellishment and change of texture; the composer created “a

multitude of views of the same object” (Swinkin 2012, 37). The appeal of variations in the later

eighteenth century was also “embellishment, which added something fresh to the melody” (Ratner

1980, 255). According to Ratner (1980), variations of the eighteenth-century created “a bit of

musical theater by dressing a tune in different ways to make a series of colorful tableaux, much as

the model in a fashion show would appear in different costumes” (259).

A number of eighteenth-century sources affirm that listeners should recognize a melody

despite variation. In 1773, Daube wrote in The Musical Dilettante: “The art of variation is no small

assistance in embellishing the main melody if it is tastefully employed! No voice part can do without

it. Yet one must not carry this to excess either, so that the main melody does not become obscured”

(Daube [1773] 1992, 136). In Hoyle’s Dictionarium Musicae, the entry for “Variation” states “as in the

first part or straing you have the plain notes of the composition, and the next part is variated, that is,

the notes here and there are altered, and more notes made than in the first part, but yet it is to the

same signification; the second part, or as many parts as ever is made, are all different, and variously
223

put together” (Hoyle 1770, 107–108). According to Koch, in his third volume of Versuch einer

Anleitung zur Composition, the “so-called theme for variations… requires a somewhat simple but

flowing melody, capable of many variations, which is at the same time so constituted that it allows

sufficient, unforced changes of harmony. In the variations themselves, the main melody should

always be recognizable” (Koch 1793; quoted in Sisman 1993, 71-72). In his 1802 Musikalisches

Lexikon, Koch summarizes this earlier understanding:

Variazionen, Variazioni….multiple immediate repetitions of a short piece, in which the


melody each time is varied through different methods of subdividing its main notes, and the
passing and neighbor notes connected therewith, yet without completely mitigating the
similarity with the main melody…. So that such variations may have a good effect, one
must… (3) retain the similarity with the main melody in each variation, so that the attention
of the listener is thereby fixed; for as soon as the similarity ends, so does the interest of the
variations, and they give the impression of a group of arbitrarily related pieces [Sätze] which
have nothing in common with each other, and for whose existence and ordering [Daseyn und
Folge] one can imagine no basis. (Koch 1802; quoted in Sisman 1993, 72-73)

Many of these quotes have a rhetoric of “recognition.” Daube ([1773] 1992) cautions against a

melody becoming “obscured” and Koch (1793) stresses a main melody “always be recognizable”

and “retain the similarity with the main melody.” So, a composer intends a listener to recognize

thematic variations as related to the original melody or theme.

Swinkin (2012) contrasts Classical variations with those of the Romantic period, writing that

nineteenth-century composers reimagined their themes significantly more than eighteenth-century

composers. In the nineteenth-century, a theme is “not so much decorated as reinterpreted: its

harmonic and melodic constituents are exposed, then reconfigured” (Swinkin 2012, 37). Swinkin

(2012) explains this dichotomy of eighteenth and nineteenth-century variations from a theme’s

standpoint:

The implicit corollary of the belief that Classical variations are essentially decorative is that
the theme in a Classical set is an autonomous entity with fixed melodic and harmonic
components, susceptible to embellishment but not reinterpretation. In other words, anyone
who presupposed a set of variations to be primarily ornamental would probably also regard
224

the theme as self-contained, self-defined and directly given—an entity whose underlying
structural properties are neither laid bare nor fundamentally altered in the course of the
variations. By contrast, anyone who granted variations greater interpretative potency would
probably regard the theme not as an a priori entity but as something whose identity is
contingent upon the processes to which it is subjected….[nineteenth-century] variations
retroactively define what the theme in fact is. (37-38, emphasis added).

Swinkin (2012) uses “autonomous entity,” “contained,” and “self-defined” to describe eighteenth-

century perspectives of themes, but “interpretive potency” and “identity contingent upon processes”

to describe nineteenth-century perspectives. Decades earlier, Meyer (1973) observed a similar shift:

“the history of the theme and variations in the eighteenth and nineteenth centuries might be

understood as the search for a way of transforming a ‘naturally’ flat, additive hierarchy (as in most

Baroque and early classical variations) into an arched, processive one – one with functionally

differentiated parts” (10).

Eighteenth-century variations tend to “embellish,” while nineteenth-century variations tend

to “reinterpret.” In my analyses, I find that an eighteenth-century “embellishment” style variations,

such as Daube and Mozart, kept relations consistent. However, nineteenth-century

“reinterpretation” style variations, such as Beethoven, often did not keep as many relations

consistent. This shift from an “embellishment” to “reinterpretation” or “processive” style may be

because composers began altering relations of relational categories, creating variation exemplars that

either were not be members of the category—inhibiting perception—or atypical exemplars. To

analyze a difference between eighteenth and nineteenth-century variations, a theorist could use the

analogy framework and look at thematic relations. Assuming nineteenth-century variations alter a

theme’s relations, then there should be greater variance between exemplars of a nineteenth-century

theme and variation form then an eighteenth.21

21I make these claims, though fully realizing that I am generalizing based on one case study of nineteenth-century
variations (Beethoven) and two case studies of eighteenth-century variations (Mozart and Daube). Although I would
225

Relational Categories: Thematic Categories and Galant Schemata

Thematic categories tend to be relational categories, which cognitive psychologists

acknowledge are schemata. Holyoak (2012) uses “schema” to label a relational category a person

learns after mapping multiple exemplars (figure 5.6).

Figure 5.6: Major components of analogical reasoning (graphic from Holyoak 2012, 236)

As an example, he discusses how scientists use a “wave” analogy. After scientists used “wave” as an

analogy for transmission of sound, they extended it for transmission of light:

Ultimately it developed into an abstract schema or relational category….In the aftermath of


analogical reasoning about a pair of cases, some form of relational generalization may take
place yielding a more abstract schema for a category of situations (as in the case of the
evolving “wave” concepts), of which the source and target are both instances. (Holyoak
2012, 235, italics in original)

It is not surprising that galant voice-leading schemata (e.g. Gjerdingen 2007a; Byros 2012) are also

relational categories, though not identical to thematic categories.

have a stronger argument with more case studies, some scholars of case study research argue that one can generalize
based on only one case study (Flyvbjerg 2006).
226

Like most thematic categories, galant voice-leading schemata have scale degrees on strong or

weak beats creating a structure upon which a composer would elaborate. Gjerdingen’s (2007a)

description of one schema found in several of Locatelli’s opening basses from 1732 (figure 5.7)

mirrors my discussion of elaborated upon structural notes in variations.

Figure 5.7: Locatelli, Op. 2, various opening basses (graphic from Gjerdingen 2007a, 14)

Gjerdingen (2007a) writes:

Though taken from movements in four different keys and five different tempos, these basses
have obvious similarities. For example, on each bass I have marked a square on beat one, a
circle on beat three, and a square again on beat seven to show that they all share, at
analogous moments, an initial C, a move to A and then a return to C. At a smaller scale, I
have marked asterisk above the stepwise descent through the tones F-E-D-C. (13-14)

Thematic categories and galant voice-leading schemata could both be analyzed in one theme and

variation movement. In one case, a galant schema could be a basis for a more specific thematic

category. In another case, a thematic category could just be a galant schemata (as in the Daube

example below).
227

Although thematic categories and galant schemata are relational categories, they do differ in

a number of ways. Thematic categories are less abstract then galant schemata. By that, I mean that

thematic categories are “smaller” patterns that can account for less features, while galant schemata

are “larger” patterns that can account for more features. Thematic categories are more likely to have

more relations and more arguments than galant schemata. For instance, galant schemata often have

four objects in a relation (e.g. Meyer, Prinner, etc.), while a thematic category could have more

objects. As its label suggestions, galant voice-leading schemata are primarily defined by relations

between scale degrees. Although this influences harmony and contour, it is scale degrees and figured

bass that are given priority. Thematic categories, on the other hand, often involve interrelated

relations of meter/rhythm, scale degrees, and more.

The most apparent difference is that thematic categories are ad hoc, while galant schemata

are conventional. On the one hand, listeners build thematic categories bottom-up by listening to

multiple exemplars. Their use may not extend beyond listening to that piece. On the other hand,

listeners learn galant schemata through exposure and use in a cultural context, but they are

conventional and not ad hoc. Since these conventional patterns are “larger,” they share relations

with hundreds of other phrases from the eighteenth-century. Thus, thematic categories align better

with bidirectional listening—analogs inform each other—and galant schemata align better with

unidirectional listening—knowledge of conventional schema acts as source and informs target

analog.

Beethoven’s Piano Sonata, Op. 109/III (1820)

According to Swinkin (2012, 38), variations by middle Beethoven and onward fall in this

latter “reinterpretation” type instead of “embellishing.” This is further evidenced by Tovey’s (1945)

comment about op. 109: “The listener who wishes to understand Beethoven’s variations had better
228

begin at once by relieving his conscience of all responsibility for tracing the melody. Moreover, he

need not worry about his capacity to trace the harmony” (125). Beethoven’s interest in variation

followed him in his late period, as final movements for his piano sonatas op. 111 and op. 109 are

both variation forms (Marston 1989, 320-321).

Beethoven closes his op. 109 piano sonata with a “reinterpretation” or “processive” style

theme and variations. After a theme, the movement contains six variations followed by a series of

cadenza-style figurations (mm. 169-187), closing with a da capo of an unelaborated theme (m. 188).

Kinderman (2009) describes this movement’s trajectory as “two cycles of transformation” where

“the first five variations recast the theme and develop its structures and character in a variety of

expressive contexts, while the sixth initiates a new series of changes compressed into a single

continuous process that is guided by the logical unfolding of rhythm development” (245). Hatten

(1994) also mentions an “elevated return of the theme in the final section” as a moment of

transcendence (88). In theory, these variations should sound similar to each other. According to

Hanninen (2012), though, some variations in op. 109 sound closer to the theme than others:

The theme of Beethoven’s Piano Sonata No. 30 in E major, op. 109 begins with the eight-
bar melody….. Core features include a sequence of two major thirds aligned with strong-to-
weak metric positions (G#-E, D#-B in mm 1-2) and the dotted rhythm in m. 1….The
opening bars of variations I-VI each retain some of these core features, but reinterpret
others and introduce new features that distinguish the individual variations…. [There are]
differences in associative proximity (e.g., variation VI is most like the theme; variation IV,
least). (133, emphasis added)

In addition to acknowledging that core features of the theme are “reinterpreted,” she observes that

some—such as variation 6 (var. 6)—are more like the theme, while others—such as var. 4—are not.

This implies a “reinterpretation” style variation since variations differ enough from each other to

recognize a scale in this difference. A thematic category for op. 109 will have to be “larger” than

others to account for such distance in variations. It also implies that not all variations may neatly
229

match an overarching thematic category. With respect to perception, listeners may not easily

perceive these variations, especially on first listen, as members of the category. Tovey (1976) fulfills

this prediction, when he considers the following about op. 109:

The student and listener must not take a mistaken view of what a set of variations is
supposed to convey to the ear. If the variations are mere embroidery, then we may be
expected to trace the melody in them. But if the principle of the variation lies deeper, we are
intended to appreciate the depths in the same way as we appreciate other depths: we attend
to what reaches our sense, and we allow the sum of our experience to tell us more in its own
good time. (251)

Since this variation is more a nineteenth than an eighteenth-century style, it may be harder to find a

single thematic category that accounts for the theme and all its variations.

In analyzing op. 109, I focused on: (a) how Beethoven alters relations so that variations in

op.109 differ from each other, and (b) how the order of variations influences perception. Since

Beethoven alters relations, op. 109’s variations are often atypical exemplars of the thematic category.

In regards to order, the variation’s perceptual features can inhibit a listener from recognizing the

relational structure. Yet, the order of the variations can help to emphasize or de-emphasize relations.

First, I discuss the construction of the theme and what segment I analyze in depth. The theme

(figure 5.8) follows a simple binary. The first part closes (figure 5.8, m. 8) with a half cadence in the

tonic key of E major (I:HC) following a Gr+6, what Gjerdingen (2007a) calls an Aug. 6th cadence

(166).22 The second part closes (figure 5.8, m. 16) with an imperfect authentic cadence in E major (I:

IAC).

22For the sake of ease, I restart the count of measure numbers for each variation. Therefore, the first four measures of
variation 1 will be mm. 1-4 and the first four measures of variation 2 will also be mm. 1-4. As I am comparing
particularly segments of the sections to each other, this makes comparison easier.
230

Figure 5.8: Beethoven, Op. 109, third movement, theme, mm. 1-16

A listener hears a basic idea (b.i.) in mm. 1-2, and then a repetition of the basic idea (b.i.’) in mm. 3-

4. Together, these groups create a presentation phrase. In this study, I limit my analysis to these first

four measures (the Mozart analysis also is limited to its first four measures). These are the first

measures a listener hears when making comparisons between a theme and its variations, or a

listener’s first exposure to something that needs categorizing. Arguably, mm. 1-2 and mm. 3-4 are

two separate groups. Yet, I analyze them together as a listener may group them together as part of a

presentation phrase, not only because of their placement, but also because they are similar to each

other. In addition, the b.i. and b.i.’ are not clearly differentiated in each variation (e.g. var. 3, var. 5),

often times fusing together, further emphasizing a four measure thematic category.

For op. 109, it is difficult to create a thematic category that accounts for all variations since

they differ from each other. Yet, a listener creates an ad hoc thematic category (figure 5.9) by

listening to the elements of each exemplar. This thematic category would not be clear from the

beginning. Instead, a listener shapes and molds this ad hoc thematic category over the course of the
231

movement, most likely arriving at this representation only by the end. At a later point, I discuss how

order influences the creation of this thematic category in op. 109.

Figure 5.9: Op. 109 thematic category, mm. 1-4

For rhythm, I chose to use a lengths approach as the rhythms did not easily fit a proportions

approach. As a reminder, I have included the key for the lengths approach from the third chapter

(figure 5.10).
232

Figure 5.10: Key for notating “lengths approach” in rhythm

In regards to scale degrees, some re-representation is necessary to account for a greater number of

exemplars (yet, it still did not capture every exemplar). Although the first note is scale degree 3 (or

“mi”), it does occur as scale degree 5. Therefore, it is re-represented as a general tonic note or [ton].

The fourth note is the same case, so it is also rerepresented as [ton]. The last note appears mostly as

scale degree 5, but sometimes scale degree 7; therefore, it is represented as dominant or [dom].

Contour is visually represented in a way that emphasizes its b.i. + b.i.’ nature, although listeners fuse

these musical events together for the presentation phrase. Finally, harmony stays consistent

throughout the theme and variations, so it is not included in the box notation.

A Venn diagram represents similarity space and the distance between each exemplar of the

thematic category. The Venn diagram highlights which exemplars have which relations of the

thematic category, and which ones do not. The circles each represent a musical parameter whose

relations were analyzed (rhythm, scale degrees, harmony and contour). An exemplar is in a circle if it
233

shares that relation with the thematic category. Thus, exemplars clustered in the center where

rhythm, contour, scale degree and harmony intersect share all these same relations with the thematic

category. These exemplars sound the most similar and will be easy to perceive as the thematic

category. Exemplars in just one circle share only one relation with the thematic category. These

exemplars sound the least similar and will be difficult to perceive as the thematic category. In theory,

variations in an “embellishment” style should cluster more in the center, as composers did not alter

relations of musical parameters. On the other hand, variations in a “reinterpretation” style should

appear more in the fringes of the Venn diagram, as composers altered relations of musical

parameters.

In figure 5.11, the Venn diagram for op. 109 has two (or three) exemplars that have the same

relations as the thematic category (most notably, the theme and the last variation fall in this

intersection). It was ambiguous whether or not some exemplars shared the same relations as the

thematic category. I placed those in both intersections with question marks.


234

Figure 5.11: Venn diagram for Beethoven Op. 109

Many of the exemplars fall in the fringes of the Venn diagram. For instance, var. 3 and var. 4 share

only harmonic relations with the thematic category and var. 1 shares only harmonic and scale degree

relations. Var. 2 and var. 5 are ambiguous; yet, var. 5’s ambiguity is whether it shares only one or two

relation with the thematic category. Thus, var. 3 and var. 4 are the least similar to the thematic

category (or the least typical), while var. 6 (figure 5.12) is the most similar.
235

Figure 5.12: Beethoven, Op. 109, third movement, var. 6, mm. 1-4, score and box notation

According to this Venn diagram, Beethoven alters many relations over each variation, which is in

line with op. 109 fitting a “reinterpretation” variations style instead of an “embellishment” one.

Contour relations is an interesting parameter Beethoven alters, especially since contour is

easy for listeners to perceive. Beethoven alters contour relations to help mask these variations as

members of a same thematic category, which is part of “reinterpretation” style variations. The

theme’s contour (figure 5.13) first descends to scale degree 5, but then ascends to scale degree 5.
236

Figure 5.13: Beethoven, Op. 109, third movement, theme, mm. 1-4, score and box notation

In var. 1 (figure 5.14), Beethoven reimagines the theme as a waltz. The first half descends as

expected, but then the second half also descends. It more closely mirrors mm. 1-2 than the theme.
237

Figure 5.14: Beethoven, Op. 109, third movement, var. 1, mm. 1-4, score and box notation

Var. 3 (figure 5.15), though, has the most striking changes to contour. The melody, transferred to

the bass, has a contour of continuous descent. Kinderman (2009) writes that this variation “suggests

a two-voice Bachian invention in invertible counterpoint” (244).


238

Figure 5.15: Beethoven, Op. 109, third movement, var. 1, mm. 1-5, score and box notation

This change in contour does not fit the contour relations of the thematic category. Thus, it could

prohibit listeners from noticing the other relations, especially since contour is so salient. A contour

change such as this one could mask how similar the scale degrees are to the thematic category.

Overall, these alterations make the exemplar more difficult to perceive as a member of that thematic

category.
239

The order of variations exposes how Beethoven can de-emphasize and re-emphasize

relations, or can influence relations a listener may attend to as they compare the variations to each

other. Order of the variations, then, may influence how easily a listener categorizes the exemplars

since listeners often compare in the order presented to them. In theory, listeners more easily

compare adjacent variations (e.g. var. 1 and var. 2; var. 2 and var. 3), then ones more separate in time

(e.g. var. 1 and var. 5). Once again, this demonstrates that a listener’s perception or understanding of

a thematic category is not an immutable object, but instead mutates and changes. The “music itself”

may not change, but what a listener attends to and perceives does. In his or her comparisons, a

listener most likely attends first to perceptual features. Yet, how Beethoven orders the variations

may guide a listener’s attention to certain relations.

Imagine a listener beginning op. 109, the first natural comparison is between the theme and

first variation (figure 5.16). Focusing first on perceptual features, this listener may notice that var. 1

turns the theme into a waltz with characteristic metric figuration (Yaraman 2002; McKee 2012).

Figure 5.16: Beethoven, Op. 109, comparing theme and var. 1


240

When comparing these musical events to each other, a listener tries to highlight relations that could

be part of an ad hoc relational category, which is a category created in the moment to understand

the theme so far. Yet, little relational structure seems mapped over in this comparison. For instance,

the contour does not stay the same, as the theme’s b.i. has a downward contour and its b.i.’ an

upward one. In var. 1, on the other hand, both b.i. and b.i.’ share a downward contour. The scale

degrees remain consistent and can be mapped over with the thematic category’s rerepresentation.

Just from this first comparison, however, several relations are significantly altered, making it harder

for a listener to: (a) create a relational category for both these musical events, and (b) harder to

perceive these musical events as members of one category. However, since Beethoven does not alter

scale degree relations, this could draw a listener’s attention to this relation over the others.

Since var. 4 (figure 5.17) is least similar to the thematic category, its order in the variations is

important for perception. This variation uses “free imitative polyphony” (Tovey 1976, 253) that

“introduces four imitative voices” (Kinderman 2009, 244). Therefore, the additional figuration

further masks and hides otherwise perceivable relations.


241

Figure 5.17: Beethoven, Op. 109, third movement, var. 4, mm. 1-5, score and box notation

In comparing var. 4 to the original theme, the only relation shared is harmony. Thus, one’s gut

instinct may be that perceiving var. 4 as a member of the same category could be near impossible.
242

Yet, this variation occurs in the middle point of the entire movement. By this point in the listening

process, a listener may have already developed a “functional” relational category. The scale degree

relations, especially, have been upheld from the theme through var. 2. Since var. 4 appears in the

middle of the form, a listener may “hear” more of the thematic category in this variation then if it

had been var. 1.

As seen in figure 5.18, the order of variations as a whole creates a trajectory of “departure”

and “return.” The movement begins with variations that closely match the relations of the thematic

category (theme and var. 6). By the middle point of the movement, however, the only shared

relation is harmony (var. 3 and var. 4). Ambiguous exemplars that may or may not share relations

with the thematic category act as transitions from the closest exemplars to the most distant

exemplars. Hence, the movement “departs” from the thematic category, but “returns” to the

thematic category in the end. Overall, Beethoven’s movement reflects a nineteenth-century

“reinterpretation” style variations.


243
244

Figure 5.18: The order of variations in op. 109. The shaded yellow sections are relations shared with the thematic

category. The shaded red sections are ambiguous whether the relations are shared with the thematic category.

Mozart’s Violin Sonata, K. 377/ii (1781)

Before Beethoven composed op. 109 in 1820, Mozart composed K. 377 in 1781. There have

been several studies of variation in Mozart (e.g. Cavett-Dunsby 1985; Ivanovitch 2004). While

Beethoven’s theme and variations reflect a nineteenth-century style of variation, Mozart’s theme and

variations often reflect an eighteenth-century style. Nelson (1948), in his The Technique of Variation,

writes that Mozart’s variations are a part of “the ornamental variation,” which has a goal of “figural

decoration of the theme” (5). Sisman (1993) holds a similar position, arguing that, “a broad view of

Mozart’s variation themes takes in characteristics shared with most other variations of the period:

that is, they are usually two-reprise structures, borrowed from popular vocal or instrumental tunes

when used for independent sets, but are newly composed for variation movements” (Sisman 1993,

198). As in the Beethoven movement, I analyze both similarities between the theme and its

variations as well as how order of variations influences categorization.

Mozart’s theme also has a simple binary form (part 1: mm. 1-16; part 2: mm. 17-32),

although the theme lacks repeat signs. Instead, Mozart writes out the repeats of the thematic

material with small changes. The first hearing of the first “part” of the simple binary also sounds as

if it doubles as an introduction, since the violin has not yet made an entrance. Once the violin enters,

it plays the melody while the piano fades into accompaniment. Both repetitions have a period form,

with antecedents (figure 5.19, mm. 1-4, mm. 9-12) ending in i:HCs and consequents (figure 5.19,

mm. 5-8, mm. 13-16) ending in i:PACs.


245

Figure 5.19: Mozart, K. 377, second movement, theme, mm. 1-16

The second part of the small binary is also repeated, once again, written out with small changes. The

second part begins with a contrasting middle with a model-sequence technique (figure 5.20). In the

repetition, Mozart uses voice exchange, so that the piano plays the violin sequences and vice versa.
246

Figure 5.20: Mozart, K. 377, second movement, theme, mm. 17-32

The variations, in contrast to the theme, have repeat signs instead of written out repetitions. I limit

myself to analyzing relations of the first four measures of the theme and all its variations.

Figure 5.21: Mozart, K. 377, second movement, theme, mm. 1-4, grouping structure
247

In analyzing melody, especially, for this theme, grouping plays a significant part, as a listener, hearing

this entire theme as one musical event, would hear sub-events per measure for each of these

motives. The motives directly connect to one another; the first is a repetition of the third, and the

second is sequenced up a step (figure 5.21).

After analyzing the relations of musical parameters in these musical events (first four

measures of theme and all variations), it is apparent that Mozart has not altered many relations,

which aligns with an “embellishment” interpretation of variations. A listener can easily construct an

overarching thematic category that accounts for every variation (figure 5.22). Here, I use a

proportional approach notation for rhythm.

Figure 5.22: Mozart, K. 377, thematic category, box notation

The rhythm of the theme has an off-kilter feel with its slight syncopation. The scale degrees stay

consistent throughout the movement, with two examples of rerepresentation with the last two notes.

A listener can abstract enough information to recognize that the last note should fit in a dominant

chord (which is indicated in the graph by the color red and in brackets). This note is scale degree 5

in theme and all variations, except for var. 5 (the major variation) where it is scale degree 7. Similarly,
248

the second to last note is rerepresented as [tonic]. The contour often keeps the following ‘M’ shape,

where the first peak and the third peak are at the same height, while the second peak is higher.

Finally, harmony alternates between tonic and dominant function measures. I only include harmony

in the box notation if it differs from the expected I and V.

In the Venn diagram (figure 5.23), many of the exemplars cluster together in the center as

they share many relations with the thematic category, further emphasizing the “embellishment” style

variations. If it was more of a “reinterpretation” style, then the pattern would have aligned closer

with the previous Beethoven analysis.

Figure 5.23: Venn diagram for Mozart K. 377


249

The theme (figure 5.24) and var. 3 (figure 5.25) share the relations of all parameters with the

thematic category.

Figure 5.24: Mozart, K. 377, second movement, theme, box notation


250

Figure 5.25: Mozart, K. 377, second movement, var. 3, mm. 1-4, box notation and score

Most of the variations cluster in the intersection of contour, harmony and scale degrees. Scale

degrees (which intimately connects to harmony and contour), especially, is a musical parameter

whose relations Mozart keeps consistent throughout the movement. It was questionable as to

whether var. 5 (figure 5.26) kept the same scale degrees, as the major mode version of the theme.
251

Figure 5.26: Mozart, K. 377, second movement, var. 5, mm. 1-4, score and box notation

Yet, a listener would hear this melody as sharing the same scale degrees, as the relations stay mostly

the same; especially, if the relation is calculated as UP-A-THIRD rather than a specific interval.

According to the Venn diagram (figure 5.23), Mozart keeps the relation of musical rhythm

least consistent throughout the movement. Indeed, theme, var. 3, and var. 4 share the rhythmic

relations of the thematic category; however, the others do not. A reader may notice that the

rhythmic notation for var. 6 is “1” and “0.5 ~ 0.5.” In this variation (figure 5.27), Mozart develops

the theme as a siciliano, a peasant dance in 6/8 (Allanbrook 2008).


252

Figure 5.27: Mozart, K. 377, second movement, var. 6, mm. 1-4, score and box notation

To keep the proportional approach, I labeled the first beat as “1,” yet had difficulty notating the

dotted-eighthsixteentheighth note. Thus, I labeled it a “0.5 ~ 0.5” to imply it was somewhat

halving the original 1, just slightly off.

Although probably unconscious, the way Mozart alters rhythmic relations seems strategic in

helping listeners perceive exemplars as members of one category. When a person compares

situations, the relational structure shared is highlighted and made more apparent. A person, listening

in the moment (Levinson 1997), compares musical events to ones previously heard in the piece. He

or she most likely compares it to a relevant musical event recently heard. In the case of theme and

variations, a listener recognizes a simple overall form of theme followed by individual sections.

Then, he or she recognizes that parts of the theme will have analogous parts in the different

variations. In hearing the beginning of var. 3, then a listener most likely compares it to the beginning

of var. 2. In hearing the beginning of var. 4, then a listener most likely compares it to the beginning

of var. 3. Depending on his or her memory, a listener may even compare it to a theme. Hence,
253

Mozart’s strategic altering of rhythmic relations; in particular, in transition from var. 2 (figure 5.28)

to var. 3 to var. 4.

Figure 5.28: Mozart, K. 377, second movement, var. 2, mm. 1-4, box notation and score

Var. 2 (figure 5.28) shares all relations with the thematic category, except for rhythm of which it has

a “1-1-2” motif. This is labeled a “1-1-2” motif, as opposed to a “1-1-1-1” motif since scale degree 5

is an elaboration of scale degree 1. Var. 3 (figure 5.25), on the other hand, shares all relations with

the thematic category, although it does not sound similar due to changing perceptual features. For

instance, var. 3 features virtuosic acrobatic runs by the pianist reminiscent of the brilliant style topic
254

(Ratner 1980, 18). This perceptual feature of figuration is noticeably absent from the theme. This

further emphasizes that, although relations can be shared, perceptual features often are not, and this

does not influence thematic categorization. According to the analysis, var. 4 shares only rhythmic

and scale degree relations (figure 5.29).

Figure 5.29: Mozart, K. 377, comparing var. 3 and var. 4

How could Mozart facilitate an analogy where listeners recognize var. 4 as a member of the thematic

category, despite being the least like the thematic category? Var. 4, since it only shares two relations

with the thematic category, needs rhythm and scale degrees to be made salient to the listener.

Therefore, Mozart precedes var. 4 with var. 3, which, unlike vars. 1 and 2, shares the rhythmic

relation. In comparing var. 3 and var. 4 (figure 5.29), a listener may take note of this rhythmic and

scale degree relations, helping him or her to categorize var. 4 as a member of the thematic category.

Making this categorization, or noticing the needed relation, may have been more difficult had var. 4
255

been followed by var. 2, since then only one relation would have been shared. Listener assumptions

or experience influence attention (Chan Barrrett 2015) and, possibly, categorization as well.

Therefore, a listener, assuming he or she knows it is a theme and variations form, would be

predisposed to categorizing exemplars as members of one category since this is part of the overall

form. Yet, in the quotes above on Beethoven, listeners not only recognize that variations do not

sound like the theme (the melody cannot be “traced in them”), but also know that it is a variation.

Knowledge of form does not make perception automatically clear. Thus, it is significant that Mozart

precedes var. 4 with a variation that shares a rhythmic and scale degree relation. This example

illustrates how perceiving a musical passage is not just about recognizing musical objects, but

participatory on the part of the listener. If a listener had compared var. 4 to var. 2, instead, the

categorization may not have been as apparent, and a listener may have begun a new thematic

category. The comparison to what precedes it shapes a listener’s perception of that musical passage,

helping a listener focus on relational information as opposed to perceptual information.

The rhythmic relations could structure a trajectory for the theme and variations movement

as a whole. Regarding theme and variations form, Caplin (1998) writes that “the basic plan is simple:

a main theme, constructed as either a small ternary or a small binary, is followed by an indefinite

number of varied repetitions” (217). Yet, other repetitions or rotations could be noticed as part of

the classical form. Since rhythmic relations is the most varying aspect in this movement, it could be

used to structure sub-sections within the form as seen in figure 5.30. The most common rhythmic

motifs are “1-2-1” and “2-1-1.” There are seven exemplars total (theme + six variations). The theme

uses a “1-2-1” while var. 1 uses a “2-1-1.” At the middle point, var. 3, the “2-1-1” returns. Yet, var. 4

and 5 have the same pattern of “1-2-1” to a “2-1-1” (figure 5.30). Therefore, it hints on how to

subdivide the structure of the movement.


256
257

Figure 5.30: The order of variations in K. 377. The shaded yellow sections are relations shared with the thematic

category.

As the variations all cluster together in the center, this movement serves as an example of an

“embellishment” type variation. Since it is composed in the late eighteenth-century, it neatly fits into

expectations for the type. Therefore, listeners, in order to categorize these exemplars as members of

one category, use analogy to recognize different relations.

Daube’s Variations from The Musical Dilettante (1773)

When he taught amateurs to compose “embellishment” style variations in his 1773 treatise

The Musical Dilettante: A Treatise on Composition, Johann Friedrich Daube’s instructions read like a

“compose-by-numbers.” He provides a numbered list of figuration styles (what he calls “melodic

motives”) for his readers to adopt in composing variation, including examples without passing tones

(figure 5.31), with passing tones, and with appoggiaturas.


258

Figure 5.31: Variation of the first chord without pasting tones (graphic from Daube [1773] 1992)

He maintains that “these variations” can be “combined at will” (Daube [1773] 1992, 139), giving

almost step by step instructions. In his guidance, Daube ([1773] 1992) mentions that, “one would

select from the accompany examples one or two types of variation and, guided by each tone of the

melody and the appropriate chord, continue to write them until the end of the piece” (140). His

instructions encourage an “embellishment” style variation since he asks his readers to “apply” one or

more of the “variations” to a single tone or chord. Sometimes he holds that a composer uses an:

Entire chord for these motives, when actually a single tone was to have been
varied….Indeed, owing to their pure blending quality, the ear is let with the impression of a
single tone when all three of them are played simultaneously on the organ, harpsichord, or
three well-tuned wind instruments. (Daube [1773] 1992, 134)
259

As part of his pedagogical initiative, Daube takes one theme (figure 5.32) and varies it while

explaining his process. Thus, this theme and variations example, like the Mozart one above, serves as

an additional example of an “embellishment” style variations.

Figure 5.32: Daube’s theme (graphic from Daube [1773] 1992, 139-140)

Unlike the Beethoven and Mozart examples above, a thematic category for Daube’s

composition is not ad hoc, but a conventional relation category instead. In particular, Daube

combines galant schemata (Gjerdingen 2007a) to compose his theme and subsequent variations.

Considering Daube’s knowledge of the thorough-bass tradition, it is not surprising that he takes

advantage of these patterns and emphasizes the bass’ importance in his instructions: “The first

concern in a piece which is to be varied… is to make certain that the bass is well composed. Then

one should carefully observe to which chords the tones of the upper voice belong in order to

distinguish these from the passing tones” (Daube [1773] 1992, 140). In writing on variations in

Verbesserungen der Forkel’schen Veränderungen über “God Save the King”, Vogler might have alluded to
260

these conventional schemata when he wrote “in order to write variations, the composer need not be

a great Melopoet, but should possess all the more Phraseology” (Vogler 1793, quoted in Sisman

1993, 26). “Phraseology” could refer to this repertoire of galant schemata, which saturates Daube’s

theme and variations.

Daube’s theme could teach schemata organization in addition to variations. I will notate

schemata in the score, first, before returning to my relations within a musical event box notation. The

theme is a rounded binary form, with a standing-on-the-dominant contrasting middle. I focus on the

first part of the rounded binary.

Figure 5.33: Daube’s theme, mm. 1-10 (graphic from Daube [1773] 1992, 139)

In mm. 1-2 of figure 5.33, Daube opens with a Do-Re-Mi schema (figure 5.34). In this schema, a

“favored opening gambit,” the first and last stages are stable tonic chords while the middle is a scale

degree 7 bass with figured bass of 6/3 or 6/5/3 (Gjerdingen 2007a, 77-78).
261

Figure 5.34: Do-Re-Mi schema (graphic from Gjerdingen 2007a, 650)

The phrase closes with a Prinner schema (figure 5.33, mm. 3-4), where the bass descends 4-3-2-1

and the soprano 6-5-4-3 (figure 5.35, Gjerdingen 2007a, 45–60). If the Do-Re-Mi is a favored

opening gambit, then the Prinner is a favored riposte.

Figure 5.35: Prinner schema (graphic from Gjerdingen 2007a, 646)

In mm. 6-8 of figure 5.33, Daube repeats the Do-Re-Mi schema, but this time closes with a I:HC in

G major. At first, the phrase seems to be heading toward a Cudworth Cadence. A Cudworth

Cadence (figure 5.36) is cadential pattern where a standard bass is in conjunction with a melodic

descent that spans a full octave, from a high 1 to a low 1 (Gjerdingen 2007a, 146).
262

Figure 5.36: Cudworth’s cadence galante (graphic from Gjerdingen 2007a, 147)

The 8-7-6-5-4-3-2 descent in the soprano implies the beginning of a Cudworth Cadence; however,

the #4 in the bass turns the cadence into a converging one. The converging cadence, so named

because the “two outer voices move toward each other” to converge on the dominant, has a distinct

ascending bassline of 4-#4-5 (Gjerdingen 2007a, 159–160). As in the previous examples, I limit my

analysis to examining the first four measures of the theme and subsequent variations in detail.

Essentially, I further explore how Daube realizes the Do-Re-Mi in his variations.

In returning to my relations within a musical event box notation, I did not need to create a

separate figure for each variation, unlike the Beethoven and Mozart examples. I only needed to

create one figure to represent the first four measures of the theme and all variations since Daube

does not alter any relations. Figure 5.37, then, is a different way to represent the Do-Re-Mi schema.
263

Figure 5.37: Daube, thematic category, box notation

Daube does not alter any relations, so no Venn diagram is needed. All exemplars would be in the

middle overlap of rhythm, scale degree, contour and harmony. The exemplars clustered together in

the middle suggests that Daube composes with an “embellishing” style variation. This thematic

category, also a schema, is simple compared to those found in the Beethoven and Mozart examples.

Only three objects create the web of relations, with an easy to perceive contour and conventional I-

V-I harmony. Thus, a listener may have an easier time perceiving this thematic category (or schema)

compared to others discussed here. Yet, not all exemplars may be as easy at first to hear as this

schema, depending on whether or not listeners recognize structural notes.

Daube’s example demonstrates how bidirectional listening (chapter 3; understanding both

analogs in terms of each other, both analogs inform each other) helps listeners hear, in retrospect,

the structural notes of a musical passage. One might assume that a theme and variations form

composed of galant schemata would have the clearest example of the schema in the theme. On first

blush, someone listening to Daube’s theme could miss scale degree 3 in m. 2 as the structural note,
264

instead prioritizing scale degree 5 (figure 5.38). Scale degree 3 only lasts a sixteenth note, while scale

degree 5 lasts at least an eighth note.

Figure 5.38: Daube, theme, mm. 1-5 (graphic from Daube [1773] 1992, 139)

Before recognizing these as a conventional musical pattern, a listener may begin forming a relational

category similar to figure 5.39.

Figure 5.39: Daube, potential thematic category after hearing only the theme

Hearing “sol” as a structural note could hinder a listener from recognizing the measures as a Do-Re-

Mi. Thus, mm. 1-2 of the theme are an ambiguous example of the Do-Re-Mi schema. Typically,

listeners expect the theme of a theme and variations to be the clearest example of the relational
265

category, especially in “embellishing” variations. In this example, however, it is the opposite. The

thematic category becomes clearer only as more variations are heard.

With bidirectional listening, a person uses knowledge about a future musical event to

retroactively re-shape his or her initial perception of a theme. To fully recognize mm. 1-2 of the

theme as a Do-Re-Mi, listeners need to make an analogy by comparing this musical event to the

variations. In mm. 1-2 of var. 1 (figure 5.40), the Do-Re-Mi is a more traditional realization. In the

original theme, scale degree 3 flew by in flash. Instead, var. 1 places scale degree 3 on the downbeat.

The figuration of scale degree 3 is reminiscent of scale degrees 1 and 2, which further emphasizes

that scale degree 3 is the structural note. Yet, var. 1 still has some emphasis on scale degree 5.

Figure 5.40: Daube, var. 1, mm. 1-10 (graphic from Daube [1773] 1992, 140)

Var. 2 (figure 5.41) solidifies the Do-Re-Mi schema by emphasizing these scale degrees only. Scale

degree 5 has been downgraded to a pedal point.


266

Figure 5.41: Daube, var. 2, mm. 1-8 (graphic from Daube [1773] 1992, 141)

Here, in var. 2, a listener hears the first clear realization of the Do-Re-Mi. If a listener compares this

musical event to the theme or var. 1, the Do-Re-Mis in these exemplars retroactively become clearer.

Through bidirectional listening, the opening measures’ ambiguity is resolved. The Do-Re-Mi is clear

in future variations (figure 5.42), as well, with the exception of some ambiguity rearing its head again

in var. 7 (figure 5.42e). Although, by that point, a listener has a clear understanding of Do-Re-Mi as

thematic category.

(a)

(b)

(c)

(d)
267

(e)

Figure 5.42: Daube, beginnings of (a) var. 3, (b) var. 4 (c) var. 5, (d) var. 6, (e) var. 7 (graphic from Daube

[1773] 1992, 141-142)

It may have been ambiguous whether scale degree 3 or 5 was a structural note in Daube’s

theme, although listeners expect clear examples of schemata in a theme. Through comparison, a

listener retroactively recognizes this passage as a Do-Re-Mi even if he or she did not categorize it

that way originally. Thus, conventional patterns are not always clear cut. Comparison and analogy,

then, can benefit a listener categorizing conventional patterns as well as ad hoc ones. After the

structural notes are clear to a listener, Daube’s theme and variations is an example of an

“embellishment” style variation since none of the relations are altered.

Conclusion

In theory, the analogy framework could be used to make several arguments in regards to

themes and variations. For instance, a theorist could use the analogy framework to argue when

listeners may be more or less likely to perceive a musical event as a variation or a new thematic

category. In this chapter, I addressed research question 1 by analyzing three theme and variation

movements by Beethoven, Mozart and Daube. I demonstrated that themes are relational categories,

which implies that analogy is used to perceive them. I also discussed how composers in different

time periods approached the theme and variations form, and how the analogy framework can be
268

used to analyze distance between exemplars of a thematic category. In general, eighteenth-century

“embellishment” style variations share more relations with the thematic category then nineteenth-

century “reinterpretation” style variations. An alternative interpretation is that eighteenth-century

structures were schematized, and so easily memorable for music insiders. Nineteenth-century

composers, on the other hand, could only rely on perceived similarity of successive features due to

their musically “unknowledgeable” audience. However, since schemata are relational categories, this

interpretation still implies that eighteenth-century variations share relations, while nineteenth-century

variations do so to a lesser degree, or not at all. As I have now analyzed these pieces, I plan to use

them (especially the Daube) as stimuli in future empirical work that addresses how listeners use

analogy to categorize themes (see chapter 6).


269

CHAPTER 6
“You’re sure to get somewhere, if you only walk long enough”:
Future Research

[Alice] went on. “Would you tell me, please, which way I ought to go from here?”
“That depends a good deal on where you want to get to,” said the Cat.
“I don’t much care where— ”said Alice.
“Then it doesn’t matter which way you go,” said the Cat.
“—so long as I get somewhere,” Alice added as an explanation.
“Oh, you’re sure to do that,” said the Cat, “if you only walk long enough.”
—Lewis Carroll, Alice in Wonderland and Through the Looking-Glass and
What Alice Found There

Dissertation Summary

My research studies how listeners use past experiences to make sense of unexpected or

unusual musical events and categorize musical patterns. I have presented an interdisciplinary

theoretical framework for analogy in listening—retrieval, mapping, and evaluation—based on

Structure-Mapping Theory (Gentner 1983) in cognitive psychology. I have argued that theorists

should consider the context in which some analytic or listening activity takes place. To account for

different contexts, listener or otherwise, theorists can use the retrieval step of the analogy

framework. I have also argued that theorists can use the mapping step of the analogy framework to

specify how listeners actively make mappings between musical events. Finally, I argued that theorists

specify what outcomes, interpretations or understandings occur due to these mappings. The

evaluation step gives a means for discussing inferences music theorists or listeners make. In this

dissertation, I have also used the analogy framework to help explain how theorists perceive musical

irony and categorize thematic variations.


270

In examining research question 1, it was suggested that participants do not categorize

musical themes in experimental settings since experiments do not facilitate analogy-making. Analogy

is used to categorize musical themes; themes are relational categories and not perceptual categories

(even though experimenters often treat them as perceptual categories in experiments). To

demonstrate that a theme is a relational category, I used the analogy framework to analyze theme

and variation movements by Beethoven, Mozart, and Daube, using these as a real world laboratory. I

found that composers of eighteenth-century “embellishment” style variations kept relations

consistent while perceptual features changed. However, composers of nineteenth-century

“reinterpretation” style variations did not keep relations as consistent as their eighteenth-century

counterparts, creating less clear relational categories. Consistent with my argument that people use

analogy to categorize thematic variations, music scholars find a difference between these variation

styles. When composers alter relations, theorists recognize this as variation in a “new style,” and

themes and their variations are not easily heard as members of one category since altering relations

blur relational category. However, altering perceptual features does not seem to damage music

theorists’ understanding of thematic categories. Since comparison is important for analogy making, I

incorporate comparison into analyses to help determine salient relations. Now that I have

established a theoretical argument for themes as relational categories, I can pursue empirical research

(see below).

When a listener hears an odd moment that cannot be mapped onto previous experiences

with music, he or she may wonder if a similar relation occurs in other domains. If so, the brain,

working as economically as possible, may use this domain to inform their interpretation of music.

According to research question 2, music theorists use non-musical analogy to make sense of

“unexpected findings” in music; in this case, music theorists use an analogy between music and
271

language to hear “unexpected” musical events as ironic since irony is easier to understand in

language than music. Drawing from empirical research in psycholinguistics on ironic language, I use

the analogy framework as well as sonata theory, formal function, and schema theory to analyze how

music theorists use analogy to perceive certain Beethoven string quartets as ironic. If theorists

encounter “unexpected” musical events, then he or she could hear this musical behavior as

analogous to linguistic behavior. The violation of expectation and flouting of Gricean maxims are

relations in ironic language that could be mapped onto relations in music, prompting a music

theorist to use an analogy to hear a musical moment as ironic.

Future Research

I plan to use the analogy framework in several future projects: an experiment, analysis that

investigates the connection between schemata and topics, and an analysis that investigates how

modern listeners hear common practice music through film music.

An Empirical Investigation into Analogy for Categorizing Themes

I plan to empirically investigate analogy as a basis for musical categorization by running an

experiment, adapting experimental designs from other analogical processing experiments (Gentner

and Namy 1999; Christie and Gentner 2010). My experimental design incorporates the following

elements from past experiments: 1) hearing multiple exemplars of a musical category, used to

facilitate analogical processing in both children (e,g, Gentner and Namy 1999) and adults (e.g. Elio

and Anderson 1984; Medin and Ross 1989), and 2) comparison between these multiple exemplars,

necessary for structural alignment (Gentner and Namy 1999; Christie and Gentner 2010; Markman

and Gentner 1993).


272

Stimuli

Stimuli for an experiment on musical analogy differs from stimuli in experiments by Gentner

and colleagues. Often, Gentner and colleagues use static pictures where participants can visually see

a relation (e.g. picture of woman giving food to squirrel; Gentner and Markman 1997). However,

this experiment would ask listeners to experience auditory stimuli. Therefore, participants may need

more exposure to the stimuli to hear the relations.

In this study, participants will listen to approximately sixteen musical excerpts. Eight of these

musical excerpts are variations of Theme A while eight of these musical excerpts are variations of

Theme B. The theme category (A or B) is a relational category. The variations differ in musical

figuration. The figuration category is a perceptual category.

I use Daube’s variations from his 1773 The Musical Dilettante: a Treatise on Composition as

stimuli. Variations were chosen as stimuli for several reasons. First, variations include a core

thematic idea that is altered on the surface. In the fifth chapter, I demonstrated that thematic

categories are relational categories using analysis, implying listeners use analogy to categorize

variations as members of one thematic category. Second, composers of theme and variations form

assume that listeners compare different variations to the theme. Even though participants in this

experiment will not know the excerpts they are hearing are from a theme and variation form, they

would still be comparing musical material as a composer might assume they would. Third, a

thematic category is an ideal example of a relational category, while musical figuration is an ideal

example of a perceptual category.

Experimental Design, Hypotheses and Procedure

I will use a 2x2 between-subjects design with categorization choice (perceptual or relational)

as the dependent variable and musical training (musician v. non-musician) and a comparison
273

manipulation (no-compare v. compare) as the independent variables (see figure 6.2 for schematic of

experimental design). The no-compare condition acts as a control, expecting results similar to past

experiments on musical categorization (e.g. Ziv and Eitan 2007; Eitan and Granot 2009; Pollard-

Gott 1983; Lamont and Dibben 2001). The following are my hypotheses:

H0: Listeners will pick the perceptual choice (based on surface-level features)

H1: Listeners will pick the structural/relational choice (based on deep-level features)

The design contains two parts: a categorization task phase (force-choice) and a rating phase.

For the categorization task, all participants will complete sixteen trials. Each trial will involve a

categorization judgment. Participants in the no-compare condition will be given one exemplar

(either a variation of Theme A or a variation of Theme B) and will be asked to choose which one of

two choices is the same category as the exemplar. One choice will be a perceptual match (has the

same surface features of the exemplar category but not the same relational structure) while the other

choice will be a relational match (has different surface features but the same structure). For

participants in the compare condition, they will hear two exemplars of a category (either variations

of Theme A or variations of Theme B) and will choose which one of two choices is the same

category as the exemplar. Before making their choice, however, participants in this condition will

actively compare the two exemplars. In theory, the active comparison between multiple exemplars

should facilitate structural alignment and prompt a listener to use analogical processing to categorize

based on relational structure instead of perceptual features. Eight of the total sixteen trials will use

exemplar(s) from Theme A and eight from Theme B. I will also gather response times for each trial.
274

Participants will then move on to phase three: ratings. In this phase, participants hear every

excerpt (of both categories) individually. They provide a series of adjectives ratings on bipolar scales

(figure 6.1; as used in both Pollard-Gott 1983 and Lamont and Dibben 2001).

Figure 6.1: Adjective pairs for excerpt descriptions (graphic from Lamont and Dibben 2001, 253)

A two-way (2x2 factorial design) analysis of variance (ANOVA) will be conducted with

categorization choice serving as the dependent variable and musical training condition (musician v.

non-musician) and comparison condition (no-compare v. compare) serving as the independent

variables.
275

Figure 6.2: Schematic of experimental design


276

Predictions

I predict that participants in the compare condition will choose relational matches

significantly more than participants in the no-compare condition. If participants in the compare

condition categorize based on relations (creating a statistically significant difference between the

compare and no-compare condition), then this implies that analogical processing plays a role in

perceiving and categorizing musical themes. Listeners, then, may be using analogical processing to

perceive musical categories based on relational structure.

Musician v. Non-musician: Previous experiments (e.g. Lamont and Dibben 2001; Ziv and

Eitan 2007) found no difference between musicians and non-musicians in their results, indicating

that musicians often categorize based on perceptual features similar to non-musicians. If this study

finds a difference between musicians and non-musicians, this may indicate the role of musical

training in categorizing thematic categories. Musical training—where humans explicitly learn labels

for categories implicitly learned—should make recognizing relational structure easier. Experts have

more relational knowledge, and so are more likely to categorize based on relational structure while

novices are more likely to categorize based on perceptual features (e.g. Chi, Feltovich, and Glaser

1981; Shafto and Coley 2003; Rottman, Gentner, and Goldwater 2012). Assuming, however, that I

find no differences between musicians and non-musicians, the results of the no-compare condition

should match results of previously experiments (that of participants categorizing based on

perceptual matches).

Time Course: In past studies on similarity, participants that responded quickly (under 700-

1000 ms) based similarity on local or perceptual matches instead of relational matches (Goldstone

and Medin 1994). Eitan and Granot (2009) recorded time spent by their participants in different

types of classification. The fastest participants classified musical excerpts using secondary parameters
277

as criterion and the slowest participants used primary parameters as criterion (165). I predict that

participants making relational matches will have significantly longer time course than those making

perceptual matches. If participants’ response times follow this prediction, then these results provide

further evidence that primary parameters are relational structures and participants could be using

Structure-Mapping Theory for categorization.

Follow-Up Studies

A scholar may argue that this experiment lacks ecological validity. A person does not listen

to music in a series of excerpts; instead, he or she listens to the piece as a whole. I plan to use the

results of this study as a springboard to two follow-up studies which examine more naturalistic

approaches to tracking melodic categorization as a piece develops. One follow-up experiment would

address the potential problems of having one condition with multiple exemplars and the other

condition with only one exemplar. The other follow-up experiment would put these musical

categories within a more natural listening context.

Connections between Schemata and Topics: Relational Categories, Perceptual Categories,

and Construction Grammar

If analogy is a cognitive process used in music perception, then one assumption is the

distinction between relational and perceptual features in music. In this dissertation, I have defined a

musical relation and discussed how it forms a relational category in music. Yet, are relational and

perceptual features already featured in current theories of pattern perception in music theory? How

would the analogy framework work with current theories to create analyses? This project investigates

a “perceptual-relational divide” in music by analyzing whether conventional musical patterns such as

topics and schemata are relational categories or perceptual categories.


278

Topics and schemata as mental representations may be “treated” as similar in cognition:

conventional patterns with associations. Therefore, a superordinate category may subsume both

these patterns: that of a “construction.” Construction Grammar in linguistics argues that language is

built from “constructions,” patterns based on frequency and use where form is conventionally

paired with meaning or function (e.g. Tomasello 2003; Goldberg 2006; Bybee 2010). If the human

mind creates these form-meaning pairings for linguistic categories, it may also do the same for

music, especially for general musical categories that span over multiple pieces and infiltrate entire

styles (see Gjerdingen and Bourne 2015). Listeners may use inter-opus analogy to perceive these

familiar and conventional patterns. The construction grammar literature looks to analogy as a way

for listeners to identify constructions (Bybee 2010, 57–75). Bybee (2010) considers analogy “the

process by which a speaker comes to use a novel item in a construction”. (57) Therefore, listeners of

music may also use analogy to recognize familiar constructions, even if there is a novel aspect

involved.

How Modern Listeners Hear Common Practice Music through Film Music

This next project ask, “How does exposure to film music change how a modern listener

hears late-eighteenth and early nineteenth-century common practice music?” How does it change a

listener’s interpretation to consider that they are hearing Beethoven’s sixth symphony after already

being exposed to the movie soundtrack for Lord of the Rings? What musical patterns or schemata do

people learn when listening to film music? And, do they make an analogy between these film music

patterns and Western, Classical common practice music? If listeners hear a common-practice pattern

similar to a more modern pattern, then they are cognitively more likely to activate the “modern”

version and its associations instead of a “historically contextualized” one. The analogy framework
279

can be used for theorists to make a cognitive argument as to when listeners may merge these

different musical events as one. While other scholars analyze pattern change through time, I instead

will analyze “backward” in time: how a modern listener exposed to current versions of these

patterns or categories may assume they are the same as historicized versions in common-practice

music.

For this project, my methodology will combine the analogy framework, score analysis and

listener reactions to a common-practice corpus and a film music corpus. For score analysis, I will

conduct a corpus study of Austro-Germanic instrumental music (1750-1830; particularly Mozart,

Haydn and Beethoven) and popular American film music (1930-2010), assembling at least thirty

exemplars of each pattern in both corpora. I will analyze common topics and schemata in the score,

how these patterns change over time but remain recognizable, and their contexts for clues to their

associations. For listener reactions, historical sources discussing these patterns in context—

pedagogical treatises and reviews—address how contemporaries may have heard them, in addition

to current scholarship connecting them to other cultural areas (e.g. literature). Film music uses these

patterns paired with images. This imagery either reinforces or differs from historical perspectives of

the patterns. If modern listeners imagined similar film visuals listening to common-practice excerpts,

then it could indicate listeners use analogy to map modern versions onto historicized ones.

Independent Classical music blog reviews and online comments reveal this imagery. If feasible, an

online survey would ask participants to write imagery while listening to common-practice excerpts.

This project gives a present-to-past, longitudinal study of style change—how patterns and their

associations change over time and how listeners’ cultural context impacts if/when they use analogy

to cognitively map them over.


280

Analogy as a Window to Musical Communication

This dissertation begins a research trajectory towards developing a cognitively-based model

of musical communication and sense-making. How do listeners make sense of music? What are

their inferences or understanding? Analogy may be a window into musical communication, though

one of many that could be used to peer inside the house.

To discover a possible “house” for musical communication, I briefly review two skills that

psychologist Michael Tomasello (2003) argues are of particular importance for language acquisition:

intention-reading and pattern-finding (3). The first—intention-reading—includes:

1) “Ability to share attention with other persons to objects and events of mutual interest,”
2) “The ability to follow the attention and gesturing of other persons to distal objects and
events outside the immediate interaction,”
3) “The ability to actively direct the attention of others to distal objects by pointing,
showing, and using of other nonlinguistic gestures,”
4) “The ability to culturally (imitatively) learn the intentional actions of others, including
their communicative acts underlain by communicative.” (Tomasello 2003, 3)

He argues that these skills are necessary in order to learn the:

Appropriate use of any and all linguistic symbols, including complex linguistic expressions
and constructions. Indeed, they basically define the symbolic or functional dimensions of
linguistic communication—which involves in all cases the attempt of one person to
manipulate the intentional or mental state of other persons. Importantly in the current
context, this functional dimension enables certain kinds of abstraction processes, such as
analogy, that can only be effected when the elements to be compared play similar functional
(communicative) roles in larger linguistic expressions and/or constructions. (Tomasello
2003, 3–4, bold added for emphasis)

He holds that use and intentionality give abstractions meaning. The second skill—pattern-finding—

includes:

1) “The ability to form perceptual and conceptual categories of “similar” objects and
events,”
2) “The ability to form sensory-motor schemas from recurrent patterns of perception and
action,”
3) “The ability to perform statistically based distributional analyses on various kinds of
perceptual and behavioral sequences,”
281

4) “The ability to create analogies (structure mappings) across two or more complex
wholes, based on the similar functional roles of some elements in these different
wholes” (Tomasello 2003, 4, bold added for emphasis).

Tomasello (2003) mentions that these skills are necessary to find patterns across different utterances

and construct abstract dimensions of linguistic competence (4). Analogy is one facet of a pattern-

finding skill.

This dissertation isolates analogy’s possible role in musical sense-making. How do listeners

use both of Tomasello’s (2003) skills, intention-reading and pattern-finding, for musical acquisition

and inferring musical communication? Understanding intention-reading in music and how it

connects to musical pattern-finding might be especially rewarding in understanding musical

communication. What is the purpose of knowing patterns if a person does not consider why they are

used and when? The analogy framework is just one facet that could be elaborated upon to tackle

larger issues of musical communication and cognition.

Conclusion

When people listen to music, somehow it seems to fill their heads with ideas; music connects

us to something deeply human and we make interpretations intuitively—meaningful connections

between a past experience and a present response. The analogy framework provides a cognitively-

informed way to analyze music as a participatory action on the part of a listener. This framework

contributes to goals of music theory—recognizing pitch, meter-rhythm and form structures. Yet, it

can also be used to analyze how listeners in different contexts perceive relations and infer

interpretations from these structures. This dissertation looks at the “somehow” of “somehow it fills

my head with ideas”—attempting to answer what that “somehow” is.


282

Works Cited

Adams, Charles. 1976. “Melodic Contour Typology.” Ethnomusicology 20 (2): 179–215.

Agawu, Kofi. 1991. Playing with Signs: A Semiotic Interpretation of Classic Music. Princeton: Princeton

University Press.

———. 2009. Music as Discourse: Semiotic Adventures in Romantic Music. Oxford Studies in Music

Theory. Oxford: Oxford University Press.

Allanbrook, Wye Jamison. 1992. “Two Threads through the Labyrinth: Topic and Process in the

First Movements of K. 322 and K. 333.” In Convention in Eighteenth- and Nineteenth-Century

Music: Essays in Honor of Leonard G. Ratner, edited by Wye Jamison Allanbrook, Janet Levy,

and William Mahrt, 125–72. Stuyvesant: Pendragon Press.

———. 2008. “Mozart’s K331, First Movement: Once More, with Feeling.” In Communication in

Eighteenth-Century Music, edited by Danuta Mirka and V. Kofi Agawu, 254–82. Cambridge:

Cambridge University Press.

Almén, Byron. 2008. A Theory of Musical Narrative. Musical Meaning and Interpretation.

Bloomington: Indiana University Press.

Anderson, Benjamin. 2012. “Understanding Music Through Mental Representations: An

Investigation of A Priori and Ad Hoc Schemas”. PhD diss., Evanston: Northwestern

University.

Anderson, Emily, ed. 1961. The Letters of Beethoven. Translated by Emily Anderson. Vol. 2. 3 vols.

London: Macmillan.

Ashley, Richard. 2002. “Do[n’t] Change a Hair for Me: The Art of Jazz Rubato.” Music Perception 19

(3): 311–32. doi:10.1525/mp.2002.19.3.311.

Attardo, Salvatore. 2000. “Irony as Relevant Inappropriateness.” Journal of Pragmatics 32 (6): 793–826.
283

Austin, J. L. 1962. How to Do Things with Words. The William James Lectures. Cambridge: Harvard

University Press.

Baldwin, Dare. 1992. “Clarifying the Role of Shape in Children’s Taxonomic Assumption.” Journal of

Experimental Child Psychology 54 (3): 392–416. doi:10.1016/0022-0965(92)90027-4.

Balter, Tamara. 2009. “A Theory of Irony in Music: Types of Irony in the String Quartets of Haydn

and Beethoven”. PhD diss., Bloomington, IN: Indiana University.

Barr, Robin, and Leslie Caplan. 1987. “Category Representations and Their Implications for Their

Category Structure.” Memory & Cognition 15 (5): 397–418.

Barsalou, Lawrence. 1983. “Ad Hoc Categories.” Memory and Cognition 11: 211–27.

Bassok, Miriam. 2001. “Semantic Alignments in Mathematical Word Problems.” In The Analogical

Mind: Perspectives from Cognitive Science, edited by Dedre Gentner, Keith Holyoak, and Boicho

Kokinov, 401–34. Cambridge: MIT Press.

Beethoven, Ludwig van. 1793. “Ludwig van Beethoven, Brief an Eleonore von Breuning in Bonn,

Wien, 2. November 1793, Autograph”, November 2. Beethoven-Haus Bonn Digital

Archives. http://www.beethoven-haus-

bonn.de/sixcms/detail.php?id=15288&template=dokseite_digitales_archiv_en&_dokid=ha:

b188&_seite=1-1.

———. 1972. Beethoven’s Letters. Translated by John Shedlock. Dent’s International Library of Books

on Music. London & Toronto, New York: J. M. Dent & sons ltd.; E. P. Dutton & co.

Bharucha, Jamshed. 1987. “Music Cognition and Perceptual Facilitation: A Connectionist

Framework.” Music Perception 5 (1): 1–30.

Bigand, Emmanuel. 1990. “Abstraction of Two Forms of Underlying Structure in a Tonal Melody.”

Psychology of Music 18 (1): 45–59. doi:10.1177/0305735690181004.


284

Bigand, Emmanuel, and Richard Parncutt. 1999. “Perceiving Musical Tension in Long Chord

Sequences.” Psychological Research 62: 237–54.

Bigand, Emmanuel, Richard Parncutt, and Fred Lerdahl. 1996. “Perception of Musical Tension in

Short Chord Sequences: The Influence of Harmonic Function, Sensory Dissonance,

Horizontal Motion, and Musical Training.” Perception & Psychophysics 58 (1): 125–41.

Bigand, Emmanuel, and Benedicte Poulin-Charronnat. 2006. “Are We ‘experienced Listeners’? A

Review of the Musical Capacities That Do Not Depend on Formal Musical Training.”

Cognition 100 (1): 100–130. doi:10.1016/j.cognition.2005.11.007.

Blanchette, Isabelle, and Kevin Dunbar. 2002. “Representational Change and Analogy: How

Analogical Inferences Alter Target Representations.” Journal of Experimental Psychology:

Learning, Memory, and Cognition 28 (4): 672–85.

Bonds, Mark Evan. 1991. Wordless Rhetoric: Musical Form and the Metaphor of the Oration. Cambridge:

Harvard University Press.

———. 2008. “Listening to Listeners.” In Communication in Eighteenth-Century Music, edited by Danuta

Mirka and Kofi Agawu, 34–52. Cambridge: Cambridge University Press.

Botstein, Leon. 1994. “The Patrons and Publics of the Quartets: Music, Culture, and Society in

Beethoven’s Vienna.” In The Beethoven Quartet Companion, edited by Robert Winter and Robert

Martin, 77–110. Berkeley: University of California Press.

Brauner, Charles. 1981. “Irony in the Heine Lieder of Schubert and Schumann.” The Musical Quarterly

67 (2): 261–81.

Brodbeck, David, and John Platoff. 1983. “Dissociation and Integration: The First Movement of

Beethoven’s Opus 130.” 19th-Century Music 7 (2): 149–62. doi:10.2307/746346.


285

Brooks, Lee, Geoffrey Norman, and Scott Allen. 1991. “Role of Specific Similarity in a Medical

Diagnostic Task.” Journal of Experimental Psychology: General 120 (3): 278–87.

doi:10.1037/0096-3445.120.3.278.

Brower, Candace. 2000. “A Cognitive Theory of Musical Meaning.” Journal of Music Theory 44 (2):

323–79.

Broyles, Michael. 1987. Beethoven: The Emergence and Evolution of Beethoven’s Heroic Style. 1st ed. New

York: Excelsior Music Pub. Co. : Distributed to the book trade by Scientific and Technical

Book Service.

Burnham, Scott. 2000. Beethoven Hero. Princeton: Princeton University Press.

Burns, James. 2010. “Rhythmic Archetypes in Instrumental Music from Africa and the Diaspora.”

Music Theory Online 16 (4).

Butterfield, Matthew. 2002. “The Musical Object Revisited.” Music Analysis 21 (3): 327–80.

Bybee, Joan. 2010. Language, Usage, and Cognition. Cambridge: Cambridge University Press.

Byros, Vasili. 2009a. “Foundations of Tonality as Situated Cognition 1730-1830: An Inquiry into the

Culture and Cognition of Eighteenth-Century Tonality, with Beethoven’s Eroic Symphony

as a Case Study”. PhD diss., Yale University.

———. 2009b. “Towards an ‘Archeology’ of Hearing: Schemata and Eighteenth-Century

Consciousness.” Musica Humana 1 (2): 235–306.

———. 2012. “Meyer’s Anvil: Revisiting the Schema Concept.” Music Analysis 31 (3): 273–346.

doi:10.1111/j.1468-2249.2012.00344.x.

———. 2013. “Trazom’s Wit: Communicative Strategies in a ‘Popular’ yet ‘Difficult’ Sonata.”

Eighteenth-Century Music 10 (2): 213–52. doi:10.1017/S1478570613000055.


286

Caplin, William. 1998. Classical Form: A Theory of Formal Functions for the Instrumental Music of Haydn,

Mozart, and Beethoven. Oxford: Oxford University Press.

———. 2005. “On the Relation of Musical Topoi to Formal Function.” Eighteenth-Century Music 2

(1): 113–24. doi:10.1017/S1478570605000278.

Carroll, Lewis. 1865. Alice’s Adventures in Wonderland and Through the Looking-Glass and What Alice

Found There. New York: Barnes & Noble Books.

Catrambone, Richard. 2002. “The Effects of Surface and Structural Feature Matches on the Access

of Story Analogs.” Journal of Experimental Psychology: Learning, Memory, and Cognition 28 (2):

318–34. doi:10.1037/0278-7393.28.2.318.

Catrambone, Richard, and Keith Holyoak. 1989. “Overcoming Contextual Limitations on Problem-

Solving Transfer.” Journal of Experimental Psychology: Learning, Memory, and Cognition 15 (6):

1147–56. doi:10.1037/0278-7393.15.6.1147.

Cavett-Dunsby, Esther. 1985. “Mozart’s Variations Reconsidered: Four Case Studies (K. 613, K.

501, K.421/417, K. 491)”. PhD diss., London: University of London.

Chan Barrrett, Karen. 2015. “Musical Counterpoint: Attention and Perception”. PhD diss.,

Evanston: Northwestern University.

Chi, Michelene, Paul Feltovich, and Robert Glaser. 1981. “Categorization and Representation of

Physics Problems by Experts and Novices.” Cognitive Science 15: 89–132.

Christensen, Bo, and Christian Schunn. 2007. “The Relationship of Analogical Distance to

Analogical Function and Preinventive Structure: The Case of Engineering Design.” Memory

& Cognition 35 (1): 29–38.


287

Christie, Stella, and Dedre Gentner. 2010. “Where Hypotheses Come From: Learning New

Relations by Structural Alignment.” Journal of Cognition and Development 11 (3): 356–73.

doi:10.1080/15248371003700015.

Chua, Daniel. 1995. The “Galitzin” Quartets of Beethoven: Opp. 127, 132, 130. Princeton: Princeton

University Press.

Clarke, Eric. 1987. “Categorical Rhythm Perception: An Ecological Perspective.” In Action and

Perception in Rhythm and Music, edited by Alf Gabrielsson, 19–33. Stockholm: Royal Swedish

Academy of Music.

———. 1999. “Rhythm and Timing in Music.” In The Psychology of Music, edited by Diana Deutsch,

473–500. Boston: Academic Press.

———. 2005. Ways of Listening: An Ecological Approach to the Perception of Musical Meaning. New York:

Oxford University Press.

Clement, Catherine, and Dedre Gentner. 1991. “Systematicity as a Selection Constraint in Analogical

Mapping.” Cognitive Science 15 (1): 89–132. doi:10.1016/0364-0213(91)80014-V.

Cobly, Paul. 2005. “Objectivity and Immanence in Genre Theory.” In Genre Matters: Interdisciplinary

Perspectives, edited by Garin Dowd, Lesley Stevenson, and Jeremy Strong, 41–54. London:

Intellect.

———. 2008. “Communication and Verisimilitude in the Eighteenth Century.” In Communication in

Eighteenth-Century Music, edited by Danuta Mirka and Kofi Agawu, 13–33. Cambridge:

Cambridge University Press.

Colebrook, Claire. 2004. Irony. New York: Routledge.

Colston, Herbert. 2001. “On Necessary Conditions for Verbal Irony Comprehension.” Pragmatics &

Cognition 8 (2): 277–324.


288

Cone, Edward. 1974. The Composer’s Voice. The Ernest Bloch Lectures. Berkeley: University of

California Press.

Corral, Daniel, and Matt Jones. 2012. “Learning of Relational Categories as a Function of Higher-

Order Structure.” In Proceedings of the 34th Annual Meeting of the Cognitive Science Society, edited by

Naomi Miyake, David Peebles, and Richard Cooper, 1434–39. Austin: Cognitive Science

Society.

Cross, Ian. 2005. “Music and Meaning, Ambiguity and Evolution.” In Musical Communication, edited

by Dorothy Miell, Raymond MacDonald, and David Hargreaves, 27–43. Oxford: Oxford

University Press.

Dahlhaus, Carl. 1991. Ludwig van Beethoven: Approaches to His Music. Oxford: Oxford University Press.

Daube, Johann Friedrich. [1793] 1992. The Musical Dilettante: A Treatise on Composition by J.F. Daube.

Translated by Susan Snook-Luther. Cambridge: Cambridge University Press.

Day, Samuel, and Dedre Gentner. 2007. “Nonintentional Analogical Inference in Text

Comprehension.” Memory & Cognition 35 (1): 39–49.

Deliège, Irène. 1996. “Cue Abstraction as a Component of Categorisation Processes in Music

Listening.” Psychology of Music 24 (2): 131–56. doi:10.1177/0305735696242007.

———. 2007. “Similarity Relations in Listening to Music: How Do They Come into Play?” Musicae

Scientiae 11 (1): 9–37. doi:10.1177/1029864907011001021.

Deliège, Irène, and Marc Mélen. 1997. “Cue Abstraction in the Representation of Musical Form.” In

Perception and Cognition of Music, edited by Irène Deliège and John Sloboda, 387–412. Hove,

East Sussex: Psychology Press.


289

Dibben, Nicola. 2003. “Musical Materials, Perception, and Listening.” In The Cultural Study of Music:

A Critical Introduction, edited by Martin Clayton, Trevor Herbert, and Richard Middleton,

193–203. London: Routledge.

Dowling, W. Jay. 1978. “Scale and Contour: Two Components of a Theory of Memory for

Melodies.” Psychological Review 85 (4): 341–54.

Dunbar, Kevin. 1995. “How Scientists Really Reason: Scientific Reasoning in Real-World

Laboratories.” In The Nature of Insight, edited by Robert Sternberg and Janet Davidson, 365–

95. Cambridge: The MIT Press.

———. 2001. “The Analogical Paradox: Why Analogy Is so Easy in Naturalistic Settings, Yet so

Difficult in the Psychological Laboratory.” In The Analogical Mind: Perspectives from Cognitive

Science, edited by Dedre Gentner, Keith Holyoak, and Boicho Kokinov, 313–34. Cambridge:

MIT Press.

Eitan, Zohar, and Roni Granot. 2007. “Intensity Changes and Perceived Similarity: Inter-Parametric

Analogies.” Musicae Scientiae 4A: 39–75.

———. 2009. “Primary versus Secondary Musical Parameters and the Classification of Melodic

Motives.” Musicae Scientiae 13 (1): 139–79. doi:10.1177/102986490901300107.

Eiting, Mindert. 1984. “Perceptual Similarities between Musical Motifs.” Music Perception 2 (1): 78–94.

Elio, Renee, and John Anderson. 1984. “The Effects of Information Order and Learning Mode on

Schema Abstraction.” Memory and Cognition 12: 20–30.

Everett, Yayoi Uno. 2004. “Parody Wih an Ironic Edge: Dramatic Words by Kurt Weill, Peter

Maxwell Davies, and Louis Andriessen.” Music Theory Online 10 (4).

———. 2009. “Signification of Parody and the Grotesque in György Ligeti’s Le Grand Macabre.”

Music Theory Spectrum 31 (1): 26–56.


290

Fincher, David. 2010. The Social Network. DVD. Columbia Pictures.

Fish, Stanley Eugene. 1980. Is There a Text in This Class?: The Authority of Interpretive Communities.

Cambridge: Harvard University Press.

Flyvbjerg, Bent. 2006. “Five Misunderstandings About Case-Study Research.” Qualitative Inquiry 12

(2): 219–45.

Fodor, Jerry. 1983. Modularity of Mind: An Essay on Faculty Psychology. Cambridge: MIT Press.

Frye, Northrop. 1957. Anatomy of Criticism: Four Essays. Princeton: Princeton University Press.

Gentner, Dedre. 1983. “Structure-Mapping: A Theoretical Framework for Analogy.” Cognitive Science

7: 155–70.

———. 2003. “Psychology of Analogical Learning.” In Encyclopedia of Cognitive Science, edited by L

Nadel, 106–12. London: Nature Publishing Group.

———. 2005. “The Development of Relational Category Knowledge.” In Building Object Categories in

Developmental Time, edited by Lisa Gershkoff-Stowe and David Rakison, 245–75. Mahwah:

Lawrence Erlbaum Associates.

———. 2010. “Bootstrapping the Mind: Analogical Processes and Symbol Systems.” Cognitive Science

34 (5): 752–75. doi:10.1111/j.1551-6709.2010.01114.x.

Gentner, Dedre, Keith Holyoak, and Boicho Kokinov, eds. 2001. The Analogical Mind: Perspectives from

Cognitive Science. Cambridge: MIT Press.

Gentner, Dedre, and Kenneth Kurtz. 2005. “Relational Categories.” In Categorization inside and Outside

the Laboratory, edited by Woo-Kyoung Ahn, Robert Goldstone, Bradley Love, Arthur

Markman, and Phillip Wolff, 151–75. Washington, DC: American Psychological Association.
291

Gentner, Dedre, Jeffrey Loewenstein, Leigh Thompson, and Kenneth Forbus. 2009. “Reviving Inert

Knowledge: Analogical Abstraction Supports Relational Retrieval of Past Events.” Cognitive

Science 33 (8): 1343–82. doi:10.1111/j.1551-6709.2009.01070.x.

Gentner, Dedre, and Arthur Markman. 1997. “Structure Mapping in Analogy and Similarity.”

American Psychologist 52 (1): 45–56.

Gentner, Dedre, and José Medina. 1998. “Similarity and the Development of Rules.” Cognition 65 (2–

3): 263–97. doi:10.1016/S0010-0277(98)00002-X.

Gentner, Dedre, and Laura Namy. 1999. “Comparison in the Development of Categories.” Cognitive

Development 14 (4): 487–513. doi:10.1016/S0885-2014(99)00016-7.

Gentner, Dedre, Mary Jo Ratterman, and Kenneth Forbus. 1993. “The Roles of Similarity in

Transfer: Separating Retrievability from Inferential Soundness.” Cognitive Psychology 25: 524–

75.

Gentner, Dedre, and Linsey Smith. 2012. “Analogical Reasoning.” In Encyclopedia of Human Behavior,

edited by Vilayanur Ramachandran, 2nd ed., 130–36. Oxford: Elsevier.

———. 2013. “Analogical Learning and Reasoning.” In Oxford Handbook of Cognitive Psychology, edited

by Daniel Reisberg, 668–81. New York: Oxford University Press.

Gibbs Jr., Raymond, and Jennifer O’Brien. 1991. “Psychological Aspects of Irony Understanding.”

Journal of Pragmatics 16 (6): 523–30.

Gibbs, Raymond. 2007. “A Brief History of Irony.” In Irony in Language and Thought: A Cognitive

Science Reader, edited by Raymond Gibbs and Herbert Colston, 3–24. New York: Lawrence

Erlbaum Associates.

Gibbs, Raymond, and Herbert Colston. 2012. Interpreting Figurative Meaning. Cambridge: Cambridge

University Press.
292

Gick, Mary, and Keith Holyoak. 1980. “Analogical Problem Solving.” Cognitive Psychology 12 (3): 306–

55. doi:10.1016/0010-0285(80)90013-4.

———. 1983. “Schema Induction and Analogical Transfer.” Cognitive Psychology 15 (1): 1–38.

doi:10.1016/0010-0285(83)90002-6.

Gjerdingen, Robert. 1988. A Classic Turn of Phrase: Music and the Psychology of Convention. Philadelphia:

University of Pennsylvania Press.

———. 1992. “Revisiting Meyer’s ‘Grammatical Simplicity and Relational Richness.’” In Cognitive

Bases of Musical Communication, edited by Mari Riess Jones and Susan Holleran, 225–48.

Washington, DC: American Psychological Association.

———. 1996. “Courtly Behaviors.” Music Perception 13 (3): 365–82. doi:10.2307/40286175.

———. 2007a. Music in the Galant Style: Being an Essay on Various Schemata Characteristic of Eighteenth-

Century Music for Courtly Chambers, Chapel, and Theaters, Including Tasteful Passages of Music Drawn

from Most Excellent Chapel Masters in the Employ of Noble and Noteworthy Personages, Said Music All

Collected for the Reader’s Delectations on the World Wide Web. Oxford: Oxford University Press.

———. 2007b. “Partimento, Que Me Veux-Tu?” Journal of Music Theory 51 (1): 85–136.

———. 2010a. “Mozart’s Obviously Corrupt Minuet.” Music Analysis 29 (1-2-3): 61–82.

———. 2010b. “Partimenti Written to Impart a Knowledge of Counterpoint and Composition.” In

Partimento and Continuo Playing in Theory and Practice: Collected Writings of the Orpheus Institute, 43–

70. Leuven: Leuven University Press.

Goldberg, Adele. 2006. Constructions at Work: The Nature of Generalization in Language. Oxford: Oxford

University Press.

Goldstone, Robert, and Douglas Medin. 1994. “Time Course of Comparison.” Journal of Experimental

Psychology: Learning, Memory, and Cognition 20 (1): 29–50. doi:10.1037/0278-7393.20.1.29.


293

Goldstone, Robert, Douglas Medin, and Dedre Gentner. 1991. “Relational Similarity and the

Nonindependence of Features in Similarity Judgments.” Cognitive Psychology 23: 222–62.

Goldwater, Micah, Arthur Markman, and Hunt Stilwell. 2011. “The Empirical Case for Role-

Governed Categories.” Cognition 118 (3): 359–76. doi:10.1016/j.cognition.2010.10.009.

Goodman, Nelson. 1972. “Seven Strictures on Similarity.” In Problems and Projects, edited by Nelson

Goodman, 437–47. New York: Bobbs-Merrill.

Grice, H. Paul. 1957. “Meaning.” Philosophical Review 64: 377–88.

———. 1975. “Logic and Conversation.” In Syntax and Semantics 3: Speech Acts, edited by Peter Cole

and Jerry Morgan, 41–58. New York: Academic Press.

Halpern, Andrea. 1984. “Perception of Structure in Novel Music.” Memory & Cognition 12 (2): 163–

70.

Hanninen, Dora. 2004. “Associative Sets, Categories, and Music Analysis.” Jounal of Music Theory 48

(2): 147–218.

———. 2012. A Theory of Music Analysis: On Segmentation and Associative Organization. Eastman Studies

in Music, v. 92. Rochester: University of Rochester Press.

Hargreaves, David, Raymond MacDonald, and Dorothy Miell, eds. 2005. Musical Communication.

Oxford: Oxford University Press.

Harrison, Daniel. 2000. “A Story, an Apologia; And a Survey.” Intégral 14/15: 29–37.

Hatten, Robert. 1994. Musical Meaning in Beethoven: Markedness, Correlation, and Interpretation. Advances

in Semiotics. Bloomington: Indiana University Press.

———. 2004. Interpreting Musical Gestures, Topics, and Tropes: Mozart, ZBeethoven, Schubert. Bloomington:

Indiana University Press.


294

Hepokoski, James. 2001. “Back and Forth from Egmont: Beethoven, Mozart, and the Nonresolving

Recapitulation.” 19th-Century Music 25 (2-3): 127–54.

Hepokoski, James, and Warren Darcy. 2006. Elements of Sonata Theory: Norms, Types, and Deformations in

the Late Eighteenth-Century Sonata. Oxford: Oxford University Press.

Hirsch, Galia. 2011. “Between Irony and Humor: A Pragmatic Model.” Pragmatics & Cognition 19 (3):

530–61.

Hofstadter, Douglas. 2001. “Epilogue: Analogy as the Core of Cognition.” In The Analogical Mind:

Perspectives from Cognitive Science, edited by Dedre Gentner, Keith Holyoak, and Boicho

Kokinov, 499–538. Cambridge: MIT Press.

Hofstadter, Douglas, and Emmanuel Sander. 2013. Surfaces and Essences: Analogy as the Fuel and Fire of

Thinking. New York: Basic Books.

Holyoak, Keith. 2012. “Analogy and Relational Reasoning.” In The Oxford Handbook of Thinking and

Reasoning, edited by Keith Holyoak and Robert Morrison, 234–59. Oxford: Oxford

University Press.

Holyoak, Keith, and Paul Thagard. 1989. “Analogical Mapping by Constraint Satisfaction.” Cognitive

Science 13: 295–355.

———. 1995. Mental Leaps: Analogy in Creative Thought. Cambridge: MIT Press.

Honing, Henkjan. 2013. “Structure and Interpretation of Rhythmm in Music.” In , edited by Diana

Deutsch, 3rd ed., 369–404. New York: Academic Press.

Horn, Laurence, and Gregory Ward, eds. 2004. The Handbook of Pragmatics. Blackwell Handbooks in

Linguistics 16. Malden: Blackwell Pub.

Hoyle, John. 1770. Dictionarium Musica, Being a Complete Dictionary: Or, Treasury of Music. A Full

Explanation of All the Words and Terms Made Use of in Music, Both Speculative, Practical and
295

Historical. All the Words and Terms Made Use of by the Italians, Are Also Inserted. The Whole

Compiled from the Best Antient and Modern Authors Who Have Wrote on the Subject. London.

http://find.galegroup.com/ecco/infomark.do?&source=gale&prodId=ECCO&userGroup

Name=northwestern&tabID=T001&docId=CW3306270950&type=multipage&contentSet

=ECCOArticles&version=1.0&docLevel=FASCIMILE.

Hunter, Mary. 2012. “‘The Most Interesting Genre of Music’: Performance, Sociability and Meaning

in the Classical String Quartet, 1800-1830.” Nineteenth-Century Music Review 9: 53–74.

Huron, David Brian. 2006. Sweet Anticipation: Music and the Psychology of Expectation. Cambridge: MIT

Press.

Hutcheon, Linda. 1995. Irony’s Edge: The Theory and Politics of Irony. New York: Routledge.

Imai, Mutsumi, Dedre Gentner, and Nobuko Uchida. 1994. “Children’s Theories of Word Meaning:

The Role of Shape Similarity in Early Acquisition.” Cognitive Development 9 (1): 45–75.

doi:10.1016/0885-2014(94)90019-1.

“Irony, N.” 2015. Oxford English Dictionary Online. Oxford: Oxford University Press.

http://www.oed.com/view/Entry/99565?rskey=50sKEi&result=1.

Iser, Wolfgang. 1978. The Act of Reading: A Theory of Aesthetic Response. Baltimore: Johns Hopkins

University Press.

Ito, John Paul. 2013. “Hypermetrical Schemas, Metrical Orientation, and Cognitive-Linguistic

Paradigms.” Journal of Music Theory 57 (1): 47–85. doi:10.1215/00222909-2017106.

Ivanovitch, Roman. 2004. “Mozart and the Environment of Variation”. PhD diss., New Haven:

Yale University.

———. 2010. “What’s in a Theme? On the Nature of Variation.” Gamut: Online Journal of the Music

Theory Society of the Mid-Atlantic 3 (1): Article 3.


296

James, William. 1950. The Principles of Psychology. New York: Dover.

———. 2000. “The Stream of Consciousness.” In Pragmatism and Other Writings, edited by Gunn

Giles. New York: Penguin Books.

Jee, Benjamin, David Uttal, Dedre Gentner, Cathy Manduca, Thomas Shipley, and Bradley Sageman.

2013. “Finding Faults: Analogical Comparison Supports Spatial Concept Learning in

Geoscience.” Cognitive Processing 14 (2): 175–87. doi:10.1007/s10339-013-0551-7.

Jee, Benjamin, David Uttal, Dedre Gentner, Cathy Manduca, Thomas Shipley, Basil Tikoff, Carol

Ormand, and Bradley Sageman. 2010. “Commentary: Analogical Thinking in Geoscience

Education.” Journal of Geoscience Education 58 (1): 2–13. doi:10.5408/1.3544291.

Jones, Mari Riess. 1976. “Time, Our Lost Dimension: Toward a New Theory of Perception,

Attention, and Memory.” Psychological Review 83 (5): 323–55. doi:10.1037/0033-

295X.83.5.323.

Jones, Timothy. 2014. “Variation Form.” The Oxford Companion to Music. Oxford Music Online. Oxford

University Press. Accessed July 11.

http://www.oxfordmusiconline.com.turing.library.northwestern.edu/subscriber/article/opr

/t114/e7067.

Jung, Wookyoung, and John Hummel. 2014. “Making Probabilistic Relational Categories

Learnable.” Cognitive Science, 1–33.

Keane, Mark. 1987. “On Retrieving Analogues When Solving Problems.” The Quarterly Journal of

Experimental Psychology Section A: Human Experimental Psychology 39 (1): 29–41.

doi:10.1080/02724988743000015.
297

Keane, Mark, and Fintan Costello. 2001. “Setting Limits on Analogy: Why Conceptual Combination

Is Not Structural Alignment.” In The Analogical Mind: Perspectives from Cognitive Science, edited

by Dedre Gentner, Keith Holyoak, and Boicho Kokinov, 287–312. Cambridge: MIT Press.

Kerman, Joseph. 1971. The Beethoven Quartets. New York: A. A. Knopf.

Kielian-Gilbert, Marianne. 1990. “Interpreting Musical Analogy: From Rhetorical Device to

Perceptual Process.” Music Perception 8 (1): 63–94. doi:10.2307/40285486.

Kinderman, William. 1995. Beethoven. Berkeley: University of California Press.

———. 2009. Beethoven. 2nd ed. Oxford: Oxford University Press.

Klein, Michael. 2009. “Ironic Narrative, Ironic Reading.” Journal of Music Theory 53 (1): 95–136.

Koniari, Dimitra, Sandrine Predazzer, and Marc Mélen. 2001. “Categorization and Schematization

Processes Used in Music Perception by 10- to 11-Year-Old Children.” Music Perception 18 (3):

297–324. doi:10.1525/mp.2001.18.3.297.

Krumhansl, Carol. 1990. Cognitive Foundations of Musical Pitch. Oxford: Oxford University Press.

———. 1995. “Music Psychology and Music Theory: Problems and Prospects.” Music Theory

Spectrum 17 (1): 53–90.

Lakoff, George. 1989. More than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of

Chicago Press.

Lamont, Alexandra, and Nicola Dibben. 2001. “Motivic Structure and the Perception of Similarity.”

Music Perception 18 (3): 245–74. doi:10.1525/mp.2001.18.3.245.

Larson, Steve. 2012. Musical Forces: Motion, Metaphor, and Meaning in Music. Bloomington: Indiana

University Press.

Lerdahl, Fred, and Ray Jackendoff. 1983. A Generative Theory of Tonal Music. Cambridge: The MIT

Press.
298

Levinson, Jerrold. 1997. Music in the Moment. Ithaca: Cornell University Press.

———. 2004. “Music as Narrative and Music as Drama.” Mind & Language 19 (4): 428–41.

doi:10.1111/j.0268-1064.2004.00267.x.

Levy, Janet. 1992. “‘Something Mechanical Encrusted on the Living’: A Source of Musical Wit and

Humor.” In Convention in Eighteenth- and Nineteenth-Century Music: Essays in Honor of Leonard G.

Ratner, edited by Wye J. Allanbrook, Janet Levy, and William Mahrt, 225–56. Stuyvesant:

Pendragon Press.

Lockwood, Lewis. 2003. Beethoven: The Music And The Life. New York: W. W. Norton & Company.

Loewenstein, Jeffrey, Leigh Thompson, and Dedre Gentner. 2003. “Analogical Learning in

Negotiation Teams: Comparing Cases Promotes Learning and Transfer.” Academy of

Management Learning & Education 2 (2): 119–27. doi:10.5465/AMLE.2003.9901663.

London, Justin. 1996. “Musical and Linguistic Speech Acts.” The Journal of Aesthetics and Art Criticism

54 (1): 49–64.

———. 2004. Hearing in Time: Psychological Aspects of Musical Meter. New York: Oxford University

Press.

Longyear, Rey. 1970. “Beethoven and Romantic Irony.” The Musical Quarterly, 647-664, 56 (4): 1970.

Love, Stefan. 2012. “‘Possible Paths’: Schemata of Phrasing and Melody in Charlie Parker’s Blues.”

Music Theory Online 18 (3).

Lucariello, Joan. 1994. “Situational Irony: A Concept of Events Gone Awry.” Journal of Experimental

Psychology: General 123 (2): 129–45.

Margulis, Elizabeth. 2007. “Surprise and Listening Ahead: Analytic Engagements with Musical

Tendencies.” Music Theory Spectrum 29 (2): 197–217.

———. 2013. On Repeat: How Music Plays the Mind. Oxford: Oxford University Press.
299

Markman, Arthur, and Dedre Gentner. 1993. “Structural Alignment During Similarity

Comparisons.” Cognitive Psychology 25: 431–67.

Marston, Nicholas. 1989. “Analysing Variations: The Finale of Beethoven’s String Quartet Op. 74.”

Music Analysis 8 (3): 303. doi:10.2307/854292.

Maus, Fred Everett. 1988. “Music as Drama.” Music Theory Spectrum 10: 56–73. doi:10.2307/745792.

McAdams, Stephen, Sandrine Vieillard, Oliver Houix, and Roger Reynolds. 2004. “Perception of

Musical Similarity Among Contemporary Thematic Materials in Two Instrumentations.”

Music Perception: An Interdisciplinary Journal 22 (2): 207–37. doi:10.1525/mp.2004.22.2.207.

McKee, Eric. 2012. Decorum of the Minuet, Delirium of the Waltz: A Study of Dance-Music Relations in 3/4

Time. Musical Meaning & Interpretation. Bloomington: Indiana University Press.

Medin, Douglas, Robert Goldstone, and Dedre Gentner. 1993. “Respects for Similarity.” Psychological

Review 100 (2): 254–78.

Medin, Douglas, and B Ross. 1989. “The Specific Character of Abstract Though: Categorization,

Problem Solving and Induction.” Advances in the Psychology of Human Intelligence 5: 189–223.

Meyer, Leonard. 1956. Emotion and Meaning in Music. Chicago: University of Chicago Press.

———. 1967. Music, the Arts, and Ideas: Patterns and Predictions in Twentieth-Century Culture. Chicago:

University of Chicago Press.

———. 1973. Explaining Music; Essays and Explorations. Berkeley: University of California Press.

———. 1980. “Exploiting Limits: Creation, Archetypes, and Style Change.” Daedalus 109 (2): 177–

205.

———. 1989. Style and Music: Theory, History, and Ideology. Chicago: University of Chicago Press.

———. 1998. “A Universe of Universals.” The Journal of Musicology 16 (1): 3–25. doi:10.2307/764076.

———. 2000. The Spheres of Music: A Gathering of Essays. Chicago: The Unviersity of Chicago Press.
300

Meyer, Leonard, and Burton Rosner. 2000. “Melodic Processes and the Perception of Music.” In The

Spheres of Music: A Gathering of Essays, 157–88. Chicago: University of Chicago Press.

Michaelis, Christian Friedrich. 1803. “Über Die Musikalische Wiederholung under Veränderung

Michaelis.” Allgemeine Musikalische Zeitung 13: Cols. 197–200.

Miller, George. 1956. “The Magical Number Seven, plus or Minus Two: Some Limits on Our

Capacity for Processing Information.” Psychological Review 63 (2): 81–97.

doi:10.1037/h0043158.

Mirka, Danuta. 2009. Metric Manipulations in Haydn and Mozart: Chamber Music for Strings, 1787-1791.

Oxford Studies in Music Theory. Oxford: Oxford University Press.

Mirka, Danuta, and V. Kofi Agawu, eds. 2008. Communication in Eighteenth-Century Music. Cambridge:

Cambridge University Press.

Monelle, Raymond. 2000. The Sense of Music: Semiotic Essays. Princeton: Princeton University Press.

———. 2006. The Musical Topic: Hunt, Military and Pastoral. Musical Meaning and Interpretation.

Bloomington: Indiana University Press.

Muecke, Douglas. 1969. The Compass of Irony. London: Methuen.

———. 1970. Irony. The Critical Idiom 13. Norfolk: Cox & Wyman Ltd.

Murphy, Gregory. 2002. The Big Book of Concepts. Cambridge: MIT Press.

Murphy, Gregory, and Douglas Medin. 1985. “The Role of Theories in Conceptual Change.”

Psychological Review 92: 289–316.

Narmour, Eugene. 1990. The Analysis and Cognition of Basic Melodic Structures: The Implication-Realization

Model. Chicago: University of Chicago Press.

Nelson, Robert. 1948. The Technique of Variation: A Study of the Instrumental Variation from Antonio de

Cabezon to Max Reger. Berkeley: University of California Press.


301

Novick, Laura, and Keith Holyoak. 1991. “Mathematical Problem Solving by Analogy.” Journal of

Experimental Psychology: Learning, Memory, and Cognition 17 (3): 398–415. doi:10.1037/0278-

7393.17.3.398.

Nussbaum, Charles. 2007. “The Musical Utterance: How Music Means.” In The Musical Representation:

Meaning, Ontology, and Emotion, 87–142. Cambridge: The MIT Press.

Ockelford, Adam. 2009. “Similarity Relations between Groups of Notes: Music-Theoretical and

Music-Psychological Perspectives.” Musicae Scientiae 13 (1 suppl): 47–98.

doi:10.1177/102986490901300104.

Oppenheimer, Robert. 1956. “Analogy in Science.” American Psychologist 11 (3): 127–35.

Palmer, Stephen. 1978. “Fundamental Aspects of Cognitive Representation.” In Cognition and

Categorization, edited by Eleanor Rosch and Barbara Lloyd, 259–303. Hillsdale: Lawrence

Elbaum Associates.

Pollard-Gott, Lucy. 1983. “Emergence of Thematic Concepts in Repeated Listening to Music.”

Cognitive Psychology 15 (1): 66–94. doi:10.1016/0010-0285(83)90004-X.

Rahn, John. 1980. Basic Atonal Theory. New York: Schirmer.

Raskin, Victor, and Salvatore Attardo. 1994. “Non-Literalness and Non-Bona-Fide in Language: An

Approach to Formal and Computational Treatments of Humor.” Pragmatics & Cognition 2 (1):

31–69. doi:10.1075/pc.2.1.02ras.

Ratner, Leonard. 1980. Classic Music: Expression, Form, and Style. New York: Schirmer Books.

———. 1995. The Beethoven String Quartets: Compositional Strategies and Rhetoric. Stanford: Stanford

Bookstore.

“Relation, N.” 2015. Oxford English Dictionary Online. Oxford: Oxford University Press.

http://www.oed.com/view/Entry/161810?redirectedFrom=relation.
302

Reti, Rudolph. 1951. The Thematic Process in Music. New York: Macmillan.

Reybrouck, Mark. 2009. “Similarity Perception as a Cognitive Tool for Musical Sense-Making:

Deictic and Ecological Claims.” Musicae Scientiae 13 (1): 99–118.

Richards, Mark. 2013. “Beethoven and the Obscured Medial Caesura: A Study in the

Transformation of Style.” Music Theory Spectrum 35 (2): 166–93.

doi:10.1525/mts.2013.35.2.166.

Riemann, Hugo. 1893. Dictionary of Music. Translated by John S. Shedlock. London: Augener & Co.

Rosen, Charles. 1976. The Classical Style. London: Faber and Faber.

———. 2002. Beethoven’s Piano Sonatas: A Short Companion. New Haven: Yale University Press.

Ross, Brian. 1987. “This Is like That: The Use of Earlier Problems and the Separation of Similarity

Effects.” Journal of Experimental Psychology: Learning, Memory, and Cognition 13 (4): 629–39.

doi:10.1037/0278-7393.13.4.629.

Rottman, Benjamin, Dedre Gentner, and Micah Goldwater. 2012. “Causal Systems Categories:

Differences in Novice and Expert Categorization of Causal Phenomena.” Cognitive Science 36

(5): 919–32.

Rudanko, Juhani. 2007. “Concepts for Analyzing Deception in Discourse Intended to Be Persuasive:

Two Case Studies from Shakespearean Drama.” Journal of Historical Pragmatics 8 (1): 109–26.

doi:10.1075/jhp.8.1.06rud.

Samson, Jim. 2007. “Genre.” Grove Music Online.

Sanguinetti, Giorgio. 2012. The Art of Partimento: History, Theory, and Practice. New York: Oxford

University Press.

Schachter, Carl. 1976. “Rhythm and Linear Analysis: A Preliminary Study.” In The Music Forum,

edited by Felix Salzer and Carl Schachter, 4:281–334. New York: Columbia University Press.
303

Schellenberg, E. Glenn, Stephanie Stalinski, and Bradley Marks. 2014. “Memory for Surface Features

of Unfamiliar Melodies: Independent Effects of Changes in Pitch and Tempo.” Psychological

Research 78 (1): 84–95. doi:10.1007/s00426-013-0483-y.

Schmuckler, Mark. A. 2010. “Melodic Contour Similarity Using Folk Melodies.” Music Perception 28

(2): 169–94. doi:10.1525/mp.2010.28.2.169.

Schoenberg, Arnold. 1967. Fundamentals of Musical Composition. London: Faber.

———. 1978. Style and Idea: Selected Writings of Arnold Schoenberg. New York: St. Martin Press.

———. 1995. The Musical Idea and the Logic, Technique, and Art of Its Presentation. New York: Columbia

University Press.

Senner, Wayne, Robin Wallace, and William Rhea Meredith, eds. 1999. The Critical Reception of

Beethoven’s Compositions by His German Contemporaries. Vol. 1. 2 vols. North American

Beethoven Studies. Lincoln: University of Nebraska Press.

Sewell, William. 2005. Logics of History: Social Theory and Social Transformation. Chicago Studies in

Practices of Meaning. Chicago: University of Chicago Press.

Shafto, P, and J Coley. 2003. “Development of Categorization and Reasoning in the Natural World:

Novices to Experts, Naive Similarity to Ecological Knowledge.” Journal of Experimental

Psychology: Learning, Memory & Cognition 29: 641–49.

Sheinberg, Esti. 2000. Irony, Satire, Parody and the Grotesque in the Music of Shostakovich: A Theory of

Musical Incongruities. Burlington: Ashgate.

Shelley, Cameron. 2001. “The Bicoherence Theory of Situational Irony.” Cognitive Science 25 (5): 775–

818.

Shinagawa, Patty. 2002. “Where No Fan Has Gone Before.” Futurama.


304

“Similarity, N.” 2015. Oxford English Dictionary Online. Oxford: Oxford University Press.

http://www.oed.com.turing.library.northwestern.edu/view/Entry/179875?redirectedFrom=

similarity.

Sisman, Elaine. 1993. Haydn and the Classical Variation. Studies in the History of Music 5. Cambridge,

Mass: Harvard University Press.

———. 2014. “Variations.” The Oxford Companion to Music. Oxford Music Online. Oxford University

Press. Accessed August 4.

http://www.oxfordmusiconline.com.turing.library.northwestern.edu/subscriber/article/opr

/t114/e7067.

Small, Christopher. 1998. Musicking: The Meanings of Performing and Listening. Music/culture. Hanover:

University Press of New England.

Snyder, Bob. 2009. “Memory for Music.” In The Oxford Handbook of Music Psychology, edited by Susan

Hallam, Ian Cross, and Michael Thaut, 107–18. Oxford: Oxford University Press.

Solomon, Maynard. 1977. Beethoven. New York: Schirmer Books.

Spellman, Barbara, and Keith Holyoak. 1996. “Pragmatics in Analogical Mapping.” Cognitive

Psychology 31 (3): 307–46. doi:10.1006/cogp.1996.0019.

Sperber, Dan, and Deirdre Wilson. 1986. Relevance: Communication and Cognition. The Language and

Thought Series. Cambridge: Harvard University Press.

Straus, Joseph. 2005. Introduction to Post-Tonal Theory. 3rd ed. Upper Saddle River: Prentice Hall.

Swain, Joseph. 1997. Musical Languages. 1st ed. New York: W.W. Norton.

Swinkin, Jeffrey. 2012. “Variation as Thematic Actualisation: The Case of Brahms’s Op. 9.” Music

Analysis 31 (1): 37–89. doi:10.1111/j.1468-2249.2012.00341.x.

Taruskin, Richard. 2005. The Oxford History of Western Music. Oxford: Oxford University Press.
305

Temperley, David. 1999. “The Question of Purpose in Music Theory: Description, Suggestion, and

Explanation.” Current Musicology 66: 66–85.

Thagard, Paul, and Cameron Shelley. 2001. “Emotional Analogies and Analogical Inference.” In The

Analogical Mind: Perspectives from Cognitive Science, edited by Dedre Gentner, Keith Holyoak, and

Boicho Kokinov, 335–62. Cambridge: MIT Press.

Tillmann, Barbara, and Emmanuel Bigand. 2004. “The Relative Importance of Local and Global

Structures in Music Perception.” The Journal of Aesthetics and Art Criticism 62 (2): 211–22.

doi:10.1111/j.1540-594X.2004.00153.x.

Tomasello, Michael. 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition.

Cambridge: Harvard University Press.

———. 2008. Origins of Human Communication. The Jean Nicod Lectures 2008. Cambridge: MIT

Press.

Tovey, Donald Francis. 1945. Beethoven. London: Oxford University Press.

———. 1976. A Companion to Beethoven’s Pianoforte Sonatas: Complete Analyses. New York: AMS Press.

Vendetti, Michael, Aaron Wu, and Keith Holyoak. 2014. “Far-Out Thinking: Generating Solutions

to Distant Analogies Promotes Relational Thinking.” Psychological Science 25 (4): 928–33.

doi:10.1177/0956797613518079.

Vogler, Georg Joseph. 1793. Verbesserung Der Forkel’schen Veränderungen Über “God Save the King.”

Frankfurt am Main: Varrentrapp und Wenner.

Watson, Angus. 2010. Beethoven’s Chamber Music in Context. Woodbridge: Boydell Press.

Wheelock, Gretchen. 1992. Haydn’s Ingenious Jesting with Art: Contexts of Musical Wit and Humor. New

York: Schirmer Books.

Wittgenstein, Ludwig. 1953. Philosophical Investigations. Oxford: Blackwell.


306

Yaraman, Sevin H. 2002. Revolving Embrace: The Waltz as Sex, Steps, and Sound. Hillsdale: Pendragon

Press.

Zbikowski, Lawrence. 2002. Conceptualizing Music: Cognitive Structure, Theory, and Analysis. Oxford:

Oxford University Press.

———. 2013. “Listening to Music.” In Speaking of Music, edited by Keith Chapin and Andrew Clark.

Bronx: Fordham University Press.

Zemach, Eddy, and Tamara Balter. 2007. “The Structure of Irony and How It Functions in Music.”

In Philosophers on Music: Experience, Meaning, and Work, 178–205. Oxford: Oxford University

Press.

Zenck, Claudia Maurer. 2008. “‘Mannichfaltige Abweichungen von Der Gewöhnlichen Sonaten-

Form’‘ Beethoven’s ’Piano Solo" Op. 31 No. 1 and the Challenge of Communication.” In

Communication in Eighteenth-Century Music, edited by Danuta Mirka and Kofi Agawu, 53–82.

Cambridge: Cambridge University Press.

Ziv, Naomi, and Zohar Eitan. 2007. “Themes as Prototypes: Similarity Judgments and

Categorization Tasks in Musical Contexts.” Musicae Scientiae 11 (1): 99–133.

doi:10.1177/1029864907011001051.

You might also like