Introduction to the Theory of Layered Tonal Functions:

Melody, Accompaniment, and Other Layers

The distinction between melody and accompaniment is a given among musicians

involved in Western music. It comfortably inhabits our practical jargon, regardless of musical

style. And for the good reason that it is useful. It permits us to make simple reference to even the

most complex of musical textures by differentiating what we regard as the central part or main

line or primary voice from any and all others. These others, as a group, are considered to play a

supporting role.

This distinction seems most pertinent and useful when it comes to genres such as the

song or the sonata, where the singer or featured instrumentalist is accompanied by another

instrument (e.g., guitar or piano). There the centrality of the singer’s melody or, for example, the

violinist’s theme is amply evident.1 Songs and sonatas are especially good examples of melody

and accompaniment because the centrality of the musical ideas expressed by the singer and the

instrumentalist and the peripherality of the accompanimental ideas are reinforced by the physical

separation in space of the musicians. Furthermore, it is clear that some genres and musical

textures reinforce the distinction more clearly than others. The voice with piano accompaniment

seems a better fit than does, say, the string quartet; and a homophonic texture better than a

polyphonic one.

Nevertheless, the usefulness of the distinction is not lost in the latter cases. And this is so

because it has its basis in the different roles performed by melody and accompaniment in musical

1
Overlooked here are the finer distinctions of the relationship such as those times when the singer will fall
silent or the theme will be taken up by the accompanist. The melody can move between parts and players can
temporarily exchange roles without reducing the utility of the distinction.
works. In general, the role of the melody or main voice is to present and elaborate the

characteristic ideas of a work. The ideas succeed each other in a logical, coherent process. They

may be repeated, varied, or developed in some way, but generally this much is assumed:

whenever and wherever a part or voice seems to stand out from its surrounding texture, because

it presents characteristic ideas against other parts or voices that do not, or that do but to a lesser

degree, it is melodic and they are accompanimental in function.2

This notion that melody and accompaniment (and other musical entities) have roles or

functions that distinguish them was taken up and begun to be formalized in the theoretical work

of Arnold Schoenberg and some of his students. This is the point of departure. It is one of the

purposes of this paper to take a different perspective on the melodic-accompanimental

relationship, based in the Schoenbergian theory of formal functions. A different perspective will

enable the description of aspects of musical organization that heretofore have been insufficiently

engaged.

To describe a music in terms of melody and accompaniment is to do a kind of basic

analysis. It can be the first step in a more detailed exposition. But it should not be

underestimated: in certain cases it is critical to good description. The distinction is by no means

trivial or dispensable, for it refers to sonic arrangements with the potential to have expressive

consequences that are unique, often unanticipated. Two such arrangements, and examples of

each, will suffice to show this.

First, take the scenario of a composer or arranger that is making a set of variations on a

melody. Doing this the composer adopts a common variational strategy and ‘reharmonizes’ the

melody, intentionally changing the chords that accompany it, sometimes quite fancifully so. The

2
Clearly, if the distinction is to be maintained in changing circumstances, it requires the careful
interpretation of details – wherein the devil dwells.
result brings about a relationship between melody and chords that is (shall we say) much less

conventional or predictable than the original, ‘plain’ setting may have been. It might be

described as a ‘looser’ relationship between melody and accompaniment. Such a strategy also

introduces certain qualities to the music that come about precisely because of this looser

relationship. What is important to note is that both the relationship and the new qualities come

about through a compositional process that begins with a distinction: melody is distinguished

from accompaniment; melody is central and leading in the process; chords (however they may be

articulated) follow, taking a supportive role. The former presents the characteristic ideas and the

latter do not. The melody remains as it is; the accompanying chords get changed. Thus the

expressive consequences of this process are not limited to chord variation just for variation’s

sake but include those new qualities the variations introduce as byproducts. Example 1 presents

just such a reharmonization of a well-known melody.

I begin with some basic observations. Consider that whereas the traditional setting (not

shown) remains in one key for its entirety and its chromaticism is minimal, this setting touches

upon a number of keys (or at least suggests them) and is quite chromatic. The purely-diatonic

melody and chromatic chords reinforce the functional difference of melody and accompaniment.

Since the melody is familiar to us, and its affirmation of a single key remains unaltered, its

divergent reharmonization can express a quality that at some points borders on the bitonal (or

perhaps puts a toe across the line). What is certain is that a looser relationship exists between

melody and accompaniment; at some points they seem as if they are trying to come apart. At

those points they are still coordinating but it would be fair to say that they are not cooperating

tonally. This is a common arrangement in some styles and a great many cases could be cited.
Ex. 1 Oh, Christmas Tree

Now, for the next example, consider the scenario of an improvising group of jazz

musicians, playing a popular tune. Once the tune has been heard, one of the players begins to

improvise on it (in some manner) as the others either play accompanimental parts, including the

chord progression belonging to the tune, or drop out. Here again the distinction is apt, for the

soloist takes a role similar to that of the violinist playing a subordinate theme or a developmental

passage of a sonata: there is no doubt that the soloist is the main voice. And here too a degree of

looseness is introduced into the relationship, by the improvising soloist, that brings with it new
qualities. In this arrangement the accompanying chords remain as they are and the melodic line

gets changed.

A discussion of the expressive qualities that come about by this arrangement (though not

couched in that terminology) is given by Peter Winkler.3 Winkler discusses a trait of popular

music wherein melodies, improvised or composed, do not always follow the chord progression

(the “changes”). His discussion begins by considering a passage from a Count Basie solo on the

1937 Decca recording of Roseland Shuffle (transcribed with reduced chords in Ex. 2).

Ex. 2 Count Basie, Roseland Shuffle, Chorus 1, mm. 9-13

According to Winkler, at this point in the solo Basie “plays a simple blues lick in

polyrhythms; the pitches do not change and clearly clash with the underlying harmonies.”4 He

notes three other areas of the solo in which there are similar clashes. Winkler says “we could

attribute this to the vagaries of improvisation, but in fact one encounters similar effects in notated

3
Peter K. Winkler. “Toward a Theory of Popular Harmony.” In Theory Only 4, no. 2 (1978), 3-26.
4
Winkler, “Toward a Theory,” 16.
ragtime.”5 As an example, he cites the end of the first strain of Scott Joplin’s The Entertainer

(shown in Ex. 3).

Ex. 3 Scott Joplin, The Entertainer, 1st strain, mm. 13-16

He goes on to say:

Here the left hand outlines a descending progression while the right hand repeats an
ostinato figure centering around C and E. The blues are full of such examples: the
melodic line often hovers around the fifth or the tonic regardless of whether it is
supported by the harmonies. Does this mean that ragtime and jazz musicians are ignorant
or careless about the proper resolution of dissonances?6

Winkler answers his question in the negative and proceeds to give an explanation that

invokes Heinrich Schenker’s notion of structural levels. He refers to various notions from

Schenkerian theory, including ‘background’ and ‘foreground’ harmony.7 What is important to

note at this point is that his explanation does not begin by considering the functions of melody

and accompaniment; and it merely mentions the “clash” and “effects” (an approximation to my

“expressive consequences”) which are a dividend of the changed relationship of the two

functions.

So, in both arrangements, one with and one without improvisation, one side of the

melodic-accompanimental distinction remained fixed while the other was subject to variation.

5
Ibid.
6
Ibid.
7
Ibid.
And the variations brought looseness to the relationship, introducing singular, perhaps

unanticipated expressive qualities, indispensable to their genres.

The wealth of music that enacts processes similar to those described above testifies to the

desirability of such emergent expressive qualities. And they come about through a relationship

that has been described as ‘loose’ but that may be characterized otherwise. If melody and

accompaniment are imagined not in terms of the firmness of their coupling (as if an assessment

were being made of the elasticity of the bond that held them together), but in terms of their

differences owing to the roles they play, they may be characterized as being functionally

independent of one another. Being independent, they may not always cooperate tonally. They

may instead interact in any number of ways.

Because they act independently as to function, they may be varied independently, and

variations in one do not alter the function of the other (though the results of their interactions

vary). And though they may act in a coordinated manner (as do melody and accompaniment),

degrees of looseness in their coupling, introduced through variation, keep them from completely

cooperating. Thus, there exists between them a gap of sorts, across which their differences,

sensed through their interactions, become expressive of new and desirable qualities. The

changing affects of a fanciful reharmonization of a familiar melody or of a uniquely-personal

improvisation that is set against the chord progression of a popular tune, come about by the

functional independence and interaction of melody and accompaniment.

I now introduce a complication: suppose I want to make a finer distinction between parts

in a musical texture than the simple binary will permit. Where, for example, is there a role for

countermelody or for a form-determinative bass line in our melodic-accompanimental

distinction? How, for example, can the roles played by the principal parts in a Souza march be
characterized, once the countermelody begins to sound? Because there may be more than two

principal parts (distinguished by their formal roles) that need describing, a more general way to

refer to them is needed. I suggest the term layers. With this term parts that have a more leading

and melodic function and those that have a more following and accompanimental function may

be distinguished. And one is not limited (except by interest) as to the number distinguished. In

fact, provision is made for those layers that seem to function in a manner lying between the

central and the peripheral.

So, the melodic-accompanimental relationship—a layered relationship—has been chosen

as the starting place for a new analytical approach. And it has been chosen for good reasons.

First, melody and accompaniment are well-known commodities among musicians of all kinds.

Second, there are a great many pieces of music whose adequate description really requires little

more than examination of the interactions between these two layers (especially when

accompanimental parts seem to form a block). Third, they are clearly distinguished by their

formal roles. Hence, they have been chosen as the basic distinction with which to typify

functional layers generally. They are the black and white that flank many shades of gray.

Using this approach, the goal is to describe certain cases of tonal music, musics that are

generally regarded as ‘tonal’ (that is, that are organized around melodies or themes and their

accompaniments, in established keys, using traditional means such as scales, chords, key

relationships, and the like) that have steadfastly resisted adequate description using the means

offered by present music theory. Their resistance to analytical probing constitutes one of the

problems this theory engages. It is my position that these cases feature layered organization with

loose coupling. And I further suggest that such organization is one of the principal reasons why
investigations of them, utilizing the extant tools of our trade, have proved, so far, to be not

wholly satisfying. There are two reasons for this.

First, it has been observed that prevailing analytical approaches lack the design features

that would enable an exploration of the kinds of layered relationships discussed thus far. This

lack is due to certain trends in Western music theory that have tried discursively to reduce

music’s complexity and to demonstrate its unity or coherence. The result has been the production

of analytical approaches that rely upon unitary conceptions of some musical elements. Foremost

among these is pitch in tonal music. A unitary view of pitch is pertinent here because the pitched

components of tonal music include melody and accompaniment.

Furthermore, there has been a tendency to take highly-reductive approaches that, besides

unitary conceptions, rely upon scale (hierarchy) in order to try to get down to the (supposed)

fundamental pitch structures of music. Predictably, this trend in analysis ends up bypassing more

‘surface’ features like melody and accompaniment, those formal functions wherein loosely-

coupled layering exists.

For example, take the typical harmonic view of tonal organization. Pitch content and

organization are assumed to be entirely governed by the harmonies identified in analysis. Chords

in the accompaniment (derived by a reductive process) are lumped together with certain tones in

the melody and the two are thought to be consubstantial. Any tones left over are dispensed with

by their consignment to the catch-all category of ‘non-chord tones.’ Hence, melody has no

essence of its own; it is a derivative of harmony. Unitary harmony swallows up every distinction

based on pitch. Clearly, monolithic conceptions focused narrowly on pitch have not been

modulated by the distinctions of the formal functions of layers.

precisely differences of pitch content and organization. Any approach that proceeds by

consolidating pitch content (rather than differentiating formal functions) will only mask these

important differences. Moreover, the expressive qualities that arise from a loose coupling

between melodic and accompanimental layers are invisible to approaches wherein differences of

pitch content and organization are erased.

For such reasons, though unitary approaches may detect some manifestations of layered

organization, they are poorly-suited, or else unable, to accommodate them theoretically or

descriptively. Since they have not theorized them, they have not been aptly described by them.

Furthermore, the theory of formal functions stemming from Schoenberg has not (so far) been

explicitly developed in the vertical dimension. Thus, an approach that permits descriptions of the

independence and interaction of musical layers is lacking.

Caplin, William E. Classical Form: A Theory of Formal Functions for the Instrumental Music of
Haydn, Mozart, and Beethoven. New York: Oxford University Press, 1998.

Caplin, William E., James Hepokoski, and James Webster. Musical Form, Forms &
Formenlehre: Three Methodological Reflections. ed. Pieter Bergé. Leuven: Leuven
University Press, 2009.

Cherlin, Michael. “Dialectical Opposition in Schoenberg’s Music and Thought.” Music Theory
Spectrum 22, no. 2 (2000): 157-176.

Cross Ian, “Pitch Schemata,” in Perception and Cognition of Music, edited by Irène Deliège and
John Sloboda, 353-86. East Sussex, UK: Psychology Press Ltd., 1997.

Cuddy, Lola L. “Melody Comprehension and Tonal Structure,” in Psychology and Music: The
Understanding of Melody and Rhythm, edited by Thomas J. Tighe and W. Jay Dowling,
19-38. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 1993.

________. “Melodic Patterns and Tonal Structure: Converging Evidence.” Psychomusicology 10

________. and Betsy Badertscher. “Recovery of the Tonal Hierarchy: Some Comparisons
Across Age and Musical Experience.” Perception & Psychophysics 41 (1987): 609-20.

________. “Tonal Relations,” in Perception and Cognition of Music, edited by Irène Deliège and
John Sloboda, 329-52. East Sussex, UK: Psychology Press Ltd., 1997.

Deutsch, Diana, “Two Issues Concerning Tonal Hierarchies: Comment on Castellano, Barucha
and Krumhansl. Journal of Experimental Psychology: General 113 (1984): 413-416.

Madden, Charles. Fractals in Music: Introductory Mathematics for Musical Analysis. Salt Lake
City: High Art Press, 1999.

Montgomery, Kip. “Schenker and Schoenberg on Harmonic Tonality.” Indiana Theory Review
15, no. 1 (1994): 53-68.

Moore, Allan F. Rock, the Primary Text: Developing a Musicology of Rock. Ashgate popular and
folk music series. Aldershot, Hants, England: Ashgate, 2001.

________. Critical Essays in Popular Musicology. Aldershot, Hampshire, England: Ashgate,

________. “The So-called ‘Flattened Seventh’ in Rock.” Popular Music 14 (1995): 185-201.

Narmour, Eugene. The Analysis and Cognition of Basic Melodic Structures: The Implication-
Realization Model. Chicago: University of Chicago Press, 1990.

________. The Analysis and Cognition of Melodic Complexity: The Implication-Realization

Narveson, Paul R. Theory of Melody: A Complete General Presentation of the Practical

Patel, Aniruddh D. Music, Language, and the Brain. Oxford: Oxford University Press, 2008.

Peretz, Isabelle, and Robert J. Zatorre. The Cognitive Neuroscience of Music. Oxford: Oxford
University Press, 2003.

Povel, Dirk-Jan “A model for the perception of tonal melodies,” Music and Artificial Intelligence
Lecture Notes in Computer Science Volume 2445, 2002, pp 144-154

Povel, Dirk-Jan, E. Jansen. “Harmonic factors in the perception of tonal melodies” Music
Perception, Vol. 20, No. 1, Fall 2002, 51–85

Schenker, Heinrich. Harmony. Edited and annotated by Oswald Jonas. Translated by Elisabeth
Mann Borgese. Chicago: University of Chicago Press, 1954.

Schmuckler, Mark A. “Expectation in Music: Investigation of Melodic and Harmonic

Schoenberg, Arnold. Fundamentals of Musical Composition. ed. Gerald Strang and Leonard
Stein. London: Faber and Faber, 1967.

________. Harmonielehre. 3d ed., rev. Vienna: Universal Edition, 1922; English trans. by Roy
E. Carter as Theory of Harmony. Berkeley: University of California Press, 1978.

________. Structural Functions of Harmony. Rev. ed. Edited by Leonard Stein. New York: W.
W. Norton and Co., 1969.

________. Style and Idea: Selected Writings of Arnold Schoenberg. ed. Leonard Stein, trans. by
Leo Black, first paperback edition with revisions. Berkeley and Los Angeles: University
of California Press, 1984.

________, Arnold, Patricia Carpenter, and Severine Neff. The Musical Idea And The Logic,
Technique, And Art Of Its Presentation. New York: Columbia University Press, 1995.

Somer, Avo. “Musical Syntax in the Sonatas of Debussy: Phrase Structure and Formal
Function.” Music Theory Spectrum 27 (2005): 67-96.

Stephenson, Ken. What to Listen for in Rock: A Stylistic Analysis. New Haven: Yale University
Press, 2002.

Temperley, David. “The Melodic-Harmonic ‘Divorce’ in Rock.” Popular Music 26 (2007): 323-
342.

Winkler, Peter K. “Toward a Theory of Popular Harmony.” In Theory Only 4, no. 2 (1978): 3-26.