The range of dictionary definitions of this most familiar of musical terms points to its essential, abiding ambiguity. Such definitions tend to proceed from textural to aesthetic interpretation: ‘combination of simultaneously sounded musical notes to produce chords and chord progressions’: ‘pleasing effect of apt arrangement of parts; agreement; concord’ (Pocket Oxford Dictionary). The ambiguity is partly aesthetic, in the implication that only pleasing concords can be properly harmonious, and partly textural, in that it seems possible to distinguish (in real music, not just technical exercises) between the harmonic ways in which simultaneous sounds relate and the contrapuntal ways in which successive sounds relate. While the entire history of music theory appears to depend on just such a distinction between harmony and *counterpoint, it is no less evident that developments in the nature of musical composition down the centuries have presumed the interdependence—at times amounting to integration, at other times a source of sustained tension—between the vertical and horizontal dimensions of musical space.
It seems natural and right that music which is not merely harmonic, but harmonious, should be highly regarded in civilized societies. Even though pure, unaccompanied vocal melody may historically have preceded it, and such music was the repository of particularly strong spiritual and lyrical qualities (plainchant, folksong), there is a clear correspondence between the concept of society as a mutually supportive commonwealth, and those manifestations of concert and theatre music which attract the collective approbation ‘civilized’. Collective performance, as in singing the same text to different but interdependent vocal lines, can be regarded as the musical correlate of civilized democracy. A hymn, or a national anthem, embodies the harmoniousness of shared rituals and beliefs, and certain instruments—organ, piano, guitar—are often known as ‘harmony instruments’, providing that chordal support without which the top-line melody on its own would seem incomplete. 20th-century developments in popular music helped to reinforce the role of basic, chordal harmonic support as an inseparable part of the musical statement, and of musical communication. Here, as in churches, the ‘harmony instruments’ may well be electronic.
2. History and theory to 1900
Since the time of the Ancient Greeks, writers on music have dealt with the acoustical nature of harmony alongside its physical and emotional effects. Discussions of how musical intervals occur, and of how particular scales or modes can be constructed, have a much longer history than studies of more than two notes in vertical combination: and it was more than a hundred years after those early 17th-century developments that brought the chordally accompanied music of recitative into the forefront of cultural practice before a major text with ‘harmony’ in its title was written. Nevertheless, the theoretical exploration of harmony began well before that text by Rameau (Traité de l'harmonie, 1722), just as it has continued after what many feel to be the last significant example of a comparable initiative—Schoenberg's Harmonielehre of 1911.
Such texts tend to share ideas and values which were closely bound up with the evolution of tonality and tonal composition between 1700 and 1900, from Monteverdi to Richard Strauss. The governing principle is that harmony embraces harmoniousness—that is, creates a pleasing effect—by observing certain compositionally established conventions. Those conventions derive from the perception that not all chords are the same and that combining different selections of notes of the scale creates different degrees of stability and instability, consonance and dissonance. The feeling that music communicates both its processes and effects through the distinction between consonance and dissonance, and through generally accepted ways of moving between these two types of harmony, had become well established before 1700, at a time when compositions were primarily contrapuntal.
But the shaping of musical forms and phrases in relation to the means of achieving convincing closure, or cadence, was a fruitful source of development in promoting harmonic theory, and the interaction between the two principles reached the height of technical sophistication and expressive power in the music of Handel and Bach. Study of the consistent procedures followed by these masters enabled Rameau, and later theorists, to codify harmonic practice, and the increasing 18th-century tendency to think of music as founded on a *fundamental bass of chordal roots, moving harmonic analysis away from the *figured bass, also contributed to the idea that rules of harmonic practice were not merely acceptable but desirable. Talented students were encouraged to proceed from simple technical exercises in the resolving of dissonances and the construction of coherent progressions to real, free composition in which those rules were still honoured, even when occasionally breached. Foremost among those rules was the celebrated prohibition of parallel movement in 5ths and octaves, aimed at promoting that degree of independence in part-writing which reinforces the interdependence of harmony and counterpoint.
3. History and theory since 1900
The potential for a realistic exploration of the nature of harmonic structures and functions in tonal composition was fully realized only in the 20th century when Heinrich Schenker gave priority to basic, background contrapuntal processes in determining those structures (see analysis (3)). The notion that the chordal progressions and relations on the musical surface, with all their particular doublings, spacings, and instrumental colourings, were less significant, less essential, than elemental two-voice fundamental structures was one of the most radical formulations of 20th-century music theory, and its presentation by Schenker was the more telling for his determination to restrict his focus to ‘masterworks’ of the Baroque, Classical, and Romantic canon. The no less radical alternative offered by *Schoenberg, that the surface was of the essence, and that—in contradiction of Schenker's elaborately wrought concepts of prolongation, of the contrapuntal elaboration of background harmony—there was no such thing as a ‘non-harmonic’ note, raised the prospect of categorizing every conceivable chordal simultaneity from three notes to 12 in terms of the disposition of intervals above the bass. But to function as the basis of effective pedagogical routine, harmonic theory always depended more on principles of progression than on strategies for chord classification.
4. Practice and principle
In many ways the most important prescription for tonal harmonic practice has to do with horizontal progression, not vertical disposition. Good harmony, the textbooks teach—and Bach's chorale harmonizations are held to demonstrate this to perfection—moves in such a way that the constituent voices of each individual chord balance independence against interaction (Ex. 1). To a remarkable extent, Baroque, Classical, and Romantic harmonic practice kept to essentially vocal ideas about acceptable principles of motion, even in a bass line that was understood to need to move more by leap than by step. And such dominant harmoniousness was reinforced by principles of rhythmic structuring and formal organization that emphasized regularity of phrase structuring as the norm and made clearly hierarchical distinctions between stronger, accented events within the prevailing metre and those weaker events which were inevitably perceived as moving between and therefore linking the strong ones in ways with which the hierarchies of tonal harmonic relationships could engage.
The devising of rule-based exercises in harmonic practice helped to reinforce the evident fact that the acceptance of harmonic laws imposed no particular stylistic constraints on the composer. Nor was it the case that composers moving from small-scale to large-scale works needed to move from harmonic conformity to harmonic freedom, or from one set of rules to another. A 20-minute symphonic movement by Mahler will range far more widely across the tonal spectrum than a ten-minute movement by Haydn, and will use a more elaborate, more chromatic harmonic palette with a higher proportion of dissonance to consonance; as the felt need to resolve the one on to the other relatively explicitly becomes increasingly eroded, the basic harmonic grammars of Haydn and Mahler are comparable—as, in the operatic sphere, are those of Handel and Verdi, Mozart and Richard Strauss. The idea that 19th-century composers were engaged in a systematic undermining of traditional harmonic principles, in order to ease Schoenberg's ultimate path to their destruction, is simplistic, since the total effect—in Wagner, above all, and also, more often than not, in Liszt, despite his late experiments—is to enhance the ultimate, delayed (but never to be doubted) arrival at consonant closure. The chromatic chords and dissonant strategies of the Romantic composers served in this way to enrich tonality, and such enrichments continued to be explored and used during the 20th century alongside more radical initiatives.
5. Harmony since 1900
Those radical initiatives are generally associated with moves from tonal to post-tonal harmony, and evolutionary models of music history explain this development in terms of the inevitability of change and the decline in concepts that had become over-familiar and—in some sense—worn out. As soon as it was possible for Mahler to end a work—Das Lied von der Erde (1908–9)—on a chord other than a pure, root-position triad (Ex. 2), or for Debussy in a piano piece from much the same time—Mouvement—to compose out a change from diatonic major tonality to whole-tone modality, then, logically, the evolutionary process could continue until tonality, and harmony, were replaced by their opposites. This model appeared to be supported by Schoenberg's ‘emancipation of the dissonance’. Nevertheless, in attempting to argue that absolute distinctions between consonance and dissonance were no longer meaningful, Schoenberg never sought to jettison the much more basic musical opposition, and alternative, between stability and instability. In ‘freeing’ the dissonance, he aimed to develop new ways of establishing coherence and stability by harmonic means: or, if it was felt that coherence could be achieved without harmonic stability, then stability, or consistency—the means whereby the music could have the potential for pleasing its listeners—must be introduced in other ways.
The simplest indication of such developments is found in a work like the third of Schoenberg's Five Orchestral Pieces op. 16 (1909), in which a single chord, dissonant by traditional standards, provides a contextually established standard of stability against which other events can be perceived as decorative, unstable (Ex. 3). Stylistically, such music is evidently not so far removed from the highly expressive continuities of late Romantic and Impressionist works which still use tonal harmony; and music since 1909 has varied greatly in the way it approaches the creation or rejection of stability and coherence in harmonic terms. Whereas many individual if not radical composers from Stravinsky to Schnittke use chords which are close enough to traditional types to set up associations with those types and their distinct functions as consonances or dissonances (even if divorced from the ‘functional’, scale-degree differentiation of truly tonal chords), others have succeeded in escaping from the need for coherence through harmony, understood as some kind of consistent chordal vocabulary. Such music (see for example Stockhausen's Klavierstück III, quoted in rhythm, Ex. 2) uses shapes, gestures, even contrapuntal relationships in ways that set up textural strategies independent of harmonic thinking. The sense of such music as extreme nevertheless offers the prospect of the judgment that an element of harmonic stability in the creation of compositional coherence could always remain a necessity for music if it is to please, and not just stimulate through provocation, more than a very small number of listeners: and this will be so even if the time-honoured harmonic distinction between consonance and dissonance is definitively set aside.