On Musical Ideas


It began for him always with the vision of some person or persons, who hovered before him, soliciting him, as the active or passive figure, interesting him and appealing to him just as they were and by what they were. He saw them, in that fashion, as disponibles, saw them subject to the chances, the complications of existence, and saw them vividly, but then had to find for them the right relations, those that would most bring them out; to imagine, to invent and select and piece together the situations most useful and favorable to the sense of the creatures themselves, the complications they would be most likely to produce and to feel. (Henry James, in the preface to Portrait of a Lady, referring to Ivan Turgeniev.)



In his essay entitled “New Music, Outmoded Music, Style and Idea, written in 1946,  Schoenberg begins with the following statement (Style and Idea, Selected Writings of Arnold Schoenberg, edited by Leonard Stein, London, Faber and Faber, 1975, p. 113, hereafter S+I):

The first three of these four concepts [i.e., in the title, (AB)] have been widely used in the last twenty-five years, while not so much ado has been made about the fourth, Idea.

Sixty years later the situation has not changed substantially: One still reads endless discussions about what is currently considered modern or old-fashioned, about questions of style, but rarely any discussion of  musical ideas as such.  To be clear: we are not talking here of ideas about music, but of ideas in music, that is to say, in pitches, timbres, rhythms, etc. .

The reasons for this are easy to discern. First, discussion of musical ideas in this sense requires far more practical experience writing music than many commentators possess. Second, it is less amenable to the aesthetic propaganda and feuds which, unfortunately, often fuel certain kinds of writing about music.

But for a real composer, once his craft is solid, thinking about the quality and implications of his musical ideas is central to his art.  However these questions are not resolved through verbal discussion, but by writing music, listening to music, trying out musical possibilities – in short, through musical thinking.

In this essay,  I hope to clarify for others a process which is fundamental to composition, but too often ignored. (And also: often lacking in the teaching of composition.) My goal is pedagogical, not philosophical: I want to complement existing literature on compositional technique, adding some material not generally known. I will not propose any comprehensive theories. Given the dearth of writing on this subject, I shall simply attempt to make a useful contribution.

As in my other online books, my discussion here is limited to western art music.  I will also avoid the issues of program music and music with text, except to note that even in these cases, only a musical idea (whatever its inspiration) can ultimately provide the impulse for a primarily musical work.

Our discussion will focus on two issues: What makes for interest in a musical idea? In what sense does the idea determine the form of the piece? These questions are not independent of one another: In fact, it is often impossible to separate a musical idea from how it is worked out.

Literature on this subject is rare. Apart from the Schoenberg essay referred to above, and some other writings by him, discussed below in more detail,  and the small book by Roger Sessions (The Musical Experience of Composer, Performer, Listener, Atheneum, Now York, 1968, p. 43 ff., hereafter “ME), the best sources are composers’ musical sketches.  Examining preliminary sketches by masters like Bach, Beethoven, Wagner, etc., and observing their progress towards their final form (as they appear in the piece),  one can learn some of these masters’ criteria for musical quality.

As Sessions puts it, talking about Beethoven:

A special fascination of his sketches is, indeed, the manner in which the various transformations of an idea always preserve and in fact progressively intensify the essential characteristics of the idea. (ME, p. 54.)

The analytical/theoretical literature rarely has much to say about criteria for the quality of musical ideas, and even less about how a given idea influences and/or determines the form of the resulting piece, except in the obvious sense of motivic derivations, and, in some more recent music, harmonic commonalities between lines and chords.  Interesting though these connections may be, they usually say little or nothing about how or why the work attracts the listener. For the composer however, finding the presentation which will have the most impact on the listener is of capital importance.

As a preliminary example, while any number of analyses have been made of the fugues in Bach’s Well Tempered Clavier, almost none of them discuss how the large design of a given fugue is related to its subject.  Motivic derivation from the theme is of course important, and usually easy to see, but motivic derivations say nothing about the temporal organization of the piece. While it would be exaggerated to propose that the entire form of a given fugue can be somehow deduced from its subject, there are definite links between the character and internal structure of a theme and the ensuing fugue’s construction.

For example, the theme and countersubject of the great B minor fugue, which ends Book 1 of the WTC, imply, due to their length and complexity, that the piece which follows will also be fairly long and elaborate, that its harmony will contain many chromatic appoggiaturas and chromatic modulations. The augmented and diminished leaps, so prominent in the subject and countersubject, will necessarily engender jagged, angular counterpoint. This material suggests an intense and very dramatic fugue.  It will also, as contrast, likely require some sort of simpler sequential episode, perhaps more diatonic, possibly returning several times in invertible counterpoint, as is Bach’s habit, to allow the overall form to “breathe”.

For a very different case, let us take the first fugue  in the same volume of the WTC, in C major.  This subject  is immediately recognizable as susceptible to numerous canonic imitations; its smoothness of line and straightforward diatonicism suggest a calmer style; it will thus require much less in the way of strong contrast: Such contrast here could overshadow the theme.  This short subject also does not suggest a very long piece.
Lest these “deductions” seem obvious or trivial, it is worth reminding the reader that several of the main traditions of teaching fugue, e.g. the French fugue d’école,  require all fugues to follow the exact same plan of entries, modulations, episodes, stretti, etc.. While such a prefabricated plan may be of some use to a beginner, the notion that all subjects should be worked out in exactly the same way is a travesty of real, imaginative composition.  By the same token, asking students to write a “standard” sonata form, without reference to what the material requires, is a mechanical task, and cannot lead to artistic thinking.
In contrast to these routine procedures, the strongest and most personal compositions always evince a certain force and individuality. They have striking, memorable, ideas, and they exploit what is most interesting and characteristic in those ideas to the maximum, thus engendering many critical decisions about what will be most salient in the ensuing form.
The Musical Idea: Basics
Defining what constitutes a musical idea is not simple, in part because it is impossible to predict what future composers will invent. Here we will suggest some lines of thought which have proven useful to composers in the past.
Let us begin with Schoenberg’s writings on the subject. His interest in this problem covers most of his career. He refers to it, here and there, very frequently. Apart from the essay quoted above, Schoenberg discusses the musical idea in more detail in his works on musical form (implying, quite logically, that a musical idea cannot be separated from its formal presentation). His two main works, available to us, are:

  • a recent compilation from a planned, but never completed, book, The Musical Idea, and the Logic, Technique, and Art of its Presentation, edited, translated, and with a commentary by Patricia Carpenter and Severine Neff, Columbia UP, NY, 1995. (hereafter TMI)
  • his textbook for American students, Fundamentals of Musical Composition, edited by Gerald Strang and Leonard Stein. (Faber, London, 1967)

While our purpose  here is not primarily to critique Schoenberg’s writing, it is important to recall his points of reference. According to Richard Taruskin, in The Oxford History of Western Music (Oxford University Press, New York, 2005, vol. 4, p. 52 ff), Schoenberg was strongly influenced by the ideas of Swedenborg, concerning mystic unity. Carpenter and Neff also mention frequently that Schoenberg tended to see musical forms in organic terms: They call this his “organicism” (TMI, p. 121).
Despite occasional tantalizing hints (for example at one point in TMI there is an empty page headed “Sonority as a formative element – and means of coherence), and his own experimental composition based mainly on fluctuating tone colors, Farben, op. 16, Schoenberg, in most of his discussions and analyses, follows the late romantic tradition and uses the term “musical idea to mean a thematic/motivic Grundgestalt (basic form), from which all the subsequent “motive-forms in a work are derived.
There is of course a good reason for this concentration on motives: By definition, a motive is a short, easily memorable musical unit. Memory being the mental faculty which enables us to perceive musical form in the first place, that which is easily memorable is a logical building block for musical form.
Further, motives are susceptible to many degrees of variation, ranging from the most obvious and elementary changes, to extremely esoteric transformations, only accessible to the eye. Of course the aspects of a musical idea which are easily audible must always have primacy. No piece is successful only or even mainly because of subtle motivic connections: By definition, these are not memorable! While a connoisseur’s experience may perhaps be enhanced by recognizing such subtler connections, they can never take the place of salient, audible structure,  since the latter is what the listener normally follows.  Put in another way, if the musical idea of a piece is not made manifest clearly and audibly, it can have no significant effect on the listener. Subtler details must enhance the effect of what is most salient if they are to be of any use at all.
Our insistence on the importance of salience perhaps makes this the place to comment on notions such as “athematic music”. While it is indeed possible to construct music without themes of the traditional type, slogans like this are deeply misleading. For one thing, they describe what the music is not, instead of what it is. Second, they treat themes like superficial features which can be removed with no effect.  If this is true, why use them in the first place? And if one is not to use themes, why not? And, most important, what will take their place in the musical architecture?
To return to Schoenberg:  Being a restless, questioning thinker, he conceded some of the problems with his excessively motivic focus. He states, in TMI:

So it is with my assertion that at least a movement if not an entire work, is formed from single motive. I could demonstrate this assertion with many examples, but many others stubbornly resist such elucidation: here I no longer see, or here there are simply other laws at work that are unknown to me. (p. 91)

Perhaps because of this problem, Schoenberg (elsewhere) proposes the following notions:

Every tone which is added to a beginning tone makes the meaning of that tone doubtful. (S+I, p.123)


Through the connection of tones of different pitch, duration and stress (intensity???) [AB: the questions marks are Schoenberg’s], an unrest comes into being: a state of rest is placed in question through a contrast. (TMI, p. 103)

The statement is a rather nice formulation, as it implies that the beginning of a piece creates tensions, which the rest of piece will work out or resolve.
However I will here adopt a slightly different approach.  I regard a good beginning as one which creates questions for the listener.  By “question”, I mean a combination of expectation and uncertainty which draws in the listener. The balance between these elements is critical: Too much predictability leads to boredom; too much uncertainty is simply chaotic, making it impossible for the listener to make meaningful connections between events, and thus to develop any expectations about continuation.
Some examples will clarify these points, and will also demonstrate how one can improve even a simple musical idea.
Ex. 1 is a simple C major scale, in steady quarter notes. Musically it is of no great interest, because the listener is left with no desire to continue, after the tonal and rhythmic resolution at the final note.

Ex. 2 increases the disorganization considerably: Apart from the equal note values, and the notes being limited to a C major scale, the notes have no discernible pattern.  This example is also of little musical interest, since it creates few coherent expectations.

Ex. 3 is similar to example 2, but in the realm of rhythm. Again, the result is simply too unpredictable; making it hard to remember, and virtually impossible for the listener to perceive coherent relationships.

Ex. 4  keeps the simple overall direction (rising C major scale), but now the details of the path upward are much less obvious.  By allowing the listener to develop some expectations about the continuation at each point (equal notes, rising line, notes from the C major scale), while adding a mild degree of unpredictability, this example gains some musical interest.

Ex. 5  takes a large step towards musically intensifying the idea.  With definite tempo and dynamics now specified, and an easily recognizable motive, resulting from repeated notes and rhythmic patterns, as well as consistent articulation, it makes a stronger, more memorable impression on the listener. Also, with its rhythmic stop at the end of bar 1, followed by renewed momentum, this example has a strong internal contrast, without becoming chaotic.  This version begins to approach a real theme: It demands continuation.

Ex. 6  makes the theme, already implicit in Ex. 5, more personal.  Instead of an anonymous C major scale, the rising profile now has a more distinctive intervallic character.  The added repetitions make the idea more memorable as well. While the rhythmic pulsation is still felt, the irregular spacing of the motive statements adds extra tension.  This version not only demands continuation, but also suggests more possibilities for dramatic development, as a result of the its internal harmonic and rhythmic tensions.

Let us explore a few of the possibilities of the idea in this form:

  • The silence in  m. 2 could subsequently take different turns: It could lead to a repeat of the beginning, it could move to other harmonic regions, it could change character unexpectedly – an even more potent device if the listener has come to know the original form of the theme and to expect the continuation already given.
  • The pause also suggests the possibility of canonic writing, imitation; or, it could be set in dialogue with itself, in various timbres and/or registers.
  • The repeated notes make inversion a useful and clearly audible derivation.
  • The theme could be presented, as a whole or in part, in another character; e.g. scherzando, by changing tempo and articulation.  (This is rather like a character in a novel, whom we encounter in different moods and situations.)
  • One could concentrate on the repeated note figure to create a kind of stubborn pedal tone, accompanying something else.
  • The leaps could be enlarged, and thereby made more dramatic.

I have deliberately pointed here to some of the less obvious possibilities, i.e. other than simply continuing to make another phrase or two in the same vein, since to make a large piece requires strong contrasts, and interesting turns to the form. Notice how many possibilities come from the fact that the theme’s overall contour is not smooth.  Such breaks allow for surprises.
Ex.  7  takes the theme still farther in the realm of internal variety.  Now the timbre is also defined, reinforcing the contrast. While based on the same contour as the previous examples, the stronger contrast here gives a sense that the work feature dramatic conflict; it will also probably  require more time to work out the tensions of the theme.

Our discussion thus far suggests a few useful criteria for evaluating a (thematic) musical idea. Criteria for such ideas depend, first, on the kind of piece the composer wants to write: A waltz requires a different kind of idea from a symphonic finale. In particular, the size and emotional range of the piece normally are connected to the degree of internal contrast in the idea:  The richer the internal contrasts, the more scope the piece will usually require to connect them convincingly, and to work them out fully.
Second, a musical idea should be memorable, and thus relatively easy for the listener to learn. It must be simple enough to grasp quickly (i.e. not overloaded with diverse information). Usually, this implies some fairly obvious repetition.
Third, it must intrigue the listener: As mentioned above, this requires a balance of predictability in its main outline, and unpredictability in its details.
Non-thematic Ideas
But what of musical ideas which are not limited to themes and motives?  Again, Schoenberg:

I myself consider the totality of a piece as the idea.  (S+I, p. 123)

This is suggestive but vague. Indeed, although Schoenberg spends great deal of time discussing formal processes which apply to different parts of the piece (exposition, development, transition, etc.) , he says nothing specific about how they vary depending on a particular idea.  But this is a critical question.
Perhaps a better way to formulate Schonberg’s notion would be: The idea and its working out are inseparable. Some aspects of the working out are easily deduced from the starting point (theme/motive), whereas others may not be. Finding the right way to throw an idea into relief, how to lead into it most effectively, how to show it in all its facets – these are all parts of composing itself.  Indeed, they take us back to the etymological root of the word: “componere, to place together”.

Sometimes […] one of the most important musical ideas, in a fundamental and motivating sense, may be not even a thematic fragment at all, but some feature of the large design […] (ME, p. 44-5).

One of the common and significant ways a musical idea can develop I call “taking a new turn. (Note that this technique does not preclude using themes/motives.)
Let us take a very well known example: the first movement of Beethoven’s Fifth Symphony.  At m. 21, the orchestra stops abruptly, leaving the first violins trailing behind with one single note; it almost sounds like a mistake.  However this gesture has important consequences in the recapitulation.  The opening material is now accompanied by more sustained wind parts, culminating in that same held note (m. 268), but now in the oboe. This flowers into a little solo phrase, adagio. This unexpected contrast creates an emotional deepening, especially after the long, incessant hammering of the repeated eighth note motive. The fact that this moment has been prepared by the very salient “question”, posed first in m. 21 by the violins, makes its meaning more profound: It stimulates the memory at the same time as it proposes a new musical and emotional direction. This very salient musical idea, even if it is not part any Grundgestalt, enormously increases the emotional richness, and unity, of the movement.
Another, different kind of example: the slow movement of Beethoven’s Piano Concerto #4. This is an interesting case, because the idea here is an alternation of characters, which in turn evolve over the whole movement.  At the start, the orchestra (strings only) plays a rough, staccato, dotted note figure, loudly, in octaves. The piano, by contrast, is sustained, lyrical, its sounds dying away sadly (a wonderful example of Beethoven using something very basic – the piano’s naturally decreasing resonance – to maximum effect). As the piano’s harmony becomes more plaintive, adding appoggiaturas, etc., the orchestral phrases get shorter and quieter, finally reduced to little two note interjections. When the orchestra finally does start playing chords, they are pizzicato, accompanying the piano’s more florid writing, now with arpeggios as well as appoggiaturas. After a short cadenza in the piano, the orchestra’s final statement is at last harmonized, sustained, arco, with the original dotted motive now relegated to the bass (this observation is Tovey’s, from his Essays in Musical Analysis, Vol. 3, Oxford University Press, London, 1963, p. 81). Piano and orchestra at the end share a simple motive. The effect, overall, is of the orchestra (the group) finally coming into sympathy with the piano (the individual); the emotional impact is enormous. The first confrontation is already potent, but the evolution from confrontation to sympathy is overwhelming.  The idea of this movement is thus not limited to its motivic construction, but rather – fittingly for a concerto – in the way the soloist and the orchestra interact.
Yet another example, with a very different kind of working out: Bach’s late E minor Prelude and fugue for Organ (the “Wedge).  The fugue subject is harmonically rich, and melodically symmetrical, although not mechanically so. (The countersubject is the lower voice here.)

This is developed for about two pages, in normal fugal fashion – if the word “normal” can be applied to counterpoint of such power.  As one would expect from the compound chromatic line in the subject, the harmony is restless, exploring the chromatic possibilities inherent in the subject, and the lines build in intensity to a powerful climax.
Then, with its final chord, the fugue dissolves into toccata-like figuration. Unlike the work’s historical precedents (e.g. Buxtehude), here there is no pause, no change of tempo, corresponding to this radical change of texture; We are clearly not starting another movement. The new section however goes on to astonish. Bach now alternates between the new, less contrapuntal figuration, and fragments of the preceding fugue, with concerto-like drama. The contrast between the two becomes dynamic, rather than static. This middle section reaches a huge climax on a dominant pedal, which, since it already is using material from the original fugue, leads easily back to a literal repeat of same, rounding off the movement.
Such a hybrid movement is not in any sense implied by the fugue subject.  It is as if Bach said “What if ….   and, seeing a possibility for architecture on a grand scale, simply allowed it to proceed. The really hard part of composing such a hybrid, however, is getting the balance right between the various elements. For example, note how:

  • the alternations between the two textures occur always at significant moments: cadences or climaxes.
  • the pacing of the climaxes is perfectly proportioned.
  • the simple symmetry of the da capo ending is exactly the right gesture to crown off the dramatic richness of the whole.
  • making the Prelude out of a (more normal) concerto-like structure, makes it substantial enough for  this gigantic fugue.

All these decisions testify to the architectural imagination of a great master. Anybody who tries to imagine the continuation from any point in the middle of the work, quickly discovers how un-obvious it all is. Only a great composer could have written the opening fugal section; the additional development via the huge toccata section is of an intensity, and a daring, which only a Bach could have achieved.
The idea for the toccata section, while not “derivable” from the fugue subject, becomes so powerful in its working out that it takes control of the form, creates an emotional and musical richness otherwise unthinkable, and does so in a way which seems, in retrospect, utterly inevitable. (We are told that Bach would listen to improvisers’ themes and then predict what they would do with them; surely this piece goes far beyond any such predictions!)
Does even this represent the complete working out of these ideas? Possibly not: There are many examples in Bach of already great works being recast in different forms (violin music becoming keyboard music, keyboard music becoming cantatas, etc.) which suggest that if Bach had had the inclination, or the occasion, he could probably have made yet something else of this, possibly even grander.  My point here is that the derivations from a Grundgestalt are only one,  fairly obvious aspect of this masterpiece; the other aspects of it are much harder to categorize.
Such examples could easily be multiplied, but they already show what is possible when one thinks in terms of simple relationships between salient moments, over fairly long time spans, rather than limiting oneself to considerations of motivic or harmonic derivation. Such ideas are specific, in the sense that they cannot be transferred to any other musical material:  The composer sees an enlarging possibility and it becomes a highlight of the piece.
Leaving aside the notion of a  Grundgestalt, or at least accepting its diminished importance, leads us to a very important conclusion:  In musical form, since it is temporal and necessarily sequential, connection is more important than derivation.  The way things are prepared and led up to – transitions, in a word – has everything to do with how effective they are.  Although I have written elsewhere on this subject (see my book on musical form), it is worth repeating here. Temporarily putting aside the elaborate baggage of thematic/harmonic derivation allows the composer to think about the destiny of his musical ideas in powerful and moving ways.
Sessions puts it well:

The musical idea […] is, then, the element which gives the music its essential character, and I have also referred to it as “the starting point of a vital musical train of thought”.    (ME, p. 52-3)

This kind of thinking is not rare in practice, but why is it so rarely a part of composition pedagogy?
One last point:  The German film-maker Werner Herzog states (Herzog on Herzog, Edited by Paul Cronin, Faber and Faber, London, 2002, p. 163):

It is always the mysterious and those things which do not perfectly fit into a story – the inexplicable images or twists in the tale – that stick out and are memorable. Sometimes I will place a scene or a shot into a film that might seem to have no place, yet that it essential to our understanding of the story being told.

Herzog’s preoccupation with mystery, as often engendering what is most memorable, is close to my heart. Of course, the key phrase here is “might seem to have no place”.  The difficulty is in finding something that seems unexpected, but which nonetheless ultimately proves to fit in, and, indeed, will enrich the whole.  Although the first impression is one of surprise, or even disorientation, subsequently – and here is where the choice of the added element is critical – one sees that it is “right” after all. Occasionally such events can even be discovered by chance, although without careful selection for effect, they will add nothing at all, or even detract.  Still, thinking over many examples of music I would call sublime, there is always this element of surprise, this sense that one would not have thought that the work, as heretofore presented, could contain this.
For a musical example, one thinks of a special category of musical idea, which engenders just such a sense of mystery: the ending as question. In pieces like Chopin’s F+ Prelude, or Schumann’s Child Falling Asleep (op. 15), the harmonically inconclusive ending leaves us unexpectedly, and magically, suspended.
There are many more examples of non-thematic musical ideas in the standard repertoire.  And often they are the moments we most remember, the one which haunt us, which stay with us and seduce us.
To conclude, some criteria for quality. If music is to communicate significantly, it needs ideas which stay in the memory.  These ideas need to be presented, and developed, in ways which throw them into maximum relief, and make their character as strong as possible.  Richness, in a musical idea, means that the idea has many, varied implications, and, usually, engenders a large work, of sustained musical thought and substantial emotional range. Details should all contribute to making the main ideas vivid and memorable. This does not mean that they must be derived from one common source, but rather that they must add something significant to the character and evolution of the work. Again, musical connections – transitions, the way the composer puts his idea in context – are far more important than derivations: A powerful contrast may make an idea far more vivid than a simple motivic or harmonic derivation. And finally, originality and personality (usually taken as indicators of quality) have less to do with surface strangeness than with the expressive impact of the work, taken as a whole. (Indeed, a piece whose language may seem superficially conventional may in fact work out in really unusual and fascinating ways; one thinks, for example, of Sibelius. This of course does not mean that what sounds conventional is necessarily good, but simply that one must dig deeper to see if the composer really has something valuable to say.)  Once again: In the final analysis, the whole piece is the idea. And the salient details of the listener’s voyage through musical time constitute its message.