This article is about my use of revision and reworking to compose a trio of closely related pieces grouped under the collective title Dance Maze: Variations for Piano, Duos for Trumpet and Piano and Solos for Trumpet. Footnote ⁹³ Such grouping finds echoes in Pierre Boulez’s and Wolfgang Rihm’s families of genetically related works, Richard Barrett’s work cycles and the interlocking polyworks of Klaus Huber. My approach differs from most of these in harnessing techniques closely associated with another composer – those outlined by Tom Johnson in his book Self-Similar Melodies. Footnote ⁹⁴ Dance Maze began life in 1994 as a solo piano piece; it was revised once in 2008 and again in 2017,Footnote ⁹⁵ by which time the idea of creating a second version by adding a trumpet part had taken hold. The trumpet part was composed using techniques from Johnson’s book and was designed to be detachable, thus turning Duos into Solos. Table 1 summarizes the form of Duos (a mobile structure in which the 15 sections may be performed in any order); shows its derivation from Variations; and refers the reader to the pages in Self-Similar Melodies used to compose the trumpet part. In the rest of this article I position my revisions and reworkings in relation to other composers’ practices; I explain in detail some of the changes made to the original piano piece (confining my comments to Variations and Duos);Footnote ⁹⁶ and I briefly discuss what drew me to revisit a work from much earlier in my output. My method might be described as automusicological: it has involved detailed study of the sketches for Duos alongside reference to written and recorded reflection from the period of its composition.Footnote ⁹⁷

The procedures I have employed to fashion Variations and Duos encompass what the Boulez scholar Joseph Salem describes as ‘basic revisions’ and ‘obvious siblings’.Footnote ⁹⁸ The former preserve the original title and make relatively small-scale changes resulting in an updated version; the latter modify the title (Boulez’s Anthèmes, for example, became Anthèmes 2), make significant changes (in Boulez’s case, various types of expansion) and usually leave the earlier version intact rather than replacing it.Footnote ⁹⁹ The way in which Duos audibly incorporates Variations is also a form of self-borrowing. Leta Miller uses this term in relation to Lou Harrison’s practice of plundering his catalogue for reusable material and associates it with the eclecticism that was so much a part of the composer’s style. I prefer to describe my self-borrowing as reworking because, despite drawing on another composer’s music, Duos is stylistically more uniform than the former term might imply.Footnote ¹⁰⁰

I will now provide some examples to explain how my processes of revision and reworking were carried out. Examples 1–3 show three versions of the same passage from Dance Maze. The revisions in Example 2 concern idiom; the performer is now given time to leap between the chords and the octave unison figuration. Example 3 shows a change of character, with a reduction in dynamic to pianissimo and much more separation between the chords (now laissez vibrer) and the figuration. This second revision, halting the forward momentum of the passage by separating out its constituent figures, hints at the deep-seated aesthetic changes manifested in Duos via the reworkings in Table 1.

Example 1 Tom Armstrong, Dance Maze, 1994 version, bars 31–8.

Example 2 Tom Armstrong, Dance Maze, 2008 version, bars 65–73.

Example 3 Tom Armstrong, Dance Maze: Variations for Piano, 2017, bars 65–79.

I will focus on two reworkings to highlight the contrasting ways in which they utilize Tom Johnson’s processes – contrast marked by the degree to which the result of each process is interfered with. Infinite Automaton is based on the transformation n → n, n + 1, n + 1,Footnote ¹⁰¹ the first level of which is seen in bar 2 of Example 4. Johnson modifies the transformation in two ways: subsequent transformations are applied only to newly added notes, and new notes are twice the duration of old ones. Infinite Automaton borrows Johnson’s process intact, differing from his example only in the whole-tone mapping and extension to the sixth level of the transformation.Footnote ¹⁰² Dragon Curve no. 9, on the other hand, reveals considerable intervention in the outcome of its generative process. This process is a version of John Heighway’s paper-folding fractalFootnote ¹⁰³ – his study of what occurs when a piece of paper is folded rightwards again and again – in which Johnson converts the left and right turns (revealed when the folds are opened to 90 degrees) into a sequence of ones and zeros. Plotted as ascending and descending scale steps, a series of phrases are produced that, starting from the same note, alternately rise or fall in a sequence of trills.Footnote ¹⁰⁴ Example 5 (from Duos) shows that the melodic curve of Johnson’s melody is drastically altered: rather than radiating upwards and downwards from the same pitch, the descending phrases are transposed two octaves higher, creating jagged breaks in the line that disrupt the melody’s smooth unwinding. Rhythmically, the regular crotchets of Johnson’s example are interrupted by pairs of quavers that mark the change to a new two-note group. The piano part is treated differently in each of these passages: Infinite Automaton maintains it intact from Variations – the level of autonomy between the instruments here creates a state of near indifference; in Dragon Curve no. 9 the piano is assimilated to the trumpet’s material and the continuous pulses of Variations (see Example 6) are disturbed by aperiodic rests.

Example 4 Tom Armstrong, Infinite Automaton, bars 1–6, from Dance Maze: Duos for Trumpet and Piano.

Example 5 Tom Armstrong, Dragon Curve no. 9, bars 15–32, from Dance Maze: Duos for Trumpet and Piano.

Example 6 Tom Armstrong, Dance Maze: Variations for Piano, bars 215–31.

The main technique used to rework Variations to create Duos is overpainting, a procedure found frequently in Rihm’s music, particularly his fleuve, Formen and Seraphim collections.Footnote ¹⁰⁵ As in Rihm, the overpainted material sometimes alters or degrades the base layer as was observed in Dragon Curve no. 9. But in contrast to Rihm, the new material is derived explicitly from the work of another composer – Johnson; Self-Similar Melodies is, after all, a kind of composition ‘manual’ that provides generous links to the author’s own music for the reader to follow up.Footnote ¹⁰⁶ As in visual art practice, overpainting changes the form of the target work. In Dragon Curve no. 9 the non-aligned trumpet part (see Table 1) extends beyond the conclusion of the piano’s music, dissipating the tension that has accrued rather than, as in Variations, harnessing it to propel the music into the next section. Here, on a larger scale, is the same attenuation of momentum observed in the revisions of Variations (Examples 1–3); such a lessening of forward drive in Duos represents a significant shift in aesthetic values between the two works that points to my reasons for undertaking such a reworking.

My decision to rework Dance Maze was driven by my own critical response to the 1994 and 2008 versions. The revision discussed earlier (Example 3) was a ‘reaction against the rather “bashy” nature of the original, even in the [2008] revised version’,Footnote ¹⁰⁷ while the later overpainting of trumpet material would ‘override the formal divisions in the piano part’ and address the ‘short breathed problem I’ve noticed in a few of my pieces’.Footnote ¹⁰⁸ It is common for composers’ revisions to be prompted in these ways, but such critiques can move beyond matters of style and technique to embrace deeper aesthetic changes. Miller shows how Harrison’s ‘compulsive retrospection’Footnote ¹⁰⁹ reached back beyond his 1947 nervous breakdown and the re-evaluation of his compositional language this occasioned; for example, the Largo ostinato (1937) was reworked as a movement of the Third Symphony (1982) with the upper parts changed significantly to reflect Harrison’s study of the Chinese cheng in the 1960s. Salem’s study of Boulez’s long decade (1948–62) is predicated on the mismatch between the composer’s aesthetics (as revealed in his early writings) and the music of formative works of the period, notably Le marteau sans maître, Structures 2 and Pli selon pli. Footnote ¹¹⁰ Boulez’s later writings, such as Boulez on Music Today, fail in Salem’s view to catch up with and adequately explain the aesthetic changes in his mature music. These changes involve a move away from an organicist aesthetics of growth towards one of proliferation involving ‘self borrowing, transcription and “open” compositional structures’.Footnote ¹¹¹ As Salem shows, Boulez’s works may be open in a variety of ways: revised over many years; sprouting from a single kernel of material; or possessing formal mobility. In my own Duos it is mobile form that is the vehicle for aesthetic critique; the fixed, regulatedFootnote ¹¹² form of Variations is ‘exploded’ into 15 discreet sections (see again Table 1) that can be played in any sequenceFootnote ¹¹³ – my reworking rearticulates an ordered whole as a disordered collection.

Duos critiques not only the integrity of a particular work, but also (through its openness) the concept of the musical work itself. Luciano Berio viewed the open work as a path to ‘recovering an ephemeral, lucid, and transitory dimension of human experience […] and educating us instead to think of the work as an agglomeration of events without any prearranged centre’Footnote ¹¹⁴ between which connections are local and not determined by formal a priori elements. A similar (that is, less singular) idea of the work is advanced by Roger Parker in his discussion of Donizetti’s sketches for ‘Al suo piè cader vogl’io’ from Adelia, in which the existence of different drafts (each perfectly viable) reminds us not to confine composer intention only to the production of authentic and final versions.Footnote ¹¹⁵ Parker further suggests, citing the interpretative interventions of the soprano on the Adelia recording,Footnote ¹¹⁶ that we should attend to the way a musical work changes over time and the way we ourselves change with it: ‘Perhaps we will ask new questions, find new meanings in objects we had once thought too familiar to excite us further.’Footnote ¹¹⁷ This leads back to revision – to revisit a piece from years past, as was the case with Variations, is to revisit a version of yourself and, in the process, to realize not only how you have changed, but how subjectivity is contingent and constructed. Perhaps, more prosaically, it is also to realize that a piece of music can always go another way.