In the String Trio op.20, completed in 1927, Webern was clearly master of the method of composition with 12 notes related only to each other. This work was a turning point in many ways. It represented a return to instrumental writing after 13 years in which the only works Webern felt to be worthy of opus numbers and publication were songs. Although the song had perhaps always been the medium in which he was most at ease, undoubtedly there was a crisis. (It is a mistake to view this long period of song production in a negative light, simply as the failure to produce instrumental music.) During this time Webern made arrangements of many earlier works, both his own and Schoenberg's, and sketched a number of movements for string quartet and trio, as well as two short piano pieces published posthumously, the Kinderstück (1924) and Klavierstück (1925), but as had been the case also with both Schoenberg and Berg, the abandonment of tonality had resulted in an inability to sustain extended forms without the aid of a text. 12-note technique provided a solution to this difficulty, and with the complete adoption of it Webern was able for the first time to write in the forms that had moulded German music for the previous two centuries, albeit considerably tailored to suit the conditions of the new method. Whereas for the slightly older Schoenberg this step represented a return to the old forms, for Webern it was a beginning: the traditional instrumental forms, while alluded to occasionally in his early music, were not followed rigorously in any of the atonal music. After his adoption of 12-note technique, although brevity would remain a cardinal feature of his works, he was to produce no more through-composed aphorisms of only a few bars. All the instrumental movements from op.20 onwards make reference to the old forms: binary and ternary forms, sonata, rondo, variations and on occasion – in the Variations op.27 for piano, the first movement of the String Quartet op.28 and the Variations op.30 for orchestra – a combination of two or more of these.
With this new-found formal stability Webern turned to writing for instruments with considerable vigour, producing three major works in four years (opp.20–22, 1927–30). The fourth big instrumental work, the Concerto op.24 for nine instruments, should have followed immediately – the first sketches run on directly from those for the op.22 Quartet with saxophone – but Webern suffered an impasse with the first movement, and its composition was to occupy him on and off for three years. As before in times of compositional crisis, he turned to song writing, composing the three songs of op.23 and the first of op.25 during intervals in his work on op.24.
Shortly after the completion of the Concerto – the second and third movements were composed in record time – Webern turned again to the poetry of Hildegard Jone (the opp.23 and 25 songs had been on Jone texts) and spent most of the following year writing his first cantata (though not so called), Das Augenlicht. Between October 1935 and August 1937 he composed two more large instrumental works, the Piano Variations and the String Quartet, and in 1940, after composing another cantata on Jone texts – the genre acknowledged in the title this time – he wrote what was to be his last instrumental work, the Variations for orchestra, in the space of eight months. These seven instrumental works, all substantial, each for a different combination of instruments and each incorporating either a sonata or a variations movement and in some cases both, comprise the entire body of Webern's mature ‘abstract’ (i.e. untexted) composition.
Although the length and formal scope of Webern's works changed significantly with his adoption of 12-note technique, most aspects of composition that had interested him from the time of the early aphoristic pieces continued to occupy him; thus many features of his style remained much the same. Two traits of his early music that might be singled out as particularly characteristic also of his later style are his special fondness for the interval of a semitone, together with its inversion and the various octave expansions of both, and his pointillistic scoring, with its resultant kaleidoscopic textures. When writing with a 12-note row, a preponderance of semitones could be assured by building them into the row. Thus this interval predominates in many of Webern's rows. The most rigorously organized example is the row for the Trio op.20, which consists of six semitones separated by larger intervals: the 48 row forms divide into two groups, in each of which all rows contain the same six semitones. One of these groups is shown in ex.2.
Scattered scoring is a feature of all Webern's 12-note works, the occasional section in which a long line is played by a single instrument (the theme, and variations 1, 2 and 6 of the second movement of the Symphony, for example) being the exception. Typically lines are divided among several dissimilar instruments and widely divergent registers, making continuity difficult to apprehend. Thus music which is already problematic for the hearer because of its dissonant intervals and its conciseness (material is not repeated or ‘played out’ in any way) is made even more formidable by its angularity and constant shifts of colour. Webern provided a good introduction to this eccentric type of orchestration (though this was not his purpose) in the arrangement he made in 1934–5 of the six-part ricercare from Bach's Musikalisches Opfer. Here, because the music itself is tonal and conjunct, and, indeed, familiar, the linear connections are easier to grasp. In one sense the most extreme example of Webern's pointillism is the second movement of the Piano Variations, where the impossibility of timbral variety is compensated for by a rigorous maintenance of registral disjunction.
Two predispositions that consistently shape Webern's 12-note writing are his propensity for canon and his fascination with symmetry. These preoccupations are interrelated, as we can see from the Symphony op.21, where they converge. This work is a brilliant tour de force of simultaneous vertical and horizontal symmetries (mirrors and palindromes) unfolding through a series of double canons. In fact, imitation in Webern is seldom direct: it is usually in inversion, sometimes in retrograde, and both of these situations result in symmetry.
The fascination that symmetry had held for Webern since at least as early as the op.5 pieces for string quartet acquired new scope with his adoption of the 12-note method of composition: now symmetry, like the semitone, could be built into the row itself if desired, as, indeed, it was in the row for the Symphony (ex.3a). This deceptively simple-looking row is symmetrical at two levels. Besides the obvious palindrome (the second hexachord being the retrograde of the first at the tritone transposition, and thus any two rows related as R6 and P0 being identical), there is a second and more subtle symmetry between any pair of rows related as I9 to P0. This can be seen in ex.3b. The two movements of the Symphony, in ternary sonata and variation form respectively, exploit these two symmetrical relationships, the first making use of the I9–P0 relationship in the outer sections and the palindrome produced by P0–P6 (i.e. P0–R0) in the central development, while the second movement offers the complement to this arrangement. The row structure of the second (variation) movement reflects the horizontal symmetry of the row itself: here not only is each of the nine sections (theme, seven variations and coda) palindromic, but the row structure of the whole movement is a palindrome, turning on the fourth variation as its axis. Characteristically, this structure is successfully obscured by the meticulous adherence to all the outward features of theme and variations: successive 11-bar sections are scored for different groups of instruments, each section heralding a change of tempo, texture and style, and these features are not symmetrically arranged.
The hiatus Webern suffered in his composition of op.24 resulted from his continued aspiration after symmetrical perfection. His row for this work was inspired by a Latin proverb (ex.4a) which can be read in four directions. The row he finally decided upon approximated the symmetry of this proverb to a very high degree. The second, third and fourth trichords represent the three permutations of the first (in the order P, RI, R, I), with the result that groups of four row forms related as P0, R6, RI7 and I1 are made up of identical trichords, and if listed in the same way as the words of the proverb they produce a similar result (ex.4b). As demonstrated here, several of the forms offered by such a row contain the same trichords in a different order or with their internal content reversed. This characteristic is exploited in op.24, where, for example, structural returns use different rows to produce the same figures on the same pitches but in a different order. The question with which Webern struggled for three years was whether to preserve the internal integrity of the row by allowing all the trichords of all rows to replicate the contour established by the first. In the end he adopted a compromise. In his exploration of symmetrical possibilities op.24 represents an extreme in its concern with microcosmic details.
The other work in which symmetry is an obvious feature is the Piano Variations op.27. In this case the symmetry is not built into the row; any row can produce palindromes and mirrors when combined with its retrograde and its inversion, and this is what happens in the first and second movements respectively.
The kind of symmetry that Webern seemed to favour in his last years is the much less obvious but perhaps more comprehensive one that combines inversion and retrograde in one operation: the rows of his opp.28–30 are all identical with their own retrograde inverted forms. This is not a relationship that can be heard, but Webern presumably saw it as a strongly unifying device. The fact that it is not audible he would have found attractive, as one of the idiosyncratic features of his writing is the careful masking – as in the case of the variations movement of the Symphony noted above – of painstakingly constructed symmetries, as if in the end they were a matter of interest to the composer alone.
In fact this inclination to conceal is one of Webern's principal traits, and is largely responsible for the listener's difficulty in coming to terms with his 12-note music. The form of these works is clearly delineated by the row structure – reprises use the same rows and the same transpositions as the original – but the musical manifestation of these is so varied as to make aural recognition unlikely. Webern does not work with themes and barely with motifs, except in the rhythmic sense. With the exception of op.30, even his variation movements do not have themes (in spite of the misleading ‘Thema’ written over the first section of the op.21 variations). These are nevertheless almost certainly the easiest of his structures to identify aurally because of the careful manipulation of secondary features; his penchant for veiling the truth in these instances serves to hide the absence rather than the presence of the traditional content.
The Variations op.30 for orchestra is perhaps the most rigorously motivic of all Webern's 12-note works. Here the entire fabric is woven from a multitude of versions of two rhythmic motifs presented in the opening bars. The motifs are derived from – or reflect – the row (ex.5). This movement illustrates a problem endemic in 12-note music: the discrepancy between variation as a technique and variation form. Music composed on a row inevitably proceeds through a process of continual variation, and this is true whatever form the movement is cast in. In spite of Webern's description of the first 20 bars of op.30 as the ‘theme’, the material that is varied throughout the movement unfolds in the first four bars, and the remainder of the ‘theme’ is already concerned with the variation of this material.