Musical Times

Prokofiev’s op.38 and op.135: Crowning the opus

Prokofiev’s op.38 and op.135: Crowning the opus

Walsh, Fiona

FIONA WALSH re-evaluates the two versions of the Russian master’s Fifth Piano Sonata

PROKOFIEV’S NINE PIANO SONATAS have never been his best-known or most deeply appreciated works, and the Fifth Sonata has been possibly the least performed and discussed of all. Although the composer himself looked back on the work in its original version, at least, as ‘intricate’ in style, and one of his `most chromatic’ essays,1 its expressive character has been variously described by other critics in terms ranging from one end of the interpretative spectrum to the other – unaffectedly simple, self-consciously complex, gentle, harsh, organically melodic, mechanically percussive, warmly natural, coldly formal. This three-movement sonata in C major was composed in 1923 as op.38, and was published in the following year; but apparently still dissatisfied with the work after some decades, Prokofiev revised the Fifth Sonata as op.135 in 1952-53, and it was so published posthumously in 1955.

The status of the original version of the Fifth Piano Sonata, op.38, has never been fully eclipsed by that of the revised version, op.135. During the 1960s the pianist Arthur Loesser forged his own solution to the question of the work’s two versions, advising his audience:

In my not too humble estimation, the composer’s revisive afterthoughts do not, for the most part, constitute improvements on the original, except, however, for the very end, which sounds more satisfactorily finale-like in the later version. Thus what you will hear will be Opus 38 almost entirely; but the last five pages will be Opus 135.2

The enduring validity of the earlier opus has been still more definitely upheld by the pianist Barbara Nissman, who performed both versions of the Fifth Sonata for her recording of the complete Prokofiev piano sonatas released in 1989,3 and likewise programmed both versions in her recital series devoted to the same works in that same year.4 In her appraisal the composer’s revisions may have produced a more pianistically effective work, but somewhat at the expense of the stylistic consistency which characterised the original: whereas op.38 is ‘classical’ in orientation, Nissman concludes, op.135 represents ‘a Romantic paraphrase of a Classical piece’.5

But I should like to argue that Prokofiev’s reworking is more than just consistent with the stylistic orientation of the original version of the Fifth Piano Sonata. It results, in fact, in a much fuller realisation of the ‘classical’ essence, most notably in the crowning and conclusion of the sonata scheme as a whole. For although the precise circumstances surrounding Prokofiev’s decision to recast the three-movement Fifth Piano Sonata may remain obscure, his priorities and strategies for revision are immediately evident in a comparison of the two versions: overwhelmingly, the changes were aimed at the final movement. indeed, judging by the overall result of his compositional rethinking, it could quite plausibly be surmised that concern for the character and role of the finale as a conclusion to the complete scheme was Prokofiev’s primary motivation for revisiting the work at all.

Even within the first movement, where the changes were by no means radical, endings at various levels of structure seem to have been something of a preoccupation, for Prokofiev concentrated his alterations at the end of the exposition and development sections of the sonata form, and in the coda. The changes he made to the second movement were more minor, but those to the finale were significant indeed, and included the complete rewriting of its coda. In the most general terms, the composer’s alterations to the Fifth Piano Sonata also produced a final movement of more refined and transparent texture, with considerably enhanced clarity of melody and accompaniment in several key passages, along with more effective definition of some key thematic and motivic elements, and a less inflexibly dissonant character. All of these effects Prokofiev was able to achieve principally through the simplification and subordination of inner parts. But as to why he might have wished to bring about such a reorientation particularly within his finale – such a question points back to historical-stylistic issues of the role of the finale within the cyclic sonata scheme. Prokofiev’s revisions to the finale of his Fifth Piano Sonata served above all to downplay any elements of turgidity within the movement, and to invoke more overtly the essentially melodic-thematic traditions of classical finale-writing, harking back to the lighter, often stylised characters and forms overwhelmingly favoured by late eighteenth- and early nineteenth-century composers as conclusions to their solo sonatas. In opting for a movement more melodically engaging and accessible, less demanding and texturally cerebral, Prokofiev chose to coax rather than coerce his audience along towards the finish – wherein, nevertheless, a measure of excitement would all the more rewardingly be found.

NOT surprisingly, bar-by-bar comparison of the finales from op.38 and op.135 reveals a large number of minor alterations, variously affecting notes, figuration, rhythmic detail (more often in the left-hand part than in the right), and some barring, particularly where small-scale insertions and excisions were made. Such changes, however, are very easily distinguished from the more significant formal revisions, and it is the latter which are of real interest in relation to issues of the movement explicitly as a conclusion to the opus. Considering the modest length of this finale, the variety of its thematic and motivic components is certainly generous, if not altogether atypically so for Prokofiev. The plan of the original version of the movement, which I have designated A-B-C-D-A-E-A-B-C-D-A-coda (see exx.1-5), was substantially preserved in the revision, but with one obvious modification: the scheme of ideas was tautened by the removal of the third A statement, resulting in the arrangement A-B-C-D-A-E-B-C-D-A-coda. Several lower-level expansions and compressions of material were also made throughout the movement, and it is noteworthy that in all of the rewritten passages the principal motive from theme A plays a less texturally pervasive role. The effect of motivic and textural integration so characteristic of the original is thus foregone, and its motivically centrifugal tendencies are suppressed; but the impact of thematic recapitulation proper, in the second half of the movement, is thereby strengthened, and in the absence of the former continuous effect of directions tried but abandoned the thematic statements themselves assume a more purposeful character throughout. The internal structural proportions of the movement also show considerable adjustment. At 140 bars, the revised version of the movement is fourteen bars shorter than the original – although not because Prokofiev simply excised so many bars: in fact he removed about twice that many altogether, but he also inserted a significant number, and most of those in a considerably extended coda, to be considered further below.

A variety of formal interpretations could plausibly be put upon the original, rather loose formal scheme, A-B-C-D-A-E-A-B-C-D-A-coda. Were it not for the consideration that the movement bears no signs of internal sectional articulation or closure, the most obvious suggestion might be that of a background rondo plan, with A as the ritornello theme, B-C-D as a somewhat kaleidoscopic first episode, and E as a second episode. Taking into account the relatively integrated, through-composed guise of the movement, however, some underpinning synthesis of sonata and rondo rationales – along the lines that Charles Rosen identifies as the classical finale-sonata form6 – would seem more likely. Some commentators, assuming a more standard, two-theme sonata-form basis to the movement (and without proposing any alphabetical scheme of thematic identification), have judged that the motive here analysed as D is a second sonata-theme proper, with C a mere transitional idea thereto – and B, presumably, a mere adjunct to theme A, although this is not explicitly said. That view, however, is perhaps unnecessarily influenced by the consideration that Prokofiev rather uncharacteristically followed an orthodox two-theme sonata formula in the first movement of his Fifth Sonata, a fact which cannot of itself be taken as the index to the architecture of his finale. Even within the original version of the third movement the ordering of thematic ideas in the second half implies that B and C are thematic elements in their own right, no less so than A and D; and this conclusion seems virtually inescapable in view of the revision to the later stages of the movement, whereby Prokofiev suppressed all references to A amidst the reprises of B and C. It would be a strange sonata-form recapitulation indeed that would still wend its way conscientiously through so much secondary material before arriving at either of the two main themes of the movement, if but two there were.

Perhaps more promising, if not yet absolutely incontrovertible, would be an interpretation of the original version of the finale still as sonata-rondo– based, but with A as a first thematic ingredient, BCD as a second group, A restated to conclude the exposition, and E-A following as the development; then a species of reverse recapitulation (themes of the second group, B-C-D, in their original order, before the final reprise of A), plus coda. In support of such an interpretation it must be emphasised that the third appearance of A, being the most tonally ambiguous and unstable of all the various thematic presentations throughout the movement, has a classically developmental orientation, and cannot plausibly be construed as initiating any sort of resolving recapitulation. On the other hand, the chief contraindication to such an interpretation of the form would be the absence of a conventionally conspicuous transition between first and second thematic constituents, for B follows with seemingly undue suddenness upon the heels of A. Considering, however, that the tonal scheme of the movement (to be discussed further below) does not at any point necessitate a transition, the absence of such a passage cannot be taken as the real key to the anatomy of the exposition.

It must be said that tonal considerations do not help much to clarify the formal-analytical issues surrounding the finale. If anything, some sort of reverse tonal principle appears to be at work. In the first half of the movement, where some opposition of tonal identities might appropriately be established, the initial presentations of thematic ideas A, B, C and D all gravitate around the pitch of C as a tonal centre (whether that is made melodically explicit, as in the case of the A theme, or is simply implied by the accompaniment, as with B, C and D themes). Theme E, likewise, is announced emphatically from the pitch of C. Then in the later stages of the movement, where some reconciliation of any tonal rivalries might be attempted, the same themes veer off in independent directions, with only the accompaniment to the last bars of B, and the final statement of A itself, finding the original tonal centre once again. The revised version of the movement features only the one change to this scheme, where the chromatic-scalar theme C, octave transposed, is recapitulated from the pitch of G, rather than G# as in the original version (see ex.6). The alteration has no obvious structural significance, unless Prokofiev intended it as a gesture of deference to the notional dominant hardly likely at so late a stage in the movement.

While the rather non-committal tonal workings of the movement are apparently less than critical to either the original scheme or the revision, in terms of architectural or structural alterations to the finale there are three passages of considerable interest. In the first of these, Prokofiev subjected the middle regions of the movement to significant contraction. From the original scheme of A-B-C-D-A-E-A-B-C-D-A– coda, not only was the third, tonally centripetal presentation of the A theme sacrificed altogether, but the second presentation was also somewhat curtailed by the trimming of four bars in the leadup to theme E. The E theme itself is worked out rather more purposefully in the revised version of the movement, and although some digressive references to the second motive of theme A are still present, they are now postponed until after the thematic identity and dominance of E have been well established; moreover, owing to a marked simplification of the texture they no longer threaten to obscure the significance of E itself, as they tended formerly to do. Overall, with the rewriting of original bars 57-89 as new bars 58-75, this inner section of the finale is reduced from thirty-three to just eighteen bars. Whether the formal scheme of the movement is construed generically as one of exposition-contrast-return, or in more sonata-specific terms of exposition– development-recapitulation, it is quite clearly the middle region which the composer deemed in need of restraint.

In the second passage of significant structural alteration, which follows directly from that just discussed, Prokofiev rewrote bars 90-130 of the original finale as bars 76-103 of the revised version. In so doing he contracted his musical material by a further thirteen bars. The passage in question covers the return of B, C and D themes, and the retransition to the final reprise of A. In terms of any sonata rationale, Prokofiev’s wish was quite clearly to tauten the first ‘half’ of his reverse recapitulation. He accomplished this purpose, as already recognised, chiefly by deleting those allusions to the principal motive of theme A which were originally interspersed between the recapitulations of themes B and C, and likewise between the recapitulations of themes C and D. Whether as a mere trade-off or in fulfilment of a more positive strategic intention, Prokofiev also worked into the revised version of the movement a deliberate, off-tonic false-start to the final reprise of theme A (bars 100-01), directing more emphasis towards the true start of the final recapitulation of that main theme (bar 104) as a point of real arrival and heightened dramatic importance.

The one remaining alteration of significant structural interest is found at the conclusion of the finale. Here Prokofiev abandoned bars 143-54 of the original version, composing afresh bars 116-40 of the revised version instead. This is the most telling of all the alterations, for as a result the rather more concise revised version of the movement boasts a much more developed coda, twice the length of the original, and evoking considerably more excitement by its expansiveness of register, insistence of ostinato figuration, and dynamic pacing. In view of these modifications, especially, it looks as though issues of the ending as a positive goal were more than simply incidental to the reworking of the movement; and in apparent confirmation of such a theory, forsaking the furtive piano, bass-register figuration of the original final bars, Prokofiev concluded his revised version far more boldly in crescendo vein, opposing bass and treble registers to one another in deliberate style with the heavily accented, meno mosso chords of the final two bars.

WHAT, then, is the overall implication of the various alterations to the finale of the Fifth Piano Sonata? The answer to that question is that in abbreviating the original thematic arrangement, A-B-C-D-A-E-A-B-C-D-A– coda, to A-B-C-D-A-E-B-C-D-A-coda, Prokofiev strengthened the structural-analogical relationship of this finale to the classical sonata form, with all its teleological implications. The original scheme unquestionably highlighted the weakness inherent in the simple rondo-based finale form that is, that thematic recapitulation tends to lose its impact as an ending strategy, if it is over-used for purposes other than those of ultimate resolution and conclusion. By excising the third appearance of theme A, and thus undercutting the rather static, literal parallel between the two large thematic blocks either side of theme E, Prokofiev remedied the suggestion of an over-reliance on his principal thematic idea. Simultaneously, he reasserted some of the traditional tautness of the middle regions of the sonata form, with a relatively brief development, and gave more overt definition to the apparent reverse recapitulation of the movement. In so doing, he also activated something of the formal dynamism and ‘onward’ momentum of the classical form. By the various other proportional adjustments already identified, he relocated the major events of conclusion – the onset of recapitulation and the commencement of the coda – somewhat further back from the very end of the movement, and in this way substituted for the rather unceremonious character of the original ending an effect of more compelling approach, expansiveness, and ultimate arrival.

All other things being equal, these measures should, in theory, have resolved any lingering weakness in the conclusion to the complete sonata. In practice, Prokofiev’s changes certainly made for a better ending, and indeed made a better movement of the finale in many respects. In the absence of a compelling tonal rationale to the movement, however, there was still some limit to what could be achieved from that basis. This observation is not to suggest that the composer should have resorted to an anachronistic system of key relations to give `meanings to his finale; however, by allowing the first, expository half of the movement to gravitate around the one tonal centre (and so foregoing the opportunity to create any effect of tonal distance at all), while permitting the second, recapitulatory half to spin off in various directions of undefined (if any) significance, Prokofiev himself forfeited the possibility of tonal resolution as the route to a really ‘final’ and satisfying finish. As a result, while structural adjustments to the later stages of the movement undoubtedly produced a proportional scheme of recapitulation and coda more orthodox from the classical-analytical point of view, the lack of formal definition and generative tension in the earlier stages of the finale furnishes little supporting framework for these important signposts towards a sonata-form ending. The reverse recapitulation of the movement is undoubtedly more evident to the analytical eye in the revised version than in the original, but its onset lacks the unambiguous tonal credentials necessary for a more definite aural impact; and the root of that anomaly lies nowhere near the conclusion itself, but somewhere much nearer to the beginning of the movement, in the fact that tonal issues were not created or exploited when they might have been.

Nevertheless, the improvements brought about by Prokofiev’s revision to his Fifth Sonata are quite definite. Above all, in comparison with the finale to op.38, the third movement of op.135 with its more dynamic formal scheme and less encumbered texture – represents a much more engagingly finale-like conclusion to the complete scheme of the work. Whatever Prokofiev’s original view of this sonata in relation to such categories (usually problematic) as modernist, neo-classical, tonal/atonal, or formalist, the later version of the work provides sure evidence of a maturation of his concept of the sonata finale in accordance with high classicism’s own aesthetic and formal traditions. The result is not only a significant improvement in the internal workings of the finale itself, but also, in consequence, a clarification and enhancement of the intermovement dynamic within the three-movement scheme of the Fifth Piano Sonata as a whole.


1. Sergei Prokofiev: Autobiography articles, reminiscences, comp. and ed.

S. Shlifstein, trans. Rose Prokofieva (Moscow: Foreign Languages Publishing House, n.d.),

pp.61, 64. The autobiographical chapters, in which Prokofiev made these comments on the Fifth Sonata, date from 1941.

2. Arthur Loesser: recital programme notes, January 1967, reproduced as notes to A recital of little known works by wellknown composers, perf. Arthur Loesser, LP International Piano Library, n.d.

3. Barbara Nissman, perf.: Prokofiev: complete piano sonatas, vol. I, Newport Classic, 1989. The last of the listed tracks, the finale to op.135, is missing from the disc.

4. See

Barbara Nissman: `Discovering

Prokofiev’s sonatas, in Musical Opinion vol.112 (1989), pp.47-49.

5. Barbara Nissman: `Prokofiev’s two versions of his Fifth Piano Sonata: facsimile and commentary’, in The Musical Times vol. 130 (1989), pp.596-99, p.599.

6. See Charles Rosen: Sonata forms, revised ed. (New York: Norton, 1988), pp.98, 123-32.

7. See, for instance, Frank Merrick: `Prokofiev’s piano sonatas 1 to 5′, in The Musical Times vol.86 (1945), pp.9-11, p.11; also Nissman: `Prokofiev’s two versions’, p.599. Merrick’s analysis obviously predates Prokofiev’s revision to the Fifth Sonata, and Nissman’s

seems to be situated within the Merrick tradition.

Fiona Walsh is a recent doctoral graduate of the University of Queensland.

Copyright Musical Times Publications, Ltd. Spring 2003

Provided by ProQuest Information and Learning Company. All rights Reserved