Is it ok to embellish Beethoven’s music?

The second volume of my recording of the complete Beethoven Piano Sonatas is being released and I want to grab the chance to make known some interpretative principles I’ve set for myself. I won’t use this forum to talk about the core of the interpretation, what I’ve wanted to make the music sound like or what kind of imagery Beethoven’s musical text evokes in me–I won’t try to repeat in words what I’ve tried to do with sound. Instead and particularly since the music is transmitted to us through historic sources and because much of the interpretative aesthetics deals with old and modern performance practices, I want to shed light on some of the more practical choices I’ve made regarding repeats and added ornamentation. I will also list the editions I’ve used for studying the music to allow people to see the issues for themselves.

As one of the primary interpretative choices, I decided to take all repeats indicated in the music, and, in so doing, depart from the expected norm of recent decades. (There was a moment in the sessions when I actually considered not repeating in the first movement of the Tempest Sonata, but, thanks to my producer Stephan Flock, I dropped that idea.) A number of my colleagues have included repeats of the second parts of outer sonata movements (such as the ones in sonatas Op. 2 No. 1 and Op. 10 No. 2) but carrying out repeats in the da capos of many middle movements (typically minuets and scherzos) has been rare. My decision to include those is based on a belief that there must have been a reason why Beethoven sometimes wrote “da capo” and, in other instances, “da capo, ma senza repetizione” at the end of the trio- or minore-sections preceding the da capos. To me that clearly indicates that sometimes he wanted the repeats and sometimes he didn’t. The custom during recent decades has been to omit repeats in da capos, but that, along with omitting the repeats of the second parts of some of the outer movements, I consider to be a modern performance practice. The main consequence of taking all the repeats is that, especially in some of the early sonatas, much of the music (obviously) ends up played two or more times.

The ornamentation I have added emerges from an understanding that the repeats in the early sonatas bear more of a rhetorical nature while in the later ones they serves as a structural device. In the very first sonata nearly all musical material is repeated but in the sonatas from ten years later Beethoven takes a unique approach to repeats in each movement. My conclusion is that by that time Beethoven no longer took anything in the “sonata protocol” for granted and the very frame of the music had as much to do with his musical message as it did with the musical material (melodies and motives). As a conclusion of that conclusion, I deemed adding some ornamentation in the early sonatas to be appropriate while I refrained from applying any in the past-1800 sonatas.

Moonlight Sonata, manuscript page

Beethoven’s manuscript page, 1st movement of Sonata Op. 27 No. 2


About the editions 

A central aspect to my process is that I study the music in many different editions. In my teens, I learned Beethoven Sonatas from Curci (Schnabel) and Peters (Arrau). The one I’ve most consulted during the past few years is the more recent ABRSM (Cooper), and I would be remiss not to mention the wonderful modern addition to the list, a facsimile reprint of all the first editions of Beethoven Sonatas by Tecla. While preparing for this recording, I also studied Dover (Schenker), Wiener Urtext (Hauschild), and Henle (Wallner). Each of the editions offers much that is good. The older ones, typically filled with a wealth of artistic advice, portrayed such noted pianists as Hans von Bülow, Schnabel, and Arrau negotiating between the historic sources they had available and their own artistic voices. I learned much about tempo, articulation, and even fingerings from them.

Even though a modern musicologist can provide more precise (and often weighted) knowledge about Beethoven’s proof-reading habits, historic inaccuracies, and recent mistakes in the scores, I wished to embrace the tradition these spectacular musicians carried forward in their own times. As a bit of a sidetrack to this, at times by certain interpretive choices I’ve wished to pay homage to pianists including Edwin Fisher and the two Wilhelms, Backhaus and Kempff (enthusiasts, feel free to find the spots!). As a result, my approach is a combination of the most recent knowledge as represented by modern Beethoven scholarship as I understand it, while acknowledging much that has been “said” about the music over the two centuries of its existence.

Recording bureau, Lentua Hall in Kuhmo

My desk at a recording session in 2010. At least 5 different editions on the table (and probably some more in the plastic bag).


Historic perspective on repeats

In the course of history, musicians have treated the repeats indicated in the score differently. Early recordings omitted many of them due to space restrictions on vinyl discs. Another approach was articulated by Johannes Brahms (of all people) earlier, when he famously stated that he no longer thought repeats in performances of his symphonies were necessary because audiences were “already familiar with them.” Hearing Wilhelm Kempff skip the repeat of the exposition of the first movement of the Tempest Sonata in his 1950s recording, one might think his decision possibly serves to underline the work’s ephemeral character; most would agree that the Tempest is a mysterious work and a repeat of the opening is likely to make the listener more familiar with its eccentricities.

Musicians who swear by Schenkerian Analysis might decide to skip the repeats of second parts of sonata movements because the enchanting analytic interpretation of Schenker likes to portray the first part of a sonata movement as a buildup of harmonic tension, which the second part serves to resolve. Repeating the process of dissolving might seem unnecessary in this context. All in all, when it comes to observing repeats, there are numerous reasons for deciding one way or another. Common to the decisions is that they tend to speak to contemporary norms and they are choices the performer is expected to make.


Repeats and phrase structure 

As concluded above, I believe that Beethoven must have had his reasons to occasionally write “da capo” and other times “da capo, ma senza repetizione” at the end of the Trio- or Minore-sections sandwiched by the first and second iteration of the movements’ main section. Much can be said about the phrase structure that has to do with this question (and I do say some of it below) but the most straight-forward point I want to make is that in most cases where the da capo-indication occurs without the ma senza repetizione -bit, the first part of the movement in question has its repeats written out due to some variation in the text and skipping the repeat of the second part in the da capo would thus create an odd misbalance in phrase structure.

There are also the cases, like in the scherzo of the Sonata Op. 2 No. 2, where the repeat of the first eight measures doesn’t strike me as a true repeat, but together with the first part creates a 16-measure phrase. Thus, omitting the repeat in the da capo would strangely dissect the phrase. Also, there are cases in the Violin Sonatas where the first part of the minuet or scherzo is written out clearly for reasons of variation (the piano plays the first phrase solo and the violin joins for the actual repeat). This happens in the second movement of the Sonata Op. 24 and in the third movement of the Sonata Op. 30 No. 2, as it does in the second movement of the Moonlight Sonata. In all these cases, the repeat of the first part is automatically played in the da capo and omitting the repeat in the second part would simply be awkward–and sure enough, the indication at the end of the trio-section in all of these cases is simply “da capo” (with no “senza repetizione” in sight).

Clarifying evidence can also be found in works where the da capos are written out (i.e. printed after the middle-section), such as the minuet movement of the Sonata Op. 31 No. 3. Beethoven probably didn’t include the da capo-section in his manuscript of the Sonata but the early editions do carry repeat signs within the written out da capo. Do we really believe that the engraver would have been unaware of the manner in which the repeats were treated? I doubt it. The case is the same in the second movement the Violin Sonata Op. 30 No. 3, where all the repeats (in the A-part, B-part, and the da capo of A-part) are written out (obviously due to slight variations in the melodic line and division between the two instruments). Here, clearly, the da capo also comes with the repeats.

On some instances Beethoven indicated that other action ought be taken, or rather, ought not to be taken regarding middle-movement-repeats. In some cases where the repeats have been written out, for example in the third movements of both the Violin Sonata Op. 30 No. 1 and the Piano Sonata Op. 28, Beethoven warns performers to take notice and obey the indication “La prima parte senza repetizione.” That together with the sometimes occurring “da capo, senza repetizione” has led me to believe that Beethoven’s indications regarding da capos and repeats were deliberate and should be followed accurately.

End of Trio-section, Mov. 3, Op. 2. No. 1

End of the Trio-section of 3rd movement of Op. 2 No. 1 followed by Menuetto Da Capo -sign (first edition).

End of Trio-section, Mov. 3, Op. 10 No. 3

End of the Trio-section of 3rd movement of Op. 10 No. 3 followed by Menuetto Da Capo ma senza Replica -sign (first edition).


More about ornamentation

Regarding added ornamentation in the repeats, I’ve established two principles. The first emerges strictly from musical considerations in the sense that I occasionally add ornaments as a way to deal with the many repeats in the early sonatas, either adding small Eingangs on fermatas or by embellishing the melodic contour. As I’ve stated above, I read the fact that repeats occur less frequently in Beethoven’s later sonatas as evidence of the repeats in the early sonatas being more rhetorical in nature and, in the later ones, a structural device. The two sonatas in F-Minor, Op. 2 No. 2 and Appassionata, serve as cases in point. In Op. 2 No. 2 nearly everything is repeated but ten years later in the Appassionata nothing in the first movement but almost everything in the second movement and only the second part in the finale are repeated–by this time the repeat wasn’t automatic but, instead, a device in service of the structural dramaturgy of the piece.

The second principle has to do with the performance culture from which Beethoven’s music emerged. Having studied a lot of the European sonata repertoire of the 1780s and 90s, including many composers unknown today, it seems obvious to me that Beethoven’s early sonatas were part of a tradition where ornamentation was still expected in performances. There were, of course, better and worse choices made (we know that Mozart thought that the clarinetist Anton Stadler sometimes ornamented too much), and my ornamentations, too, may be criticized. To demonstrate the historical stylistic change in performance culture, especially regarding to the performer’s “handprint,” I ornament more in the early sonatas and gradually do so less as the cycle progresses. Beyond the Sonata Op. 28, I add no ornaments that would change the text, so to speak.


I plead guilty

These days we tend to look for the composers’ consent for our handling of their music, and, in the absence of more concrete input, we too often turn to anecdotes of the kind with which Beethoven-literature is virtually brimming. Some relate Beethoven’s reactions to performers taking liberties in his music. Interestingly, the two most famous stories testify to contrasting reactions by the composer. The first, related by Ferdinand Ries, concerns the first performance of the Kreutzer Sonata with Beethoven at the piano and George Bridgetower playing violin. According to Ries, Bridgetower’s surprise-move to add a little cadenza in the repeat of the first part of the sonata prompted Beethoven to stop the performance and enthusiastically embrace the violinist and shower him with superlative praises, “my dear boy …”

On a different occasion Beethoven was said to have lost his temper when his student Carl Czerny “took too many liberties” in a  performance of the Quintet Op. 16. There is no record of what exactly those liberties might have been, but a study of the quintet reveals that first of all, there are only two short repeats in the whole work and it is conceivable to speculate that Czerny was unlikely to have limited his inventiveness to them. Also, there are a number of written out cadenzas spread through the piece (much like the one added by Bridgetower in the Kreutzer Sonata) and if Czerny went beyond the cadential elaborations already provided by Beethoven, the reaction of the composer can be forgiven. In a letter dated a few days after the incident, Beethoven apologized for his anger to Czerny and wrote that while he held Czerny’s artistry to be of the highest degree, he was at times bothered by his apprentice taking too many liberties.

While I know that I won’t have my day in court with Beethoven as a member of the jury, I have decided to take my chances and go the route demonstrated by George Bridgetower and to thus do my best to serve the music as I conceive it.

Share:Share on FacebookTweet about this on TwitterShare on LinkedInEmail this to someone

Subscribe to new posts by email:


Leave a Reply

Your email address will not be published. Required fields are marked *