The mistaken art

1. Introduction

In a much-quoted letter from 1920, George Bernard Shaw, the famous playwright and music critic, cautioned Jascha Heifetz, the young violin virtuoso, in jest to remember his own humanity:

If you provoke a jealous God by playing with such superhuman perfection, you will die young. I earnestly advise you to play something badly every night before going to bed, instead of saying your prayers. No mortal should presume to play so faultlessly. (McLellan, 1987)

As a college student in the 1990s, fiendishly discovering and exploring the wonders of early recorded performances, my experience listening to Heifetz on record repeatedly brought Shaw’s quotation back to mind, so “perfect” were his performances. Initially this became one of my boasting points as I annoyingly promoted his playing to my friends. “He’s so great,” I argued, “that he never missed a note!”

It was within this context and stage of life that I excitedly listened to an unreleased recording of the Sibelius Violin Concerto Op. 47 with Heifetz, Leopold Stokowski, and the Philadelphia Symphony Orchestra from 1934, the only recorded collaboration between these two artistic giants. It was clear that they had a different conception of the concerto. There were moments where it was almost like a horse race as one tried to outdo the other. And it was, perhaps for this reason, that Heifetz refused to let the recording be published at the time. Yet for me it was thrilling. The tension was palpable (even as a studio recording!) and the friction between the competing musicians, exhilarating. But above all else, for the first time as a listener, I heard Heifetz in a state of imperfection. Not every note was exactly in tune or executed to precision. Compared to his later recordings of the concerto (which I knew intimately and adored), Heifetz revealed a different side of himself, his fallibility or, perhaps, his humanity, or that’s how I saw it at the time and as I revisit the performance today. And it was difficult to separate my delight in the performance from the imperfection of his playing.

I begin this essay, by way of this personal anecdote, to introduce the well-known but contradictory perception of musical errors alongside conflicting views of perfection. Mistakes, almost by definition, represent a violation of expected norms or rules and even human frailty. These stand in opposition to conceptions of fidelity to artistic, compositional truths or the honoring of superhuman ability and discipline. It was a combination of these qualities that, for a time, justified my personal worshipping of Heifetz. And, yet, depending on the contexts, these musical miscues also became, ironically, a source of celebration. In an age of mechanical reproduction, and an increased awareness of studio edits and manipulation, displays of imperfection can often be celebrated as an appealing or even desired aesthetic unto itself for a variety of reasons, many of them stemming from the rhetorical honoring of human spirit over brute force or “liveness” over editorial artifice. Thus inspired by Shaw’s oft repeated quotation, the revelation of imperfection in Heifeitz’s playing also demonstrated his humanity, suggesting he was no mere machine or god. Indeed, as Alexander Pope famously wrote in his An Essay on Criticism, “To err is human, to forgive, divine” (Pope, 1711).

Pope’s sentiment is lovely, and one that our society at large would benefit from heeding. However, for the purposes of this essay, it is not forgiveness (on behalf of the musicians I discuss) or even excuses that I’m seeking. Rather, I would like to explore the different ways musical mistakes or missed notes, especially those captured on record–both in studio and from concert performances–have been or could be embraced, celebrated, and even integrated into the aesthetic experience and critical evaluation of the recorded performance. I will also focus predominantly on recorded performances since the recording as a form of mediation cannot be separated from our experience as listeners, even if we often try. In fact, as I will explain below, listener expectations for recordings (and the aesthetic categories they generate as mechanised objects) are significantly different from, but not unrelated to, the ways that many experience concert performances. Nevertheless, I will still refer to broader anecdotal histories of artistic mistakes for two reasons. First, they are entertaining and in many of these instances are central to biographical lore. Additionally, they amplify and complicate what it even means, for both the producer of musics but also our critical responses, for something to be mistake in the first place.

Finally, I will be focusing predominantly on classical music since, as I will discuss below, the histories of classical music, in the last two centuries, have created a hierarchy that positions the performer well below the composer. Furthermore, the concept of Werktreue (faithfulness to the “work”) and the idea that a score truthfully represents the intentions of the genius composer further reinforces the idea that a given performance, if it doesn’t follow the instructions in the notation, is inherently “wrong,” and thus a mistake.

2. Technology and Perfection

With the introduction of recording technology towards the end of the 19th century (including not only gramophone machines but also the variety of piano players or pianolas), we start to see a growing binary emerge between humans and machines that often shapes the way performers are viewed.² In some ways, the binary has centuries-old roots when we consider early mechanised instruments or, more broadly, the ways that critics and audiences evaluate performers like Paganini or Bottesini in relation to their instruments. Caricatures of the virtuoso Dragonetti, for instance, merge his bass and body in a manner that reflects an identity that is “half man and half instrument” (Nachtergaele, 2024, p. 219).

And yet as binaries go, the human/machine relationship is anything but consistent as technologies, aesthetics, and listening habits change. As Greg Milner argues in his perfectly titled book, Perfecting Sound Forever: An Aural History of Recorded Music, the presence, use, or awareness of technology can be seen as both an asset and a liability. On one hand, recording technology can be seen as sterilizing or impersonal, by removing any sense of time and place. Moreover, given the mechanical nature of reproduction and production itself, recordings challenge the very nature of “originals” (and the Benjaminian concept of “aura”) that might not have ever existed in “real life,” given the prevalence of post-production work, editing, etc. On the other hand, recording technologies not only gives us access to the past, enabling us to hear performers no longer alive or concerts thousands of miles away, but they present us with the ability to listen differently, with repetitions, and to hear potential aspects of a performance that might not have been possible to hear from an audience. In this way, we hear more and we hear differently, allowing us to analyze and study aspects that are not perceptible to the naked ear or single-time listening opportunities.³

Obviously, this long-standing tension is far more complicated and, in many cases, genre specific since there are many styles where the heavy hand of production or post-production is expected as opposed to those where it is stigmatised. As Amy Blier-Carruthers (2024) has argued, this stigma, the expectation of perfection, is especially prevalent among classical musicians where the aesthetic of “live-ness” (i.e. the concealing of any production work) is combined with the expectation of note-perfect execution (e.g. no mistakes). This creates a tension that stems entirely from the impact of recordings on the act of performance, where performers increasingly fear making mistakes. Because recordings can be played back repeatedly, where each iteration is exactly the same as the previous one, they become associated with a sense of fixed-ness. For this reason, any mistake that, in a concert, might be easily dismissed or forgotten by the performer or audience, becomes permanently engrained in perpetuity, to be heard over and over upon replay. It is for this reason that so many performers struggle to achieve “perfection” when playing in the studio while also taking advantage of editing processes to remove any imperfections, blemishes, or mistakes. With the exception of such radical and courageous performers as Glenn Gould, who openly embraced and exploited the creative possibilities of studio editing, most classical performers seek to conceal or minimize the post-production editing in order to maintain the aesthetic of “live-ness.”

But at the end of the day, it is an aesthetic of “live-ness,” the projection of time and place onto recordings, especially of classical music, that we need to attend to. This aesthetic can be found most clearly in often over-simplified debate about “live” vs. studio. I say “over-simplified” for a number of reasons. First and foremost, “live” performances released on recording are still distributed and circulated as well as edited and, in some cases, even mixed from different performances. In other words, the label “live” instills a sense of naturalness or authenticity for marketing purposes, but it is never devoid of technological mediation. A performance may have been recorded in 1981, but the listening experience can happen over and over on any other time, date, and location. To make things even more complicated, a studio recording could be anything as simple as a single performance recorded without multiple takes but without an audience or as complicated as multiple takes with splices and edits.

In short, these false binaries of “live” vs. studio, or natural vs. edited/mediated, or authentic vs. inauthentic, are arbitrary boundaries we establish as criteria to elevate or denigrate a given performance or recording. It has everything to do with what to expect and to what degree our desires or needs are being fulfilled or denied. Besides, as Glenn Gould playfully demonstrated, most listeners are unaware of or unable to identify professionally executed splices when asked to listen for them (Gould, 1984a). Consider the marketing behind so-called hi-fidelity technology that sells itself by, as it were, denying its own material and sonic presence by presenting the listener a performance as faithfully as possible, without the distracting sounds of surface noise. By comparison, as Alexandra Supper demonstrates, other recordings are promoted specifically for their “lofi” qualities as a way to celebrate their authenticity, that is, their location in time and space, their spontaneity, or their naturalness. (Supper, 2018) The surface noise or extra-musical sounds (or lack thereof) are the same in both cases, except in one case they’re a cause for criticism and in other cases a reason for praise.⁴

3. Mistakes in Performance

The arbitrary or, at the very least, genre/style/artist specific-way one determines value in technological mediation is similar to the way that many evaluate mistakes in performance, especially in classical music where fidelity to the intentions of a composer, a musical work, and/or the score is often viewed as sacrosanct.⁵ Consider, for instance, the almost sla-vish devotion to “wrong” notes as a measure for value in musical competitions.⁶ A few slips can knock the most musical musician out of contention with ease. It’s for this reason that competitions are often criticised for emphasizing perfection (i.e. correct notes) over musicality. Indeed, Glenn Gould’s satirical essay, “We Who Are About to Be Disqualified Salute You!” makes this argument in spades. But in the defense of competitions’ judges, the issue is symptomatic of the need to make music competitive in the first place. After all, interpretation is highly subjective whereas a wrong or right note can be compared in order to justify more “objective” decisions.

The ubiquity of recordings has also, of course, played no small part in the expectations of many listeners, who are often used to note-perfect renditions of music.⁷ Indeed, listeners who learn a composition from a recording or become accustomed to a note-perfect, even if highly edited performance, often find fault with concert performances which are rarely note perfect. But just as surface noise can be seen as a blessing and a curse, as described above, so, too, can wrong notes. While there are endless examples where flawless renditions are celebrated, there are also many for whom the absence of perfection is, itself, celebrated an aesthetic asset.

Consider, for instance, the response by a variety of musicians to the performances, on record and in concert, of the pianist Alfred Cortot whose penchant for finger slips, especially as he got older, was commonly acknowledged. The well-known pianist and author Alfred Brendel writes of his predecessor:

Where the leading of voices, the grading of dynamics, the control of character and atmosphere, timbre and rhythm are handled with the mastery of Cortot at his best, it appears to me that a few missed notes are not only irrelevant but almost add to the excitement of the impact. (Brendel, 2015, p. 336)⁸

Brendel does not apologise for Cortot’s mistakes. Rather he reframes them as part of the musical experience. Implicit in Brendel’s description is a kind of tight-rope walking tension that comes from a lived experience. That is, it’s not that the mistakes, themselves, are desired but that they’re an expected component of the overall experience, a performative form of lo-fi that reinforces Cortot’s “authenticity.” Or in other words, Brendel projects a kind of “live-ness” to the studio recording experience to give it value.

Music critic Harold Schonberg made a similar point when he wrote of Cortot’s mistakes: “One accepted them, as one accepts scars or defects in paintings by an old master” (Bambarger, n.d.). Schonberg’s comments resonate with the aesthetic of imperfection that we see in writings about visual art, where the intentional or unintentional signs of the artist (or imperfections on an otherwise polished work of art) identify the humanity and genius of the artist even while the art seeks the verisimilitude of the image or subject.⁹ That is, the mistakes or defects are still errors, but they offer the work a kind of aura that separates it from perfection represented by mechanical recordings.

Pianist Philippe Entremont amplifies the connection between recordings and mistakes further, especially the impact of recordings on performance style, by celebrating the mistakes in Cortot’s recordings in opposition to the sterile “perfection” that more modern recordings present:

I am more fascinated by some of the old recordings of Cortot than by most of what I hear on recordings made today. Today's recordings tend to be extremely well-executed—I use the word 'executed' on purpose, because sometimes it is a deadly execution—with a kind of cool perfection, and a good sound. But there is an antiseptic quality about them….Pianists are not willing to take the risks they once did…But that would not happen with a pianist like Cortot—even his mistakes were fabulous. …Once when I was in Yugoslavia playing a concert. I heard them on the radio in my hotel room. I was beside myself, absolutely spellbound by the courage of his playing, the nonconformity, the fabulous drive—the poetry of the music was airborne. (Dubal, 2000, pp. 157-158)

Entremont’s assessment of Cortot implies that the expectations of perfection on recordings have had a detrimental effect on all subsequent playing and that the excitement that comes from the risk of concert performances is lost as a result. And there’s no question that recordings have had and continue to have a radical impact on both how listeners evaluate performances (recorded or in-concert) but also how performers play, both in concert but also when recording.¹⁰

However, it’s also clear that elements of this aesthetic existed without the interference of recordings, that is, an appreciation for the all-too-human element of performance that stand implicitly in contradistinction from more mechanically, note-perfect but “less musical” renditions. At the turn of the 20th century, for instance, long before recordings had had a significant impact on listening or performing expectations, the pianist Percy Grainger reported on his impression of the pianist Eugen D’Albert:

When I saw d’Albert swash around over the piano [playing his sonata] with the wrong notes flying to the left and right and the whole thing a welter of recklessness, I said to myself “That’s the way I must play.” I’m afraid I learnt [sic] his propensity for wrong notes all too thoroughly. (Grainger, 1999, p. 342)

Graingers response is a mixed bag. As I’ll discuss later, Grainger’s embrace of “wrong” notes or, rather, his potential “righting” of these notes, might in some ways anticipate his future views on music. On the other hand, Grainger’s description anticipates the same kind of aesthetic, previously cited, that was used to justify Cortot’s performances. Grainger is sheepishly apologetic for the errors but is more than willing to embrace them as part and parcel of a larger performative experience. Indeed, it is D’Albert’s “recklessness” that excites him, the kind of thrill that comes from a performance in real-time that could potentially fall apart at any moment but for which the danger, the precarity of virtuosity, the high-wire act of dramatic flair, trumps any errors along the way.

4. Re-centering the Performer

One of the central reasons that Cortot and D’Albert’s performances, and their technical inconsistencies, aren’t necessarily cited as obstacles to enjoyment by their critics or highlighted as errors, stems from the fact that their aesthetic focus was different from more contemporary historical perspectives. In other words, Grainger’s views of D’Albert and Brendel and Entremont’s views of Cortot were, as it were, just that, evaluations of performers and their interpretations. As opposed to most histories and theories of music written over the last two centuries, which largely ignore the role of performers or relegate their status to that of conduits between composers and listeners (i.e. passive or empty vessels), these pianists celebrated the creativity of performers and what they bring to any concert.¹¹ When composers and their scores are the exclusive focus on historiographic and music analytical work, it becomes all too easy to see the interpretive work of performers as nothing more than a fetish obsession, a violation of compositional intention, or a deviation. Increasingly, in the 19th century, musical scores became fixed representations of musical works. Recordings only, for some, built on this idea. For some composers like Igor Stravinsky, recording technology was embraced as a way to further exclude interpretive interference from how the “work was supposed to go.”¹² In other words, one could use a recording as an opportunity to record a definitive sonic representation of “the work,” a model of perfection.

In the views of Grainger or Brendel, however, it’s not necessarily the case that the composer is lost in the conversation but, rather, the resulting performances are framed as collaborative or co-creative efforts, where the concept of the “musical work” is less text or score-based but, instead, more like an incomplete script. When the expectations of the listener shift from any idealised sense of a “musical work” to a performance, “mistakes” are no longer necessarily mistakes but, instead, a reflection of the performer’s style, artistry, humanity, or craft or a larger reflection of the historical or cultural moment in which the performance took place. When the performer becomes the protagonist of the story, the presence of right or wrong notes can be either irrelevant or, in fact, an opportunity to honor or recognize the genius or greatness of the performer.¹³ Anecdotal examples of this kind abound.

Consider, for instance, Albert Spalding’s famous account of the Adelina Patti during one of her final concerts:

She was reckless enough to include an old-horse, ‘Il bacio’. There were notes that simply could not be reached, scales and roulades that creaked at the hinges. It promised to be lamentable. But we reckoned without Patti. When she approached a passage where the apprehended difficulty, or perhaps disaster, she employed her fan with telling results. She would start the scale or arpeggio with aplomb, the fan in her outstretched arm slowly unfolding. This would continue to register beyond which lay danger. Then with a sudden gesture, the arm would fly up, the fan snapped shut with a click, the audience would burst into a tumult of applause drowning out both orchestra and voice, and triumph greeted a fioritura or a high note that was never heard. (Scott, 1998)

The missing note is not, in this story, a performative vice even if, in the most technical sense, she failed to faithfully render the notes of the aria. Rather Patti’s embodiment of the note and her masterful and “playing” of and with the audience’s applause renders the moment nothing less than a musical triumph.

But where Patti’s dramatic performance, or the account thereof, extols the sonic and physical illusion, thus rendering the “mistake” non-existent, there are any number accounts of musicians whose memory slips or finger flubs become the catalyst for even greater improvisatory brilliance that transforms a mistake into an opportunity for the performer to shine. A story of Ferruccio Busoni having a memory slip resulted in one such improvisatory moment, one in which the aura and creative genius of the performer outstripped any need to adhere to Chopin’s score:

Finally Busoni said he would play something which he had not played for thirty-five years; sat down to the pianoforte then and there and played the C sharp minor Scherzo of Chopin. Philipp said he had never heard Busoni play Chopin like that before. In common with many of Busoni’s friends he had always found Busoni’s Chopin impossible to accept. But this impromptu performance was something quite different. It was not very accurate; it was an impression of the work, almost a transcription; he played it ‘as he remembered it, as he felt it’. But before the concert Busoni practiced it. The magic of the ‘impression’ was gone; the work was distorted by over-intellectualisation—‘he had thought it over’. (Dent, 1966, p. 267)

The description of Busoni, in many ways, anticipates the sentiment Entremont describes about Cortot by comparison with more safe or note-perfect performers. Busoni taps into an almost seance-like, spiritual connection that many performers professed to project even as they intentionally (or unintentionally) disregarded the literal representation of the script. In fact, it’s almost as though the correct notes are an obstacle to penetrating to the inner essence or spirit of Chopin’s Impromptu. Here, the composer is not ignored. Busoni was not Glenn Gould who believed he knew more about Beethoven than Beethoven or cared little about compositional intentions. The authority of the composer was respected but, as a performer, it was Busoni’s ability to “transcribe” Chopin’s music into his own performative language, not unlike an actor freely rendering a script in order to understand the deeper meaning, that was celebrated. Like Entremont, correct notes or perfection stand in, symbolically, with superficiality.

Finally, we find both stories and more contemporary recordings which celebrate the performer’s ability to adjust in the moment, especially after an error has been made or a memory slip leaves the performer in a moment of precarity. In his autobiography, Artur Rubinstein tells a self-congratulatory story about an early performance of Mendelssohn in which his memory failed him. But the failure only opened the door for a demonstration of his ability to channel the spirit of the composer:

There was a terrific ovation, with cries for an encore. Professor Barth was so nervous that he remained backstage, but now calm, and satisfied, he told me to play “The Duet,” a song without words by Mendelssohn. By this time I was completely relaxed and drunk with my victory, and I ignored all his warnings, began my piece, smiled at my friends, and thought about everything but the music. Suddenly, bang, there was a catastrophe. My mind became a blank: I couldn’t remember a single note. All I knew was that the piece was in A flat, and so, without stopping, my heart frozen, I began to improvise. I developed a theme in A-flat, all right, but it had nothing to do with Mendelssohn. After a few modulations, I invented a second subject, in minor, for contrast, elaborated it for a while, and returned to the romantic A flat. The coda was a delicate arpeggio, played pianissimo with soft pedal.

Naturally, and for a good reason, the audience did not know the piece, and I was received with the same enthusiasm as before. It was lucky, but I hardly dared to take a bow, and I was shaking with fear when I returned backstage to be slaughtered. Well, I shall never get over my amazement when I saw Professor Barth, instead of raising the ax, come happily toward me, shake my hands, and exclaim, his eyes shining, ‘Teufesjunge, you are a rascal—but a genius! I couldn’t have pulled that trick in a thousand years.’ (Rubinstein, 1973, p. 40)

Rubinstein’s story was anything but lacking in modesty and his two volume autobiography presents delightful stories that don’t always resonate with what we know of the pianist from other sources.¹⁴ Nevertheless, a concert performance of Chopin from 1964 in Moscow [Clip 3] presents a recorded moment in which Rubinstein overcame his own memory lapse in a manner similar to how the pianist described his earlier days. Not only does this “evidence” suggest that Rubinstein was more than capable of keeping his cool but that his deep connection to the style of Chopin and birds’ eye perspective on where he needed to go formally enabled him to improvise (as he suggested he did with Mendelssohn and as Busoni’s performance suggested) in the spirit of the composer, where Rubinstein, in essence, co-created the passage as though he were channeling Chopin’s spirit. Obviously, for musicians intimately familiar with the score (i.e. those who centered the text of the composer), we can see the error of Rubinstein’s way. But the fact that the audiences, perhaps less familiar with the particular Song without Words but cognizant of Mendelssohn’s general style, received the performance with as much enthusiasm as if it were Mendelssohn, suggests the pianist’s failure but anything but. And yet, even Rubinstein’s teacher, who was aware of what transpired and what was supposed to happen, celebrated his student as a “genius.”

For pianofiles and devotees of recordings in general, one might find endless examples of moments like the one described by Rubinstein, of pianists who transformed a brainfart into a singular moment that wouldn’t be replicable at any other times specifically because of how the musician recovers. Indeed, Youtube is abound with them. Some of them are similar to Rubinstein’s example. In a concert performance of Chopin’s Piano Sonata No. 3 Op. 58 from 1985 [Clip 4], for instance, Jorge Bolet lost his way in a transitional moment, one in which the sequential nature of the passage (not unlike the one that faced Rubinstein in 1964) led him astray, and simply improvised a new version that non-experts might not have recognised from the “original.”

Perhaps even more popular on Youtube, among so-called “classical fails” we find the pianist Maria João Pires faced with an unexpected crisis as the orchestra started the introduction of a different piano concerto by Mozart from the one she had prepared [Clip 5]¹⁵. Viewers of this video, numbering over 2.6 million times (on this particular version- it’s posted on multiple links with over another 800,000 views collectively), marvel in astonishment at her recall, her ability to transform in real time with on safety net, and sing her praises. It’s telling that Pires’s “fail” has 100,000’s of more views, on Youtube, than her less eventful but equally sensitive performances of the same concerto.

Pires’s fail and those in which musicians simply fall apart represent two sides of the same coin. On one hand the memory slip reminds us of the precarity of performance, the risk and tension of the moment, and the expected virtuosity of the performer. Whether it reminds us how difficult a moment is (when they play wrong notes) or the challenges of extreme memorisation, the failures humanise the musicians even when we empathetically cringe. On other hand, when we witness recoveries like those Bolet, Pires, and Rubinstein, we’re offered a singular moment in which the genius of the performer (not the composer) rewrites the wrong into a “right” and, as it were, makes musical lemonade out of lemons. Like lo-fi recordings, these performances present a kind of aura that resists the perceived aesthetic of recordings as sterile and conceals their nature as mass produced and distributed.

5. Embracing Mistaken Possibilities

Even with Rubinstein’s improvisation or the celebration of Patti’s unsung high note, the celebration of the performer results in the celebration of an illusion, the channeling of a compositional spirit or the projection of unheard songs. How might one, however, consider these mistakes as simply another part of the sonic “text,” as it were. Richard Beaudoin has argued, for instance, that when we listen to recordings, whether from concert performances or studio recordings, that we embrace ALL of the sounds that we hear, whether they be the clicks of the fingernails, the breaths or grunts of a performer, or the squeaks of a chair under performers as they play. There is a value, Beaudoin claims, to recognizing “sounds as they are” (Beaudoin, 2024).¹⁶

Like many of the responses described earlier, Beaudoin emphasises how these non-notated sounds amplify the human-ness of the moment, the traces of the space and body that help remind a listener of the performer’s presence even if presented, on record, as a disembodied sound. Beaudoin, however, also encourages listeners and analysts to consider the ways in which these “extra-musical” sounds might interact with “the music itself” in such a way that these two categories (that I have intentionally framed in scare quotes) become blurred together. In other words, what we hear is what we hear and, as recordings, the variety of sounds will not change from one performance to another.

I would like to embrace Beaudoin’s approach to recordings but in a manner that doesn’t necessarily integrate “sounds” with “music” but, rather, asks us as listeners to consider the possible “rightness” of wrong notes, that is, to embrace the aesthetic possibilities that come from musical mistakes. The remainder of this article is, above all else, an application of this approach, taken from personal experience. I do not presume that my experiences are more or less correct than any other listening or that anyone else hears things the way I do. Instead I hope that, instead, readers might explore the possibilities that come from re-orienting more conventional habits of listening and judging.

Consider a concert performance of Ravel’s Tzigane, recorded in 1949 with the violinist Ginette Neveu with the New York Philharmonic, conducted by Charles Munch. About halfway through the solo violin part that introduces the work (described by the composer as a “quasi cadenza”), Ravel develops a dramatically virtuosic buildup to a reprise of the opening bars in B minor. The harmonies are relatively simple but conventionally effective, as a predominant E-minor builds to an F# dominant double stop that uses the subsequent rest (drawn out by a fermata) to enhance the anticipation of the tonic reprise (see Example 1). The virtuosity of this passage draws attention to the performer’s skill, their rapid movement across strings and from high to low, and the swooping glissando (a momentary non-tonal gesture) that culminates fiercely on the leading tone, A# before building right back up, through minute chromatic alterations, to a dominant chord.

It is Neveu’s performance of this chord, when I listen to this performance, that always delighted me. Three years before her recorded concert performance, Ne veu recorded the composition in the studio, accompanied by her brother, Jean [Clip 6]. This recording, too, is intense beyond belief (as was in keeping with the violinist’s musical temperament) even while her rendition of this particular passage is, as it were, “faithful” to the score. That is, she plays the notes on the page.

But in the concert performance [Clip 7], Neveu struggles with the dominant chord and doesn’t land on it as cleanly or neatly as she did in the studio or as the score suggests. Instead, she falls back from the initial F# to an unwritten E natural before pushing back up to the correct F#. It is not, like Busoni or Bolet’s performances, an improvised moment to cover up the error. It is clearly a mistake, and not enough to be passed off as graceful ornament. But in the context of the moment, it’s marvelous and as a listener, I’ll listen to this performance over other performances a dozen times over.

It’s not only because Neveu is human. Rather it’s what the mistake brings. The entire passage is one of building tension, of growing dynamically, performatively, and harmonically to a precipice. The slip demonstrates the danger, how close Neveu was from almost completely falling apart: except she doesn’t. Like a tightrope walker wavering in the wind, Neveu’s slip shows us how close to failure she might have come. Instead of collapsing, she surges back, the E to which she falls driven back up with forceful determination, correctly playing the dominant chord as we expect but with even greater purpose. Based on her earlier performance, Neveu is likely not screwing this up on purpose. But the effect feels, at least to my ears, so much more “right” than when she doesn’t because of the amplified tension of the moment.

Is it possible, however, that musicians would make mistakes on purpose. And, if so, do we still conceive of them as mistakes? Anna Scott, for instance, in her work on re-creations of early pianists challenges us to rethink “what we actually mean by mistake” whether these “errors” are in fact accidents, habits, or deliberate (Scott, 2023). In keeping with this question, consider the Australian pianist, Percy Grainger, discussed earlier. He had a bewilderingly contradictory and complicated relationship to the Grieg Piano Concerto Op. 16. On one hand, Grainger was one of the early proponents of the concerto while also using it, and the music of Grieg, to help launch aspects of his career. He first performed in 1905 and made a recording in 1908 of the first movement cadenza, a year after the composer died. A little over a decade later, he recorded a piano roll version, one that has been re-performed multiple times with modern symphony orchestras in the last few decades [Clip 8]. Furthermore, when Grieg heard Grainger perform the concerto, he was reported to have responded with wonderful enthusiasm. On the other hand, Grieg acknowledged that his interpretation was not what he had necessarily expected or intended. And, yet, ironically it is Grainger’s edition of the score that has become one of the most used versions over the last century.¹⁷

There exist, also, multiple recordings of the Grieg concerto made late in Grainger’s life. Grainger’s reckless and rambunctious approach to the pianist remains on full display in these performances and, like his idol, D’Albert, wrong notes were tossed off left and right. There are, however, also some odd moments that fall somewhere between a mistake and an intentional deviation from the score. Towards the end of the third movement (see Example 2), Grieg indicates a hemiola to interrupt the flow of the triple metre. The pianist is instructed to play the hemiola with sforzandos which, in turn, are emphasised by the rhythmic placement of the orchestra. While Grainger maintains these markings in his edition of the score, he chooses, with every performance, to ignore the sforzandos and pushes ahead with the triple metre. This results in an even greater metric tension, as the hemiola of the orchestra grates against the triple patterns of Grainger’s playing [Clip 9]. But it’s more than just a momentary error; It’s clear from multiple versions that Grainger does this on purpose. Perhaps to amplify the tension? One can only speculate.¹⁸

At the start of the same movement, however, Grainger does something even more error-ridden, at least on the surface. After a brief orchestral passage, the piano comes in with a torrential, two-handed arpeggio that harmonises the predominant augmented sixth chord before resolving to the cadential 6/4 chord (see Example 3).

Even if Grainger’s playing had declined in the later age (and it had), and even if he was prone to making mistakes now and then (which he did), it strikes me as odd that in each performance of this arpeggio, the pianist seems to play all the notes that aren’t notated as well. Here [Clip 10] is Grainger in 1945, with Leopold Stokowski at the Hollywood Bowl, from 1950 and here with Richard A. Morse and the Southeast Iowa Symphony Orchestra from 1956 [Clip 11] and, finally, from 1957 with the Aarhus Municipal Orchestra with Per Drier on the podium [Clip 12]. The effect is surprisingly dramatic, which would appear to be the effect of the bold announcement by the piano in the first place.

In short, it begs the question, did Grainger just screw it up or intentionally “enhance” the sensation of the moment?¹⁹ My answers are both personal and speculative, although not entirely historically ungrounded. On a personal level, not unlike my response to Neveu’s amplification of the double-stopped dominant, I adore this moment for a variety of reasons. First and foremost, it is unexpected and very different from the dozens and dozens of other recorded performances of this canonic work. In the spirit of Glenn Gould who argued that there is no point recording a composition unless one had something distinctive and new about it, I find Grainger’s bash through arpeggio exhilarating, not only because of the reckless energy but because it has the very real effect of shocking me more than the more “coiffured” performances of the concerto that I had come to expect. Along the same lines and in the spirit of Busoni, Grainger’s spectacular rendering of this bar has the same aesthetic effect as Grieg’s original notation. In that sense, Grainger might be thought of as having changed Grieg’s music in order to enhance Grieg’s compositional intention.

But for years I have speculated about this moment, first introduced to me in the studio of Ward Marston in the late 1990s when he transferred the 1945 performance for release on Biddulph. I wondered if there might be a connection between the lateness of these performance (i.e. late in his life) and his developed thinking about “free music,” that is the desire to create music and sound that escaped the parametres of conventional musical systems (e.g. an augmented 6th chord). In 1938 Grainger wrote: “Out in nature we hear all kinds of lovely and touching 'free' (non-harmonic) combinations of tones, yet we are unable to take up these beauties and expressiveness into the art of music because of our archaic notions of harmony” (Sowa, 2019). Moreover, Grieg’s own music had, for Grainger, connections to this liberating aesthetic. Malcolm Gillies and David Pear remind us that “[i]n his twelve broadcast lecture-recitals for the Australian Broadcasting Commission during 1934-5, he mentioned Grieg in at least six of them, in particular seeing his interest in nature music, wailing sounds and gliding tones as foreshadowing Grainger’s own development of a ‘Free Music’, that is, music freed from the fixed steps both of pitches and rhythms” (Gillies & Pear, 2007b). Additionally, Grainger, in his piano editions, encouraged a kind of recklessness or, at the very least, unconventionally extreme techniques. Consider, for instance, his famous Country Gardens. Grainger instructs the pianist repeatedly to play the moment “violently” or, when rolling chords in the left hand, that they be “violently wretched.” And at the very end of the piece, he solicits an almost percussive, non-tonal sfff, by telling the pianist to play a note with their fist (Grainger, 1919).

In short, Grainger embraced, in part from Grieg, non-harmonic or non-tonal experimentation that was in keeping with his desire to develop a “free music.” As a pianist, going all the way back to his response to D’Albert and his virtuosic and pianistic arrangements of folk songs, Grainger pushed the limits of the instrument, especially to see in what ways it could generate sounds beyond the more traditional Romantic language of Chopin or Schumann (whose music he played). There is, to my knowledge, no evidence suggesting that Grainger made specific connections to this aesthetic of free music and this particular interpretation of the Grieg concerto. But given the fact that he bangs his way through these arpeggios with such abandon, such disregard for the individual notes but with a keen ear to the dramatic purpose of the passage, I cannot help speculate that even if there weren’t an intentional connection on the part of Grainger, that my knowledge of his aesthetic shapes my way re-thinking and re-hearing this cacophony of notes as a spectacular freeing from tonal constraints, a temporary moment of chaos that is then recaptured by the cadential 6/4 chord, the descending scale, and the commencement of the A-minor dance that follows. In other words, his wrong notes don’t feel wrong, his mistakes feel right, and Grainger’s exploration of “free-er” music is on full display.

6. Conclusion

I would like to visit Alfred Cortot one last time, by way of concluding, in order to explore a later recording (1957) of Chopin’s Ballade n. 2 Op. 38. Like Heifeitz’s recording of the Sibelius Concerto, Cortot did not consent to the recording’s release at the time, imperfect as it was. Like Grainger when playing Grieg, Cortot’s technical facility was failing him. And in keeping with the many descriptions that we’ve already read of Cortot, his performance is anything but note perfect, especially compared with his studio recording from two decades earlier. Moreover, as an inverse to Patti’s illusion, Cortot’s “mistakes” open up the possibility, on the part of the listener, to perceive things that for most never previously existed.

Cortot, like many pianists, was keen to highlight internal or unexpected melodies from within complicated textures. It is a style of performance that transcended stylistic schools in the early twentieth century, something one finds in the playing of pianists like Josef Hofmann, Vladimir Horowitz, and Leopold Godowsky (indeed, all three of these pianists, in their arrangements of other music, were celebrated by the way they wove in new and unexpected countermelodies) as well as more recent pianists like Glenn Gould, Vladimir Feltsman, or Awadagin Pratt. The first few minutes of the Ballade only encourage the exploration of inner voices. It is written in four (or occasionally five) voices that maintain a relatively homophonic texture (not unlike a chorale or hymn). A comparison of performances reveals a delightful variety of paths through this counterpoint. Some pianists select individual motives, others highlight certain voices in order to create the illusion of dialogue, and others, like Cortot, weave a variety of voices together, moving in unexpected ways from top, to bottom, to middle voices.

It is when Cortot gets to the Presto con fuoco (Example 4) that matters go from “merely” gorgeous to revelatory. At this moment the technical requirements of the composition ultimately outstrip the aged pianist’s ability. The torrential passagework in the right hand is, to be candid, a mess. And I would guess the pianist was well aware that his right hand couldn’t keep up. As a result, he often smudges through many of the sections in order to maintain the momentum. The resulting performance of the right hand might be imagined as somewhat painterly or impressionist, the gesture and the effect of the arpeggios taking necessary precedence over the correctness of the notes. These are some of the effects that, most likely, pianists like Brendel and Entremont speak of.

The re-imagining of the right hand, however, has a remarkable effect. Like what happens with an amputation (in the Deleuzian sense) or the loss of a particular sense, Cortot’s chaotic right hand forces his left hand to take up the slack and to compensate for a loss.²⁰ In his earlier recordings (1929 [Clip 13] and 1933 [Clip 14]), Cortot plays m. 48 (Example 4) (and similar passages later in the Ballade) in much the same manner as other recorded pianists, namely, Cortot synchronises the semiquaver pattern of the left hand with the semiquavers of the right hand, creating a multi-layered but rhythmically and texturally homophonic and homorhythmic ascent, almost like climbing a ladder or stairs (while maintaining the sense of the 6/8 metre) back to the start of a new wave.

In the recording from 1957 [Clip 15], however, the right hand is a shambles. Notes are missed, rhythmic articulation is distorted. To make up for this loss, however, Cortot transforms those notes played by the thumb, in his left hand, to generate an entirely new alternative melodic path through this passage (see notes circled in m. 48 of Example 4). The effect is remarkable. Not only do the notes in the left hand, previously subsumed by those played in the right, emerge with melodic potential but Cortot also, intentionally or otherwise, reveals an increased rhythmic complexity to the pattern, that is, the possibility of a 12/16 pattern or the insertion of a four-beat per bar phrase from within the otherwise complex triple metre. Ironically, it is only by dissolving the strength of the right hand (by smearing the notes) and separating the left from the right hand, does this new possibility come to exist. And, yet, at least within the left hand, Cortot articulates notes that are, in fact, written in Chopin’s score even if, in the hands of a younger Cortot and most other pianists, their coherence, melodic potential, or dialogic expression is otherwise lost when played “correctly.”

This final relationship between score and performer is perhaps what makes this moment distinctive to Cortot’s examples and many of other performances discussed in this essay. While in non-classical musics, mistakes can and have been differently framed, often as serendipitous moments that become iconic in their own right, classical musicians, even those that seek to challenge tradition [Clip 16], are pushed to adhere, much more “faithfully” to the notes in the score, where missed notes, errors, memory slips are, more often than not, a source of condemnation.²¹ What I hope to suggest, however, is that such a narrow view of “wrong,” “imperfect,” or “mistake” can, if listened to with a more open ear, an attention to the performer, or an awareness of audience delight, be reimagined as “right” if only because of the new possibilities they reveal. It is not just adherence to the notes of a score that grows stale, especially for such canonic works, but that disassembling, reframing, or re-arranging by mistake shifts our attention to the immediate intentions and needs of the performer to create something dramatic, powerful, provocative, and moving, often in ways that transcend what we might have expected in the first place.

Although most of this article has focused on aesthetics and listening, it is my hope that performers might take solace in knowing that embracing “mistakes” or re-imagining interpretive possibilities, could be freeing for classical performers. Not only, as Daniel Leech-Wilkinson (2023) suggested, does this new aesthetic liberate performers from the yokes of tradition or the gatekeeping of critics, but it is my hope that performers and pedagogues might also welcome mistakes as simply a part of the music making in general. The striving for perfection and the grading down of errors in musical festivals, auditions, and competitions, can be damaging to the mental health of performers and, more importantly, to the overall enjoyment of making music. If the recordings discussed in this article show us anything it is that there is a long history of musical precedents that celebrate the performance of imperfection and, moreover, a similarly long history of celebrating mistakes as well.

Acknowledgements

I would like to express my greatest appreciation for Mine Doğantan-Dack who originally invited me to write this essay on a topic I've been percolating on for over a quarter century. I'm also grateful to Antonio Grande's guidance and patience, the comments from the two anonymous reviewers, and the encouraging and critical support of Richard Beaudoin, Adam Behan, Kelly Gross, Cecilia Oinas, and my family.