“You can be early, you can be late, the two hands are not in phase; then you make a compensation which re-establishes the ensemble.”

George Mathias (pupil of Chopin, teacher of Raoul Pugno and Isidor Philipp)

This article examines closely three historical recordings of Chopin’s B minor Prelude Op.28 No.6, made by Vladimir de Pachmann (in 1927), Moriz Rosenthal (in 1935) and Raoul Koczalski (in 1939), with extensive audio and musical examples. Both Pachmann and Rosenthal had contact with Liszt, who could apparently imitate Chopin’s playing very well, Rosenthal in particular receiving extensive instruction from him. In addition, Rosenthal and Koczalski were taught as boys by Chopin’s disciple, Karol Mikuli.  Clearly then these pianists deserve to be taken seriously from a stylistic point of view.  The B minor Prelude is chosen for its brevity, the relative simplicity of its texture, and for the similarities of approach noted in the recordings, all of which use asynchrony as an expressive device in stark contrast to the majority of modern recordings.

Asynchrony is a general term which is used to describe playing notes in a separated or not-quite-together fashion where they are written as if they should normally be played at the same time in the score, for example a chord to which an arpeggiation is applied, or a left-hand bass note and right-hand melody note both written on the same beat but actually played with one hand being placed slightly before the other. 1 It is apparent on many early recordings made of pianists who were born in the nineteenth century and has been the subject of detailed analysis in recent years (see Peres Da Costa’s ‘Off the Record’ cited in the bibliography).  It is an area of performance practice that I find personally very interesting for its role in some of the most exquisite and in other instances most eccentric-seeming performances recorded by such artists.  By incorporating it into my own playing I have found it of great effectiveness in realising the music of Chopin in particular.

After sketching some important background information on the earliest-born of these pianists, Pachmann (b.1848), I map out the asynchrony in his recording, representing it in score form and attempting to reproduce it on a modern piano. I then go on to compare his recording with the other versions by Rosenthal (b.1862) and Raoul Koczalski (b.1884) noting various similarities.  I refer to a number of theories that may account for how and why these pianists feature asynchrony, including its role in articulating: metre and phrasing, accents, and enhancing the audibility of the melodic voice.  I argue further that some form of asynchrony must have been an integral aspect of Chopin’s rubato where accompaniment and soloist were not tethered precisely together.  Such ‘Classical rubato’ is suggested in some passages from piano roll recordings by Saint-Saëns and was still practiced in a form captured on record by duos performing in the early twentieth century.  Finally, recordings of the Prelude by Alfred Cortot and Mark Hambourg, in which points of asynchrony are scarce, but still apparent, are used to reflect further on the gradual changes of style. A further recording of great interest by Robert Lortat reveals yet another highly asynchronised version from 1928.

Vladimir de Pachmann

Vladimir de Pachmann’s birthdate (1848) overlaps with the final year of Chopin’s life and although this obviously cannot count for any direct influence, the sense of a very distant and at times unfamiliar form of pianism may surprise the otherwise seasoned Chopin listener who delves into Pachmann’s recorded legacy.  Pachmann’s decidedly eccentric stage antics have proved something of a distraction from serious appraisals of his playing, yet his artistry has been re-evaluated in recent years, and his complete recordings reissued with restored sound.  His reputation as the Chopin player and specialist of the late nineteenth century was built on a lifetime of accumulated study and experience, first with Joseph Dachs at the Vienna Conservatory, and then through witnessing the playing of Tausig, Anton Rubinstein and, in particular, Liszt.  No less a pianist than Moriz Rosenthal, whose recording of the B minor Prelude is compared with Pachmann’s later in this article, visited Pachmann on one occasion with the purpose of gaining insight into his Mazurka playing2.  What is more, he was widely respected as a Chopin interpreter by other great players of Chopin such as Godowsky, Ignaz Friedman and even Liszt himself.3

Mention should also be made of research carried out by Allan Evans, who interviewed Roman piano teacher, Aldo Mantia.  The latter claimed to be Pachmann’s student in the late 1920s, and maintained that Pachmann sought advice from one of Chopin’s teaching assistants, Vera Rubio, in 1879, facilitated by a year-long stay in Florence.  This was during Pachmann’s break from the stage in which appears to have done nothing but practice for a period of about six years, and during which there is little evidence of his whereabouts.  The story of this encounter with Rubio is contested as nothing more than a ‘rumour’ by Pachmann’s biographers, Blickstein and Benko, due to the lack of any other corroborating evidence, which they would have expected to unearth during their own course of research (Blickstein and Benko, 409-410). 4 

Pachmann’s Recording

Provided in Ex 1.0 below is an annotated score showing how Vladimir de Pachmann used asynchrony between the hands for his recording of 1927 (Victor 1459).  Similar examples throughout this article allow the reader to compare these features with those of other pianists. In the recording below Pachmann’s touch is fascinating but it is probably the asynchrony that is perhaps most striking to a modern listener.

Ex 1.0 Asynchrony in Pachmann’s Recording (1927)

Listening to this recording, which is available on the YouTube link above, and consulting Ex. 1.0, reveals the surprising frequency of notes in the left-hand melody that do not coincide precisely with the right-hand accompaniment.  Altogether there are about twenty-five instances, in addition to which arpeggiations are added to the chord at the beginning of bar 6 and in the middle of bar 8, neither being indicated in Chopin’s score.  The diagonal lines in Ex 1.0 show whether the left-hand notes are early or late in relation to the right hand, and the placement of note-heads in the score have been adjusted slightly to reflect the nature of the displacement.  From this it can be seen that of these, the majority of displaced notes in the left hand melody are played before the right hand (twenty in total) whereas five notes are played a after the right hand.  The degree of displacement has not been quantified exactly, but some displacements are certainly more obvious than others due to the varying time lag between the hands (for example, the downbeats of bar 1 and 4 are particularly striking, whilst the downbeat of bar 3 is less noticeable because the time-lag between hands is only slight).  Moreover, in bar 7 the displacement of the quaver A# in the left-hand actually produces a different rhythm to that of the original score.  The question of whether a given melodic note is early or late relative to the prevailing pulse is and interesting one, but again such information is not encapsulated in the diagram, and to an extent it is hard to be objectively precise about such a judgement.

I recorded the piece myself on my Weber 150 piano, copying Pachmann’s playing as far as possible.  Most of the final edit was taken from a recording in which I played alongside Pachmann’s recording in order to maximise the correspondence in timing.  I also used annotations in my own score to try to reproduce as many details as possible, such as pedalling, touch and rubato, in addition to the asynchrony.  The aim of the recording is firstly to show the overall effect in modern sound, although as this is done in my home studio there is scope for greater opulence in the recording quality and in the quality of the instrument itself.  Secondly, it allowed me to enter into the spirit of this different style of playing, hopefully leading to greater understanding of the style. 

Pachmann’s 1927 recording of Chopin’s B minor Prelude reconstructed on a modern baby grand by Charles Tebbs.

Many pianists new to this style may be surprised or indeed perhaps irritated by the effect of this kind of constant asynchrony, and wonder what effect Pachmann was trying to achieve.  However, note that this sort of thing was quite normal in his playing.  Greater familiarity with recordings and pianists of this period will be required until the aesthetic of the effect can truly be appreciated, and its possible relation to Chopin’s style taken seriously.  For those who are open minded it is worth listening a number of times and considering the extent to which the asynchrony allows Pachmann to play the regular quaver ostinato fairly steadily, whilst the left hand melody pursues its own individual rubato.  This is of great significance when considered alongside reports of Chopin’s own playing.  Here are two such reports, though there are several more (see Eigeldinger, 49-52):

‘…the hand responsible for the accompaniment would keep strict time, while the other hand, singing the melody, would free the essence of the musical thought from all rhythmic fetters, either by lingering hesitantly or by eagerly anticipating the movement with a certain impatient vehemence akin to passionate speech.’

(Mikuli, 3)

‘Chopin, as Mme Camille Dubouis explains so well, often required simultaneously that the left hand, playing the accompaniment, should maintain strict time, while the melodic line should enjoy freedom of expression with fluctuations of speed.  This is quite feasible: you can be early, you can be late, the two hands are not in phase; then you make a compensation which re-establishes the ensemble.’

(Mathias, 5)

We will return to this notion later.  However, now let us turn to other pianists that recorded the piece, two of whom were taught by Mikuli, Chopin’s assistant, whom I have just quoted, and both of whom do indeed also exhibit similar kinds of asynchrony at similar points in this work. 

Rosenthal’s Recording

Both Rosenthal and Koczalski (see below) were students of Karol Mikuli. The significance of this in terms of pianistic style can hardly be overstated, and in addition to similarities identified in this essay between Rosenthal and Koczalski, there is further evidence of similar styles of asynchrony in the playing of another Mikuli student, Michalowski, though sadly he did not record the Prelude in question5. According to Eigeldinger, ‘of all the students of Chopin who took up the profession, [Mikuli] worked the hardest to hand down his teacher’s tradition’ (Eigeldinger, 172). Between 1844 and the 1848 Revolution Mikuli studied with Chopin in Paris, was among his official copyists, and was also allowed to attend the lessons of others. He toured Eastern Europe as a pianist, settling in Lviv in 1858 and becoming director of the Conservatory there. In 1880 he published his edition of Chopin’s works in seventeen volumes, still considered authoratative, taking into account the advice of the aforementioned Mme Rubio and other important students, as well as the annotations on his own French editions that had been added by Chopin. We have already heard from Mikuli how Chopin would have played melody and accompaniment in differing execution between the two hands, but specific indications are lacking. Indeed, Cortot says even less of this in his editions, even though he used asynchrony frequently. It seems it could only be transmitted as an oral (or aural) tradition, not put simply into words.

Rosenthal studied with Mikuli between 1872 and 1874 when still a boy of only ten, and recounts in some detail the nature of the instruction in his partial autobiography (see Mitchell and Evans, 25-28) although his comments do not imply that teacher and pupil concentrated particularly on Chopin’s music. It is possible that his lessons with Liszt were of greater importance in this repertoire. Rosenthal says little about it, but Pachmann recounted that Liszt ‘showed him many secrets – nuances and pedal effects – that Chopin had used himself’ (Blickstein and Benko, 38). Rosenthal’s studies with Liszt took place in two phases, 1876-1878 and 1884-1886.

There are a total of eight recorded takes of the Prelude made by Moriz Rosenthal on three separate occasions (see Mitchell and Evans, 151), six of which are available (Spotify has Rosenthal’s ‘complete recordings’), and there are strong similarities between these.  For this purpose I have selected one recording by him for close examination, recorded in 1935 (Victor 14300) when the pianist was sixty-three.  There is an atmosphere of resignation and stillness in the way this performance moves that is quite captivating.  Rosenthal draws out great breadth in his phrasing, with ritardandi and tenuti used at several points, marked in the example below.  He also adds some further rhythmic character of his own, overdotting the dotted rhythms (or at least giving the effect of shortening the semiquaver a little). 

Ex 1.1 Asynchrony in Rosenthal’s Recording (red) Compared with Pachmann (black)

Moriz Rosenthal 1935 (Victor 14300)

As a student of Liszt, Rosenthal might be expected to demonstrate an affinity with Pachmann in his recording of the piece, and indeed it does exhibit a similar number of points of asynchrony – a total of twenty-eight displacements of the hands!  Example 1.1 shows his use of asynchrony in red, overlaying the earlier example of Pachmann (his displacements are still shown in black).  From this it will be seen there are about twelve identical locations where asynchrony is used by both pianists, although sometimes the order of the hands is exchanged at these points (for example, where the red line in the middle of bar 3 crosses in the opposite direction to the black line, it shows the left hand F# being placed after, rather than before the right hand in Rosenthal’s recording).  Dotted lines have been used to show very slight asynchronies; these were checked by listening to the recording at reduced speed.  

As with Pachmann’s recording the majority of these instances are of the left hand played before the right hand, with seven that work the other way around (only four of these are articulated with a very noticeable time lag between the hands, however).  The quantity of displacements is comparable with Pachmann’s, but the quality is slightly different, with a less noticeable time-lag between the hands generally as well as the even slighter asynchrony noted in the dotted lines. The broader phrasing gives more of an air of melancholy here than in the urgency of Pachmann’s rendition, and some may prefer it for its more held-back style of rubato and also the more transparent quality of the sound.

Koczalski’s Recording

Raoul Koczalski (1885-1948) was a Polish pianist-composer and teacher.  He became Karol Mikuli’s protégé as a young boy, attending daily two-hour lessons for four successive summers from 1892 to 1895.  Koczalski was no ordinary child; by the time his studies with Mikuli began in 1892 he had already performed his own opus 46 and experienced the life of a touring child prodigy.  Jacques Eigeldinger, a significant scholar in the field of Chopin and performance practice, who created a new French edition of Koczalski’s book on Chopin as well as compiling the fascinating ‘Chopin: Pianist and Teacher…’ (see bibliography), is himself an advocate of Koczalski as ‘the one to have preserved this living tradition most purely.  In this respect his recordings constitute a document of prime importance.’ (Eigeldinger, 97).  Robert Philip, an expert and engaging writer on the history of recording in classical music, agrees:  ‘given the many descriptions which Chopin employed in his own playing…it seems quite likely that Koczalski’s freedoms to some extent derived from that tradition, even though they must undoubtedly be his interpretation of such freedoms rather than Chopin’s or Mikuli’s.’ (Philip, 187).

Ex 1.2 Asynchrony in Koczalski’s Recording (green) Compared with Pachmann (black).

Raoul Koczalski 1939 Deutsche Grammophon / Polydor (67505B).

Example 1.2 shows Koczalski’s displacements in the Prelude in green as against Pachmann’s during the first twelve bars of the piece.  This is based on Koczalski’s only recording of this piece, part of a complete cycle of the Op.28 Preludes he recorded in June 1939 for Deutsche Grammophon / Polydor (67505B).  From this it can be heard that there are some similarities with Pachmann, although the total number of instances of asynchrony is reduced considerably during the first twelve bars from around eighteen in Pachmann (and a similar number in Rosenthal) to only around seven in Koczalski. 

Both Pachmann and Rosenthal begin the second phrase in bar 3 by placing the left hand B before the right hand, and this note is also marked out by Koczalski with a notable dynamic accent as well as asynchrony.  Bars 7 and 8 are also points where asynchrony is used more densely, in the form of melodic displacement and arpeggiation.  Koczalski arpeggiates the first chord of bar 8 (where Pachmann arpeggiates the third chord).  He also plays the grace-note at the end of the previous bar in a style that Chopin himself advocates, writing it in one of his student’s scores: there should be an arpeggiation upward to the grace-note from the bass (see Eigeldinger, 109 where he gives an example from the Dubois score).  The low G initiating the phrase in bar 11 is played before the right hand by both pianists, and in the same manner by Rosenthal and Cortot (see below). 

Koczalski sometimes uses only the slightest whisper of asynchrony between the hands, as with the high left hand G of bar 5 which is placed almost undetectably after the right-hand accompaniment.  Pachmann also displaces this note, but more obviously, and before the right hand rather than after it.  There is also just a hint of asynchrony with the high tenor E of the left hand in the last bar of the Koczalski example (not marked).  The rest of the piece, although it does contain displacements in Koczalski’s version, does so with such slight discrepancies between the hands that it is difficult to demonstrate clearly what is going on here, so I have omitted it from the example.  But perhaps the fact that the asynchrony is so fractional here may represent something Chopin would have done.  If the impression here is of a more restricted used of asynchrony compared with the other pianists, elsewhere Koczalski’s interpretations are thoroughly on a par with Pachmann’s recordings in the quantity and  degree of the effect (for example, the versions of the Nocturne in B major Op.32 No.1 or the Raindrop Prelude).


The early placement of the bass or left-hand accompaniment relative to the right hand (or the late arrival of the melody, depending on the nature of the particular example and how one hears it in relation to the pulse) is the most obvious form of asynchrony in early recordings.  The opposite presentation, with the bass or accompaniment arriving after the melody is more unusual, although it is also perhaps less noticeable, and some artists became quite adept at slipping it in (for instance, it happens quite frequently in Ignaz Friedman’s famous Mazurka recordings, and Hamilton has identified it as a feature in Pachmann’s recording of the Raindrop Prelude (Hamilton, 152)).  The more understated nature of left-hand-after-right-hand asynchrony may account for its use by many modern pianists (listen to the opening of Hélène Grimaud’s recording of the Raindrop Prelude, which can be slowed down on Youtube to hear the effect more clearly).  But why use asynchrony at all?

There is an increasing body of research into the employment of asynchrony by pianists which is based on practical experiments using pianos that are hooked up to computers to record timing data, whilst also using interactive ways to test the response of listeners.  This research has certainly demonstrated the importance of asynchrony in piano playing, whether players are conscious of it or not, yet further research also indicates that slight asynchrony may merely be a by-product of the melodic voice being played louder, the hammers that articulate the melody therefore arriving slightly before those of the accompaniment (see Goebl).  Moreover, most of the discussion has centred on the notion of ‘melody lead’, where the melody arrives first, slightly before the accompaniment.  In these early recordings, however, the predominant kind is of an opposite order. Nevertheless, some findings seem to indicate that listeners hear melodic details more clearly or with a greater sensation of accent when particular notes are desynchronized in this way from accompanying chords.  ‘A melody that was both delayed and louder in intensity was rated significantly louder than a melody that was simultaneous and louder’ (Goebl and Parncutt, 1).  This coincides with my personal sense of the effect.

Robert Philip has also drawn attention to a case where accents are interpretted using asynchrony in the form of hand displacement instead of using markedly heavier touch, something that was perhaps more effective on a pianos of Chopin’s period where a heavy touch easily became harsh and mechanical.  Philip refers to the melodic bass line in bars 23-24 of the G minor Ballade Op.23 (Example 1.3), where he observes that Koczalski places the notes that have accents slightly before the right-hand accompaniment (this is the case in both of his recordings of this work, incidentally).  In fact, from my own observations this passage was rendered in a similar fashion by other pianists of the period: Benno Moiseiwitsch, Mark Hambourg, Horszowski (all Leschetizky students), Cortot (in all his versions of the work) and Horowitz (in his 1947 RCA recording).  Here then is another instance where the left hand, when it had melodic material, was distinguished from the right hand by subtle desynchronization.  From an acoustical point of view this avoids what is called ‘masking’, the concealment of one sound by another (and any music teacher will know the difficulty students have in hearing and singing back the lower of two voices compared with the upper one).  Could that also explain why it is so prevalent in the B minor Prelude, where the left-hand could otherwise be masked by the right?

Ex 1.3 Asynchrony in Chopin’s Ballade No.1 as played by Moiseiwitsch, Horszowki, Koczalski, Hambourg, Horowitz and Cortot

Conversely, the right-hand melody can be made to stand out and ‘sing’ if the left hand is struck at a slightly different point in time: both Hamilton and Peres Da Costa refer to Malwin Brée’s Die Grundlage der Methode Leschetizky of 1903 in their discussions of asynchrony.  Brée was Leschetizky’s assistant, and he seems to have given full endorsement to her book. Leschetizky (1830-1915) may not have heard Chopin play but he became close friends with one of his most talented students, Carl Filtsch, who died young.  Hamilton quotes the relevant advice:

The fundamental bass note and the melody note must also not always be taken at the same time; rather, the melody note should be struck shortly after the bass, by which method the melody rings out more clearly and sounds softer. Nevertheless, this may happen only at the beginning of a phrase, and mostly only with important notes and strong beats of the bar (Weak beats are better struck precisely together.) The melody note must be brought in so quickly after the bass that this is hardly noticeable for lay listeners, for example Chopin’s Nocturne [in D flat]

(Brée, quoted in Hamilton, 144)

Reflecting on this in the light of the recordings examined here, of all the points of asynchrony in Pachmann’s recording, 14 (about half) do indeed occur on downbeats, though not only in places at the beginning of a phrase.  Every single phrase also begins with such asynchrony, however (interpreting the phrases from Chopin’s slurs), and even though the melody is in the bass in this case, it is still the bass that enters before the right-hand in the majority of cases (the phrase in bar 19 begins with an anacrusis and it is on the anacrusis entry that Pachmann places the asynchrony here as seen in Ex 1.0, reversing the order of hands on the downbeat of bar 19 itself). In Rosenthal’s version too, 10 displacements occur on downbeats, and all the phrase entries except bars 15 and 19 are marked in this way.  In Koczalski’s recording all except one displacement of the hands occurs on the downbeat and one of the two added arpeggiations is also added in this location, and he twice places asynchrony at the start of the phrase (bar 11 receives a new phrasing slur in the original manuscript, an oversight in the examples reproduced here). 

Considering these observations in the light of Brée’s statement we can see that her advice is only partially reflective of actual practice, in that there are many other places in the phrase and in the bar where asynchrony occurs, not just at the start of phrases or downbeats – Koczalski’s recording sticking more closely to this principle, however.  To be fair, none of the artists considered thus far was a student of Leschetizky, although Leschetizky’s own piano roll of the Nocturne in D flat is scrutinised closely by Peres Da Costa and found to contain many other instances of asynchrony than those that might be inferred from Brée’s written advice.  This then calls into question the extent to which an advanced player would really be restricted to downbeats or phrase entries in their use of hand displacement.6  Nevertheless, the use of asynchrony on downbeats lends weight to the idea that at least one of its functions for a pianist was to enable the creation of a metric type of accent.  This is explored in some depth in Mark Arnest’s article (see bibliography).  He refers to the notion of a ‘downbeat area’ being spread between the hands rather than a precise moment of downbeat.  The idea certainly makes sense to me, having attempted to copy Pachmann and incorporate asynchrony of this sort in other pieces.  ‘One effect of any type of noncoordination is to expand the beat, giving it a sense of duration; that is, the romantics favored a pulse-like area over a point-like beat.’ (Arnest, 3) And he adds:

When used melodically, a noncoordination was usually an accent. Earlier pianists were especially fond of marking downbeats by delaying the melody slightly. This clarified the rhythmic structure without resorting to a crude dynamic accent. Especially expressive moments, such as melodic leaps or dissonances, were also often delayed. Even on weak beats, romantic pianists often used a melodic delay as a substitute for – or an intensifier of – a dynamic accent. This gave romantic performances extremely flexible emotional character, with subtle distinctions between different kinds of accents.  

(Arnest, 3)

This idea that asynchrony can create an accent resembles another more familiar one, the ‘agogic accent’, which is achieved by a slight lengthening of a note instead of (or sometimes in addition to) a dynamic accent, and is a well-known technique in harpsichord and organ playing where dynamic accents cannot be produced by striking the key harder.  Instead of simply lengthening both the bass and the melody notes together in this way, the interval of lengthening is effectively spread between the hands.

We can hear the combination of dynamic and ‘downbeat area’ accent in Koczalski’s rendition of the third bar of the Prelude, where the effect on the first bass note is rather spontaneous and attention-grabbing, and seems to echo his own advice on the matter, where the (bass) melody ‘gripped by the warmth of the feelings that it expresses, it tries, hesitatingly or energetically to free itself’ from the accompaniment (Koczalski, 56).  Also, consider the fact that asynchrony is used by both Pachmann and Rosenthal (the former using arpeggiation) to emphasise the chord in the middle of bar 8.  This occurs at the crest of a small crescendo and marks the arrival on a strong-to-weak cadence.  Having attempted to copy Pachmann myself, the sensation is that the arpeggiation adds emphasis here.

One of the best examples of the effect of the ‘downbeat area’ accent occurs in bar 13 (Example 1.4).  Here, Pachmann and Rosenthal both bring the low C bass notes in before the right hand, creating a ‘downbeat area’ emphasis on both.  This underlines the peculiar metrical disruption the composer introduces by repeating the ascending arpeggio motif unexpectedly on the third beat of the bar.  Although, if we are to believe Koczalski’s interpretation is likely to have been closest to Chopin own we must be disappointed here because his recording does not reproduce this effect.  (Cortot also does this on the first of these Cs in his 1933 recording)

Ex 1.4 Bars 13-14 of the Prelude No.6 (Metrical Disruption)

with Asynchrony by Rosenthal (red) and Pachmann (black)

In addition, in Pachmann’s version the notes on the weak beats (the tenor Es) are played slightly after the accompanying quavers.  The first of these high Es is also accented (dynamically) quite heavily, but what is the function of the asynchrony on this note?  Does it reinforce the accent, or rather allow the note to stand out in its own space without being masked by the simultaneous striking of the right-hand chord, or both?   The kind of complex interplay of early and late asynchrony in these bars, as played by Pachmann and Rosenthal, returns us to the quote at the head of this article from George Mathias’s preface to Isidor Philipp’s book of daily piano exercises to prepare the student for Chopin, in which he states, ‘you can be early, you can be late, the two hands are not in phase; then you make a compensation that re-establishes the ensemble.’ (Mathias, 5).  Mathias was a student of Chopin.  To explore this notion in more detail we have to refer to a related idea that has all but disappeared from interpretations of 18th-century music on modern piano recordings, ‘Classical rubato.’

Classical Rubato

Chopin inherited a type of rubato that was commonplace in the 18th century but was rapidly displaced by a newer type of continuous tempo fluctuation in the early part of the 19th century (see Eigeldinger, 118-119 and Rosenblum, 34-43, available as a pdf below).  He was known to have modelled his Nocturne style not only on the examples of Field, who invented the genre, but upon the bel Canto of the opera he so regularly attended at the Warsaw National Theatre.  As Eigeldinger notes, it is probable that the older style of rubato described as far back as Tosi’s 1723 treatise was predominant here, where the rubato freedom of the singer was superimposed over a stricter accompaniment.  Such a style is captured in the two quotations earlier cited from Mikuli and Mathias, which describe Chopin’s own playing.  This might be termed ‘Classical’ as opposed to ‘Romantic’ rubato7; in the modern interpretation of rubato the tempo shifts uniformly with both hands adjusting to faster or slower pace so that there is no asynchrony.  In Chopin’s day and in various treatises throughout the 18th century the earlier kind of rubato was described, and of course if one hand maintains time whilst the other deviates, there are going to be many moments when the hands do not quite coincide.

Born in 1835, Saint-Saëns was one of the oldest pianists to record in the piano roll medium.  In his roll recording of the Adagio cantabile from Beethoven’s Sonata Op.31 No.1, there are a number of places where he plays with this older style of rubato, keeping the accompaniment figure going in a strict style while the melody is pulled around.  One of the most striking is analysed in Example 1.5 (bars 91-94).  Here the theme is being presented for the fourth time, albeit now with a more animated accompaniment than before, so perhaps Saint-Saëns was trying to add some variation of his own using the technique.  Lest it be thought that his aged fingers could not get round the notes here his execution of the extremely fleet tuplet figures in bar 95 is delivered with uncompromising speed (and accuracy).  Ex 1.5 shows the extraordinary shifting about of the rhythms in relation to the strict left-hand accompaniment.  Here the effect seems to give the melodic notes their own rhythmic space, varying the way the end of the trill is played in relation to the left-hand and causing the melody to lag behind slightly in the second and fourth bars of the example.  A slowed-down audio track is also given to reveal greater detail (50% speed).

Saint-Saëns: Welte roll (1905) of Beethoven Sonata Op.31 No.1
bb.91-94 (original speed)

Ex. 1.5 Saint-Saëns’s piano roll recording

Saint-Saëns: Welte roll (1905) of Beethoven Sonata Op.31 No.1
bb.91-94 (50% speed)

Other similar touches of asynchrony are introduced in bar 13, 27, 31, 65, 68, 71, 77 and in the coda.  Blink and you’ll miss them (the one in Example 1.5 is the clearest) but they are there.  In many other places he also mostly begins the right-hand trills slightly earlier than the beat as defined by the left hand (again as in Example 1.5).   This serves to make the trill as long as possible, a trick that is encountered a good deal in recordings by 19th-century pianists.

One struggles to find many modern players who attempt to reconstruct this rubato touch in familiar 18th-century repertoire, exceptions being found here and there among those who play period instruments.  Two striking examples are given below (Mozart and Bach) – one by a little-known fortepianist, Ivo Sillamaa, and another by Wolfgang Rübsam performing on the lautenwerk or lute harpsichord.  Eigeldinger quotes Mozart’s letter to his father of 24 October 1777: ‘They are all amazed that I play accurately in time.  They can’t grasp that in tempo rubato in an Adagio the left hand goes on unperturbed; with them the left hand follows suit’.  After suggesting this style for parts of Mozart’s slow movements, Eigeldinger observes ‘how closely Mozart anticipated Chopin!’ (Eigledinger, 119).  This can be felt in a few moments of subtle asynchrony in Sillamaa’s recording below:

Ivo Sillamaa – Mozart Sonata in F K 280 (Adagio, bb.51-54)

Rübsam’s recordings on the lautenwerk are of great interest because he uses copious amounts of asynchrony to distinguish polyphonic voices, to create agogic accents and overall to impart a sensuous poetry to his interpretations of Baroque music perhaps a little at odds with received notions of the style, but suited well to this instrument.  Imagine such playing transposed to the piano8 Here the tempo of the accompaniment is relatively steady, with asynchrony used to set the melody in relief.

Wolfgang Rübsam – Bach, Prelude in E minor BWV 855

Saint-Saëns knew the singer Pauline Viardot who also described the earlier form of rubato (she had known Chopin and played duets alongside him).  He wrote that

Through Mme Viardot…I learned the true secret if tempo rubato…[in which] the accompaniment hold its rhythm undisturbed while the melody wavers capriciously, rushes or lingers, sooner or later to fall back upon its axis. This way of playing is very difficult since it requires complete independence of the two hands; and those lacking this give the illusion of it by playing the melody in time and dislocating the accompaniment so that it falls beside the beat; or else – worst of all – content themselves with simply playing one hand after the other. It would be a hundred times better just to play in time, with both hands together.

(Saint-Saëns, 386-7)

These comments are fascinating to read in the light of Saint-Saëns’s own piano roll recording of Chopin’s Nocturne Op.15, no.2.  To be fair, one is witnessing a recording here that cannot necessarily be taken at face value in the way one might listen to a conventional recording.  A musician will sense an untidiness and hardness in the execution (beyond the asynchrony itself) which is a typical inadequacy of the piano roll technology of the time to capture a full impression of the pianist9  However, even a single listen-through is sufficient to establish that 1) the hands are rarely perfectly aligned, 2) that there is a lot of un-notated arpeggiation in the left hand, and 3) that the tempo shifts markedly at several points, undermining somewhat the idea of an ‘undisturbed’ accompaniment mentioned by Saint-Saëns.  One of the clearest instances in which the hands deviate in the manner of Classical rubato is when the right-hand glissando scale in bar 18 spills over past the third left-hand quaver chord with which it should synchronise according to the score; instead, Saint-Saëns obstinately keeps time with his left hand and effectively allows the right-hand to finish in its own time, slightly later, just as he does with the fast descending scale in the example from Beethoven shown earlier (in both instances the sheer speed of his performance combined with an unrelenting accompaniment seems to necessitate this).  He does this again with the return of a similar figure two bars later (Example 1.6). 

Saint-Saëns: Welte piano roll of Chopin
Nocturne in F# Op.15 No.2 (1905) bb. 17-24

Ex 1.6 Saint-Saëns’s timing in bar 18 of Chopin’s Nocturne Op.15 No.2

Saint-Saëns: Welte piano roll of Chopin
Nocturne in F# Op.15 No.2 (bar 18)

A facet that links Pachmann in particular with the notion of a Classical rubato is that Koczalski and Rosenthal introduce more tempo fluctuation than he does.  For example, Rosenthal slows down markedly at the end of bars 8 and 15 (marked in Ex 1.1).  After a little hurrying in the beginning (end of bar 2 and bar 4) the overall tempo is more stable in Pachmann’s version, and the spacing of the accompaniment chords is quite even, this aspect making his version easier than the others to mimic (having tried to play along with all three versions).

There is a decided lack of much recorded evidence that betrays accompaniments quite so ‘strict’, neat and orderly, as implied by the use of that word – even in Pachmann’s B minor Prelude (and certainly in Saint-Saëns’s Nocturne) there are fluctuations, as noted above.  Nevertheless, there is Classical poise and steadiness to Chopin’s accompaniment figures as played by a number of the earlier-born pianists in some telling examples which I have compiled into a Youtube playlist.

First, three versions of the Fantasie-Impromptu Op.66 show a similarly straightforward approach to the main theme (that is, without any of the pronounced rubato that is sometimes used here) whilst the middle section is only slightly more flexible but marked in the first two cases by many points of asynchrony between the hands, few of which are particularly extreme in time-lag however. The first example is by Marguerite Long (b.1874) who left frustratingly few Chopin recordings. She studied with Marmontel (1816-98) who heard Chopin frequently from 1832 onwards.10 The second example is by Emil von Sauer (b.1862), a student of Nicholas Rubinstein (and later, for two years, of Liszt).  This no-nonsense approach can be heard also in Scharwenka’s (b.1850) piano rolls of the Fantasie-Impromptu and Scherzo No.2. As well as a noted composer, he was co-editor, with Karl Klindworth, of the most widely circulated Chopin edition of the late nineteenth century.

Although there are similarities in all the Fantasie-Impromptu recordings here it is worth stating that asynchrony is far less frequent in Scharwenka’s version; nevertheless in the lyrical theme of the B flat Scherzo it is far more in evidence, both in arpeggiations on the downbeats and hand displacements that give voice to the bass line in a way not dissimilar to Koczalski’s version (though bear in mind the dynamics are coarse in the Scharwenka as this is a piano roll recording). Hearing the detail of the asynchrony is much more challenging in faster-moving music such as this, and it helps to devote particular attention on the bass.

In Arthur Friedheim’s (b. 1859) recording of the ‘funeral march’ from the Second Sonata, which can be likened to Pachmann’s own fairly steady rendition, the right hand is often played slightly after the left hand accompaniment on downbeats in the lyrical middle section (Pachmann’s asynchrony is remarkably restrained by comparison – for him). I personally find Raoul Pugno’s (b.1852) version more touching and imaginative in its subtle but varied asynchrony here in spite of the rather awful quality of the 1903 recording – however, whether this is really classicist might be open to debate as Pugno adds a deliberate and gradual accelerando towards the end of the middle section (Pugno was another Mathias student).

Koczalski (Mikuli’s student), although not averse to tempo modification, tends towards a rather punctual performance style, especially for example in waltz-type accompaniments, without the sort of inflections of pulse and rhythmic placement in which Cortot and Friedman regularly indulge (the beginning of Cortot’s G minor Ballade is a case in point, or any of the waltzes recorded by Friedman – compare these with Koczalski in the same passages). I have included on the playlist a further example of Koczalski’s playing in the form of the Nocturne in G minor, Op.37 No.1 where I feel his classicism comes to the fore, but which also shows a high level of asynchrony, the right-hand often lagging behind the left though not in a predictable fashion (in my opinion).

Carl Friedberg (b.1872), a student of Clara Schumann, also plays with a more reserved sense of tempo modification (as might be expected given Clara Schumann’s reputation as something of a stickler): in this playlist you can compare his recording of the Nocturne in A flat to that of the Liszt student Frederic Lamond (b.1868), which is about an extreme a comparison as could be made, Lamond lurching around all over the place in addition to some quite marked asynchrony between the hands. 

In spite of this classicism (and indeed perhaps because of it) touches of asynchrony are thus still to be found ‘little and often’ in almost all the aforementioned examples.  It is worth remarking that a rediscovered piano roll recording of the Liszt Sonata by Arthur Friedheim (b.1859), arguably one of Liszt’s closest students, has revealed the likely importance of asynchrony – both arpeggiation and the asynchrony between hands – in Liszt’s own style (Carter and Adler, 24-28) and this is also known from several other recordings of Liszt’s music by his students (Friedheim’s was Liszt’s favourite interpretation of the piece).  Of course, this says nothing of Chopin’s style, yet Liszt was known to be a very good mimic of that, and it is likely to have some elements in common.  The point is that ‘Friedheim adopts a more ‘classical’ approach than other recorded Liszt pupils although it is in certain respects more ‘romantic’ than that of modern-day pianists.’ (Carter and Adler, 57).  

This seems to point to the possibility that we cannot simply conflate both extreme rubato (in the usual sense) and asynchrony with late-Romantic practices that had little to do with any of the styles prevalent in Chopin’s day, which some might be tempted to do in order to dismiss it 11.  What it rather suggests is that asynchrony had always been there as a consequence of 18th-century ‘Classical’ rubato practices that predate even Chopin.  Indeed, Chopin himself advised using exactly this kind of rubato in Weber’s music (for example, according to George Mathias, in Weber’s Sonata in A flat Op.39 – see Eigeldinger, 50)! Saint-Saëns’s piano roll of the Beethoven Adagio grazioso embodies significant traces of this type of execution as we have seen. 

To return to the main case study in hand, these recordings of the B minor Prelude could provide strong evidence of Chopin’s general manner of touch at the keyboard with regards to the application of asynchrony.  Moreover, early piano recordings in general could provide evidence of a sort of combination of the two kinds of rubato, with the asynchrony that results from a split between melody and accompaniment being traceable further back in time to a Classical type of rubato, now lost except for a few glimpses like the Saint-Saëns above.  We can discover accounts of Chopin’s playing that state quite clearly that he used both kinds of rubato, so to an extent we might expect a muddying of the water, a mixing of the two, making it harder to unravel the distinctions:

Chopin was far from being a partisan to metric rigour and frequently used rubato in his playing, accelerating or slowing down this or that theme. But Chopin’s rubato possessed an unshakeable emotional logic.  It always justified itself by a strengthening or weakening of the melodic line, by harmonic details, by the figurative structure. It was fluid, natural; it never degenerated into exaggeration or affectation.

(Mikuli, quoted by Michalowski, translated in Eigeldinger, 50).

No wonder that attempts to find evidence of a rubato that uses a strict accompaniment have proved elusive, for by the time the first recordings were made nineteenth-century pianism had become a melting pot of so many influences from the history of the keyboard, arguably reaching right back into the techniques used on the early fortepiano and even the harpsichord (such as arpeggiation). 

Against Asynchrony

When we consider that norms were defined through the hands of amateur musicians much more than today (who of course did not have recordings or, necessarily, definitively good teachers, on which to model their playing) the backlash against the ‘limping’ style of asynchrony that can be found in the advice of pianists – Raoul Pugno, Mark Hambourg, Josef Hofmann and Busoni, can be considered in its due context. Some of the contradictions in these author’s comments against asynchrony are discussed in Hamilton’s chapter on ‘A Singing Tone’ (pp.139-178) – a must-read for anyone interested in the phenomenon.  He quotes famous advice given by Hofmann in his manual for pianists (‘this “limping” as it is called, is the worst habit you can have in piano playing’), yet points out Hofmann’s admiration for Rosenthal’s playing (Hamilton, 149).  Peres Da Costa’s analysis of Raoul Pugno’s recording of Chopin’s Nocturne Op.15 No.2 also shows that Pugno’s own advice to keep the hands absolutely together is not born out by his recording, showing the relative nature of such advice (Peres Da Costa, 80).

There is ample evidence that, at least in certain works, these players employed asynchrony when the mood took them, and especially so in Chopin’s music.  Here is a striking example from Chopin’s Nocturne Op.9 No.2, where Hambourg loosens his right hand from the accompaniment in bars 13-14, effectively introducing a little variation to this phrase on its second occurrence (the first time being more synchronised) – see Example 1.7.  The asynchrony is more easily heard at a slightly slower speed – and in this case to maintain clarity at a resulting lower pitch:  

Mark Hambourg: 1927 HMV (C 1416) 78 rpm
Chopin’s Nocturne in E flat, Op.9 No.2 (bb.12-15)

Ex 1.7 Asynchrony in Mark Hambourg’s Recording of Chopin’s Nocturne Op.9 No.2, bars 12-15

Hambourg: 1927 HMV (C 1416) bb.12-15
showing asynchrony and added arpeggiation including some additional notes (in brackets)

Note that, in addition to the asynchrony between melody and accompaniment there are left-hand arpeggiations of chords throughout (indicated below the score on the example), a practice that was very common in early recordings of this work, not a single one being indicated by the composer, as well as playing quite a few of the low bass notes, especially on the main downbeats, slightly ahead of the right-hand (also de rigueur).

Returning to the notion of Classical rubato it is revealing to compare the above snippet with the same passage played in Sarasate’s arrangement for violin and piano, recorded by Mischa Ellman (b.1891) in 1915 (Victor 74052). No accompanist could be expected to follow such free soloistic playing perfectly, thus the two diverge considerably, and it is impressive that they arrive together unanimously for the cadence in the middle of bar 14 at all:

Mischa Elman: 1915 Victor (74052)
Chopin arr. Sarasate Nocturne in E flat Op.9 No.2

Hambourg’s interpretation of the same passage, whilst not as extreme, certainly intimates the effect of a soloist loosening themselves slightly from the accompaniment, though in perhaps more tasteful proportions.  In fact, listening to the two excerpts side by side there are some strong similarities. But if Ellman’s take on the piece is closer (for some listeners) to a caricature of late-Romantic excess, between the two examples perhaps lies some vestige of a style that Chopin himself may have recognised.  Robert Philip has also drawn attention to striking examples of the earlier form of rubato in early recordings of instrumentalists and singers, including violinist Eugène Ysaÿe and singer Adelina Patti – Ellman’s is not an isolated instance (Philip, 110-112).

Robert Philips’s books on early recordings give a good overview of how Romantic and modern practices diverge. He observes that, in allowing the hands to drift apart a little at times, pianists were mirroring the approach of other ensembles, where simultaneity was not necessarily as highly prized then as it later became during the age of recording (Philip, 129).12  After reviewing recorded evidence from a number of string quartet examples, he writes, ‘It is inconceivable that modern ideas of ensemble, which have only recently developed, existed sometime in the nineteenth-century or earlier and were lost by the time of early recordings.’ (Philip, 123).  I would argue that this conclusion is apposite too for the notion of asynchrony on the piano.  Considering a wealth of recorded examples and understanding how the pianists are situated along the timeline, I would surmise that in Chopin’s day the likelihood of simultaneity in piano playing being some kind of norm is extremely remote.  I must also stress as Philip does, that it is also likely that there were a greater range of interpretations and approaches than exist today, so that among performers, some were surely a lot looser than others (this trend towards homogeneity has to do with steadily increasing global communications and travel, not to mention the increasing availability of recordings and, if you will, the professionalization of music-making). 

What should pianists do?

In one of his books on Chopin, Koczalski himself alludes to the effect of asynchrony between the hands and his description gives voice to the expressive purpose behind it in a statement that echoes the quotation included earlier from Mikuli, his teacher. He tells us that when the right hand plays ornamental passages, or,

‘gripped by the warmth of the feelings that it expresses, it tries, hesitatingly or energetically to free itself from the left hand, only then can there be a difference in the attack of both hands.’

(Koczalski, 1936, 56)

This typically Romantic statement perhaps leaves us perhaps none the wiser as pianists and teachers attempting to recreate meaningfully Chopin’s music.  Without guidance from the score on the matter, how often and how much should we free the hands from each other?  In a slow cantabile piece like this Prelude, should we indulge frequently as Pachmann and Rosenthal do, or be more moderate, as Koczalski does, or play it safe and not use asynchrony at all? 13 Some might prefer the latter – after all, ensemble playing no longer tends to exude the looseness it once did, so why should pianists get caught up with a deliberately blurred sense of ensemble?  Of course, synchronised playing is still largely the default these days, but are we missing out on something – perhaps the very soul of Chopin’s rubato? 

Before concluding we should ask whether there exist versions of this work by pianists who were born in the nineteenth century who did not introduce asynchrony.  Mark Hambourg (b.1879), who studied with both Anton Rubinstein and Leschetizky, recorded a version which exhibits far less asynchrony, but it is still present in three clear instances (annotated in blue on Ex 1.8).  Note that two occurrences coincide with Pachmann’s (in black). There is also a remarkable alteration of rhythm in the left hand with four of the quaver F sharps in the left hand of bars 15, 16, 19 and 20 played a semiquaver later than written.  This is certainly a form of displacement as it means these notes no longer coincide with the right-hand accompaniment.  Perhaps Hambourg wanted to give these F sharps clarity without emphasising them dynamically (as there are at weaker points in the metre) but he achieves the effecct within a framework of more or less exact rhythm rather than a staggered form of asynchrony between the hands. 

Ex 1.8 Asynchrony in Hambourg’s Recording of Chopin’s Prelude No.6

Chopin: Prelude in B minor Op.28 No.6
showing asynchrony applied by Mark Hambourg (blue)
compared Pachmann (black)

Arthur Rubinstein does not use any asynchrony, which is to be expected.  It is unfortunate that a number of other key figures from this period did not leave recordings of the work, such as Hofmann, Rachmaninov or Moiseiwitsch.  Isidor Philipp’s student Guiomar Novaes did record the piece but does not make use of asynchrony, in spite of her teacher’s connection to Chopin’s student, George Mathias.  Yet in recordings of other pieces she makes some subtle use of both arpeggiation and desynchronization of the hands.  In a similar way, in other Chopin pieces Hambourg, Koczalski and Cortot use asynchrony more extensively than they do in the B minor Prelude (we have already heard Hambourg doing so in Op.9 No.2).

Cortot’s 1933 recording of the work has only three clear instances of asynchrony between the hands, yet all in places where we have seen already it used – on the first beats of bars 5, 11 and 13 where the low bass notes initiate the main arpeggio motif, serving as a ‘downbeat area’ accent.

Among Cortot’s teachers was (if only briefly) Emil Descombes, who studied with Chopin, whilst another of his teachers, Louis Diémer (b.1843), left an early disc recording of Chopin’s Nocturne Op.27 No.2 that is quite overwhelming in its use of asynchrony (see Links below).  If Cortot is quite sparing here, in another Prelude from Op.28, the E minor, he uses considerably more asynchrony, delaying the right hand melody in relation to the repeated chordal accompaniment on many occasions, as well as adding quite a lot of additional arpeggiation to the left-hand chords. He seems to have imparted some of the flavour of his asynchrony in this work to his student Samson Francois judging also by his recording the piece.

We are left with a puzzle that can only be expected when a style of playing is not fully embodied in written scores and can only be passed down from teacher to pupil – and then only to pupils that either have the capacity to employ the technique and to whom it also appeals in a given piece.  How can modern pianists be true to a style when it appears and disappears before our very eyes (one should say ‘ears’), as one artist uses asynchrony in a piece whilst another seems less interested in emphasising that feature, and another omits it altogether?  Indeed this same problem was reported in ever-varied renditions that Chopin himself gave to demonstrate his music to students, and of course in the problems editors face with assimilating the many variants that exist in different manuscripts of some of his pieces.


The fact remains, however, that in the B minor Prelude piece at least five artists born in the 19th century, and all with connections to students or contemporaries of Chopin (Liszt, Mikuli, Descombes and Leschetizky) are all using asynchrony several times during a short piece and in various interrelated ways, some significantly, others less so, but it is always there in some measure.  Considering this from a more subtle historical angle, we might note some of the conclusions drawn by Mark Arnest in his fascinating paper on asynchrony, whose research appears to have encompassed a vast array of 19th-century pianists (he uses ‘noncoordination’ instead of asynchrony):

The general trend was towards homogenization. No pianists born in the 1870s used noncoordination as much as Pachmann or Paderewski had before them; but at the same time, few are as restrained as some of the pianists born in the 1860s, such as Friedheim or da Motta. Pianists born in the 1880s are more alike than pianists born in the 1870s, and, with notable exceptions such as Friedman and Koczalski, less romantic in approach. Some pianists born in the 1890s abandon noncoordination altogether.

(Arnest, 9)

This overview would predict that in the B minor Prelude, Pachmann (1848) and Rosenthal (1862) would give us the most amount of asynchrony (true), that Hambourg (1877) and Cortot (1879) would be less inclined to break the hands in this way (also true), and to introduce less or no arpeggiation (true in this particular case but not always so with Cortot), and that Koczalski would be the odd one out (true, because he uses more asynchrony than Cortot and Hambourg, in spite of a slightly later birthdate (1888) and arguably perhaps due to his association with Mikuli from whom he picked up on Chopin’s asynchrony).

Arnest also suggests that the 1850s and 60s could have been the period of greatest ‘excess’ with regard to asynchrony and rubato fuelled also he suggests by the influence of Liszt. This possibility is intriguing, because in Liszt’s pupil, Rosenthal, we have heard perhaps the most intense combination of both of these traits (rubato and asynchrony) in the B minor Prelude, remembering that Rosenthal grew up in the 1860s and received instruction from Mikuli in 1872-74 and Liszt from 1876-78 (and again from 1884-86).  However, Liszt’s students represent a contradictory picture overall with regard to Chopin – the subject for a future project!

Asynchrony is an issue that ought to get pianists listening, thinking, talking (and playing) in response to some very interesting evidence, but at the same time the evidence is capricious, will-o-the-wisp-like (to borrow a favourite nineteenth-century image).  It is the pianist’s equivalent of a magician’s sleight-of-hand, the slightest amount being much easier to spot on a modern recording than on an early twentieth-century one.  It is easy to miss, hard to quantify, elusive, requiring many repeated listenings to pin down in its details.  The reward is to begin to understand a much-misunderstood (and maligned) phenomenon, whose artistic merit has not helped by the distortion of dynamics and touch in piano rolls that makes the effect seem blunt and ugly in that medium; yet this is counterbalanced by some exquisite examples in the disc medium.  Pianists today certainly use it, mostly in a very subtle and occasional way, but early recordings provide a window on an era in which its use was cultivated far more.  As so much in music comes down to taste it is easy to see how the technique has become lost along the way as fashions inevitably change.  This needs to be seen in context: ‘expression’ itself has changed, rubato is less volatile, and professional pianism in general exhibits much more control in the age of recording than the freedom it once had (again, see Robert Philip’s book).  

The elevation of composers above performers and consequent attempts to sanction only the letter of the score can also make it very tempting to disregard the kind of Chopin-playing that was heard in the most distant past rather than that we which we now infer from the dots on the page or our immediate teachers.   But the Chopin-playing of Koczalski, Pachmann, Paderewski, Pugno, Friedheim, Horszowski, Michalowski, Bauer, Friedberg, Cortot, Long, Rosenthal, Friedman, Rachmaninoff – all of whom incorporated asynchrony – is also a collective phenomenon, a group of pianists all speaking related languages. Their playing belonged to an age where music was heightened emotion and every nuance in the pianist’s toolkit was turned in that direction.  Now their recordings can be accessed at the touch of a button.  As a consequence today’s more attentive and adventurous pianists are indeed listening and responding, in spite of what I perceive as resistance from a conservative establishment14. I have included a number of examples in the Links section below, including a few of my own recordings.

The idea of recreating some sort of Chopin school would perhaps have appeared nonsense to those who heard the composer play.  His touch, as Mathias concedes, was intrinsically inimitable.  But what are we left with? The stark score on the one hand with its precision notation, and the recorded legacy of nineteenth-century pianists on the other, offering us Chopin ‘as heard’ in a myriad of reflections, none of which quite measure up to the exactness of the written score which we have perhaps made the mistake of taking too literally.  Although in many pieces asynchrony is only a mere whisper here and there, it is a lost accent (quite literally) in modern piano playing, and one that deserves to be more thoroughly accounted for. In the B minor Prelude we see at least some evidence of its expressive role. A performance without it, whilst surely valid, is going to sound and feel rather different.

The asynchrony in these recordings of the B minor Prelude represents the tip of a very large iceberg, so the picture needs to be deepened.  Arnest’s and Peres Da Costa’s work certainly contain great insight, especially the latter’s in-depth consideration of two Chopin Nocturnes.  However, no author focusses exclusively on asynchrony in Chopin, and little wonder, for the plethora of different interpretations really makes it hard to martial evidence and draw conclusions.  Yet my own research suggests there are many similarities between interpretations of some of the most frequently recorded pieces of Chopin, and in more aspects than just asynchrony.  Other un-notated tendencies to do with other aspects of rubato and tempo modification reveal themselves too.   I have created below some playlists for these works (see Links, below), including my own recordings of certain familiar pieces that are based on observations from my ongoing research into early recordings.  I can only hope that at some point there will be chance to present in detail my findings for each piece.


Since writing this I have been made aware (by a twitter post from Mark Ainley) of a wonderful recorded set of Chopin’s Preludes from 1928 by Robert Lortat (b.1885), who attended Louis Diémer’s class at the Paris Conservatoire along with Cortot. His recording of the B minor Prelude adds further weight to the evidence surrounding a tradition of asynchrony in this piece. In fact, unlike Cortot, Lortat’s recording is unbridled in its asynchrony, situating it firmly alongside Rosenthal and Pachmann’s recordings. Without further analysis I leave listeners to appreciate this recording, for educational and research purposes:

Robert Lortat: 1928
Chopin, Prelude in B minor Op.28


  • Arnest, Mark: Why Couldn’t They Play With Their Hands Together? available as online pdf only: https://www.lib.umd.edu/binaries/content/assets/public/ipam/resources-reviews-and-links/arnest-hands-together-article-pdf-5-15-12.pdf
  • Blickstein, Edward. and Benko, Gregor: Chopin’s Prophet: The Life of Pianist Vladimir de Pachmann, Rowman and Littlefield, Maryland, 2013
  • Carter, Gerald and Adler, Martin: Liszt Piano Sonata Monographs: Arthur Friedheim’s Recently Discovered Roll Recording, Wensleydale Press, Sydney, 2011
  • Eigeldinger, Jean-Jacques: Chopin: pianist and teacher – as seen by his pupils, Cambridge University Press, 1986
  • Goebl, Werner: The Role of Timing and Intensity in the Production and Perception of Melody in Expressive Piano Performance, (Doctoral thesis). Karl Franzens-Universität Graz, 2003  http://www.ofai.at/cgi-bin/get-tr?download=1&paper=oefai-tr-2003-28.pdf
  • Goebl, Werner and Parncutt, Richard: ‘Asynchrony versus Intensity as Cues for Melody Perception In Chords and Real Music’, Conference Proceedings, 2003 https://pdfs.semanticscholar.org/92ae/6d365e11a2bd81b3daf534c4326adcd87edc.pdf?_ga=2.82099039.83971444.1559380391-257773849.1559380391
  • Hamilton, Kenneth: After the Golden Age, Romantic Pianism and Modern Performance, Oxford University Press, 2008
  • Koczalski, Raoul: Frédéric Chopin. Betrachtungen, Skizzen, Analysen, Tischer & Jagenberg, 1936
  • Mathias, George: Preface to Isidore Philipp, Exercises quotidiens tires des oeuvres de Chopin, Hamelle, Paris, 1897
  • Mikuli, Carl: Vorwort to Chopin’s Pianoforte-Werke, edited by Mikuli, Kistner, Leipzig, 1880
  • Mitchell, Mitchell and Evans, Allan: Moriz Rosenthal in Words and Music, Indiana University Press, 2006
  • Peres Da Costa, Neal: Off The Record: Performing Practices in Romantic Piano Playing, Oxford University Press, 2012
  • Philip, Robert: Performing Music in the Age of Recording, Yale University Press, New Haven, 2004
  • Rosenblum, Sandra: ‘The Uses of Rubato in Music, Eighteenth to Twentieth Centuries’, Performance Practice Review Vol. 7 No.1, 1994 https://core.ac.uk/download/pdf/70972184.pdf
  • Saint-Saëns, Camille: ‘Quelques mots sur l’exécution des oeuvres de Chopin’, Le Courrier musical, XIII/10 (1910), pp. 386-7

Links: Modern Recordings with Nineteenth-century levels of Asynchrony

Related Links

  1. It is also referred to by the terms ‘desynchronization’, ‘noncoordination’, ‘hand displacement’ or ‘dislocation’ (the latter two being distinct from arpeggiation)
  2. Rosenthal was aware of Pachmann’s renown as a Mazurka player but had never heard him play one, so visited Pachmann’s Berlin house unannounced early in 1902. As reported in Huneker’s book, Variations, Pachmann played a trick on him, refusing to give up his secrets.  He played through the entire collection but Rosenthal noted to Huneker that ‘not an accent was correct, the phrasing was falsified, though the precise notation was adhered to…The joke was later when Rosenthal teased Pachmann about his trickery and the Chopinzee absolutely grinned with joy.’ (Blickstein and Benko, 197).
  3. Pachmann’s association with Liszt is the subject of a short chapter in Benko and Blickstein’s detailed account of his colourful life, but they admit that obtaining a conclusive picture is problematic due to some of the exaggerated stories Pachmann himself told.  What is sure is that Pachmann heard Liszt on several occasions, witnessed his masterclasses in Weimar and seemed to receive high praise from him regarding his own Chopin interpretations (Blickstein and Benko, 37-43).  There is even the suggestion that some of his bizarre stage antics were inspired by Liszt.
  4. The story seems not to have come from Pachmann himself but through Mantia’s mother, Ida Bosisio, herself a student of a Liszt’s pupil, Sgambati.  Evans nevertheless claims to have found the story also corroborated by a Florentine colleague of Mantia’s mother, so it perhaps it remains a tantalising possibility. The pair of women apparently visited Rubio and heard the story directly from her, although she herself never mentioned it to anyone.
  5. Horszowski’s mother also studied with Mikuli and taught her son piano to begin with – he can be heard in concert using typical Mikuli-school asynchrony in Chopin Nocturnes as late as the 1980s – see Links
  6. The contradiction between verbal advice and actual practices (in many cases the practitioners are the authors themselves) comprises an overarching theme in Peres Da Costa’s book.
  7. Rosenblum uses the term ‘contrametric rubato’ whilst Peres Da Costa uses ‘metric rubato’
  8. In fact it is of course perfectly possible to play the piano in a way saturated in asynchrony and the B minor Prelude as played here is but one example.  One of the most remarkable examples I have come across is Mompou’s Variations on a Theme of Chopin played by the composer himself, saturated throughout with tiny asynchronies, only a few of which are actually necessitated by the stretch of a ninth or more. This is available on YouTube (see Links below).  Can it be a coincidence that Mompou was a student of Isidor Philipp, student of Chopin’s own student Mathias?  Indeed Philipp encouraged him greatly in his efforts as a composer.  It remains very frustrating that Philipp himself left no recordings of Chopin’s music.
  9. This is especially down to the dynamics, which could at this stage only be very generally recorded or were added afterwards. With many of the accompaniment notes played unduly loudly any hint of sensuousness is lost, not to mention that there is also a lack of dynamic variation in the melody; the playback speed of the roll might also be questioned although Saint-Saëns was a notoriously fast player.  However, it is maintained by several experts in this field that piano rolls are fairly accurate when it comes to timing, and therefore a reliable guide to asynchrony
  10. According to Eigeldinger he lived very near Chopin’s home at the Square d’Orleans (see Eigeldinger, 131)
  11. Perhaps the renown of Paderewski’s playing has not helped here: although there are some great examples, he certainly tends to mix quite volatile Romantic rubato with pronounced asynchrony of the hands – a combination that proved too heady for many and became identified with a sort of false ad libitum approach to Chopin
  12. One of his major insights hinges on the idea that the process of recording actually lead to greater precision of ensemble in a sort of feedback loop in which players and ensembles gradually became aware of a need to tidy up looseness in the ensemble as a result of repeatedly recording and listening back to themselves, not to mention listening to the recordings of others.
  13. In fact, Koczalski’s behaviour in relation to this phenomenon is not as restrained generally as this particular example or his statement above suggest. He applies asynchrony to bass notes frequently in many slow and lyrical passages throughout his large recorded output of Chopin’s music, and at times the hands drift apart even in faster music.
  14. Pianist Dimitry Ablogin achieved only an honourable mention as finalist in the new Chopin Pleyel competition in spite of his entirely appropriate use of asynchrony and his equally authentic preludising between pieces, neither of which was practised by other contestants. Elsewhere flawless and highly expressive playing made one wonder to what extent the unusualness of his use of asynchrony made judges discriminate against him