dismiss icon

Win a FREE year of tonebase!

Enter our raffle for a chance to win free access to 500+ courses, weekly LIVE events, a vibrant community, and more!

Enter To Win A Free Year Of tonebase Piano
Discover The Genius of Bach's Keyboard Works

Discover The Genius of Bach's Keyboard Works

A thorough introduction to Bach's keyboard music, available now for free.

Watch Now →

If you’re an early to intermediate-level piano student, you may be encountering the word “rubato” for the first time. 

Or, if you’re a longtime pianist, you may have heard this word tossed around but without much instruction on what it means for you as a performer. 

Well, that’s what I’m here to discuss today. 

What is rubato? How do we play it on the piano, and how do expectations for rubato usage differ by style period? 

Firstly, “rubato” literally means “stolen” in Italian (from rubare), and “tempo rubato” means “stolen time”—that you may take time away from another moment in the piece by lingering or you may push forward into a moment a little earlier than the metronome may suggest. 

The idea is that, as in satisfying endings to TV crime dramas, what’s stolen will eventually be returned: the fluctuations of tempo should be compensated for by the end of the piece. However, there are differences in how the rubato idea is interpreted in the various style periods that most pianists encounter. 


A little background

In the preface to his excellent book Stolen Time: The History of Tempo Rubato, Richard Hudson cites Pier Francesco Tosi’s discussion of “musical theft” in 1723 as the first mention of rubato usage (throughout this article, unless otherwise noted, I will refer to page numbers in Hudson’s book even when I am quoting others he has also quoted, just to make these easier for you to look up should you desire to). 

1723 is around the late Baroque and the very beginnings of the Galant style—think Domenico Scarlatti and the very young Haydn, if you’re a pianist. This means people were thinking about rubato even at that time. 

Hudson defines two types of rubato. The first is an earlier rubato developed in vocal and violin music in the early 18th century in which “…some note values within a melody are altered for expressive purposes while the accompaniment maintains strict rhythm. 

This type of rubato continues in vocal and violin music well into the 19th century. Keyboard music incorporates this earlier type of rubato during the second half of the 18th century” (p. 1). 

The second, which he calls later rubato, is defined as follows: “…tempo rubato begins to refer to rhythmic alterations not only in the melody, but in the tempo of the entire musical substance. 

For at least the first half of the 19th century both types of rubato existed concurrently, but later in the century the earlier type disappears” (p. 1). The second sentence in this quote is of particular interest to pianists: Chopin, of course, lived almost exactly in the first half of the 19th century (1810-1849) and that’s the period in which both rubato types circulated. No wonder there are so many questions about how to execute rubato in Chopin’s music!

The Earlier Rubato 

(melodic fluctuation in note values over a steady accompaniment) – relevant to the Baroque, Classical, and Early Romantic Eras

This type of rubato traces back to techniques of melodic variation seen as early as 1320 (p. 13) that could include varying the rhythm of the written melody. 

One type of rhythmic manipulation that fits the definition of earlier rubato, particularly prevalent in the French Baroque, is notes inégales—something like today’s swung 8th notes in jazz and blues. 

Just as in swing, a string of even 8ths or 16ths might be rendered “long-short long-short” almost like the first and third 8th notes in a set of 3 triplet 8ths. Wikipedia gives a nice illustration in notation here.

As notated:

As sounded if played as notes inégales:

Hudson calls this “inequality” to translate the term into English (p. 26). The Italians Caccini and Frescobaldi mention inequality as well (p. 26), but it’s probably best known in the French repertoire. Enjoy this Gavotte by François Couperin as an example:

Click here for the score if you’re curious.

In this earlier rubato, the underlying pulse (what you would tap your foot to) is not affected. It’s the stuff on top that gets manipulated. 

The Later Rubato 

(tempo fluctuation in the whole texture)

Although Hudson calls the whole-texture rubato “later rubato,” it has existed in various forms for much longer than the name suggests. 

Performance directions to indicate tempo changes have existed since Haydn and Mozart, but tempo modifications themselves have existed since the days of Gregorian chant (p. 4) for expressive reasons, like emphasizing specific points in sung text. 

This can be heard most often today in Baroque and Classical opera recitative, which is so free in its sung rhythms that the collaborative pianist or harpsichordist needs to follow the text to know where the chords will go. If you’d like to deep-dive into what a recitative is, here’s a nice overview on YouTube (timestamped, not mine): 

The cadenza (virtuosic improvisational passage for a soloist) or eingang (lead-in) in the Classic-era concerto is another great example in which time is practically suspended, in contrast to the usually very regular meter and rhythms in the rest of the piece. 

Christian Kalkbrenner mentions tempo flexibility as one type of rubato in 1789, but he cautions that it should be limited to a measure long, at most, before returning to the typical tempo (quoted on p. 140). 

Daniel Gottlob Türk (a name you may know from some of the student-level pieces he wrote and we still play frequently) also refers to “intentional hastening and hesitating” while clarifying that the term “tempo rubato” usually only refers to the hesitation (quoted on p. 142). He continues by calling out virtuosi and solo singers who tended to hurry or hesitate too often and at inappropriate places (quoted on p. 143).

How the Earlier Rubato Applies to Keyboard Music and How to Play It

Wolfgang Amadeus Mozart points out how people didn’t tend to grasp the idea of the earlier rubato style (as mentioned earlier in his father’s treatise) on the solo keyboard, saying “people cannot grasp…that in tempo rubato in an Adagio, the left hand should go on playing in strict time” (quoted on p. 113). 

That’s because this earlier rubato is very difficult to play on the same instrument by the same person, a bit like patting one’s head while rubbing one’s belly—two very different things at once. 

As Hudson mentions later on the same page, this may be one reason that “rubato” came to mean a general slowing down or speeding up, because a keyboardist’s attempt to play a rhythmically free right hand alongside a strict left hand frequently resulted in both hands drifting in tempo together instead.

Luckily for those of us who weren’t alive in the 18th century, and probably because of the aforementioned difficulty, the earlier style of rubato was often written into the score, or at least approximated, using articulation markings or changes in note values. 

18th-century theorist Friedrich Wilhelm Marpurg gives some helpful examples in three different publications, which are reproduced and compared by Hudson (p. 116) here:

piano rubato examples by Friedrich Wilhelm Marpurg

Ernst Wilhelm Wolf, writing in 1785, also points out that this type of rubato is often written out (quoted on pp. 116-117). 

We have a record of Sigismond Thalberg treating rubato in this way during his lifetime, which spanned most of the Romantic era (1812-1871), and Adolph Friedrich Christiani mentions that when he “…embellished the melody, the accompaniment proceeded with steady, unwavering precision, unaffected by the emotion displayed in the solo parts” (quoted on p. 128).

Therefore, if you encounter a piece with rhythmic figures like in Marpurg’s examples, you can reasonably assume they are outlining a degree of written-out rubato, and as long as you can play those rhythms clearly, you will be executing something fairly close to the earlier-style rubato. 

A clear and accessible example of written-out earlier rubato can be found in the slow movement of Haydn’s early Keyboard Sonata in Bb (referred to on IMSLP as a “Partita”), which starts on page 5 of this PDF.

And here’s tonebase artist Marc-André Hamelin performing the sonata:

How the Later Rubato Applies in Keyboard Music

The situation is much less clear regarding the later-style rubato, where the whole texture slightly slows or accelerates, in keyboard music during the early 19th century. 

Hudson addresses this confusion on p. 150: “...many of the types of rubato could occur either as a compositional device indicated in a score by a composer, or as an expressive nuance added [by] a performer.” 

Carl Czerny, another name that’s likely familiar to modern piano students, outlines some structural situations within a piece where some time-taking is appropriate, like the return of an important theme (like the opening theme or a transition to a different rhythmic movement) (quoted on p. 148). 

Overall, at this point, later-style rubato was considered appropriate only briefly and for carefully selected places in the music (p. 151). It was meant to be treated as an expressive device, or more of a spice than a foundational ingredient, but musicians and critics even back then were complaining about performers misusing it. 

These ideas about rubato more generally began to evolve further as the 19th century wore on, so it’s probably more useful to focus on a few specific composers and their usage of both rubato types, and that’s where I’ll turn now.

And That Guy You Probably Came Here For: Chopin

And what about the most “rubato-ish” composer in most pianists’ minds, Frédéric Chopin?

Well, as Hudson points out, accounts of Chopin’s playing seem to conflict with each other. 

He spends a whole chapter on Chopin’s rubato if you are interested in a deep dive. Several contemporaries of his note his rhythmic flexibility in his own playing, including the (in my opinion underappreciated) composer Ignaz Moscheles: 

“the ad libitum playing, which in the hands of other interpreters of his music degeneratese into a constant uncertainty of rhythm, is with him an element of exquisite originality” (quoted on p. 177). 

He also compares Chopin’s playing to “some singer who troubles himself very little about the accompaniment, and follows his own impulses” (quoted on p. 192). 

This agrees with my own experience as a vocal repertoire-focused collaborative pianist, in which I’ve often had to reinforce the meter of the accompaniment when the singer needs to do different things like stretch a syllable because of a particular consonant or expressive point in the text. 

Chopin loved opera and the accounts of the vocal repertoire’s influence on his writing are widespread.

In addition to Moscheles, Carl Mikuli notes that “in keeping tempo Chopin was inflexible, and it will surprise many to learn that the metronome never left his piano” (quoted on p. 178). 

Chopin’s student Wilhelm von Lenz remembered Chopin saying “The left hand is the conductor, it must not waver, or lose ground; do with the right hand what you will and can” (quoted on p. 191). This sounds like a description of the earlier rubato, with a steady accompaniment.

But Mikuli also notes that “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” (p. 178). 

This sounds more accurate as a description of later rubato, judiciously applied. This makes sense given that, as Hudson clarifies, “Chopin lived at a time when classic restraint still existed side by side with the newly emerging freedoms of romanticism” (p. 179).

So what does that mean for us, since we are probably wondering about how to apply the later rubato in Chopin? 

Well, Chopin is generally very clear in using markings like ritenuto for relatively sudden hesitations, rallentando for more gradual unwinding after the peak of a piece, or stretto for tightening/compacting (among many other examples). 

As long as you’re using an edition that hasn’t undergone heavy-handed editing—most competitions recommend Jan Ekier’s Polish National Edition—carefully observing Chopin’s expressive markings will be helpful in figuring out where the later rubato should apply. 

However, the majority of the quotes above (and many more in Hudson’s chapter that I am leaving out for space reasons) remark on Chopin’s special skills in the earlier-style rubato in which the left hand is kept rock-steady. 

Chopin illustrates this, too, in his infamous fioritura—those odd-numbered tuplets (I’ve seen 11, 17, and 19-tuplets!) that tend to terrify pianists new to his music. 

Those tuplets themselves create a sense of departing from the pulse because they don’t fit in our typical meters. They’re meant to sound “off” compared to our steady 4/4 or 3/4 time signatures, so we don’t need to do much to them, nor do we need to worry about them being timed down to the microsecond. They should just sound like improvised ornaments.

Therefore, I would suggest dedicating more energy to aiming for the earlier rubato: keeping the accompanimental textures steadier in tempo except at the designated points marked by Chopin’s performance directions while letting the melodic passages wander—sometimes behind the beat if you want to communicate a feeling of lingering and sometimes pushing ahead if you want to communicate more urgency. 

Keep in mind that this isn’t always going to be in the right hand. His Etude op. 25 no. 7 puts that more flexible melody in the left hand, so in that piece, the right hand needs to assume the role of the conductor! I would also say this about sections of the Polonaise in C minor, op. 40 no. 2.

Late Romantic Rubato

In the later Romantic period, mostly the second half of the 19th century, Hudson notes that “the word rubato now referred generally to the later type” and that “the various methods of applying it were transmitted by Liszt and his numerous pupils” (p. 300). 

But as these ideas spread, they also evolved into different ideas about what rubato meant, making the whole business of figuring out rubato’s meaning even more challenging for us today. 

One English-language musical dictionary by John Stowell Adams says the following in 1851: 

“Tempo rubato…implies that the time is to be alternately quickened and retarded [slowed], but so that one process may compensate for the other” (quoted on p. 317).

But then Paderewski rejects this idea in 1909: 

“…The making up of what has been lost is natural in the case of playing with the orchestra…With soloists it is quite different. The value of notes diminished in one period through an accelerando, cannot always be restored in another by a ritardando. What is lost is lost” (quoted on p. 320).

Hudson also mentions (on p. 334) the frequent practice of arpeggiating or “breaking” chords among early 20th-century performers, a famous example of which is heard in Rachmaninoff’s recording of his famous Piano Concerto no. 2, op. 18. 

Rachmaninoff easily had the hand size to strike all the chord tones easily if he had wished to (by some accounts, he could reach a 13th), but he places the bass notes just before the rest of the opening chords’ tones: 

piano rubato in Rachmaninoff Piano Concerto No. 2

This is also a type of rubato. While it is not always considered fashionable to play with these types of chord breaks today, it’s certainly a relief for those of us with small hands to hear the composer himself treat the chords in this way!

Post-Tonal/Contemporary Rubato

In short: with the increasingly specific notation that composers begin to adopt in the post-Romantic era, it becomes easier for us as performers: we can usually just follow what the composer says. 

As Hudson notes, “We have already noted the general tendency during the course of the Romantic period for the composer to increase the number of tempo directions” (p. 356). 

On pages 340 and following, Hudson elaborates on Debussy’s specific tempo and rubato markings, which he needs to include because he is no longer really following a tonal system that sets up harmonic expectations on which we can base tempo choices. 

Another advantage of this era, as we saw in the earlier Rachmaninoff example, is that composers like Bartók and even Debussy lived long enough to have their performances recorded. (A few years ago I consulted, among others, Debussy’s own recording of his art song “Green” to prepare for recording it at Oberlin.) 

For example, while Bartók and Debussy wrote numerous tempo change directions, Stravinsky described his own music (at least in his middle and late career) as containing “metronomic strictness, no rubato…mechanical regularity” (quoted on p. 382). 

He makes it clear in his markings that tempo inflections should not be layered on. As a lifelong composer myself and someone who undertook significant undergraduate and graduate coursework in composition during the early 2000s, I’ve always found that we are encouraged to be extremely clear in our notation of all matters, including tempo and all types of rubato, since we’re currently living in such a musical “melting pot” of styles and aesthetics. We can’t rely on players being conversant in a single tonal language like composers in 18th-century Germany or Austria could.


Rubato is a multilayered concept that we can’t cleanly divide into each of the style periods our audition and competition programs ask for. 

After all, the style periods were never that cleanly divided in the first place: as I jokingly say to my students, Bach’s body (d. 1750) was still warm when Mozart was born (b. 1756). Beethoven and Chopin overlapped fairly significantly in their active compositional years, too, not to mention people like Schubert, Rossini, and Czerny. 

I find Hudson’s concepts of the earlier rubato and later rubato more helpful, since they are different techniques that have been available throughout the common-practice period and don’t require us to put each style period into its own box. 

I know this discussion may have raised more questions than it answers—but good discussions usually do! 

I hope this article has provided some inspiration for your own playing and perhaps piqued your interest in some composers you hadn’t checked out before.

Did you learn something new?

If you're a pianist looking for the best destination for all of your piano practice needs, I strongly encourage you to check out tonebase Piano

The team over at tonebase has created the best piano learning platform on the market, with hundreds of exclusive courses from the biggest pianists in the world, from Emmanuel Ax to Jean-Yves Thibaudet and more. 

As a bonus, they offer weekly livestreams with the artists on the platform, as well as an active forum of passionate pianists and custom annotated practice workbooks and scores. 

If you'd like to access these benefits, click here to sign up for a free 14-day trial. 

Happy practicing!

Learn From The World's Leading Pianists

Online lessons, courses, and interviews with the greatest minds in classical piano.

Get Started
Enter your email below to receive free lessons, PDFs & more!

Or, see how tonebase can take your practice to the next level today!

learn more →
Share the learning:
facebook logotwitter logolinkedin logo

"I don't regret for a minute having spent the money on the membership. There's something for every musician on tonebase – I recommend you give it a try."

Photo of Dave
Dave McLellan

Concert & Chamber Musician

Join over 10,000 fellow musicians improving every day on tonebase.