Mind the Gap! – Practising the Piano

There is a problem with certain performance directions composers write in the score – or the lack of them. Especially articulation marks. Some of the markings we find can easily be misinterpreted, especially when applied generically across the style periods or from one type of piece to another.

The staccato dot is one example – sometimes it means a literal shortening of the note, other times it might mean a type of accent. And it does not mean we can’t use the pedal either. Notice the dot on each of the main beats in the LH of Chopin’s E flat Nocturne. This signifies a slight stress on each bass, and it is evident not only from Chopin’s pedal markings but also from the harmonic context that the bass notes are to be caught in the pedal.

In the absence of specific markings, is a piece legato or staccato? May we add slurs and phrasing of our own? This depends on the speed and character, as well as its style period. There is only a certain amount of information a composer can write on the page, and in much music from the earlier periods performance decisions were either left up to the good taste of the performer, or players would have known how things were supposed to be done.

Bach left a certain amount of articulation markings but a glance at a page of any Urtext edition may show none at all, or very few. In music from later periods, particularly the Classical period, a question that regularly comes my way is how to play the slurs and short phrases that are marked in the score. Do we always need to make a gap in sound between the note at the end of the phrase mark and the next note, or is it OK to connect? I’m afraid I can’t offer you a neat one-size-fits-all solution – it all depends on context.

From the numerous treatises of the Baroque and Classical eras, it is possible to draw some very practical conclusions about certain matters relating to style and performance practice. When it comes to slurs, Richard Troeger states:

The first note under a slur receives a slight accent, and the ensuing notes are successively softer. Thus, a slur implies (indeed, indicates) a slight accent followed by a slight diminuendo to the end of the slurred groups, as well as a legato connection. (The diminuendo helps the keyboard player to approximate a real legato, whereby the attack of one note matches the dynamic level reached at the end of the previous note’s sounding.) Naturally, the degree of accent must be made relative to the context. (Playing Bach on the Keyboard, P. 104)

We can all feel how the two-note drop-roll slurs work in lively or stylised dances, such as minuets. We play the first note with emphasis, and the second note shorter and softer, lifting out of the keyboard as we release. There is a distinct gap in sound at the end of the slur. When it comes to allegro movements in Mozart, for instance, the slurs and short phrases can sound so choppy and disruptive to the line when they are played with such pedantic separations. Ignoring them is equally wrong and unstylistic, so how do we deal with them? One thing we must do is to play the notes under the slur legato (except when staccato dots are present too – in which case we use a mezzo staccato or portato touch).

Check out the two-note slurs at the start of Mozart’s Sonata in B flat, K. 570.

On the modern piano the opening phrase can sound very choppy when the crotchets are lifted. However, they work fine on the type of piano Mozart was writing for, where the attack is more immediate and the decay more sudden. Listen to how Malcolm Bilson articulates them on his fortepiano:

András Schiff also observes the articulations but not pedantically so. The first and last slurs are clearly punctuated, the middle one less so (I am wondering whether this might have something to do with string bowing, presumably down bow on the first and third bars, up bow on the second and fourth).

Maria João Pires does the exact opposite, playing the whole phrase under one legato slur and stressing the supposedly weak crotchet:


But does this make Bilson and Schiff (and others who play the slurs literally) “correct” and Pires (and others who go for the long line) “wrong” or unobservant? Hardly. Pires is a superb artist who must have considered this opening, presumably reaching a conclusion that satisfied her musical intentions. While many would disagree with her, there is no black and white solution. She may have decided to apply the longer slur because in later instances when the theme returns, it is indeed marked as such (possibly as a foil for the articulate counterpoint that joins it?).

In this short video clip, Bilson explains how the short articulations Mozart writes very clearly in all his piano music might not always work so well on the type of piano we play today.

If the context makes a literal gap in sound clumsy or impossible, we can still observe the slur by thinking of it in terms of the string player’s bowing, or the wind player’s tonguing. A string player would start a new bow at the beginning of each slurred group, and the wind player would articulate the breath by tonguing a “t” syllable (for a sharper attack) or a “d” syllable for a softer one. Thinking like this is very helpful for us pianists too. Thus groups of four slurred notes might be articulated by thinking in the following ways:

ta-la-la-la, ta-la-la-la (sharper)

da-la-la-la, da-la-la-la (softer)

ta-la-la-la, da-la-la-la (a mixture for longer groups of notes)

wa-la-la-la (even softer)

shma-la-la-la (leaning and loitering a bit on the first note)

A little background may help here. As Malcolm Bilson explained, in the Classical period music was highly articulated. We pianists generally feel more comfortable with the tradition we inherited – the long legato line that was in place by the end of the 19th century. In Mozart’s day nonlegato touches were more prevalent, and composers were compelled by tradition to show metrical accentuation (the hierarchy of the beats), meaning slurs almost always stopped at the barline. Here is a good example from the slow movement of the A minor Sonata, K. 310.

It is pretty obvious that the performer would slur over the barline (bar 22 into 23) in exactly the same way as has been set up in the first two beats, only Mozart was not permitted to write this because of the conventions of the day. Many later editions “corrected” this to make it more user-friendly for today’s players. Here is the Peters Urtext (ed. Carl Martienssen), c. 1938:

If a composer wanted a prevailing legato he was not permitted to write slurs that crossed the barline, so he had to use bar-length slurs until Clementi decided this was silly and confusing, so he changed it.  Sandra Rosenblum, in her discussion of this subject (Performance Practices in Classic Piano Music, p. 172-183), shows how Clementi’s modernised the notation of the slurring in the 1807/8 revision of his Sonata op. 13, no. 6 (first published in 1785, in the old style).

I have heard performances of Classical period music where the pianist seemed preoccupied with carefully curating the articulation marks, but there was little else going on in terms of characterisation. All the t’s were crossed and i’s dotted, but this alone was not enough for the musical message to come across. This should remind us that strict observance of the letter of the law is far less important than being true to its spirit.

I have more to say on this subject, in relation to Brahms and Ravel, but I think this post is now in danger of getting too long. Look out for Part 2, next week…