Domenico Scarlatti - the Sheet Music

The purpose of this publication is to enable as many harpsichordists as possible to play Scarlatti's music as he would have played it.

It is formatted so it can be printed portrait orientation on modern A4 or North American letter paper. (The original manuscripts are mostly 37x26 cm landscape.) The files are in Postscript code, which may be printed on almost all computers and printers with GhostScript.

These manuscripts are produced by the LilyPond typesetting system, which is still under development. Some note durations are confusing because I can't stop LilyPond putting note heads and beams on top of one another. And, part rests are missing because LilyPond has no mechanism yet to position rests. However, they are playable as they are, with care.

Download them here

Why Yet Another Scarlatti Edition?

First, this one is free, it doesn't cost a thousand dollars or more. Specifically:

This edition is Copyright © John Sankey, 2000, under the Berne Convention in order to protect the right of all to continue to freely use it. You are free to make and distribute as many copies as you wish of all or any part of this edition, electronic or on paper, provided that the notice of copyright and permission to further copy is preserved on all copies. You are also free to modify all or any part of this edition and distribute it as much as you wish in any manner, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one. You are not permitted to deprive anyone of the right to continue to use and copy it freely, by further copyright or any other means.


[a fragment of the manuscript]

But there is another reason. In the original manuscripts of the Scarlatti sonatas, notes are divided between the two staves at middle C, not to indicate whether the right or left hand should be used. Note stem directions were mostly chosen for maximum readability of the score. The apportionment between hands in the Longo edition was done by Longo, writing as a piano player long after Scarlatti's style had fallen into disuse, and omits most of the hand assignments explicitly written in Scarlatti's original texts, let alone showing the variety of hand-over-hand variations natural to his music. This edition presents the results of my study of Scarlatti, by writing out not only the hand crossings explicitly written in the manuscripts, but also examples of alternative hand crossings in the styles I consider Scarlatti likely used while playing. These alternatives are firmly based on my study of all the 555 sonatas, and of the keyboard techniques of Scarlatti's time used on single-manual harpsichords of his time.

There are many treatises of Scarlatti's time on ornamentation, phrasing, variation and other such essentials of musicianship. But, these all omit Scarlatti's hand crossing style because his playing and music were unique in this respect. Almost all of his music is written with implied pauses in one hand or the other that permit variations in hand crossings - right over or under left, left over or under right, even intermixed. From the consistency of these pauses throughout his music, I am convinced that this was the primary variation technique that he used. I believe that you should be able to play most Scarlatti phrases with either hand interchangeably, and make them sound exactly the same either way, if you are to understand his keyboard technique.

No one who knows only Czerny-derived technique, with its expectation that arms remain at right angles to the keyboard, will be able to play even half the variations shown in the manuscripts. You must learn to play with arms completely parallel to the keyboard, with finger action like legs walking, in order to play with hands crossed to the extent Scarlatti wrote. A bass interval played with the right hand is done with thumb on the upper note, not the lower. Left hand bass octave runs are played with thumb action similar to the heel-and-toeing of organists and 5th finger crossing over the 4th, not thumping the way pianists are taught today. Rapidly repeated notes on a harpsichord may be smoothly played with one finger, as if they were half a trill. Scarlatti explicitly noted when he wanted the rougher sound of 'changed fingers', a two-fingered 'trill' on one note - most can not be played as written with four fingers flailing the way pianists are taught today. And so on. Using techniques of Scarlatti's time, on a keyboard of his time, these sonatas can indeed be played the way they are written here. However, when I made my recordings, I used fingerings that made them sound their best. Wild hand crossings belong in front of a live audience, intricate ones are for private enjoyment. Neither work in a recording studio.

There were several pianofortes and at least one 2-manual harpsichord at the Spanish court, and some of the sonatas appear to have been written for each. As Kirkpatrick has shown, however, no instruments at the court other than the Spanish single-manual harpsichords had the note compass needed to play all of the sonatas. One might think that occurrences of things like an 8th note + 8th rest in parallel with a 16th rest + same note 16th + 2 other 16th notes would indicate 2-manual usage. But, when I go through all the sonatas in detail, I'm convinced that it didn't to Scarlatti: I can find no correlation with overlapped notes vs. non-overlapped. This usage seems designed to say something about musical context, not about manual usage. What I can say is that about a fifth of the sonatas work with two manual overlapped-note sound, but that I find the one manual sounds to be more consistent across the oeuvre.

Editorial Philosophy

A manuscript, originally presented to Queen Maria Barbara of Spain, is preserved in its original 16 volume binding at the Biblioteca Marciana in Venice, Italy. A second copy, mostly in the same hand and apparently made at the same time for Scarlatti's own use, is kept at the Biblioteca Palatina in Parma. Ralph Kirkpatrick arranged for the publication in facsimile of the Parma manuscript, and of some other Scarlatti pieces preserved only in derivative collections. As a result of Kirkpatrick's publication, under the United States copyright law of the time, the content of his manuscripts is free from copyright and ownership restrictions, and is the authority for this edition.

Printing is black and white, and music isn't. A musician has to peer through the limitations of notation to hear a live person, the composer, then bring that composer's ideas to life in the performer's world. Successful users of this edition must do the same. In Scarlatti's time, a performer was expected to be as able a musician as the composer. Music was read and re-created, not memorised note for note. I do not know when it became acceptable to play absolutely identical repeats, but it was after Scarlatti's time! Notation was intended to communicate the essence of the music rather than to prescribe how to play it. Performers were reminded of the composer's usual stylistic elements only occasionally, rather than by writing them out in full. This edition does the same - it is not for neophytes, it is for musicians. (If you play the piano, I recommend the Longo edition - Longo was a good piano teacher.)

Every note and duration from the Parma manuscript is preserved. You will find no modern 'interpretation' or 'corrections' here in that respect. I have avoided like the plague that panacea of small minds, 'regularization' of original variation. Beaming is preserved whenever possible, as Scarlatti often used it to indicate articulation. Triple time was indicated then with an extra beam (e.g. 3 beams for triplet 16ths) compared to today's practise - this has been modernised. (In a few instances, I have had to simplify Scarlatti's beaming in complex passages to make the score legible with modern typesetting.)

Scarlatti wrote almost all note heads with separate stems, as if they were separate voices, even when they obviously aren't. Close-spaced notes of identical length that sound as chords are stemmed together here for clarity. Hopefully, you will find the result visually matches the hand motions required to play them, and clearer to read than the original.

The Parma manuscript was obviously used by several players. In some cases, notes in the manuscript have been lengthened by converting a solid note to an open one, or adding a dot. Since these changes are essentially all in the direction of converting the notation from a part-oriented one to an as-played one, and appear to have been done at the time, I have used the modified durations - this is a player's edition. However, several of the sonatas also have romantic-style slur marks - since these are not typical of the harpsichord and are not matched in other manuscripts, I have omitted them.

Scarlatti used what is now interpreted as a double-sided repeat bar (:||:) to end both the first and second halves of almost all of his sonatas. He was otherwise so meticulous about writing all repeated passages out in full that I consider that the :||: bar was used primarily as an ornamental device. In order to avoid confusion with modern use of the symbol, I have used a plain double bar to separate the first half from the second, and an end bar to end each whole sonata. Occasionally, a single bar, often of fractional length or containing solely the part for one clef, is enclosed in :| and |: brackets between the first and second half. These are obviously optional passages, to be used when one half is being repeated, but the usage is so inconsistent and sketchy that I felt it best to omit them. Readers should remember that performers were expected to provide graceful transitions between sections of music, cadenzas even, and not be rigid in their performance of this music.

Scarlatti's music ranges over the keyboard to such an extent that he had to write his music on a staff pair, using ledger lines only for middle C and for notes below the bass or above the treble clef limits. If one staff is used for each hand, his music becomes an illegible mess of clefs (top right, from the Santini copy) or ledger lines (middle right, Longo edition). If each hand change is marked throughout, the music also can be confusing to read (bottom right, from the Fitzwilliam manuscript). However, showing detailed hand dispositions by two note shapes works well - I am indebted to the skilled programmers of Lilypond for modifying their programs to provide them for this edition. A diamond note head indicates the left hand, the usual oval head the right. Notes that Scarlatti did make concerning hand crossings are preserved in the form he wrote them, D (dritta, right) and M (manca, left), in the first example of their occurrence in each sonata.

Scarlatti showed a variety of articulation for many passages - I do the same for hand assignments. Readers are expected to play enough of the sonatas from this edition to be able to choose an effective performance style for them, and to not parrot the samples as shown. The right and left hand note shapes chosen have similar visual weight to facilitate this. Usually, due to the short length of the sonatas, a single pattern of variation should be chosen for consistent use throughout a piece for performance, using a complementary variation for repeats if they are played. Mostly, I show examples that are visible to an audience. There are hundreds of other workable ones that are fun to use among friends, if the reader chooses.

Variations in hand usage and articulation imply variations in the lengths used for notes. Appropriate note durations in Scarlatti depend very much on the way the piece is played. In particular, lengthening notes is an essential aspect of good harpsichord sound. But, I have a horror of editions by, in Thurston Dart's memorable phrase, "Herr X, amended by Dr Y, and thoroughly revised by the eminent pianist, Mr Z" - I have seen too many of them. Scarlatti normally set note lengths to form separate, although often fragmentary, voices rather than to prescribe the length to be played. Some must be modified on a single-manual instrument. Some must also be modified to match the hand disposition chosen. His durations are perfectly legible to the modern reader, so I leave it to you to turn them into music as he did, on your instrument, in your surroundings, as you play it.

Scarlatti generally used the same notation for acciaccatura and appoggiatura - a small 1/32nd note. Since modern players are mostly unaccustomed to the distinction, I have written them as they should be played, according to my judgement. Those I interpret as acciaccaturas are written as simultaneously struck full-length notes - Scarlatti often wrote them as such when they formed part of chords. Those I interpret as appoggiaturas are realised in the standard baroque manner. Both such usages appear with a smaller note head than the rest of the text, as Scarlatti wrote them, to indicate that they should not be played rigidly as written.

Scarlatti used the baroque trill symbol and tr, for brief tremulo, apparently interchangeably as a general sign for a note to be stressed. When he wanted a continuous trill, he seems always to have written the word tremulo out in full. Since modern players are accustomed to play trills in the later classical manner, I have used the baroque symbol throughout. The symbol usually means a trill, starting on the upper note and not terminated by a turn - Scarlatti wrote most terminations out using 1/32nd notes. However, it can instead imply an acciaccatura, appoggiatura, inverted mordant (several of which appear in other copies of the sonatas made at the time), short trill, or nothing at all, depending on the context, the instrument used, its acoustic surroundings, and the tempo of performance. So, an editor cannot pre-choose their performance. Players who wish performance examples are referred to my recordings.

Accidentals here follow modern usage - an accidental continues until the end of the bar. Scarlatti normally used a flat to cancel a key signature sharp, but a natural to cancel a key signature flat. Sometimes, a natural was used to cancel an accidental preceding it in the bar, but usually accidentals were written on each affected note and did not apply to other notes. However, when space was tight, and it was considered obvious to a player of the time what the proper pitch was, accidentals were omitted. I have used my best judgement to interpret ambiguous cases. Scarlatti's E# and B# are replaced by F and C respectively, following modern practise. I have retained Scarlatti's other enharmonics (i.e. Eb vs. D#) except when it would impair legibility of the music with modern notation. I note that Scarlatti's music shows no evidence that enharmonics were important to him - he composed almost totally by sound, at the keyboard, and all of his music is suited perfectly to a closed 12-tone tuning.

For the sonatas clearly intended for solo and continuo, I have provided a realization suitable for a student and teacher format. Despite the thousands of pages written on the subject, there is only one real 'rule' for realizing a continuo - it should make the soloist sound fantastic. That depends so much on the instruments, surrounding acoustics and social ambiance that no editor can do it for you. In form it can range from totally self-effacing support to a full-blown trio sonata for solist and harpsichord or gamba. If you can, you should adapt the examples I give, or write your own, for your performances.

I am greatly indebted to Jim Bailey, who has undertaken the demanding task of proofreading these scores against the original manuscripts. I thank the many helpful people on, comp.lang.postscript, and comp.text.tex who have assisted me with this project, especially Han-Wen Nienhuys, Jan Nieuwenhuizen and Mats Bengtsson. See the Mutopia project and the GMD music archive for more free sheet music.

John Sankey
other notes on harpsichord playing