Like any harpsichordist, I worship at the altar of Domenico Scarlatti. The high point of my devotions was in 1985, when the tercentenary of his birth was celebrated, together with those of J. S. Bach and Handel. I got the task of distributing 555 sonatas among 60 performers, to be performed in Amsterdam the space of a week. In 1991 I recorded 18 sonatas for TelDec – a CD that, within months of its release, went down with the label when it sank under the control of Warner Music. Since a mild stroke robbed my right hand of all velocity I’ve had to move on to other, more ancient and less athletic deities. But while surveying Gustav Leonhardt’s recordings of the composer, I was struck by one of Mimmo’s most profound utterances: number 52 in Ralph Kirkpatrick’s catalog.
I would have chosen to record this Andante moderato, but I was too much in awe of the piece at the time and couldn’t decide how I wanted to interpret it. The unusual tempo indication was one obstacle. Did Scarlatti mean faster than a normal andante (and what is that, precisely?), or slower? Another problem was general content and texture. It reads and feels like a slow movement from a Handel concerto grosso, or like what one of his famous improvised organ voluntaries might have been. I thought, and still think, that it must be an echo of the famous competition between Handel and Scarlatti, where the Saxon was adjudged winner on the organ after a tie on the harpsichord.
K52 still being within my diminished technical means, I wanted to see if I could now settle on a way of playing it, with an eye to a possible recording for this website. I struggled for awhile with the slow side of andante moderato, motivated mostly by the extraordinary pedal-point passage towards the end, which Kirkpatrick aptly described as “Brahmsian”. Then I checked the other sonatas with the same tempo marking, and found that I had the wrong end of the stick. The operative word is moderato, tempered by andante. J. C. Bach sometimes prescribes Andante molto, and one finds Andante con moto frequently in the Viennese Classics. These tempi are slightly on the slow side of an Allegro moderato. It makes the many examples of figura corta (two 16ths preceded or followed by a longer note) found in K52 much easier to play than when taken slower.
Checking Kirkpatrick’s catalog for the source, I was surprised to find that K52 is present in two versions in the earliest (vol. XIV, 1742) of the 15 volumes of Scarlatti sonatas now kept in the Biblioteca Marciana in Venice, which were luxuriously bound for Scarlatti’s patroness, Queen Maria Barbara of Spain. A second volume (Venice XV) is dated 1749. The volumes numbered I to XIII containing later works were copied in relative haste, possibly in view of the composer’s deteriorating health. The compact dating (1752-1757) ends in the year of Scarlatti’s death. The court copyist responsible for these 13 volumes, José Alaguero, produced another large series of manuscripts now in Parma. There are 15 volumes in the Parma collection as well, but they contain fewer sonatas in total than Venice. Add to this the fact that their outward appearance is modest, and it is understandable that Venice was chosen by Kirkpatrick as his primary source. A later editor, Emilia Fadini, made the same choice. Her complete edition (Ricordi) is the only one than can be called “critical”. Both the Parma and Venice sets were willed to the castrato Farinelli, long a veteran of the Spanish court, when the queen died.
The two earlier Venice volumes (XIV / 1742 and XV / 1749) were originally unnumbered, but were given higher Roman numerals (in spite of their containing earlier works) by Alessandro Longo, the first Scarlatti scholar to attempt a complete edition.
***
A student of a great composer wants to know how his subject developed. A chronology of the corpus is obviously the first objective. For music of the distant past, this can be surprisingly difficult. In the case of Domenico Scarlatti, it is unattainable, except along the broadest of lines. Let us consider the question before returning to K52, since it will help to put that excellent piece in perspective.
Kirkpatrick, faced with a difficult task, chose to go by dated sources, starting with Scarlatti’s own printed Essercizi of 1738, followed by Roseingrave’s pirated edition of 1739, which included 12 sonatas which that editor had collected himself. The catalog then continues with the dates of the Venice collection (XIV / 1742, XV / 1749, I / 1752, continuing consecutively to XIII / 1757), interspersed with various secondary sources. The famous total of 555 sonatas is chimerical. It includes nine movements for melody instrument and continuo, as well as a considerable number of more or less obvious fakes. There have been numerous claims of “new” sonatas since Kirkpatrick, but I find none of them convincing.
Venice XIV / 1742 is, at 61 sonatas, the largest volume of the two big Spanish collections. It starts with a series of 15 masterpieces, 12 of which also were accepted into the Parma canon. These 15 may have been intended for a new publication. After no. 15, there is a clear change of intent. No.16 (K58) is a weak, early attempt at a chromatic fugue. For the rest, this long concluding section is chaos until the final no. 61 – a collection of mostly early works, few of them worthy of remembrance, obviously collected here as an archive of things past. Some of them were probably teaching pieces, which Scarlatti may have wanted to preserve for the queen. Kirkpatrick was aware of the chronology problem thus presented, but says he thought it “unwise” to depart from secure dates in favor of stylistic considerations.
This long later section includes the nine movements with continuo, and a few flashes of mature genius hidden within the general desolation. K64 is bizarre gavotta, possibly from the Venice period. K84 is fine C minor Allegro. Special mention must go to K69 and K87 — two gorgeous slow movements, similar in their unusual style and texture. These last three possibly came directly into José Alaguero’s hands during the copying process, i.e., around or before 1742. Dare I suggest that the profound sorrow which characterizes K69 and 87 might be connected with the death in 1739 of Catalina Scarlatti, the mother of the composer’s five children?
Venice XV, finished in 1749, is the archive of the first Spanish style, from the time Scarlatti spent in Seville beginning in 1729. Its extravagant leaps and passages taken from folk music and flamenco were not fit for publication in 1738. They may have also been reserved for the exclusive pleasure and use of the Crown Princess, as she then was. Such a restriction would explain the near-total lack of contemporary Iberian manuscripts of these and all the later sonatas.
Venice I (1752) commences the archive of later pieces. It begins with simpler works, likely used for teaching Maria Barbara. There is a rough consecutiveness between the later Venice volumes and Parma. It may reflect to a considerable degree the order of composition. One could imagine a collection of autographs in Scarlatti’s house in Madrid, which grew as compositions were added, and which were handed over approximately in that order to the copyist. But it is patently absurd to think, as Kirkpatrick theorized, that the nearly 400 sonatas in the royal archive Venice I-XIII, dated 1752-7, were actually composed in those years. For one thing, Scarlatti was already an elderly man by then. His only surviving letter (1752) states that he cannot leave his house. That this was for reasons of health is confirmed by his reference to the good health of his recipient, the later Duke of Alva. For another thing, there are too many differences of style for such a short period of evolution. For a third...the idea defies common sense.
With five exceptions, K1-30 (the printed Essercizi of 1738) do not appear in either Parma or Venice, because they weren’t needed for these archives. The book was dedicated to Maria Barbara’s father, and I imagine copies were present at all four of the major Spanish palaces visited by the court in a fixed seasonal rotation every year. The five sonatas from the Essercizi which do appear in Venice XIV are represented by earlier versions in the latter section of that manuscript. Scarlatti’s 1738 publication probably represents work produced in Lisbon in the 1720‘s, under the patronage of its dedicatee, King João V; mature pieces, but as yet unaffected by the onslaught of Spanish folk music which Scarlatti underwent when he followed Maria Barbara to Seville.
Kirkpatrick’s second, dated section (K31-42), the Roseingrave print (London, 1739*), contains besides the 30 pirated Essercizi twelve sonatas of clearly earlier date. These are relics of the publisher’s sojourn in Venice sometime before 1720, when he was closely acquainted with Scarlatti. One of them (K39) is an early version of one of the Essercizi (K24). Scarlatti clearly saw potential in the piece, especially the opening motif, and completely reworked it. This is to my knowledge the only example of such a reworking in the Scarlatti corpus.
Five of Roseingrave’s twelve “new” sonatas also appear in Venice XIV, once again in the later section. One of these (K37) shares characteristics unique in Scarlatti’s oeuvre with other works in Venice XIV: there are numerous fragments which are so obviously transposed to a wrong octave that editors should have noted the fact, but in all but a few cases failed to do so. A comparison with Roseingrave’s edition shows that in the case of K37, where he uses many C-clefs, we are dealing with simple mistakes in transcription from C- to G- and F-clefs. This may seem surprising in a court copyist, but there is no other possible explanation. Scarlatti at some point stopped using C-clefs, and began notating everything on two staves in the modern fashion, with the hands moving freely throughout them. Oddly enough, Roseingrave includes an example of the shift: the first two sonatas in his edition are two versions of the same piece (K8). The first is his own souvenir de Venise, the second is the version from the Essercizi, which was obviously supervised by the composer, and contains no false octaves.
One case of false octave transcription, which was so obvious that Kenneth Gilbert changed and annotated it, occurs in K76, a unicum in Venice XIV. In bars 26 and 27, eight notes in the midst of a fast downward scale, which must have been in the alto clef in Alaguero’s model, appear an octave too high, interrupting the tirade. This is the only one of many errors of this nature which Gilbert corrected. Fadini goes much farther, if not far enough, and we can be grateful to her for finally giving us critical notes.
This brings us back to K52, one of the works in Venice XIV which have false octave transpositions but are not in Roseingrave. In the earlier version of the two found in Venice XIV (no. 61) they are of greater extent than in any other sonata, especially towards the end. They comprise some 2/5 of the whole piece.** It may be of significance that this earlier version is the last work in the manuscript, after a fuga which doubtless was intended to close a collection of 60. (The 30 Essercizi end with a fuga, as do Venice VIII and IX and each half of Sebastian Albero’s 30 Sonatas.) That Alaguero nodded at the end of his huge task can be excused. Perhaps more puzzling is the presence of a revised version at no.10. Why the late duplication, with so many octave errors? Perhaps the old version mistakenly remained at the bottom of the pile of manuscripts left him by Scarlatti for copying, and Alaguero, working by candlelight with his jug of Rioja beside him with the prospect of an end in sight, made too long a night of it. During one of our discussions of these problems, John Koster pointed out an interesting parallel: the chaotic final section of the original edition of Bach’s Kunst der Fuge mistakenly includes (pp. 45-7) an earlier version of Contrapuntus X. This was a result of a similar misunderstanding; manuscripts connected with the project were in disorder and Bach’s musical heirs didn’t know what his intentions were.
Three pieces from the “good” first 15 sonatas in Venice XIV didn’t make it into the Parma manuscripts. K45 might have been an early-period candidate for revision, like K52. It has one of the long wavy lines (omitted by Gilbert), like the many in K52, which have been the subject of much speculation. K51, a somewhat inferior work, is in the key of E-flat, which Scarlatti may have ultimately considered too unusual for a prospective publication. That leaves K52. Why would Scarlatti have held back one of his most beautiful works from inclusion in the big Parma archive?
There is at least one reason why he may have wanted to revise it a second time, and it is the element which led me to this investigation in the first place. His revision of what might henceforth be called K52a led to a grievous error in counterpoint: in no. 10, bar 39 has a doubled dissonance***, followed by near-parallel octaves. In no. 61 the passage is flawless. Scarlatti’s other revisions are all to the good. The task of revising K52 might then have been forgotten, along with K45 – left in a cubbyhole, to be eventually destroyed along with every single one of his autograph manuscripts, to the enduring grief of all Scarlatti scholars and all true music lovers.
October 5, 2023
*Only one piece from of either of these publications made it into the Parma archive: K 41, a Handelian fuga.
**Other interesting aspects of no. 61: the double repeat in the middle is impossible. It simply doesn’t work, neither for the first return nor the end of the piece. No. 10 provides first and second endings. There is, however, a “D.C:” marked at the end of no. 61, as is the case with almost all the Venice XIV sonatas. This sign, omitted in all editions, would indicate a repeat (or third repeat in “normal” sonatas) from the top, much like those mandated in d’Anglebert’s print, and prescribed by Georg Muffat. No. 61 also lacks the flat in the key signature, an archaic usage by the time K52 was first revised. Fadini thinks no. 61 is a valid version, and presents both in her addition. She attributes an unplayable passage towards the end to the use of pedals. There were, in fact, early Neapolitan harpsichords with a few pull-down pedals, and larger Italian organs had simple pedalboards. But there is too much wrong with no. 61 to consider it anything but a failure of transcription from an original with C-clefs. The mistakes were corrected in no. 10, and many details were improved.
***He saw an opportunity for a powerful third-inversion dominant 7th, and put g in the bass on the second beat. He forgot the g’’ in the discant. Both resolve downward within an 8th-note of each other.
click  to listen (mp3 file)
|