Klarinet Archive - Posting 000268.txt from 2007/05

From: "Lelia Loban" <lelialoban@-----.net>
Subj: [kl] R. Kell, Part 5: Corelli
Date: Tue, 29 May 2007 14:31:38 -0400

Part 5: Corelli
The Decca set includes Reginald Kell's own transcription of Arcangelo
Corelli's Gigue, from the Sonata in C major (Kell transposes it into
orchestral A-flat), Op. 5, No. 9. Kell plays B-flat clarinet, with a
rather shrill tone. There's no room for notey-ness in this brisk piece,
but Kell does use that style I've referred to elsewhere as "slangy," with
exaggerated phrasing. Brooks Smith accompanies inconspicuously on piano
(1957, from Decca LP DL 9926). Corelli (1653-1713) wrote the sonata for a
choice of violin or viola da gamba with harpsichord continuo. Corelli
wrote the Gigue in 12/8 but Kell transcribes it into 6/8.

My copy of the version for clarinet is Kell's edition for B-flat clarinet,
published by International in 1958. (I said I didn't have any more Kell
editions, didn't I? I lied.) Are the Decca liner notes correct that this
sonata was originally in C? I compared Kell's edition with a well-
regarded one for violin and piano (Vol. 2 in the 2-volume set, "Corelli
Twelve Sonatas Op. 5 for Violin and Piano," from Schirmer's Library of
Musical Classics, no. 1904), edited by Gustav Jensen (1943-1895). Jensen's
edition is in orchestral A major. Transposing the piece from C (D for the
clarinet) to A- flat (E-flat for the clarinet) wouldn't make a lot of sense
to me in this piece, because those clarinet keys are equally convenient to
play with good intonation; but if the original key was A, then I can
understand why even such a highly-skilled professional clarinet player, who
should have no trouble at all in *reading* the score in any key signature,
might prefer not to record such an exposed clarinet part in the (clarinet)
key of B, a key that's hard to pull into tune with a piano.

This piece is easy to adapt for single-voiced solo instruments, because
Corelli, a violin virtuoso who pioneered the use of double-stops
(multiphonics) on violin, included only one double-stop, easily expendable,
in this three-movement piece (Largo, Giga, Tempo di Gavotta). Since
Corelli also wrote this gigue within the compass of nearly any instrument,
it's been transcribed for almost every instrument, including the trombone.
Kell's transcription rates about average for folding, spindling and
mutilating.

I don't know what musicologists think of Jensen's phrasing marks, but
Kell's are drastically different from Jensen's and from general performance
practice on violin, FWIW. I'm not going to try to compare and contrast the
numerous similar examples, but the most startling difference between
Jansen's edition and Kell's edition and performance is in the signature
6-note opening phrase, often repeated later. Jansen phrases it,
di DAH-yi, DAH-yi, DAH
(slurs from the second note to the third and from the fourth note to the
fifth, cutting the third and fifth notes slightly short for emphasis)
but Kell phrases it,
di-YUP, di-YUP, di-YUP
(slurs from the first note to the second, the third to the fourth and the
fifth to the sixth, cutting the second, fourth and sixth notes quite a lot
short for emphasis).

When Kell's edition came out from International, the year after the
recording, he had already changed his mind about much of his phrasing--or
maybe he didn't think it was important. For instance, he writes slurs over
the first two notes of the three-note phrases from bars 4-9, but in the
performance, he tongues those notes. He plays the slurs in the phrases
constructed the same way at bar 11, then reverts back to articulating. He
exaggerates his tonguing and the di-YUP contrast between tongues and slurs
ending in shortened notes. Those exaggerations sometimes make the music
sound sing-songy. It sounds as if he wants to end every sentence with an
exclamation point.

A wind instrument transcriber often needs to build in some places to
breathe, but the violin version of this score wouldn't need changes for
that reason until the second half of the second ending--and, even there,
it's easy to find good phrase-endings where a clarinetist could sneak
enough air with a breathing mark, not a rest. Kell doesn't write in
breathing marks or extra rests (in fact, he removes rests--see below), but
he cuts so many notes so extremely short that the performance sounds as if
it's full of rests. He gives himself an eighth-note silence at the
beginning of bar 49 that's not written in his score or in Jensen's violin
score. The silence sounds okay, although from the evidence elsewhere, Kell
had good breath control and didn't need to catch some air there--certainly
not that much air. His breathing is rarely audible and then only
slightly--he's not a gasper. Maybe he just thought the caesura sounded
nice and he figured he knew better than Corelli.

Corelli didn't write this sonata as an equal collaboration. It's for
soloist with continuo. However, Kell's version pushes the keyboard even
farther into the background than normal. Compared to the Jansen version,
Kell's leaves out a lot of the pianist's notes. Many bars in which
Jensen's score gives the pianist passagework have only chords in Kell. He
also gives himself more notes, in (Kell's) bars 29-36, for instance
(filling in the violinist's pair of 8th-note rests with two extra 8th notes
in each of these phrases); in (Kell's) bar 42, where he gives himself an
extra note; and at (Kell's) bar 49, where he gives himself four more notes.
In each of those cases, he gives the pianist considerably fewer notes than
Jansen's pianist gets. I don't know which version comes closer to
Corelli's original. Since Baroque composers normally sketched in the
continuo and left it up to the performer to improvise (musicians complained
to ecclesiastical authorities about J. S. Bach's habit of trampling on
musicians' prerogatives by writing out all the parts, including continuo,
in detail), the rather large difference in the number of notes between
Jansen's piano part and Kell's may mean that Kell stays more faithful to
the original keyboard notation, while Jensen tales the more typical 19th
century attitude that he should fill out the part instead of trusting the
accompanist to improvise on it--but I think it's likely that, in places
such as bars 29-36, Kell adds notes for himself into rests Corelli leaves
silent in the solo, with conspicuous gaps in phrases, because he means for
the harpsichord--not the soloist--to fill in those silences.

I'm surprised Kell passes up another opportunity to upstage the keyboard:
He eliminates the violinist's pair of trills in (Kell's) bar 36 and raises
the second of those formerly-trilled notes half a step in the Kell score.
Kell alters management of the repeats. In the Jensen score, written in
12/8, this section before the first repeat ends with bars 18 and 19. Kell
re-writes these bars considerably, and raises the note half a step on what
was formerly the second trill. Cumulatively, these re-writes have the
effect of eliminating 4 bars from Jensen (the equivalent of 2 bars in
Kell's 6/8) before the first repeat. In Jensen, there's an entire section
repeated (with a double-dot bar and a pickup note) from bar 20 to bar 45,
the last bar of the piece. Instead of using a double-dot repeat sign (as
he and Jensen both do in for the first repeat), Kell writes separate first
and second endings. Repeating from his bar 78 and then going through the
second ending adds 4 Kell bars. Thus Kell's score (minus 2 Kell bars
before the first repeat, then plus 4 Kell bars in the second section's
second ending) comes out to 92 bars when Jensen's 45 bars, if put into 6/8,
would be 90 bars.

Normally I don't trust 19th century editions, and I don't know which of
these two versions (maybe neither of them!) scholars regard as more
faithful to Corelli's original. I'm tempted to trust Jensen more, because
he constructs his version of the gigue as a typical Baroque binary sonata
(compare with the structure of Scarlatti's 550, for instance), in effect a
sonata-within-a-sonata. I'm thinking that's more typical of Corelli's
period than Kell's construction of the first section with a double-dot
repeat and the second section with separate first and second endings.
Usually, in an Italian or Spanish Baroque binary format, either the first
and second sections repeat themselves identically, both with double-dot
repeat signs, or both the first half and the second half have first and
second endings. They're not written half one way and half the other, the
way Kell's got them here.

Kell does the diva bit, playing (and writing) the first note of bar 8 up an
octave and the first note of bar 61 up an octave. He often switches the
keyboard player's octaves, too--by lowering them! No surprise at the end:
After taking full advantage of the instruction that appears in Jensen's
edition, "II-da volta più largamente" with a big, dramatic rallentando
(although, oddly, Kell doesn't include that instruction in his own
edition), he plays (and writes) his last three notes an octave higher than
in the Jansen version, and holds the last, high note in double forte
triumph, like a Broadway star belting "The Star Spangled Banner" at a
baseball game. Of course, in the original version, that isn't the last
note, because the sonata proceeds on to the Gavotte.

I might enjoy Kell's Gigue more if I weren't familiar with the way
violinists usually play the sonata. If Kell had regarded this piece as
nothing but a showy trifle to toss off for an encore, I'd expect him to
play it faster (he's a bit slow, compared to most violinists), but still, I
think it's a somewhat tasteless performance.

Lelia Loban
"Corelli's excellence lies in the chastity of his composition."
--John Gregory, "The State and Faculties of Man," 1766
What does an unchaste composition do, then? I'm picturing that scene with
the cellist and "just your average horny little devil" in "The Witches of
Eastwick"....

------------------------------------------------------------------