Transcribing, arranging, translating: reflections on Albéniz on the guitar

In the shadow of Gödel, Escher, Bach’s extraordinary intellectual and linguistic virtuosity, Douglas Hofstadter’s later work Le Ton Beau de Marot is less well known. It combines the high-minded playfulness of the earlier work with deep reflections on the nature of language, meaning and translation, with intellectual autobiography, and touching and finally heartbreaking personal reminiscence. The thread of the book consists of multiple translations, by Hofstadter and many others, of a gem-like lyric by the Renaissance French poet Clément Marot, from which are hung — amongst other digressions — commentaries on what constitutes translation, particularly of poetry: where the meaning of a poem resides (clue: not just the words) and what equivalents for rhyme, metre, style, sound (and more) are available to the writer brave enough to attempt to create a poetic equivalent in another language.

All this has been at the back of my mind for many years as, like many other guitarists, I study and perform others’ arrangements of music for the guitar, and as I make my own. The case of Albéniz is particularly interesting: here is music which in arrangement has become core to the guitar repertoire, originally written for the piano, but in an idiom suffused with the textures and figurations of Spanish flamenco and folk guitar. I’ll touch on some of the consequences in the notes below.

Torre Bermeja

Here’s the opening, from the earliest edition (1912) available on IMSLP:

Choice of key/transposition is a critical initial decision. I’ve not seen this piece arranged or played in anything other than D (with ⑥=D tuning), though it is instructive to read through it in the original key to see where difficulties lie. In D, the two tonic-note open strings on ⑥ and ④ and the dominant on ⑤ make for fluent readings in the rich textures of the main minor section, and resonant tonic and dominant chords throughout.

Here is how Miguel Llobet arranged this (UME 19511, assumed arranged in the early years of the twentieth century):

Arranging is about choice, and we can identify a number of choices here.

• Key — D major, as noted.
• Tempo, dynamics — at least in this extract, rather faithfully reproduced (note however the absence of the rapido annotation in m.2, and the different starting points of the hairpins in m.6 and 7)
• Opening flourish — this is where things get interesting. There’s no reason not to transcribe the intervals in the original exactly. It’s eminently playable, and perfectly idiomatic. Looking at Llobet’s right hand fingering, his goal appears to be to create a dramatic opening gesture with a glissando between the first two notes, in spite of the fact that the original marks each of the notes with a staccato. Coming back to Hofstadter, this feels to me like a real translation — substituting for a word from the idiomatic language of the piano (an upwards rush in a double arpeggio) one from the language of the guitar. To my mind this is an unnecessary substitution: the original word, as it were, is quite satisfactory when expressed through the medium of the guitar. However, we can make a case in Llobet’s version for some essential part of the meaning of the gesture being preserved.
• Chord at start of bar 2 — re-voiced as a 4-note chord. No pianist would arpeggiate this chord, nor others where a big resonant attack is required. Most if not all guitarists at this point arrange/play a 6-note D chord using all the strings, strummed with the RH thumb and rapidly damped. If the word we’re translating here means “play a big, resonant chord” then the strummed guitar chord makes sense, though it’s interesting that Llobet doesn’t do this. It’s worth playing this chord several ways — 4 notes, 6 notes strummed, 6 notes together with p (④ — ⑥) i m a, to see if there’s another way to translate this.
• Triplets — Llobet has recast the triplets, changing the order of their constituent notes. This transformation preserves some part of the meaning, for sure: the harmonies and rhythms are unchanged, the first note of each triplet is implicitly accented (via the double stem), and presumably for Llobet this was in some sense enough, this was what the passage meant to his ears. But Albéniz notates this section quite clearly, with the inner part of the harmony taking the first position in the triplet, implicitly emphasised through playing with the left hand (M.I.). Losing this loses something fundamental to the meaning of the music at this point, so I contend that this is an uncharacteristically poor piece of arranging on Llobet’s part. Most guitarists nowadays have come round to playing the triplets in Albéniz’s original ordering, though it’s still easy to find recent performances (on youtube, for example) that play either Llobet’s version or one which reverses the last two notes (p m i rather than p i m)

Here’s another decision. In the original, m.30–34:

It’s possible, though awkward, to reproduce the triplet inner texture as written. Here’s my own alternative:

The meaning preserving choices I’ve made here are:

• The rhythm of the triplet configuration is absolutely primary, and cannot be lost without damaging the piece
• First note of each triplet is as in the original, important given the echo of these in the following measure
• Ability to sustain the the melody note above, and the low F beneath the figuration

What I’ve consciously lost:

• The contrary motion in the inner figuration — the pattern in the RH of the piano original moves down while the LH moves up, in both m.30 and 32. This is characteristic of the texture throughout the section, and it pained me to let it go, but playability and the sustaining of the melody and bass trumped the inner voice
• The b-flat in the RH in m.31 — playable, but with the loss of the bass F
• Here (and elsewhere) the notational idiom in the original, where the triplet notes are not beamed, but grouped with a slur and the triplet indication. This pattern Albéniz uses consistently in the piece, with the intention of distinguishing the distribution between LH (first note) and RH in the figuration, and of implying a non-legato articulation. While I’m usually strict about preserving notational aspects of an original in an arrangement (see also the later discussion of the Prelude — Asturias: a composer’s notational choices are rarely completely arbitrary), in this case the compression of the music onto a single stave makes keeping these notes unbeamed visually problematic, and unnecessary from a performer’s standpoint on guitar.

Granada

Here’s a prime example of the inertia of a familiar arrangement. Albeniz’s piano original opens its meno mosso middle section like this:

Note the harmony in the 6th measure of this example. Francisco Tarrega set the ball rolling with arranging this piece: his middle section opens thus:

Fortunately, no recording or performance I’ve heard reproduces the first couple of measures here, which compress Albéniz’s first four, and messes with the rhythm to boot. The glissandi are pure 19th-century, and make of the original something idiomatic for the guitar of the period, but which moves away from transcribing to something like recomposing. The gentle barcarolle ostinato in the piano LH is absent.

Segovia, in his version, restores the missing bars at the beginning of the section, and the barcarolle rhythm for these four measures, at least. This arrangement is still regularly played, and as we’ll see continues to influence other arrangers.

In the 6th measure, we get a nod towards the barcarolle rhythm, but with a change of the harmony — a simple dominant seventh acquires a flat ninth, lending this version a dark colour missing in the original’s harmony. Other liberties are taken with the left hand figuration in measures 7 and 9, preserving rhythm but losing the motivic consistency of the original.

Here’s Manuel Barrueco’s version of the same passage (we pick this up in the 3rd measure of the meno mosso, the third and fourth measures repeat the first two, as in the original):

Here we see the extraordinary persistence of Segovia’s arrangement: while the indignities visited on the fourth measure have been effaced, the change in harmony beneath the second measure of the melody is still there — that annoying c-natural! — and the changed inner parts in the measures that follow (although in a nod to the original pattern in the LH the order of the notes on the 2nd and 3rd eighth-notes is reversed). You might conclude that Barrueco’s version here is an arrangement of Segovia’s transcription as much as of Albéniz’s work.

Another set of choices is possible here. For example, you might (as I have done) decide to avoid changing the harmonies, and to keep the barcarolle pattern to the maximum extent possible:

We’ve gained fidelity to the original, but what is lost here? Firstly, given the compression required by the smaller range of the guitar, the barcarolle pattern crosses the melody at times, so we lose the separation in register of the melody and accompaniment. In performance it’s important to keep the melody to the fore, but the persistence of the accompaniment creates its own continuity in a way that feels close to the piano version. What’s more, I rather like the sound of the delicate interweaving of the melody and accompaniment in (for example) m.50–52.

Secondly, the implied sustain of the low E in the pattern is lost when the B on the second eighth note is taken on ⑥ (m.47, 55). This is a straight trade: I’d rather keep the pattern going as written than sustain the E, but of course another arranger is at liberty to decide otherwise. (A player with larger hands might well find it possible to take the option of playing the B on ⑤.)

Preludio (Asturias)

This was going to be a long commentary on the piece and the distortions imposed on it in its long tradition of performance on the guitar, but Stanley Yates gets there first in a long and insightful essay on his re-thinking of the piece for the guitar. Once again, the benefit of going back to the original (or an original, see below) version of the piece is in peeling back years of familiarisation to approach the piece with new ears and a new sense of what’s possible.

Of Yates’ many recoveries, the most satisfying to me is avoiding the triplets in the outer sections. The piano original at this point intensifies the texture by adding an octave to the pedal-point dominant:

Segovia’s and subsequent arrangers’ choice here is to finesse this textural intensification into a rhythmic one, but Yates makes this observation with respect to the original:

Albéniz’s original keeps to sixteenth-note rhythm throughout the Allegro section. The effect is one of a gradually accumulating insistency as the opening melodic motive is repeated over and over again, supported by the relentless, unchanging rhythm of the upper pedal. (In other words, I felt no need to incorporate the triplet figuration of Segovia!)

The texture is, essentially, perfectly playable as written, and taking this choice remains faithful to an aspect of the music (the persistent 16th-notes) that for me (and clearly for Yates, and clearly not for Segovia) must be preserved. If you’ve only ever heard a standard guitar version of this piece, discovering this in the original can feel like a real shock.

A couple of points remain, however. I prefer the last version below of the final flourish in the opening section to Yates’ alternative: it keeps the original’s relationship of chord notes and beats intact, and arrives in the same way as Albéniz at the final high dominant chord:

Albéniz

Yates

Harvey

With respect to the dramatic chords of the opening section: Yates is right to ask them to be damped immediately: the rhythmic effect of this damping is to articulate what becomes the following sixteenth-note rest as an element of the rhythmic continuity, which to my ears almost (but not quite) makes up for the gap in the texture. The unidentified editor of the 1918 Hofmeister edition of the Suite Espagnole had the same thought: in the early Spanish editions (this one is the 1922 reprint of the 1900 Casa Dotesio edition of Op 232) these chords:

change to the following, with the right hand part delayed:

This simplifies things for the pianist, and makes it easier to maintain the flow of 16th-notes within a stable tempo. (Note too that the Ped indication in the Hofmeister edition is missing in the early Spanish edition. Just because an edition is old doesn’t make it right.)

For me this suggests a possible alternative to Yates (and others) pacing of the equivalent bars in the piece, in which one plays the chords twice, in the 16th-note rhythm, with a rapid down-up stroke:

Again, to think about choices, what I’ve done here preserves something that’s missing in the other arrangements (an attack on each 16th-note pulse, to avoid a ‘hole’ in the flow) at a cost (duplicating the chord in a way that’s absent in the original, although it preserves something of the feel of the unknown Leipzig editor’s version).

Finally, a comment on the appearance of Yates’ edition. Given some of the work I’ve been involved with in notation technology, you won’t be surprised to find me picky about the visual meaning of the notational choices composers, engravers and editors make. There’s a whole other article to write on engraving guitar music beautifully, but I can’t help thinking that, apart from anything else, the look on the page of Yates’ arrangement of the piece loses much of the meaning communicated in a direct and visual way by the piano edition — Albéniz would not for an instant have conceived of writing the piece out this way, the visual language of this engraving (as in the opening section’s closing flourish shown previously, or at the start of the piece as below) speaks the language of Berio and Stockhausen, not that of the late 19th century:

Albéniz

Yates

The piano version has two staves to pay with, and the cross-staff beaming idiom to represent both the flow and the physical distinction of left and right hands is of course common. The visual points here are firstly the clear sense that the pulse is moving in 16th-notes, which Yates’ polyphonic separation confuses: secondly, the beaming of the original affects the balance of light and darkness on the page, which is one of the subliminal triggers that makes a score easier or harder to read. Preserving this in an idiomatic setting for guitar leads me to something like this:

The double-stem notational idiom is familiar to guitarists, does the duty of the cross-staff beaming in the original, and preserves both the connection of and distinction between the moving line and the pedal note. Additionally, there seems to me to be no reason not to include the original’s distinctive articulation for the first LH note of each measure (the marcato indication doesn’t, after all, need to be religiously interpreted as a combination of an accent and a staccato, and Yates’ collision of the dot and the accent is visually graceless).

Envoi

Guitarists still learn and play old arrangements, without thinking or questioning the choices made by the original arranger, and as we’ve seen in the case of Barrueco’s Granada, continue to base new transcriptions on the compromises made by previous generations of arrangers. Thirty years ago one might have accepted this state of affairs: copies of the originals would have had to be sought out in libraries, or ordered and bought: even then, for much of this repertoire there’s still confusion around editions and their accuracy. Now, resources such as IMSLP, Spotify and YouTube give us instant access to these materials, and to a wide range of performances.

One goal of an arrangement (or of one kind of arrangement — others exist) is to produce a natural and vivid musical experience for performer and audience, but we almost always have multiple choices to make when we approach the task of reinventing a work in an alternative medium. This is particularly intriguing in the case of Albéniz, whose piano music characteristically reimagines the textures and articulations of the guitar on piano: fidelity to an original in this case might sometimes seem like counterproductive, where a less strict alternative might in some senses be experienced as more ‘guitaristic’ than a version closer to the original text. Nevertheless, when we arrange we apply numerous values and principles, with dimensions to each: we can be more or less faithful to the original, more or less accommodating some notion of suitability to the instrument, or to playability. If we’re aware of these we can make these choices consciously and consistently, and consequently explore the possible spaces around the original in which different kinds or qualities of arrangement can exist. This for me is Hofstadter’s lesson in Le Ton Beau de Marot: here we’ve been concerned with versions of works which largely preserve faithfulness to their originals, but there are of course many other ways of reimagining the musical substance of a work.