A tale of two editions

So recently (well, not so recently) I was grumbling about the quality of a recent edition of a major guitar work, rediscovered in manuscript and published in, frankly, a shocking and slapdash form. One thing I really wanted to see was the composer’s manuscript of the work, which in the absence of commentary on the edition (or even if there were a critical apparatus) conveys so much of value to a performer.

With Angel Gilardino’s Bèrben edition of The Segovia Archive we at least have these. This is a groundbreaking series, showing us both well-known works in their early stages and also works which their dedicatee, for whatever reason, chose not to publish. I can’t imagine programming any of these pieces without having these at hand, and as I’m about to revisit in a couple of recitals Federico Mompou’s wonderful Suite Compostelana of 1962, thought it worth sharing some thoughts on the manuscript, Segovia’s original edition (published by Salabert in 1964) and the version prepared by Giladino and Luigi Biscaldi which accompanies the facsimile.

The first thing to note is just how close Segovia’s edition [S] is to the manuscript [M]. Looking at other originals of works prepared by Segovia for publication there are significantly fewer alterations, in more or less all respects. In comparison with (say) the Thème variée et finale of Ponce (as edited by Tillman Hopstock and published by Schott), or the Joan Manén Fantasia-Sonata (also in the Bèrben series), [M] and [S] show few differences, even down to voicing of chords. There is, of course, no way of knowing exactly what the state of composition Mompou’s manuscript represents — it has all the appearance of being a fair, final copy — nor without more commentary than is provided in the edition can one be sure that this state wasn’t arrived at after many sketches, interactions and conversations. Gilardino points out in the introduction

…correspondence between the two musicians…shows beyond any doubt that Mompou granted Segovia total freedom to adapt the work (p.6)

I like to think that Mompou’s refined and sparing musical style meant that the work as we have it represents something very close to his original conception.

The significant textual differences between [S] and [M] can be summarised as follows

• Music removed (ex)
• Rhythmic changes (rh)
• Pitch changes — Octave displacements (po)
• Pitch changes — Chord voicings (pv)
• Pitch changes — Obvious wrong notes in [S] (pw)
• Pitch changes — other… (pz)
• Phrasing (fr)
• Dynamic (dy)
• Articulations (ar)
• Tempo (tm)

Here’s a detailed look at the Preludio, offered here both as an example of what I’d expect a decent editor to document when preparing or comparing editions, and as a useful overview for anyone performing the piece — there are always decisions to be made, particularly in the case of odd or unclear readings or suspicious and contradictory notations, so the more data you have the better. Measure numbers are measures in [S] (as this is the edition a guitarist is most likely to have), c4 is the c notated on the first ledger line below the staff (so in standard tuning, open strings are e3, a3, d4, g4, b4, e5). q3 means 3rd quarter-note in measure, 9s the ninth sixteenth note, and so on.

Preludio

m.1–15 (fr) [M] has phrasing marks throughout, initially measure-by-measure (each phrased into the first note of the following bar) then shortening as the cadence to m.11 approaches. These expressive indications should be taken seriously — many recordings just treat this as an undivided sequence of notes, or worse still as a kind of toccata. In general in this movement Segovia or his engraver has omitted phrasing marks: additionally, none of the note-to-note slur indications in [S] are present in [M]

m.5 (fr) Amusingly [S] has a vestige of one of the phrasing marks in the manuscript leading into the first note, f#5, of the system.

m.7–9 (ar) [M] has tenuto articulations g5-f#5, e5-d5.

m.10 (rh) [M] sticks to the previous measure’s 9/8 time signature, a3-b3 dotted half/dotted quarter respectively.

m.14–17 (dy) [S] has decrescendo from m.14 to mf in m.15, then p in m.17. [M] starts the decrescendo at the very end of m.15 to the mid-point of m.17, with no dynamic indications.

m.15–16 (pv) [M] simply has the rhythm of these measures played on e4 throughout, [S] adds the dotted quarter e3 to each beat.

m.17 (pv) [M] has e5/e4, [S] e4/e3.

m.15–16 (ar) [M] has strong accents (‘upside-down v’) on the first beat of each measure. [S] has a standard accent on the first 16th of m.15, nothing in m.16 (apart from a spurious and incorrect string indication). Mompou uses the strong accent symbol throughout [M], [S] mixes both, with a tendency to use standard rather than strong accent symbols.

m.18 (ar) [S] has harmonics on q1–4.

m.18–19 (dy) [M] has crescendo on m.18 q1–4, decrescendo m.19 q1–4.

m.22 (dy) [M] has mf, no dynamic in [S].

m.24–25 (pv/ar) e3 on q4 of each measure added in [S], strong accents on first beat of each measure in [M].

m.26 (rh) e4 is half-note in [M]

m.26–29 (ex) [M] repeats m 24–25 between the two halves of this phrase (inserted at m.27).

m.27 (pv) [M] has f5 stemmed above the e4 on the first beat of the measure.

m.28 (pv) [M] ties e5 of q3 to a quarter note on the next beat.

m.34 (pv) [M] repeats b#4 on q3, ties it to a quarter note on the next beat.

m.28 (pv) [M] stems the a4 in the chord to the lower two notes, with a duration of a half-note. The lowest note b#3 of the chord has a dot in [M].

m.37–40 (dy) In [M] it is not particularly clear which system owns the extended crescendo, but the dim which follows in ([S])m.45 clearly belongs to the upper stave in [M], so the crescendo also. In [M] it starts at s4 and extends to m.42: in [S] the start is delayed to s8 and is written to end at the end of m.40.

m.43–47 (po) [M] lower part has e5, e4 | e4, e3 | d5, d4 | d4 in the bass, with a crossed-out e3 dotted-half-note at the beginning of m.47. [S] has a single dotted-half d4 at the beginning of m.45, I suggest that another d4 should appear at q4 (corresponding with the rhythm of [M] and the octave displacements in the previous bars in [S]). There is a space in the bar at this point rather than a notated rest: the omission of this d4 therefore seems to me more likely to be an engraving accident than an editorial change.

m.46 (pz) [S] places the dotted-half-note d4 in parentheses. I’d always wondered about this — did the composer really mean this note to be optional? Looking at [M], however, one quickly realises that Mompou’s convention for notes which take a cautionary accidental is to place the accidental and the note in parentheses. In general, [M] changes these to conform to regular practice (see m.13, m.35): this is just one that was missed.

m.47 (pw) In [M] s1 is clearly c#5 (which also maintains the sequence of the previous bars) and not b#4. s8 is d4 and not e4 ([S] even fingers this correctly as an open string).

m.55–57 (ar) Tenuto marks in [M] c6|b5 and a5|g5 (corresponding to the equivalent passage in m.7–9).

m.58 (rh) [M] stays in 9/8, d4 is dotted half-note, e4 dotted quarter.

m.59 (tm) [S] is missing a Tempo indication, present in [M].

m.59–60 (po) a3 in [S] in the lower part are both a4 in [M].

m.63–64 (rh) In [M] the a3 dotted-half-notes at the start of m.63 is tied to that in m.62.

m.68 (pv) [M] has an unplayable 6-note chord: from the top g5/b4/f4/d4/a3/e3. [S] omits the low e3 and revoices the chord as g5/d5/b4/f4/a3.

m.68–69 (ar) [M] has arpeggio lines before the two final chords, missing in [S].

The Gilardino Edition

As you would expect, the general content of Angelo Gilardino’s edition [G] follows [M] more closely than does [S]. It’s important to keep in mind that it’s goal is to provide

…a working edition that is conceived not as an individually elaborated text for a personal use of the editor (as in the case of Segovia’s edition) but as a text ready for performance which lies as close as possible to the original (p.5)

This is a laudable aim — I’m a fan of what you might call “performance urtexts” and have produced one or two myself — but I find myself wondering whether this has been achieved in this case, or alternatively whether once again an opportunity has been missed. I have concerns in several areas, some clearly more subjective than others, but to summarise:

Arbitrary editorial decisions: [G] shows many cases where a notation is changed for no apparent reason. A few examples (not exhaustive: measure numbers refer to [G]):

Preludio, m.21: an arbitrary revoicing that changes the harmonic and melodic meaning of the measure.

Coral, m.10 and elsewhere, the original’s notation is changed unnecessarily from 2 to 3 parts, which also clutters the appearance of the score.

Coral, m.10–16, this section indicated with repeat marks in [M], but written out here.

Cancion, throughout: [M]’s expressive original notation (the upper interval chords are notated as quarter-notes, with l.v. ties indicating sustain) is replaced prosaically with half-notes.

No indication of editorial changes: With a facsimile accompanying the score, all it would take would be a light (*) or similar to indicate a divergence — there aren’t, after all, that many. An interested player would then be in a position to check against the facsimile.

Differences between [G] and [S] not highlighted: although the text claims to indicate all extra measures removed by Segovia in his edition, this is not in fact the case. The only measures so indicated in [G] are those which are explicitly struck out in [M] — other omissions (for example, the interpolation removed in the Preludio m.26–29 of [S], and some repeated cadential chords in the Coral) are not identified.

Incorrect engraving: while some of this is indeed subjective, there are some well-established practices which [G] regularly flouts (in general, [M] and [S] both reflect standard practice and are reasonably consistent). Some examples

Preludio, m.1–5: phrase marks on the wrong side of the notes.

Preludio, m.9–11 and elsewhere in the suite: tempo indications below the stave and associated with dynamics/expression.

Recitativo, m.49 incorrectly positioned fermata (below the note: in [M] it’s above the centre of the measure. Once again, there’s surely no reason to change this…)

Metronome marks, throughout — these are set with the rhythmic value in full-size rather than small notes/noteheads.

Obtrusive fingering: partly to do with the very dense black used for the print (see below), but the fingering (LH and string numbers) is in a heavy font which could do with either being made un-bold, or brought down a point size or so, or both. I find much of the fingering itself frankly bizarre, but that’s another article…

The look/production of the edition: to my eye the edition is disturbingly unattractive. Part of this is down to the engraving issues identified above (like bad kerning, once you’re aware of this it’s impossible to un-see), but in general:

• There is too much whitespace on the page. [S] presents the suite in 12 pages, [G] takes 20. The volume is significantly larger than A4 size, with a page that’s as tall as [S] though slightly narrower, but is set throughout with six, sometimes five systems per page (where [S] has eight or seven). The staves can be smaller, which will improve white/black balance and ease page turning.
• The edition’s white paper is too white, and the black ink too black. No, seriously: next time you look at a decent piece of published music by an established publisher, try to notice the colour of the page (it’s never bright white, in some cases it’s visibly ivory or cream-coloured) and the ink (in general, printers of text, as well as music, never use completely black ink: it’s always a few shades towards grey, The best internet typography, like here on Medium, follows this convention). The paper is also glossy, which I assume is to support printing the facsimile, but which is bad to read from.

Implications for performance

I have a couple of recitals coming up with the suite on the program. Having worked through all three sources, this is my approach:

Incorporate [M] phrasing, dynamics, articulation on the basis that these originate with the composer.

Correct wrong notes and omissions — specifically, in the Preludio as indicated, in Cuna (in the middle section, m.41, where a chord with c5 and f5 in [S] is missing the accidentals for both pitches found in [M]), and in the Recitativo, where [M] features a correction from ab3 to c4 with the word “do” (i.e. “c”) at m.25. (Interestingly, in his recording of the Suite Segovia plays the ‘wrong notes’ in Cuna m.41 — whether this was an editorial change that he made, with or without Mompou’s approval, or whether he played from a manuscript or the printed copy that accidentally missed out the the accidentals, is a matter for conjecture. The recording is of course yet another source for consideration, but that would be a whole new article…).

And that’s all. Essentially, this means I play what I consider to be a ‘corrected’ version of [S]. Others might take a different approach, of course, but on balance I don’t think I can recommend the Gilardino edition as a performing score: if I were planning to play a ‘composer’s cut’ I would without a doubt make my own clean version from the manuscript.

The issues with engraving standards are increasingly important to me. Notation software has put printed music production in the hands of many more people than previously, and while some software is better than others at automatically producing an acceptable standard of engraving none is without its issues, and all produce work that, without some level of care and adjustment, falls significantly short of even run-of-the-mill engraving of the pre-computer era. If we don’t pay attention to these things, we’ll end up with a generation of musicians who don’t know what good engraving looks like, we’ll be condemned to editions that — maybe without us being able to say why — we find awkward and difficult to play from, and as a culture we’ll have lost all the design skills of the engravers craft.