This passage in Giulio Regondi’s Reverie, Op.19 is a sort of music-analysis puzzle:
Giulio Regondi: Reverie Op.19, mm. 88-90
There must be a misprint somewhere in the group marked with X. As it is printed, it hardly constitutes a valid harmonic progression in the practice of the time. We can have certainty that the last chord of the group is D7, which serves as dominant to the G-major harmony in the next bar. Then, we can safely guess that the C’s of the previous two chords are also natural (a D7 – C – D7 progression). These two chords could work with C# as well (F#m-A7-D7 progression), but I think it is more likely that they have C’s, and virtually every guitarist plays C’s. The C# alternative comes from the C# in the first half of the bar; at some point the pitch C# has to transition to C within the bar, but we cannot say exactly when, since the natural sign is missing.
The first three chords of the group are the puzzling ones. The first chord, C#-minor, can work as such (notice that I assume the top note to be C#, for there was a C# in the same bar). The second chord, however, is not a valid one here. It has the structure of a dominant-seventh chord, but it is neither spelled nor resolves as such. The spelling is that of an augmented-sixth chord, but then it should resolve to a D or G chord, which again is not the case. Furthermore, the chord leaps to an A-minor triad, which makes the passage even more clashing as it is written. Considering the C# variable, this third triad could be A-major, and this possibility will be briefly considered below. Again, I have only heard C and not C# for that and the successive triads.
So, there is a misprint somewhere. It can be an accidental missing or misplaced, a note misplaced, or a combination of them. So how do we “correct” this group?
Let’s consider some possibilities. Perhaps the sharp and natural signs of the G’s are misplaced and actually belong to the C’s. There are a few points that support this: (1) the C#-C descent must happen at some point and it is not marked; (2) as written, the first and second triads repeat the C#, which is strange within a descending group, especially considering that this descending gesture occurs in other parts of the work and never features repeating melodic notes; (3) the G#-G motion of the middle voice is actually awkward in the context of the other voices. So perhaps what Regondi intended is:
This progression works (can be an acceptable harmonic progression within the style), but the parallel C- and A-minor triads do not sound too satisfactory, at least to my ears. Many guitarists play a similar version, with the original motion of G# to G, which renders three parallel minor triads:
Though certainly left-hand-friendly, it is doubtful that Regondi intended such progression.
Many guitarists–I might dare say most–play a progression that sounds well, but looks completely different than the printed edition. (This is the version David Russell plays and that is probably why it became the most popular one):
This version sounds nicely, but changes many things. Since it only needs a natural sign on the C, it assumes that all other accidentals on the score are misprints. It also changes the notes of the second chord. Therefore, we can hardly consider it to be Regondi’s intended progression. Truth is, if we take the liberty to remove or change so many signs from the score, then we can build other versions, some of which sound well and appear more similar to the original. For example:
Or one with accidentals that resemble a bit more those in the original:
There are a few other options that would work. For example, in both of these cases, the first chord can have A instead of G (A-minor and A-diminished triads respectively), and/or the second chord can substitute the G# for F.
In the last three examples, we have considered the possibility that the first chord of the group already contains C instead of C#. We should note that this means that the other two notes of the chord cannot be E and G# as printed. They form an augmented triad that is unprepared and hardly justified by the voice-leading or as some sort of passing sonority. We can alternatively assume that both first and third chords have C#, and then a few possibilities open for the second chord, albeit hardly resembling the original. This one, for example, works nicely and foreshadows the sequences that start at measure 97:
But we might have delved into too many changes; let’s consider the original group again, and try to move from C# to C in the first two chords. In “Giulio Regondi: The Complete Works for Guitar” (Monaco, Chanterelle ECH415), Simon Wynberg mentions a piano transcription by Frederic Alquen, published in 1871, that has this version:
Again, I am not convinced by the sonority of the two neighbor minor chords. (Nothing wrong with neighbor minor chords–Regondi includes them, for example in measure 101–but for this particular progression I think he would have written something more elegant.) However, this version is particularly interesting in that it maintains the Eb throughout the group. Furthermore, we could assign it some sort of historical validity, for Alquen might have based it on a performance he heard at the time. With all this in mind, the following two versions have the same “amount” of accidentals as the original and create tasteful harmonic progressions:
Now let us go back to the original group. What if the natural sign of the third triad alters not the E, but is actually a courtesy sign for the C? A few points would support this option: First, alterations on the same name note across different registers (same “pitch class”) are specified for each of the registers. Look at measure 88 in the first example. The middle and high C#’s both have sharp signs; we should then expect that both the high- and middle-register C’s in the next bar carry natural signs. Second, this option maintains the Eb throughout the group, like Alquen’s version. Third, it looks closer to the printed score than most others. Voilà:
Although this is the version I play and personally favor, my aim is not to argue for a correct, definite version. By presenting all of these options, I simply hope to help guitarists avoid versions that do not work and/or do not make justice to Regondi’s musical language and ingenuity. Have fun analyzing and experimenting with these and other alternatives, and let me know your thoughts!