Let’s argue about this one weird chord in the Brahms Intermezzo in B-flat minor

I have some aspirations in music theory pedagogy, and toward that end, I’m learning more about Schenkerian analysis. If I’m going to resist it, I should at least be conversant in the thing I’m resisting, right? So I’ve been reading Schenkerian Analysis: Perspectives on Phrase Rhythm, Motive and Form by David Beach. One of his examples is the opening nine bars of the Brahms Intermezzo in B-flat minor, Op. 117: No. 2. Listen here:

Brahms isn’t my cup of tea in general, but this piece was immediately ear-grabbing. I put it into Ableton along with the MIDI file so I could listen closely. Then, since I already had it warped out, I did a remix. Van Cliburn plays beautifully, but as a stylistic thing, all the rubato gets on my nerves. I’d rather hear this music on the grid.

Below, you can see Beach’s Roman numeral analysis of the first nine measures of the Intermezzo. It’s the first 34 seconds of the Van Cliburn recording, or the first 24 seconds of my remix.

If you don’t read figured bass notation, here’s Beach’s analysis as a chord chart:

| Bbm/Db | Bbm7 | Ebm7 Ab7 |
| Dbmaj7  Gbmaj7 | Cm7b5 | F7 |
| F7 | F7 | Bbm/Db |

This all makes perfect sense to me until bars eight and nine. Beach says they continue the F7 from bar seven, but I hear an unambiguous Gbm6 chord. The first three notes of bar eight outline a literal Gb minor triad, which sustains all the way through that bar and most of the next. That chord is the most striking harmonic event in the whole piece, because it’s so conspicuously outside the key of Bb minor. I mentally refer to this particular thing as the Darth Vader chord because it’s most familiar to me from John Williams’ “Imperial March.” How could anyone mistake it for boring old V7?

Beach didn’t label those two bars as V7 by accident, though. In the text, he goes to some length to explain why.

The addition of the seventh (Ab5) to the tonic harmony in measure 2 introduces the upper neighbor of F5—namely, Gb5. However, rather than return to a stable F5, the sequence continues beyond this point to the Eb5 on the downbeat of measure 5. Thus this Gb5 is not marked as a neighbor note, but rather as the initiating point of a descending third leading to the Eb5, which is notated by a slur. The arrival at this point coincides with the initial statement of the descending third Eb5-Db5-C5, articulated by a voice exchange within the ii7 harmony, which is subsequently repeated, now supported by the dominant. Our expectation, of course, is that this Eb5, the dissonant seventh of the dominant, will resolve to a stable Db. Instead, what follows is an enlargement of the descending third supported by an extended dominant that is transformed into a diminished seventh chord in four-three position leading to a repetition of the opening gesture. It is this manner of return to the opening gesture that leads to the idea of the opening chord being an incomplete diminished seventh chord.

One might be tempted to interpret the harmony in measure 8 as a minor chord built on Gb, thus supplying consonant support for the Db, but the continuation makes it clear that Brahms means the A natural he has written, not Bbb. The Db is not stable; it is an extended passing tone within the restated third Eb5-Db5-C5 (p. 35).

In other words, even though it sounds like a Gbm chord, Beach argues that the chord is actually an Adim7 with its flat fifth on the bottom. In jazz terms, Adim7 is a rootless voicing of F7b9. (Jazz theory would also invite you to think of the Gb minor triad as a rootless voicing of F7alt.) If Brahms had actually meant Gbm, he would have written Bbb rather than A natural, but he spelled it A natural because he was thinking Adim7 and/or F7 the whole time. Also, the bottom note of the whole figure is E-flat, and that fits better as a bass note for Adim7 than for Gbm.

Except that… it is so obviously a Gbm chord! How am I supposed to ignore that? It doesn’t just flit by momentarily, either; it takes up a good long chunk of the passage that Beach is analyzing. Even if Brahms did mean that the Gbm is really an extended voicing of F7, did he really expect us not to hear it as a trip outside the key? As a Schenkerian, Beach wants to explain everything in terms of tonic and dominant, but in doing so, he steamrolls over the most striking and beautiful event in the whole passage. I’ve heard other Schenkerians interpret interesting chromaticism as “really” being plain vanilla functional harmony. For all I know, they’re “correct” in their understanding of the composers’ intentions. But Beach’s analysis defies common sense. Besides, if we’re going to try to read the composer’s mind, I would point out there’s no F anywhere in bars 8 and 9, and given how many other notes the line sweeps through, it kind of seems like Brahms was avoiding F deliberately.

Here’s how I analyze the chords.

I don’t know why the Ebm6 chords don’t merit Roman numerals in Beach’s analysis, they feel significantly present to me. I hear the Ebm7 and Ab7 in bar three as ii-V in the key of Db, and the Cm7b5 and F7 in bars four and five as returning us to Bb minor. That’s not meaningfully different from Beach’s analysis, but I would consider Dbmaj7 to be a local I chord, not bIII in Bb minor. More to the point, I hear the Gbm6 as being a resolution, rather than a continuation of the F7’s tension. It’s like a deceptive cadence. The top note of the chord in bar eight is Db, which you’d expect to hear in the tonic chord in Bb minor. But the rest of the chord is an entire major third below what you’re expecting.

The melody also supports a hearing of Gbm6 as a side trip into a new universe. For the first seven bars, the melody has all been short arpeggiated figures, just three or four notes long each. But in bars eight and nine, Brahms suddenly unleashes a stream of nineteen notes that sweep down and up across multiple octaves. Then in bar nine he’s back to calm little arpeggios. Also, he brings you back home to Bb minor via the same little Ebm6 arpeggio from the very beginning–that chord is a neat parallel to Gbm6.

The debates in the music theory literature between Schenkerians and non-Schenkerians seem to mostly boil down to one question: how much should we be thinking about the horizontal axis compared to the vertical one? Schenkerians lean hard toward the horizontal view. Whatever might be happening on the surface of the music, the important thing for them is the large-scale structures of tension and resolution, and in the Western classical canon, that means dominant chords and tonic chords. Schenker thought that all music worth analyzing could be explained in terms of a kind of battle of wills between dominant and tonic, a battle that mirrors the dominance struggles in social hierarchy. My main problem with Schenker is the idea that the only valid music fits the harmonic mold of the Western European art canon, it’s atavistic and racist. But even within the Germanic masterworks themselves, Schenker’s approach is oppressively narrow-minded. It’s interesting to think about the contrapuntal function of notes in a melody, but we don’t need to explain away every deviation from the basic tonic-dominant model.