Is this really better analyzed in G minor than in Bb?

So I have been analysing Rondo a Capriccio by Beethoven and I reached a point where a so called "expert's" opinion and mine differ drastically. It starts at bar 57 and continues to bar 68. After the last eighth in bar 68, my opinion and their opinion as far as the analysis are the same. First I will show both my opinion and the expert's opinion as far as analysis.

Here is mine:

As you can see, my opinion as far as analysis is concerned has 2 harmonic areas, Bb and Gm. Now here is the expert's opinion about these same measures:

As you can see, the so called expert thinks that the entire passage is in Gm. It makes for there being fewer modulations, but what previously made total sense is nonsense now. Like the vi ii V I, one of the most common chord progressions becoming i iv VII III, which is just total nonsense, even in a minor key. Also, the resolution to III diminishes the sense that you are in a minor tonality to begin with.

My way of analysing it avoids this problem by having the section from bars 57-68 be in Bb until it reaches the last eighth note of bar 68, at which point it pivots to Gm and it only makes sense in Gm.

Here is my reasoning behind analysing bars 57-68 in Bb instead of Gm:

Yes, you would expect to hear Bb major chords in a G minor piece. But in a G minor piece, if there is a resolution to III, that typically is a modulation to Bb and thus is truly a resolution to I, not III. VI and VII on the other hand can be resolved to without a sense of modulation in a minor key, just like how in a major key, you can resolve to vi without actually modulating to vi. So it just makes more sense to me for bars 57-68 to be in Bb, regardless of whether that means there is a modulation later on or not. And besides, Rondo a Capriccio has frequent modulations anyway.

Also, typically, in a minor key, III does not get inverted whatsoever. It usually is in root position. Here, if the whole passage is analyzed as being in Gm, you get all 3 inversions. There are only 2 chords(3 if you extend it to seventh chords) that typically get all the inversions. Those being I and V in major and i and V in minor(and also vii°7 if you extend it to seventh chords). Other chords like IV or vi typically only get 2 of the 3 inversions and a few others like III typically aren't inverted at all. So to see all 3 inversions of what is supposedly III in Gm makes me doubt that it is in Gm at all and instead think that it is in Bb.

So am I right in saying that it starts in Bb and then pivots to Gm? Or is the expert right in saying that the entire passage from bar 57 is in Gm? Which one makes more sense harmonically speaking, my analysis or the expert's analysis?

• In Bb I’d say the second half of bar 63 is V7 because of the Eb, which further strengthens the Bb major feel. Aug 16 '19 at 21:27
• The i iv VII III chord progression is not total nonsense in a minor key. The VII-III chord progression is very common in classical music from my findings, and I used it effectively myself in one of my earlier classical music compositions. Stringing it together with i-iv is not implausible at all. In fact, i-iv-VII-III makes a very powerful sequence. Aug 17 '19 at 0:08
• May I ask who the expert is?
– Richard
Aug 17 '19 at 1:49
• @Richard Teodor Lontos on Musescore.com is who I'm referring to. He commented on my harmonic analysis score there that bars 57-74 should all be analyzed in G minor instead of part of it being analyzed in Bb. Here is the link so you can see both his comment and my response to it: musescore.com/user/50070/scores/5667301 Aug 17 '19 at 1:58
• This question highlights a limitation of functional harmonic analysis. It's like listening to a piece in triple time and then arguing whether it is in 3/4 meter or 3/8 meter: it's more about the conventions of analysis and notation than it is about the music itself. It's a piece in G minor that spends some time in the tonal center of the mediant. Whether a chord is labeled VII or V/III or is analysed in the context of B flat make instead of G minor is not of great concern. Aug 17 '19 at 8:55

Using the word "key" is too definite. The music progresses through lots of different tonal centers but most of them don't last long enough to be worth calling "keys".

Both of you are wrong about the whole passage, but the "expert" is more wrong.

The point which you both missed is the sequence in bars 66-70 with a V-I cadences in Bb major, C minor, and D major - which of course is the dominant of G, so we arrive back home.

Equally, trying to pretend that a G minor chord after a full cadence in G major is in any tonal center other than G is just ignoring what the music sounds like.

So, somewhere between bars 57 and 65 (and don't forget the passage is repeated) the tonal center shifts from G minor to B flat major. Exactly where does that happen? Frankly, who cares! If you like, you can claim you are both right, and the first time through the repeat is in G minor, but the second time in is B flat major. There is no rule that music has to have a unique, unambiguous analysis according to the "rules" in some textbook, and most good music is ambiguous to greater or lesser degree. The chord in "bar 57 following bar 56" doesn't necessarily have the same harmonic function as the chord in "bar 57 following bar 64" even though the notes are the same.

Incidentally, claiming that bar 75 is anything except a fresh start in Eb major is just as nonsensical as the debate in the OP. Once again, stop looking at the score, and start listening!

I suggest that it's ridiculous to analyse 57 as anything but i in G minor, but equally ridiculous to analyse the cadence at 64/65 as anything other than V-I in B♭ major. So have both. G minor at 57 with an immediate modulation to B♭ major. After that, at 64, it's a bit more up for grabs. We know we're in G minor by bar 72. What tonal centre do you prefer at bar 65? A new one every couple of bars or not? Perhaps Roman numerals aren't all that useful in this section!

Look at the whole piece. Up to bar 57 we have been firmly in G major, key signature one sharp, and apart from an occasional 'V of V' completely diatonic. The change at 57 is clearly to i of G minor (not vi of B♭ major), and your 'expert' has preferred to treat the entire section up to the return of G major as based on that key. I see his point, from a structural viewpoint, though I think it results, when analysing the detail, in Roman numbers that aren't always very helpful.

https://musescore.com/classicman/scores/201476

• but i iv VII III, which would be like vi ii V I but in minor, makes no harmonic sense because VII isn't a dominant function chord in minor. And the resolution to III and III showing up in all 3 inversions both weaken the minor tonality. The presence of III, III6, and III64 in particular makes it sound like the true tonic is actually Bb for that entire section, at least until bar 68 where afterwards, it only makes sense in G minor. If bars 57-68 are analysed in Bb major, these problems don't exist and it makes much more harmonic sense. Aug 16 '19 at 20:43
• I was actually largely agreeing with you! I've edited my answer to hopefully make it clearer. Aug 16 '19 at 20:54
• @Caters If you want to say it’s in Bb right at 57 then it’s a “crash and go” modulation, which seems reasonable for Beethoven, but at least as reasonable is a modulation to Bb over a few bars (57-60) via a smoother modulation to the parallel minor. In the end, as long as you see what’s going on I think you can be comfortable with your analysis. There’s not necessarily one right answer for this kind of thing. I agree with this answer that the cadence at 63-64/65 is very strong in Bb and also once you get the F#s in 70 that really makes it sound like G Minor. Aug 16 '19 at 21:34

Look at how the excerpt falls into sections/phrases. Harmony often isn’t static, each section in a single key: it moves, on various timescales, through phrases and sections of a piece.

The first 8 bars of the excerpt, bars 57–64 (plus repeat), are most naturally read as starting in g minor, and ending in Bb major. Then the next 8 bars, 66–75, go from Bb major back to g minor. So when seen together as a 16-bar section (or 32-bar including repeats), it goes (g minor) –> (Bb major) –> (g minor). The g minor is the key this section starts from and returns to; so it’s the key that makes the most impression on the ear, and this is why it would generally be analysed as the main key of the section, although it’s certainly true that in the middle, the section moves to a (shorter-timescale) tonal centre of Bb major.

Going back to the first phrase now, to see why it would usually be analysed as “starting in g minor, and ending in Bb major”, rather than as “in Bb major, starting on the 6th”: When you first hear the g minor beginning, there’s been nothing before it to suggest Bb major; it and the next bar clearly set up the idea of g minor. Unless something later gives a strong reason to view things otherwise, the natural way to analyse this is that the phrase starts in g minor. At this point, the listener has no reason to hear it as anything else. Now, as you say, there is some reason to consider the alternative view, since the phrase immediately moves towards a clear cadence in Bb major. However, it doesn’t stay there — as quickly as it has moved to Bb major, it then moves back in the next phrase to an equally clear cadence in g minor. So the Bb major isn’t established for any longer than the g minor was; so it doesn’t really give a compelling reason to reanalyse the start of the phrase in hindsight.

An example of the contrary, where the start of a phrase really does deserve reanalysis based on what comes after it, would be the Rondo of Beethoven’s 4th Piano Concerto in G. The first couple of bars seem to establish C major; but the rest of the theme, and the continuation, establish G major more strongly and for longer. With so much surrounding context in G, it becomes natural to hear the opening bars as really being the subdominant of G, rather than establishing C as a tonic as they seem to on first hearing. However, in your excerpt, there isn’t anything like so much surrounding context to establish Bb major as a more significant tonal centre than the initial g minor.