I saw someone make the vague claim that in Debussy you can often identify passages with predominant function, even if they don't represent chords that are typically considered predominant chords. I realized that, as far as I can tell, there's no single scale degree, pitch class, or interval movement which identifies a chord has having predominant function (unlike dominant function, which is arguably defined by leading tone to tonic or a tritone resolving inwards). Some predominant chords look exactly like standard dominant functions applied to chords which actually have dominant function in the tonal context, like V/V, but other predominant chords do not. In addition, it's even more complicated in instances like Debussy where there are totally nonfunctional passages, in which case obviously not being tonic or dominant does not necessarily imply predominant function. What criteria can I use to identify a chord as having predominant function?
First, I think any attempt to understand functional harmony will be frustrated by using examples from Debussy! I don't mean there isn't functional harmony in Debussy, but a key element of his style was non-functional harmony. Composers like Bach, Handle, Mozart, or Haydn will provide much clearer examples of functional harmony.
Various textbook offer up various harmony progression 'rules' or conventions. But, I think this sums up the basic functional idea...
tonic > (pre-dominant) > dominant > tonic I > (x) > V > I
A pre-dominant chord must be optional or else the definition wouldn't fit any actual music.
I used and 'x' as a symbol for the pre-dominant to show many chords can be used, it isn't realistic to say any chord can be used. For starters it categorically should not be a
I or a
V, because those are the chord for the other two functions.
Now we get to an interesting point. The three function are different in the number of chords that fulfill the function:
- tonic : really just
I. I suppose someone will say
vifor a deceptive cadence, but I would disagree as the whole point is the deception is because of the non-tonic identity of the cadence chord.
- dominant : the
viiochords with leading tones. (Regarding other scale degrees: despite the function name, the
^5scale degree, the dominant, may not appear, like in the case of
viio, except for the dominant, no other tones from the tonic chord should appear, therefore
iiiis not a dominant despite the leading tone being in the chord.)
- pre-dominant : lots of chords...
If saying 'any chord not a tonic or dominant is a pre-dominant' is a problem, then typical textbook harmony would give something like...
- categorically exclude
viio, what remains will be...
ii iio bII Fr+6 iii* bIII** IV iv It+6 Gr+6 v*** vi bVI bVII
Those would be the typical pre-dominant chords. Inversion and seventh chords aren't enumerated.
iii contains the leading tone, but doesn't function as a dominant meaning it doesn't lead to a cadence on
I, normally it would go to
bIII might not lead to
V normally and could be part of a modulation from minor to relative major, or might just lead to
(***) minor dominant doesn't have a leading tone, doesn't form a cadence to
Some don't like prescriptive lists of harmony in which case I suppose the '...any chord...' definition will work.
Pre-dominants make much more sense in a diatonic context. It's tricky with Debussy where sometimes he doesn't have a clear tonic to begin with (i.e. the whole-tone stuff). But roughly speaking, my understanding is that a pre-dominant is a chord that typically precedes the dominant. If someone plays a piece up to that chord and asks folks to sing (arpeggiate) the next chord, they'll sing a dominant. This can be V/V for sure, but also typically ii or vii. Even vi (or VI) is sometimes considered pre-dominant but usually I think of a pre-dominant as being free of the tonic note, unless it's a ii7.