The question

Does music theory provide a language for directly expressing voice leading? Roman numeral analysis (RNA) encapsulates voice-leading expectations, but these are built in by implication (indirectly) to what is a harmonic (vertical) language. Is there a more directly melodic (horizontal) language?

The origin of the question

In Bach's setting of the chorale "Ach Gott, vom Himmel sieh' darein" (BWV 2) (IMSLP: chorale #3) (YouTube), the final phrase is harmonically complex, but voice-leadingly simple(r). A harmonic (Roman numeral) analysis doesn't convincingly capture (IMO) what's actually happening compositionally.

Does "music theory" provide an analytical language for voice leading outside of the voice-leading expectations built into RNA? That is, is there a way to express voice leading directly?

If yes, a demonstration of how that language would be applied in this situation would be very helpful. Here is the score of the phrase in question. (Note: The preceding phrase cadences in A minor.)

Choral #3, mm. 8–10

Other questions of interest

  • 2
    Interesting question! Can you clarify what such a language might say or do that our current system(s) don't already provide?
    – Richard
    Commented May 22, 2022 at 23:26
  • @Richard I've been wondering that myself, since the current language really did develop out of a desire to express voice-leading relationships between different parts — first intervalically a la Fux, and later, chordally with RNA. As a starting point, though, one thing that strikes me about the Bach excerpt is how, in the alto, F natural is used to place us in C major (in my analysis), but then F#-G# is used to run us toward A — but the note A rather than the key of A. This can be explained harmonically, but the explanation is, IMO, overly complex and even misleading.
    – Aaron
    Commented May 23, 2022 at 0:06
  • 1
    @Aaron The chorale is in fact in A minor (easily seen as each cadence is clearly A minor). It is common practice in german baroque notation to notate music in minor keys with doric key. This can be well seen in the same collection, number 8: The chorale is obviously in F minor, but it only features three bs, as would be expected from C minor.
    – Lazy
    Commented May 23, 2022 at 10:38
  • 1
    @Aaron Which is the reason why I’m posting this as comment. As already said in my comment it was common practice in german baroque notation to notate minor in doric notation, thus with one step towards the sharp side. For A minor this means we get a single sharp, which is what you interpret as E minor.
    – Lazy
    Commented May 23, 2022 at 12:58
  • 1
    @Lazy indeed I am not. Chorale denotes the melody, whether or not it is harmonized.
    – phoog
    Commented May 23, 2022 at 13:10

1 Answer 1


RNA will tell us unambiguously what the chord roots are.

RNA will tell us almost unambiguously what the bass line is through the figured bass inversion symbols. It won't tell us octave or anything about non-chord tones, so the line is at most a sketch of the bass.

The example from Bach in RNA is Am: i | V6/3 i iv6/3 viio7/IV | IV viio7/V V | which as a sketch bass line, without the non-chord tones and moving smallest distances, is...

enter image description here

I think it's possible to use standard theory terms to give a sketch of a soprano line, at least in some cases. You could say something like: harmonize the passage in conjunct, contrary motion. With the exception of the first beat of the second bar, that would give us...

enter image description here

...which isn't too far off of the actual line.

But, harmonize the passage in conjunct, contrary motion could also give us...

enter image description here

...which isn't like the original soprano. The pitches are different, but the contour is similar. Then again, the bass line, derived only from the RNA isn't all that accurately representing the original.

So, you could use standard terminology:

  • line type: conjunct/disjunct
  • relative motion: similar, contrary, oblique
  • direction: ascending/descending

...to describe how the harmonizing soprano moves.

But that will get cumbersome quickly. Why not just annotate the Roman numeral root with one more symbol, the principle soprano tone..?

   ^1  |^5    ^3  ^4     ^5        | ^4  ^3       ^2 |
Am: i  | V6/3  i   iv6/3  viio7/IV |  IV  viio7/V  V |

...that isn't too much extra writing and it's fairly unambiguous.

That's more or less how Robert Gjerdingen expresses his voice leading schemata in Music in the Galant Style, except he doesn't use RNA for the bass, he just writes the scale degree number for both the soprano and the bass.

  • One problem with trying to define "well-formed" melodies (analogously to logical Well-Formed-Formulas) is that almost any irregularity can be made regular (or at least musical) by repetition.
    – ttw
    Commented May 23, 2022 at 22:54
  • I'm not saying anything about "well-formed", only that you can give short hand for a harmonizing line, well-formed or not. Commented May 24, 2022 at 13:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.