# Should sensible analysis show secondary functions fulfilling the supposed function?

This question started for me with an analysis using `IV/IV`.

It was analysis in Kostka/Payne for Mozart's K545, mv II, mm. 69-71...

`G: I | IV6/IV V6/IV | IV`

`G: I |C: IV6 V6 | IV ... G:...`

First that made me wonder how to show a tonicization rather than a modulation. I thought the `C:` change was easier to read, but is that reserved for the specific meaning of "modulation?" If tonicization is meant, the secondary slash notation should be used?

Next, should the secondary role be fulfilled? In the Mozart example it is. It eventually gets to `IV`.

By comparison, if I label a `V/ii` shouldn't it fulfill its dominant function in some way? Like `V/ii-ii` or `V/ii-VI/ii`. Hypothetically, if I had `C A7 Am` and labeled it `I V/ii vi` or `I V/ii v/ii` isn't that kind of silly? In what way is that `A7` in context acting as a dominant?

If I had `C f#o B7`, wouldn't `I iio/iii V7/iii` at least make some sense? Because the subdominant function of `iio` at least fulfilled its role and moved to the dominant of `iii` - in other words `f#o B7` is `iio V7` in `E` minor.

• – Dom
Jan 4 '19 at 19:41
• @Dom yes, your question was similar. Although, after this discussion, I now would say'at least two chords are needed to show a real function.' How to label something that' non-functional is another question. Jan 4 '19 at 19:48

Short Answer: Yes, the secondary role should be fulfilled. In other words, the analysis should show how a chord is functioning. It's not really functioning as a V/ii if it doesn't go to ii (or a VI/ii, or some other possible resolution).

One caveat to this is the notion of the back-relating dominant. In short, a back-relating dominant is a V chord that's connected to the I chord that precedes it, not a I chord that follows it. We can imagine something like `C Dm A7 G`, which could be viewed as `I ii V7/ii V`, with that A7 functioning as a back-relating dominant. It's more rare, but it does happen.

Longer Answers: In the Kostka/Payne example, I view that as what we often call an "extended tonicization." To me, a modulation must achieve a cadence in the new key, and a cadence must have both a root-position V and a root-position I. (Some people will disagree with me on those points, but the scholarly world of music theory is basically in agreement. Also keep in mind that this is Mozart, so the style is different than it is in 70s rock.) Since this example has an inverted V, it won't be a modulation; but since we can view multiple chords as pertaining to that IV, it's an extended tonicization.

The `C:` is typically reserved for modulation, but you can bracket underneath the applied chords to show an extended tonicization:

```G: I       | IV6   V6 | I
\_____________/
IV
```

And yes, your hypothetical E-minor example is just another instance of what I would call an extended tonicization. Though out of context, since it's a root-position V7, it could be a modulation, but of course we'd need a real example (with meter, etc.) to determine whether it was a modulation or a tonicization.

• That bracket is much easier to read that several slashed chords. Thanks! Jan 4 '19 at 19:08