My friend's chord progression.

Niliov,

Seems to me I said as much :)

Riemann analysis seems simple and straightforward, but I've actually found it much more difficult to follow for even straightforward jazz tunes than simply looking at things in terms of keys of the moment.

It amounts to the same idea, but calling a iii chord a sub-dominant chord confuses the issue for people who have learned to call the IV chord sub-dominant.

It's a semantic change that frankly, in my mind, doesn't have any real upside over a more classical approach to looking at harmony.

Was that simple?

I got Tom's explanation but yours threw me for a loop. Do you have any links like Rieman's notation for idiots? I'd like to read up on it.

Wel Kingpatzer,

You are partly right:

Rieman has troubles with dominant chains, but the numeral thing has problems with Napels and altered subdominants.

Example:

s-6 D T

There is no way to write this in Roman Numerals except the faulty:

bII6 V I

which gives the impression that in the key of C the first chord would be Db/F ->aaaaaahhhhhhhhhhh
This creates confusion since the root is not a Db but an F.

What about the widely used progression in jazz:

C-: C- Ab7 C- (for example: You Dont Know What Love is)

In roman numerals that would add up to:

I [V]td -> V(ommitted) I

The above would give the totally wrong impression that the second chord is a dominant (and a tritone subst) -> aaaaahhhhhhhhh
Of course it is not a dominant but an altered sub dominant (F# double dim with an Ab in the bass)

so correctly in Riemann:

C-: I S(dd) I

I decided a long time ago that in fact we should use both systems to complement each other, with the emphasis on Riemann!

Niliov

I just simply disagree. But I'm still learning this stuff and I'll admit my view is very much subject to change. Most of my view is coming from a heavy reading of various doctoral thesis' on the subject, and since I have spent no time in formal training it's entirely concievable that I've missed something important.

Rameau, Riemann, Schoenberg and others who have long sought out a syntax of chord progressions all fail. They either fail because they merely place the chords in categorical boxes (ala Riemann) or they dump everything into patterns of root movement (Schoenberg).

Now, both of those have advantages to merely looking at chords in terms of scale steps codified as roman numerals. However, simply categorising chords as tonic, sub-dominant, dominant says absolutely nothing about the ways in which the chord flows from each other or how those chords relate to the musical phrases.

As such, Riemann, Schoenberg and Rameau (and others) all fail to actually deliver what they promise. They give a system of categorization but not a syntax. They ignore that harmony and melody are two sides of the same coin and that if what you're saying about the harmony doesn't actually give you more information about the song the chord progression is from it doesn't add value.

Such analysis of progressions says nothing about which progressions are right in certain situations. This is no syntax. Riemann fails to meet his own criteria!!

Finally, these models always ignore the importance of linear progressions
particularly as they relate to melodic phrasing. And they do little to provide insight into ontrapuntal formulations and harmonies, something Schenker points out. They dont' account of modulations. They don't account for chromatic harmonies.

There are relationships between harmonic functions and musical phrases. THat is what is important in a song. Riemann and his ilk's preoccupation with 'functional analysis' and categorisation of chords totally fails to see these patterns.

Compare that with someone like Schenker's theory that the key issue is in the relationship between chord progressions and linear progressions. Are chord progressions created from linear progressions or do linear progressions just decorate underlying chord progressions? Riemann and Schoenberg and others don't even demonstrate an awareness of this sort of question.

Focus on Riemann if you want. But I think that's a dead end.

Okay this is getting to be interesting:

I like what you said especially:

"...As such, Riemann, Schoenberg and Rameau (and others) all fail to actually deliver what they promise. They give a system of categorization but not a syntax. They ignore that harmony and melody are two sides of the same coin and that if what you're saying about the harmony doesn't actually give you more information about the song the chord progression is from it doesn't add value..."

You are right this information alone doesn't accomplish anything but trying to find a relation between chords, which is worthless if the piece of music the harmony comes from well....sucks!

BUT: If you want to learn about harmony and not get stuck on things like: A G D E which we both agree is one of the simplest chord progressions known to popmusicians all over the globe harmony analysis is invaluable.

I am not saying using Riemann will solve all your problems, musically or domestically for that matter but is it a nice system which could really open one's eyes to the simplicity of harmony. In practice I also use mostly chord symbols, I'd even write Ab7/Gb when I know it is suppossed to be F#dd.

Whenever I get a new classrooms full of theory students one of the first things I tell them is:
- There are many theory scholars whose music could be flushed down the toilet immediately
- The most beautiful music was written by people who knew nothing or very little about harmony
And I play them a Pierre Boulez piece (the first category) and a Beatles and Hank Williams song (second category).

Niliov

There's certainly some beautiful music that was written by people uneducated in harmony, but that's not really surprising - because almost* all theory follows common practice. (*almost - exceptions would be things theory predicts based on 'holes' in common usage, like the Locrian mode). There's also plenty of beautiful music that's been written by people well versed in traditional harmony. Like any creative field, it's best to have both the art and the craft at your disposal, but art alone works too.

I agree with Niliov that the typical systems of Roman numerals or the related 'Nashville' notations don't work terribly well for music with non-diatonic changes. I also agree with King that Riemann's system has shortcomings.

The basic problem, I think, is that there are simple systems that work for simple music... and that's what folks are likely to encounter. Those systems start to collapse when you're looking at German sixths, the Tristan chord, quartal harmony etc.

Efforts like Riemann's strive for a master key to harmony. It's a nice idea... but then so is Esperanto :)

The major disadvantage of Riemann's system is that it's not limiting in what I'd call its predictive ability. Like many analytical tools, it's great after the fact - but most budding songwriters are going to be much more comfortable with something that says "you've just used vi and then ii - the next chord should probably be V".

Riemann is useful for getting a 30,000 foot overview of a harmony's motion. But I don't see it as being a great starting point to understanding harmony - I think the logical route is the natural harmonic series, then secondary dominants, and then you'll be better grounded for wider approaches.

It's a lot like teaching time signatures. We all tell beginners "the top number is the number of beats per measure" when that isn't universally true. But it gives beginners something to hang their hats on, and discussions of compound meters can come a bit later.

Ok, we are getting a bit of track maybe, mostly my fault of course.

Let's get back to that chord progression since I was not very clear in my explanation as it seems.

A G D E

If we say the whole thing is in A the only problem is the G chord.

Of course you could try to prove that it is borrowed from a different mode (like A mixolydian) or that it is in two keys (we say "spilchord" in Dutch, translated that would be something like "dualchord", i.e.: a chord which can be seen in two keys) but then we are talking about a modulation for one chord which seems a bit odd, especially because the whole thing sounds so logical!

Whenever you get stuck analyzing it is mostly a good idea to look at the chord as a collection of leadtones going to the next chord.

Lets break down the G chord to D chord: G, B, D going to D, F#, A
So: G -> F# , B -> A, the D stays

Without context our ears recognize these solutions as IV going to I, right? (in this case G is IV and D is I).

So what's happening here is that we hear the D chord as a temporary I chord with the G chord as IV. Whenever this happens we write the extra IV between [] so we can see that it is relative to the next chord and not the main key of the progression.

thus: I [IV] IV V

Why do I think it is better to look at the progression in this way instead of looking at it as a progression in two keys? Two reasons:

1. The progression is too short and sounds to simple to actually contain some kind of modulation!
2. Now we can very easily add even more chords without leaving the key for example:

Lets say the [IV] chord (G major) is also a temporary I then we could put another IV before it resulting in:

A C G D E

Of course we should try to put the C and the G chord in one bar (two beats for each chord, or 3/1, or 1/3) s not to spoil our four bar set up (if each other chord lasts one bar). In functions the chord progression now would look like:

A: I [IV] [IV] IV V

This way we've never really left the key of A and things become a bit simpler. At least it looks that way to me, but maybe thats because I'm used to it!

Niliov

That amounts to exactly what I'd said... "Next, you've got closely related keys - the chords in D are D, G, A.. the chords in A are A, D, E. Two chords overlap. So you can be bouncing back and forth between two keys. "

Using secondary chords requires an overlap between two keys. Piston calls the overlapping chords 'chords of dual function' I think. It's also called 'pivot chords' by some writers, and the main term in most books is 'the secondary dominant principle' (because it's usually done with dominant chords, but you can use subdominants as well).

Here you've got A (the tonic), G (the secondary subdominant), D (the secondary tonic, and the principle subdominant) and E (the principle dominant). D appears in the keys of both D and A, making it the pivot.

Yes, you can say the entire piece is in A - but you can do that no matter what method you're using to assess the progression. I mean, Beethoven's first sonata is in Ab, but large sections are in Bb minor and C minor - even though the whole piece uses the key signature of Ab.

And no matter what method you use, you're going outside the key for the G chord - because you don't have a G note in key of A. Calling it 'IV of IV' is just another way of saying 'we're jumping into the key of D for a moment'.

In other words, it's a "IV of IV" chord.

Sometimes there's a beauty in simplicity. Many guitar players are most familiar with stringing together a group of "IV of" chords in the song, Hey Joe, where the chord progression is:

C - G - D - A - E

The use of "IV of IV" is prevalent in rock and pop music, almost a standard thing in a lot of rock, but you can find this type of progression in just about every genre of music.

Lunchmeat, you made a great observation of the fact that it wouldn't sound right if you started in your original progression (A - G - D - E) on the G. Did you try it starting with the D? That progression, D - E - A - G, would give you another typical short modulation:

I (D) - V of V (E) - V (A) - IV (G)

It's great that you're listening and experimenting with this sort of thing. Hope you don't let all the "technical talk" scare you off! :wink:

Peace

Noteboat writes:

"...That amounts to exactly what I'd said... "Next, you've got closely related keys - the chords in D are D, G, A.. the chords in A are A, D, E. Two chords overlap. So you can be bouncing back and forth between two keys... "

Actually yes and no but more no. We do agree that the Gchord has to be seen in the key of D but I make it point in my mail to state that the G is not a "pivot" chord which would indicate we are dealing with two keys here. You are always using examples from classical music (Beethoven seems a favorite) but then you should also know (as I do not doubt you do), that a pivot chord is the start of a modulation and that the modulation needs to be confirmed with a V-I cadense. So in the above chord progression we are not dealing with to keys because we are not modulating.

You write:
"...And no matter what method you use, you're going outside the key for the G chord - because you don't have a G note in key of A. Calling it 'IV of IV' is just another way of saying 'we're jumping into the key of D for a moment'...."

You seem to miss my point which is: that the Gchord is just a collection of notes which nicely lead to the notes of a Dchord because it happens to be the sub dominant in D, but it is not a modulation. I think there is a subtle difference which manifests itself big time in the analysis, because now we can avoid the involvement of "pivot chords", bitonality and keep things simple!

Niliov