Numbering chords

I'm sure Noteboat will impart some greater wisdom here, but this is my take on it:

1 - It's in E - the whole tonal pull is back to E - it starts with E and look at the perfect cadence (B7->E) in the last line.

2 - G# - there's no Ab in E major (or C# minor come to that) so let's stay within the key. Likewise, it's C# minor not Db minor because there's no Db in the key of E

3 - Let's look at the notes you've got - A, F#, C, E. We've got two stacked minor thirds of F# to A and A to C to make F#dim. The E is the 7th of F# giving us F#dim with an augmented 7th (a dim7 chord would need D# instead of the E to keep the stack of minor thirds - source for that one is the Chambers dictionary of music.) Going the other way, A, C and E make a simple chord of A minor, F# is the 6th of A in the key of E, so Am6 works for me.

4 - well, what have we got? Keeping it in E throughout:

1st line - I - III7 - vi - vi
2nd line - bVII9 - III7 - vi - I9
3rd line - IV - III7 - vi - iv6
4th line - I - V7 - I

A :-)

I'm with Alan on it being in E major. You'll often have shifts to a relative minor key for a while, but the last cadence does tie things up.

He's also got the G#7 covered. When you're in a sharp key, you wouldn't use enharmonic flats in the spellings.

Am6 and F#m7b5 are enharmonic. When you've got a choice between two chords, the root that makes sense is the correct name. With the exception of the first chord, the root motion mirrors the first line, so it should be A - and that makes it Am6.

Thanks. That cleared some things up for me. I was seeing chords that aren't diatonic to E major and thinking I should change the value of "i/I" to reflect the apparent change in key. I didn't know that something like bVII9 was allowed. But thinking about it now, if the whole purpose of writing out chord progressions in roman numerals is to communicate to other musicians how to play the song in any key, it makes perfect sense to write it that way. I have a few follow-up questions if you guys don't mind:

1) When is it appropriate (if ever) to change the root of the i/I chord? I imagine there would be some song that starts in one key (say E major), and then in the middle of the song switches to another key (say G Major) without ever returning to the original key. In a case like this, would you change the I chord (from E to G)? If so, how would you notate this?

2) Let's say I'm Mark Knopfler and I'm writing this song, and I have the first line (E - G#7 - C#m - C#m ) in the bag. How do I come up with D9 (bVII9) for the next chord? Is there something in music theory that predicts this chord or explains why it's a good choice? I think it sounds great, but I'd like to know why. bVII9 seems random to me. I notice that it is almost the same as the Am6 that comes later in the song (Am6 is like a D9 without the D). Any ideas?

Thanks!

1. The Roman numeral system only works within the context of a home (I/i) key, so it's not really appropriate to change your I/i chord. But more complicated tunes often modulate to new keys, so that leaves us stuck - how can we use a diatonic system to notate chords that span two keys?

One solution I remember from college was the use of two lines of Roman numerals... it's probably better suited to analysis than notation, but for what it's worth I'll present it here.
Chords: C F G7 C7 F Bb Gm etc.
Key of C: I IV V I7
Key of F: V/V V I IV ii

Not exactly elegant, but in the context of analyzing a piece it works. It's absolutely no help in communicating the progression to the rest of the band, which is why we use charts with chord symbols instead of Roman numerals for the more complicated stuff!

2. Explain, yes - predict, not really. The job of theorists is to explain what's already been done... the job of composers is to do new things. Composition always takes the lead, and composers craft not-too-helpful statements like McLaughlin's "any chord can follow any other chord" to enlighten us to their thinking. Lots of composers do things that leave us scratching our heads for an explanation at times. But what Knopfler has done here has a pretty standard explanation, so I'll just work backwards from the theory side of things.

You've got just one dominant chord (G#7) in the progression so far, and it resolves down to C#m. That's pretty standard - dominant chords go down a fifth, and it sounds pleasing to our ears. We've been doing it for about 600 years or so, and we've come to expect it. But we can drill down and ask WHY the G#7 needs to resolve.... and the answer is the chord has a tension - a tritone between the B# and the F#. That's a b5, which is a pretty dissonant interval. We want to hear that tension resolved.

What happens inside the chord change is a resolution by chromatic movement, with each note of the tritone moving a half step in opposite directions (in counterpoint, it's called contrary motion). In a major key resolution, B# moves up to C#; F# moves moves down to F (enharmonic to E#) and you get a G#7 to C# resolution (G#-B#-D#-F# -> C#-E#-G#). That's so ingrained we expect a C# root chord to come next, so the C#m works as a parallel minor substitution.

But a tritone is exactly half an octave. And that means it doesn't really matter which note moves up and which one moves down - B#->C# and F#->F could also be resolved as B#->B and F#->G. We hear a titone resolving to a third either way. And that leads us to a thing called the tritone substitution... instead of resolving from G#7 to C#, we could just as easily go from G#7 to D (G#-B#-D#-F# -> G-B-D - a G major chord. And what's the tension that resolves to G? It's a dominant D chord!

So by extending the D7 to D9, he's giving our ears the OTHER option for a resolution to G major. We hear it as a substitute for G#7, because either chord can end up in the same place (G). Since he's not actually going to go to a G chord, he bounces back to the G#7 - which we've already heard - and resolves it as we expect. And having prepared our ears to hear a D9 sound there, he's free to use part of it (Am6) later on, and we hear it functioning as a D9 - because short-term memory plays a big role in music; without it, no notes would ever be out of place.

Anyway, that's just one possible take on the inner workings of the chord progression. In many cases, deciphering how a composer arrived at a specific choice is a lot like a ground beef = cow problem: if you put a cow into a grinder, you know exactly how you got to what's on the other side... but from the other side, you can just as easily imagine it beginning as a pig (or a possum!) So any analysis like this is wide open to other explanations. :)

One solution I remember from college was the use of two lines of Roman numerals... it's probably better suited to analysis than notation, but for what it's worth I'll present it here.
Chords: C F G7 C7 F Bb Gm etc.
Key of C: I IV V I7
Key of F: V/V V I IV ii

Yep - this is the way I learned too

Thanks NoteBoat! That makes perfect sense. Looking at it again I see that G#7 and D7 have the same tritone interval to resolve (same tones, with 3 and b7 positions swapped). Is it common practice to swap a V7 chord with the Dominant 7 chord a tritone away? I'm going to have to play around with that when I get home from work and see if I can make something cool with it.
you get a G#7 to C# resolution (G#-B#-D#-F# -> C#-E#-G#).

A little question about spelling. You wrote the G#7 chord as G# - B# - D# - F#. I know that B# and C are enharmonic, but why did you decide to go with B# rather than C? Is it because the iii7 chord in E is normally G#min7 and you are indicating that you raised the third to make the G#7? Can you imagine a situation where you'd spell a G#7 chord with C instead of B#?
Anyway, that's just one possible take on the inner workings of the chord progression. In many cases, deciphering how a composer arrived at a specific choice is a lot like a ground beef = cow problem: if you put a cow into a grinder, you know exactly how you got to what's on the other side... but from the other side, you can just as easily imagine it beginning as a pig (or a possum!) So any analysis like this is wide open to other explanations. :)

Yeah, for all we know Knopfler could have written a bunch of chords on little pieces of paper and pulled a random one out of his hat. I've never seen him wear a hat though, just a headband when he was younger, so I'll put my money on your explanation. :)

Yes, the tritone sub is very common - especially in jazz.

For the spelling, chords are built in thirds. Since a 7th chord has a formula of 1-3-5-b7, we have to apply that to the major scale of the root. In this case, the root is G#, so we use a G# major scale: G#-A#-B#-C#-D#-E#-Fx-G# (Fx = F double sharp)

There aren't any situations where music theory would allow the substitution of an enharmonic tone. But in practical terms, sheet music will often do that - so if you've got a Cº7 (C-Eb-Gb-Bbb) you'll often see the Bbb written as A. That can make it tricky at times to convert the written music into chord symbols, but the theory behind the chord spelling never changes - no matter how awkward the spelling gets.

A little question about spelling. You wrote the G#7 chord as G# - B# - D# - F#. I know that B# and C are enharmonic, but why did you decide to go with B# rather than C? Is it because the iii7 chord in E is normally G#min7 and you are indicating that you raised the third to make the G#7? Can you imagine a situation where you'd spell a G#7 chord with C instead of B#?

He had to go with B# to be correct. In the case of any G triad the third will only and always be some kind of B. As soon as you say C you are naming a 4th/11th

Because G A B...see how B is the 3rd letter...that's why its the only interval/note that can be named as a 3rd, and when he gives a Dominant 7 triad, you have 1 3 5 b7...so as you can see, a 4 is not resident in that formula.

It doesn't matter if its a Bb or even B# or B, in G any G kind, the 3rd will always be a B of some kind as well. This keeps things clean and clear, as you never have to second guess "Shouldn't it be a C?"

Now that said, I tell my students, if you're ever in a situation where you are playing and showing someone how to play a part which uses intervals and notes such as E7#9, call it by the note that's appropriate to the level of the other guitarist...don;t make it complex on a guy who has had no chance to understand these kinds of concepts.

Hope this helps.

Sean