I mean the non-diatonic kind. Like for example if you had the progression C, Am, F, G7 you could use roman numerals to classify the progression as I, vi, IV, V7. But what do you do for non-diatonic chord progressions? Like the end "starship trooper" (an easy example) the chords are G, Eb, C and since the chords don't all naturally belong to one key, do you just pick one and specify which chords are altered? You could pick the key of Bb and say the progression is a VI, IV, II, or pick Eb and say it's III, I, VI. Does it matter? Maybe I'm looking too much into it but lately I've been going through some of my favorite songs and analyzing them, so thats the reason behind this.
Determining the key of that G Eb C progression is difficult and I will not discuss that part (mainly because I'm kinda stumped), but if you consider G the key, you would write that progression as I bVI IV.

Here are some exercises. Assume the first chord gives the key.

1. E G D A

2. A Am G Fdim

3. Am A Bm B Bmaj7 B7#9

4. F G A Ab

5. B F C E# B#m

These are supposed to be hard, but try your best. I'll correct any errors you make. Also, I'm picking random chords with no regard for sound.
k here i go.

1. I bIII bVII IV
2. I i bVII bvio (not sure how to do the "o" superscript for diminished chords")
3. i I ii II IImaj7 II7#9
4. I II III iii
5. I bV bII bV bii

do people normally choose the key in a way that will make the numerals easier? for instance the first one would be really simple if D was chosen as the key and in that case it would be II IV I V.
Quote by einspacetime

5. I bV bII bV bii
This is wrong. While E# and B# sound the same as E and F, their ROman numerals are different. Please make the correction.

Quote by einspacetime
do people normally choose the key in a way that will make the numerals easier? for instance the first one would be really simple if D was chosen as the key and in that case it would be II IV I V.
No. The key isn't really chosen. It's the note towards which the progression gravitates.
Ummm how about....I bV bII #IV #i
Regarding the Starship Trooper progression:
No two of those chords belong to the same key. An idea would be to analyze the solo, and see what you can infer from his note choice. From what I can see, he's treating the G and the C as dominant 7th chords and the Eb as a major 7th chord. Think of the chord progression as an implied G7-Ebmaj7-C7 progression. He makes use of the blue 3rd on both the G7 and C7, so keep in mind that you can use the G and C minor pentatonic scales, respectively.

The way that I would treat this progression is almost as a blues progression, where you're moving between V7 and I7, with Eb being the gateway. I think Eb works because it is used so briefly and when you're moving to C7, the Eb note resolves to E which is the 3 of C (not with the voicings that he's using but in spirit), and since Ebmaj7 has the notes Eb-G-Bb-D, with G being the 5 of C, Bb being the b7 of C, and D being the 9 of C(a consonant note), it creates a bit of extra movement without moving too far outside the lines.

All that being said, I would call C the tonal centre, and with roman numeral notation, the progression would be: V7-bIIImaj7-I7.
known as Jeff when it really matters
Last edited by titopuente at Jul 22, 2008,
I realize that he's only playing triads, but in the solo he plays a Bb over the C, and the only time he plays a B is over the G, which to me implies that he's treating the C as a C7.
known as Jeff when it really matters
Quote by einspacetime
the chords are G, Eb, C and since the chords don't all naturally belong to one key, do you just pick one and specify which chords are altered? You could pick the key of Bb and say the progression is a VI, IV, II, or pick Eb and say it's III, I, VI. Does it matter? Maybe I'm looking too much into it but lately I've been going through some of my favorite songs and analyzing them, so thats the reason behind this.

As far as the G Eb C progression this is rather ambiguous. Both the G and C can sound very stable with these chords. If pushed at first glance and hearing it for the first time I would probably describe it as a C progression using as a V-bIII-I. I say this because the root movements outline a Cm triad.

However there is a common note between the three triads ->G and if we threw this in the bass could we possibly have some pitch axis?

Anyway that is not a concept I am all too familiar with so I analyse the chords the way which is regardless of what the key or roman numerals are. Through studying individual movements of notes and chords and how each contribute to the overall effect.

This particular progression for example uses two movements down a third which create a sense of forward momentum before pulling in the reins with a move up by a fifth to create some tension and interest.

The first movement is G->Eb is a movement down a major third
The note movements could be viewed in the following way...
G->Eb
G->G (The root of the G becomes the 3rd in the Eb)
B->Bb (The 3rd in the G chord moves down a half step to the 5th in the Eb)
D->Eb (The 5th of the G chord resolves up a half step to the root of the Eb)
The G remains in place while the remaining two tones move an opposite direction by a half step each. I personally find a half step up to provide a great sense of resolve while a half step down tends to be kind of ugly - but that's just me.

The next change from Eb to C sees a root movement down by a third again but a minor third this time.
Eb->C
Eb->E (The root of the Eb resolves up a half step to the 3rd of the C)
Bb->C (The 3rd from the Eb triad moves up a whole step to the root of the C chord)
G->G (once again the G remains as the 5th this time in the C chord)

The final movement is rom the C back to the G. This is a movement up by a perfect fifth which is a strong chordal movement. I typically view movements up a perfect fifth as a pull against the natural flow of the V toward the I.
The individual note movement might be viewed as...
C->G
C->B (the root C moves down a half step to the 3rd of the G)
E->D (the 3rd in the E moves up a whole step to the 5th of the G)
G->G (once again the G remains constant moving from the 5th in the C triad to the root, obviously, of the G)

Now I believe if we were to use the G in root position the Eb in 1st inversion and the C in second inversion we might get some some pitch axis? I'm guessing this would imply a root of G?
Though I'm not 100% on this and would love for someone to fill me in.

Either way, whatever root you use the relationships between the individual chords doesn't change.

Hows that for an over analysis of three chords?
Si
wow. thanks guys.