# Trouble reading roman numeral notation with flats

I am working my way through an article on transformations in rock harmony, but I am getting stuck trying to read the author's roman numeral analysis. I would really appreciate some help understanding how to read flattened roman numerals. Here is an example of what I mean:

Thank you!

TLDR;

a sharp or flat before numeral mean raise or lower the chord root from its normal diatonic spelling

In the usual Roman numeral analysis we have these conventions:

First we have to give a label for a key like `E:` for E major or `Em:` for E minor. Then we have these points...

• capital numeral means major triad
• lower case numeral means minor triad
• 'o' after numeral means diminished triad
• '7' means a diatonic seventh above root
• a sharp or flat before numeral mean raise or lower the chord root from its normal diatonic spelling

That last rule is the crux of your question.

In `E` major the diatonic chord of the sixth degree is `C#` minor and gets labelled `vi`.

In `E` minor the diatonic chord of the sixth degree is `C` major and gets labelled `VI`.

If the `VI` from `E` minor is used in `E` major the chord root is altered from `C#` to `C` natural. (This is called chord borrowing.) When in `E` major that change in the chord root is made explicit in the Roman numeral figure by using the flat to show the change of chord root: `bVI`.

The reason your book's example is confusing is because the key signature on the staff is `E` minor, but the analysis key shows `E` which is read as `E` major (it should be `Em:` to mean minor) and then the Roman numeral figures are labeled with flats which makes things seem like the assumed key is `E` major. Basically the labeling is a mess and it makes confusing the use of sharps and flats on Roman numeral figures.

As others point out, the book doesn't even label the chords as 7th chords!

The book seems sloppy.

@Tim makes a good point that Roman numeral analysis can be tricky with rock music. The real conundrum to me is the final chord. It's an `E` dominant 7th with a `G` natural on top. That is a very bluesy sound. It's a very familiar sound, but Roman numerals are a poor way to label it. There is no way to indicate the simultaneous `G#` and `G` natural... and the minor 7th. About the only thing the `I` label gets right is a root on `E` and a perfect fifth of `B`. Tread carefully with harmonic analysis of rock music. You might find a lot of inappropriate labels.

• Thank you for the great response! Your answer was very clear--I appreciate it! – 286642 Mar 19 '19 at 19:42
• That's the Hendrix chord. E7#9 So there's a 7th - D, and a #9 - Fx, which isn't exactly G, more F##. Can't think of any RN that describes it! – Tim Mar 19 '19 at 20:52
• Using the 7#9 is a common way to fit the blues major-minor chord into roman numeral analysis. In equal temperament, the distinction between a double sharp 2 and a flat 3 become less important. The notes are the same, so they are named either for function or even convenience. – trlkly Mar 19 '19 at 21:32
• @trlkly, `7#9` as in `E7#9` would be a jazz symbol rather than a Roman numeral symbol. But it is interesting the #9 in a dominant chord versus a flat (minor) third over a major triad. I think when I have seen the latter it has been for a tonic chord. I've not seen a rule stated, but #9 to me says 'dominant' and flat third over major triad sort of 'tonic' – Michael Curtis Mar 20 '19 at 2:46
• Trouble is, in relation to this music example at least, the answer's wrong. The example uses the 'mode-blind' convention of chord numbering. Everything is named relative to E major, even if the tune's in E minor. – Laurence Payne Mar 20 '19 at 14:51

The notation seems a bit confusing. The first chord is a D7 in the key of either E-Major or e-minor; the author isn't quite so clear. The key signature indicates e-minor in which case the chord is VII (no flat needed.) If the author shows keys in ALL CAPS (which has been outdated since 1834), the chord VII7 but that disagrees with the single sharp signature. The author also leaves off the seventh indicator.

Perhaps we can assume the author uses E irrespective of Major or minor mode and that he leaves off the seventh. This still leaved the bVII vs VII so it's still still. The (V) must mean the seventh of G.

The author shows a descending scale (the caratted numerals above some of the chords) so perhaps that may help. I think he's using the Roman numerals associated with E-Major but the score is written as e-minor; he's assuming that all chords have sevenths (OK, but limiting). Then we have (if the score were in e-minor): VII7, VII7, VII7, VI7, III, VII7, VI7, V7, IM7 (at least as I would notate it.)

• Your 3rd line - did you mean 'chords' not 'keys' in ALL CAPS? – Tim Mar 20 '19 at 9:36
• You'e right; I meant chords, not keys. It's harder to read Roman numeral analysis with all cap chords. Older versions of Piston's Harmony do this. Marx's 1830s book does use more modern notation and is easier to read. – ttw Mar 20 '19 at 13:47

Are you confused because the song's in E minor, but the Roman numeral chord names are relative to E MAJOR? Yes, that's how it works. The numbers only care that E is the tonic. E minor triad is i, E major triad is I. The uppercase Roman number doesn't mean 'diatonic', 'in key' or 'correct', it just means 'major'.

Similarly for modified numerals. G is the 'correct' third note of E minor, but G B D is called ♭III. That's ♭ because G is the flattened 3rd of E major, III because it;s a major triad.

G B♭ D would be ♭iii. That's ♭ because it's built on the flat 3rd of E major, iii because it's a minor triad.

I repeat, these names would be the same whether we were in E major, E minor or E anything-else.

RNA isn't designed for analysing 'rock' music. It sort of works when non-diatonic chords need naming, as we use 'b' in certain circumstances. As in key E. There is no D and no C, only D# and C#. So, we have to call the D chord bVII and C chord bVI. It gets as close as we can.

The bIII represents G, again because there's no G in key E: G# is iii (III) but we need the flattened G#, making G. Thus bIII.

The trouble is:

The key sign is G-major/e-minor: (in e-minor the flat should be a natural sign to explain the VI and VII degree are not of the harmonic scale!)

Indeed this progression is notated in respect of E-major!

D - D - D - C - G - C - D - E

If you transpose this figure to C you will understand that the flats are concerning

Bb - Bb - Bb - Ab - Eb - Ab - Bb - G - C (with the blue note b10)

• 'In Em, the flat should be a natural sign...' I don't understand! – Tim Mar 20 '19 at 8:30
• As in e-minor (harmonic and melodic key) the VII is D# a flat wouldn't be quite the chorrect sign, in my opinion there should be the sign the accidental corresponding to D natural). – Albrecht Hügli Mar 20 '19 at 8:54
• Except we don't necessarily expect E minor to have D# - it could well be D nat. Although using the major as a datum point, D# would be the diatonic. But diatonic doesn't feature well in rock music, so RNA is never going to be a good idea. – Tim Mar 20 '19 at 8:58
• RNA is roman numbers? Yes, I agree. To me it's only useful for historical interests. The Chord signs will fit here better. – Albrecht Hügli Mar 20 '19 at 9:07
• Except we don't necessarily expect E minor to have D. Have you seen the remark below the time 4/4? it says E: (if this means E-major the VII and VI would be D# and C#. – Albrecht Hügli Mar 20 '19 at 9:12