About the new (add♭9) Roman numeral display

In both common practice and popular modes, the accidental in the (add9) embellishment does not apply to Roman numerals containing an accidental on the root. All the chords containing a minor third should add a sharp or a flat to the (add9); in the link above, the first ♭i(add9) should contain a flat, and the last two ♯i(add9) should contain a sharp.

This is not required for the added ninth on chords with a major third (yet), for the same reason minor seventh chords such as the borrowed iv7 in major do not contain a flat in the extension (yet), because no ambiguity could arise since Hookpad does not support non-diatonic alternatives like maj(add♭9) and mM7 (both would appear in harmonic minor by the way, but 7♭9 is more appropriate than the former for most situations).

In any case, the modern chord symbols still do not contain (add♭9) for iii(add9) and viio(add9) in the major key, e.g. Em(add♭9) and Bdim(add♭9) in C Major.

Also, while you are at it, this should be consistently extended to sus4 and sus2 as well.

1 Like

@HertzDevil, thanks for pointing this out.

We originally posted this change to enable some examples that we are preparing for our second theory book, Hooktheory II. Since then, we’ve realized that labelling accidentals on the add9 is a little more delicate than we had originally thought.

The issue really has to do with the fact that we are merging Roman numerals, a classical theory device, with (add9), which is really a popular music concept (same issue with sus4 as well). The question is whether it is more useful for the “9” in add9 to refer to the interval with respect to the root of the chord (in this case, a major 9th could be add9, a minor ninth would be add♭9, and an augmented 9th would be add♯9), or should the “9” refer nominally to the note that is 9 scale degrees up from the root, in the given scale (in this case iii(add9) would be understood to contain a minor 9th, since that’s the interval that is made between scale degrees 3 and 4 in the major scale).

From a purist perspective, it would seem that the latter interpretation is more correct. After all, we don’t qualify the figured bass 642 with accidentals depending on the quality of the 6th and 2nd intervals. The only way we know that the “2” in V42 is different from the “2” in I42 is that although both V and I are major chords, we recognize that that 4 → 5 is a whole step and 7 → 1 is a half step. This is not dissimilar from recognizing that the difference between a vi(add9) and a iii(add9) is that 6 → 7 is a whole step but 3 → 4 is a half step. In Hookpad 2 we are discussing generalizing the theory to work with arbitrarily defined scales, and would seem that needing to qualify figures relative to the major scale in some sense hinders our ability to try and internalize a new sonic space that is provided by a nontraditional scale.

However from a practical perspective, many people think of (add9) as part of a quality. Along these lines, one would expect minor chords embellished with (add9), for instance, to always sound the same, regardless of what the root scale degree is. I can certainly sympathize with wanting to know that a ♭♭VI(add9) and a ♯III(add9) have the same quality without having to figure out the context of how that chord came to pass.

Ultimately we haven’t finalized this decision, and appreciate feedback on the issue.


Although the added ninth as an embellishment is a popular music concept, the 9-8 suspension certainly is not, nor are the 4-3 suspension and the 2-3 retardation (2-3 suspension is not yet possible in Hookpad since it calls for a slash chord); in these contexts the numerals represent scale degrees also.

The “practical” approach might work if it applies consistently to all extensions, e.g. the minor seventh is always 7, the major seventh always M7 or Δ7. Although some modern music theory books do present Roman numerals in this way, they do not cope well with inversions (some use a slash plus the bass scale degree as though they are true slash chords rather than just inversions, others might use “b”, “c”, “d” etc.), and in Hookpad they are redundant anyway; as Hookpad always displays modern chord symbols along with Roman numerals, said distinction between a minor ninth and a major ninth embellishment could be succinctly done in the chord symbol part.

V24 and I24 are rather subtle, since they could have come from Lydian and Mixolydian respectively. For seventh chords in the third inversion, this is rarely a problem in practice, since Roman numerals are almost always accompanied by the actual bass notes in some representation (staff notation or, in Hookpad, the chord colouring); where absolutely no ambiguity shall arise, one would use V♯7d and I♭7d. The issue does arise for other inversions of extended chords, e.g. IV56 represents both Cmaj7/E and C7/E. Whatever notation it be, symbols should not be indistinguishable unless their interpretations are fully replaceable with each other, which in this case probably are not. The developers have most likely addressed this in some way already when the change log mentions that a way to display which mode a chord comes from has been added to Hookpad, but such option does not seem to be available for users right now.

Just saw the new modal chord displays in the editor. You might want to add an option where modes are relative to the root of the chord rather than the current key, although that would not be quite useful until Hookpad begins to deal with jazz harmony.

Well, that a neat thing. I also notice that the second half of this tab (which is entirely borrowed chords in order to get the key to g#m) shows “bor”. Interesting.

EDIT: It seems to be that Lydian-Locrian is bor. I’ll make a list for other supermodals later.

EDIT 2: Here’s a link to a Google Spreadsheet. Yes, it’s editable. Major and all borrowed modes are done, but I’m too tired to do all 7 modes.