Early

Journal Content on JSTOR, Free to Anyone in the World

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in
the world by JSTOR.
Known as the Early Journal Content, this set of works include research articles, news, letters, and other
writings published in more than 200 of the oldest leading academic journals. The works date from the
mid-seventeenth to the early twentieth centuries.
We encourage people to read and share the Early Journal Content openly and to tell others that this
resource exists. People may post this content online or redistribute in any way for non-commercial
purposes.
Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early-
journal-content.








JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people
discover, use, and build upon a wide range of content through a powerful research and teaching
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please
contact [email protected].

NOTES ON THE NATURE OF HARMONY

By OTTO ORTMANN

HEword "harmony,"in its broadestconnotation,applies to

any combinationor successionof simultaneoustones. Such

an applicationof the word includes regionsof tone hitherto
musically unexplored, virgin-fields that but await the touch of
the capable and observingcomposer before unfolding their beauties; beauties other than those to whichwe have been accustomed,
but beauties none the less. Strangelyenough,the particularfield
selected for recent exploitation-the whole-tone scale-is by its
very nature the least rich in internal variety of tone color.
There are many other scales-Scriabine has used some-that offer
distinctly greater advantages in this respect. I shall forego the
pleasureof exploringthese inviting regions here. Instead, I shall
content myself with an analysis, albeit fragmentary and somewhat belated, of what is really at the bottom of the system of
harmony which, founded by Rameau, has given us such works as
the B minor Mass, the Beethoven Symphonies,Feldeinsamkeit,
and Don Juan.
This system, as everyone knows, is based upon a relationship
of Fifths, numericallyrepresentedby 2:3. The octave, it is true,
represents a simpler ratio, namely 1:2, but in harmony, octaves
are conceived as identicals, which precludes their being used as
basis for the development of a system. Their identity, however,
permits an octave transpositionwhich does not involve a change
in the nature of the chord thus transposed. Such a transposition
I shall make use of later. But the Series of Fifths, in itself, does
not constitute the entire basis of our harmonic system. Taken
alone, such a scale of fifths embraces a wider range than that
which includes all that we call good writing. Here again is an
unworkedfield, waiting for appropriatemusicalmanipulation.
With the physiologicalbasis of harmony I am not here concerned. The promising theories of Shambaugh, Kishi, and
Hardesty may lead to an adequate physiological explanation,
impossible in terms of the basilar membrane theory of Helmholtz, notwithstanding certain peculiar advantages of the latter
theory.
The concept whichgives our system of harmonyits individual
character, a physiognomy which, though clearly defined, is yet
366

Notes on the Nature of Harmony

367

ignored with annoying frequency, is the concept of Tonality.

Tonality or Key, is the selection of one tone as principaltone
and of certain other tones as more or less subordinate tones. This
subordination involves relationship, and in this relationship is
found the key to the nature of harmony built upon such a scheme.
They who have read their d'Indy, Gariel, Riemann, Strube, and
Robinson will understand the importance of the principle of
relativity embodied in the tonality concept, and the logic of the
harmonic analysis here attempted. The principles advanced,
accordingly, are not new, but it is hoped that the following sketch
may throw light from a new angle upon the realization that a
fundamental shift of viewpoint is necessary for the teaching and
appreciation of harmony as an element of artistic music. Among
other things the famous "Don't Trespass" sign, long posted on
areas reserved for the great composers-although the latter never
as much as suggested reservation-must come down. It should
never have been put up.
*

The present chordal basis of harmony is a triad, generally

speaking, either major or minor. This basis is not permanent.
Twelve centuries have seen its progress from the simplest ratio
of 1:2 in primitive octave music, to the 2:3 and the 3:4, the fifth
and the fourth relation, respectively, of the Organum; thence
through the 3:4:5, the fourth and sixth ratio of the Fauxbourdon,
to the 4:5:6 ratio of our major triad. The Tierce de Picardie
still reflects the reaction to the minor triad as a dissonance; the
present time shows signs of accepting other, more complex chords,
as complete consonances, and the future will record the passage
of subjective consonance to yet higher ratios. The socalled major
and minor duality of our harmonic system is not as sharply defined
as we are often led to believe, for the differences in consonant or
dissonant character of the two chords are too great to permit
interchangeable use. Here the attempt at analysis will be made
from the basis of the major triad only. The conclusions reached
apply, in general, to the minor triad also, allowing for thedifferences in the scales of the two modes and in the ratios of the two
chords.
Let us choose C as the central tone of a tonality or key which
we call C major. Since our simplest ratio is the 2:3 the next
related tone will be G. Not lower F, for C would be related to F
as 3:2. We take the fifth in an ascending direction because

368

The Musical Quarterly

whatever be the modifications it has met with, the harmonic

basis in nature is the harmonic series of overtones, not undertones,
in spite of Riemann's splendid hypothesis. As we continue the
selection of tones on the basis of ascending fifths, we get
8va...........
Fig.1

Another fifth, Ff, would take us out of the tonality of C major.

The tones given, therefore, form the tonality as far as a series of
fifths can do so, and since there is no other source from which to
draw, these tones must form the entire harmonic basis.
By building a triad on each of the given tones, we get:
sVa.............

2
Fio.2

f-i_

The triad on B, since it is a diminished triad, must be excluded from the real harmonic basis, which is limited to major
and minor. It is included here merely to bring out more clearly
certain relationships described later.
Obviously, at the pitches indicated, the sounds are musically
not serviceable, or at least less so than in some other relation.
Accordingly, we transpose by octaves, a transposition, which,
as we have seen, does not alter the nature of the chord. Thus
transposed, our series becomes:
Fig.3

I ]I
M

^
0

This gives us a central chord U, to which five other chords are

more or less related. As we pass from U to M, we pass upward
in fifths, away from the key-tone or the key-chord. As we pass
from MIto U, we pass downward in the series of fifths and toward
the key-chord. The chord S is nearest related to the key-chord,
the chord R next and so on. In this relationship is the first
harmonic principle: musically a chord is always a part of a
tonality, and bears various relations to the other chords of that
tonality.
The second principle is that all chords tend to pass into the
key-chord. The latter is a point of rest to which the other more

Notes on the Nature of Harmony

369

or less transient chords tend to lead. Consequently, triads differ

not only in their structure, but also, and this is the important
point, in function. In the key of C major, every triad other than
that on C is harmonically, musically, or tonally, a dissonance,
and demands harmonic resolution as much as a seventh or a
suspension demands its non-harmonic resolution. The musical
purpose of a succession of chords is the establishment of the key
of which they are a part, for tonality is one of the two basic harmonic concepts. Such fixing of key is accomplished by progressing
toward the key-chord and not away from it. And it will be found
that any succession of chords passing along the series from 0 to U
is a harmonically acceptable progression. It is this because it
adheres strictly to the true harmonic progression, a progression
in descending fifths.
Chords stand in true harmonic relation only if their roots
are a fifth apart. The basic interval of chord structure, however,
is not a fifth, but a third. Accordingly, the third is a harmonic
interval. It cannot be primary because it is not present in the
original series of fifths. The E in the series, Fig. 1, for example,
has a different harmonic function from that of E used as third of
a triad on C. But it can be secondary, because the chords
which we built upon the single tones of Fig. 1 contain not only a
fifth, but also the interval of a third. By placing triads in thirdrelationship to those already derived, we get as our complete
chordal series:
Fig.4

f]
M

^A tv
OP

The added chords are N, Q, T. They could not come at any

other points in the original series because they would then not
stand in harmonic relation to the neighboring chords, the basis
of primary harmonic relation being a fifth, that of secondary
harmonic relation, a third. The primary harmonies of C major
are the chords 0, P, R, S, U. And the secondary harmonies
are the chords N, Q, T. But is not N the same as S, and T the
same as O? Not at all! This point illustrates the most fundamental error to which the usual treatment of harmony has been
subject. It results from the mistake of considering chords as
separate entities, irrespective of their environment, from placing
their structure first and their function second. Harmonically
speaking, an isolated chord does not exist.

370

The Musical Quarterly

Since there is but one basic harmonic relation, that of the

fifth, all harmonicprogressionsare primarilyfelt in this relationship. Thus the secondaryrelation of the third, is an incomplete
relation of the fifth, and tends so to be felt. That is, N, in Fig.
4, has the harmonicfunction of 0; Q has the harmonicfunction of
R; and T has the harmonic function of U. N could not have
the harmonicfunction of M, nor Q of P, because that direction
is anti-harmonic,away from the key-chord instead of toward it.
Now we are back to our originalseries of fifths:
M

6
Fig.5"5

[]

a harmonic h.s.
step

Ii

t?j
h.s.

h.s.

?
h.s.

and we see that it is the environment which makes the chord.

The musical value and nature of the latter changes with each
change in the environment. In the series given, N is not S but
an incomplete 0; T is not 0 but an incomplete U. Q, generally
viewed as sub-dominant,demandssomefurtherexplanation.
There is no self-existing harmonic sub-dominantfunction in
music. For we have seen that our series of chords of primary
relationship leaves no room for an F as independent tone. It
has been explained as a secondary relation. But what is more
important, is, that, thus explained, it has a dominantfunction,
because it is related to R, the second chord in the dominant series
of fifths. The F-A-C when used as a sub-dominant,is an entirely
different chord in function. The sub-dominant effect, which is
not really a progressionanti-clockwisearound the circle of fifths,
depends upon a precedingdominant effect. The progressionIV-I
is really only felt as IV-I when it has been precededby some V-I
progression. Played alone, without accent, IV-I is usually felt as
I-V. (The numeralsused merelydesignatethe chordalstructure).
Thus the effect of the Plagal Cadence depends upon a preceding
form of the Authentic Cadence, for the true characterof the subdominant is a deviation from the dominant, or harmonic progression,and wherethe latter has not existed, the formerobviously
cannot exist. The charming,often unusuallybeautiful effect of a
sub-dominantcoda, depends upon the contrast with the preceding
dominant effect. Why, for example is the typical harmonicform
of a periodI-V-I-IV-I and not I-IV-I-V-I? The chordQ therefore
(Fig. 5), has not a sub-dominantfunction at all, it is a dominant
chord, that is, it belongs to the dominant series. Its use as a

Notes on the Nature of Harmony

371

sub-dominant falls on the other side of the chord U (Fig. 5), and
naturally introduces other anti-harmonic relationships.
By considering the secondary harmonies as incomplete
primary harmonies, the weakness of certain chord progressions,
and the strength of others may be explained. Thus the progression M-N-O or S-T-U is harmonically weak because it is really
a harmonic reiteration, V-I-I, and reiteration stops harmonic flow.
The value of chord reiteration is found in the melodic aspect of
music, hence the rule to use chord repetition in places where a
marked melodic shift is required or desired. On the other hand,
as has been pointed out, any chord progression such as Q-S-U
is good, because Q being harmonically R, produces the perfect
primary progression R-S- U. The succession [M]-O-P-R-S- U with
its modification [M]-N-P-Q-S-U or any part thereof, taken in the
given direction, is the only fundamental harmonic progression in all
music.
All other chord progressions are melodic, not harmonic in
function. They have their essence in diatonic or chromatic progression, as the true harmonic progressions have their essence in
the harmonic fifth-relationship. It is true, that even in the harmonic series, some of the upper voices, but never the bass, will
move in diatonic steps, that is, melodically, but it is also true that
such progressions are determined far less by this factor than by
the harmonic relation existing between the chords. Proof of this
is found in the greater freedom with which primary harmonic
progressions are connected as compared to the non-harmonic
progressions. In the latter case, it is the melodic steps which
make the progression possible. As a consequence, wherever there
is diatonic progression on a bass other than a harmonic bass
(fifth-relationship), the true harmonies are not represented by
the chords as written.
All connections involving inversions of triads; seventh, ninth,
and higher chords, with their inversions; suspensions, and altered
chords, are melodic, not harmonic progressions. As such they
obey the psycho-physiological laws of melody, which are essentially
different from those of harmony as here outlined. As soon as
we add a seventh, or a ninth, or any dissonant interval to a triad,
this added tone has a melodic function.
This leaves the clock-wise, anti-harmonic fifth-progressions,
such as S-R-P in the Figures, to be explained. These, too, without
exception, are melodic in function, for they violate the second
principle of harmonic progression, which demands movement
towardthe key-chord. But what of the progression

372

The Musical Quarterly

whichcertainlyis musicallyacceptable? This is likewisea melodic

progression. Harmonicallythe sub-dominantcannot be explained
as the under-dominant. This becomes evident when we consider
the different treatments which the IV-I and the V-I demand in
music, and the differencein their effects, one of which is not the
opposite of the other but involves a differencein kind. The subdominantstands in no harmonicrelationto the tonic; the relationship is melodic. Far from being of almost the same harmonic
importanceas the dominant,the sub-dominant,throughits melodic
proximity to the tonic, is merely the most important example
of the anti-harmonicfifth-progressions. The musical satisfaction
which IV-I gives, is not the result of the harmonic progression
of IV into I, but the result of the IV already being a I. That is,
the

is a discord, whether F is in the bass or not, a

double suspensionresolving directly into the key-chord

As a discord, its function is melodic. When introduced, in the

coda, for example,by the seventh chord, (in C major by C, E, G,
Bb) it is also melodic, for all sevenths are melodic tones. The
sub-dominanteffect, therefore,is not a harmoniceffect opposite to
the dominant effect; it is a melodic effect of a different kind.
Thereis butone harmonicrelationin music, and thatis thedominant
tonic relation.' All else is melodic. The Circleof Fifths can only
be explainedanti-clockwise. Never in the reversedirection. It is
not to be conceived as a figure symmetrical with respect to any
F
BMr

diameter.

In the tonality of C, not as

E6

'Strube: A Treatise on Harmony.

c
_,s

G
\D

Notes on the Nature of Harmony

373

B
-E
-A
-D

but rather as
-G

-C

The third harmonicprinciple,which is but a further application of the two principlesdescribed,is that every discordis musically felt as a part of the tonality of which the next tonic harmony
is the key-chord, regardlessof the number of chords intervening,
and often changing in function with the onset of a new phrase.
The key-chord will consist of an uninvertedtriad, tonic function,
on some metrically accented beat. (This last statement is but a
generalization. It must suffice here, for even a brief treatment
of the rhythmic aspect of harmonic progression,would take us
too far afield.)
The first result of the applicationof these principles,is that we
may no longer consider such chords as the cadential four-six and
the passing four-six,or the dominantIV, representedin the figures
by Q, and the sub-dominantIV, which representsa melodic relationship, as one and the same chord. Musically, and that, after
all, is the point that counts, the forms of chords mentioned are as
different as the major and the minor triads. Such a conception
necessarilychanges our system of figured bass completely. Fortunately, this change involves no great difficulty. The plan
which first suggests itself is to number the central or key-tone I,
then to numberall other tones belongingto that tonality, that is,
grouping themselves around this tone, which becomes both their
aim and end, in relation to this I by counting the intervals from
the key-toneinstead of fromthe bass-tone. Thus:

,J JJ J
.

I1
3
5

I2
6
7

instead

1:

of I-V6-

I6 And as cadence:

(V2)('
It
3
6

374

The Musical Quarterly

instead of IV-I.

II
3
5

I1
4
6

Next, for the sake of simplicity, we may dispense with the

repetitions of I, with the understandingthat a tonality continues
until cancelled by another Roman Numeral. In like manner, I
alone may stand for I1, and I1, for I1. The figuringfor the first
3

6
6
is
is
then
It
example given
necessaryto include 1 in all
I-I.-I.
5

7
Such a plan may probablybe modified to ad-

cases except Il.

3
5

vantage, and is given merely as a possible solution of the problem

of figuring chords when conceived in relation to their tonality.
Figured bass, after all, is but a convenient form of numerical
analysis. The musical function of a chord, since it representsa
subjective reaction, cannot adequately be represented by any
fixed system of figures.

:J?,J J J, rj~
J
;p
E

I3II3
5
7

1
3

r<

V1
4

fr ~

The figuring as given

here ignores
the metrical
aspect and indicates only
the purely harmonic.

rJn r,.Jd

V
3
55

I!
3
5

or simplified
6

II-V1-I.
7 4
6

The C in the first chord is a seventh. But since sevenths stand

in no harmonic relation to the chord (they are melodic tones),
we need not be troubled by the ascent into D, for there is no melodic rule prohibiting ascent. The laws governing melody, even
in terms of the Lipps-Meyer theory, involve factors essentially
extraneousto pureharmonicprogression.
Finally, by taking a more complex example for illustration,
the extent to which true musical function differsfrom the figuring

Notes on the Nature of Harmony

375

of chords as generally viewed, will be made clear. Several musical conceptions of the phrase, with the resulting changes in
harmonic function, are given, and the unabridged figuring for
one of them is included.
Schumann.

VI
V
4 3
65
6
7
G minor I
C minor V

VI

VI

21.3
6
i
6

'2
It
5
3
7(
4
^
61.

12 It
4k 3i
5

I
64
63.

12
5

aj

12
7
57

1M

7ll 8

Bunte Blatter

2 b2
4
5
7ll

4
6

It

The brevity with which all dissonant tones are disposed of

as melodic tones may need a word of explanation in its defense. An
analysis of the melodic principles was not attempted since we were
concerned with the nature of harmony, and the very definition
which we gave to harmony, precludes melody from being treated
from any other than an almost opposite viewpoint. It may be
said in closing, however, that the melodic conception of dissonances
will be found to lead to a rational and musically adequate explanation of their use. One interesting result is the manner in which
a melodic analysis explains the "raisons d'etre," if any, of the
rules of harmony governing dissonances, inversions, and chord
successions, many of which are inexplicable on the basis of harmonic relation. As long as we conceive sevenths and suspensions
as different in musical nature and function; treat chords and
their inversions as harmonic identicals; and speak of harmonic
relation, such as the sub-dominant, where no harmonic relation
exists, we cannot hope to arrive at any satisfactory conclusions.
The plan here outlined, at least places the study of harmony
upon a musical, and relatively fixed harmonic basis, namely:
tonality and the descending series of fifths.