The Psychoacoustics of

Harmony Perception
Centuries after three-part harmony entered Western music, research is starting

to clarify why different chords sound tense or resolved, cheerful or melancholy

Norman D. Cook and Takefumi Hayashi

ing your favorite college fight song ness or tenderness, plaints, and mourn­ We believe, however, that the dif­
S or the United States national an­
them to a suitable instrumental accom­
ful songs."
The major/ minor distinction entered
ferent emotional responses to minor
and major have a biological basis. But
paniment, and the chances are that you Western music during the Renaissance before we venture into such controver­
will hear lots of stirring major chords. era, as composers moved away from sial territory, we propose to answer a
The Star-Spangled Banner is a perfect the monophonic melodies and two­ simpler question first: Why do some
example: When you sing "Oh say, can part harmonies used, for instance, in chords sound stable and resolved, and
you see?" you are singing the three Gregorian chants and embraced har­ give a sense of musical finality, where­
notes (one of them raised an octave) of mony based on three-tone chords (or as other chords leave us in the air and
a major chord. triads). Composers found that triadic expecting some sort of resolution?
Now think of a wistful, pensive harmony allowed them to tap a deep­ Psychophysical research has pro­
song, and there is a good chance that er range of emotions, of conflict and vided part of the answer. More than
the mood will be set by minor chords. resolution. That is why, to the modem a century ago, Hermann Helmholtz
For example, in the Beatles' Yester­ ear accustomed to chords, Gregorian identified the acoustic basis of musical
day, when Paul McCartney intones chants sound curiously monotonous dissonance. There is more going on in a
"Why she had to go, I don't know, she and emotionally flat. triad than mere dissonance or conso­
wouldn't say," the notes "why-had­ Major and minor chords remain ab­ nance, however; some relatively con­
go" form a minor triad. solutely central to Western music, as sonant chords nevertheless feel unre­
Music theorists were, of course, well as to non-Western traditions in solved. We have therefore developed an
aware of the different emotional reso­ which three-tone chords are not used, acoustical model of harmony percep­
nance of major and minor chords long but short melodic sequences often im­ tion that explains harmony in terms of
before Sir Paul wrote his opus. Jean­ ply major or minor modes. And yet the relative positions of three pitches.
Philippe Rameau, the French compos­ the psychological effect remains unex­ In particular, we have identified two
er and author of an influential book on plained. Today, this question has some­ qualities that we call tension and valence,
harmony, wrote in 1722: "The major how become an embarrassment to the­ which together explain the perception
mode is suitable for songs of mirth and orists. For example, in a book on music of "stability" and explain how major
rejoicing," sometimes "tempests and psychology, John Sloboda makes brief chords differ acoustically from minor
furies," and sometimes "tender and reference to research indicating that the. chords. This model will give us a basis
gay songs," as well as "grandeur and major and minor modes elicit positive for speculating on the reasons for their
magnificence." The minor mode, on and negative emotions in both adults different emotional connotations.
the other hand, is suitable for "sweet- and children as young as three years,
but neglects to discuss this remarkable Upper Partials
fact (Exploring the Musical Mind, 2005). The scientific explanation of music be­
Norman D. Cook is a psychologist interested In David Huron's Sweet Anticipation gins with the wave structure of tones.
in the higher cognitive functions of Homo (2006), the entire issue is relegated to Even a single isolated tone is more com­
sapiens, notably music perception, speech a single footnote. Most theorists are plex than it appears, due to the pres­
prosody, pictorial depth perception and the adamant that the association of ma­ ence of so-called upper partials (or high­
reverse perspective illusion. He is a prOfessor jor keys with positive emotions, and er harmonics). This one fact of physical
of informatics at Kansai University, Osaka,
Japan. Takefumi Hayashi, who received his
minor keys with negative emotions, acoustics was unknown to Renaissance
PhD. in engineering from Nagoya University is a learned response. It is simply the theorists, but is easily studied today
in 1992, does research in human psychophysics "Western idiom," and pointless to ex­ with a laptop computer and appropri­
and computer graphics. He is also a prOfessor plain in the same way that it is point­ ate software. The effects of the upper
of informatics at Kansai University. Address less to explain the conventions of Eng­ partials underlie many of the subtler
for Cook: cook@res.kutc.kansai-u.ac.jp lish spelling or grammar. phenomena of musical harmony.

www.arnericanscientist.org 2008 July-August 311

Figure 1. Music theorists have long been aware of the different emotional resonances of, for example, major and minor chords. The distinc­
tion between these two chord types entered Western music during the Renaissance, when two-part harmonies were supplanted by three-tone
chords. Although most theorists maintain that human responses to these chords are learned responses, the authors argue for a biological basis.
Their acoustical model explains harmony in terms of the relative positions of the three notes in a triad and how their complex higher harmon­
ics, or upper partials, interact with them. Barbershop quartets place particular emphasis on achieving the greatest possible amolmt of conso­
nance in their three- and four-part harmonies. They also exemplify the use of upper partials; accomplished quartets can lend the impression
that there are actually five voices. (Photograph courtesy of the Barbershop Harmony SocietylMax Duryea.)

The basic pitch of an isolated tone .dle-C on a piano, for example, sowld tal frequency is middle-C, then Fl is
can be described in terms of its "funda­ different from the same note played on an octave above middle-C (written C').
mental frequency" (denoted Fa, and ex­ a saxophone. In general, the upper par­ However, the next partial, F2, is between
pressed in terms of cycles per second, or tials become weaker and weaker and one and two octaves above middle-C,
hertz [Hz]). The Fa can be illustrated as can eventually be ignored, but at least because its frequency is only 3/2 the
a sine wave, as in Figure 3. Associated the first five or six partials have a signifi­ frequency of FI. In Western music, this
with the Fa are several upper partials­ cant effect on our perception. tone is called G'. Thus, as illustrated
Fl, F2, F3 and so on-which are sound The "upper partial story" would be in Figure 3, the middle-C on a piano
waves that vibrate at multiples of the easy if all of the partials were separated comprises a mixture of tones: C, C', G',
fundamental frequency. For example, if by octaves, but that is not the case, be­ C", E", and so on. This surprising fact
the Fa is middle-C (261 Hz), then Fl is cause pitch perception scales logarith­ makes the phenomenon of harmony
522 Hz, F2 is 783 Hz and so on. mically. That is, although the first upper more complex, but at the same time far
Any musical sound (other than a pure partial falls one octave higher than more musically interesting.
sine wave) will necessarily be a combi­ the fundamental frequency, further
nation of these partials. The number and multiples of the Fa fall at gradually Consonance and Dissonance
strength of the various partials give each smaller and smaller intervals above Like isolated tones, two-tone inter­
note its unique timbre, and make a mid­ that (Figure 3b). Thus, if the hmdamen­ vals are normally described in terms

312 American Scientist, Volume 96

of their fundamental tones. But when
a piano player strikes two notes on
the keyboard, a smorgasbord of upper
partials enters into the listener's ears
(see Figure 3c).
Beginning with Hermann Helm­
holtz in 1877, several generations of
experimentalists have studied the per­
ception of consonance or dissonance
of different intervals. They have con­
sistently found that normal listeners
hear an "unpleasant," "grating" or
"unsettled" sonority whenever two
tones are one or two semitones apart.
(One semitone is the interval between
two adjacent notes, white or black, on
the keyboard.) In addition, two tones
separated by 11 semitones are also no­
tably dissonant, despite the fact that
they do not lie close to one another on
a keyboard, and an interval of 6 semi­
tones is perceived as mildly dissonant
(see Figure 4a).
In 1965, psychologists Reinier Plomp
and Willem Levelt explained the experi­
mental perception of dissonance by us­
ing a theoretical curve (see Figure 4b) to
represent the dissonance between two
pure sine waves. This curve does not
explain the dissonance of large intervals
such as 6 or 11 semitones. However,
when Plomp and Levelt added more
and more upper partials, the "total dis­
sonance" gradually came to resemble
the empirical curve very closely. As
shown in Figure 4c, the model of Plomp
and Levelt predicts small decreases in
dissonance at or near to many of the
intervals of the diatonic scales (3, 4, 5, 7,
9 and 12 semitones).
The match between the minima of
dissonance and the tones of the most
common musical scales means that the
spacing of the tones in scales is not an
arbitrary invention. On the contrary,
it is a consequence of the way that the
human auditory system works, and
it is no surprise to see the same inter­
vals used in different musical cultures
arOlmd the world. Some tone combi­
nations have lesser dissonance, and
music that is constructed with these
less dissonant intervals is more pleas­
ing to the human ear. Of course, the
creation of "pleasant music" requires
much more than simply avoiding dis­
sonance. Indeed, some musical tradi­
tions or styles may actually encourage
dissonance. Nevertheless, the amount
of consonance or dissonance employed
will always be an important factor in
how the music is perceived.
Triads
The perception of chords-whether
they are 3-tone triads, 4-tone tetrads or Figure 2. Jean-Philippe Rameau, a French
more complex chords and cadences-is composer and author, wrote his Treatise on
likewise influenced by upper partials. Harmony in 1722, one of the first and most in­
In a triadic chord, as in a 2-tone inter­ fluential studies of harmony in Western mu­
val, the frequencies with the greatest sic. His book noted the profound emotional
difference between major and minor chords:
amplitude are usually those of the ftm­
"The major mode is suitable for songs of
damentals, the three distinct notes that mirth and rejoicing," he wrote, while the
are written in the musical score. The minor mode was suitable for "plaints, and
upper partials usually have smaller mournful songs."
amplitudes, but give the chord a rich
feeling that we might call its overall in Figure 5, with the size of the lower
"sonority." On rare occasions-such as interval shown on the vertical axis and
in barbershop quartet singing-the up­ the size of the higher interval on the
per partials may reinforce each other horizontal axis. (As before, these in­
to such an extent that they are almost terval widths are expressed in semi­
as strong as the fundamentals, and this tones.) For example, a major chord in
creates the much-coveted illusion of a "root position" has a lower interval of
"fifth voice." 4 semitones and an upper interval of 3
For simplicity, though, let us begin semitones (grid position 4-3). Any oth­
the discussion of triadic harmony by er triad in Western music can also be
considering only the fundamental specified by its location on the triadic
frequencies. The three pitches can be grid. Other musical cultures employ
plotted on a "triadic grid," as shown different scales, and may thus have

a b c
FO

II I I II nI I II .11.1 •.
mmJl!lI!ml
C C' G' Coo E"
tone 1: FO F1 F2 F3 F4 tone 1: FO F1 F2 F3 F4 F5F6
tone 2: FO F1 F2 F3 F4 F5F6

Figure 3. The fundamental frequency (FO) of a tone can be described as a pure sine wave (a). Above FO lie a number of upper partials with high­
er frequencies (Fl-F4) that contribute to our perceptions of nearly all musical tones. F1lies one octave (double the frequency) above FO and, in
this case (b), is described as C'. F2, F3 and F4 are further multiples of FO, in this case G', C" and E". When two fundamentals are included (c), the
upper partials interact in complex ways. The relative strengths of tones and upper partials are shown by amplitude bars (C =red, F = blue).

www.americanscientist.org 2008 July-August 313

 Ql Ql Ql
u U U
C
ttl
C c
ttl ttl
C
o c c
(/) o(/) 0
(/)
(/) (/) (/)
'6 '6 '6

o 1 2 3 4 5 6 7 8 9101112 o 12 3 4 5 6 7 8 9101112 0 12 3 4 5 6 7 8 9101112

a interval size (in semitones) b interval size (in semitones) c interval size (in semitones)

Figure 4. Perception of a chord as dissonant or consonant depends on the intervals between tones and upper partials (a). In empirical tests, the
dissonance reported by listeners is greatest when two musical tones are separated by one or two semitones (the distance between two adjacent
piano keys, white or black). Dissonance also peaks at an interval of 11 semitones and, to a lesser extent, at 6 semitones. Plomp and Levelt's
model explained the perception of dissonance by taking the effect of upper partials into account. They postulated a theoretical dissonance
curve (b) that peaks at 1 to 2 semitones but does not exhibit the 6- and 11-semitone peaks. When the upper partials were added (c), however, the
result began to resemble the empirical curve. Dips in this curve (that is, particularly consonant intervals) lie at or near many of the intervals of
the diatonic scales (3,4, 5, 7, 9 and 12 semitones).

chords that lie in the gaps of this grid. and beauty, as well as certain chords Dissonance in Triads
(For example, Arabic and Turkish mu­ that are simply avoided in most types Structurally, each triad contains three
sic use a scale with 24 tones in an oc­ of music. distinct intervals, so the obvious first
tave, compared to only 12 in Western The triadic grid provides a useful step in trying to explain their sonor­
music, and thus enjoy a greater variety framework for studying how the in­ ity is to add up the dissonance of
of possible harmonies.) clusion of the upper partials affects the these intervals to obtain the total dis­
Figure 5 shows various inversions of .harmonic sonority of a 3-tone chord. sonance. Figure 6a illustrates the to­
the major and minor triads, in which This framework will enable us to ad­ tal dissonance of all the triads on the
one or two notes are raised by an oc­ dress the two main questions we re­ triadic grid, taking into account only
tave. The six types of chords shown ferred to in the introduction: Why are the fundamental frequencies. The
in this figure provide the harmonic certain triads perceived as more or less figure shows two strips of relatively
framework for nearly all Western clas­ stable, and how can we account for strong dissonance, corresponding to
sical and popular music. The other the commonly perceived positive and triads that contain an interval of one
locations on the triadic grid include negative emotional valence of the ma­ or two semitones. An oblique view of
many other chords of varying utility jor and minor chords? the graph shows the dissonance even

5-3 5-4

major (second inversion)

5 4-5
4-3 (ii
~
Ql
~ 4
Qj
3:
.2 3
major (root) minor (first inversion)
2

o
o 2 3 4 5 6 7
upper interval
minor (root) major (first inversion)

Figure 5. Triads can be described as a pair of intervals on a grid. The interval between the two lower tones is plotted on the vertical axis and the
upper interval is plotted on the horizontal one. Thus a major triad (in root position) has a lower interval of 4 semitones and an upper interval of
3 semitones. A first inversion raises the lowest tone by one octave; a second inversion raises the lowest two tones by one octave. The six major
and minor triads provide the harmonic framework for nearly all Western classical and popular music.

314 American Scientist, Volume 96

more clearly. We can see an extremely 13

steep peak of dissonance when both 12

intervals are one semitone in magni­ 11

tude, and two high ridges of disso­ 10

nance when one of the intervals is less 9

~
than two semitones. The remainder 2: 8

Q)
of the triadic grid is a valley of con­ C
7

sonance-and this is where all of the ill 6

;;: 5

common triads lie. .2

When we add one set of upper 4

partials to the calculation of total dis­ 3

sonance, the "valley of consonance" 2

splits into two regions (Figure 6b). As 1

we add more upper partials, the fine 0

a
0 1 2 3 4 5 6 7 8 9 10111213

structure of the maps gradually gets upper interval

more complicated, but the general pat­

13

tern remains more or less the same

12

(Figure 6c). That is, there are regions of

11

strong dissonance (when either inter­ valleys of

10

val is small) and expanses of relatively consonance

9

strong consonance (where all of the ~

common triads lie). 2: 8

Q)
7

Clearly, an explanation of harmony .~ ,

6

~
in general cannot rely solely on the to­ Qj
;;:
tal dissonance of triads, because such a .2
5
~\~
4
Z>~
view would imply that all of the com­
3
~ ~
monly used triads have more or less '6l't.\
\
2

the same sonority. Perceptually, that c;i>/\

1
\
is simply not true. Major and minor b
0

chords are commonly described as 0 2 3 4 5 6 7 8 9 10111213

stable, final and resolved. Other triads, upper interval

even those that do not contain any 1- or

13

2-semitone intervals, are heard as tense 12

or unresolved. A study published in 11

1986 by Linda Roberts, an expert in au­ 10

ditory perception at Bell Laboratories, 9

showed that these perceptions were ~

consistent among musicians and non­ 2: 8

Q)

musicians; others have tested children .S 7

and adults, and people from the West ill 6

;;: 5

and Far East with similar results. Thus, .2

factors other than dissonance must be

3

involved in the sonority of a chord. 2

Realizing that "sensory dissonance"

1

can explain only so much, music

0

psychologists such as Sloboda, Huron, 0 1 2 3 4 5 6 7 8 910111213

David Temperley and Klaus Scherer upper interval

have assumed that normal listeners be­

come "brainwashed" to hear the major
and minor chords as stable and re­
solved, simply because they are so fre­
quently employed in all kinds of popu­
lar music. Because the other chords are
used less often, they maintain, listeners
hear them as unfamiliar, and therefore
as ambiguous, unresolved and "musi­
cally dissonant/' even though tlley are
not acoustically dissonant.
In essence, these theorists invite us
to consider all aspects of music percep­
tion as being social constructs, and to
believe in the overpowering influence
of learning and culture. The antidote
Figure 6. Sonority of triads containing intervals of 0 to 13 semitones can be depicted as a three­
dimensional surface, with dark blue indicating low dissonance and red indicating high. When
only the fundamentals are considered, small intervals exclusively produce dissonance (a). But
when one set of upper partials is added (b), the "valley of consonance" splits into two regions.
As more upper partials are added (c) the surface further subdivides but the general pattern
remains more or less the same. All of the significant chord types congregate in and around
areas of relative consonance, where M denotes major; m, minor; A, augmented; d, diminished;
and S, suspended fourth.
to such ideas is simply to play, for ex­ and unstable about it-something un­
ample, an augmented chord (CE-G#) mistakable in its acoustical structure
followed by a major (C-E-G) or minor that even people with minimal expo­
(C#-E-G#) chord ... and listen. Even sure to music hear and feel. It is that
though all of the intervals in the aug­ "common perception that music psy­
mented chord are consonant, there is chologists would like to explain on an
something inherently tense, unsettled acoustical basis.

www.americanscientist.org 2008 July-August 315

 augmented (4-4) figure is, in. essence, a visual represen­

tation of Meyer's argument from 1956,
a including its apparent flaw.
Figures 8b and 8c show how the
addition of upper partials vindicates
~ '­ Meyer's theory. Even with only one
interval 1 interval 2
upper partial, as in Figure 8b, all of
the diminished and suspended-fourth
chords lie on ridges of high tension.
When we add more partials, as in Fig­
ure 8c, we find that the major and mi­
nor chords continue to lie in blue val­
leys of stability.
toward toward
minor major
b ~-\:F----+---'----+--~7---
C The Instability of Triads
-1.0 0.0 1.0 The tension model indicates that the di­
minor (3-4) major (4-3)
difference of intervals minished, augmented and suspended­
(interval 2 - interval 1) fourth triads have high tension-in all
of their inversions and when played
over one or two octaves. Thus, the total
harmonic "instability" of triads is a con­
/' /'
interval 1

"
interval 2 interval 1
"

interval 2

Figure 7. Certain triads invoke tension, which is definitely not consonance but also not exactly
dissonance. It is a feeling the listener hopes will be resolved to stability. Tension is enhanced
when the intervals in a triad are equal. In example a, the augmented chord has lower and up­
sequence of two independent acoustical
factors. The first is intenral dissonance
and has been acknowledged to be an
important part of music perception at
least since Helmholtz's experimental
per intervals of 4 semi tones. A minor chord (b) has intervals of 3 and 4, whereas a major chord
work in the 19th century. The second
(c) has intervals of 4 and 3. (Adapted from Cook 2002.)
factor is triadic tension, which is explic­
itly a three-tone effect (Meyer, 1956).
Tension in Chords a psychophysical model of Meyer's Going one step further, we can es­
If 2-tone dissonance does not com­ theory by defining an abstract tension timate the total harmonic instability
pletely explain the sonority of a triad, curve for triads (Figure 7). The curve of any triad by adding together the
the next step is to examine the 3-tone has a peak when the two intervals in dissonance and tension factors, while
configurations of chords. In his classic the triad are equivalent. When one in­ gradually including the effects of more
book, Emotion and Meaning in Music, terval is greater than the other by at and more of the upper partials. Figure
psychologist Leonard Meyer suggest­ least a full semitone, breaking the sym­ 9 shows the results for 3-tone chords
ed that the tension in certain chords metry, the tension drops to zero. with up to four partials. Again, we see
arises from the equivalence of the two The original version of Meyer's the­ that the major and minor chords lie
intervals contained within them (for ory seems to have one flaw in it. Be­ in regions of relative stability, in all
example, 3-3, 4-4 or 5-5 semitone sides the augmented chord shown in of their inversions and when played
spacing). According to Meyer, when Figure 7, many other triads also have over one or two octaves. The other, less
a chord or short melody of three notes an unsettled, tense character-for ex­ commonly used chords lie on ridges or
contain.s two neighboring intervals of ample the so-called inversions of the peaks of instability.
precisely the same size (in semitone diminished chords. Yet these triads do This model implies that the Renais­
units), the tonal focus becomes am­ not appear to satisfy Meyer's descrip­ sance musicians of the 14th century
biguous and the music takes on an tion of intervallic equivalence. were not simply the lucky inventors
unsettled character. In other words, The resolution to this puzzle begins of a musical idiom that proved to be
the listener perceives tension because to become apparent when we bring popular. On the contrary, they were
it is tmclear how to group the equal­ the upper partials into consideration. discoverers-musicians who were sen­
ly-spaced tones. In contrast, when a In Figure 8, we have plotted the "to­ sitive to the symmetry or asymmetry
3-tone combination has two unequal tal tension" in each chord, using the in the acoustical patterns of th.ree-tone
intervals and no dissonance (that is, theoretical tension curve from Figure configurations, whereas their medieval
3-4, 4-5, 5-3, 4-3, 3-5, 5-4 semitone 7. When we compute the total tension predecessors had remained enthralled
spacing), the listener hears stability. using only the fundamental frequen­ by lower-level interval effects.
Of course, most people do not con­ cies, as in Figure 8a, we see a ridge of Questions concerning the relative in­
sciously think about the relative spacing high tension that corresponds to all fluence of intervals and chords in mu­
or "grouping" of tones. Nevertheless, the symmetric chords. The augment­ sic are still debated passionately today,
the human auditory system has evolved ed chord (A), one of the diminished but it is clearly a misunderstanding
the ability to notice it subconsciously. In chords (d) and one of the suspended to maintain that either effect alone ex­
the conclusion of this article, we will fourth chords (5) lie on this ridge; on plains harmony. When music employs
discuss one possible reason why. the other hand, inversions of the di­ intervals, consonance is the most im­
Inspired by Plomp and Levelt's ap­ minished and suspended fourth lie in portant issue, and the tuning should
proach to dissonance, we developed the blue valley of low tension. This seek the sweetest, most consonant com­

316 American Scientist, Volume 96

bination of the two tones. But when
music includes triads, the tuning of the
chord as a chord becomes the primary
perceptual event. It is then the relative
spacing of the intervals, not the location
of the tones relative to the tonic, which
becomes of central concern. Thus, the
Renaissance discoverers of harmony
achieved a shift in focus-away from
the "perfection" or "imperfection" of
intervals and toward the symmetry or
asymmetry of 3-tone configurations.
The Affective Valence of Triads
If interval dissonance and h'iad tension
were the only factors determining the
sonority of triads, we should expect
that all of the major and minor chords
would sOlU1d rather similar. Yet there
is ample evidence that they do not.
Children as yOlU1g as three years old
will associate pieces in a minor mode
with a sad face, and pieces in a ma­
jor mode with a smiling face. Casino

13

12 c

o
11 'iii
c
10 2

(ij
2: 8
OJ 7
.~
6
Q)

~ 5

Q
4

2
o 2 3 4 5 6 7 8 9 10 11 12 13
1 upper interval

0 1 2 3 4 5 6 7 8 9 10111213

a upper interval

13 . c

12 o

'iii

11 c

2
10 13~_
9 11

(ij
8 _9

2: (fJ
OJ 7 2:7
.~ 6
.2
.£5

Q)
Q
~ 5

4
13
2

o 2 3 4 5 6 7 8 9 10 11 12 13
1

upper interval
0

0 1 2 3 4 5 6 7 8 9 10111213

b upper interval

c
o
'iii
c
2
13~~
11
19~~~"~
.37
c
;:-5
ff
23

2 3 4 5 6 7 8 9 10

upper interval

123 1 2 3 4 5 6 7 8 9 10111213

c upper interval

Figure 8. The theoretical tension of triads with equal intervals is high on a simple diagonal (red) when only the fundamental frequencies are
considered (a). This explains the unsettled character of some of the suspended (S), augmented (A) and diminished (d) chords. When both the
fundamental and the first upper partials are included (b), further intersecting lines of tension appear, accounting for the high tension of the
other suspended and diminished chords. Note also the peaks in the three-dimensional view. When the fundamental and four upper partials
are considered, the "map" becomes more complicated (c) but the S, A and d chords still occupy high-tension zones. These added demarcations
also bound areas of low tension, where the major (M) and minor (m) chords lie.

www,americanscientist,org 2008 July-August 317

operators fill their casinos with slot
machines that play tones in C-major­
hoping to create a comfortable, reas­
suring acoustic environment for gam­
blers. NBC-TV's signature three-tone
cadence forms a major chord. Even
the labels "major" and "minor" sug­
gest something perceptually distinct
about these two classes of chords. In
English, French and Italian, the major/
minor distinction suggests differences
in size and strength. In German, Our
and Moll mean hard (durable) and soft
(mollify).
13
The emotional valence of major and
minor chords can of course be sup­
pressed and even reversed through
rhythms, timbres or lyrics that tell a dif­
ferent story. However, if all else is held
constant, major triads will be heard as
"positive," whereas minor chords have
a "negative" affect. That difference is
one of the longest-standing puzzles of
Western harmony. It is also one of the
most important, because the emotions
evoked by major and minor harmo­
nies help give music its meaning. They
distinguish music from the unfocused
meandering of birdsong or the cacoph­
ony of a city street.
mental frequencies leads to an overly
simple picture that misclassifies certain
chords. In Figures lOb and 10c, the tri­
ads composed of a 3-semitone interval
and a 5-semitone interval (either 3-5 or
5-3) are located in regions that are nei­
ther orange nor blue-neither major
nor minor, which contradicts what we
know from musical experience.
When we include the upper partials,
however, the total valence scores are
remarkably consistent with our per­
ceptions of major and minor triads.
Even with only the first set of upper
partials (Figure lOd), we find peninsu­
las of positive (orange) and negative

We have seen that the relative size (blue) valence at all of the major and
12

of the two intervals was the key to lm­ minor triads. The tension chords (d, A,
11

derstanding triadic tension. Moreover, S), on the other hand, fall in between
10

9
from a state of "intervallic equiva­ the regions of positive or negative mo­
lence" (with its inherent perceptual dality, as would be expected from tra­
~8 tension), there are only h.vo directions ditional harmony theory.
~7
c
": 6 of pitch movement that can reduce the Thus, among the upper partials of all
Q)
tension. Either the lower interval can of the major chords, there is a predom­
~ 5
-4 be made greater than the upper inter­ inance of triadic structures where the
3 val, in which case the chord resolves to lower interval is one semitone larger
2 a major triad, or the lower interval can than the upper interval. Minor chords
1 be made smaller, which corresponds to show the opposite structural feature.
o a minor triad. The brain could, in theory, identify the
o 1 2 3 4 5 6 7 8 9 10 11 1213 This reasoning suggests that we major or minor nature of a chord sim­
a upper interval should reformulate the tension mod­ ply by summing the valences of all
el so that the direction of motion away the possible three-tone combinations
13,.....,..........,~......

from symmetry indicates the degree of of partials.

12
"majorishness" or "minorishness" of
11
any 3-tone chord. Conclusion

10
9
Thus, in Figure lOa, we propose a Now that we have a model of how lis­

modality curve to distinguish the two teners identify a chord as major or minor,

~8
~ 7
types of resolution. The horizontal we may take the final step and speculate

c
": 6 axis shows the difference between the as to why the acoustical valence carries

Q)
two intervals in the chord, in semitone an emotional valence as well.

~ 5
-4 units. When there is no difference, the We contend that the emotional sym­
3 chord is ambiguous, and its "valence bolism of major and minor chords has a
2 score" on the vertical axis is zero. The biological basis. Across the animal king­
1 valence score rises or falls to a maxi­ dom, vocalizations with a descending
o mum or minimum when the difference pitch are used to signal social strength,
o 1 2 3 4 5 6 7 8 9 10 11 1213 between the intervals is exactly 1.0 or aggression or dominance. Similarly, vo­
b upper interval -1.0 semitone (points a and b). The va­ calizations with a rising pitch connote
lence score falls to zero again if the dif­ social weakness, defeat or submission.
13 .--.---:-:-­
ference is two or more semitones. Of course, animals convey these mes­
12
As in the dissonance and tension sages in other ways as well, with fa­
11
models, considering only the funda- cial expressions, body posture and so
10
on-but all else being equal, changes in
9
Figure 9. The total harmonic instability of the ftmdamental frequency of the voice
~8 any triad can be shown by adding together have intrinsic meaning.
~ 7
c dissonance and tension factors. Once again, This same frequency code has been
": 6
Q) the picture is relatively simple when only FO absorbed, though attenuated, in hu­
~ 5 and Fl are included (a), with major and mi­
- 4 man speech patterns: A rising inflec­
nor chords lying in low dissonance/tension
3
areas (blue). But when four upper partials
tion is commonly used to denote ques­
2
are added to the picture (b), major and minor tions, politeness or deference, whereas
1
chords continue to congregate in low zones, a falling inflection signals commands,
o whereas suspended (5), augmented (A) and statements or dominance. How might
o 1 2 3 4 5 6 7 8 9 10 11 1213 diminished (d) chords (c) lie in areas of high this translate to a musical context?
c upper interval instability (red). If we start with a tense, ambiguous

318 American Scientist, Volume 96

 13 ,-,r--;--r.....,..-r--;--,--r-~-;­
12
11
10
(ij9
~8
30 ~7
Qi 6
~5
4
3
difference of intervals 2
in semitones 1
0
0 1 2 3 4 5 6 7 8 9 10111213
a b upper interval c

Bibliography
Cook, N. D. 2002. Tone of Voice and Mind. Am­
sertdam: Benjamins.
Cook, N. D. 2007. The sound symbolism of
major and minor harmonies. Music Percep­
tion 24(3):315-319.
Cook, N. D., and T. X. Fujisawa. 2006. The psy­
chophysics of harmony perception: Harmo­
ny is a three-tone phenomenon. Empirical
Musicology Review 1(2):106-126.
2 3 4 5 6 7 o
2 3 4 5 6 7 8 9 10 111213 Cook, N. D., T. X. Fujisawa and K. Takami.
d upper interval e
upper interval 2006. Evaluation of the affective valence
j
of speech using pitch substructure. IEEE
Figure 10. Interval dissonance and triad tension separate chords into categories but fail to Transactions on Audio, Speech and Language
explain why major and minor chords don't sound similar. The authors propose that the direc­ Processing 14:142-151.
tion of motion away from interval symmetry <tension) determines the degree to which a chord Helmholtz, H. L. F. 1877. On the Sensations of
is "majorish" or "minorish." If the two intervals of a triad are equal, both the difference in Tone as a Physiological Basis for the Theory or
intervals and the "valence score" (vertical, or modality, axis) are zero (a). As the lower interval Music. Mineola, N.Y.: Dover Publications,
minus the upper interval approaches +1, the modality approaches +1, or major. As the lower 1954.
interval minus the upper interval approaches -1, the modality approaches -1, or minor. When Kastner, M. E, and R. G. Crowder. 1990. Per­
only the fundamentals of the triad are considered, the valence values are as shown in (b) and ception of major / minor: IV. Emotional con­
(c), and only four of the six major and minor chords have appropriate valence scores. But when notations in young children. Music Percep­
F1 is added (d), all of the minor triads have negative valence scores and all of the major tria~s tion 8:189-202.
have positive valence scores (red). On the expanded triadic grid, when Fo-F4 are included (e) Meyer, L. B. 1956. Emotion and Meaning in Mu­
all 12 of the major and minor chords have positive and negative valence scores, respectively. sic. Chicago: Chicago University Press.
Narmour, E. 1990. The Analysis of Cognition of
Basic Melodic Structures. Chicago: Chicago
chord-for example, the augmented have taken this historical trend to its University Press.
chord containing two 4-semitone in­ logical extreme, and produced music Plomp, R., and W. J. M. Levelt.1965. Tonal con­
tervals-and decrease anyone of the that fanatically avoids all indications sonances and critical bandwidth. Journal of
three fundamentals by one semitone, of consonance or harmonic resolution. the Acoustical Society of America 35:548-560.
the chord will resolve into a major Such surprisingly colorless "chromat­ Roberts, L. 1986. Consonant judgments of mu­
key. It will then have a 5-4, 3-5, or 4-3 ic" music is intellectually interesting, sical chords by musicians and untrained
listeners. Acuslica 62:163-171.
semitone structure. Conversely, if we but notably lacking in the ebb and
Sethares, W. A. 199. Tuning, Timbre, Spectrum,
resolve the ambiguous chord by rais­ flow of tension and resolution that Scale. Berlin: Springer.
ing anyone of the three fundamen­ most popular music employs, and that
tals by a semitone, we will obtain a most listeners crave. Whatever one's
minor chord. The universal emotional own personal preferences may be for
response to these chords stems, we be­ dissonance and unresolved harmonies,
lieve, directly from an instinctive, pre­ some kind of balance between conso­
verbal understanding of the frequency nan.ce and dissonance, and between
code in nature. One of us (Cook) has harmonic tension and resolution,
explored this in more detail (see the seems to be essential-genre by genre, For relevant Web links, consult this
bibliography). and individual by individual-to as­ issue of American Scientist Online:
Individual tastes and musical styles sure the emotional ups and downs that http://www.americanscientist.orgl
vary widely. In the West, music has make music satisfying. Issue TOC/issue/11 01
changed over the centuries from
styles that employed predominantly Acknowledgement
the resolved major and minor chords Thanks for critical comments on the visual
to styles that include more and more display of harmony go to the participants
dissonant intervals and unresolved at the Image and Meaning Workshop 2.4
chords. Inevitably, some composers at Harvard, October 26-27, 2007.

www.americanscientist.org 2008 July-August 319