Article

Psychology of Music
1–15
Perception of six basic © The Author(s) 2010
Reprints and permission: sagepub.
emotions in music co.uk/journalsPermission.nav
DOI: 10.1177/0305735610378183
http://pom.sagepub.com

Christine Mohn
Department of Psychology, University of Oslo, Norway and Vestre Viken Hospital Trust, Norway

Heike Argstatter
German Center for Music Therapy Research (Viktor Dulger Institute), Heidelberg, Germany

Friedrich-Wilhelm Wilker
Department of Music Therapy, University of Applied Sciences, Heidelberg, Germany

Abstract
A test of the ability to perceive six basic emotions (happiness, anger, disgust, surprise, sadness, and
fear) in music was presented to 115 participants. Eighteen musical segments, lasting 3–5 seconds,
were designed for this test using a variety of solo instruments. The results show that six basic
emotions are perceivable in musical segments previously unknown to the listeners, although there
was large variability in the percentage of correct classification of each of the segments comprising
each emotion, and happiness and sadness were easier to classify than the other emotions. Moreover,
the ability to classify musical emotions was not related to childhood or youth musical instruction or
personality traits (assessed by NEO-PI-R).

Keywords
emotions, music, personality

Introduction
According to the neuro-cultural theory of Paul Ekman (Ekman and Friesen, 1971; Ekman,
Levenson, & Friesen, 1983), the six emotions – happiness, anger, disgust, surprise, sadness,
and fear – are associated with separate autonomic activation patterns and facial expressions.
Moreover, the ability to identify these emotions through facial expression seems to be universal
among humans (Ekman, 1992; Elfenbein and Ambady, 2002), suggestive of evolutionary
theory strongly contributing to explaining the origins of emotion. Specifically, facial emo-
tions are assumed to have evolved in order to allow rapid communication of danger or safety
(Ekman, 1993).

Corresponding author:
Christine Mohn, Department of Pscyhology, University of Oslo, PO Box 1094, Blindem, 0317, Oslo, Norway.
[email: christinemohn@hotmail.com]
2 Psychology of Music

Ekman suggests, in accordance with Tomkins (1963), that each of the emotions that may be
expressed facially also has a vocal expression (Ekman, 1993). This notion is important regard-
ing individuals whose visual recognition of emotions is impaired, for example because of
permanent blindness or situationally restricted vision, who may still be able to judge the
emotional content of their communication partners through characteristics of speech, such as
loudness, rapidity, and prosody. Moreover, where vision is unobstructed, the recipient may be
unsure of what to believe if the verbal content of the message diverges from the emotional con-
tent being transmitted, for example if the message ‘Everything is OK’ is spoken in a shrill, rapid
manner. There is some evidence that we tend to focus on the emotional tone of the voice rather
than the verbal content of the message in cases where these modes of communication are
incongruent (Powers and Trevarthen, 2009).
Emotion regulation seems to be one of the primary reasons for the use of music in everyday
life (Juslin and Laukka, 2004). Several studies have reported that humans are very adept at
identifying the emotional content of music (Sloboda and Juslin, 2001). This ability may be so
pivotal for human communication that it is manifested early in childhood (Nawrot, 2003). For
example, Dalla Bella, Peretz, Rousseau and Gosselin (2001) reported that 5-year-old children
are able to discriminate between happiness and sadness using information about tempo, and
that a year later they start taking mode into consideration when making the same classification.
However, most of the studies on emotion recognition in music have concentrated on a very
limited number of emotions, usually the happiness–sadness dichotomy. It seems that several
other emotions may be identified acoustically. A survey of the emotion ratings made by music
professionals demonstrated that all of the six basic emotions seem to be represented in Western
art music, although happiness and sadness are probably much more common as musical
themes than – in decreasing order of importance – anger, fear, surprise, and disgust (Kallinen,
2005). In addition, Vieillard et al. (2008) reported that happiness, sadness, fear, and peaceful-
ness may be readily recognized in relatively short stimuli (9–16 seconds). The present study
aims to extend the work of Vielliard et al. (2008) by assessing the ability to detect all six basic
universal emotions (Ekman, 1992) in unfamiliar musical stimuli.

Personality traits and emotion perception in music

Individual differences may influence the perception of emotions in music. A study of participants
with unknown musical experience demonstrated a significant relationship between emotional
intelligence and the ability to identify the correct emotions of tempo- and loudness-manipulated
excerpts of Western art music (Resnicow, Salovey, & Repp, 2004). Moreover, several studies have
been performed on the relationship between personality traits and artistic interests (McAdams,
2006). Most seem to conclude that openness (O) is the most relevant trait, in that individuals high
in O tend to value fantasies, dreams, artistic leisure activities, and creative and intellectual types
of work (McAdams, 2006). In addition to a positive relationship between O and self-reported
creativity, Furnham, Zhang, & Chamorro-Premuzic (2006) found a significant negative associa-
tion between conscientiousness (C) and creativity.
In a study of the relationship between personality traits and music preferences, Rawlings
and Ciancarelli (1997) reported that high levels of extraversion (E) corresponds to a preference
for popular music (such as pop, rock, and ‘easy listening’ music), while high levels of O were
related to a wide range of preferences (rock, jazz, folk, classical, and electronic music). A series
of six studies (Rentfrow and Gosling, 2003) revealed a robust underlying personality structure
of preferences. A preference for cheerful, upbeat (often vocal) music was positively correlated
Mohn et al. 3

with E, and a preference for reflective and complex (often instrumental) music was related to O.
Not only preferences for broad genres of music but also for musical elements and structure may
be influenced by personality traits. Kopacz (2005) found that, regardless of genre, both extra-
verted and open-minded and creative individuals tended to prefer a high number of melodic
themes. In addition, extraverted individuals preferred fast tempos.
North and Hargreaves (2008) noted that different types of music may compensate for
certain aspects of an individual’s personality. Extravert individuals may prefer upbeat music
with a fast tempo in order to provide strong stimulation of a brain characterized by relatively-
low baseline arousal, while introvert people, assumed to display enhanced cortical activity at
baseline, may prefer calmer music in order to enjoy themselves. Moreover, music preference
may reflect aspects of personality (North and Hargreaves, 2008). Individuals high in O may
prefer complex music simply due to their being creative and intellectual.
To the best of our knowledge, only one study exists on the influence of personality traits
on musical emotions classification accuracy. Chamorro-Premuzic and Furnham (2007)
reported that, while high neuroticism (N) was correlated with the self-reported tendency to
use music for mood regulation rather than intellectual stimulation, no trait correlated with
the self-reported ability to classify musical styles and composers correctly. However, this
study did not present the participants with sound stimuli to be classified, and no listening
procedure took place. Hence, the investigation of the relationship between personality traits
and emotion recognition in musical segments previously unknown to the participants seems
warranted.
Moreover, high levels of E seem to correlate with social competence and the ability to predict
and detect emotional reactions in others (Costa and McCrae, 1992). In this respect, one should
expect individuals high in E to perform well on tests of emotion perception. However, there
seems to be a general tendency for individuals high in E to emphasize positive affective experi-
ences (Costa and McCrae, 1992). It is therefore possible that E may bias individuals into overes-
timating the presence of positive emotional stimuli. The present study will therefore investigate
the effect of personality traits on the ability to classify emotional stimuli.

Rationale of the present study

In order to explore the possibility of six basic emotions being recognized in acoustic stimulation
unknown to the participants, the German Center of Music Therapy Research developed a test
of emotion perception in music (Busch et al., 2003). A pilot study was used to validate the use
of 18 musical segments (described later) for studies of perception of emotions in music. This
study demonstrated that these six emotions were identifiable in musical stimuli, and that there
was no significant difference in detection accuracy between music therapy students and controls
without musical or psychological training (Busch et al., 2003).
The sample size of the pilot study (Busch et al., 2003) was relatively small (eight patients
with mental illness, 18 music therapy students, and 20 control subjects). Therefore, the present
study aims to put these preliminary results on a more secure footing by testing a larger number
of participants as well as investigating the effects of personality traits on the identification
of emotions.
Due to the exploratory nature of this study, no specific hypotheses were formulated. The fol-
lowing research questions were asked: (1) Are the six basic universal emotions perceivable in
music unknown to the listeners?; and (2) Are personality traits related to the perception of six
basic universal emotions in music?
4 Psychology of Music

Methods
Subjects and procedure
The participants were 115 undergraduate and graduate students (N = 41, 35.7% males,
N = 74, 64.3% females) recruited through advertisements and email lists at the University of
Oslo. Exclusion criteria were (self-reported) hearing loss and inability to understand spoken and
written Norwegian. Demographic characteristic and data on music education and listening
habits are presented in Table 1.
The participants were tested individually or in groups of up to four in a non-soundproof
classroom at the Department of Psychology, University of Oslo. The music segments were
played on a ghetto blaster placed on a table that was 3 metres from the participants. The sound
volume at the position of the participants was kept constant at 60 dB. After the musical emo-
tion test, lasting 10 minutes (see later), the subjects filled in questionnaires on demography and
personality traits. The entire session, consisting of the music emotion test and the question-
naire completion, lasted one hour. This study was approved by the Regional Committee for
Research Ethics (REK-Sør), and all participants signed consent forms prior to the test.

Test of emotion perception in music

The current test of perception of emotions in music was the same as the one used in our pilot
study (Busch et al., 2003). Twenty-four music segments (six trial segments and 18 test seg-
ments) were composed by members of the academic staff at the German Center for Music
Therapy Research in Heidelberg, Germany. The segments, all tonal, were intended to represent
the six basic emotions, happiness, anger, disgust, surprise, sadness, and fear, and were recorded
in a studio. The segments broadly represent classical or jazz music. The musical stimuli were
provided by three musicians (pianist, percussionist, and cellist) and four music therapists. The
performers played their favourite instruments out of a free choice of instruments. They made no
account for their choice. Every performer was instructed to improvise short musical pieces on
the six basic emotions in a way that a listener should be able to decode the intended emotion.
This procedure led to a great variety of possible improvisations. Stimuli not fulfilling the formal
criteria (for example lasted more than 10 seconds or were of pure acoustic quality with no
melodic line) were eliminated. For the remaining musical excerpts, distinctiveness and typicality

Table 1.  Demography and music habits of the participants

Agea 24.7 (SD 7.2) years

Years of music instruction as child/youtha 6.1 (SD 5.1) years
Listening to music
   2–6 times / week 15.7%
   Daily 80.9%
Preference for music type
   Classical 8.7%
   Popular 30.4%
   Both classical and popular 60.9%
Number of concerts attended last 12 monthsa 11.3 (SD 16.4)
a Numbers in mean. Classical music: classical and opera music. Popular music: rock, pop, blues, jazz, techno, and hip
hop music.
Mohn et al. 5

were rated by the improvising performers and additionally by five trained music therapists. This
led to the final set of segments.
Apart from the musical instruments, the equipment consisted of two large-diaphragm
microphones, one hard disc recorder, and a mixing board. The recorded material was trans-
ferred from the hard disc recorder to two computers, and edited in a digital sound studio. The
segments were transferred to a compact disc (CD) with the order of presentation randomized
(Table 2). There was a 10 second pause between each segment.
All participants listened to the same CD and were thus exposed to the segments in the same
order. The participants were instructed to try to classify each segment as one of six emotions

Table 2.  Characteristics of the musical segments (N = 115)

Number Emotion Duration Instrument Musical characteristics

 1 Fear 1 4 seconds Cello Short, “shivering” vibrato, low volume, fast

tempo
 2 Happiness 1 3 seconds Tuba Vivid expression, staccato, broad timbre, high
volume, fast tempo
 3 Sadness 1 5 seconds Electric bass Legato, light, subdued ascending and
descending tones, slow tempo
 4 Disgust 1 5 seconds Violin “Schreeching”, medium volume, several
variations with changing expression and
emphasis
 5 Anger 1 3 seconds Piano Hard touch, staccato, loud volume, rapidly
ascending tempo, dissonant harmony
 6 Surprise 1 4 seconds Electric bass Short tones, staccato, jumping ascending
dynamics, medium volume
 7 Sadness 2 5 seconds French horn Minor modea, stepwise intervals, weak touch,
medium volume, consonant harmony
 8 Disgust 2 5 seconds Cello Uncontrolled tones in rapid succession,
ascending and descending movements
 9 Fear 2 4 seconds Guitar Very rapid touch, ascending volume, tempo,
and dynamics
10 Anger 2 3 seconds Tuba Staccato, low pitch, short intervals between
tones, loud volume
11 Surprise 2 4 seconds Piano Major mode, jumping, ascending melody,
broad expression, crescendo
12 Happiness 2 5 seconds Guitar Major mode, dance-like 3/4 rhythm, large
intervals, loud volume, no dissonances
13 Surprise 3 3 seconds French horn Major mode, staccato, jumping ascending
melody, medium volume, crescendo
14 Disgust 3 3 seconds Electric bass Weak touch, subdued timbre, slow tempo, low
volume, diminuendo
15 Fear 3 5 seconds Tuba Rapid, unregular vibrato, low pitch, medium
volume, from crescendo to decrescendo
16 Happiness 3 5 seconds Piano Major mode, strong timbre, vivid expression,
rapid tempo
17 Anger 3 5 seconds Cello Minor mode, staccato, low pitch, strong
vibrato, rapid tempo
18 Sadness 3 5 seconds Piano Minor mode, weak touch, low volume, slow
tempo with large variations
aMode is described only for those segments that a clear major or minor melody is identifiable
6 Psychology of Music

and to mark the most appropriate emotion category on a questionnaire. They were informed
that they had to make a decision and that they had to choose only one emotion for each segment.
Before the test, the subjects familiarized themselves with the task by trying to identify one
segment each of the six emotions. These trial segments were different from the segments used
in the test. The entire musical emotions test procedure lasted 10 minutes.

Demographic characteristics and music preferences

After the musical emotion test, the participants filled in a form with questions regarding gender,
age, general education, and music experiences and preferences. They were asked to state:
(1) whether they had received music instruction on childhood or youth and whether they were
still singing or playing an instrument (if so, which type of instrument and number of years of
instruction); (2) how often they listened to music at home; (3) number of concerts they had
attended during the last 12 months; and (4) which type of music they preferred. Preference for
music could be classical instrumental music and opera (referred to as the category ‘classical
music’), or rock, pop, blues, jazz, techno, and hip hop music (referred to as the category ‘popu-
lar music’). The participants were allowed to state several preferences, and in 60.9% of the
cases, both classical and popular music were preferred (Table 1).

Personality traits
The revised version of the NEO Personality Inventory (NEO-PI-R, Costa and McCrae, 1992;
Norwegian version by Martinsen, Nordvik, & Østbø, 2003) was administered to assess the
personality traits neuroticism, extraversion, openness, agreeableness, and conscientiousness.
The NEO-PI-R is one of the most widely-used personality inventories. It is reported to have high
reliability and validity, and has been validated both cross-culturally, by self-ratings, and by
ratings by peers and spouses (Wiggins, 1996). The items of the questionnaire are presented as
statements, for example ‘I am not a person that worries’, or ‘I like being surrounded by people’.
Responses are made on a five-point Likert scale ranging from ‘strongly disagree’ to ‘strongly
agree’. The full version of the NEO-PI-R, consisting of 240 items, was used in this study.

Results
All statistical tests were conducted using SPSS for Windows, release 14. Chi-square tests, t-tests,
and correlation analyses were two-tailed.

Perception of musical emotions

Percentages of correct and incorrect classification of musical emotions in the 18 segments are
given in Table 3. In order to test whether there were statistically-significant differences in clas-
sification rate across the three segments of each emotion, repeated measures analysis of vari-
ances (ANOVAs) with post-hoc comparisons were performed with scores calculated as
percentage correct hits. These analyses revealed that anger 1 was significantly more difficult to
classify than anger 2 and anger 3 (F = [2] 14.92, p < .001). Disgust 1 was significantly easier
to classify correctly than the other two examples of this emotion (F = [2] 32.87, p < . 001). All
three segments of surprise were significantly different from each other in terms of classification
accuracy (F = [2] 56.27, p < .001). Fear 3 was significantly more difficult to classify as the
other to fear segments (F [2] = 76.59, p < .001).
Mohn et al. 7

Table 3.  Recognition and confusion of emotion in musical segments (N = 115)

Perceived  emotion

Happiness Anger Disgust Surprise Sadness Fear

Expressed emotion
Happiness 1 70.0% 2.6% 7.8% 15.7% 3.5% 0.0%
Happiness 2 83.5% 0.9% 0.0% 15.7% 0.0% 0.0%
Happiness 3 93.3% 0.9% 1.7% 3.5% 0.0% 0.0%
Anger 1 20.0% 25.2% 3.5% 41.7% 0.9% 8.7%
Anger 2 4.1% 59.1% 24.3% 3.5% 5.2% 3.5%
Anger 3 6.1% 47.0% 7.0% 1.7% 7.8% 30.4%
Disgust 1 0.0% 9.6% 70.2% 2.6% 6.1% 11.4%
Disgust 2 2.6% 28.7% 30.4% 3.5% 2.6% 32.2%
Disgust 3 1.7% 7.0% 25.2% 5.2% 47.0 % 13.9%
Surprise 1 16.5% 0.9% 3.5% 78.3% 0.0% 0.9%
Surprise 2 83.3% 0.0% 0.0% 16.7% 0.0% 0.0%
Surprise 3 52.2% 0.0% 0.0% 46.1% 0.0% 1.7%
Sadness 1 7.0% 0.9% 13.0% 11.3% 65.2% 2.6%
Sadness 2 0.0% 0.0% 0.0% 0.0% 100% 0.0%
Sadness 3 3.5% 0.0% 0.0% 0.9% 95.7% 0.0%
Fear 1 0.0% 7.0% 8.7% 3.5% 0.9% 80.0%
Fear 2 0.9% 14.8% 1.7% 7.0% 0.0% 75.7%
Fear 3 0.9% 38.3% 34.8% 1.7% 5.2% 19.1%
Note: Bold numbers represent the accuracy of perceived emotion. Non-bold numbers represent false hits. The
numbers may not add to 100% because of rounding.

Visual inspection of Table 3 suggests that confusions of classification, such as instances

where anger is incorrectly perceived as fear, may be statistically-significant. Therefore, Chi-
square analyses were performed on the nominal data representing the scores of the test,
comparing correct hits with false hits. This procedure revealed that anger 1 was significantly
misclassified as surprise (c2 [1] = 4.69, p < .05), disgust 3 was significantly misclassified as
sadness (c2 [1] = 7.53, p < .01), and that fear 3 was significantly misclassified as both anger
(c2 [1] = 8.14, p < .01) and disgust (c2 [1] = 5.92, p < .05).
Next the three musical segments representing each emotion were combined into six musical
emotion indices by simple aggregation and the results of the identification accuracy given in
percentage correct answers (Table 4). A repeated measures ANOVA revealed that happiness

Table 4.  Perception of musical emotions: indices (N = 115)

Musical emotion index Mean (SD) percent

Happiness 82.6 (22.2) %

Anger 43.8 (27.7) %
Disgust 41.9 (25.0) %
Surprise 47.0 (24.6) %
Sadness 87.0 (18.0) %
Fear 58.3 (22.9) %
Note: Each index represents the mean accuracy detection score of three musical segments.
8 Psychology of Music

and sadness were significantly easier to classify correctly than the other four emotions, happi-
ness was as easy to classify correctly than sadness was (F = [5] 84.14, p < .001).
The number of females was significantly higher than the number of males (c2 [1] = 9.47,
p < .01). Therefore, a series of independent-samples t-tests with gender as the grouping vari-
able and the indices as dependent variables were run. The results yielded no effects of gender
in the ability to classify musical emotions.
Although our previous study did not find significantly different detection accuracy
between music therapy students and controls without musical training (Busch et al., 2003),
others (Bigand, Vieillard, Madurell, Marozeau, & Dacquet, 2005) have reported a small effect
due to musical experience. Therefore, Pearson’s correlation analyses of the association
between musical instruction in childhood/youth and the emotion indices revealed a statisti-
cally-significant relationship between years of music instruction and the classification of
happiness (r = .28, p < .01). When running logistic regression tests on the three items of the
happiness index, it was revealed that only happiness 1 (played by a tuba) was related to years
of music instruction (β = .14, p < .01).
Finally, a series of correlation analyses were run aiming to test the relationship between the
number of concerts attended in the last 12 months and the musical emotion indices. These
tests generated no statistically-significant results, and this relationship was not studied further
in regression analyses.

Personality traits and perception of musical emotions

Hierarchical regression analyses were used to test the association between personality traits
and the classification of musical emotions. In each regression model, one of the six musical
emotion indices was the dependent variable, and the five personality traits of NEO-PI-R (Costa
and McCrae, 1992) were the independent variables. This procedure revealed a significant con-
tribution of O to the perception of happiness when O was entered in combination with N and E
(β = .28, p < .01) and N, E, and A (β = .28, p < .01), respectively.

Discussion
The main finding was that the participants were able to identify six basic universal emotions in
completely unknown musical stimuli, with happiness and sadness easier to classify correctly
than the other emotions. In addition, there was great variability in the means of percentage
recognition, indicating that some of the musical stimuli should be altered or substituted in
order to convey the intended emotion in a better way. Moreover, the ability to perceive emotions
in music was not more than marginally related to music instruction in childhood or youth and
the personality trait openness (O).

Perception of musical emotions

The participants of the present study correctly classified the emotional content of unknown
musical stimuli, in most cases well above chance level. Thus, we have replicated the results of a
pilot study (Busch et al., 2003) with a larger number of participants, putting the previous find-
ings on a more secure footing.
In the present study as well as in the pilot study (Busch et al., 2003), sadness and happiness
were more readily identifiable than the other emotions, and disgust, surprise, and fear the least
Mohn et al. 9

identifiable. These results are consonant with those of Kallinen (2005), who argues that
sadness and happiness are the emotions most often expressed in Western art music. Sadness
and happiness are relatively easy to express musically, due to their fairly consistent different
characteristics in terms of mode and tempo. Moreover, most Westerners learn at an early age to
associate slow music in the minor mode with sorrow and faster music in the major mode with
joy due to the large amount of exposure such pieces of music get relative to pieces expressing
other emotions or pieces with less clear-cut emotional content (Kallinen, 2005). In this respect,
it must be kept in mind that the musical segments of the present study were based on Western
tonality; even though they were unknown to the participants, their emotional content may
have been recognized because the participants were all Westerners familiar with Western
tonality. However, this possibility does not preclude an interpretation of the present study in
terms of the neuro-cultural theory of emotion (Ekman and Friesen, 1971). This theory is not
completely universalistic, but postulates that cultural factors contribute to the recognition and
display of emotions, as evidenced by recognition rates of facial emotions rarely reaching 100 %
(Ekman, 1994). If each facial emotion may be expressed by sound (Ekman, 1993), one may
assume that the ability to classify emotions through sound is a characteristic of all humans, but
that the recognition rate will be highest when the relevant sound stimuli contain certain
aspects characteristic of one’s particular cultural auditory environment.
In contrast to sadness and happiness, which are either negative or positive emotions,
surprise may be difficult to identify musically because this emotion may be regarded as both
positive and negative (Kallinen, 2005). In addition, the classification of disgust may be difficult
because this emotion is unexpected within the framework of Western tonality, as Western
music traditionally has been used for individual enjoyment or to foster social cohesiveness
(Kallinen, 2005).
In cases where the recognition rate was less than 50% for the negative emotions anger,
disgust, and fear, they were most often mistaken for another negative emotion, for example
anger being misclassified as fear. Possibly, this may be the result of the fear and anger segments
sharing several musical elements, such as loudness, rapid tempo, and changes in dynamic.
Alternatively, the evolutionary purpose of emotions may offer an explanation. Both fear and
anger are elicited in dangerous situations requiring rapid action. Perhaps rapid action, such as
fight or flight, in such situations is so important to our survival that it is difficult for us to ponder
the finer points separating the strong, negative emotions.
Anger 1 was often mistaken for surprise, an emotion that is not necessarily negative.
According to Juslin (2001), anger may be expressed musically through a high sound level,
staccato articulation, and a fast mean tempo. These characteristics could easily be perceived in
anger 1. However, anger 1 did not contain the relatively-large variations in tone duration that
may be an additional prerequisite for anger to be expressed in music (Gabrielsson and Juslin,
1996). Possibly, variation in tone duration is not necessary for anger to be perceived providing
the other characteristics are present.
A different explanation may come from those claiming that music is incapable of inducing
or conveying all the discrete, universal, strong emotions similar to those exhibited by the human
face. Scherer (2004) argued that the tendency for cognitive appraisal of aesthetics renders the
musical experience an inherently subjective and private one, and that more subtle, changing
feelings are associated with music to a much larger degree than is a fixed set of categories of
emotions with strong action tendencies. This line of thought may account for the present
difficulty in classifying fear and anger correctly. These emotions may simply not be detected in
a musical context because they are not expected, as listening to music is a pleasurable activity
10 Psychology of Music

for most people. On the other hand, others (Juslin and Laukka, 2004) have demonstrated that,
relative to more complex affective states, the basic emotions, including anger and fear, were
more frequently chosen as examples of what music may express. However, giving examples of
emotions that may be expressed in music in general is not the same thing as having to judge
the emotional content of a specific piece of music one is hearing for the first time, as our partici-
pants were instructed to do. In this respect, it is not possible to conclude that the present study
provides definitive support for the theory that the perception of emotions in music is best viewed
within a universalistic, categorical framework.
Participants who had received music instruction in childhood or youth were significantly
better at classifying one of the happiness segments. This segment was played by a tuba, and it is
possible that music played by this instrument is somewhat difficult to interpret emotionally for
musical novices. Compared to other instruments, there are very few pieces written for tuba
solo, and the tuba usually enters orchestral pieces during forte segments or when the composer
wishes to emphasize the bass line. Thus, most musical novices may have learned to associate
the timbre of the tuba with darker or perhaps scary musical segments, while experienced musi-
cians may have learned to pay more attention to the mode and tempo of the segment instead of
the instrument on which it is performed. On the other hand, the tuba features prominently in
brass band music, a popular genre in Norway, where many children and teenagers receive
extra-curricular music instruction in school bands. In this respect, the tuba could be associated
with joyful experiences. In contrast to tempo, loudness, and mode, the emotion-expressing
aspects of timbre is not well explored (Gabrielsson and Lindström, 2001). Therefore, in the
absence of studies of the relationship between the timbre of a large number of instruments and
the emotions perceived in the listeners, this explanation remains speculative. Moreover, the
music preferences of the present participants seemed broad, in that most of them stated a pref-
erence for both classical and popular music. Hence, they seemed to have been exposed to a large
variety of music genres and were presumably able to determine the emotional effects of different
musical cues.
The accuracy of recognition of the emotion indices in our study was less than typically
found in studies of facial emotions, which normally range from 97% (happiness) to 67% (anger)
(Ekman, 1992; Elfenbein and Ambady, 2003). The accuracy of sadness recognition, however,
was in line with studies of facial emotions (87%; Ekman, 1992; Elfenbein and Ambady, 2003).
One explanation may be that emotion is more difficult to identify in acoustic than in visual
stimuli, as the visual system is the most advanced of the human senses. A second possible
explanation concerns the different characteristics of facial emotion stimuli (Ekman, 1992) and
musical emotional stimuli. The facial emotions are spatial, and we are immediately able to
perceive the totality of the face and the characteristics of the different facial structures produc-
ing the emotional expression. In contrast, music has a prominent temporal characteristic and
is perceived as a sequence of sounds relating to each other. In our study, the maximum length
of a musical segment was 5 seconds. While there is evidence that happiness, sadness, and fear
are recognizable in musical segments of less than 4 seconds in duration (Vieillard et al., 2008),
emotions with less strong action tendencies, for example surprise, may require stimuli of longer
duration to be identified.
A third explanation may be that emotions are as easily recognizable in music as in facial
expressions, but that this first version of the test contains items that are not sufficiently repre-
sentative of the emotions they intend to express. According to Ekman (1993), each emotion
correlates with unique physiological responses. Happiness, for example, is conveyed facially by
increased muscular activity around the lips and eyes. According to Juslin (2000), it is possible
Mohn et al. 11

to isolate musical parameters that symbolize happiness, sadness, fear, and anger. A future task
for our laboratory is to identify the parameters, such as mode, tempo, timbre, volume, and
dynamics, of the stimuli that uniquely relate to disgust and surprise.
Fourth, although Ekman (1993) postulates that each emotion expressed by the face may be
expressed through sound, his focus is the human voice. The expression of emotion though art
music may, in this perspective, be less related to the need for rapid communication. Therefore, it
may be harder to identify the emotional content in instrumental music than in facial expres-
sions. A way to test this hypothesis would be to develop a vocal version of the musical emotions
test, in which the emotional segments are sung rather than played. There is evidence that
happiness, sadness, anger, and fear are readily detected in spoken messages (Laukka, 2005).
Studies of the emotional content of non-verbal vocal stimuli have, to the best of our knowledge,
not been undertaken.

The relationship between personality traits and perception of musical emotions

There was only a weak association between personality traits and the recognition of musical
emotions, in that O contributed significantly to the prediction of happiness when entered as an
independent variable together with other personality traits. This is not unexpected, in that
O denotes a tendency to creativity and intellectual curiosity. Individuals with high levels of
O may be better able to identify emotions simply because they tend to possess prominent
intellectual skills (McAdams, 2006).
However, in the absence of other studies in this field, this interpretation remains tentative.
Possibly, personality traits are much more relevant for music preferences than for emotion
detection accuracy in musical stimuli (Rentfrow and Gosling, 2003).

Strengths, limitations, and future studies

One strength of the present study is its connection to the neuro-cultural theory of emotion
(Ekman and Friesen, 1971), providing a well-established framework for the interpretation of
the data and contributing to the theoretical economy by systematizing future research efforts
in this field.
Second, we employed custom-made musical stimuli, thus ruling out the effects of previously-
formed associative emotional responses to the items of the test. It may be argued that, due to
our choice of mostly acoustic instruments and improvisations within the tonal system, the
sound clips would automatically create associations to classical music and thus evoke emo-
tional memories of previous classical listening experiences in the participants. In this respect,
our stimuli may not be regarded as truly unknown. In order to control for this possibility, future
studies of this test should ask the participants to report whether the sound clips reminded them
of something they had heard before. However, the ultimate consequence of this line of reasoning
would be that one could never avoid associative emotional responses to test stimuli, rendering
the study of emotional responses virtually impossible. In this light, our test is a clear improve-
ment in the field of musical emotions research in that it attempts to reduce the influence of the
associations the participants carry with them.
Despite the strengths of the study, certain limitations must be kept in mind when interpreting
our results. First, the musical segments were based on the Western tonal system. This is a
potential problem for the cross-cultural study of musical emotions. Due to the large cultural
variations in musical expressions, this test may not be suited for cross-cultural research on the
12 Psychology of Music

acoustic recognition of emotions. However, there is evidence that humans are able to recognize
the intended emotions in music from unfamiliar cultures. Balkwill and Thompson (1999)
reported that Western listeners successfully identified joy, sadness, and anger in Hindustani
ragas, and that the musical elements most relevant for this process were tempo, melodic com-
plexity, and timbre. Similar data were obtained from a study of Japanese listeners detecting
emotions in Japanese, Western and Hindustani music (Balkwill, Thompson, & Matsunaga,
1999). A study of individuals from the African Mafa population, who had never been exposed
to Western culture, showed that this group readily recognized happiness, sadness, and fear in
Western art music as well as in their own traditional music (Fritz et al., 2009). Nevertheless,
our test should be employed in non-Western cultures in order for us to conclude that it assesses
universal emotions.
Second, although we instructed our participants to report the emotions they recognized
in the musical stimuli and not the emotions those stimuli evoked in them, it is possible that,
in some cases, the participants may have identified the emotion they felt rather than the one
the music expressed. Whether such possible confusions has a tendency to occur when taking
this test, and whether felt emotions influence the classification process, are not topics of the
present study. Nevertheless, it should be mentioned that individuals, although they tend to
feel positive emotions stronger than negative emotions induced by music, are capable of clas-
sifying the different emotions correctly (Gabrielsson, 2002; Kallinen and Revaja, 2006).
Moreover, a recent investigation suggests that confusion of expressed and felt emotions in
music may be modest and not pose a grave challenge for research in this field (Vieillard et al.,
2008).
Third, the forced-choice method employed in the current study is a natural choice with a
categorical emotions approach as a theoretical point of departure, but may have led to higher
recognition rates for emotions that may be relatively hard to classify, such as disgust or anger.
In order to provide a clearer picture of classification difficulties as a part of further development
of the musical emotions test, Likert scale approaches will be considered.
Fourth, the order of presentation of the 18 musical segments was identical for all partici-
pants. It is not inconceivable that this may have generated biased results regarding the classifi-
cation of the last segments. The participants knew that they had to choose between six emotions
for each of the 18 stimuli, and if they remembered that they had classified sadness, for example,
only twice, it is possible that they felt compelled to perceive the 18th segment as sadness in order
to create numerical balance. Whether the order of presentation influences the perception of
emotions using this test should be subject to further study.
Fifth, there were several statistical tests performed with a relatively small sample. This
increases the risk of Type I error, where a true null hypothesis is rejected. This point seems
particularly relevant regarding the somewhat scattered results of the analyses of associa-
tions between musical education experience, personality traits, and emotion recognition
indices. This suggests that our findings of statistically-significant associations between
theoretical background, personality, and emotion perception should be interpreted with
caution.

Conclusion
In conclusion, the results from this first major study of this musical emotions test suggest that
the six basic universal emotions are detectable in musical stimuli, and that the ability to do so
does not seem to be influenced by musical experience or personality traits.
Mohn et al. 13

Acknowledgements

This study was supported by the Department of Psychology, University of Oslo and the German Center
for Music Therapy Research. Mr Dag Erik Eilertsen provided valuable statistical advice. Comments and
suggestions from Dr Anne Leins, Prof Dr Hans Volker Bolay, Prof Dr Thomas Hillecke, Ms Nicole Meissner,
and Mr Tanjef Gross have been greatly appreciated. Financial support for the development of the musical
emotions test was provided by Mr Reinhard Walter, FOM Future Office Management GmbH, Heidelberg,
Germany. We are grateful for the comments and recommendations of three anonymous reviewers on an
earlier version of this manuscript.

References
Balkwill, L.L., & Thompson, W.F. (1999). A cross-cultural investigation of the perception of emotion in
music: Psychophysical and cultural cues. Music Perception, 17(1), 43–64.
Balkwill, L.L., Thompson, W.F., & Matsunaga, R. (2004). Recognition of emotion in Japanese, Western,
and Hindustani music by Japanese listeners. Japanese Psychological Research, 46(4), 337–349.
Bigand, E., Vieillard, S., Madurell, F., Marozeau, J., & Dacquet, A. (2005). Multidimensional scaling of
emotional responses to music: The effect of musical expertise and of the duration of the excerpts.
Cognition & Emotion, 19(8), 1113–1139.
Busch, V., Nickel, A.K., Hillecke, T.K., Gross, T., Meissner, N., & Bolay, H.V. (2003). Musikalische und mimi-
sche Emotionserkennung: Eine Pilotstudie mit psychiatrischen Patienten [Musical and facial emotion
recognition: A pilot study with psychiatric patients]. Zeitschrift für Muzik-, Tanz- und Kunsttherapie,
14(1), 1–8.
Chamorro-Premuzic, T., & Furnham, A. (2007). Personality and music: Can traits explain how people use
music in everyday life? British Journal of Psychology, 98(2), 175–185.
Costa, Jr., P.T., & McCrae, R.R. (1992). NEO-PI-R, professional manual. Odessa, FL: Psychological Assessment
Resources.
Dalla Bella, S., Peretz, I., Rousseau, L., & Gosselin, N. (2001) A developmental study of the affective value
of tempo and mode in music. Cognition, 80(3), B1–B10.
Ekman, P. (1992). Are there basic emotions? Psychological Review, 99(3), 550–553.
Ekman, P. (1993). Facial expression and emotion. American Psychologist, 48(4), 384–392.
Ekman, P. (1994). Strong evidence for universals in facial expressions: A reply to Russel’s mistaken cri-
tique. Psychological Bulletin, 115(2), 268–287.
Ekman, P., & Friesen, W.V. (1971). Constants across cultures in the face and emotions. Journal of Personality
and Social Psychology, 17(2), 124–129.
Ekman, P., Levenson, R.W., & Friesen, W.V. (1983). Autonomic nervous system activity distinguishes
among emotions. Science, 221(4616), 1208–1210.
Elfenbein, H.A., & Ambady, N. (2002). On the universality and cultural specificity of emotion regocnition:
A meta-analysis. Psychological Bulletin, 128(2), 203–235.
Elfenbein, H.A., & Ambady, N. (2003). When familiarity breeds accuracy: Cultural exposure and facial
emotion recognition. Journal of Personality and Social Psychology, 85(2), 276–290.
Fritz, T., Jentsche, S., Gosselin, N., Sammler, D., Peretz, I., Turner, R. et al. (2009). Universal recognition of
three basic emotions in music. Current Biology, 19(7), 573–576.
Furnham, A., Zhang, J., & Chamorro-Premuzic, T. (2006). The relationship between psychometric and
self-estimated intelligence, creativity, personality and academic achievement. Imagination, Cognition
and Personality, 25(2), 119–145.
Gabrielsson, A. (2002). Emotion perceived and emotion felt: Same or different? Musicae Scientiae, Special
Issue, 123–147.
14 Psychology of Music

Gabrielsson, A., & Juslin, P.N. (1996). Emotional expression in music performance: Between the performer’s
intention and the listener’s experience. Psychology of Music, 24(1), 68–91.
Gabrielsson, A., & Lindström, E. (2001). The influence of musical structure on emotional expression. In
P.N. Juslin & J.A. Sloboda (Eds.), Music and emotion. Theory and research (pp. 223–248). Oxford: Oxford
University Press.
Juslin, P.N. (2001). Communicating emotion in music performance. In P.N. Juslin & J.A. Sloboda (Eds.),
Music and emotion: Theory and research (pp. 301–337). Oxford: Oxford University Press.
Juslin, P.N. (2000). Cue utilization in communication of emotion in music performance: Relating per-
formance to perception. Joural of Experimental Psychology: Human Perception and Performance, 26(6),
1797–1813.
Juslin, N., & Laukka, P. (2004). Expression, perception, and induction of musical emotions: A review and
a questionnaire study of everyday listening. Journal of New Music Research, 33(3), 217–238.
Kallinen, K. (2005). Emotional ratings of music excerpts in the Western art music repertoire and their
self-organization in the Kohonen Neural Network. Psychology of Music, 33(4), 373–393.
Kallinen, K., & Ravaja, N. (2006). Emotion perceived and emotion felt: Same and different. Musicae
Scientiae, 10(2), 191–213.
Kopacz, M. (2005). Personality and music preferences: The influence of personality traits on preferences
regarding musical elements. Journal of Music Therapy, 42(3), 216–239.
Laukka, P. (2005). Categorical perception of vocal emotion expressions. Emotion, 5(3), 277–295.
Martinsen, Ø., Nordvik, H., & Østbø, L.E. (2003). NEO-PI-R. Oslo: Gyldendal.
McAdams, D.P. (2006). The person. Hoboken, NJ: John Wiley.
Nawrot, E.S. (2003). The perception of emotional expression in music: Evidence from infants, children,
and adults. Psychology of Music, 31(1), 75–92.
North, A., & Hargreaves, D. (2008). The social and applied psychology of music. Oxford: Oxford University
Press.
Powers, N., & Trevarthen, C. (2009). Voices of shared emotion and meaning: Young infants and their
mothers in Scotland and Japan. In S. Malloch and C. Trevarthen (Eds.), Communicative musicality.
Exploring the basis of human companionship (pp. 309–340). Oxford: Oxford University Press.
Rawlings, D., & Ciancarelli, V. (1997). Music preference and the five-factor model of the NEO personality
inventory. Psychology of Music, 25(2), 120–132.
Rentfrow, P.J., & Gosling, S.D. (2003). The do re mi’s of everyday life: The structure and personality cor-
relates of music preferences. Journal of Personality and Social Psychology, 84(6), 1236–1256.
Resnicow, J.E., Salovey, P., & Repp, B.H. (2004). Is recognition of emotion in music performance an aspect
of emotional intelligence? Music Perception, 22(1), 145–158.
Scherer, K.R. (2004). Which emotions can be induced by music? What are the underlying mechanisms?
And how can we measure them? Journal of New Music Research, 33(3), 239–251.
Sloboda, J.A., & Juslin, P.N. (2001). Psychological perspectives on music and emotion. In P.N. Juslin &
J.A. Sloboda (Eds.), Music and emotion: Theory and research (pp. 71–104). Oxford: Oxford University
Press.
Tomkins, S.S. (1963). Affect, imagery, consciousness: Vol. 2. The negative affects. New York: Springer.
Vieillard, S., Peretz, I., Gosselin, N., Khalfa, S., Gagnon, L., & Bouchard, B. (2008). Happy, sad, scary and
peaceful musical excerpts for research on emotions. Cognition & Emotion, 22(4), 720–752.
Wiggins, J.S. (1996). The five-factor model of personality. Theoretical perspectives. New York: Guilford Press.

Christine Mohn is a clinical psychologist and a senior research fellow at the Department of
Psychology, University of Oslo, and Vestre Viken Hospital Trust. Her research interests include
general psychology of music, cognitive neuropsychology, and the psychology of chronic pain.
Mohn et al. 15

Heike Argstatter is a clinical psychologist as well as a music historian. Currently, she is a

postdoctoral research fellow at the German Center of Music Therapy Research in Heidelberg.
Her main research interests are music therapy in tinnitus and other neurological disorders.

Friedrich-Wilhelm Wilker is a clinical psychologist and currently professor of medical psychology

at the Faculty for Music Therapy, SRH University of Applied Sciences in Heidelberg. His other
areas of interest are general psychotherapy and music therapy.