1

Kings College London

Faculty of Arts & Humanities
Coversheet for submission of coursework
(Undergraduate & Taught Postgraduate)

Complete all sections of this form and ensure it is the first page of the document you submit. [Note:
either copy and paste this page into the front of your work, or write your work on subsequent pages
of this form]
DO NOT WRITE YOUR NAME ON YOUR WORK.
Pages should be clearly numbered.
Failure to attach the coversheet as required may result in your work not being accepted for
assessment.

Candidate no.

Module Title:

Issues and Topics in Music 4

Module Code:
(e.g. 5AABC123 )

4AAMS164

The word count, which should preferably be calculated electronically, must be stated accurately above.
For details of how to calculate the word count, please consult the Faculty handbook. No penalty is exacted for
work up to 5% above the word limit. Thereafter two marks will normally be deducted for every 5% above the
word limit, until 50% is reached. After 50%, three marks will normally be deducted for each additional 5% above
the word limit. These regulations are laid down by the Boards of Examiners in the Faculty of Arts & Humanities.

DECLARATION BY STUDENT
This assignment is entirely my own work. Quotations from secondary literature are indicated by the use of
inverted commas around ALL such quotations AND by reference in the text or notes to the author concerned.
ALL primary and secondary literature used in this piece of work is indicated in the bibliography placed at the end,
and dependence upon ANY source used is indicated at the appropriate point in the text. I confirm that no
sources have been used other than those stated.
I understand what is meant by plagiarism and have signed at enrolment the declaration concerning the
avoidance of plagiarism.
I understand that plagiarism is a serious examinations offence that may result in disciplinary action
being taken.
Your assignment may be used as an exemplar of good practice for other students to refer to in the
future. If selected, your assignment will be uploaded to the Kings E-Learning and Teaching Service
(KEATS). The assignment will be anonymous and not include feedback comments or the specific grade
awarded. Your participation is entirely optional and will not affect your grade.
If you DO wish your assignment to be used in this way then please put an X in the box

Are Musical Emotions Universal?

For centuries, music has been at the heart of all cultures, and emotion has
been perceived and induced so naturally, that only recently have investigations
begun that question why music can touch us so deeply, and whether emotional
responses to music are universal. In particular, two experimental articles will be

discussed in this essay; Universal Recognition of Three Basic Emotions in Music

and Universal and culture-specific factors in the recognition and performance of
musical affect expressions. I will be examining the probability of musical emotions
being universally induced, with emphasis on the former article, as I found it to be
less conclusive, and therefore more open for improvement and subsequent
discussion.
In the first article, a group of psychologists in Leipzig, Germany in 2009 were
interested in investigating the probability of musical emotions being universally
perceived. This resulted in two comparative cross-cultural experiments between an
isolated group of the 21 native African Mafa participants and a group of Western
listeners.
The first experiment asked participants to listen to two types of music: their
native music, and the other groups respective music, and indicate the emotion
best perceived, by specifying the corresponding facial expression. These consisted
of three images, taken from the Ekman archive, representing happy, sad, and
scared. The results concluded that the both the Mafa and Western participants
recognised the happy, sad and scared emotions in Western music. The Western
participants rate of recognition was substantially higher than the Mafas, but the
Mafas recognition were nevertheless above chance 1. These results concluded that
the rates of identification were of a high enough consistency to demonstrate
universal emotional recognition.
One of the main considerations of the experiment was the mode and temporal
cues1; both the Mafa and the Western participants identified a faster tempo with
happy music, and a slower tempo with scary or fearful music. Tonally, a major
piece was considered happy, and a minor, sad. However, results show that the
Mafas differed from the Westerners in that they did not associate a typically
western sad piece to a slower tempo.

1 T.Fritz,.S.Jentschke,.N.Gosselin,.D.Sammler,.I.Peretz,.R.Turner,.S.Koelsch, Universal
Recognition of Three Basic Emotions in Music,p.573

So what music did both parties listen to? The Western music consisted of
classical piano excerpts, however this was played through a midi recorder. The
flaw in this is that the Western ear perceives emotion not just through notes but
also through dynamic, melodic phrasing and expression. Contrastingly, the
equivalent music from the Mafas was a live recording. It would be prudent to
assume that because the Mafas had not previously been exposed to Western
classical music, that they would not better perceive emotions in a more
emotionally convincing performance.
So would the choice of recording have affected emotional prosody to an
appropriate extent to affect the results? Emotional prosody is defined as an
important feature of language, comprising intonation, loudness, and tempo 2. The
article thus implies that the notion of emotional prosody enables us to recognize
emotion in music, because music and speech use the same emotion-specific cues.
The results showed that there was a correlation that proved a universal capability
to recognize nonverbal patterns of emotional expressiveness 3 but when the
Western music was played to both groups, the results were p 0.001 for the Western
participants, but p 0.05 for the Mafas. If p 0.001 is the value statistically significant
enough to say something is true, then for the Mafas there is an above-chance
level, but not statistically significant enough. To conclude, it is therefore possible
that the choice of recording may have contributed to the Mafas high p score.
Another obvious limitation of this experiment is the nature of the forced-choice
design. According to Ekman (with help from Wallace V. Friesen and Richard J
Davidson in 1978), there are 6 basic emotions of joy, disgust, surprise, fear, anger
and sadness. With only 3 emotions, participants would be choosing a result that
they did not necessarily agree with, but which best fitted the options available.
Ekmans research in fact stemmed from Silvan Tomkins, who developed the Affect

2 R.L.C Mitchell, The neural response to emotional prosody, as revealed by functional

magnetic resonance imaging,(2003),p.1
3 Fritz,.Jentschke,.Gosselin,.Sammler,.Peretz,.Turner,.Koelsch, Three Basic Emotions,p.574

Theory4: this stated that there were only nine affects (or emotions) that were
biologically

based.

The

basic

six

are

interest/excitement,

enjoyment/joy,

surprise/startle, distress/anguish, anger/rage, and fear/terror, with the last three

being shame/humiliation, dissmell and disgust, and that the first word is the less
intense version of the second word 1. It raises the issue of why exactly the
psychologists conducting the experiment limited their range to 3 simple emotions,
when 6 would surely not be too overcomplicating.
Another less obvious issue to be considered, which also highlights the issue of
the small Mafas participant count, is Autism. According to Tomkins, experiments
monitoring the amygdala region of the brain concluded that if participants were
shown images, for example, of scared people (fear), a response was stimulated in
the amygdala. But when Aspergers candidates were shown the same images, no
responses were stimulated. According to the 2010 CDC data analysis, 1 in 68
people worldwide were shown to have autism 5, resulting in a 1.47% probability
that one of the 21 Mafa participants would be autistic. Since the experiment relied
on recognising a facial expression associated with the music, these findings could
be enough to skew the results.
Although the article advocates the monitoring of perceived emotions, it is
important to realize that the border between the two alternatives is somewhat
blurred In most situations, listeners are somewhere along this continuum. 6
Provided this statement is convincing, then there would be some form of brain
activity, differing in intensity. Building on both this idea, and the experiment
designed to monitor activity within the amygdala, perhaps an additional method to
more accurately achieve results could be to monitor the amygdala regions in order
to build a far more complex spectrum of emotional intensities. This would be
implemented alongside the original experiment of immediate human response, but
further improved with twelve images, in line with Tomkins Affect Theory. Every two

4 S.Tomkins, Affect Imagery Consciousness: The Positive Affects, Vol. 1(1962)

5 CDC, [Online]http://www.cdc.gov/ncbddd/autism/data.html,(Accessed 1/5/15)
6 A.Gabrielsson, Emotion perceived and emotion felt: same or different?(2002) p.124

images would be a more intense/less intense combination of the 6 main affects.

The cross-cultural goal could also be further improved by travelling to at least four
different remote regions, to gather a broader range of data with more participants,
and to experiment with more than two contrasting music traditions.
The second experiment focussed on whether a piece of dissonant music was
an equally perceived as unpleasant cross-culturally. The dissonant music consisted
of the original recording overlaid with 2 pitch-shifted records; up a semitone, and
down a tritone. So if in C, you would hear C, C#, and Ab simultaneously. Wholly
dissonant seems a fair description, because there would be occasions where the
notes of all three layers could temporarily line up to form a consonance, but this
proportion is insignificant. This type of manipulation is fair because it allows both
the western and mafas music able to be manipulated in the same way.
Predictably, the Mafas preferred the original recordings, but at 0.022% error, it
was much higher than the 0.001% of the Westerners. This was because dissonance
was not only factor: the spectrally manipulated version resulted in more
frequencies [being] simultaneously audible in Cameroon, traditional tribal music
is centred on percussion, so more frequencies gives the impression of more people
participating, and a consequently stronger tribe.
The second and arguably more in-depth article consisted of two experiments,
the first of which researched how emotions and related states can be
communicated through music within and across different cultures 7. It used eleven
affects (emotions) with professional string instrumentalists from traditional
Japanese culture, Swedish folk, Classical Hindustani and Western Classical culture.
Both the first and second experiment showed results above chance for recognising
emotions, but that there was an in-group advantage 8 emotional recognition when
the participants listened to music of their own culture.

7 P.Laukka,.T.Eerola,.N.S.Thingujam,.T.Yamasaki,.G.Beller. Universal and culture-specific

factors in the recognition and performance of musical affect expressions;
Emotion(2013).p.435
8 P.Laukka,.T.Eerola,.N.S.Thingujam,.T.Yamasaki,.G.Beller. Emotion,.p.435

The second experiment builds on the first using a lens-model approach 9,

which examines the way in which the end coders (performers) suggest an
emotional cue in the decoders (listeners). The acoustic and musical cues test
whether the listeners are recognising specific features in the music that are
consistent, and the cues are compromised of dynamics, rhythm, timbre, register,
tonality, and structure.

It would have been interesting to have repeated the

experiment again using different timbres perhaps wind, in order to get more
rounded results. Nevertheless, the second articles results were far more
conclusive for one, they [assessed] emotion recognition accuracy 10 where first
article did not.
After analysing both articles experiments, I have come to the conclusion that I
agree with the second articles theory in that musical emotions are, to some
extent,

universal,

although

music-emotional

recognition

improves

after

assimilating the cultural-specific factors. Experimentation therefore needs to work

directly towards the core of musical expression, by essentially becoming a multimusical-linguist so that musical emotions can be fully understood from unbiased
cultural perspectives. Ethnomusicologists should also make use of the most
advanced technologies available to back up their physical data with statistics, thus
creating reliable results.

9 P.Laukka,.T.Eerola,.N.S.Thingujam,.T.Yamasaki,.G.Beller. Emotion,.p.435
10 P.Laukka,.T.Eerola,.N.S.Thingujam,.T.Yamasaki,.G.Beller. Emotion,.p.436