British Journal of Music Education

http://journals.cambridge.org/BME
Additional services for British

Journal of Music Education:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

Composing, performing and audience-listening as indicators of musical

understanding
Keith Swanwick and Cecilia Cavalieri Franca
British Journal of Music Education / Volume 16 / Issue 01 / March 1999, pp 5 - 19
DOI: null, Published online: 08 September 2000

Link to this article: http://journals.cambridge.org/abstract_S026505179900011X

How to cite this article:
Keith Swanwick and Cecilia Cavalieri Franca (1999). Composing, performing and audience-listening as indicators of musical
understanding. British Journal of Music Education, 16, pp 5-19
Request Permissions : Click here

Downloaded from http://journals.cambridge.org/BME, IP address: 200.130.19.173 on 29 Oct 2014

B. J. Music Ed. (1999) 16:1, 519

Copyright # 1999 Cambridge University Press

Composing, performing and

audience-listening as indicators of
musical understanding
Keith Swanwick and Cecilia Cavalieri Franca

Although it is often suggested that there are important connections between composing,
performing and audience-listening, there is little evidence as to the nature of this relationship.
In this paper we report research into the extent that musical understanding is symmetrically
revealed and developed across the three activities. Our theory of musical understanding is
based on the work of Swanwick, and assessment criteria derived from the work of Swanwick
and Tillman (1986) made it possible to compare levels of musical cognition whatever the
specic activity. A study of the musical work of twenty children in a music school in Belo
Horizonte, Brazil, suggests that performance usually elicits lower levels of musical understanding, signicantly different from either composing or audience-listening. It seems that
performance can be problematic within the music curriculum unless students are able to work
at a technical level where they are able to exercise interpretative judgement and make musical
decisions. The ndings support the claims for an integrated music curriculum comprising all
three activities.

It is widely believed that composing, performing and audience-listening should all be

included in the school music curriculum and they are somehow interactive.1
However, little systematic work has been done to support this belief and the nature of
any interaction among the activities is by no means clear. There is little research to
support the view that these three activities are `dependent' on one another (Mills,
1991), or to conrm that they `share many skills' (Gane, 1996). Interpretation of any
evidence here would require a reasonably robust theory of musical mind, a valid
account of underlying musical understanding (a term preferred to `cognition', which
still carries a suggesting of separating thinking from feeling) whatever the specic
activity.
In this paper we describe research which may illuminate something of the relationship between composing, performing and audience-listening. It is suggested that there
are cognitive structures that underlie the different modalities of musical engagement
(Hargreaves and Zimmerman, 1992; Hargreaves, 1996). We argue that underlying
the different modalities is the dimension of `musical understanding', the quality of
musical thinking or cognition that is manifested in the activity. However, musical
understanding may not be consistently or symmetrically revealed across activities,
since the demonstration of understanding is dependent upon the nature of the specic
task. Each modality of musical behaviour is articulated through specic processes and
practical skills. Interaction among music curriculum activities takes place among are
the interpretative schemata of an individual, what we call musical understanding.
This interaction may be fairly straightforward when the task is at a technical level that
the person can control. For instance, a child listens to an ostinato and manages to use

http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

this device in a composition. Perhaps the same child also perceives the expressiveness
of a large crescendo in a recorded piece of music and may want to incorporate this in
a performance. This will only be successful if he or she has mastered technique
necessary to play a graduated crescendo. Interaction among the activities is optimal
only when there exist the necessary technical conditions for the accomplishment of
the different tasks.
The skills, procedures, techniques and experiences of composition, performance
and audience-listening are very different. For instance, a skill in composition may
include the ability to develop motifs; in performance, the capability to move the
ngers and arm in a particular way to produce a crescendo; or to identify a musical
style in audience-listening. Reporting audience-listening experiences involves still
other skills, often oral/linguistic, drawing, notating, moving/dancing to music. The
particularity of skills can also be seen, for instance, in the case of an accomplished
composer who does not have highly developed instrumental skills and may not
therefore be able to demonstrate musical understanding to the same extent when
attempting to play a virtuoso instrumental piece.
This kind of discrepancy between conceptual understanding and the means of
articulation is most obvious in the area of ordinary language. A pre-school child's
speech may indicate quite advanced understanding of the functioning of a language
even though, he or she may not have begun to learn written language skills. At this age
written text would be a poor indicator of linguistic understanding. The reverse may
happen with an adult in learning a second language. Someone who happens to have
learned this language mostly from books may be able to communicate reasonably well
through writing but not through speaking, a very different skill in which there may not
have been enough practice.
The demonstration of the extent of one's understanding of music similarly depends
on renement of the skills particular to that activity. We can only assess with
condence the extent of a student's musical understanding provided that the skills
specic to the activity are within reach. Furthermore, if students are not working at
the highest levels at which they can exercise musical judgement and take musical
decisions, how can they be said to be developing their musical thinking?

The aim of the study

The study described here investigated the relative levels of musical understanding
demonstrated through composing, performing and audience-listening activities.
Should the level of understanding revealed in one activity be able to predict the level
to be revealed in other activities? Under what conditions is there likely to be such
symmetry? It is hypothesised that children may reveal the same underlying level of
musical understanding across the three activities when the complexity of musical tasks
is controlled.

The research instrument

The instrument to measure students' understanding across the three activities was the
three-fold set of criteria statements derived from Swanwick and Tillman's (1986)
spiral model of musical development. This is based on a comprehensive theory of
musical mind (Swanwick, 1979 and 1983) which is further developed through careful
observation of children's music-making. The model identies the elements of musical
experience, the awareness of which unfolds according to a developmental sequence:
6
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

sound materials (Sensory and Manipulative levels), expressive characterisation (Personal and Vernacular levels), form (Speculative and Idiomatic levels) and musical
valuing (Symbolic and Systematic levels). Although the model does not set up a xed
developmental timetable, it does imply an invariable sequence which has `considerable predictive power' (Swanwick, 1991). Subsequent research suggests that the
achievement of higher levels of musical development may be facilitated by education
and training (Stavrides, 1995; Hentschke, 1993).
The reliability of the criteria for the assessment of composition and performance
has been established earlier (Stavrides, 1995; Swanwick, 1994). The audiencelistening criteria, however, had not been subject to a reliability test prior to this study.
A similar procedure to the one described by Swanwick (1994) for testing the
performance criteria was used here to test the extent to which the audience-listening
criteria attract interpersonal agreement. These criteria are given below.

Audience-listening criteria
Sensory The student recognises sound qualities and effects, perceives clear differences of loudness level, pitch, timbre, tone colour and texture. None of these is
technically analysed and there is no account of expressive character or structural
relationships.
Manipulative The student perceives steady of uctuating beat, identies specic
instrumental and vocal sounds, devices related to the treatment of musical material,
such as glissandi, ostinati, trills; yet does not relate these elements to the expressive
character and structure of the piece.
Personal The student describes the expressive character, the general atmosphere,
mood or feeling qualities of a piece, maybe through non-musical associations and
visual images. The student relates changes in the handling of sound materials,
especially speed and loudness, with changes of expressive level, but without drawing
attention to structural relationships.
Vernacular The student identies commonplaces of metric organisation, (sequences,
repetitions, syncopation, drones, groupings, ostinati) and perceives conventional
musical gestures and phrase shape and length.
Speculative The student perceives structural relationships, the ways in which musical
gestures and phrases are repeated, transformed, contrasted or connected. He or she
identies what is unusual or unexpected in a piece of music; perceives changes of
character by reference to instrumental or vocal colour, pitch, speech, loudness,
rhythm and phrase length, being able to discern the scale in which changes take place,
whether they are gradual or sudden.
Idiomatic The student places music within a stylistic context and shows awareness of
technical devices and structural procedures which characterise an idiom, such as
distinctive harmonies and rhythmic inections, specic instrumental or vocal sounds,
decoration, transformation by variation, contrasting middle sections.
Symbolic The student is aware of how sound materials are organised to produce a
particular expressive character and stylistically coherent formal relationships. There
7
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

are individual insights and independent critical appraisal. A feeling of valuing of

music comes over through an account of personal involvement in a chosen area of
music-making and/or a sustained engagement with particular works, composers or
performers.
Systematic The person reveals a profound understanding of the value of music due
to a developed sensitivity with sound materials, the ability to identify expression and
comprehend musical form. There is a systematic commitment to music as a meaningful form of symbolic discourse.
Twelve judges, experienced musicians who were not previously acquainted with these
criteria, were asked to independently rank sets of randomised cards into a qualitative
hierarchical sequence. Each card contained a single criterion statement. An appropriate statistical test for the level of inter-judge agreement is highly signicant.2 These
results further support the reliability of the three-fold set of the criteria statements as
an instrument for assessing musical understanding wherever it appears. We therefore
felt condent to observe and assess students' musical `products' across the three
activities maintaining an element of rigour.

Context
The main study was carried out in Belo Horizonte, Brazil, over a ve-month period of
teaching in a non-specialist music school. This school provided the range of music
education that enabled us to investigate students' achievement across the three
activities. Students attend weekly a one-hour `musicalisation' lesson (in groups from
four to eight), which consists of general musicianship activities, and a one-hour
instrumental lesson (in pairs). The study proceeded in an ecologically friendly way,
where music-making took place in a normal classroom environment.

Sample
A purposive sampling method was used (Robson, 1993; Cohen and Manion, 1994).
The population consisted of twenty students at the school who met all the sampling
requirements. They were aged between eleven and thirteen-and-a half, an age at
which they are expected to have developed a reasonable level of instrumental
performance skills as well as language skills to articulate their audience-listening
experiences. Potential technical differences across instruments are controlled by
choosing only pianists with a minimum of three years attendance at classes at the
school. Students' social background was considerably levelled, as it is a private school
that attracts high-middle-class students. A further element of control was the use of
repeated measures. We collected three `products' by each student in each activity.
There were thus nine observations for each student, giving measures both across the
conditions and within the conditions, thus increasing the internal validity of the study
(Coolican, 1994).

The data the work of the students

All compositions were produced individually, during the student's regular piano
lesson at different occasions over the ve months of teaching. Time spent making a
composition was limited to a maximum of twenty minutes. This was thought to be
8
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

sufcient for students to work out and organise musical ideas into pieces without
these getting too complicated or long. Although students had already developed
reasonable notational skills, oral composition was preferred, since notation tends to
impose several constraints over the process. The initial stimulus for each composition
was at the level of the Manipulative as, for example, syncopation, particular intervals,
or the `chop-sticks' technique. The students determined everything else and they
could even move away from the initial stimulus. Once the student said the composition was nished it was recorded. A second recording was made for those who were
not satised with the rst one.
The students' performance repertoire in this school is agreed between them and the
teacher, and rehearsed over the semester. The three performances by each student
were recorded by the end of the teaching period. The collection of data in the
audience-listening modality was through an individual structured interview. Students
were to listen to three pieces of music, three times each, and report in turn. Whole
pieces (not fragments) ranging from 1'30'' to 3'0'' minutes were chosen from different
Brazilian musical styles. They were instrumental and musically interesting but did not
require complex listening skills. The objective was not to test students' discriminative
skills, but to check what dimensions of music criticism captured their attention. To
minimise the possibility of bias, broad questions were posed, like `What can you tell
me about this music you have listened to?', `If you had to describe this music to a
friend who had never heard this music before, what would you say about it?'
(Swanwick 1994; Hentschke, 1993). Students were given the opportunity both to
take notes and talk about the pieces.

Assessment of students' work

Eight experienced music teachers acquainted with the criteria were divided into two
groups of four. Each judge independently assessed students' `products', assigning the
highest evident criterion level of musical understanding revealed in each product. The
scores represent the criteria levels, with 1 standing for the lowest level (Sensory) and 8
standing for the highest (Systematic). In terms of inter-judge reliability the results are
highly signicant.3

Results
The scores of all judges for all the products were computed to show their range and
distribution. This distribution is given in Table 1 and Figure 1.
In composition and audience-listening the raw scores are clearly concentrated
around the Speculative and Idiomatic levels (the Spiral layer of form). There is a
greater dispersion of lower levels in performance than in the other two modalities.
Table 1. Distribution by raw scores

COM
AUD
PER

MAN

PERS

VERN

SPEC

IDIOM

SYM

0
0
1

0
3
19

27
9
115

117
138
74

81
90
27

15
0
4

9
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

140
120
100
80
60
40
20
0
MAN

PERS

VERN

Composing

SPEC

IDIOM

Audition

SYM

Performing

Fig. 1. Distribution by raw scores

The `rule' for summarising data

To compare achievement for individual students across the three activities, the raw
scores had to be transformed into one derived score for each student in each activity.
A common procedure would be to nd averages, particularly the median or the mode,
which are appropriate measures of central tendency for ordinal level of measurement
(Coolican, 1994). However, to reduce these scores to averages could be misleading. It
is convenient to take a hypothetical example to demonstrate the effect of reducing the
scores to averages. Take, for instance, the three compositions by a student; with four
judges making twelve observations in all.
Table 2. Hypothetical judges' scores for the three compositions by a student (J=`Judge')
Composition 1
J1
5

J2
5

Composition 2
J3
5

J4
4

J1
5

J2
5

Composition 3
J3
6

J4
5

J1
6

J2
7

J3
6

J4
6

In this instance, should either the mode or the median be used, the score for that
student in composition would be 5 (Speculative) despite the work having been
assigned to level 6 (Idiomatic) four times, and once to level 7 (Symbolic). Both the
median and the mode are unaffected by extreme values in one direction (Coolican,
1994). Consequently, they would have given a distorted picture of the student's
understanding of music by leaving precisely the highest scores the evidence this
research was looking for out of the derived scores.
10
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

A more coherent procedure was to set up a `rule' for transforming the raw scores
into more valid derived scores. There were two critical points. First, how much
evidence was needed to say that a particular quality of thinking was revealed across a
student's products? Or, how frequently was that quality of behaviour occurring so that
it evidenced the consistency of that quality of thinking? Secondly, the rule should also
be able to accommodate the possibility that a student may have produced, for
instance, one Idiomatic and two Speculative compositions (as suggested in the
example above). That is why it was imperative to have several (nine) observations per
student, representative of a range of musical behaviours, thus lending the measures
more validity.
Thus the following operational `rule' was followed:
For a student's products in an activity, nd the highest score assigned at least
three times out of the twelve judgements.
A higher score includes the previous ones, e.g. a score of seven includes level six.
The levels are essentially cumulative.
This `rule' was applied to all students' `products' in each of the three activities. The
derived scores obtained correlate signicantly with the raw scores from which they
were drawn; i.e. the rule does not signicantly distort the raw scores.4 The distribution of students' derived scores by level is shown in Table 3 and Figure 2.
Table 3. Distribution by students' derived scores

COM
AUD
PER

MAN

PERS

VERN

SPEC

IDIOM

SYM

0
0
0

0
3
0

0
0
7

4
6
10

15
14
2

1
0
1

Once again, performance is the activity with the greatest dispersion of lower levels,
ranging from the Vernacular to the Symbolic. The derived scores for composing and
audience-listening are again clearly concentrated in the layer of `Form', that is to say
at the Speculative and Idiomatic levels. The differences between the three activities
are statistically signicant (see Table 4). There is no simple symmetry across the
activities.5
Table 4. Difference between the three activities
Variables

Chi-Square

Signicance

composition/aud-list/performance

13.8250

p<.001

A Friedman Two-Way Anova was calculated for each pair of variables to check if the
difference could be explained by one single variable. The results are given in Table 5.
Table 5. Difference between each pair of activities
Variables

Chi-Square

Signicance

Composition/audience-listening
composition/performance
audience-listening/performance

0.4500
8.4500
9.8000

p<1 (n.s.)
p<.01
p<.001
11

http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

16
14
12
10
8
6
4
2
0
MAN

PERS

VERN

Composing

SPEC
Audition

IDIOM

SYM

Performing

Fig. 2. Distribution by students' derived scores

The difference between composition and audience-listening is not statistically

signicant. The results clearly show that performance is responsible for the variance
over the three activities.

Age and gender effects

A Chi-square test detected no age effects within the group, thus to some extent
legitimating the status of the age cohort as a single group. Although the study was not
designed to examine gender effects, we found no signicant association between
students' scores and gender, though the small numbers involved suggest some caution
in this interpretation.

Discussion
The relationship between musical understanding and skills
We argued that musical understanding is an underlying conceptual dimension that is
manifested through the various `channels' composing, performing and audiencelistening. The symmetrical results across composing and audience-listening seem to
support the assumption that musical understanding is a broad conceptual dimension
that operates across more than one activity. There are instances in the data that
illustrate interaction among the activities within the dimension of understanding
(musical materials, expression or form).
The non-symmetrical ndings in performance support our argument that the
12
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

revealing of musical understanding is constrained when tasks are not appropriate and
accessible. This embraces two distinct but overlapping issues: the level of practical
skills involved in the tasks and the distinctive nature of the activities.

Skills in composition and performance

The problem of skills seems to have had a greater impact in the performance. It was
the activity which presented the greater distribution of lower levels and the poorest
indicator of students' understanding of music. Should those students be assessed only
through their instrumental performance, the extent of their understanding of music
would be seriously underestimated. Just the opposite happened in the composition.
Students were able to demonstrate understanding of musical form through their
compositions even producing idiomatic, quite stylised pieces. They failed to demonstrate this quality of thinking in their performance, despite the fact that compositions
were produced within twenty minutes while performances were rehearsed over ve
months.
It seems that behind this higher achievement in composition (as in audiencelistening) is the level of technical articulation involved. In the composition tasks
students themselves determined the technical boundaries within which the pieces
would be drawn. The stimuli given as starting points set up no technical constraints
neither compositional skills nor performing skills. When these were specic piano
technical points, they were carefully drawn from the range of technique they had
mastered (e.g. `thirds', broken chords). The students themselves determined everything else. Obviously, they drew on their actual technical possibilities to produce the
compositions. Consequently, their understanding of music was given breadth to
expand closer to the upper limit of development set up by maturation. Compositions
drawn from the same stimulus (e.g. `chords', `chop-sticks technique') are so different
from each other that they do not leave a clue of what the stimulus was like. Students
explored the stimuli given whatever way their intellectual and affective impulse led
them to, letting intriguing musical richness, variety and interest ourish. (A compact
disc containing samples of this work will accompany BJME 16.3.)
There seems to be a difference between performance skills in composition and in
performance itself. In composition they drew on instrumental skills they had
mastered, as well as on a range of possibilities of melody, texture, phrase shape,
chord patterns from the group compositions, from the audience-listening repertoire,
and a lifetime of informal musical nourishment. The problem of (in)appropriateness
of skills in instrumental performance needs to be put into perspective. What is being
suggested is that no matter the level of technical complexity people are working at, if
technical demands are higher than what people can manage, they may not be able to
show the extent of their musical understanding. The crucial point thus seems to be
the choice of the performance repertoire so that it adequately and comfortably
allows students to reveal the quality of their musical thinking. Certainly it is
necessary at times to assign a student a piece or exercise, for example, in string
playing `to develop the fourth position'. But some teachers may present one new
technical challenge after the other, with no opportunity in between for the student
to exploit that technique to make music with a sense of expressive and structural
purpose.
There are clear instances of students transferring their performance experience into
their compositions. Many students borrowed materials from their performance
repertoire; others reproduced the expressive character of a piece under study. Some
13
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

took advantage of their idiomatic awareness and produced compositions in the same
style as their repertoire. In these cases, they showed the knowledge of how materials,
especially certain harmonic colours and textures, work to create mood and stylistic
authenticity. Their accumulated experiences in performance and audience-listening
enabled them to put musical ideas together creating interesting new structural
relationships, but the realisation of these compositions through their playing was in
general constrained by their instrumental technical accomplishment. In some cases,
the interaction includes also the transfer of technical uency, motor control, quality of
touch and articulation. Although performance was in general the poorest indicator of
children's understanding of music, these musical examples point to a very positive
effect it made on composition: it enabled students with the technique to realise
compositions in a consistent and developed manner.

Audience-listening and the layers of musical criticism

It may also be that integrating performance with composing and audience-listening
may encourage a productive relationship between imitation and imaginative play.
Audience-listening was the modality that presented the greatest homogeneity of
response: it had the smaller range, with all derived scores concentrated within the
Form layer of musical criticism (six students achieved the Speculative and fourteen
achieved the Idiomatic level). This may be explained by the fact that it was the only
activity in which students were experiencing the same three pieces of music. This
means that the task factors were held constant for all students, i.e. `music' variables
were controlled. What followed was that all students referred to the three layers of
musical criticism, materials, expression and form, in all three pieces of music. This
happened in spite of audience-listening being the activity which presents the more
distinctive procedural nature. In this modality the assessment of musical understanding is indirect, i.e. through students' reports. The pieces selected for the
interview presented clear communicable musical ideas and no matter whether a
student had associated the piece with a `dark corridor' or a `rain forest', with a
`primitive tribe' or a `battle', their reports gave clear clues as to which piece they
referred to. One of the criteria for selection of the pieces for the interview was that
they should be developed enough to prompt responses at all levels of musical
criticism yet not involving complex discriminative listening skills in order to facilitate
appraisal. Should the complexity of the audience-listening tasks have been higher, we
could have expected lower achievement in this modality, or, at least, more heterogeneous results.

The psychological nature of the activities

The level of skills involved partially explains why students achieved higher scores in
composition and audience-listening, and one or two levels removed in performance.
But it may be illuminating to recall Swanwick's (1983) assumptions as to the
psychological nature of the activities. Audience-listening has a heavier bias towards
accommodation. Conversely, composition is to a greater extent assimilation, an
exercise of imaginative play. While in audience-listening responses were conned to
the Speculative and Idiomatic levels, in composition there were a number of
judgements at the Symbolic level. As Vygotsky (1978) has suggested, in imaginative
play children may override the expected level for their age cohort. Composition gave
the opportunity for assimilation to take over, as it was the student who decided
14
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

either consciously or not what musical elements would be articulated and how. This
does not diminish the importance of the consistency achieved in audience-listening,
but it certainly highlights the capability of students to compose at a higher level.
Traditional instrumental performance also involves a great deal of accommodation to
a musical product that would have been produced by someone else, in another time
and place. The individual has to adjust to a series of external constraints it is a
struggle to master various skills, from notation (if this is the case) to idiomatic
features. One will only be able to make performance decisions within a range of
performing skills that one controls.
There seem to be also important psychological differences between performing
one's own composition and performing someone else's. It is important to recall that
our students were to perform their own compositions. It is striking and almost puzzling
that in many cases they were able to play their own compositions more meaningfully
and sensitively than their `normal' piano repertoire this despite the fact that they
had practised their performance repertoire for many weeks, even months. There were
instances of students who could not shape a phrase in their performance, but who
produced expressively shaped phrases in their compositions. This may be due to the
fact that while composing they are using their skills with a direct musical purpose to
achieve a particular result or effect that is in their mind. The analysis of individual
students' `products' show many cases in which all the renement, careful touch,
phrase shape and direction, the structural articulation, ending gestures, many times
revealed in their compositions, disappeared in their performances.
The assimilation/accommodation issue involves not only technical but also intellectual and affective aspects, as to personal preference and taste in terms of expressiveness and style. The distinctive nature of the activities sets up different levels of
freedom in regard to choices and decision making over musical discourse. When
performing someone else's piece, students must rst `conform' to what is being
dictated by the score and hopefully give their individual interpretation which would
involve no more than subtle deviations of timing and loudness. But while playing their
own compositions, they are playing what is technically appropriate for their ngers
and hands, and expressing their own ow of ideas, with their meanings, shapes,
character, personality, emphasis: they are speaking for themselves.

Composition as a leading activity

Among the three activities, composition is the one that has the stronger assimilative
nature, involving a greater extent of imaginative play and allowing more freedom than
the others. When composing, students may practise their repertoire of musical
schemata, reorganising and extending them. This certainly has an impact on
curriculum practice and lends some empirical and psychological support the long
commitment of some writers notably John Paynter to musical composition as a
central activity in the music curriculum. Although the difference between composition
and audience-listening in our study was not statistically signicant, qualitative analysis
of compositions gives some indication of students achieving slightly more in this
activity. This seems to be a relevant area for further research, using larger samples and
cross-sectional or longitudinal designs. An important implication of this study for
further research in the eld is the need to consider the appropriateness of musical
tasks; otherwise validity of such measures would be threatened.

15
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

Getting performance into perspective

Instrumental teaching carries the heavy tradition of the virtuoso performer, which can
help to perpetuate an equivocal conception of what musical experience is about.
Depending on the way it is conducted and assessed it can lead to a result which is not
always musically and educationally valid. The performance tasks (the repertoire)
often stretch students' skills beyond a level they can afford. How can one make
expressive decisions without control? What is the point in pushing children into tasks
that they carry out mostly with no meaning, no shape, no sense of expressiveness, no
coherence? How could it allow them to experience the power of music to enhance
intellectual and affective life? All the pleasure, the enjoyment, and the fullment of
musical experience may be easily replaced with a mechanical and dull practice of a
narrow repertoire. This is not to say that musical accomplishment can be achieved
without hard work; but the nourishment of the affective side and opportunity for
aesthetic insight can be easily neglected.
It is essential to nd a balance between the need to develop understanding and
develop skills in the choice of repertoire. Such balance will lean towards one pole or
the other according to the nature of the educational setting whether it tends more or
less to specialist objectives. But the undesirable situation is that in which across her
whole repertoire the child is never given the chance to perform pieces which he or she
would play more comfortably with expression, insightful touches and a sense of style,
so as to feel how fullling musical experience can be. Pieces should not be always so
easy that they do not offer challenges for further development, but also not so difcult
that they are beyond students' capabilities. This may result in a pointless and
frustrating effort. There is no harm in assigning students technically easier pieces, in
parallel to more demanding technical exercises. They can approach these more
musically, exploring different loudness levels, articulation, speed, and making expressive decisions over a range of skills they can control. The same applies for the artistic
quality of the repertoire they play. If the pieces themselves do not require or permit
the student to exercise judgement at a higher level, how can such quality of musical
thinking be developed? The balance could thus be achieved across several pieces:
some more technically demanding, allowing a student to `develop the fourth position',
and others involving technical elements which students can manage more comfortably. The psychological motivation towards mastering and the accommodative effort
of playing could be balanced with an imaginative and insightful interpretation of the
piece.
There is also a sort of fascination surrounding staff notation, and the initiation into
it represents a great achievement for children. Such `power' of notation is overwhelming and it soon takes over, and the time allowed for creative `playing around' on
the instrument is gradually reduced and replaced by `playing middle C'. It divorces
the `playfulness' of improvising from the `seriousness' of reading a score. Although
notation can open many new possibilities, it takes students years until they can read
and play music just as interesting, rhythmically as rich and as wide in tessitura as
music they used to play in their rst year, by ear, imitation or through improvising. It
is for teachers to help them to keep the sense of imagination and playfulness in their
performances, to restore and refresh the component of assimilation necessary to bring
about an insightful interpretation.

16
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

Conclusion
It is important to acknowledge the possibilities of generalisation. The small sample
was studied in some detail, and this enabled us to address the relationship between
understanding and skills across the various activities, and to address the relative role
of these in musical development. This was possible because all students selected for
the sample were productively involved in an integrated approach to the three activities
and could offer a range of products across which we could assess their understanding
of music. Hence, there is a relevant scope for generalisation at a theoretical level:
students may achieve at their optimum level when tasks are appropriate and accessible. The
modality in which tasks were less appropriate and accessible (performance) the
revealing of students' understanding of music was constrained. Should the composing
tasks have pre-set more complex techniques, a particular style or another medium
students were not acquainted with, the results could have been quite different the
same for audience-listening.
The ndings of this study support the idea that music education should provide a
wide range of musical experiences. It was evident that different forms of musical
behaviour may be either better or poorer indicators of students' optimal level of
understanding, according to the nature of the task and the skills involved. Also,
different students may have different degrees of ability or interest for one activity or
another. The data illustrate that the three activities nourish students' experiences
from differing angles. Through audience-listening students may expand their musical
horizons and understanding. It enriches their repertoire of options upon which to act
creatively, transforming, reconstructing and reintegrating ideas into new shapes and
meanings. But it allows decision-making to a lesser extent. Also, there are no signs in
our data of any student reaching the Symbolic level in audience-listening, which did
happen in performance and composition. Performance allows decision-making to
some extent, for the interpretative component that it involves. But it is composition
that allows more breadth for decision-making over a much wider range and thus is
particularly powerful in facilitating the development of musical understanding (Swanwick, 1994). This study indicates that an integrated approach to the three activities
may foster development when it engages students in all the layers of musical understanding. But in order to maximise any interaction, it is necessary to enable students
with activity-specic skills.
An important outcome of this work is that it draws attention to the relationship
between revealing understanding and developing understanding. It is necessary to
disentangle the level of technical skills involved in a task and the level of understanding being fostered through that task. If pupils are not working at a level where
they can exercise judgement and take decisions, how can they be developing a more
sophisticated quality of musical thinking? How can students develop a higher level of
musical understanding if they are not given the opportunity to work or `function' at
that level? How can students have an aesthetically rewarding experience when they are
still worried about `the third or fourth position'? How can students engage in music as
symbolic discourse if they never have the opportunity to produce a musically meaningful statement? If education is concerned with the potential of musical experience to
develop the mind, to open new horizons, to deepen intellectual and affective life,
there can be no compromise.

17
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

Notes
1
2

3
4

See Leonhard and House (1959/1972), Swanwick (1979 and 1994), Regelski (1975),
Plummeridge (1991), Mills (1991), Glover and Ward (1993), Gane (1996) among others.
The Kendall Coefcient of Concordance W' is a non-parametric test which determines the
degree of association among several sets of rankings; it is particularly useful to measure interjudge agreement (Siegel, 1956). In this case there is a W value of 0.9193, at p<0.0001. The
order of the sums of the ranks matches perfectly the predicted order of the criteria.
Each of the two groups of judges was tested by the Kendall Coefcient of Concordance, at
p<0.0001 for both groups. `W' values are 0.7866 for Group 1 and 0.6780 for Group 2.
The two `distributions of scores', the one before the rule (Table 1), and the one obtained
from the rule (Table 3) were subject to a correlation test. A Spearman Correlation Coefcient
gives a value of 0.7274, p<.001.
Students' derived scores in the three activities were submitted to a Friedman Two-Way
ANOVA. The signicance level is p<.001.

References
Cohen, L. & Manion, L. (1980/1994) Research Methods in Education. London: Routledge.
Coolican, H. (1994) Research Methods and Statistics in Psychology. London: Hodder &
Stoughton.
Gane, P. (1996) `Instrumental Teaching and the National Curriculum: A Possible Partnership?'
British Journal of Music Education,13/1, pp. 4965.
Glover, J. & Ward, S. (1993) (Eds) Teaching Music in the Primary School. London: Cassel.
Hargreaves, D. (1996) `The Development of Artistic and Musical Competence', in I. Deliege
& J. Sloboda (Eds) Musical Beginnings: Origins and Development of Musical Competence. New
York: Oxford University Press.
Hargreaves, D. & Zimmerman, M. (1992) `Developmental Theories of Music Learning', in
R. Colwell (Ed) Handbook of Research on Music Teaching and Learning: a Project of the Music
Educators National Conference. New York: Schirmer Books.
Hentschke, L. (1993) `Musical Development: Testing a Model in the Audience-Listening
Setting.' Unpublished PhD thesis, University of London Institute of Education.
Leonhard, C. & House, R. (1959/1972) Foundations and Principles of Music Education.
McGraw-Hill Book Company.
Mills, J. (1991) Music in the Primary School. Cambridge University Press.
Plummeridge, C. (1991) Music Education in Theory & Practice. London: The Falmer Press.
Regelski, T. (1975) Principles and Problems of Music Education. New Jersey: Prentice-Hall.
Robson, C. (1993) Real World Research: A Resource for Social Sciences and PractitionerResearchers. Oxford: Blackwell.
Siegel, S. (1956) Nonparametric Statistics for the Behavioral Sciences. New York: McGraw-Hill
Book Company.
Silva, C. Franca (1998) `Composing, Performing and Audience-listening as Symmetrical
Indicators of Musical Understanding.' Unpublished PhD Thesis, University of London
Institute of Education.
Stavrides, M. (1995) `The Interaction of Audience-Listening andCcomposing: A Study in
Cyprus Schools.' Unpublished PhD thesis, University of London Institute of Education.
Swanwick, K. (1979) A Basis for Music Education. London: Routledge.
Swanwick, K. (1983) The Arts in Education: Dreaming or Wide Awake? Special Professorial
Lecture. University of London Institute of Education.
Swanwick, K. (1991) `Further Research on the Musical Developmental Sequence,' Psychology
of Music, 19/1, pp. 2232.
Swanwick, K. (1994) Musical Knowledge: Intuition, Analysis and Music Education. London:
Routledge.
18
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173

Composing, performing and audience-listening Keith Swanwick and Cecilia C. Franca

Swanwick, K. & Tillman, J. (1986) `The Sequence of Musical Development: A Study of
Children's Composition,' British Journal of Music Education, 3/3, pp. 30539.
Vygotsky, L. (1978) Mind in Society: The Development of Higher Psychological Processes. Cambridge, MA: Harvard University Press.

19
http://journals.cambridge.org

Downloaded: 29 Oct 2014

IP address: 200.130.19.173