Content uploaded by Anna Rita Addessi
Author content
All content in this area was uploaded by Anna Rita Addessi
Content may be subject to copyright.
B. J. Music Ed. 2005 22:1, 21–46 Copyright
C
2005 Cambridge University Press
doi:10.1017/S0265051704005972
Experiments with a musical machine: musical style
replication in 3 to 5 year old children
Anna Rita Addessi1and Franc¸ ois Pachet2
1Dipartimento di Musica e Spettacolo, Universita di Bologna, via Barberia 4, 40123 Bologna, Italy
2Sony-Computer Science Laboratory, 6 Rue Amiot, 75 005, Paris, France
addessi@muspe.unibo.it
The relationship between new technology and learning is gaining increasing relevance
in the field of music education (Webster, 2002; Folkestad
et al.
, 1998). However, only a
few studies have considered the nature of the interaction between children and musical
machines. This article describes an observation study of children aged 3–5 years confronting
a particular interactive musical system, the Continuator, which is able to produce music
in the same style as a human playing the keyboard (Pachet, 2003). The analysis of two
case studies suggests that the Continuator is able to develop interesting child/machine
interactions and creative musical processes in young children. It was possible to observe
a ‘life cycle’ of interaction, as well as micro-processes similar to those observed in child/
adult interactions (Stern, 1985; Imberty, 2002). The ability of the system to attract and
hold the attention of children has been interpreted through Csikszentmihalyi’s (1990)
‘flow theory’.
Introduction
The present study deals with an area still under-researched, that of interactive musical
systems, and attempts to understand in what way these systems can affect the learning
and musical creativity of children. In particular, we chose to study young children aged
3–5, because at this age the problem of the interaction between child and machine takes
on a fundamental role in the learning process. According to some developmental theories
(Fogel, 2000; Stern, 1985; Trevarthen, 2000), the infant/adult interaction plays an important
role in the affective and cognitive development of the child. Imberty (2002), in accordance
with Stern (1985), describes the musical development of a young child as based on the
vocal interaction with the mother, characterised by fluent mechanisms of repetition and
variation and affect attunements. Recent studies dealing with musical invention in very
young children (2–4 years) have suggested that the origin of new musical ideas is structurally
anchored in the sympathetic interaction established between the adult and the child while
playing with educational musical instruments (Young, 2004). Mazzoli (2003) refers to the
Vygotskian concepts of modelling, scaffolding, and mirroring, and the relative mechanisms
of interaction they involve, to provide an operative definition for the role played by the
educator in musical games in the nursery school. The interesting question therefore arises
as to which models of development and learning are produced when these relationships
are established not between two human subjects, but between a child and a machine. An
experimental protocol was therefore designed to observe young children playing with a
musical machine.
21
Anna Rita Addessi and Franc¸ois Pachet
This paper will present an overview of the literature on musical education and new
technology, a description of the interactive system we used, the observation setting, and
an analysis of two case studies. We will then draw some conclusions regarding the
psychological and pedagogical implications of our study.
Music education and new technology
The relationship between new technology and learning is gaining increasing relevance in
the field of music education. The new technologies have been viewed by researchers as
pedagogical tools (Webster, 2002), or as ‘transparent’ instruments that allow children, even
outside the schoolroom, to make and produce music (Folkestad
et al
., 1998). Many studies
have also documented the impact of new technologies on the music curriculum and its
teaching methods (e.g.
British Journal of Music Education
, 14/2, 1997;
Les Dossiers de
l’ing´
enierie educative
, 43, 2003).
This subject also has a more theoretical aspect, regarding the relationship between
new technological language and the development of knowledge (De Kerckhove, 1993;
Maragliano, 1999; Turkle, 1984). The questions most frequently asked are: What are the
learning modalities determined by digital communication? What are the new ‘cognitive
frames’ that new technologies are creating, in other words the context in which children
develop the processes of music learning and perception?
Within the issues presented above is an area still rather under-researched, that of
interactive musical systems. From this point of view, studies in the domain of artificial
intelligence are producing interesting results (e.g. Camurri & Coglio, 1998; GRM, 2000;
IRCAM
-
MusicLab, 2002). Only a few studies have observed the nature of the interaction
between children and musical machines ‘in the field’, especially in the pre-school age range
(e.g. Mazzoli, 2001).
Interactive reflective musical system: the Continuator
In our study we used a particular system, the Continuator, developed at the Sony Computer
Science Laboratory in Paris (Pachet, 2003), which was originally designed in the context
of developing new tools for improvised music. This system is able to produce music in
the same style as a human playing the keyboard, like a sort of sound mirror. An important
consequence of this design is that the phrases generated by the Continuator are
similar but
different from
those played by the users.
To illustrate the working of the Continuator, a simple musical example is given in
Figure 1. The example is notated exactly as it is played, i.e. without rhythmic quantisation.
This shows how the Continuator adapts quickly to arbitrary styles and is able to generate
musical material that ‘sounds like’ the user input on a relatively small scale. In the standard
mode, the system receives musical Midi input from one musician. The output of the system
is itself sent to a Midi synthesiser and then to a sound reproduction system. The system
is essentially a sequence continuator: the note stream of the musician is continuously
segmented into musical phrases. Each phrase is sent asynchronously to a phrase analyser,
which builds up a model of recurring patterns. In reaction to the played phrase, the system
immediately generates a continuation, according to the database of patterns already learnt.
Issues related to capturing higher-level structure are not discussed here as they are not
22
Experiments with a musical machine
Fig. 1 A simple melody (top stave) is continued by the Continuator in the same style
relevant for our purpose (see Pachet, 2003 for more details). The Continuator keeps on
learning from whatever input is given. As a consequence, the behaviour of the system
improves over time.
The basic playing mode of the Continuator is a particular kind of turn-taking between
the user and the system that is governed by three principles:
1. Automatic detection of phrase endings. The Continuator detects phrase endings by
using a (dynamic) temporal threshold (typically about 400 milliseconds). When a time
lapse exceeds this threshold, the Continuator takes the lead, and produces a musical
phrase.
2. The duration of the phrase generated by the Continuator is parameterised, but in most
cases the duration is set to be the same as the duration of the last input phrase.
3. Priority is given to the user. If the user decides to play a phrase while the Continuator
is still playing, then the system will stop and return to listening mode (and eventually
apply principle 1 again).
Experience with the system has shown that these rules are easily learned by most users in
an implicit way – the behaviour of the system is usually obvious, even for children.
Many projects have been undertaken to propose ways of enabling young children
to play music, with the goal of developing musical abilities early by designing musical
instruments that are easier to play than conventional ones (IRCAM-MusicLab, 2002;
Weinberg, 1999), or by developing tools that allow children to become instrument designers
themselves (Resnick
et al
., 1996). Many of the features we thought were exciting for
professional musicians, such as the organic capacity of the system to learn musical styles
and its ability to respond in real time, proved just as exciting for non-musicians and young
children. In all cases, the main lesson learned from these experiments is that it is worthwhile
to design and use a particular class of
interactive systems
– we call them
reflective
–for
music education: systems in which the user, regardless of skills, competence level, and
23
Anna Rita Addessi and Franc¸ois Pachet
musical goals, is confronted by a developing ‘mirror’ of him- or herself. This unusual
situation creates strong subjective feelings that we believe can be exploited for enhancing
musical creativity and for teaching musical skills in general. The core concept of this
approach is to teach powerful but complex musical processes indirectly by putting the user
in a situation where these processes are developed not by the user, nor by the machine,
but by the
interaction
between the two.
An experiment was conducted by the second author in Paris with eight children aged
3 and 4 years, who were invited to play a keyboard. When the children showed signs of
boredom, the keyboard was connected to the Continuator. A certain number of interesting
points were observed in this experience, relating to the power of attraction/addiction, the
increase in attention span, and the development of analytical behaviour (Pachet & Addessi,
2004).
The study with children and the Continuator: method
An experimental protocol was subsequently established to observe systematically some
interesting aspects of behaviour noted during the first experience, in relation both to the
age of the children (3–5 years) and to the amount of exposure to the experience, and
also taking into account whether the children were playing alone or together with another
child. From a psychological and pedagogical point of view, the general aim was to study
the nature of the interaction between the children and the system, the kinds of musical
behaviours that developed, and how interactive systems can be used in the educational
field to stimulate creativity and the pleasure of playing.
The observation was trialled in the nursery school ‘La Mela’ of Quarto Inferiore
(Bologna, Italy), in collaboration with the Istituto Comprensivo of Granarolo. The
collaborators consisted of a teacher from the school who is also a lecturer at the Faculty
of Education, University of Bologna, two graduates in Education, the other teachers of the
school and the children’s parents.
Observation of video recordings and photos
Taking into account the young age of the children, and wishing to provide a setting where
they felt at ease, we chose the observation method, which allowed us to observe the
conducts of the children without disrupting their daily routine at the nursery. The term
‘conduct’ refers here to the French term ‘conduite’ as used and scientifically defined by
Pierre Janet, Jean Clapar`
ede and Jean Piaget.1Controlled observation is possible, according
to Piaget’s ‘quasi experimental’ procedure, by means of continuous and systematic
observation of the children’s conducts in the field, taking into account various hypotheses
and monitoring defined variables (Camaioni
et al
., 1988). In our experiment the variables
were: the ‘partners’ with whom the children were invited to play (the keyboard alone, the
Continuator, another child), the exposure to the experience (once daily for 3 consecutive
days), and the age of the children (3–5 years).
Other data collected
We also collected
audio recordings
of the improvisations played by the children and
the Continuator, to aid the analysis of the musical ‘process’ (i.e. the transformations of
24
Experiments with a musical machine
the children’s musical improvisations that took place during each successive session) and
‘product’ (i.e. the improvisations themselves) (Mialaret, 1997; Folkestad
et al
., 1998).
The children were asked to
draw
the experience one week after the video recording.
The parents were asked to complete a
questionnaire
about the musical taste and
experience of their children, and about their interaction with computers, TV and hi-fi.
The teachers provided psychological and educational
profiles
of all the children
involved.
Equipment used
We used: the Continuator, a laptop computer, a Roland ED PC-180A keyboard as the
interface, a Roland expander, a pair of amplified loudspeakers, a video camera, and a
digital camera. The basic playing mode of the Continuator was the particular kind of turn-
taking described above. The duration of the phrases played by the system was set to be the
same as the duration of the last input phrase played by the child. These parameters were
set without explicitly telling the users.
Procedure
The observation was preceded by a short meeting between the operators and the children.
The meeting lasted about 20 minutes and was held in groups (two groups of approximately
15 children each). The aims of the meeting were to introduce the staff to the children, to
get to know the children, and to prepare them for the experimental activities. During this
meeting games were played, also involving the keyboard and the Continuator.
The video observation took place in the days following the meeting. Video and audio
recordings were made in the small library of the school. The keyboard was placed on a
table in front of the children and the portable computer on a nearby table. A video camera
(not visible to the child) was positioned in front of him/her, in order to record both hands
and face. One collaborator worked with the video camera, while another worked with the
children and the computer. The sessions were individual (1 child) or in pairs (2 children).
The children were supervised in the library by the operator or by the teacher. The operator
gave the assignment to the child (if necessary he turned on the computer), and while the
child was working, he either stayed in the same room and kept busy (reading, tidying), or
left the room. The children were left increasingly on their own until the third session, when
they were alone in the room. The children were asked to play in four different ways: with
the keyboard alone, with the Continuator, with another child, and finally with both another
child and the Continuator. The operator asked the child to perform the following ‘musical
games’:
The child alone:
Task 1
: ‘Play the keyboard for as long as you like. When you are tired, call me’.
Task 2
: ‘Play the keyboard for as long as you like and it will answer you. When you
are tired, call me’. (For this task the operator activated the Continuator through the
computer.)
The child with another child:
Task 3
: ‘Play the keyboard together for as long as you like. When you are tired, call
me’.
25
Anna Rita Addessi and Franc¸ois Pachet
Task 4
: ‘Play the keyboard for as long as you like and it will answer you back. When
you are tired, call me’. (For this task the operator activated the Continuator through
the computer.)
The tasks were given in random order to make sure that the different tasks were not always
performed in the same order. If the children asked to play with a friend or to start with one
particular task they were allowed to do so. The list of tasks was in any case modified so that
in the next session a child would not repeat the tasks in the same order. All sessions were
recorded on video. The music played by the children and the systems was recorded by the
same system. After one week the children were asked to draw a picture of the experience.
At the same time, their parents were asked to complete a questionnaire. The teachers were
asked for a profile of each child.
Participants
The observation was carried out with 27 children (13 boys, 14 girls). Nine children
performed all 4 tasks in all 3 sessions: 3 aged 3 years (1 girl, 2 boys), 3 aged 4 years
(2 girls, 1 boy), and 3 aged 5 years (1 girl, 2 boys). The other children only took part in
the tasks in pairs, or in free sessions, which were also recorded on video. The children all
participated on a voluntary basis and were from two different school classes. Care was taken
not to diverge from their normal daily routines. The job of the operator was to check that
the children completed the protocol, with the tasks in random order. This fluid organisation
of the protocol made it possible to fit the experiment into their normal school routine in
a natural way, allowing us to observe them under everyday conditions and encouraging
the children to express themselves in a relaxed and spontaneous fashion. This was made
easier by the organisation at La Mela school, which was based on modules and open class
projects, offering the children a high level of autonomy.
Data analysis
The data collected were wide-ranging and interesting, and stimulated us to go more deeply
into certain aspects, some of which had not been foreseen in the original protocol.
The video recordings were analysed as follows. First of all, they were watched by
two independent observers who extracted the most significant parts and divided them into
thematic areas: ‘Interaction with the system’, ‘Listening and aesthetic experience’, ‘Ways
of playing’, ‘Exploring the instrument’, ‘Joint attention’, ‘Musical improvisations’. This step
was taken in order to give us an overall idea of the material we had gathered and to
highlight any aspects not foreseen in the original objectives. The videos were edited to
enable us to study side by side the 9 children who had completed the whole protocol. After
watching the videos, and referring to the notes taken while making them, two case studies
were selected for their interesting and contrasting characteristics. They were both boys of
the same age: the marked contrast in certain of their conducts led us, in fact, to rethink our
conclusions about the variables of sex and age. The analysis focused in particular on the
nature of the interaction between child and system, and on the evolution of this interaction
during the course of the three sessions. A descriptive analysis was made to help us select
some of the most significant moments, varying in duration from around 20 seconds to
26
Experiments with a musical machine
2 minutes, which were then subjected to a micro-analysis; this also enabled us to build a
dynamic profile of the evolution of the interaction. The duration of each task was measured
to assess the attention span in relation to the different partners the child had played with.
A descriptive analysis was then made of the 7 other children who had completed the
whole protocol. The attention span of the children was measured for each task. The most
interesting data observed in the two case studies were compared with the other children
by means of an observation grid. In this paper we report some examples of the thematic
areas that emerged, together with the results of the analysis of the two case studies.
Results of the study: general aspects
Interaction with the system
The most common types of interaction identified were sensory-motor (touching and
handling), symbolic (e.g. dramatisations, ‘let’s pretend’) and rule-based. The children
danced, sang, listened, went through different emotional tones, and often expressed
aesthetic opinions. They tried to understand the rules of the system. They listened carefully
in order to create ‘musical’ dialogues with the system. They sometimes even narrated a
story while listening to the Continuator (Fig. 2a).
Relationship between pairs of children and the system
Of particular interest are the relationships established between the two children when
playing together, and between them and the system: playing, listening, exploring together,
watching the partner’s reactions, playing separately, alternating, or conflicting (Fig. 2b). A
typical situation encountered was the phenomenon of ‘joint attention’: more precisely, one
of the children would force the other to stop playing in order to listen to the situation. We
call this situation ‘Aspetta’ (the Italian word for ‘wait’) (Fig. 2c).
Fig. 2a Narrator, player, listener
27
Anna Rita Addessi and Franc¸ois Pachet
2b 2c
Fig. 2b They listen to the keyboard as it answers, and share their perplexity; Fig. 2c Joint
attention – ‘Aspetta’: one girl stops her friend in order to listen to the machine
Listening and aesthetic experience
The listening was very careful, both to the replies given by the system and to their own
work. Sometimes the listening gave rise to moments of sheer ecstasy, sudden outbursts of
joy, and excitement (Figs. 2d and 2e). Another child, Alberto, listened to the system and
exclaimed: ‘E bellissimo!’ (‘It’s wonderful!’).
2d 2e
Fig. 2d Listening and ecstasy; Fig. 2e Listening to the Continuator
Ways of playing; exploring the instrument
The children explored the keyboard and means of making sound in a myriad different
ways: with their elbows, heads, bottoms or forearms, with their hands in their sleeves,
chopping, with just one finger, several fingers, the palm of the hand, facing backwards,
rubbing, alternating the hands/fingers (Figs. 2f and 2g).
28
Experiments with a musical machine
2f 2g
Fig. 2f Exploring the machine with her elbow
...
;Fig.2g
...
and with his fist
Musical improvisations
A preliminary analysis of the improvisations revealed rhythmic and melodic patterns,
synchronisation on the same pulse, forms of song and accompaniment, individual
improvisation styles, brief formal constructions based on imitation, repetition, alternance
and contrast.
Two case studies: the ‘life cycle’ of interaction
The first phase of observation allowed us to observe an initial dynamic curve that
moves from
Surprise
(the
Aha effect
) to a phase of
Excitement
, followed by a period
of
Concentration and analytical behaviour
: we called it the ‘life cycle’ of interaction. In
the two case studies we concentrated on the quality of the analytical behaviours and the
ways in which the child/computer interaction starts, develops, and ends. We will present
the
attention span
(the time of every ‘game’) and the
dynamic profile of interaction
,thatis,
the development of the interaction over the three sessions.
Jerry: case study no. 1
Jerry is 5 years and 10 months old. In the preliminary meeting, he was immediately
interested in the ‘keyboard that answers’, and passed quickly from surprise (Aha effect)
to a more careful and analytic approach, commenting aloud: ‘It repeats ...but isn’t exactly
the same’. His interest in the new ‘instrument’ was stimulated, among other things, by a
knowledge of musical instruments uncommon in children of this age group: his drawing
included a wide range of instruments (drums, guitar, flute, violin, trumpets), and from the
parents’ questionnaire we learned that he listens to classical music. The ‘technological’
aspect, on the other hand, seems to have taken second place. He wanted to manage his
‘meetings’ with the machine by himself. Jerry asked to start the first session in pairs, with
29
Anna Rita Addessi and Franc¸ois Pachet
the Continuator, and then he asked to stop. In session 3 he asked to start by playing alone
with the system.
The order of the ‘games’ was as follows:
Session Tasks
14
2 1,2,3,4
3 2,1,4,3
Attention span
By ‘attention span’ we mean the subjects’ tendency to persist in their contact with the
objects or activities, irrespective of any underlying aim. Jerry’s attention span was measured
for each task, as shown in Figure 3. In the first session, Jerry chose to perform only task
Fig. 3 Jerry’s attention span. The session numbers are shown on the horizontal axis, the
duration in minutes on the vertical axis
4, one of the tasks that lasted the longest. In session 2, though, he clearly preferred task 2,
alone with the system, and this was the task that went on for the longest of all, performed
with great concentration. In session 3 Jerry was again interested in working with the system
alongside a friend, and even though task 2 was again the longest, the difference between
tasks 1 and 4 was very small. His attention was almost identical in sessions 2 and 3 when he
played alone without the Continuator (task 1), and then with a partner and the Continuator
(task 4). Nevertheless, when playing alone Jerry often appeared to get bored, stopping
frequently and sometimes waiting for the system to reply.
An unexpected result was the shorter length of the task involving the partner but
without the Continuator (task 3), where the children seem to have had less fun and did not
listen so carefully. The system therefore appears to motivate not only individual children,
30
Experiments with a musical machine
4a 4b 4c
Figs. 4a/b/c Session 2, task 2: (a) Jerry plays with one finger only, then (b) listens to the
Continuator, then (c) plays again using all his fingers
but also those working in pairs, thus stimulating the socialisation aspect of the musical
experience.
To conclude, then, the attention span analysis would lead us to hypothesise a greater
interest on the part of Jerry in the tasks involving the system, especially when playing alone.
Dynamic profile of interaction
We will now consider the analysis of various fragments that show the dynamic evolution
of the interaction between Jerry and the system over the three sessions. Jerry tended to
start immediately, displaying concentration and analytical behaviour: he observed and
experimented with the rules of the system. To illustrate this conduct, we will start with a
fragment taken from session 2.
Concentration and analytical behaviour
Session 2, task 2 (Figs. 4a/b/c)
. Jerry proceeds
by trial and error, respecting the turn-taking with the system. He starts by systematically
playing first with his index fingers, then two fingers, then with the palm of the hand,
exploring the whole range of the keyboard. The procedure develops in a linear fashion,
almost going from the simple (one finger, middle register) to the complex (two fingers, the
palm, etc.; middle, low and high registers). He always stops and listens to the system’s reply
(turn-taking). He listens carefully, unhurriedly, with a concentrated expression. He behaves
like an observer introducing variables and trying to understand the results. His relationship
with the system might be called ‘symmetric’, a definition used by Fogel (2000) for the
face-to face communication between mother and child. In our particular case, though,
there is not yet any real dialogue between the two partners: Jerry systematically introduces
new elements, eager to deepen his exploration of the system, but the replies given by the
Continuator do not always reflect Jerry’s input.
Repetition and variation: the dialogue
Session 2, task 2 (Figs. 4d/e/f)
. After some
minutes, Jerry plays one note at random (G3
staccato
) and is about to fold his arms and
listen to the machine’s reply, but the Continuator plays back the same note and merely
adds the octave (G3–G4). Jerry recognises his own note as in a mirror: he is surprised,
looks at the keyboard, lifts his hand and then immediately replies with the same note and
a variation (G-G-A-A-B-cluster). This marks the start of a real dialogue based on repetition
31
Anna Rita Addessi and Franc¸ois Pachet
4d 4e 4f
Figs. 4d/e/f Session 2, task 2: (d) Jerry recognises his own notes played by the Continuator,
then (e) a musical dialogue starts based on repetition/variation; (f) the dialogue ends
and variation: Jerry and the system reply to each other and add variations in register,
rhythm, modes of playing (e.g. Jerry plays G
staccato
, Continuator: G-G
staccato
; Jerry:
G-G-A-A-B-cluster, Continuator: cluster/rising arpeggio; Jerry: short cluster, Continuator:
cluster, rising 3rd, etc.). This type of interaction gives us a good idea of how the system is
able to imitate and vary the child’s proposals, and how this aroused in the child a sequence
of emotions going from surprise and interest, to curiosity, which encouraged him to turn
a random single note (G) into an alternating succession of variants of a rhythmic-melodic
cell, making up an interesting, albeit brief, musical dialogue. It is interesting to observe how
this type of interaction seems very close to that occurring between an adult and a child:
‘in the exchange and vocal games [of the child] with the mother, the maternal voice ...
acts as a sound mirror that reflects the vocal experiences of the child and reinforces them’
(Imberty, 2002).
In our example, when the system’s replies became very varied and the mirror effect
vanished, Jerry lost interest, the dialogue ceased, and he asked for the game to stop.
Listening automatism
Session 3 (Figs. 4g/h)
. Jerry asks to start the game alone with
the system (task 2). He starts a strict dialogue with the system, consisting of exploration,
repetition, variation. Jerry plays the first phrase of ‘Fr`
ere Jacques’ (C-D-E-C/C-D-E-C), the
system repeats with variations (C-D-E-F-G-A-B ..). During the next game, without the
Continuator (task 1), a sort of ‘listening automatism’ is instigated: Jerry knows the system
is not connected, but still plays and then stops, automatically waiting for the Continuator
to reply, which does not happen. Jerry starts playing again alone, but still waits every now
and then for a reply (e.g. he puts his hand to his ear). A kind of automatic expectation has
been instituted. The somewhat sad expression that appears on his face, his head leaning
on his hand (Fig. 4h), shows his emotional anticipation, part of a more general affective
relationship that the children had established with the Continuator, which they treated
with care and attention, just like a playmate, as if the machine were human, or at least
a living creature: ‘it answers by itself’ was the most frequent comment heard during the
performances of the system. Jerry’s sad expression brings to mind certain descriptions
of ‘attachment’ observed in children at school when separated from their mothers. The
reference to the theory of attachment is also suggested by applications of this theory in the
field of artificial intelligence, for example to the virtual animal-toys Tamagotchi or Aibo
(Kaplan, 2001).
32
Experiments with a musical machine
4g 4h 4i
Figs. 4g/h/i Session 3, tasks 2, 1 and 4: (g) Jerry faces the Continuator (task 2); (h) he plays,
then listens, but the Continuator does not reply (task 1); (i) Jerry observes the surprise of his
friend on hearing the reply of the Continuator (task 4)
The observer
Session 3, task 4 (Fig. 4i)
. The interaction continues in the following task
(4). Jerry plays with another child and the Continuator. This time Jerry does not only observe
the system, but also his friend’s interaction with the system: he tells him to play, to wait for
a reply from the system, and watches his reaction of amazement. In this phase Jerry shows
some moments of excitement.
Tom: case-study no. 2
Tom is also 5 years 10 months. He soon becomes involved in the sound, and learns very
quickly how to interact ‘in real time’ with all the stimuli that he receives from the system and
from the whole set of equipment (computer, loudspeaker, expander, wires). He manages
with ease to integrate all the various elements into the games with the keyboard and his
partner. Once again the parents’ questionnaire provides useful information, telling us that
Tom knows how to work the TV and the computer. Very significantly, in his drawing he
shows the computer, the wires, the loudspeakers and the expander, just as meticulously as
Jerry drew the musical instruments.
The order of the games was as follows:
Session Tasks
1 2,1,3,4
2 4,3,1,2
3 1,2,4,3
Attention span
Right from the first session we can see that task 4 (with another child and the Continuator)
lasted the longest (Fig. 5), its duration remaining fairly constant throughout the following
sessions, even though the time spent on the other tasks varied considerably: in session 2,
task 2 was the longest (of all the sessions), while the length of the others decreased. On the
other hand, in session 3 task 2 was much shorter, while the others increased.
33
Anna Rita Addessi and Franc¸ois Pachet
Attention span
Tom
0
2
4
6
8
10
Sessions
Duration
(min.)
Task 1 4 1,33 5,38
Task 2 6,52 2,31 4,29
Task 3 3,55 8,57 1,31
Task 4 8,52 7,39 9
I II III
Fig. 5 Tom’s attention span. The horizontal axis shows the sessions, the vertical axis the
duration in minutes
The fact that task 4 lasted the longest, not only in sessions 1 and 3 but also in terms
of the total overall time for each task (1=11.11 minutes; 2 =11.52; 3 =14.22; 4 =25.31),
confirms the trend, already noted in Jerry, that the system enhances socialisation and the
sharing of the experience. In Tom’s case, task 2 was not always very long but was quite
significant in that his attention was completely captured, leading to rich interaction with
the machine. Nevertheless, unlike Jerry he showed a general preference for the interaction
involving all three participants: himself, his partner and the system.
Dynamic profile of interaction
Tom soon achieved a good level of dialogue with the system, learning how to make it
imitate him and how, in turn, to imitate the system. In the last session, as we can see in the
following examples, he reopened his dialogue.
Surprise and assessment
Session 1, task 2 (Figs. 6a/b/c)
. He plays a few notes, the
Continuator replies. Tom recognises the repetition and shows surprise and excitement. The
next reply of the Continuator is much longer than expected and Tom shows disappointment,
saying ‘It never ends’. He nevertheless waits for the system to stop before playing again, thus
respecting the implicit rule of turn-taking. There follows a series of improvisations during
which Tom uses various styles and listens carefully. At a certain point the Continuator
begins to repeat the same note continuously, as if the machine were stuck (maybe actually
due to an error in the working of the system). Tom notices something is wrong and blocks
his ears with an expression of irritation.
Excitement and learning by ‘immersion’
Session 1, task 4 (Figs. 6d/e/f/g/h/i)
.Tom
teaches his friend the rules: play, wait and listen to the reply, as the keyboard ‘plays by
itself’. An intense moment of interaction begins: they play and listen, bringing their ears
34
Experiments with a musical machine
6a 6b 6c
Figs. 6a/b/c Session 1, task 2: (a) Tom recognises his own notes played by the Continuator
(surprise and excitement); (b) Tom listening to the long reply by the Continuator: ‘Non si
ferma’ (‘It never ends’) (turn-taking); (c) Tom puts his fingers into his ears when the system
repeats the same note like a blocked machine
Figs. 6d/e/f Session 1, task 4: (d) Tom stops his friend and teaches him the rules of the
system and turn-taking: ‘Suona da sola’ (‘It plays by itself’); (e) They use the ability of the
system to imitate the sounds they produce in order to enjoy themselves: they play funny
sounds with the aim of (f) exciting and sharing the excitement, listening to the equally
funny reply by the Continuator
closer to the speakers; they play with their hands, heads, bottoms; they even introduce
the ringing of a cell-phone into the game. The rules apply, but also listening, touching,
discovering, playing, having fun, provoking amazement and pleasure. Tom often imitates
his friend. They discover that the system repeats what they play and learn how to make
the system imitate them: the most exciting game is to produce strange sounds (brief
sequences of strong, fast and irregular clusters) for the pleasure of hearing the Continuator
repeat them – just like laughing at your funny faces in the mirror.
The moment of
excitement also becomes the moment of learning
(learning by ‘immersion’: Maragliano,
1999).
‘Dead moments’
Session 2, task 4
. During this phase Tom and his partner interact with
the system, but from afar. They speak to each other while the Continuator is playing, as if
not wanting the system to hear them. They then approach the machine, look beneath the
keyboard, and ‘bite’ the wires. They begin to play again, but not together; they pause, Tom
35
Anna Rita Addessi and Franc¸ois Pachet
6g 6h 6i
Figs. 6g/h/i Various interactive conducts of Tom and his friend: (g) They listen to the
speakers, (h) play the keyboard, with the Continuator and the phone; (i) they both play and
move their heads in synchronisation
plays restlessly, his partner seems to be looking for something, makes several attempts on
the keyboard, plays with his head, his behind, tries out some short rhythmic and melodic
patterns, the system replies with just a few notes, while Tom looks through a book. The
interaction could be said to be passing through a moment characterised by a slowing-
down in the turn-taking, a discontinuity of attention, irregular exploration and a sort of
disorientation. The function of this moment is perhaps to allow the children to readjust
their approach to the interaction, which is soon taken up again with a new type of game.
We have called this phenomenon the ‘dead moment’, or ‘readjustment’, due to its functional
similarities with what Stern speaks of, using the same terms, as a period of adjustment in
the interaction between mother and child (Stern, 1985).
From turn-taking to role-taking
Session 2, task 2 (Fig. 6.j)
. Tom plays with more
concentration and analytically; he tries to understand the system by looking at the screen of
the computer. He plays short musical improvisations. He does not explore, he immediately
creates music with the Continuator: they play short rhythmic and melodic patterns, repeat
and elaborate them, then play short but complex musical phrases. At a certain point he
moves towards the lower register and plays C1. The system responds with C4-A5. Tom
recognises that the system has played the same note as he had but at a higher register and
6j
Fig. 6j From turn-taking to role-taking: Tom plays the C5 just played by the system in reply
to the C1 played by Tom (session 2, task 2)
36
Experiments with a musical machine
says ‘High’; he then goes to the upper register and, imitating the system better than the
system had done with his proposal, plays C5, and then goes away saying ‘Finished’, while
the Continuator plays Bb-A.
Tom has understood the system, has played with it, has learned to make it imitate him
and to imitate the system. It is here, then, that we observe a transition from
turn-taking
,
the alternation between two interlocutors, to
role-taking
, i.e. the moment when one of the
two interlocutors takes the partner into account and as a consequence regulates their own
behaviour according to that of the other. Children are, for example, able to adapt their
language when speaking to children younger than them (Emiliani & Carugati, 1985).
‘Attunement’ and climax
Session 3, task 2
. Tom begins the last session by playing
energetically, the Continuator replies with similar intensity; Tom relaunches softly and
delicately, the Continuator responds with soft and slow notes. For a while they adapt to
each other, not with exactly the same notes, but adopting the same ‘mode’ of playing and
following the sequence of question – answer – relaunch. A written description, time frame
analysis, and formal structure of the dialogue are provided in Figure 7.
What strikes one most of all in this sequence is the regularity in the timing of each
turn, and the fact that both ‘partners’ follow the same dynamic pattern: first
forte
,then
decrescendo
and weak, and finally
forte
again. The overall effect, underlined by the
expressions appearing on Tom’s face at each step, is one of strong initial impetuosity
that is slowly subdued, becoming softer and calmer, but then returning to the initial energy,
and concluding with a reinforcement. It is interesting to note the formal structure of the
overall
dialogue
:
A(a + a a + a b + b’ b” + b”’)
B(c + c’)
A’ (a+aa+aa+a a+a a’+a’)
We are witnessing the creation of a true and proper musical sentence, with questions
and answers, repetitions of short phrases, intermediate variations in dynamics and density,
giving rise to an overall ternary structure, ending with a finale that presents all the qualities
of musical finali: repetition, increased intensity and longer duration. The musical structure
has its origins in the form that the dialogue between Tom and the Continuator takes on
over time: the rules governing the interaction (turn-taking, regular timing turns, imitations,
variation, contrast) also become musical rules.
In this sequence, the system seems able not only to mirror the notes played by the
child, but also to some extent to reflect his affective and dynamic profile. We would not,
of course, wish to infer that a machine could have the intentionality of a human. What
allows the machine to behave in this way is also the speed with which it is able to adapt
to the child’s input and to produce a suitable reply. Without this ability a situation such as
the one described above could not exist.
Since we are not simply dealing with a matter of repetition/variation, but rather
with a situation that is more complex on account of the dynamic, formal and temporal
aspects taken into account, where musical form and emotional experience proceed in
37
Anna Rita Addessi and Franc¸ois Pachet
Total time: 37 seconds.
The turns last around 1 second each, apart from the last ones, which take about
5 seconds each. A very short pause (less than 1 second) sometimes precedes Tom’s
turn.
Sec. Written description Formal structure
0.00 Tom starts by playing energetically with a cluster of
notes,
forte
the Continuator replies with a loud cluster short pause
a+a
0.03 Tom repeats the same thing
so does the Continuator
a+a
0.06 Tom plays a
glissando
, again loudly
the Continuator plays a few notes, loudly
short pause
b+b’
0.09 Tom plays two
glissandos
simultaneously,
decrescendo
the Continuator plays a short sequence of several
notes, continuing the
decrescendo
short pause
b” + b”’
0.12 Tom delicately plays 1 note
ppp
, in the middle
register, and listens to the Continuator’s reply: two
notes,
ppp
, in the lower register
c+c’
0.15 Tom starts playing note clusters again, energetically,
agitated,
forte
the Continuator replies with loud clusters
a+a
0.17 This type of exchange is repeated three more times
Tom listens to the ‘angry’ reply and proceeds with an
irritated and defiant expression on his face. He puts
too much energy into his playing and hurts his elbow
a+a
a+a
a+a
0.26 The 5th time Tom’s input is much longer, then he gets
up to look at the computer screen
The Continuator’s reply is also longer
a’ + a’
Fig. 7 ‘Attunement’ sequence
a reciprocal and parallel fashion, we have called this phenomenon ‘attunement’, due
to certain similarities with the ‘affect attunement’ described by Stern in the interaction
between mother and child. This phenomenon features a type of imitation that is not simply
the repetition of the manifest behaviour of the child, but is to a large extent transmodal,
being based on the correspondence of a state of mind (inferred or directly understood),
and is a process that takes place extremely quickly, with regular timing turns. The three
dimensions of such correspondence are essentially musical qualities: intensity (vocal or
kinetic), rhythm and form (Stern, 1985).
After this phase of ‘attunement’ the dialogue is resumed. Tom gets up, jumps from the
computer to the keyboard, and his movement is mimicked in the music he and the system
38
Experiments with a musical machine
play (Fig. 8a). Delightful and amusing to see, it is truly a moment of genuine creativity. Tom
is no longer exploring the system, they are making music together – a real jam session.
Fig. 8a Session 3, task 2 – climax: Tom plays and moves quickly between the keyboard
and the computer
Relaunching
Session 3, task 4
. In the final session, Tom relaunches the repeated
notes that he had considered a fault in the machine during the first session. A three-
sided interaction is set up based on this musical idea: sequences of two notes repeated
rapidly, slowing down, speeding up, followed by clashes between the groups of notes
played by Tom in the middle range and single notes repeated in the high register by his
partner, interspersed with occasional moments of synchronisation when they both follow
the same pulse. The Continuator replies with a long sequence using a single repeated note,
which stimulates the children and prompts them to carry on with this new musical idea. It
is interesting to see how Tom develops a musical idea that stemmed from an error of the
machine, thus turning it into a ‘creative error’.
This particular performance of task 4 provides a good instance of ‘scaffolding’, where
the Continuator not only re-proposes the musical ideas to the child who first produced
them (as happened in the individual sessions), but also ‘refracts’ them on the other partner,
and in so doing encourages both children to contribute to the building up of an overall
musical construction. This is a fine example of how to set up group improvisations.
The session concludes with a relaunching of the previous explorations, interspersed
with pauses when they discuss what to do. Their finale offers a triumph of technology: the
Roland expander, with its lights, becomes a bomb to be defused (Fig. 8b).
Fig. 8b Session 3, task 4: Tom and his friend are defusing the ‘bomb’ (Roland expander)
under the table. They play and watch the lights on the screen of the expander
39
Anna Rita Addessi and Franc¸ois Pachet
Discussion
In the two case studies presented above we can observe an interaction between the children
and the system that builds up over time, passing through various dynamic states which do
not necessarily follow a linear order. We shall now underline the significant aspects of these
two cases, interpreting them in the light of developmental theories and more especially of
musical development in children. The interpretation of these two case studies allows us
to give some answers to the questions raised at the beginning of this paper regarding the
nature of interaction between a child and a machine, especially with the particular system
used in this experiment.
Nature of the interaction
The two case studies would suggest that the Continuator is able to develop interesting
child/computer interaction, very similar to that between humans. This phenomenon seems
to have its origins in the ability of the system to replicate the musical style of the children.
The interaction based on repetition/variation allows the children to organise their musical
discourse, passing, as in the case of Tom, from exploration to genuine musical invention.
In particular, we note that the moment of climax arrives when the two partners adapt to
each other’s ‘style’ of producing sound and accelerate the times of the turn-taking; once
this has been achieved the interaction is concluded, almost like a gesture of liberation from
the accumulated tension. A similar structure based on repetition and variation, pauses
for readjustment, and temporal dynamics has also been observed by Stern (1985) in the
vocal relationship between mother and child, and by Imberty (2002) in the field of music.
Similar interactions have been observed recently in young children and adults playing a
xylophone (Young, 2004): this kind of interpersonal dimension is a potential source of
musical creativity for young children. The very fact that the interaction is so similar to that
of humans may perhaps explain why the children find it so exciting: just like in cartoons,
where the thing they like most is that ‘it seems real because it’s fake’ (Mattia, 3 years old,
in Mazzoli, 2001).
In fact, despite the apparent simplicity of the mechanism, the Continuator generates
very complex reactions, where the children are expected to form judgements about ‘self’
and ‘other’, and to assume the point of view of the other in order to judge their own self.
In the literature these passages are considered crucial for the building of the child’s self:
the Continuator, by means of its mirror effect, could be said to represent the construction
of a ‘musical’ self, or, in the words of Turkle (1984), a ‘Second self’, where the machine not
only seems to ‘think’, but thinks like the user.
Rules of the system and musical rules
The children learned the rules of the system: it replies by playing alone, it replies when
you stop playing (turn-taking), repeats what you play, repeats with variations (or ‘errors’),
is capable of establishing a dialogue made up of repetition/variation, it does not always
respect the rules, you can teach the system, and the rules of the system can be taught to
others. During this process the children reacted if the system did not respect the rules: Tom
40
Experiments with a musical machine
showed disappointment when the system began to repeat the same note like a blocked
machine, and corrected it when it repeated inaccurately what he had played. In the end,
though, he reused an error of the system, the repeated notes, to invent music.
We observed two kinds of learning style. In the first case study Jerry learned the rules
of the system in a ‘linear’ way, moving from the simple to the complex, by trial and
error; in the second case Tom learned to use the system by putting all his senses into
his involvement with the system and other instruments, and as a result the moments of
excitement and learning coincided. We are witnessing two different styles of learning,
which have been defined as ‘linear’ and ‘by immersion’. The former is more typical of
the technologies associated with writing, such as books, while the latter is more linked to
multimedia technologies (Maragliano, 1999; Mazzoli, 2001). In both cases the system has
stimulated a learning strategy for problem solving: during the interaction the children not
only identified the problems of interacting with the system, including the rules governing
musical language, but discovered a solution to these problems too.
In pairs or alone, with or without the Continuator
The two tasks involving the system gave rise to the longest attention spans and showed how
most children reach a stable level of attention characterised by a strong and continuous
interest in the interaction. The time also increased considerably in task 4 when the
two children played with the system. The system therefore appears not only to motivate
individual children, but also children working in pairs, thus stimulating the socialisation
aspect of the musical experience. These data, together with the phenomenon of Surprise and
Excitement observed, could be interpreted as signs of
intrinsic motivation
that stimulated the
children’s interest and pleasure in using the machine and its musical and interactive games.
From a pedagogic point of view this aspect is of utmost importance since it stimulates
learning and creativity, as well as encouraging an interest in musical instruments, which
are normally less attractive to such young children (O’Neill & McPherson, 2002).
Listening
The listening conducts are particularly rich and varied: concentrated, analytical, but also
symbolic and creative. The children often ‘dramatise’ the sounds they hear, giving them a
narrative form or an expressive representation. We can see here an important difference
regarding the educational strategies of mirroring, modelling and scaffolding described
by various authors, where the sound is just one perceptual channel among others. In
our case hearing is the main channel, and at times the only one, through which the
children communicate with the system. This factor gave rise to some particularly careful
and prolonged bouts of listening, encouraging the children to ‘think in sound’. A further
significant aspect is the quality of the children’s listening to their own productions while
they played, heightened by the interactive element that encourages them to listen carefully
so as to compare their own pieces with the reply and new proposal of the system, and to
identify repetitions and differences. As has already been reiterated many times in the world
of teaching, listening to one’s own musical productions while playing is one of the main
objectives of music education (Delalande, 1993; Frapat, 1994). The quality of the child’s
41
Anna Rita Addessi and Franc¸ois Pachet
explorations, improvisations, compositions and performances depends essentially on the
child’s ear, and it is this aspect that prompts the child to reproduce an invention, vary it,
play with the variation and create special effects.
Music-makers with style
The system stimulated and reinforced conducts of an explorative type, during which the
child’s actions were coordinated with the purpose of exploring the new partner, and which
were characterised by the systematic introduction of new and different elements. But it
also prompted inventive conducts, where the aim of the child’s actions appeared to be to
elaborate particular sounds and musical ideas and to undertake a dialogue with the system
through the sounds.2Both in the exploration and in the improvisations themselves, we
can see very personalised styles in the children’s approach to producing sounds, in their
handling of the instrument and other equipment, and their working out plans of action
to satisfy their own goals. The Continuator is able to reinforce these individual styles,
and allows them to develop and evolve. We have observed that the ‘teaching method’ is
based on turn-taking and regular timing turns, on the strategies of mirroring, modelling
and scaffolding, and on starting up ‘affect attunement’, intrinsic motivation, collaborative
interaction and joint attention. One of the most interesting aspects is that the invention is,
in the end, not individual but collective: the child is playing along with the machine, in a
pair, like two musicians improvising together. In case study 2, in particular, we see how the
system teaches the child to play with it, by guiding him from exploration towards musical
invention, just like a real teacher. The analysis of these two case studies has provided us with
a basis on which to perform a systematic study of the improvisations of the other children
who participated. The way the children play also shows their stylistic competence, not only
as listeners, as previous research has found (Addessi
et al
., 1996; Hargreaves & North, 1999;
Marshall, 2001), but rather as music-makers: Tom played standing up, moving a lot, his
sleeves pulled down over his hands, displaying an intense physical relationship with the
instrument; Jerry always played while seated, in a composed state. The questionnaires
tell us that Tom’s father is an expert in rock music, whereas Jerry listens to classical
music.
The Continuator as a flow machine
Finally, in case study 2 it is possible to recognise the conditions described in the theory of
flow by Csikszentmihalyi (1990): distractions are excluded from the consciousness; action
and awareness are merged; there is immediate feedback to one’s actions; step by step, the
activity becomes autotelic. The notion of flow defines the so-called ‘optimal experience’ as
a situation in which people obtain an ideal balance between skills and challenges. We can
say that in both case studies a balance between challenges and skills is achieved. Other
states can also be described in terms of the balance between skills and challenges, such
as anxiety or boredom, but the flow state is fundamental in order to develop creativity.
We might think of the Continuator as a flow machine in the sense that it produces a
response corresponding to the skill level of the user (see Pachet & Addessi, 2004 for further
discussion).
42
Experiments with a musical machine
Conclusion
In this paper we have presented a study carried out at the University of Bologna about
children working with a particular interactive system, the Continuator, created at the
Sony-Computer Science Laboratory in Paris. We have described the setting in which our
observation took place and offered an analysis of two case studies. The results provide some
important categories for further observation and interpretation of data and make it possible
to formulate various hypotheses about the nature of the interaction between children and
interactive systems, in particular between children and the Continuator.
The Continuator would appear to be able to elicit interesting child/computer
interaction, and seems to activate interactional microsystems similar to those observed
in infant/adult communication (we refer in particular to Stern, 1995 and Imberty, 2002),
but with a more mechanical and computational approach (Turkle, 1984), which would
perhaps explain why the children found it so exciting. The system’s ability to maintain
the children’s attention for long periods of time, remarkable for this age, together with the
phenomenon of Surprise and Excitement observed, suggest that while interacting with the
system the children reach high levels of well-being and creativity, similar to those described
in the theory of flow (Csikszentmihalyi, 1990). As such, the Continuator is just one instance
of a larger class of systems that could be called
reflective
, in which users can play with
virtual copies of themselves, or at least agents who have a mimetic capacity and can evolve
in an organic fashion. We believe that most of the interesting properties highlighted in our
experiment probably stem from this particular characteristic. From this point of view, the
system has proved to be an excellent research tool for studying the genesis of the processes
of invention and musical interaction in children.
From a pedagogical point of view, one of the most significant results is that the children
are helped to develop very attentive listening skills, as well as creative musical conduct and
a personal music improvisation style, based on their own ability and musical knowledge.
This, in our opinion, is no small result, since despite its great importance the teaching of
improvisation is still rarely tackled in Western formal music education (McPherson, 1994;
Kenny & Gellerich, 2002).
As far as the use of new technology for the purposes of music education is concerned,
we think that the encouraging results obtained were also due to other features of this kind
of system. A full analysis of the properties of reflective interactive music systems as tools
for enhancing creativity can be found in Pachet (In press). We shall limit ourselves here
to pointing out some of the most obvious features. First of all, the children only interact
by playing, without other graphic or mechanical interfaces (e.g. mouse, buttons, switches,
etc.). The system therefore possesses the properties of
transparency
, involving ‘a shift from
the representation of music to the music itself’ (Folkestad
et al
., 1998: 95), and
reflection
,
in the sense that it is the system itself that helps the user to understand the mechanism
of interaction. Second, it avoids the monotony of mere repetition: inane repetition has
been indicated as one of the negative aspects of many automatic systems (De Kerckhove,
1993). Third, it is not programmed with fixed musical objectives (e.g. ear training, chord
recognition, etc.), and therefore fosters the pleasure of not knowing what will happen.
Finally, the children’s timing seems to be respected: the overall interaction remains fluid,
passing through moments of serious concentration, but also moments where the interaction
43
Anna Rita Addessi and Franc¸ois Pachet
dwindles. Furthermore, the children are able to interrupt the game when they want, thus
preserving the factor of ‘distance’ between child and machine, vital from an aesthetic and
pedagogic point of view (Bertolini & Dallari, 2003).
We are now preparing an observation grid to analyse the performance of all the children
that took part in the protocol and to check the categories established so far, including the
influence of age. We are analysing the musical improvisations in detail and checking all
the psychological states described by the theory of flow.
In the light of these results, the project foresees the experimentation of new protocol
for interaction and new variants to be applied to interactive reflective musical systems. We
believe an approach consisting of the close integration of psychological experiments and
system design to be very productive and one that should be pursued.
Acknowledgements
We would like to thank the director, the teachers, the children and their parents of the Scuola
Materna Statale ‘La Mela’ (Bologna, Italy); Simona Carlotti, Laura Ferrari and Nicoletta Rossi
for their help in realising the protocol; Fabio Regazzi from the University of Bologna for
technological assistance.
Notes
1 French:
conduite
; Italian:
condotta
;German:
Betragen
; English:
conduct
. The term appears above all in
the literature of countries with Latin-based languages and is often used as a synonymn of ‘behaviour’.
However, ‘It can be distinguished from the latter in that behaviour refers to the set of habitual actions
and reactions of an organism in an environment where objective observation is possible, whereas
conduct refers to a deeper interior level where these actions and reactions originate’ (Galimberti,
1992: 214). Piaget defines conducts as the ‘behaviours, including the conscience’ (Piaget & Inhelder,
It. 1970: 7). In the musical field the concept has been used by Delalande (1993): ‘Reasoning in terms of
conduct as opposed to behaviour means trying to understand the function of the acts. When someone
picks up their instrument, prepares to play and then plays, what are they looking for, what do they
expect from this set of coordinated actions? It is the purpose itself that helps us define the musical
conduct’ (p. 45).
2 The distinction between exploration and invention is not always clear cut, and from a theoretical point
of view scholars are not in agreement on this matter. Our standpoint is based primarily on what we have
observed, and takes inspiration from the differences proposed by Delalande (1993) and Frapat (1994)
between
exploration
and
invention
, and by Kratus (1994) between
exploration
and
development
.
References
ADDESSI, A. R., BARONI, M., LUZZI, C. & TAFURI, J. (1996) ‘The development of musical stylistic
competence in children’.
Bulletin of the Council for Research in Music Education
,127, 8–15.
BERTOLINI, P. & DALLARI, M. (2003) ‘A proposito di giudizio estetico e mass media’, in A. R. Addessi &
R. Agostini (Eds.),
Il giudizio estetico nell’epoca dei mass media
, 93–118. Lucca: LIM.
CAMAIONI, L., BASCETTA, C. & AURELI, T. (1988)
L’osservazione del bambino nel contesto educativo
.
Bologna: Il Mulino.
CAMURRI, A. & COGLIO, A. (1998) ‘An architecture for emotional agents’.
IEEE Multimedia
, October, 2–11.
CSIKSZENTMIHALYI, M. (1990)
Flow: The Psychology of Optimal Experience
.NewYork:Harper&Row.
DE KERCKHOVE, D. (1993)
Brainframes: Mente, tecnologia, mercato
. Bologna: Baskerville.
44
Experiments with a musical machine
DELALANDE, F. (1993)
Le condotte musicali
. Bologna: CLUEB.
EMILIANI, F. & CARUGATI, F. (1985)
Il mondo sociale dei bambini
. Bologna: Il Mulino.
FOGEL, A. (2000) ’Oltre gli individui: un approccio storico-relazionale alla teoria e alla ricerca sulla
comunicazione’, in M. L. Genta (Ed.),
Il rapporto madre-bambino
, 123–57. Roma: Carracci.
FOLKESTAD, G., HARGREAVES, D. J. & LINDSTR ¨
OM, B. (1998) ‘Compositional strategies in computer-
based music-making’.
British Journal of Music Education
,15, 1, 83–97.
FRAPAT, M. (1994)
L’invenzione musicale nella scuola dell’infanzia
. Bergamo: Junior (first published 1990).
GALIMBERTI, U. (1992)
Dizionario di Psicologia
. Torino: UTET.
GROUPE DE RECHERCHE MUSICALE (GRM) (2000)
La musique ´
electroacoustique
, CD-ROM, INA-GRM.
HARGREAVES, D. J. & NORTH, A. C. (1999) ‘Developing concepts of musical style’.
Musicae Scientiae
,3,
193–216.
IMBERTY, M. (2002) ‘Il bambino e la musica’, in J.-J. Nattiez (Ed.),
Enciclopedia della Musica
,2, 477–95.
Torino: Einaudi.
IRCAM-MUSICLAB (2002)
6 Interactive Music Applications for Music Teaching in the National Education
,
http://www.ircam.fr/produits/technologies/multimedia/musiclab-e.html.
KAPLAN, F. (2001) ’Artificial attachment: will a robot ever pass Ainsworth’s Strange Situation Test?’, in
S. Hashimoto (Ed.),
Proceedings of Humanoids 2001: IEEE-RAS International Conference on Humanoid
Robots
, 125–32.
KENNY, B. J. & GELLERICH, M. (2002) ’Improvisation’, in G. McPherson & R. Parncutt (Eds.),
The Science
and Psychology of Music Performance: Creative Strategies for Teaching and Learning,
117–34. Oxford:
Oxford University Press.
KRATUS, J. (1994) ’The ways children compose’, in H. Lees (Ed.),
Musical Connections: Tradition and
Change
, 128–41. Proceedings of the 21st ISME Conference, Tampa, Florida. Auckland, NZ: Uniprint,
The University of Auckland.
MARAGLIANO, R. (1999)
Nuovo manuale di didattica multimediale
. Bari: Laterza.
MARSHALL, N. A. (2001) ‘Developing concepts of musical style’. Unpublished PhD thesis, University of
Durham.
MAZZOLI, F. (2001)
C’era una volta un re, un mi, un fa
...
Nuovi ambienti per l’apprendimento musicale
.
Torino: EDT.
MAZZOLI, F. (2003) ’I suoni abitati’, in F. Mazzoli, A. Sedioli & B. Zoccatelli,
I giochi musicali dei piccoli
,
7–17. Bergamo: Edizioni Junior.
MCPHERSON, G. (1994) ’Improvisation: past present and future’, in H. Lees (Ed.),
Musical Connections:
Tradition and Change
, 154–62. Proceedings of the 21st ISME Conference, Tampa, Florida. Auckland,
NZ: Uniprint, The University of Auckland.
MIALARET, J.-P. (1997)
Les explorations instrumentales chez les jeunes enfants
.Paris:PUF.
O’NEILL, S. & MCPHERSON, G. (2002) ’Motivation’, in G. McPherson & R. Parncutt (Eds.),
The Science
and Psychology of Music Performance: Creative Strategies for Teaching and Learning,
31–46. Oxford:
Oxford University Press.
PACHET, F. (2003) ‘Musical interaction with style’.
Journal of New Music Research
,32, 3, 333–41.
PACHET, F. (In press) ’Enhancing individual creativity with interactive reflective musical system’, in
I. Deli`
ege & G. H. Wiggins (Eds.),
Musical Creativity: Current Research in Theory and Practice
.Hove:
Psychology Press.
PACHET, F. & ADDESSI, A. R. (2004) ‘Children reflect on their own playing style: experiments with
Continuator and children’.
ACM Computers in Entertainment
,1, 2, 14–14.
PIAGET, J. & INHELDER, B. (1966)
La Psychologie de l’enfant
. Presses Universitaires de France: Paris (It.
trans. 1970).
RESNICK, M., BRUCKMAN, A. & MARTIN, F. (1996) ‘Pianos not stereos: creating computational
construction kits’.
Interactions
,3, 6, 40–50.
STERN, D. (1985)
The Interpersonal World of the Infant
. New York: Basic Books.
45
Anna Rita Addessi and Franc¸ois Pachet
TREVARTHEN, C. (2000) ‘Musicality and the intrinsic motive pulse: evidence from human psychobiology
and infant communication’.
Musicae Scientiae
, Special Issue 1999–2000, 155–215.
TURKLE, S. (1984)
The Second Self: Computers and the Human Spirit
. New York: Simon & Schuster.
WEBSTER, P. R. (2002) ‘Computer-based technology and music teaching and learning’, in R. Colwell &
C. Richardson (Eds.),
The New Handbook of Research on Music Teaching and Learning
, 416–39.
Oxford: Oxford University Press.
WEINBERG, G. (1999)
Expressive Digital Musical Instruments For Children.
MS thesis. MIT Media
Laboratory.
YOUNG, S. (2004) ‘The interpersonal dimension: a potential source of musical creativity for young children’.
Musicae Scientiae
, Special Issue 2003–4, 175–91.
Video and musical examples
Example of music created by the Continuator can be found on the web site,
www.csl.sony.fr/∼pachet. Video examples will be included on the next
BJME
CD.
46
