Proceedings paper

 

MUSICAL SCHEMATA IN CHILDREN FROM 10 TO 12 YEARS OF AGE:

A STUDY ON SEGMENTATION AND MENTAL LINE ELABORATION

Dimitra Koniari (1), Marc Mélen (2 & 3) and Irène Deliège (3)

(1) Department of Musical Studies

Aristotelian University ofThessaloniki

(2)Chercheur qualifié au Fonds National de la Recherche Scientifique

(3)URPM - CRFMW

Department of Arts and Sciences of Music

University of Liège

 

  1. Introduction
  2. Numerous developmental studies have shown that at least some of children's knowledge is organized as schemata forms for familiar events, objects, people or places and have focused on the role of schemata organisation in children's memory (see for a review Davidson, 1996). The present study is concerned with the ability of musician and non-musician children (from 10 to 12 years of age) to form musical schemata during listening to a piece. As stressed by Neisser (1976, p.54)  A schema ... is internal to the perceiver, modifiable by experience and somehow specific to what is being perceived. The schema accepts information as it becomes available at sensory surfaces and is changed by that information.  Thus, the term schemata refers to mental structures which organise information received from our senses and are continuously altered by that information. Regarding music information derived from listening to a piece, the elaboration of a schema, as pointed out by Deliège (1997), should be understood as a reduction of the musical piece, rather than the reconstitution of its score.

    The main problem in understanding the cognitive processes underlying real-time listening are related to memory for events evolving in time. In the late eighties, Deliège (1987 for a first sketch) suggested that listening should be considered as a schematisation process built on cues picked up from the musical surface by an abstraction mechanism (referred to as cue extraction in Deliège 1987; 1989; Deliège & El Ahmadi 1990). The role of this mechanism is to provide landmarks of the temporal flow of the musical piece and to generate the segmentation and categorization of the musical structures (for a more general view of the model, see Deliège a & b, and Mélen & Wachsman, this symposium). More closely in relation with the mental processes involved in the present experiments, i.e. the elaboration of a mental line of a piece of music, are the concepts of "cognitive maps" and of "carte mentale" put forth, respectively by Tolman (1948) and Pailhous (1970), as cited by Deliège (1991; 1998), which suggest that the animal or the individual are building up some kind of maps to summarize a more important amount of information. Indeed, this idea is likely to be rather close to the processes involved in the elaboration of a mental line during listening to music.

    The validity of this last proposal was tested, with adults musicians and non-musicians, using pieces from the contemporary repertoire (Deliège 1989) and the cor anglais solo from Tristan und Isolde by Wagner (Deliège 1998). However, the development of this process has not been studied in details yet. Mélen (1999; Mélen & Wachsman, this symposium) found evidence that elements of the cue abstraction mechanism should already exist in infancy. These findings are in accordance with the general remark of Dowling (1999) that many of the perceptual mechanisms of adults for music processing are found to be built on elements already present in infancy. Deliège & Dupont (1994) and Mélen and Deliège (1995), in a series of experiments with children and adults, in relation to cue abstraction, also supported evidence that this mechanism might already be present at these ages. Giving these arguments, the questions which are still waiting for response are: (i) what is the development of the cue abstraction process and (ii) to what extent does music experience and training influence its role in the mental representation of a musical piece during real time listening.

    Several experimental methodologies have been elaborated to investigate the cue abstraction hypothesis. In the present study two methods are used: a) the segmentation process, which underlines the organisation of musical events into groups and b) the mental-line procedure which examines the participant's ability to reconstruct the piece after listening. These methods constitute the experimental designs from Deliège's study with adults (1998).

    The pieces used in the experiments reported here are chosen in a children's repertoire: the Rondo Finale of the Sonatina no 2 in C by Anton Diabelli and the Laendler no 10, extract of the Dances for Piano vol 1, D145, op. 1 by Franz Schubert. These pieces were chosen, on the one hand, as it was very unlikely that they would be familiar to any of the participants and, on the other hand, because they presented an archetypal structure in respect to the classical forms of tonal compositions (alternation of variations of an A and B motif).

     

  3. The segmentation task
  4. 2.1. Method

    Participants: There were forty-one (41) children attending Grade 5 classes and drawn from a primary school and a Musical Academy in Brussels (20 girls, 21 boys ; mean age 11 years 1 month, ranging from 10 years 2 months to 11 years 8 months). They were divided in two groups depending their music training: twenty-one children had never followed music lessons (NM) and twenty children had received a continuous music training, two hours per week, for at least 2 years and maximum 3 years (solfège and instrument) (M). Within the groups, children were assigned to two conditions depending on the familiarisation factor (i.e. number of presentations of the piece - 1 or 3 - before the segmentation task). The groups will be hereafter refered to, respectively, following the factor training and the number of listenings, as: NM1, NM3, M1, M3. All children participated in the experiment of segmentation of two short musical pieces, alternating within groups the order of presentation of the pieces.

    Experimental materials and equipment: Two short pieces from the classical piano repertoire:

    a) the Rondo Finale of the Sonatina n 2 in C by Anton Diabelli (total duration 24'') performed by Pietro Galli (Cassiopée 965256) and b) the Laendler n 10, extract of the Dances for Piano vol. 1, D 145, op. 1 by Franz Schubert (total duration 25'') performed by Alice Adler (Chant du Monde, LDC 278876).

    The presentation of the music was generated using a recording of real piano sounds. For playback during the task two loudspeakers connected with a Macintosh Power PC were placed equidistant from the participant. The MAX software (version 2.5) controlled the data collection.

    Procedure: All participants were tested individually in one session for both pieces. They first completed a questionnaire concerning their identity and musical experience. The volume of the music to be listen to during the tasks was adjusted to a comfortable level. The tasks for the two pieces (segmentation and reconstruction) were then carried out: first the segmentation task, second the reconstruction of the piece. A ten-minute break was taken after the testing of the first piece. The two pieces were presented in alternate order for each participant.

    For the segmentation task, the experimental design was built in two phases. In the first phase a performance of the entire piece was presented 1 or 3 times, depending on the group. Participants had to listen carefully to the music, as if they would be listen to a story, and to locate in their mind (inherently) the moments corresponding to a punctuation. In the second phase the performance was repeated three times and participants had to introduce their punctuations (here refered to as "segmentations") by pressing the space-bar of the keyboard. The first time was to become familiarised with the experimental material, and the two following ones constituted the experimental data and were thus recorded.

    Results and comments: Figure 1 and 2, respectively, show a general view of the segmentations of the piece by Diabelli and Schubert. Black and white columns represent, respectively, the segmentations of the first and second recordings, and dashed columns are related to the number of segmentations that were confirmed by the same participant and at the same place of the piece. As can be seen, a maximum of 12 segmentations were recorded for the Schubert piece and 14 for Diabelli, but all of these were not necessarily introduced by each participant. However, it is worth noting that all the segmentations that were recorded, coincided with the main articulations of the piece, as they would appear in a classical morphological analysis: i.e ends of musical phrases and motifs.

    (a) M1 & M3 (b) NM1& NM3

    Figure 1: Segmentations of Diabelli by musician (a) and non-musician (b) children. Black columns represent segmentations during the first recording, white columns segmentations during the second recording. The dashed columns indicate segmentations that were confirmed by the same participant at the same moment.

    (a) M1 & M3 (b) NM1& NM3

    Figure 2: Segmentations of Schubert by musician (a) and non-musician (b) children. Black columns represent segmentations during the first recording, white columns segmentations during the second recording. The dashed columns indicate segmentations that were confirmed by the same participant at the same moment.

     

    a) The segmentation choice

    Again, as in previous experiments with adults (Deliège 1989, 1998), some segmentations are more often chosen (see figure 1 and 2). For example, in Schubert (figure 2), it appears that four segmentations (ns 3, 6, 9 and 12, respectively, end of bar 4, end of bar 8 - central cadence to the dominant -, end of bar 12 and end of the piece - final cadence) are common in most participants' choices. In Diabelli (figure 1) only two segmentations predominate in all groups: n 7, end of bar 8 (central cadence to the dominant), and n 14, end of the piece (final cadence).

    Although, the two pieces share the same morphological structure (two periods of 8 bars divided into two phrases of 4 bars, constituted by 2 motifs A and B, of 2 bars each, and their variations), participants appeared more sensitive to the four main articulations of the piece by Schubert than by Diabelli. A possible explanation for this might reside in the difference of tempo of the pieces: the Schubert's tempo being slower than the Diabelli. Additionaly, the flow of Diabelli's temporal rate is more continuous than the Schubert one, in which some kind of pauses are heard, indicating appropriate segmentation places.

    No significant effects were found in the 2(training) x 2(familiarisation) x 2(composer) ANOVA mixed-design test (p>0.05), showing that the segmentation processs is not significantly influenced neither by musical training, nor by familiarisation or by composer. This tendency was already observed in adult musicians and non-musicians results, a reason for Deliège (1998) to consider the segmentation process as a rather automatic psychological behaviour.

    b) The stability of the participants in their segmentations

    As Deliège (1998) has observed in experiments with adults, the number of segmentations seems to be related to the personal psychological behaviour of the participant. Results in this study show the same general tendancy in children (with mean distance average 1.2). A child who is segmenting generously during the first segmentation task, will be repeating a similar behaviour in the second one; and in the opposite, a more economical behaviour will be also maintained. However there were two exceptions in the M1 Diabelli group, i.e. children who segmented 9 and 13 times and 11 and 5 times respectively in the first and the second segmentation task.

    An analysis of the coherence of the segmentations was developed. In a given overall segmentation task of each child the number of the non-repeated segmentations of the first task was added to the number of the new ones founded on the second segmentation task. By this way the distance of each child from a stable performance was calculated. The results were analysed by an overall mixed model ANOVA 2(training) x 2(familiarisation) x 2(composer). Data showed significance only for training (F(1,36)=6.278, p=0.0169), supporting the hypothesis that musician children are more stable in their segmentation choice. The difference in segmentation between participants of the two familiarisation groups was not significant and no interaction was observed.

     

  5. The reconstruction task: "mental line process"
  6.  

    3.1. Method

    Participants: The children having participated in the segmentation task were here employed.

    Experimental materials and equipment: The same pieces by Schubert and Diabelli were each cut into 8 segments of different lengths at the ends of musical phrases. The segments were transmitted via MIDI interface and the MAX software into 8 keys of a device named ScaleGame (for details see Deliège, Delges, Oter & Sullon, 1998), which permitted the real time listening of the musical information sent in each key.

    Procedure: All participants were tested individually. They had been informed that they would be invited to reconstruct the piece, after 4 or 6 listenings, by using 8 keys of the device ranged in a different random order for each subject. The duration of the task was not limited. The MAX software (version 2.5) collected the data.

    Results and comments:

    a) Schubert: 3 children out of the total of 41 rebuilt correctly the piece (7.56%). They were all members of the musicians group (1 of the M1 group and 2 of the M3 one). Within the 38 wrong reconstructions an additional analysis was performed to observe if the children had been sensitive to the deep structure of the piece, i.e. alternance of variations of motifs A and B. These results are interesting but not yet significant. 10 children in the remaining 38 (28.9%: 5 M and 5 NM) were sensitive to the deep structure of the piece. However, as in similar research with adults (Deliège 1998), an effect of primacy and recency was found. 14 children (36.8%: 9 M and 5 NM) chose correctly the first key for the beginning and 18 children (47.3%; 9 M and 9 NM) chose the last key to finish the rebuilding of the piece. Table 1 shows the distribution parameters -primary mode, mean, and mean distance- for the key positions chosen by all the participants. Table 2, on the other hand, refers to the possible locations chosen by the participants for each key.

     

    Table 1

    Distribution parameters -primary mode, mean, and mean distance- for the key positions chosen by all the participants in the Schubert piece.

     

    M1

     

     

    M3

     

     

    NM1

     

     

    NM3

     

     

    Segm

    Modes

    Mean

    M. Dist.

    Modes

    Mean

    M. Dist.

    Modes

    Mean

    M. Dist.

    Modes

    Mean

    M. Dist.

    1

    1

    3

    2

    1

    2.5

    1.5

    1

    2,9

    1.9

    7

    4,3

    3.3

    2

    2/6

    5.1

    3.1

    2

    4.3

    2.3

    2/6

    4,5

    2.5

    2

    4,1

    2.3

    3

    3

    4.9

    2.1

    3

    4.5

    1.7

    5

    4,5

    2.4

    4

    5

    2.2

    4

    4

    4.9

    1.3

    4

    4.7

    2.1

    2

    3,3

    1.6

    5

    5,2

    1.8

    5

    3

    4.1

    1.2

    5

    4.4

    0.8

    6

    5

    2

    5/7

    4,7

    1.9

    6

    1

    4.9

    2

    6

    5.2

    1.6

    8/7

    4,8

    2.3

    3

    3,5

    3

    7

    2

    3.7

    3.3

    7

    4.7

    2.4

    7

    5,4

    1.7

    2/1

    2,8

    4.2

    8

    8

    5.2

    2.7

    8

    5.7

    2

    8/4

    5,5

    2.4

    8

    6,4

    1.6

     

    Table 2

    Localisation attributed by participants to the eight segments

    In relation with the eight possible locations (bold characters), the numbers in the columns indicate which segments have been localised in that place, and the number of participants (in parentheses) who have attributed this location to this segment.

    Segm

    1

    2

    3

    4

    5

    6

    7

    8

     

    1(17)

    1(1)

    1(2)

    1(5)

    1(3)

    1(6)

    1(7)

    1(1)

     

    2(1)

    2(12)

    2(4)

    2(4)

    2(3)

    2(5)

    2(9)

    2(4)

     

    3(8)

    3(3)

    3(10)

    3(5)

    3(7)

    3(6)

    3(2)

    3(1)

     

    4(2)

    4(3)

    4(4)

    4(9)

    4(5)

    4(2)

    4(5)

    4(12)

     

    5(8)

    5(4)

    5(5)

    5(5)

    5(12)

    5(4)

    5(4)

    -

     

    -

    6(13)

    6(5)

    6(4)

    6(5)

    6(7)

    6(2)

    6(6)

     

    7(6)

    7(2)

    7(9)

    7(4)

    7(4)

    7(6)

    7(11)

    -

     

    -

    8(4)

    8(3)

    8(6)

    8(2)

    8(6)

    8(2)

    8(18)

     

    b) Diabelli: Results are slightly better for both groups but the dominance of the musician group is relevant. 3 musicians (M3) and 1 non-musician (NM1) out of 41 children correctly reconstructed the piece (9.75%). Within the 37 wrong reconstructions, 8 respected the deep structure (21.6%; 7 M and 1 NM). As for the primacy and recency effect, 23 children (62.1%; 10 M and 13 NM) put correctly the first key and 33 children (89.1%; 16 M and 17 NM) ended correctly with the last key. Table 3 shows the distribution parameters -primary mode, mean, and mean distance- for the key positions chosen by all the participants. Table 4 shows the possible locations chosen by participants for each key.

     

    Table 3

    Distribution parameters -primary mode, mean, and mean distance- for the key positions chosen by all the participants in the Diabelli piece.

     

    M1

     

     

    M3

     

     

    NM1

     

     

    NM3

     

     

    Segm

    Modes

    Mean

    M. Dist.

    Modes

    Mean

    M. Dist.

    Modes

    Mean

    M. Dist.

    Modes

    Mean

    M. Dist.

    1

    5

    3

    2

    1

    2

    1

    1

    3.2

    2.3

    1

    1.5

    0.5

    2

    2

    3.8

    2

    2

    3

    1.2

    5

    4

    2.2

    5

    4.7

    2.5

    3

    3

    3.3

    1.5

    3

    3.4

    1

    3

    3.7

    1.4

    3

    4.1

    0.9

    4

    4

    4.5

    1

    4

    4.3

    0.7

    2

    3.3

    1.5

    4

    4

    1

    5*

    6

    4.1

    1.5

    5

    4.1

    1.1

    5

    4.6

    1.3

    5

    4.5

    0.9

    6

    6

    5.2

    1.4

    6

    4.8

    1.6

    7

    4.9

    1.8

    6

    4.7

    1.7

    7*

    3

    3.5

    3

    7

    6.6

    0.4

    4

    3.7

    3.3

    7

    4.5

    2.1

    8

    8

    7.6

    1

    8

    7.6

    0.4

    8

    7.6

    0.4

    8

    7.6

    0.4

     

    Table 4

    Localisation attributed by participants to the eight segments

    In relation with the eight possible locations (bold characters), the numbers in the columns indicate which segments have been localised in that place, and the number of participants (in parentheses) who have attributed this location to this segment.

    Segm

    1

    2

    3

    4

    5

    6

    7

    8

     

    1(23)

    1(2)

    1(5)

    1(1)

    1(3)

    1(3)

    1(4)

    -

     

    2(4)

    2(13)

    2(3)

    2(8)

    2(4)

    2(6)

    2(3)

    -

     

    3(2)

    3(4)

    3(15)

    3(2)

    3(5)

    3(4)

    3(9)

    -

     

    -

    4(3)

    4(5)

    4(16)

    4(4)

    4(2)

    4(7)

    4(4)

     

    5(8)*

    5(11)*

    5(9)*

    5(8)*

    5(18)*

    *

    *

    -

     

    6(4)

    6(7)

    6(4)

    6(5)

    6(6)

    6(15)

    -

    -

     

    *

    *

    *

    *

    *

    7(10)*

    7(18)*

    -

     

    -

    8(1)

    -

    8(1)

    8(1)

    8(1)

    -

    8(37)

    * Segments 5 and 7 are identical. In order to calculate the distribution parameters, n 7 is considered as n 5 when is located in positions 1-5, and n 5 as n 7 when is located in positions 6-8.

     

    c) General remarks

    The absolute values of the difference between the right key location and the location chosen by each child were calculated and analysed with a 2(training) x 2(familiarisation) x 2(composer) x 8(segment's position) ANOVA. Results revealed a main effect of the segment's position (F(1,37)=6.258, p=0.0001), corroborating the primacy and recency effect already mentioned above, and of composer (F(1,37)=22.670, p=0.0001) indicating better results for the Diabelli piece. The difference between musicians and non-musicians was not significant (F(1,37)=1.696, p=0.2009). Perhaps musician children's training, in terms of years and hours of tuition, was not yet sufficient to reach significatively different results. Familiarisation was not significant either (F(1,37)=1.540, p=0.2224). However, the performance of the M3 group was better than the non-musicians', as shown especially in tables 1 and 3, where it can be observed that for both pieces this group of participants realised a completely correct mode regarding the rebuilding of the piece.

     

  7. General Discussion

The aim of the present study was to investigate the role of early music practice in the schematisation process by children, during real-time listening. Musician children who participated in our experiments were at the first steps of their music practice and presented a mean time of 2.5 years of music training. Data collected from the analysis of the results showed that their performances differenciated slightly from the non-musician children's ones. In similar studies with adults, musicians and non-musicians, a greater differenciation in the performance of the two groups was observed (Deliège 1989; 1998). However, it is worth noting that data analysis reported the same results' pattern, as the one observed with adults in all previous experiments for both tasks of segmentation and reconstruction of a piece (Deliège 1989; 1998). Music practice, even from its early stages, seems to influence processes involved during real-time listening. Additionally, the factor familiarisation had a more prominent role in musician children's memory, showing better skills to grasp the musical features.

Analytically, data from the segmentation task showed no effect of the music training factor in the process of grouping formation, i.e. the segmentation, during listening. Similar groupings were observed in both groups of musician and non-musician children and they were in accordance with the main articulations of the piece. However, differences in the coherence of the performance between the first and the second segmentation by musicians and non-musicians provide evidence that processes during a re-representation of a musical piece might be influenced by musical experience. Musician children, even in the first stages of their music practice, are more stable in their groupings. This provide evidence that, although segmentation process is suggested as a rather automatic psychological behaviour (Deliège 1998), music practice has an effect in stability of its results in knowledge's representation.

Performance in the reconstruction task was slightly better for musicians. In addition, they seemed more sensitive to the deep structure of the piece than non-musicians. An effect also of familiarisation was founded mainly in the piece by Diabelli. Children in the NM3 and M3 groups performed better than children from the NM1 and M1 groups. In Schubert, the comparison between the results of the two familiarisation groups showed a slightly better performance for the groups which received three previous listenings. Children's ability to remember musical schemata occurring in listening seems highly related with familiarisation, music practice and the particularities of the attended piece. As in general memory investigation, an effect of primacy and recency also appears in memory for musical structures. Children presented better results in the reconstruction task for the begining and the end of the attended pieces.

Participants' performance differenciated also in accordance with the musical piece that had been listen to. Results from Diabelli were clearly different from those collected for the Schubert piece. This differenciation might be highlighting the fact that even in musical pieces of the same style and period, particularities in their surface involve a different impact in listeners' processing. Cue abstraction, schematisation process and memory of the resulted schema of a musical piece is influenced by its characteristics as, for instance, the flow of the temporal rate.

In general, processes exhibited by both categories of children listeners, musicians and non-musicians, in the formation of musical schemata are not different. They are simply used more efficiently (and perhaps result in more explicit representations, from a cognitive point of view) by children with music training than by children without music training. Similar observations have been reported in previous studies with adults (see Deliège & Mélen, 1997). These remarks, underlined also by the fact that experiments with infants (Mélen, 1999; Melén & Wachsman, this symposium) have shown evidence of the presence of the cue abstraction mechanism already in the initial state of human cognition, suggest that this mechanism might be a predisposition of the human mind/brain, which during development and experience is modularized, in the sense of Karmiloff-Smith's theory (1992).

Karmiloff-Smith developed the hypothesis that the human mind/brain starts out with cognitive predispositions that already exist in its early age. As development proceeds, these predispositions are moduled from external influences and result to specific brain circuits that are activated in responce to domain-specific inputs and, in certain cases, to relatively encapsulated modules' formation. This modularization process of the human mind/brain sustain the structure of its behaviour and is responsible for its particular way of acquiring knowledge. During modularization, on the one hand, representations of the information already stored (both innate and acquired) are continuously altered via a process of redescription or, more precisely, of an iteratively re-representation of knowledge in different representational formats, and on the other hand, implicit information from these procedural representations is rendered via a process of "explicitation" in a more explicitly one.

Thus, in the light of Karmiloff Smith's theory, the cue abstraction mechanism might be an innate predisposition of the human mind which during development is influenced and modulated by environmental constraints such as experience, training or culture. However, additional researches are needed to validate this hypothesis and to provide more evidence for the role of the cue abstraction mechanism in the schematisation process of children of other ages and different levels' of musical experience.

Bibliography

Davidson, D. (1996). The role of schemata in children's memory. Advances in Child Development and Behavior, 26, 35-58.

Deliège, I. (1987). Grouping conditions in listening to music: An approach to Lerdahl & Jackendoff's grouping preference rules. Music Perception, 4(4), 325-360.

Deliège, I. (1989). A perceptual approach to contemporary musical forms. In S. McAdams et I Deliège (Eds). Music and Cognitive Sciences. Contemporary Music Review, 4, 213-230. Trad. Frs : Approche perceptive de formes musicales contemporaines. In S. McAdams et I Deliège (Eds) La Musique et les Sciences Cognitives. Bruxelles: Pierre Mardaga, pp. 305-326.

Deliège, I. (1991). L'organisation psychologique de l'écoute de la musique. Des marques de sédimentation - indices et empreinte - dans la représentation mentale de l'uvre. Thèse de Doctorat. Université de Liège: Non publié.

Deliège, I. (1997). Similarity in processes of categorisation : Imprint formation as a prototype effect in music listening. A preliminary experiment. In M. Ramscar, U. Hahn, E. Cambouropoulos & H. Pain (eds), Proceedings of SimCat 1997: An Interdisciplinary Workshop on Similarity and Categorisation. November 28-30, Edinburgh University, Edinburgh, pp. 59-65.

Deliège, I. (1998). Wagner "Alter Weise": Une approche perceptive. Musicae Scientiae, Special Issue, 63-90.

Deliège, I. (this symposium, a). Prototype effect in music listening. An empirical approach on the notion of imprint.

Deliège, I. (this symposium, b). Perception of Similarity and Related Theories.

Deliège, I., Delges, P., Oter, J-C. & Sullon J-M. (1998). Annexe: Le ScaleGame, un outil MIDI multi-fonctionnel. Musicae Scientiae, Special Issue, 117-121.

Deliège, I. & Dupont, M. (1994). Extraction d'indices et catégorisation dans l'écoute de la musique chez l'enfant. In I. Deliège (Ed). Proceedings of the Third International Conference on Music Perception and Cognition/Actes de la Troisième Conférence Internationale sur les Sciences Cognitives de la Musique. Brussels: ESCOM Publications, pp. 287-288.

Deliège, I. & Mélen, M. (1997). Cue abstraction on the representation of musical form. In I. Deliège & J. Sloboda (eds). Perception and Cognition of Music. East Sussex: Psychology Press, pp. 387-412.

Deliège, I., Mélen, M. & Bertrand D. (1997). Development of Grouping Process in Listening to Music : An Integrative View. Polish Quarterly of Developmental Psychology, 3(1), 21-42.

Dowling, J. (1999). The development of music perception and cognition. In D. Deutsch The Psychology of Music, San Diego: Academic Press, 2nd ed, pp. 603-625.

Karmiloff-Smith, A. (1992). Beyond Modularity: A developmental perspective on cognitive sciences, Cambridge MA: MIT Press.

Mélen, M. (1999). Les principes organisateurs du groupement rythmique chez le nourrisson. Musicae Scientiae, III(2), 161-191.

Mélen, M. & Deliège, I. (1995). Extraction of Cues or Underlying Harmonic Structure : Which Guides Recognition of Familiar Melodies? European Journal of Cognitive Psychology, 7(1), 81-106.

Mélen, M. & Wachsman, J. (this symposium). Categorisation of musical structures in 6- to 10- month-old infants.

Neisser, U. (1976). Cognition and Reality. New York: WH Freeman & Co.

 

 

 Back to index