This paper analyzes children's spontaneous graphical representations while listening to and making sense of musical fragments. Starting from the theoretical and empirical state-of-the-art, we set up a study with a view to investigate the impact of the following four factors on the level of sophistication of children’s graphical representations: (i) the children’s age, (ii) their musical background, (iii) the complexity of the musical fragment to be graphically represented and (iv) the specific musical parameter being prominently present in these musical fragments. Hundred-and-sixteen children divided into four more or less equal groups (8-9 yrs olds and 11-12 yrs olds with and without extra musical education) were exposed, in the context of a whole-class test, to six musical fragments that differed in terms of complexity (three simple and self-composed fragments and three complex ones taken from existing musical compositions) and of musical parameter being prominently present in the fragment (melody, rhythm or dynamics). Children’s representations were categorized by means of a classification scheme (that was developed in a previous pilot study) that distinguishes between (i) global notations, where the fragments are represented in a holistic way by one overall pictorial image (like a musical instrument or a real-world scene) and (ii) differentiated notations, which try to capture one or more musical parameters in their temporal unfolding (like an informal or formal pictorial notation of the development of the melody). As expected, we found a significant impact of age and music education: older children and children following extracurricular music education used more differentiated notations than younger children and children without such extra music education. Moreover, complex musical fragments elicited less differentiated notations than simple ones. We also found several interaction effects between the subject and task variables, – expected as well as unexpected ones. Theoretical, methodological and educational implications will be discussed.

Background and questions

The study aimed at analyzing children's spontaneous graphical representations of musical fragments. The background was provided by both theoretical (mainly neo-Piagetian) models and empirical studies of musical development, describing how subjects of different ages and with different musical experience apprehend and represent music, with a gradual shift from an intuitive and holistic approach towards a more formal and analytical notational approach as they grow older and/or get more musically experienced. Complementary theoretical input came from recent work on symbolizing, modeling and tool use in other domains like mathematics and science education.

Starting from this background, two pilot studies were set up with elementary school children: one wherein different groups of children were confronted with simple self-composed musical fragments and another one involving more complex existing fragments. These pilot studies, first, confirmed the rich variety in children's spontaneous ways of perceiving and representing music. Second, they resulted in a categorization scheme for graphical notations consisting of ten categories divided in two groups: (i) global notations, where the fragments are represented in a holistic way by one overall pictorial image (e.g. musical instrument, real-world scene, musical icon) and (ii) differentiated notations, which try to capture one or more musical parameters in their temporal unfolding (e.g., analogue image, non-formal notation, formal notation). Third, using this categorization scheme, we found some tentative evidence for the impact of children’s age and formal music education as well as of some musical features (i.e., the complexity of the fragment and the dominant musical parameter in the fragment) on the nature of their representations. The goal of the present study was to investigate the impact of these different subject and task variables in a more systematic and rigorous way.

Design

We set up a new study to investigate the impact of the following factors on children’s graphical representations: (i) children’s age and (ii) extracurricular musical experience, (iii) complexity of the musical fragments and (iv) musical parameter prominently present in a fragment.

Hundred-and-sixteen children divided into four groups (8-9 and 11-12-yrs-olds with and without extracurricular musical education) were exposed, in the context of a whole-class test, to six fragments differing in terms of complexity (three simple self-made fragments and three complex fragments from existing compositions) and of musical parameter being prominently present in the fragment (melody, rhythm or dynamics).

Children were asked to make drawings so that someone else could ‘reconstruct’ afterwards each fragment by looking at these drawings. Children were allowed to represent graphically whatever they wanted, except verbal annotations. Each musical fragment was played back four times. After having represented all fragments, children were invited to write down a short explanation of each notation at the bottom of each representation. The simple and complex fragments were, respectively, composed or selected with a view to accentuate one of the three musical parameters.

Children’s representations were categorized by means of a slightly adapted version of the above-mentioned classification scheme. The data were subjected to a quantitative and a qualitative analysis. The quantitative analysis focused on the impact of the four above-mentioned factors on the global versus differentiated nature of the representation. Given both the repeated and dichotomized nature of the measures, we used a Generalized Estimating Equations (GEE) approach to conduct a logistic regression analysis.

Results

First, we observed, as expected, an effect of age: 8-9-year-olds generated 29.5% differentiated notations, whereas for the 11-12-year-olds it was 43.9% (p<.01).

Second, the expected impact of musical education was found too: 26.1% differentiated notations for children without musical training versus 45.6% for those with such training (p<.01). There was also an interaction (p<.01) between this variable and the variable ‘musical parameter’, revealing that this main effect for musical background was especially due to the fragments about ‘melody’..

Third, we found the anticipated impact of the task complexity: simple fragments elicited 58.6% differentiated notations versus 14.1% for complex fragments (p<.01). The impact of this task variable was even greater than we had expected, since there was no interaction between task complexity and children’s age or musical background, indicating that also older and musically more experienced children had great difficulty with generating differentiated representations for the complex fragments.

Fourth, with respect to musical parameter, we found that children generated fewer differentiated notations for fragments about melody (27.1%) than for the other two parameters (40.1% for rhythm and 41.8% for dynamics) (p<.01), with no significant difference between the latter two.

Besides this quantitative analysis of the impact of the four factors on the global versus differentiated nature of children’s graphical notations, we also performed a detailed quantitative analysis at the level of the different subcategories of notations, which revealed, for instance, a remarkably low number of conventional formal musical representations, even among the older children with several years of musical training, as well as a qualitative analysis, which revealed a rich variety of graphical notations (some of which were of remarkable originality, esthetic beauty and/or representational sophistication).

Conclusion

The present study further documents the rich variety in children's spontaneous ways of perceiving and representing music. Compared to previous research, the most striking finding from our study is the demonstration of the very strong impact of task complexity (as operationalised in our study) on children’s representations. Many children who were able to generate differentiated representations for simple fragments reverted to less sophisticated global ones when confronted with more complex fragments. It remains unclear whether this difference is caused by the multidimensionality of the fragment (i.e., the fact children had to perceive and graphically represent several musical parameters simultaneously) or to the fact that these complex fragments were taken from ‘real’ compositions and therefore elicited different listening and/or representation strategies (e.g. more dominated by the emotions, fantasies, memories that these fragments evoked in the children). Once we understand better the representational features of these representations, we will better be able to explore their potential role in pedagogical attempts to foster the development of musical perception and representational competencies in children.