GENERIC REPRESENTATIONS IN CHILDREN’S DRAWINGS

Abstract

1. Generic representations in children's drawings It was suggested already by Luquet (1927) that drawing development proceeds over four phases. First scribblings or involuntary designs are followed by fortuitous realism, after that intellectual realism develops, and finally drawings are characterized by visual realism. In other words, there is a phase in children's drawing development where they they rather than what they see. Later studies have demonstrated that even though relatively young children can draw they see (they may include view-specific information in their drawings) there is still a shift in emphasis from one kind of depiction to another (e.g. Cox 1992, 1993). It is also noteworthy that drawings of young children often do not refer to any particular object but to a generic type (Gardner 1980, Luquet 1927, Milbrath 1998). In order to differentiate between drawings that represent a generalization and drawings that represent particular models, it is necessary to know exactly model a child drew. In case of free drawings it is not very often clear which model the child intended to draw. In case of model drawings, however, it is also not easy to differentiate generalized drawings from exemplar drawings. Some models may not have distinct characteristics. Even if a model has distinct characteristics it would always be possible that a child's attentional capacities are limited, and distinctive characteristics were not drawn because a child drew a generalization, but rather because she or he forgot to draw those particular attributes. This obstacle was taken into account in a study by Toomela (2003b). He presented children a model of an unusual doll, which had fewer attributes than usual dolls. Drawings of a doll were coded as follows: scribbles or patterns; prototypical drawings of a doll where more attributes are drawn than characterize the model; 2D-exemplar drawings where only attributes of the model are depicted, but the drawing is not in perspective; and 3-D-exemplar drawings of the model drawn in perspective. It was found in this study that children before the age of three years and ten months drew only scribbles or prototypical drawings of a doll. Before the age of eight years, the most typical drawings were also prototypical. In this study a cube was also presented to children as a model. Cube drawings were coded as suggested by Toomela (1999) into four categories: scribbles and patterns, single squares, figures differentiated into faces with visually not realistic relationships between faces, and integrated wholes drawn in parallel or convergent perspective. It turned out that the two coding systems, that of a doll and that of a cube, were significantly related even when the effect of age was partialled out. Thus, prototypical drawings of the doll were accompanied by earlier stages of drawings of a cube. This is indirect evidence that earlier stages of cube drawings may actually be drawings of a cube generalization rather than the particular model. In the studies of drawing development, cube has been one of the most popular models. Therefore, it would be interesting to establish whether drawings of a cube may also be generalizations in earlier stages of drawing development. Drawings of a usual cube cannot be informative in that respect because there are no distinct attributes of the model that could be omitted or added. In order to determine the child's intention in making an outline drawing of a cube, Moore (1986) used a colouring task, in which each face of a cube presented to children was painted a different colour. She discovered that 7-year-old children included more hidden faces of the object in their drawings than 9-year-old children, but the colours used were always correct colours. Younger children drew in addition to viewpoint specific faces also faces of the model hidden from their viewpoint. In addition, it was also found in this study that outline drawings comprising a single square represented in some cases, especially in younger children, the whole object rather than a particular face of the model. …