21
Home Introduction Keynote Speakers Symposium Overview Papers
Drawing on the imagination in young children
N. H. Freeman
Abstract
There has been a tradition of regarding the period when children are aged something like six or seven years as the time when great discoveries start to be made in drawing. It is true that a flowering can occur then. We review recent experiments on asking children to draw fictional characters, experiments that demonstrate how much support five-year-olds need before they become relatively self-reliant in using their imaginations to generate brand new drawings. But that can only be part of the story. Another part of the story concerns how children get themselves organised in the period before schooling, when drawing emerges largely untaught from scribbling. We review recent simple techniques that show how pre-school children monitor what they are doing with the pencil and categorise shapes on the page in the course of acquiring what deserves to be termed a 'theory of pictures'. Once we understand that theory, then we can understand what it is that children are grappling with in the period when they are on the threshold of discoveries that generate systematically inventive drawing, and why engagement with understanding art is a lifelong enterprise.
Children differ amongst themselves in their understanding of art. Part of that variation might be due to the fact that tastes differ. No one would deny the fact. But current conceptions of taste are pretty limited. The question is, what do children conceive of the pictures about which they express differing tastes. And we must add to that the question of how children conceive of their role as picture-producers. The heart of the matter is that children grow into an intuitive theory of art as they develop (see Freeman, 1995). Presumably, encounters with challenging artworks advance the development; but whether an individual's development be fast or slow, the first task is to formulate what it is that develops. Any developing theory of art will come to encompass relations between at least four different elements: (a) the picture plane itself with markings on it, (b) what, if anything, the markings represent, (c) how the organisation of markings was arrived at, usually through the agency of the artist, and (d) the role of the viewer, whether that role be conceived of as that of a passive observer or of an active interpreter. Our task is to take a look at how we can specify changes in the organisation of the relations between the four things, in the light of the opinions that children express about their own drawings and other people's drawings.
The phrase 'theory of art' refers to the theory that young people acquire in the course of their daily experiences and their encounters with artworks. It is right to call it a 'theory'. Although the term sounds pompous, the consensus is that the term 'theory' should not just be reserved for professional researchers and intellectuals, as though they were the only ones to acquire organised structures of beliefs and knowledge (see Wellman, 1990). In the course of everyday life, people acquire theories of various domains, such as the worlds of biology, physics and everyday mathematics. The domain of the arts is no exception. Conceptions of the visual arts involve more than developing matters of taste; people work out deep-seated beliefs about their relation to art, and they vary in the assumptions they are readiest to make. People acquire a set of concepts, beliefs and desires about visual art, a conceptual framework that they use to explain to themselves what art can do for them. People do not keep to sharp divisions in their everyday thinking. For example, Belver (1989) found very high correlations between people's judgements of how beautiful they thought particular pictures were and how interesting they found them. There is no sharp frontier between the two types of judgement. People vary in how developed or coherent their theory may be. A huge variation in conceptual organisation is to be found in people of different ages. One looks to developmental psychology, with its wide range of difference across huge differences in age, to expose normal differences in thinking patterns.
It is not so long ago that there was a phase in education where the arts were considered to centre on matters of feeling and intuition that foster creativity and self-expression. So the arts can. But the arts were not thought of as involving the sort of intelligence fostered by other school subjects. That cannot possibly be true. Coming to terms with the meanings of artworks, as both producer and viewer, involves children in making deep intellectual discoveries. It is easy for commentators to underestimate very young children's discoveries. In one sense the discoveries are simple, but a better term is 'fundamental', because those early discoveries serve as the bedrock of an elaborate, developing theory of art. Let us start with discoveries in production, noting along the way how they attest to conceptual development as much as to mastery of the trade.
Children in nursery school experiment eagerly with drawing materials. But the children are often more interested in the process of making marks than in what the marks look like when they have finished. They enjoy the activity but will also end it abruptly in favour of other activities. They characteristically pay rather fleeting attention to the finished result of their activity. What would impel their attention to the finished product? It seems that some conception of a fulfilled intention is a crucial development. Let us consider a bit of evidence from children aged two and three years.
The simplest quasi-representational drawing by two-year-olds uses the picture-plane in pretend play. Thus, children may use a marker as a pretend rabbit, say, and make it hop across the paper, leaving traces that are a record of the hopping, like footprints in the sand (see Freeman, 1993; Gardner and Wolf, 1987). Three things come together in such a case: (a) the activity involved in production, (b) the referent that lends meaning, and which is (c) used as the instrument that makes marks. Along with such activity comes scribbling. That may be done for the fun of it, or with representational intent, or both. Even when an independent observer cannot tell what the scribbles represent, children may claim representation if asked what the drawing is of. But often enough no answer is forthcoming. It has been difficult for observers who want to make statements about the roots of representation to formulate what is going on in the scribbling phase. But in 1997, a paper by Adi-Japha, Levin and Solomon appeared that did give an incisive indication of a nascent representational process at work. It transpires that experimenters had been asking children the right question but under the wrong conditions. At first sight, the paper we are going to consider seems to be a trivial advance in technique. But the advance enabled the experimenters to introduce a battery of new conditions that cumulatively led to an incisive suggestion. The technique was to ask two-year-old scribblers about parts of their scribble rather than about the finished product as a whole. The experimenters were careful to point to examples of two types of mark in the scribbles when they asked children what those parts were. One type of mark was a smooth sweep, defined in terms of the geometry of production as recorded from the electronic bitpad on which the children drew. The results here were the usual: sometimes some children would say that the mark was something, a cat or a hat or whatever, and sometimes no answer was made. But when the type of mark was inflected with a sort of V-shape somewhere in it, or was a broken line, the number of representational claims made by the children significantly increased. The suggestion is that the act of making an inflection in a line involves some investment of attention, as can be detected from the kinematics of the children's production on a bitpad (their speed of pen-sweep, for instance). When a line discontinuously changes direction, an effort of close control is needed in wielding the pen. The suggestion is that this effort of the producer primed a willingness to make an act of representational interpretation.
Is that suggestion warranted? On its own, the inflection-advantage could mean almost anything. The significance of the finding lies in another condition which the experimenter ran. The inflection-advantage disappeared if the experimenter herself produced a tracing of the child's scribble and questioned the child about that. So it was not the case that the inflection-advantage was due entirely to differences between smooth and jagged marks. That is fully congruent with the suggestion that the crucial factor is the child's registration of her own effort as a producer. There is no suggestion that the representations claimed by the children were anything other than post-hoc. But that is not a weakness in the finding. Indeed, there is a sense in which the finding is stronger if it is the case that the children were not recalling a representational intention but were seeing a significance in the mark because of a memory of an effort of production that was fulfilled in a change in direction on the picture plane. It was suggested above that some conception of a fulfilled intention is important in coming to value the product. Two-year-olds' representational conception may be entirely or mostly rationalisation after the act; but that is consonant with a long tradition of arguing that young children learn by doing. If the child is aware of having concentrated on controlling the direction of the pen, she takes it that effort merits significance. Or maybe that one effort should be followed up by another effort, this time in studying the mark to see what its significance might be when the adult shows that he is interested in that mark. The mentality of two-year-olds is a bit obscure: it is a real achievement when experimenters find out anything much.
It should be noted that the inflection-advantage also disappeared if there was a long delay (long for a two-year-old, that is) before the questioning was done. That is further evidence that the fresh memory of production was important, not just the shape of the mark. It is not until some three years of age that evidence emerges for a conception of the fulfilment of a prior intention. For such evidence, we can look to a paper by Bloom and Markson (1998), where again the suggestion is that doing is important, not just the shape of the mark on the paper.
Let us set the work in the context of a traditional problem in philosophical aesthetics. Is an artist's involvement in production a necessary condition for something to be a representational picture? It is, in principle, possible for an ant to leave a trace on the sand that looks very like Winston Churchill (see Putnam, 1981). The pattern on the sand is a natural kind, produced by a force of nature. Here nature supersedes the role of the artist. Someone might be willing to claim the role of artist and thereby turn the sand pattern into a picture. One could take a cast of the sand pattern and put it on display. Such is the logic behind the gallery display of objets trouvés. The resulting products rely for their effect on the ambiguity between attributing responsibility to the natural producer and to the keen eye of the artist. It is no accident that Damien Hirst's shark, framed in a tank, and Carl André's unframed bricks, challenge the hard-won reasoning of a vast section of the viewing public. The challenge is to believe that there is a representation to be discussed. How do three-year-olds respond to a child-centred analogue of the problem? Gelman and Ebeling (1997) showed three-year-olds a variety of crude pictures, each of which could be seen as representational of an object (a bear, for instance). A set of the pictures was given to a group of children with the remark that the pictures were intentionally produced, so those pictures represented the efforts of an artist (even if the pictures were not very good). The same set of pictures was shown to a matched group of children with the introduction that the pictures had been produced by accident, like paint being spilled onto the page. The data came from asking children to describe each picture after it had been introduced as being intentional or as accidental. The crucial data are the extent to which children mentioned the name of the object that could be seen as represented in the picture. That tendency was much stronger in the intentional condition. Naturally, this finding is open to different interpretations. Perhaps the experimenter's remarks about prior intention alerted children that something or other was being attempted, so they had the mental set of looking for something to say about a picture, and the representational content grabbed their attention. Yet if the representational content was that highly attention-grabbing, one would expect children in the accidental condition to comment on it. There is a way of starting to investigate how the effect works. It would be to vary the recognisability of what could be seen in pictures, so one could see if (a) the prior-intention effect applied across the board to pictures of different levels of recognisability, just occurring more frequently in one group than the other, or (b) scaled out so that one group lowered their criterion for depiction and accepted cruder pictures as representational. That is, one can start to model what is going on in the child's mind by varying intention against apparent successfulness of the product as it looks to a viewer's eyes.
One reason for doing that bears on another fundamental problem in philosophical aesthetics. That problem concerns what we have been cavalier about in the above exposition: the way that marks themselves have effects regardless of the artist's intention. Pictorial realism can be specified as 'the quality of a picture that allows us quickly and easily to recognise what it is a picture of' (Sartwell, 1994, p. 5). Realism lies 'not in quantity of information but in how easily it issues' (Goodman, 1976, p. 36). It is theoretically most interesting that characters in caricatures, which are clearly fictional transformations of referent appearances, can often be more quickly identified than photographs, so caricatures would be categorised by philosophers as more realistic than 'naturalistic likenesses'. Rhodes and McLean (1990) showed that untrained adult viewers recognised caricatured birds as accurately as they recognised undistorted images. Bird-experts found caricatures even more recognisable than undistorted images. That is a neat finding because the bird-watchers had not been specially trained in caricature interpretation: their expertise in discriminating the referents gave positive transfer to picture interpretation. Caricatures are obviously examples of artists investing intentional effort.
Let us put together the points that (a) three-year-olds seem to approach picture-description in a representational way if a prior intention has been announced, with (b) whether children take a realist stance on pictures. Bloom and Markson (1998) asked 24 three-year-olds and 24 four-year-olds to make a pair of drawings of either a balloon and a lollipop, or of the experimenter and a self-portrait. Such pairs of drawings usually turn out pretty indistinguishable when you are that young (and sometimes the drawings are not that recognisable to an independent viewer, either). The children then did something else for 7 minutes, and were then asked to name their drawings. There was significantly greater than chance appropriate naming: the three-year-olds named drawings in accord with prior intention 76% of the time, and that rose to 87% in the four-year-olds. Clearly, when shape does not distinguish the pairs of drawings, so that each is equally recognisable as representing each referent, children take prior representational intent as the discriminating variable. Intent is taken as a criterion when realism is equated. The authors produced a formulation that rather reverses the usual order of exposition of why shape is so important in naming pictures as representational: 'Children might call a picture that looks like a bird "a bird" not merely because it looks like a bird, but because its appearance makes it likely that it was created with the intent to represent a bird. In general, appearance - and shape in particular - is seen as an excellent cue to intention' (Bloom and Markson, 1998, p. 203). Perhaps that bold formulation overstates the case? Again, one would want to see if the intention effect scaled out against realism. An earlier attempt had been made by Bloom and Markson (1997), but they made realism too strong. Thus, they had an artist looking at a fork and carefully drawing, but the finished drawing looked like the spoon that lay to one side. When asked to name the drawing, almost everyone at any age called it a spoon rather than the fork at which the artist had been looking whilst she drew. Under a condition where the drawing was made in such a way that it was about equally plausible as a contender for being a fork or a spoon, almost everyone, even four-year-olds, called it a fork in accord with their construal of the artist's intention. But one should be wary, as Bloom and Markson (1998) pointed out, because it is a sort of trick situation really. However, it does seem reasonable to suppose that children see it as unlikely that an artist would make a drawing that very clearly triggered a recognition of a referent without intending to have that effect on a viewer.