Moving into Literacy: How it All Begins

Lesley Lancaster. Handbook of Early Childhood Literacy. Editor: Nigel Hall, Joanne Larson, Jackie Marsh. Sage Publication. 2003.

There is very little research into the development of literacy much before the age of three. There are a number of studies of its development between the ages of three and five or six (Torrey, 1973; Clark, 1976; Bissex, 1980; Luria, 1983; Ferreiro and Teberosky, 1982; Kress, 1997; 2000), and although there are references to earlier stages of literacy in some of these, there is no substantive study of literacy below this age. This is perhaps unsurprising since very young children are not likely to be literate in the commonly understood sense of the term, and therefore, so the argument might go, there is not a great deal of evidence to consider. However, this does leave a significant gap in the account of children’s development of literacy. It would be surprising if, from an apparent position of no serious engagement with print at all, children are suddenly able to explode into the type of rampant, creative, independent and reasoned relationship with the medium which is described in most of the studies cited above. It appears as if much groundwork has been laid before. However, there are both theoretical and practical reasons why this has not invited a great deal of attention in educational circles.

The influence of Jean Piaget on educational thinking about children’s cognitive development has been considerable over the years. Critiques of aspects of his work have been accepted for some time now (Donaldson, 1978; Deloache and Brown, 1987; Feldman, 1987; Thelen and Smith, 1994), but the legacy has persisted. Two claims of Piaget can be held to partial account for the lack of attention to early literate activity. First, there is the suggestion that learning happens in fairly discrete stages, with the very earliest involving little more than reflexive responses to the external environment. In other words, children have to reach the right stage before they are able to cope with the physical, cognitive and symbolic demands of reading and writing. Secondly, there is the view that a common process of development pervades all areas of cognition, rather than there being different domains of cognition and learning. So, very young children would not be capable of recognizing features of literacy as a unique symbolic system, distinguishable from other such systems.

A second factor influencing the investigation of early literacy comes from the study of the development of writing itself, and in particular its relationship to speech. A traditional view holds that writing is predominately a transcription of speech (Havelock, 1976; Gelb, 1952); that it developed in order to enable talk to be written down. It would therefore be quite reasonable to suggest that until children have reached a fair level of proficiency with spoken language, they are not going to have the requisite skills to understand or produce written language; they cannot read or write down talk if they are not yet producing it. The marks that children do make prior to being able to make this connection are frequently described as ‘scribble,’ with its rather unhelpful association with marks that are purposeless, illegible or meaningless. Within the literature, it tends to refer specifically to a stage of development prior to there being any real connection being made between letter forms and spoken meaning; Sulzby refers to the way in which children of five or six who are aware of ‘conventional spelling’ will revert to the use of ‘lower-order forms like scribbling’ (1986: 70) in certain situations. The suggestion here is that where the connection between spoken and written modes is not made, these features of children’s ‘written’ productions have ‘lower-order’ communicative intention and representational purpose.

The third reason why this early stage of literacy has not been researched to the same extent as literacy beyond infancy is a practical one. Once children reach the age of three, many attend nursery classes and playgroups. The population of children under two, however, is very widely located, with children mostly being looked after at home or by childminders. Access, in other words, is difficult. It is also problematic to work with children this young without the cooperation and help of someone the child knows and trusts. Whilst this can have many benefits, it does add practical and organizational difficulties to research. The other side of this is that the relative ease of working with children in institutional settings also contributes to what Street (1995) calls the ‘pedagogization’ of literacy: the reduction of reading and writing to social practices predominantly associated with schooled learning. This contributes to the exclusion of the literacy practices of infants and children who are not yet part of the formal educational system from serious consideration.

In spite of the sparsity of research studies concerned exclusively with children’s development of literacy before the age of three, studies from disciplines like art education show that these children are capable of a range of representational behaviour. Far from this being a quiescent period, children of this age avidly explore graphic systems, much as do children later on. In this chapter, I shall look at evidence that they are able to systematically explore graphic systems as a way of representing significant features of their personal, social and cultural experiences, long before any direct connection with language is made. I shall show that they are capable of recognizing the unique features of different domains of representation—writing, drawing and number—and I shall look at how even the very earliest marks made by children reflect intention and meaning. Finally, I shall discuss the view that they are already actively acquiring parallel systems of representation within these domains between the ages of 12 and 24 months. In other words they are already actively producing and interpreting different levels and genres of written language and different means of representing them.


In certain very obvious senses, it can be said that literacy begins at the beginning. At the start of the twenty-first century, most children are born into cultures which are in one way or another driven or affected by a complex array of literacy practices. The social and material evidence of their operation is pervasive, although it goes without saying that individual exposure to this is variable, both within and across cultures and communities. And long before they might be expected to understand much about it, babies and toddlers are on the receiving end of a kind of intense semiotic acculturation, which includes writing of many kinds in many forms. Clothes, bed covers, eating and drinking utensils, toys, videos, books: a host of merchandise covered in pictures, logos, numbers and print of all kinds, representing characters and objects from media events of the moment. These are part of a huge industry producing an array of items which are directed at the families of very young children and ultimately, of course, at young children themselves. Marsh in this volume demonstrates the extent of the modes and media involved.

Initially it is the adults who respond to the particular characters, narratives and images signified by these various objects. However, their cultural significance is mediated through talking, showing, shaking, touching, and the countless other ways in which adults communicate with babies about significant things. The baby’s first response is likely to be to their material qualities, and its exploration to involve senses like vision, touch and taste. However, the fact that, initially at least, a baby finds putting an object in its mouth an effective mode of exploration, does not preclude it from being responsive to its other signifying features. Nor does it preclude it from directing the focus of its attention to objects and images not in the least designed to be interesting or attractive to an infant. Very young children are notorious for finding unpredictable things salient. I recall my son, as a young baby, screaming when being moved from one part of the room to another so that he could watch the movement of what I considered to be a really interesting mobile. I eventually realized that in his original spot, he was watching the movement of the leaves of a tree, reflected through a sunny window as flickering shadows onto the ceiling. I had interrupted some serious semiotic work.

Even the best intentioned adults have a tendency to disrupt this kind of intense engagement which young children have with seemingly unimportant or irrelevant objects and activities often, of course, for their own protection and safety. My mother remembers an incident which happened when she was probably less than a year old. She was crawling along a beach when she spotted what she considered to be the most brightly coloured and desirable object she had ever seen. She crawled towards it with delight and anticipation, and was just about to grab it when a hand came down and snatched it away from her grasp. What the object had been, she realized much later, was a Swan Vestas matchbox, with its highly coloured images and lettering. Eighty years later, she can still recollect the feeling of devastating loss and disappointment!

These examples illustrate some important things about how very young children engage with visual and graphic representations of meaning. In both cases, the activity is independently constructed; it is purposeful and compelling, involving, potentially at least, a commitment of time and effort to find out more. In the case of the matchbox, the compulsion was also a physical and bodily one: to hold and scrutinize the desired object. Watching shifting patterns of shadowy images, on the other hand, requires prolonged visual concentration. It also requires considerable intellectual engagement with what amounts to a fluctuating sequence of abstract images. This kind of intense desire to find out more, the need to be physically involved, and the extended concentration and interest in things at an abstract level, are also the qualities which very young children require to start engaging with all the cultural paraphernalia and attendant semiotic systems which surround them; including, of course, systems of writing. Whether they are just generally aware and responsive to all of this, or whether they start to sort out how these systems work as soon as they are aware of them, is central to a consideration of how literacy develops.

Recognizing Different Domains

Harste et al. (1984) claim that by the time they reach the age of three, children are quite insistent about there being a distinction between drawing and writing. Evidence comes from the marks that children make and what they say about their function and purpose. Although there is not necessarily a consistency between the types of marks which are used to represent these different modes, nevertheless, whatever form is chosen is used consistently and systematically. So, children might use circular shapes for drawing and linear strokes for writing. Whatever and however marks are produced, they are always ‘a serious expression of meaning’ (Goodman, 1986: 7). A number of factors can be significant in how this choice of marks is made, including the dominant writing script in their communities (Harris and Hatano, 1999). Useful evidence comes from the productions of slightly older children who have access to more than one writing script from an early age. Gregory and Williams (2000) show the skill and proficiency with which such children move between two very different writing systems that use marks in quite different ways, Bengali and English. Kenner (2002; and Kenner and Gregory, in this volume), in her study of young children learning to write in two different writing systems at the same time, demonstrates that at five and six, children are able to clearly articulate distinctions between the semiotic principles operating for the different systems they are using. This includes being able to point out that there is a distinction between a logographic system like Chinese and a pictorial system. In other words, the decisions which children make at this stage about which system is which and how they are constructed are consistent, rational and well informed. As has already been suggested, it looks as if in order to have the capacity to reach such a sophisticated conceptual level by the time they are five or six, children will have been engaged in serious analyses of the systems of representation around them for some time. In contrast with Piagetian thinking about this, they will have recognized different domains of cognition and learning and made a start on working out their distinctive features long before they reach the age of three.

Deloache et al. (1979) show that already at five months old, children are able to recognize relatively abstract pictures of objects. Karmiloff-Smith (1992) notes that by the time they are 10 to 18 months, children readily differentiate between drawing and writing in their productions. They are ‘adamant,’ she says, about the distinction between a mark that is a drawing and one that is writing. However, the relationship between process and product has to be taken fully into account when considering the intentions of children this young. The mark on its own is likely to be ambiguous without the context of its production, since toddlers go about the processes of drawing and writing differently, even though the outcome might be largely undifferentiated. So, for example, they tend to lift the pen from the page much more frequently when pretending to write than when pretending to draw. Lancaster (2001) shows that before they are two, children are also able to use other modes of communication to help ascribe meaning to their notations: language and gesture, for example, can be used to ascribe identity to otherwise undifferentiated marks, along the lines of, ‘This mark I am pointing to is writing because I say it is writing, this one is a cat because I say it is a cat.’ The interpersonal interactions involved in this kind of symbolic activity are central to the process of production (Halliday, 1975; Trevarthen, 1990). Both communicative processes, in the form of the informal involvement of adults and other children, and processes of physical and material production are an integral part of the outcome. The final production cannot really be understood without taking into account the circumstances which gave rise to it. However, it is interesting to note that whilst this multimodality is central to very young children’s ability to communicate distinctions of meaning, it is not a quality that simply disappears as they develop fuller linguistic and graphic repertoires. Mackey (2003), in describing a project set up to explore readers processing texts in a variety of media, shows how children of 14 still manifest an understanding of the story they are reading in a variety of tactile ways. They too use gaze, a range of gestures, and their hands and fingers to process the print. In other words, the use of the body to mediate symbolic meanings (see Johnson, 1987) is not something which children move on from as they develop the ability to talk, draw, and read and write in more adult ways, but a process which has a developmental momentum of its own.

Karmiloff-Smith (1992) has also looked at slightly older children’s ability to discriminate between drawing, writing and number, as well as between drawing and writing. Working with children of four years, she found that they apply clearly differentiated constraints to the three modes of notation: drawings are not acceptable as either number or written language; single elements are acceptable as number, but not as writing, as are repeated identical elements; linkage between elements is accepted for writing, but not for number; and a limited number of elements is accepted in a written string, but the same constraints do not apply to numbers. Clearly this involves sophisticated analyses and a considerable degree of understanding of the salient features of the different modes of notation. It is unlikely that the ability to additionally recognize numerical properties is something which happens suddenly at this later stage, particularly in view of evidence which suggests that babies as young as four months old can distinguish between different number arrays (Treiber and Wilcox, 1984), and by six months they can detect numerical correspondence between auditory and visual inputs (Starkey et al., 1985). It also seems unlikely, given a very young child’s ability to deploy the most salient resources to deal with the task in hand (Kress, 1997), that such useful insights would be ignored. Whilst children under two might not be able to articulate the differences between writing, drawing and number in the same way that the four-year-old can, this does not mean that they are not aware of them; more likely the problem lies in the difficulty of gaining access to what it is they know and how they come to know it. What it does suggest is that children’s learning about domains of symbolic representation is a continuous, developing and expanding process, which starts very early on in their lives.

Expanding Ways of Meaning in Each Domain

If children are able to distinguish between writing, drawing and number before they are two, then it follows that this is based on their experience and perceptions of how each of these domains individually represent meaning. This is far from being a straightforward thing to sort out, however, since the domains also have features in common, at least in certain circumstances, and sometimes they do not behave ‘in mode’ at all. So writing can be represented graphically, two-dimensionally on a page, using a simple tool like a pencil, as can drawing and numbers; numbers can also be represented by writing; writing and numbers can be incorporated into drawings; but then writing can also be read aloud, numbers can be represented three-dimensionally by an abacus, quantity and spatial organization can be incorporated into written meanings, drawings can be made to move in cartoons, and so on.

If one starts to consider the sheer range of possibilities which each domain of symbolic representation presents to young children at the outset of the twenty-first century, then it is not difficult to reach the conclusion that to view them as principally pedagogical practices is to significantly misrepresent the nature of the task facing children, as well as the complex nature of symbolic systems involved. Indeed Teale (1986) concludes that for preschool children, 90% of their reading and writing experiences occur as part of the daily lives of their families and communities, rather than as being seen and acknowledged as literacy activities per se. As Goodman (1986) points out, children know that most things in their lives are organized systematically, and this insight is applied to the complexities of symbolic representation as they present in the course of daily routines and interactions. Pahl (1999) shows how they interpret things according to the information and resources to which they have access at the time, and according to what is currently salient to their thinking. She cites the example of a child in a nursery copying every feature of an adult’s rendition of his name, including mistakes and superfluous dots; the same child could also produce his own version quite independently. Pahl’s explanation is that to the child, these tasks are quite different: the first he constructed as a design task, whilst the second ‘required subjective decisions about what constituted his name’ (1999: 63). In other words, to see this task as a process with a single outcome is to entirely misunderstand the complexity of the child’s understanding of the ways in which meanings can be represented graphically.

Both writing and drawing suffer from being dominated by pedagogical concerns and by being frequently reduced to a single endpoint: in the case of writing, correctly making links between sounds and letters; and in the case of drawing, the achievement of realistic picturing. Wolf and Perry (1988) suggest that in fact, drawing involves a wide repertoire of visual languages. In investigating how young children develop drawing skills, they identify three broad areas of development: the invention of drawing systems; the construction of distinctions between different graphic genres (between maps, drawings and diagrams, for example); and the evolution of specific renditions of visual images. Here, their definition of ‘drawing’ also included features of the process of production. Looking at the creations of children between 12 and 15 months, they identified the appearance and use of a series of distinguishable systems.

Significant analogies can be drawn here with features of writing development. Wolf and Perry describe one of the earliest of these systems, evident at between 12 and 14 months, as ‘object-based representations,’ where drawing materials and tools are substituted for the referents which the child has in mind. So the pen might be rolled up in the paper to signify the object ‘hot dog,’ rather than being used to make marks which represent it graphically. Kress (1997) describes a range of creative constructions devised by children moving into literacy, which could be similarly characterized. Rowe (1994) also shows how immediate and transformative children can be at this stage: turning a paper plate into a note by pencilling across it, giving it to someone, and then throwing it into the air to demonstrate how it flies. Labbo (1996) shows how children use the computer screen to similarly explore symbols and objects. These children are slightly older, between three and five, but they are still using the same multimodal strategies as the younger children, though arguably at a more developed level. The system is also broadly the same whether it is being considered as drawing or writing. As Kress points out, the principles deployed in the learning of writing are much the same as for other sign systems, ‘employing the strategy of using the best, most apt available form for the expression of a particular meaning’ (1997: 17). This is central to all children’s early representations of symbolic meaning.

Another system which Wolf and Perry identify as appearing at about the same age is that of ‘gestural representations,’ where gestures are incorporated into the process of signifying meanings. They give the example of a child saying ‘bunny’ and hopping a marker across the page, making a trail of dotted footprints. Vygotsky (1978) suggests that gesture is also linked to the development of written signs in much the same way, and many of his examples are very similar in structure and concept to those cited by Wolf and Perry. Lancaster (1999) describes a child of 23 months using gesture to ‘read’ writing. Looking at a line of print on a title page, she moves her pointing hand rhythmically back and forth along the line, saying in unison with the movement, ‘says writing David’s, writing David’s.’ The gesture models the movement of her older brother’s hand as he writes. These signs, with their precise integration of language and gestural action, comprise the most useful and appropriate forms available to these children at these times to express these meanings. As to whether a distinction between gestural signs which are characterized as drawing and those characterized as reading and writing can be made, more research would need to be done. However, gestures of different kinds are an important means by which very young children signify meanings, both within and across both these domains.

The third system of drawing activity which Wolf and Perry identify, they describe as ‘point-plot representations.’ At around 20 months, children start to make ‘planful’ use of graphic properties. They are able to record the number and location of an object’s features by linking the graphic properties of the marks with the spatial features represented on the surface of the paper. So a human body might be represented by a line above another line and two lines at the bottom: top half, bottom half, and feet. This ‘planful’ linking of marks, properties of marks, and spatial organization is also well documented in the development of children’s writing between three and school age, and is discussed in detail in a number of the studies referred to at the start of this chapter. Marks are made and placed systematically to express ideas and concepts. Ferreiro and Teberosky (1982) describe how a child might make a very large mark because it signifies a big animal, or conversely a small one because it signifies a small animal; and a mark might be placed adjacent to another mark because it signifies something which is part of or owned by the object or person represented by that mark. Luria (1983) shows how children also ascribe meaning to otherwise very similar marks by how they are located on the page.

Thelen and Smith (1994) point out that the facility to organize and classify space symbolically is clearly evident by the time children are 18 months old. Lancaster (1999) shows the same kind of systematic organization of marks and space at 23 months. The child in this study is making a card for her mother, with the support and collaboration of her father. The page she is working on is organized so that the same mark can be used to represent different things according to how it is placed on the page: for example, one side of the page has ‘cat’ marks and the other ‘drawing’ marks; ‘writing’ marks are placed at the very bottom of the page, as close to the edge as possible. These constraints reveal that she already understands significant things about symbolic practice and representative domains. A graphic mark can be used to represent different systems (drawing and writing) and objects (cats); these have different relationships to the surrounding space, with drawing and object marks being less constrained and linear in their relationship than writing. Kress (1997), Matthews (1994) and Pahl (1999) have also observed the keen interest which very young children have in organizing space, including the ways in which they physically inhabit it, and the ways in which this experience informs their perception of symbolic space. Space is of itself a significant meaning making resource (see Kress and van Leeuwen, 1996), which young children use to very good effect.

In my example, principles of spatial organization are used to structure a simple narrative based around a cat involving a sequence of possessive ‘events’ where activity is implied by this ownership: cat’s tree; cat’s house; cat’s treehouse. For very young children, ownership tends to involve a physical relationship with an object or person: holding, touching or being very close to what is desired. Tomasello (1992) reports on how such social encounters and relationships are reflected in the ways in which children construct a linguistic system. He notes that by 17 months, a child is able to express possession by saying the one ‘next’ to the other, as in ‘mummy sock.’ So, both physical experience and language provide vital tools in dealing with the difficult problem of how to represent a narrative with a temporal organization involving a sequence of different events, on the fixed and two-dimensional space of the page, in a consistent and repeatable way. The principle of adjacency, of locating a mark representing someone close to the mark representing the thing that they own, is one approach adopted by the child in this study. This suggests the beginnings of an insight into two significant and distinct features of writing: generic structure and grammatical organization. This is derived from thoughtful and reasoned activity, and investigation of consistencies and regularities in these domains, long before it might considered that any explicit connections could be made between different graphic and linguistic forms and modes.

Making Marks

Consistent reference has been made throughout this chapter to the marks that children under two make, and the systematic ways in which they are generated and presented. Reference has also been made to the way in which they are frequently described as ‘scribble,’ with all the negative connotations associated with the term. None of the evidence discussed so far suggests that this characterization is likely to be an accurate one. The fact that children’s earliest marks are seen as falling short of intentionally representing alphabets, or images which closely resemble real objects, has often meant that they have been regarded as simply part of a stage to be passed through, on the way to the real thing. However, whilst there is limited research on how young children construe the material structure of written notation, what there is provides evidence of their systematic engagement with its basic elements.

Dissanayake (1992) says that children of this age require the physical trace of a marker in order to remain interested in what they are doing. If the marker fails to leave a mark, they lose interest in the activity. The related actions and movements are not in themselves enough: the visual sign is after all an essential characteristic of writing and of artistic practices, and children are surrounded by evidence of these in every possible medium. A significant distinction between these domains, however, is that in the case of print, its signifying status remains fixed in spite of local variation: letters of the alphabet, logograms and ideograms are identifiable as such whether they are handwritten, printed or typed, and in spite of variations in handwriting, print type, size or location (see Goodman, 1976). It is most unlikely that children’s early ability to discriminate between drawing and writing does not include a level of understanding about this defining feature of systems of written notation.

According to Kellogg (1970), there are 20 distinct kinds of markings which have been identified in the graphic productions of children of two years old and under. On the basis of the overall direction of the movement of the hand making them, these can be grouped into six categories: vertical, horizontal, diagonal, circular, alternating, and no line movement; the category ‘alternating’ includes wavy lines and zigzags; ‘no line movement’ refers to dots. These categories also include single line markings and multiple line versions: so vertical, horizontal and diagonal lines can be represented singly or multiply. The alternating lines include a ‘roving’ line which meanders and one which doubles back on itself, creating enclosed areas. One might surmise, though, that this is more likely to be the case when they are ‘drawing’ than when they are ‘writing.’ Fein (1993) has identified similar categories of markings, and her evidence also shows interesting individual variation within categories. So, for example, a meander (Kellogg’s ‘zigzag’ within the ‘alternating’ category) can be made with either a rounded or a sharp turning point. In this case the mark remains fixed in spite of local variation, presaging systems of print: handwriting of a sort. Baghban (1984) shows evidence of this range of marks being used in the productions of her daughter between the ages of 18 and 24 months. Both Kellogg and Fein show how these marks continue to be used and to develop in complexity as children get older, and Kellogg points out that the different categories of mark evolve into the construction of both written forms and forms used for artistic expression and design.

Whilst the marks themselves do not constitute a recognizable system of written notation, nevertheless, qualities intrinsic to them can be incorporated into the structure of the ‘point-plot organization’ which I have described. In the example discussed previously, (Lancaster, 1999), a single type of mark, a zigzag, is used throughout the making of the card. The use of a constant category of mark means that it is its placement that is the variable factor which signifies distinctions of meaning. Variation to qualities of the mark itself has the potential to introduce a further level of distinction. An extended zigzag, going from the top to the bottom of the page, is assigned the ascription ‘tree’; with its length and span, it also resembles a tree. Acknowledging the resemblance, the little girl adds ‘branches’ at the top; this time the zigzag is curved in an wide arch going from the ‘tree’ into the middle of the page, maintaining a link between representation and resemblance. The ‘writing’ at the bottom of the page also has a structural resemblance to that which it represents, being a linear sequence of small zigzags. These intentional and systematic variations to the graphic device itself suggests the operation of a basic morphology.

Continuous Development

The evidence already discussed in this chapter suggests that children between the ages of one and two have the capacity to distinguish confidently between writing and drawing, and to recognize and produce certain features, qualities and marks, characteristic of both these modes; even well before this age, they have the capacity to recognize many of these features. One construction which has been put on the relationship between these modes is that writing evolves from drawing, both developmentally and historically. Martlew and Sorsby (1995) suggest that children’s ability to draw writing provides evidence of this. However, the boundaries between writing and drawing are far from being fixed and absolute, and children move comfortably around them, using whatever is most salient and useful to interpret and communicate what is needful. In doing this, they are using not only elements which are characteristic of each mode, but also those which they have in common; and at this stage in life, they will commonly experience texts where writing and pictures operate in tandem. To draw writing is as reasonable as to write about a drawing, and would seem to demonstrate a sound understanding of the semiotic functions which are distinct to each mode, as well as those which are common to both.

A parallel can be drawn with interpretations of early writing systems, which suggest that they evolved through a highly pictographic stage to become increasingly abstract systems. However, evidence from the very earliest systems suggest that this is not the case (Schmandt-Besserat, 1978), and that graphic marking has always been used for different purposes, with drawing and writing having different representational functions, then as now. Writing developed to communicate information by graphic means (Harris, 2000; Gaur, 2000; Olson, 1994), whereas drawing developed for the purposes of aesthetic expression, even though the two modes might have drawn on similar types and methods of marking. It is likely that young children’s recognition of the distinction between the modes is also based on an understanding of this fundamental difference in function. Kenner’s example of a young Chinese boy who dismisses a fellow pupil’s idea that an oval shape could represent a mouth in Chinese, on the basis that in Chinese you have to write, ‘not draw pictures,’ demonstrates this point nicely. However, this still leaves open questions about development in each domain: what happens to these very early representational methods and structures which children construct and operate. Wolf and Perry claim that children’s early drawing systems are frequently regarded as being simply preparatory. Luria makes a similar point with respect to early writing systems, describing them as ‘primitive techniques’ that are ‘similar to what we call writing’ and which ‘served as necessary stages along the way’ (1983: 237). In other words, the outcomes of all that intense semiotic exploration simply wither away. The conclusions which Wolf and Perry draw about children’s early drawing systems, however, suggest a much more likely developmental path. Each of the systems which they develop and use is not so much a stage on the way to realism, as a system which continues to evolve and to remain useful; nothing is wasted. Mackey’s (2003) work demonstrates a similar path for the early use of bodily modes like gesture in communicating graphic meaning, with it continuing to have a significant communicative role to play in the interpretative strategies of much older children. At the same time, new systems are acquired, as and when they are needed: what Pariser calls ‘a kitbag of graphic strategies’ (1999: 104).

Important parallels can again be drawn. The ways in which infants and young children set about the interpretation and production of writing suggest that, like drawing, writing is also a multiply constituted mode. By the time they are two, children have been grappling with different genres, ways of displaying and representing meanings and information, spatial organization, systems of marking, fixed and flexible representations, morphological distinctions, and the movement between the interpretation and expression of writing. Each of these constituents of early writing continues to be used and to evolve. To reduce writing to the making of correct links between letters and sounds is to misrepresent both the nature of writing and children’s ways of constructing it. This is not to underestimate the significance of this feature of writing, but to make the point that it is one of many. Children discover a great many things about writing as a system before they reach the point of needing to investigate its relationship to speech. Development, in other words, is continuous. Far from this being a stage of limited semiotic activity, restricted by a lack of cognitive, social and linguistic development, it is a time when children are actively and independently interested and involved in representational matters.