The first thing one learns about the origins of written languages is that pictographic signs play a very small part in their development and represent a very primitive stage. Then come cuneiform signs, hieroglyphs, logo-grams, logo-syllabic scripts, alphabetic scripts.
So the relationship of words, either spoken or written, to “things”, is a world away from that of visual sensation to its pictorial presentment. The pictographic representation of a tree has a morphological link to its object – this means that its significatory function is radically different from that of sign to “thing” in writing. In developed languages the link between signs and their objects has become arbitrary; not the case when it comes to painting. It is much closer to “reality” (however defined) than is the word.
Drawings on Mayan ceramic vessels are like a strip cartoon, with glyphs acting like bubbles above the figures; they have a marvellously fluid line-drawing style of great economy in deliminating figures. From an aesthetic perspective, decipherment of the glyphs (partly logographic, partly phonetic apparently) adds little to our appreciation of this art, or even its “meaning”, though of course of great interest from an anthropological point of view. Its “meaning” as art is not circumscribed by its meaning as historic record, and is, of course, entirely non-verbal.
The pictorial world is a direct modification of visual sensation concentrated on the one plane, akin to the construction of perception-confusing rooms in experimental psychology. Think of Piero della Francisco’s Flagellation of Christ, or Vermeer’s The Love Letter.
How words came to stand for things is one of science and philosophy’s impenetrable mysteries, involving as it does incantation, song, rhythmic chant and the kind of poetic imagination hinted at in the following passage from Robert Caldwell on the origins of the Dravidian language (and the Indus script, as yet undeciphered):
“Who that has seen the phosphorescence flashing from every movement of the fish in tropical seas or lagoons at night, can doubt the appropriateness of denoting the fish that dart and sparkle through the water, as well as the stars that sparkle in the midnight sky, by one and the same word – a word signifying that which glows and sparkles.” 
This is more suggestive of the way language developed, an exclamation of wonder, naming, and its connection to poetic vision, than any of the more prosaic speculative accounts currently in circulation.
Painting is not “structured like a language”, nor does sculpture share a syntax with language, nor does music. Hans Keller in his last book Criticism, says that music is not a language since it is cannot be translated (and as yet no written language can be deciphered unless it has links to another known language). Therefore structuralist and post-structuralist linguistic theory and literary criticism are irrelevant to the analysis of painting. All figures of speech, and the metaphoric, metonymy, synecdoche, trope etc., – indeed meta-anything, should be banished from critical discourse as a gross misapplication of language if applied to visual art.
And dreams are not “structured like a language”; only the interpretation of dreams. If they were, Finnegan’s Wake would be much more accessible than it is, and everyone their own self-engendered Joyce. Whereas, of course, only Joyce had the genius to run with the conceit that language could mirror the night-time journeying of everyman, an imaginative creation of his own vision of Irish archetypes in the phylogenetic origins of language, and the infinite malleability of “meaning”. The Wake is an extreme point of style and it has had no progeny.
The allure of subjectivity biased, obscurantist, pseudo-scientific French literary theory for nice-but-dim art theorists has wreaked untold damage to the tenor of art-critical discourse. Subjective obscurity in poetic literature is one thing, but it has no place at all in criticism. This is a clear case of the consequences of mis-applying one kind of discourse, however flawed, to another, in which it has no valid place – unless you want to resurrect the whole edifice of narrative symbolism that necessitated the forked-tongued double-voice with which even the greatest painters were obliged to address their patrons before the cleansing hand of “modernism”.
Noam Chomsky, a rationalist committed to a scientific exploration in structural linguistics, and somewhat of an expert in the field, concludes that the surface-structure solipsism of Saussurean semiotics is “totally inadequate” to deal with even the most elementary of the challenges to understanding thrown up by more recent and more sophisticated research in structural linguistics, grammar and syntax; let alone semantics. And he is skeptical as to the value of linguistics in shedding light on the perennial problems in philosophy, the concepts of the mind, thinking, the relationship of thought to experience. The way in which language “mirrors” the world is as daunting as ever, and how it originated, evolved, a surpassing mystery still.
Chomsky’s position is that instead of trying to make language conform to some devised system of logic, structural linguistics should study the hidden logic revealed by analysis of grammar as it occurs in enduring language use. Behind this surface-structure, there lies a deep structure which conforms to the logic of the mind as it tries to order experience, and to which all human languages conform. It has become innate over many centuries, such that a child, exposed to a very partial and fragmentary encounter with language use is more-or-less able to create new and meaningful sentences; i.e. learning not by rote, but by internalising the principles of grammatically correct speech, principles that are already potential at birth – just as birds are able to migrate long distances without being taught to do so.
And that words like “sign” or “symbol” do not come near to capturing how words actually do work, both in the formation of sentences, concepts, or in referring to aspects of “the world”, or experience. In writing what may have begun as a pictographic sign is quickly modified by the act of writing itself, the flow of the implement used etc., into a kind of short-hand in which the original sign is transformed until its pictorial element is lost. Not so in art, painting or drawing; here the short-hand – dots and dashes of paint or line retain their direct visual role and are continually brought back into a correspondence with the “facts” of visual sensation (even in abstraction).
Chomsky does not even mention the likes of Derrida, Lacan, Baudrillard, Lyotard and co., for the obvious reason that they operate in the utterly unscientific, disreputable genre of literary theory and cultural criticism which has its own hidden animuses and ideological agendas. Post-structuralism is as disreputable in relation to structural linguistics as post-modernism in relation to “scientific” or “rationalist” modernism; i.e. Cubism, Mondrian, de Stijl, though both rationalist and irrationalist are in operation – architecture and vertigo apply not only to Picasso’s analytical cubism, but to all great painting.
Yet in the shallows of art-theory the post-structuralist and post-modernist confusion still holds sway. I read recently in one such tract the announcement that “children paint symbols, not what or how they see”. The glib assurance of this mangling of fact and fiction is only made possible by an amateur dabbling in recent popularising dilutions of the arcane world of linguistics. What the author meant, I suppose, is that there is a conceptual element to every act of drawing, however innocent, but this act of simplification does not make it a symbol, or even a sign (in the terminology of linguistic theory at least) – it is a vision, a memory trace of the immediate coup d’oeil of sensation – “the humble transcription in terms of paint, of sensation itself”, reduced to its simplest cognitive terms – unconscious memory – and concentrated on the plane surface.
When Picasso said that it had taken him a life-time to learn to paint like a child (or un-learn), he paid tribute to the essential paradox of cognition versus impulse which lies at the heart of the modernist enterprise.
And when it comes to painting, the sheer delight in applying, or dabbing down blobs of brightly coloured pigment occludes signification, and thus brings us back to David Webb’s toy elephant. Is it possible to speak of a pre-significatory stage in child-art, and if so, how would we know, since the moment we ask a child “what it is”, we have tainted the process Heisenberg-style beyond hope of discovery? The last person I would consult about the stages of development in child art would be a linguist, just as the last person I would consult on the analysis of “expression” in music would be a semiologist.
My grand-children paint or draw houses, birds, mummy and daddy, snowflakes… as and how they “see” them, with wonder, as the fish in the lagoon and the stars in the midnight sky were seen – and named – with syncretistic vision (modified by millennia of phylogeny a la Chomsky).
But as already discussed, naming is not the same as drawing. A drawn sign for a tree, a flower, a bird, does not stand in relation to its object in the same way as the words do to theirs. And as Croce would have it, perception involves an intuition, irreducible to words, which can only be experienced in the “expression” of it – i.e. through an artistic transformation. It is essentially an aesthetic act. That is one way of putting it – there may be others as persuasive. So even cognition, perception involves the intuition of gestalts. To extend that experience over something as complex as a sequence of “effects of sensations apprehended in their relations” (Roger Fry) is what painting is all about. In the abstract picture, this “gestalt” will be relational, and spatial, and involves a plastic purpose.
We can always ask, “What is the significance of that signification?” If a child asks, “Why is the sky blue?”, our answer will rub up against the limits of scientific knowledge, if we have any. But if in a painting, the child asks “Why is the sky blue?” we will give quite a different answer. And more interestingly, if the question is “Why is the sky red?” – in, say, a painting by Altdorfer (Rome is burning, the sky is lit up by flames) – we will be forced to give an answer which involves not “What does it mean?”, but “What is the artist trying to do?” (Or if we are very clever we will ask the child back, “Why do you think it is red?)
When a child keeps on asking “Why?” they are asking for a more extended version of the connection between words and experiences, not just an account of how words or concepts interconnect, i.e. not just an exercise in semantics, grammar or syntax – though it may seem like that when one answers them. Underlying their questioning is an attempt to understand the world, and connect language to other parts of their experience – their emotions, fantasies, play, reverie, dream.
A symbol has a historically determined and culturally accepted “meaning” or complex of meanings (standing for different concepts in different contexts), whereas in modernist art it is precisely the novelty of the context in which pictographic signs find themselves that frees them from conventional “meanings”. The artist “personalises” them, renders them malleable to his or her pictorial world (even if quoting from some past references, or using them ironically), so their exegesis should primarily concern their autonomous function in the life history of the artist and in his or her dialogue with other artists; for example, the role played by apples in the private iconography of Cézanne and its “referencing” by Matisse later, in works like Bowl of Apples on a Table (Chrysler Museum, Virginia) or Apples (both 1916). What do these latter “symbolise”? – They symbolise an affectionate tribute to Cézanne, and to the health giving power which comes from great art.
Children’s art, apart from “signifying” people and things, is driven by emotions, fears, anxieties, and these are conveyed directly through the seismographic effect of welding an implement. The child’s emotion is telegraphed directly through the concentratesd effort of psychic energy released onto the page by the implement in hand. And it is this semi-conscious expression that is the envy of the mature artist.
 Such as those reproduced in Andrew Robinson, Lost Languages: The Enigma of the World’s Undeciphered Scripts, Thames and Hudson, 2009, pg 139
 Caldwell is quoted in Robinson, Lost Languages