Robert R. Verbrugge once wrote: “Identifying the origins of metaphor is a perplexing puzzle. The assumption is often made that metaphoric language is a special competence, peripheral to the chief vocations of language and acquired relatively late in development. Support for this view has derived from studies of metaphoric tasks on which only older children show success; estimates of the critical age have ranged from seven to twelve years or even older (Asch and Nerlove, 1960; Billow, 1975; Richardson and Church, 1959; Winner, Rosenstiel, and Gardner, 1976). But alternative views can also be advanced. One alternative is that metaphoric processes are present even before the onset of language learning, that they are exhibited throughout development, and that their variable appearance in behavior reflects a variety of cognitive and contextual constraints. One’s viewpoint on this question of timing is closely related to how one approaches several fundamental issues: the relation of linguistic metaphor to perception and action, the definition of metaphoric phenomena, the structure of metaphoric resemblances, the nature of comprehension processes, and the pragmatic context of metaphoric activity.” (p.77) I couldn’t agree more!
(Writing in 1979,) Verbrugge also notes: “One problem that has not received sufficient attention is the developmental relation between linguistic and nonlinguistic metaphor. Metaphoric processes can be manifested in perception (for example, when viewing a cloud as a lion), in action (when growling or springing like a tiger), and typically, in both at once (when searching for an appropriate stick and riding around on it as if it were a horse). These processes of play are as distinctive and difficult to define as linguistic metaphor. Distinguishing metaphoric extensions from other extensions predicated on similarity is of course, problematic; the observer seeks evidence (often in vain) that the child recognizes not only the appropriateness but also the inappropriateness of the familiar schema to the
new content. One example sometimes cited as a clear case is Piaget’s (1962) observation of little Jacqueline, who laughingly cradled her spoon like a doll (see Gardner and others, 1978).” (p.78)
“…it is very possible that these playful metaphors of perception and action are the prototypes of genuine metaphoric processes- not simply their fragmentary precursors. These nonlinguistic processes may be fundamental to linguistic metaphor in at least two respects. First, qualitatively similar experiences accompany the production and comprehension of metaphoric language. Language operates as an elaborate system of constraints that, among other things, can guide the reintegration of types of experiences specified originally in perception and action. Thus, language can catalyze imaginal experiences qualitatively indistinguishable from nonlinguistic play. In this sense, play is not an abandoned precursor of metaphor but recurs whenever people comprehend or produce it. An understanding of metaphoric styles of perception and action is an essential prerequisite for understanding their control of and by language. In short, the study of linguistic metaphor may have to begin in the turbulent dynamics of children’s fantasy.
A second sense in which nonlinguistic metaphor is prototypical derives from the fact that language itself is perceptually guided action. The perceptual information guiding some particular linguistic action, such as naming, can be preserved even when the substantive content changes in unusual ways: For example, a child may call a wind-swept field of grain an ocean or speak of a chimney as a house-hat. These overgeneralizations have provided a delightful headache for researchers on lexical development; it is very difficult to detect
when such expressions are only partial or playful applications of the child’s current schemata, rather than full, literal extensions of those schemata. Even when such a decision cannot be made with confidence, the child‘s expressions can be very informative because they are symptomatic of the kinds of schemata currently governing word use. They reveal the kinds of perceptual information to which the child is currently sensitive, and they constantly remind us how abstract and relational that information can be. An important corollary is that children cannot be expected to produce or comprehend metaphors based on information that exerts no control over their more conventional word use.” (pp.78-79)
“Focusing on play as prototypical metaphor alters one’s perspective on definitions of metaphoric competence. It leads one to see the preservation of a type of psychological process throughout the course of development, a continuity that is not as obvious when one takes adult metalinguistic competence as the departure point. This orientation also serves to redirect one’s attention away from linguistic forms to the psychological relations that give them significance.” (p.79)
“Metaphoricity is not a property of sentences as objects, but is a type of dynamic relation holding over utterances, language users, and perceived or imagined settings. The risk in treating metaphor as a preeminently linguistic phenomenon is that a particular linguistic attitude is adopted: Meaning can be ascribed to sentence-objects abstracted from communication settings. In defining metaphor, Gardner and others (1978, pp. 5-6) begin with the standard linguistic: taxonomy: Metaphors contain autopic” (or “tenor”) that is described in terms of a second element, the “vehicle”; the common property is the “ground.” However, they temper this definition with some psychological constraints: The linkage between the two elements must be ‘‘intentional and conscious,” and it must evidence an “awareness of tension.”” (p.79)
“When immersed in such an experience [as dreaming, reverie, fantasy, and active play], there need be no awareness of tension, no pangs of apology over barriers broken. In many cases …no incompatibility is present; the experience is simply unusual. In other cases, an incompatibility will only be experienced when one adopts a different attitude – after the fact. Thus, it seems too strong a test of the child at play, the dreamer, or the poet that they should be aware of a tension before we credit them with metaphoric thought.” (p.80)
Ref: Robert R. Verbrugge (1979) The Primacy of Metaphor in Development New Directions for Child Development, 6, pp.77-84