"No people are cold!": On Young Children's Rejection of Metaphorization

One of the most influential books of the last decades is Lakoff and Johnson's Metaphors We Live By (1980), which lays the foundations for the theory of conceptual metaphor. In Metaphors We Live By, Lakoff and Johnson demonstrate that much, perhaps most, of our discourse, from small talk to philosophy, is undergirded by a vast network of metaphors. Unlike conscious, "literary" metaphors, these unconscious "conceptual" metaphors function not to create stylistic effects but to structure our thinking. They tend to be rooted in the domain of the physical, often in basic bodily experience. They allow us to make use of the physical to help conceptualize more abstract domains. We tend, to give just one simple example, to conceptualize abstractions like ideas as physical entities, and we often see them, via the so-called "conduit metaphor," as being placed into containers so as to be transmitted across space, as in "It's hard to put my ideas into words" or "It's hard to get that idea across to him" (Lakoff & Johnson, 1980, p. 11; Reddy, 1979). In the years since the appearance of Metaphors We Live By, the study of conceptual metaphor has generated a voluminous and ever-expanding body of work, not only in linguistics, but also in such fields as philosophy, psychology, and, especially, cognitive science.

The assumptions and the findings of the branch of cognitive linguistics that studies conceptual metaphor correspond in a number of ways to those of psychoanalysis. Both disciplines are concerned with the infrastructure of human thought processes. Both see those thought processes as largely unconscious (though cognitive scientists and cognitive linguists tend, mistakenly in my view, to see the "cognitive unconscious" as thoroughly distinct from the more dynamic unconscious that psychoanalysis is concerned with). Both disciplines assign privileged status to early experience, especially to early experience of the body. Indeed, a number of attempts have been made both to relate and to apply the theory of conceptual metaphor to psychoanalysis (Borbely, 1998; Carveth, 1984; Casonato, 1994, 2003; Holland, 1999; Lakoff, 1993; Melnick, 1997, 2000; Modell, 1997, 2000a, 2000b, 2003; and Rosenbaum & Garfield, 1996).¹ Very little has been written, however, that applies psychoanalysis to the unconscious processes connected to conceptual metaphor. The present article attempts to do just that. Raising the issue of how conceptual metaphor functions within the psyche of the individual, it focuses on the question of why young children, as a well-known field study indicates, reject or deny the very concept of metaphorization. The answer that it proposes, which points to an intertwining of cognitive and psychodynamic processes, has implications for theories of mental development both in psychoanalysis and in cognitive science.

* * * * * * * * * * * * * * * * * * * * * * * *

Cognitive linguistics has convincingly demonstrated the existence of conceptual metaphor as a phenomenon in language. But, though researchers have begun to qualify their claims concerning this particular issue (Gibbs, 1999; Gibbs & Steen, 1999), the theory of conceptual metaphor also implies that conceptual metaphor is a phenomenon within the individual psyche, significantly grounded in the experience—in particular the early experience—of the individual (Gibbs, 1994, p. 415; Johnson, 1987, esp. ch. 2; Lakoff, 1987, pp. 380-415; Lakoff & Johnson, 1980, pp. 57-59, and 1999, ch. 4; Melnick, 1997; Tolaas, 1991).² There is a problem associated with this claim, however, a kind of chicken-and-egg paradox. For in what sense can any individual say that conceptual metaphor reflects her infantile experience? Infantile experience is largely pre-verbal. Conceptual metaphor, however, is embedded in language, which the individual does not invent but acquires ready-made from outside herself.³

There is in fact a fairly simple, common-sense answer to the question of how conceptual metaphor can be acquired ready-made from outside ourselves while still expressing something meaningful about the early experience of the individual. For psychoanalysis has always supposed that we unconsciously retain some kind of memories of our early experience, even to some extent of our pre-verbal experience. The exact nature of those memories, to be sure, has never been perfectly clear. Today it is often supposed that memories, very early memories in particular, are not precise, fixed photographic residues, but, rather, plastic, changeable, post-factum constructions or reconstructions or, to use Modell's term (a translation of Freud's Nachträglicheit), "retranscriptions" (1990, esp. pp. 15-19 and pp. 62-64).⁴ But if we do have memories of preverbal experience—however shadowy, incomplete, plastic, and "reconstructed" they may be—then, when we come to learn language, there will necessarily be some kind of fit between those memories and our newly acquired linguistic categories and relations, be they syntactical, semantic, or metaphorical. This statement is in fact a truism, saying little more than that language, which we acquire from outside ourselves, nevertheless has some correspondence with what we perceive our own experience to be. To an extent, of course, this correspondence exists because our perception of our experience is largely structured by language. Indeed, it may be (in consonance with the theory that memory is reconstruction) that the acquisition of language itself restructures—and therefore distorts—our memories of pre-verbal experience. Such a phenomenon would help to explain infantile amnesia—we cannot remember experience which is outside the (largely linguistic) categories into which we organize experience. But the restructuring of early experience due to the acquisition of language cannot be total. If it were, infantile amnesia would also be total, and experience in any meaningful sense would begin only with language. In such a case, furthermore, a number of paradoxical questions would arise, not only about infants' acquisition of their mother-tongues (e.g, can one learn a language that has no significant links with one's previous experience) but also about the origins of language in general.

The common-sense view, then, would be that when we acquire language, we have some crude, basic sense that, at least to some extent, it corresponds to and expresses relations which we are already aware of from our previous, non-linguistic experience.⁵ And what is true of language in general is presumably true of conceptual metaphor. Presumably we possess, pre-linguistically, a kind of associative substrate⁶ in which, say, the experience of warmth is linked with the experience of affection.⁷ Thus, when we come to learn the linguistic expressions which follow from the conceptual metaphor affection is warmth, those expressions unconsciously feel right to us, even though we are not consciously aware of the connections involved. This is, of course, very similar to the process which lies behind much unconscious symbolism of the type "food = love" or "feces = gift."

* * * * * * * * * * * * * * * * * * * * * * * *

This common-sense explanation appears, however, to be contradicted by a well-known study of children's perception of conceptual metaphor carried out by Asch and Nerlove in 1960, two decades before the term "conceptual metaphor" was invented.⁸ Asch and Nerlove are interested in what they call "double-function terms"—words, like sweet or warm, that describe, on the one hand, "the properties and activities of things" and, on the other, "psychological activities or the properties of persons." The existence of such terms, they note, appears to be psychologically meaningful—"a previous study (Asch, 1955, 1958) has demonstrated that languages belonging to different families possess such terms, and that the languages agree significantly in the meanings assigned to them" (p. 47). Thus, in investigating children's understanding of double-function terms, Asch and Nerlove are attempting to "clarify the cognition of psychological events, and its relation to the cognition of physical events" (p. 47). That objective can, of course, easily be re-expressed in terms of the theory of conceptual metaphor—Asch and Nerlove are interested in how words that express our physical, bodily experience of the world come to be applied to the more abstract domain of character. All of the double-function terms they investigate—hard, cold, soft, bright, deep, and crooked, in addition to sweet and warm—can in fact be expressed as conceptual metaphors, whether of a low-level type like agreeable is sweet or of a more generalized type like character is a tactile experience. Indeed, at the most generalized level, the very concept of a "double-function term" as Asch and Nerlove define it can be seen as expressing a single conceptual metaphor, character is a sensory experience.

Working with fifty American children ranging in age from three to twelve, and divided by age into five groups, Asch and Nerlove asked each child a series of relatively abstract questions about whether this or that of the eight words tested could be applied to people, and if so, how and why. Once it was established that a child understood the physical meaning of the term, he would be asked, for example (p. 50), "Are people, too, sweet? Do you know any sweet people? How do you know they are sweet? What do they say or do when they are sweet?" Asch and Nerlove found the following three "trends":

(1) Children first master the object reference of double-function terms. (2) They acquire the psychological sense of these terms later, and then apparently as a separate meaning, as if in independence of the object reference the term already possesses. (3) The dual property of the terms is realized last, and then not spontaneously as a rule. (pp. 55-56, my italics)

Asch and Nerlove comment:

The striking finding is not that psychological meanings appear later in development, but that they are initially divorced from the corresponding object reference. The acquisition of psychological meanings does not, it appears, make contact with the physical meaning that the terms already possess. Taken at face value, this result signifies that, for the child, double-function terms are initially homonyms, and that only later he reaches a stage when he can discern the relation between them. (p. 56)

Now, "classic" though Asch and Nerlove's study may be (Gardner, Winner, Bechhofer, & Wolf, 1978, p. 7), there is something problematic about how it has been interpreted and how it has influenced subsequent research. In order to make my own objectives clear, I need, at this juncture, to clarify three points: (1) what Asch and Nerlove's purpose was in carrying out their study, (2) what sort of research into children's use of metaphor followed Asch and Nerlove's work, and (3) what specific use of Asch and Nerlove I make in the present article.

Prior to 1960, Asch had already written two papers (Asch, 1955, 1958) which in retrospect appear very much to prefigure the theory of conceptual metaphor. In the “explorative” (p. 48) and methodologically rather simple study that he carried out with Nerlove in 1960, Asch was explicitly interested in discovering what he could about how the use and understanding of “dual” or “double function” terms—and also the comprehension of their metaphorical import—developed in individual children. He was in fact attempting a rudimentary empirical investigation into the ontogenesis of conceptual metaphor, and one of the questions he was looking into was the one from which the present article stems: the extent to which conceptual metaphor functions actively within the individual psyche. Inevitably, the thinking set out in the present article makes significant use of what Asch and Nerlove found in 1960.

With the single exception of Özçaliskan, 2005 (see note 22), the questions that Asch and Nerlove investigated and the further questions that arise from their work have not, as far as I can see, been revisited by empiricists interested in the theory of conceptual metaphor, although they are essential questions which it would be extremely useful to re-investigate with the conceptual and empirical tools available today. In the 1970’s, though, and to a lesser extent in the 1980’s there was a great deal of interest in studying the ability of children to create, use, and understand metaphor, analogy, and analogical reasoning. A number of the investigations from this period are mentioned below, both in my text and in my notes. Only one of these further studies, however (concerning an attempt to train children in the use and understanding of double function terms [Lesser & Drouin, 1975]) was a direct outgrowth of Asch and Nerlove’s. To what extent the other studies of the 1970’s may have been actually inspired by Asch and Nerlove is a moot point. All of them, to be sure, make due acknowledgement of Asch and Nerlove, but their conceptual orientation is usually quite different from Asch and Nerlove’s—understandably so, since at the time Asch and Nerlove’s orientation must have appeared to be highly idiosyncratic. Often these later studies imply or appear to imply that Asch and Nerlove somehow accused children of having no metaphoric ability, and the subsequent studies, which tend to avoid direct questioning of children about what they have understood, strive to show that, though children may lack the intellectual ability to speak metalingually about metaphor (as Asch and Nerlove had in fact asked them to do), they nevertheless possess, in one way or another, many of the competencies called for in the production and understanding of metaphor. Valuable as these subsequent studies are in other respects, none of them is concerned with the ontogenesis of conceptual metaphor.

In point of fact, one of the two questions that Asch and Nerlove set out to elucidate—i.e, to what extent children are aware of the connection between the two senses of a “double function” word—is a metalingual question,⁹ one that is, moreover, of central importance for understanding how conceptual metaphor functions within the individual psyche. Asch and Nerlove’s answer to it, we have seen, is that until around the age of eleven, children are simply unaware of the connection between “physical” and “psychological” meaning (and that when older children become aware of it, they do not usually do so “spontaneously”). Though that fact surprised Asch, it is not on the face of it especially surprising. There is, furthermore, no evidence to the contrary (although it is true that no one has seriously looked for any). It is not a difficult conclusion to accept, and I have adopted it as a working hypothesis. Nevertheless, it is not that particular conclusion (which the authors themselves considered to be their most salient point) that I find most interesting in Asch and Nerlove’s article. What I find most striking is something that the authors themselves mention only incidentally—the surprising amount of affect displayed by a number of the youngest children in the study (the three and four year olds) in asserting the lack of connection between the two senses of a double function term. This was not something that Asch and Nerlove set out to look for, nor was it something to which they attached any very great significance. It is, rather, something that I found potentially significant as I reflected on their article. Combined with other, psychoanalytic, considerations, it has led me to a series of speculations, set out below, about the relation between metaphor and certain psychodynamic processes in very young children.

The affect that Asch and Nerlove report is related to their third conclusion, about which they note, as mentioned above, that "for the child, double-function terms are initially homonyms." Once again my initial question arises. For if the speakers of a language acquire conceptual metaphors as mere homonyms, what is the psychological status of the metaphorical relation involved? Clearly that metaphorical relation exists in the language. But does it exist in any active way in the minds of individual speakers?

The question may be clarified if we remember that conceptual metaphors normally function unconsciously. "Unconsciously" here is not a psychoanalytic concept—the relevant unconscious is the so-called "cognitive unconscious." Simply, according to cognitive scientists, most cognition takes place without our being aware of the processes involved. When we speak our mother tongue, for example, we make unconscious use of its syntax (and do so correctly, even though we may not be able to articulate the principles behind the syntax we use). Similarly, the theory goes, we employ conceptual metaphor without being aware of what we are doing and without necessarily being able to express what we are doing. Asch and Nerlove recognize this, and explicitly point out (p. 57) that double-function terms normally operate unconsciously, in adults just as much as in children.

They also point out, however, an interesting difference between adults and younger children. Adults are not usually consciously aware of the relations between the physical and psychological meanings of double-function terms, "but once their attention is called to it, they are quite capable . . . of realizing and explaining them" (p. 57). Indeed, so too were the oldest children in Asch and Nerlove's study. The younger children, however, were incapable of explaining the relations between the two meanings, and at times emphatically denied even the theoretical possibility of such relations.

Asch and Nerlove reach the conclusion, later confirmed by Lesser and Drouin (1975), that the "segregation of the dual meanings is much stronger at earlier stages of development" (p. 57) and that children's perception of double-function words develops over time. But what I find particularly interesting is that the attitude of children towards the relationship between the physical and psychological meanings of the same word also develops over time. Among three and four year olds Asch and Nerlove found a tendency to reject—sometimes vehemently—the very idea that a word like deep might be applicable to people. Five and six year olds made only rare use of the psychological meaning of double-function words. Children of seven and eight tended to recognize that double-function words had both physical and psychological meanings, but usually failed to acknowledge any relation between the two. Eleven and twelve year olds, however, showed not only "a noticeable advance in the comprehension of the dual function" (p. 53) but also, often, a keen interest in it. These children gave "few reports of having earlier thought about the dual meanings," but when that question was brought to their attention, many of them "were completely fascinated" (p. 54).

Asch and Nerlove's study thus seems to contradict the common sense explanation I outline above of how conceptual metaphor can express the pre-linguistic experience of the individual. If, to remain with my previous example, the association between affection and warmth is rooted in very early experience, then one would imagine that the youngest children would be the ones with, in some sense at least, the strongest awareness of it. It is easy to understand that their cognitive development might not yet enable them to state the connection for themselves, but there is no apparent reason why they should not at least see it or respond to it when it is pointed out to them. (As already mentioned, an extensive body of empirical work, summarized in Winner [1988] and in chapter 9 of Gibbs [1994], shows that although the metalinguistic ability to talk about metaphor develops relatively late, many of the capabilities involved in metaphorization are present from a very early age. Gardner, Kircher, Winner, & Perkins [1975] found that in a metaphor-production exercise pre-school children produced a greater number of appropriate metaphors than primary school students. See also Gardner et al. [1978] and Gardner & Winner [1979, esp. pp. 130-134].) And yet the youngest children not only fail to perceive any such connection—they refuse, often emphatically, to envisage even the possibility that any such connection might exist. "No people are cold!" says one of them. "I never heard of deep people anyway!" protests another (p. 50). Only towards the age of eleven do Asch and Nerlove's subjects become receptive to seeing the connection, and even then they see it only when they are invited to.

* * * * * * * * * * * * * * * * * * * * * * * *

The very vehemence with which the youngest children reject the possibility of any relation between physical and psychological meanings may provide a key for understanding their perceptions of metaphor. The "surprised and indignant" (p. 50) reactions of these children are strongly colored by affect, and psychoanalysis is perfectly familiar with the phenomenon of affectively charged refusal to envisage the possibility of certain truths. Within psychoanalysis such a refusal is likely to be seen as denial, a defense mechanism. And indeed in the reactions of the three and four years olds interviewed by Asch and Nerlove there is something—less the refusal itself than its accompanying affect—that does feel like a defense. It is thoroughly possible that denial, perhaps accompanied by other types of defense mechanisms, is operating in their responses.¹⁰ If so, a psychoanalytic explanation may be called for.

It is true that for the affect involved in the responses of the young children two non-psychoanalytic explanations are also available, but neither of them is really adequate. The first is that the demands made on them by Asch and Nerlove are just too advanced for children of three and four¹¹—that the cognitive abilities of children that age do not allow them any metalingual understanding of metaphorization, so that they are simply discombobulated by Asch and Nerlove's questions. This explanation (which is similar to one often given for the extreme literalism of the autistic) is commensensical enough, but it begs the question of why, in this particular instance, young children seem to perceive their own cognitive incapacity as particularly threatening. There are a great many concepts that three and four year olds do not understand, but they do not react to all of those concepts, as some of them react to the possibility of certain metaphors, with rejection and indignation.¹²

The second non-psychoanalytic explanation for the children's response is that our culture actively promotes literalism in children, encouraging them to equate literalism with being grown up. That fact could indeed account for some of the affect found in their rejection of metaphor. But there is a flaw to this explanation. For the principal cultural factor that fosters literalism is schooling, early schooling in particular. Gardner et al. note "that figurative language declines during the early years of school" (1978, p. 13),¹³ and Glicksohn and Yafe stress that "pressure to be verbally proficient in a literal fashion" (1998, p. 202) begins in Grade 1. Logically, then, cultural encouragement of literalism should be less likely to play a major role with the pre-school children under discussion than with slightly older children. But in fact the older children that Asch and Nerlove tested (including kindergartners and second graders) did not display the same sort of vehement denial found in some of the pre-schoolers. Asch and Nerlove note explicitly that the "strong denial of applicability disappeared in the older groups" (p. 50).

It seems, then, worthwhile to seek a psychoanalytical explanation, one that supposes that in their affectively charged rejection of certain metaphors (and perhaps of metaphorization in general), Asch and Nerlove's pre-schoolers are utilizing some kind of defense. But what is being shut out? If a defense has been called up, just what is being defended against? To that question I in fact find two distinct answers, each of them connected to a separate mechanism. One of the answers has, as indicated above, to do with denial. The other, which I will take up first, has to do with simple repression.

I have already alluded to the similarity (or identity) between the acquisition and use of conceptual metaphor and the acquisition and use of a certain kind of crude unconscious symbolism. Unconscious symbolism of that type—pillows or balloons for breasts, guns or rocket ships for phalluses—remains subject to repression not, presumably, because there is anything especially forbidden about the symbolism itself but rather because its subject matter (breasts, phalluses, etc.) is closely bound up with frightening, perhaps forbidden feelings or wishes. That same explanation may well hold for a great deal of conceptual metaphor. That the symbolic process involved should be unconscious we can take for granted, since most cognitive processes are unconscious. What requires explanation is young children's vehement refusal to admit the possibility of certain metaphors, e.g, of sentences like "She's a deep one." And with young children the explanation could well come, as with unconscious symbolism, from the content of the words in question. Among Asch and Nerlove's eight terms, bright and crooked are perhaps less emotionally charged, but soft, warm, sweet, hard, cold, and deep are all words that are easily associated with libidinally invested infantile experiences having to do with (or fantasized in terms of) objects like the breast and the phallus, and so are likely to have close associations with potentially frightening infantile feelings and with repressed infantile wishes.¹⁴

It may not be entirely fortuitous that the particular metaphors that Asch and Nerlove chose should tend to have those associations. For although the "target domains" of conceptual metaphors may include some very sophisticated activities (like philosophy or mathematics), their "source domains" appear to come down to simple, basic experiences. Mark Johnson (1987) argues that behind conceptual metaphor there lies a certain number of fundamental structures of perception (e.g, container or blockage or near-far) that he calls "image schemas." He lists twenty-seven of them (p. 126), but suggests that there may be others. In a similar vein, Grady (1997) argues that conceptual metaphors result from the combination and recombination of certain elemental metaphors (like existence is visibility [p. 284] or quantity is size [p. 285]) that he calls "primitives." The implication of Grady's work and perhaps of Johnson's as well is that all conceptual metaphor can ultimately be traced back to a relatively few very elementary experiences. Since they are elementary experiences, they would be (as some theoreticians have explicitly said) early experiences, i.e, experiences going back to infancy.

Given the limitations on the cognitive development of the human infant, given the particular emotional investments of the infant (and their particular nature), and given the importance of the primary process in infantile thinking, these elementary experiences necessarily partake of the irrational infantile fears and wishes that are so large a part of the raw material of psychoanalysis. An obvious example would come from the much-analyzed "container" schema, present not only in a sentence like "The milk is in the fridge," but also in one like "Is that meaning really in the poem, or just in the reader's mind?" (The container schema is also present—and not merely in the quoted examples—both in the sentence you have just read and in the one you are reading now.) But, as Turner has pointed out, our initial awareness of the container schema has a great deal to do with our "early experience with objects going into or out of the body" (1994, p. 96). Now, of that early experience a very high proportion concerns the libidinally invested activities of eating and excreting. As a consequence, our very concept of a container is colored from the outset—and will forever remain colored, to one degree or another—by the infantile fantasies and wishes and fears connected with eating and excreting. The general point exemplified here—that cognitive perceptions and operations are unconsciously affected by often irrational infantile emotions—is a commonplace of psychoanalytic thinking, though it has too often gone unnoticed by cognitive scientists. Certainly that fact could help to explain why nursery school children seem to have strong defenses against speaking metaphorically of people as soft or warm or hard. Children of nursery school age are, after all, just emerging from a time when they passionately experienced people as having those qualities literally. They may still be too close to the passionate feelings and fantasies involved to want to be reminded of them.¹⁵

This explanation does require one qualification, however, in that one of Asch and Nerlove's double-function terms—sweet—did not follow the same pattern as the others. It was the only word that the children in the youngest group "applied to persons in a sense other than physical." Seven out of ten of them used it of people, with the general meaning of "good," "nice," or "likable" (p. 50). Lesser and Drouin similarly found that for all ages concerned sweet was the "easiest" of the nine double-function terms that they tested and taught (p. 292). The reason, as Lesser and Drouin suggest (p. 297), probably has to do with the children's linguistic experience rather than with psychodynamics. Young children are themselves so habitually addressed as "sweetie," "sweetie pie," "honey," etc. that repression would have to be of a truly adamantine severity for them not to realize that "sweet" is a word applied to people whom one likes or loves. But obviously the repression involved cannot be as adamantine as that—if it were, it would not be overcome as effortlessly as it eventually is, with nothing but the passage of time (and concomitant maturation) to bring it about. Presumably if the standard pet names for infants expressed softness, or warmth or brightness, then very young children would also acquire the double functions of soft, or warm, or bright.¹⁶

It does not seem, then, that the early use of sweet in reference to people argues against repression's being the reason why young children deny the existence of double-function terms. The use of sweet shows simply that the repression involved in the denial of double-function terms is not so strong that it cannot be undone. In point of fact, it is naturally and spontaneously undone in the normal course of events—and the reactions of Asch and Nerlove's oldest subjects to achieving a conscious understanding of double-function terms perhaps afford further indication that repression has indeed been at work. Asch and Nerlove report that the "distinguishing feature" of the oldest group of children that they tested (essentially eleven year olds) "was their interest in the questions and in the general subject of the investigation. Many were completely fascinated by the problem and often turned from the consideration of one term to another with great reluctance" (p. 54, my italics). It seems difficult to explain the children's intense interest on purely cognitive grounds. But the fascination of these sixth graders does recall the intellectual fascination that an analysand may display with material that has recently become accessible to consciousness—a feeling that so much that had previously been taken for granted and left unexamined can now be actively and consciously understood. A reaction of this sort, which requires that previously repressed material have entered into awareness, is highly gratifying, intellectually as well as emotionally. Reactions of this type probably created some of the early fascination with psychoanalysis itself. Today they are helping, I suspect, to create the fascination that one finds in certain fields with the theory of conceptual metaphor.

But all of this concerns repression only. It says nothing about the denial which appeared so prominently in the responses of Asch & Nerlove's youngest subjects—and which seems to me to be the stronger of the two mechanisms involved. A more complete explanation requires going beyond the content of the metaphors involved (or of the infantile feelings and fantasies connected with them). It also requires examining the nature of the metaphoric link—a link which, there is reason to believe, children unconsciously perceive as frightening in and of itself.

* * * * * * * * * * * * * * * * * * * * * * * *

When psychoanalysts write about conceptual metaphor, they tend to stress the tendency of analytic patients (and others) to "literalize" metaphor. Though each uses different terms for the phenomenon, Carveth (1984 and n.d.), Modell (1997), and Enckell (2002) all write about the (usually unconscious) tendency to take metaphor as expressing genuine equivalence or identification rather than mere similarity.¹⁷ (Borbely's discussion of "analogical repetition" [1998] follows very similar lines, although it lays less stress on equivalence or identification as such than on the loss of polysemy involved.) This phenomenon is so common—it is, among other things, the basis of transference—that any analyst can supply examples of it. Carveth (n.d.) mentions a patient who perceived the defense of his doctoral dissertation as submission to a gang-rape and another who identified his penis with a Cruise missile, and goes into some detail about a third patient, who, in dangerously "concrete" fashion, saw his wife as a vampire. Campbell and Enckell (2005) recount how with one of their patients actual violence resulted from the "concretised" feeling that his girl friend was suffocating him (p. 811). In another, particularly vivid illustration (Segal, 1975, cited by Modell, 1997, p. 226), a schizophrenic patient, upon being asked why he had stopped playing the violin, replied, "Do you expect me to masturbate in public?"

That vivid examples should come from schizophrenic patients is not, of course, fortuitous—the "degeneration" (Carveth, 1984, p. 507) of metaphorical resemblance into literal identity is, as Carveth, Modell, and Enckell¹⁸ all point out, well known "as part of the thinking disorder that characterizes schizophrenia" (Modell, 1997, p. 226). Carveth and Modell are, however, mainly concerned with the manifestation of that degeneration in neurosis (as is Borbely). They see it as a rigidification of metaphor, which transforms "living" metaphor into something fixed and unchanging. Carveth thus speaks of "dead," and Modell of "foreclosed" metaphors. All four of the authors I have cited consider that one of the functions—very possibly the function—of analysis is to revitalize these "dead" metaphors or, more generally, to revive the capacity to create "living" metaphors.

Analytic writing on conceptual metaphor stresses, then, a psychic potentiality, particularly visible in schizophrenia and very widely present in neurosis, to perceive absolute, rigid equivalences between entities which do not in reality support so totalizing an identification (though they may bear some metaphorical resemblance to each other). This phenomenon has in fact been studied in great detail, albeit from an idiosyncratic point of view, by Matte Blanco (1998), who refers to it as "symmetrization"—an expression of what he calls the "principle of symmetry." (In his explication of Matte Blanco, Rayner defines "symmetrization" as the "process of ignoring certain asymmetrical or difference relations," and explains that "the process could be referred to by such neologisms as 'identicalization'" [1995, p. 25].) Matte Blanco points out, among other things, the importance of his "principle of symmetry" in schizophrenic thinking (pp. 37-39, 54-55) and in primary process thinking, and makes a plausible argument that it underlies Freud's description of the "special characteristics of the system Ucs" (pp. 41-47). He himself sees it—this is, indeed, the thesis of his dense, lengthy book—as one of two fundamental principles of mental functioning.

However it is termed, "literalization" or "symmetrization" or "identicalization" means a failure to discriminate or differentiate. This sort of non-differentiation is found, we have seen, in many neurotic symptoms and probably in dreams as well. It is connected to the primary process, and it is very visible in schizophrenia. Though it is potentially present throughout life, it originates, presumably, in infantile experience. Indeed, at one time it was a commonplace in theories of development that the infant's earliest perceptions are non-differentiated—i. e, that the infant fails to discriminate one object from another, internal reality from external reality, self from other. Today, to be sure, we have abundant evidence that no completely "non-differentiated" stage exists, that from birth on (and possibly before) the infant makes a significant number of discriminations (Maurer & Maurer, 1988, p. 10, pp. 84-85, p. 108ff, p. 126, p. 145, p. 148 and passim; Stern,1985, p. 10 and passim). Nevertheless, an infant's ability to differentiate is unquestionably weaker than an adult's or an older child's—and it may be different. There is good reason to believe, for example, that infants live in a largely "cross-modal" world (Maurer, 1993; Maurer & Maurer, 1988, pp. 65-67, pp. 196-98; Meltzoff & Moore, 1993; Stern, 1985, p. 51ff, pp. 154-56). Indeed, it has been hypothesized that an outright synesthetic stage is part of infancy and that when synesthesia fades from the scene, it does so largely on account of anatomical changes in the brain that occur as part of the normal developmental process (Maurer, 1993, pp. 112-13).

This would mean that at a certain point in development the older infant or very young child would find himself perceiving the world in a very different—and more highly differentiated—way than in the recent past.¹⁹ If at this point (or later) the infant or child retained any kind of memory or pseudo-memory of his earlier, less differentiated modes of perception, that memory would necessarily (on account of the development of the brain, and the changes in its circuitry) be of a very shadowy and insubstantial nature. Given that, and given our tendency to "retranscribe" our memories (Modell, 1990, 1997), as well as our tendency to structure experience in terms of polarities, the older infant or very young child might well retranscribe this tenuous pseudo-memory into a vague, pre-conscious or unconscious sense of an earlier time characterized by "non-differentiation" or "unity" and contrasted with the "differentiated" present. (It may be that theorists have been so ready to assume an early "undifferentiated" period precisely because they themselves, like all of us, have retained an oversimplified, partially fictitious pseudo-memory of such a time.) In this oversimplified and partially fictitious representation of the past, the earlier period would no doubt appear to have had advantages lacking in the present: fusion with the environment (especially the mother), omnipotence, etc. But at the same time the representation of that earlier phase is likely to be colored by a feeling of having had, during that earlier period, even more difficulty than at present in making sense of the world, and also of having been even more helpless and dependent.

Now, such non-differentiation as in fact exists in early infancy is not of a type in which one discriminated object is interchanged with another. On the contrary, it is probably of a type in which objects are simply not distinguished as such, apart from the unified "events"²⁰ in which they participate. This failure to distinguish objects is likely, however, to be perceived in retrospect (once the concept of "object" has itself been established) as implying a confusion or interchangeability among objects, such that any object can be any other. This latter principle—the interchangeability of objects—is, I suspect, very closely associated with the pre-conscious or unconscious representation that the older infant or young child retrospectively (and, in part, inaccurately) constructs of his own mode of perception during the first months of life. In itself, of course, the principle of interchangeability is indispensable in enabling the process of metaphorization, which consists of saying or implying that one thing is another thing. But though it is a necessary condition for metaphorization, it is not a sufficient one. Genuine metaphorization also requires a principle of differentiation or contrast. For if a statement of the type "A is B" is to make sense, A and B must at least seem to be different. "Magruder is a lion when attacked" makes sense since we know that Magruder is a human being, not a lion. But we would not normally say, "Simba is a lion when attacked."

The principle of the interchangeability of objects enables us, then, to make statements of equivalence; but in order to interpret those statements of equivalence as metaphors (i.e, comparisons) rather than as literal identities we need the principle of differentiation. And if the principle of differentiation operates only weakly in an infant's earliest perceptions and gains in strength over time, then there is bound, in the early stages of psychological development, to be some confusion concerning metaphorical assertions of equivalence, both explicit and implicit. Intellectually, much—or most—of that confusion will have dissipated by kindergarten age, although vestiges of it may remain.²¹ But the effects of that confusion on unconscious psychodynamics will be much longer lived.

Most adult logic and perception depend on an "axiom of non-contradiction," according to which, for example, a proposition cannot be simultaneously true and false, or (to extend the axiom more broadly) two objects cannot occupy the same space at the same time. Obviously the axiom of non-contradiction (the absence of which Freud lists among the "Special Characteristics of the System Ucs." [1915/1961, pp. 186-87]) cannot operate without logical and perceptual differentiation. Given that, whenever one feels any significant uncertainty about one's own ability to make correct "adult" discriminations, then metaphorical utterances, which on the face of it deny differentiation and discrimination, may re-kindle unconscious fears of losing one's grip on the adult perception of reality, which depends so largely on the ability to differentiate and discriminate. In an adult, to be sure, any such fears are greatly attenuated. If they occur, they may take the form of fear of madness (schizophrenia). But in a young child, whose mastery of differentiation is more recent and hence less confident, unconscious fears of this sort are likely to be stronger and more frequent. They would usually take the form of fear of regression to an earlier phase of life—a phase which may have its appeal (connected, among other things, to memories of one-ness with the environment) but which is nevertheless frightening in that it is remembered as a time of utter dependency, the result of one's intellectual and physical helplessness. (The same retranscribed memories may also color an adult's apprehension of madness.) In the actual course of development these fears may be relatively slow to arise, as the opposition undifferentiated/differentiated cannot be constructed until after the child has already had a certain experience of differentiation. But such would be the case with those three and four year olds who responded with such vehement denial to Asch and Nerlove's suggestions that people could be deep or hard.

It would seem, therefore, that two reasons, both connected to psychodynamics, come together to make young children particularly resistant to metaphor. One of these reasons is that so many conceptual metaphors—certainly most of those that Asch and Nerlove asked about—are connected to libidinally invested activities and libidinally invested parts of the body. To admit them would be to reawaken a whole slew of powerful feelings and wishes, including many (which may be linked to Oedipal conflicts) that are frightening or forbidden or both. The other reason is that a premature acceptance of metaphor—especially of the kind of metaphor that is embedded in one's language (as opposed to the kind that one deliberately creates for special effect)—provokes unconscious fears about regressing to a (partially fantasized) time, largely remembered as one of helplessness and dependency, before we had begun to master the most basic cognitive skills²² that adults require in order to cope with the world.

This second hypothesis would also explain the re-awakening of interest in metaphor towards the end of the latency period. For if the latency period has done its work, then as the child emerges from it, she will feel confident of her ability to use language literally and to analyze the world in terms of the conventional adult categories of her culture. (See, on this point, Gardner et al., 1978, p. 20.) Under normal circumstances, the fear of regression to a "non-differentiated" phase of development will no longer be significant, and the pre-adolescent will be free to discover, if her attention should be called to it, the metaphorical relations embedded in her language. Indeed, she may be free to appreciate metaphor—to experience the gratifying "shock of recognition" that metaphor produces—as she has not previously been. For much of the feeling of numinousness involved may well come from a connection between the operation of metaphor and unconscious memories (fictitious though they may in part be) of the positive aspects of the time of non-differentiation—i.e, of unity with one's surroundings linked to a kind of mental omnipotence.

* * * * * * * * * * * * * * * * * * * * * * * *

I have, in the course of this article, proposed four inter-related hypotheses:

1. Ther exists a kind of associative substrate which links the individual's pre-verbal experience to subsequently acquired conceptual metaphors.
2. Until puberty a kind of repression operates that interferes with conscious understanding of many conceptual metaphors.
3. One reason for this repression is that many conceptual metaphors arise out of experience involving parts of the body that are libidinally invested during early development.
4. The nature of the metaphorical link, which implies the possibility of full identity between two different things, is itself frightening, since it recalls a very early period of life characterized (in memory) both by a relative inability to differentiate and by nearly total helplessness.

All four hypotheses are speculative, of course. But some of them may be testable. If the third hypothesis is true, then children ought to respond with less inhibition to metaphors that are further removed from conflicts concerning libidinally invested body parts and with greater inhibition to metaphors that are closer to them. Criminal is crooked (Cienki, 1998, p. 128), for example, should provoke less inhibition than the body is a container for the emotions (Lakoff, 1987, p. 383). On the face of it, testing that prediction should be relatively easy. It might also be possible to test the fourth hypothesis. One conceivable procedure for that hypothesis has to do with the difference between metaphor, which ostensibly states an identity, and simile, which states only a similarity. The hypothesis is about fear of non-differentiation. It would, therefore, predict that in principle pre-school children will be more receptive to simile ("John is like a lion") than to metaphor in the strictest sense ("John is a lion" or "John roared").²³ There is in fact evidence that autistic adults, who are known for their literalism, use simile much more readily than metaphor (Happé, 1991, pp. 237-238),²⁴ and that children understand simile more readily than metaphor (Reynolds & Ortony, 1980; Vosniadou, Ortony, Reynolds, & Wilson, 1984). My fourth hypothesis deals, of course, not with use or with comprehension as such but with metalingual acceptance of the two types of figurative language. Testing it would no doubt require a complex set of experiments. Nevertheless, if workable tests were devised and if even one of these hypotheses were to be confirmed, that would have significant implications for future thinking and research about the ontogenesis of conceptual metaphor. Any such confirmation would suggest that, in at least in some important instances (and perhaps, as I have argued elsewhere [Melnick, 2000, p. 242], in all instances) there is considerable interaction between the "cognitive" unconscious of cognitive science and the "dynamic" unconscious of psychoanalysis.

Acknowledgement: For their comments on earlier drafts of this article I am greatly indebted to Donald Carveth, Henrik Enckell, Norman Holland, Arnold Modell, Tim Rohrer, and the anonymous reviewers of the Canadian Journal of Psychoanalysis. I am also grateful to those who responded to abridged versions of the article presented at the Nineteenth International Conference on Literature and Psychology (Arezzo, 2002) and the Conference on Language, Culture and Mind (Portsmouth, England, 2004). For bibliographical and other information, I owe thanks to seyda Özçaliskan.

References

Asch, S. (1955). On the use of metaphor in the description of persons. In H. Werner (Ed.), On expressive language: Papers presented at the Clark University Conference on Expressive Language Behavior (pp. 29-38). Worcester, MA: Clark University Press.
Asch, S. (1958). The metaphor: A psychological inquiry. In R Tagiuri & L. Petrullo (Eds.), Person perception and interpersonal behavior (pp. 86-94). Stanford: Stanford University Press.
Asch, S., & Nerlove, H. (1960). The development of double function terms in children: An exploratory investigation. In B. Kaplan & S. Wapner (Eds.), Perspectives in psychological theory: Essays in honor of Heinz Werner (pp. 47-60). New York: International Universities Press.
Baron-Cohen, S. (1996). Is there a normal phase of synaesthesia in development? Psyche, 2 (27). Retrieved October 5, 2006, from http://psyche.cs.monash.edu.au/v2/psyche-2-27-baron_cohen.html
Borbely, A. F. (1998). A psychoanalytic concept of metaphor. International Journal of Psychoanalysis, 79, 923-936.
Campbell, D., & Enckell, H. (2005). Metaphor and the violent act. International Journal of Psychoanalysis, 86, 801-823.
Carveth, D.L. (1984). The analyst's metaphors: A deconstructionist approach. Psychoanalysis and Contemporary Thought, 7, 491-560. Republished in PSYART (2001). http://www.clas.ufl.edu/ipsa/journal/2001/carvet04.htm
Carveth, D. L. (n.d.). On therapeutic deconstruction: Resurrecting dead metaphors in psychoanalysis and religion. Unpublished manuscript.
Casonato, M. (1994). Metafore. Rome: Nuova Italia Scientifica.
Casonato, M. (2003). Immaginazione e metafora. Rome: Laterza.
Cienki, A. (1998). Straight: An image schema and its metaphorical extensions. Cognitive Linguistics, 9,107-149.
Enckell, H. (2002). Metaphor and the psychodynamic functions of the mind. Kuopio University Publications D. Medical Sciences 265. Kuopio, Finland: Kuopio Yliopisto [Kuopio University].
Fast, I. (1985). Event theory: A Piaget-Freud integration. Hillsdale, NJ: Erlbaum.
Freud, S. (1961). The unconscious. Standard edition (Vol. 14, pp. 159-215). (Original work published 1915.)
Gardner, H., Kircher, M., Winner, E., & Perkins, D. (1975). Children's metaphoric productions and preferences. Journal of Child Language, 2, 125-141.
Gardner, H., & Winner, E. (1979). The development of metaphoric competence: Implications for humanistic disciplines. In S. Sacks (Ed.), On metaphor (pp. 121-139). Chicago: University of Chicago Press.
Gardner, H., Winner, E., Bechhofer, R., & Wolf, D. (1978). The development of figurative language. In K. E. Nelson (Ed.), Children's language (pp. 1-37). New York: Gardner.
Gentner, D. (1977). Children's performance on a spatial analogies task. Child Development, 48: 1034-1039.
Gibbs, R. W. Jr. (1994). The poetics of mind. Cambridge, UK: Cambridge University Press.
Gibbs, R. W. Jr. (1999). Taking metaphor out of our heads and putting it into the cultural world. In R. W. Gibbs Jr. & G. J. Steen (Eds.), Metaphor in cognitive linguistics (pp. 145-166). Amsterdam: Benjamins.
Gibbs, R. W. Jr., & Steen, G. J. (1999). Introduction. In R. W. Gibbs Jr. & G. J. Steen (Eds.), Metaphor in cognitive linguistics (pp. 1-8). Amsterdam: Benjamins.
Glicksohn, J., & Yafe, T. (1998). Physiognomic perception and metaphoric thinking in young children. Metaphor and Symbol, 13, 179-204.
Grady, J. (1997). Foundations of meaning: Primary metaphors and primary scenes. Dissertation Abstracts International, 59 (3), 803. (University Microfilms International No. AAC 9828708).
Happé, F. G. E. (1991). The autobiographical writings of three Asperger syndrome adults: Problems of interpretation and implications for theory. In U. Frith (Ed.), Autism and Asperger Syndrome (pp. 207-242). Cambridge, UK: Cambridge University Press.
Holland, N. N. (1999). Cognitive linguistics. International Journal of Psychoanalysis, 80, 357-363.
Johnson, C. (1999). Metaphor vs. conflation in the acquisition of polysemy: The case of see. In M. Hiraga, C. Sinha, & S. Wilcox (Eds.), Cultural, psychological and typological issues in cognitive linguistics: Selected papers of the Bi-annual ICLA Meeting in Albuquerque, July 1995 (pp. 469-480). Amsterdam: Benjamins.
Johnson, M. (1987). The body in the mind. Chicago: University of Chicago Press.
Lakoff, G. (1987). Women, fire, and dangerous things: What categories reveal about the mind. Chicago: University of Chicago Press.
Lakoff, G. (1993). How metaphor structures dreams: The theory of conceptual metaphor applied to dream analysis. Dreaming, 3: 77-98.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press.
Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh: The embodied mind and its challenge to Western thought. New York: Basic Books.
Lakoff, G., & Turner, M. (1989). More than cool reason: A field guide to poetic metaphor. Chicago: University of Chicago Press.
Lesser, H., & Drouin, C. (1975). Training in the use of double-function terms. Journal of Psycholinguistic Research, 4: 285-302.
Matte Blanco, I. (1998). The unconscious as infinite sets: An essay in bi-logic (Rev. ed.). London: Karnac Books.
Maurer, D. (1993). Neonatal synesthesia: Implications for the processing of speech and faces. In B. Boysson-Bardies, F. de Schonen , P. Jusczyk , P. McNeilage, & J. Morton (Eds.), Developmental neurocognition: Speech and face processing in the first year of life (pp. 109-124). Dordrecht: Kluwer.
Maurer, D., & Maurer, C. (1988). The world of the newborn. New York: Basic Books.
Melnick, B. A. (1997). Metaphor and the theory of libidinal development. International Journal of Psychoanalysis, 78, 997-1016.
Melnick, B. A. (2000). Cold hard world\Warm soft mommy: The unconscious logic of metaphor. Annual of Psychoanalysis, 28, 225-244.
Meltzoff, A. N., & Moore, M. K. (1993). Why faces are special to infants: On connecting the attraction of faces and infants' ability for imitation and cross-modal processing. In B. Boysson-Bardies, F. de Schonen , P. Jusczyk , P. McNeilage, & J. Morton (Eds.), Developmental neurocognition: Speech and face processing in the first year of life (pp. 211-225). Dordrecht: Kluwer.
Modell, A. H. (1990). Other times, other realities. Cambridge, MA: Harvard University Press.
Modell, A. H. (1997). Reflections on metaphor and affects. Annual of Psychoanalysis, 25: 219-233. Republished in PSYART (2001). <http://www.clas.ufl.edu/ipsa/journal/2001/model01.htm>.
Modell, A. H. (2000a). Are mental functions hierarchical? Annual of Psychanalysis, 28, 127-135.
Modell, A. H. (2000b). The transformation of past experiences. Annual of Psychoanalysis, 28, 137-149.
Modell, A. H. (2003). Imagination and the meaningful brain. Cambridge, MA: MIT Press.
Özçaliskan, Ş. On learning to draw the distinction between physical and metaphorical motion: Is metaphor an early emerging cognitive and linguistic capacity? Journal of Child Language, 32, 291-318.
Rakova, M. (2002). The philosophy of embodied realism: A high price to pay? Cognitive Linguistics, 13, 215-244.
Rakova, M. (2003). The extent of the literal: Metaphor, polysemy and theories of concepts. Houndmills, UK: Palgrave-Macmillan.
Rayner, E. (1995). Unconscious logic. London: Routledge.
Reddy, M. J. (1979). The conduit metaphor: A case of frame conflict in our language about language. In A. Ortony (Ed.), Metaphor and thought (2d ed.), (pp. 164-201). Cambridge, UK: Cambridge University Press.
Reynolds, R. E., & Ortony, A. (1980). Some issues in the measurement of children’s comprehension of metaphorical language. Child Development, 51, 1110-1119.
Rosenbaum, B., & Garfield, D. (1996). Containers, mental space and psychodynamics. British Journal of Medical Psychology, 69: 281-297. Republished in PSYART (2001). http://www.clas.ufl.edu/ipsa/journal/2001/rosenb01.htm
Stern, D. N. (1985). The interpersonal world of the infant: A view from psychoanalysis and developmental psychology. New York: Basic Books.
Tolaas, J. (1991). Notes on the origin of some spatialization metaphors. Metaphor and Symbolic Activity, 6: 203-218.
Turner, M. (1994). Design for a theory of meaning. In W. F. Overton & D. S. Palermo (Eds.), The nature and ontogenesis of meaning (pp. 91-107). Hillsdale, NJ: Erlbaum.
Vosniadou, S. (1987). Children and metaphors. Child Development, 58: 870-885.
Vosniadou, S., Ortony, A., Reynolds, R. E., & Wilson, P. T. (1984). Sources of difficulty in the young child’s understanding of metaphorical language. Child Development, 55, 1588-1606.
Watt, D. F. (2004). [Review of The blank slate: The modern denial of human nature]. Neuro-Psychoanalysis, 6: 116-118.
Winner, E. (1988). The point of words. Cambridge, MA: Harvard University Press.
Winner, E., Rosenstiel, K., & Gardner, H. (1976). The development of metaphoric understanding. Developmental Psychology, 12, 289-297.

Footnotes