PART II   /  A Psychology of I

4  /  Symbols

Freud dreamt about spinach, Shaw wrote about spiders and lionesses, and Iiro talked about ducks and shoes. Symbols all. Objects all. Shoes and spiders and spinach are the stuff of reality but also functions within the symbolic worlds that Iiro and Shaw and Freud inhabited. As fish know the world as sea, we humans swim in of symbols that we both live in and wonder about. They are us and we are them, yet we make and unmake them. How?

If I trace shoes in the all-too-convenient "Index of Symbols" that the final volume of the Standard Edition of Freud's works provides, I will read: "Shoes and slippers are female genitals" (1916-17, 15:158). "A spider in dreams is a symbol of the mother, but of the phallic mother, of whom we are afraid; so that the fear of spiders expresses dread of mother-incest and horror of the female genitals" (1933a, 22:24). Can Iiro's duck be considered the same as the vulture that nibbled at Prometheus's liver ("a penis--a meaning which is not strange to it in other connections, as we know from legends, dreams, linguistic usage and plastic representations in ancient times" [1932a, 22:189])? Or is that duck the phoenix ("which probably bore the significance of a penis revived after its collapse" [1932a, 22:190-1])? If so, then it too must symbolize a penis. Lions, on the two occasions Freud treated them as symbols, stood for big men or fathers (1900a, 5:462, 1918b, 17:39, 112). Presumably, then, Shaw's lioness is a mother.

Decoding symbols this way can become a bit bizarre. For example, when I used to walk down to the corner store to get my New York Times, I would pass lampposts (phalluses), cats (pubic hair), trees (phalluses), trash cans and cartons (uteruses), overcoats (genitals--ambiguous gender), hats (male genitals), and two churches (female genitals). Why wasn't that walk more interesting then? Why didn't I feel aroused when I passed the elaborate locks on the back door of the drugstore (female genitals)? Why didn't I feel anxiety when I passed the tailor's shop (castration)?

I have to ask (as indeed Freud imagined me asking), "Do I really live in the thick of sexual symbols?" "Are all the objects around me, all the clothes I put on, all the things I pick up, all of them sexual symbols and nothing else?" (1916-17, 15:158.) Freud answered his question only to the extent of demonstrating these and other symbolic equations from a great variety of[85] sources, fairy tales, myths, jokes, slang, and folklore of all kinds. He could present a great deal of evidence, and he did.

Further, these symbolic equations can make a lot of sense. We know Shaw was much preoccupied with his mother. He was likely to translate any female animal into his concern. It seems equally likely that Iiro would translate his feelings of inadequacy about his hands and feet into a feeling that he was inadequate as a male child with a penis: hence his first association to a duck and later associations to eels and a sword.

I don't think these symbols are "wrong" so much as insufficiently spelled out. Because of their one-to-one quality--duck equals penis--they are substitution theories: the duck replaces the penis. It makes more sense to me to think of the symbol as a dual unit, involving both symbol and symbolized. The duck may represent a penis, but it is a duck (Noy, 1973). In general, one-to-one symbolic substitutions do not allow room for any complexity of the dream, joke, novel, symptom, or squiggle, the symbolic structure itself, or its relation to its human symbolizer (Rodrigué, 1956). It is all right for a duck to have webbed toes. Might not that be the reason Iiro imaged himself as a duck? Merely to say shoe = vagina would leave out the complex feelings Iiro must have had toward shoes, because he had to wear orthopedic shoes. Could he hide his webbed toes in ordinary shoes? Could he have imagined that if he gave up his masculinity for femininity (became a shoe instead of a duck) his hands and feet would be cured? Similarly, in Shaw's phrase to the lover, "Your head is in the lioness' mouth," the lioness, be she mother or not, is primarily an eater. Freud's table d'hôte dinner might well symbolize a debt--Freud was certainly preoccupied with moral and monetary debts--but it could also represent gratification or nurture.

Ernest Jones set out the classical theory of psychoanalytic symbolism in 1916. In the first-phase manner (see p. 331), he proceeded by assuming an opposition between conscious and unconscious that led to other sharp boundaries.

First, he distinguished psychoanalytic symbolism from all other kinds. "True" symbols always represent repressed unconscious themes, as opposed to the flags, traffic lights, or trade marks we call symbols in the ordinary sense.

Second, he said, true symbols always have a constant meaning. He was following Freud:

In this way [interpreting symbols] we obtain constant translations for a number of dream-elements--just as popular 'dream-books' provide them for everything that appears in dreams. You will not have forgotten, of course, that when we use our associative technique constant replacements of dream-elements never come to light (1916-17, 15:150).[86]
Thus Jones and Freud find these symbolic equations quite opposite to free association, the usual evidence of psychic connection in psychoanalysis.

Third, said Jones, all symbols express themes about the bodily self, immediate blood relatives, or the phenomena of birth, love, and death. People often translate that crudely: "Freudian" symbols are "sexual."

Fourth, symbols translate thoughts that were previously in existence in another form but repressed. Hence, interpreting symbols always points us toward the past and usually the very archaic past of childhood fantasies and beliefs. Jones calls them "the most primitive ideas and interests imaginable."

In interpreting symbols Jones observes, as Freud had before him, that symbols often follow linguistic connections through etymology or slang. Similarly, symbols have parallels in myth, legend, folklore, fiction, and poetry, although (as Freud pointed out) not all these cultural symbols appear in dreams nor do all dream symbols appear in cultural settings. Clearly, then, one can compare "Freudian symbols" to a language (as in the various dictionaries of dream symbols), and Freud did--

the dreamer has a symbolic mode of expression at his disposal which he does not know in waking life and does not recognize. This is as extraordinary as if you were to discover that your housemaid understood Sanskrit, though you know that she was born in a Bohemian village and never learnt it (15:165).

Before Freud recognized unconscious symbolism, he had spoken of unconscious activity only in terms of mental actions: the constant effort of repression, the conversion of feelings into body symptoms, the dream "work."

Now, however, it is a question of more than this, of unconscious pieces of knowledge, of connections of thought, of comparisons between different objects which result in its being possible for one of them to be regularly put in place of the other. These comparisons are not freshly made on each occasion; they lie ready to hand and are complete, once and for all (15:165).
Like a language. Or like the ready-made sentences or arithmetical calculations that a dreamer recalls into his dream.

To explain this quasi-language, only part of which dreams use, Freud and Jones resorted to a theory of the Swedish philologist Hans Sperber, that the sounds of speech originally served to call the speaker's mate. Then these same sounds, rhythmically uttered as our Neanderthal ancestors hunted and gathered, came also to denote acts of work. Hence we have many verbal roots (about spearing, say, plowing, or planting seeds) that denote work but that can be used as sexual symbols.[87]

Farfetched as this theory sounds, some contemporary analysts have found a core of truth in it. If I think "Jane," I not only bring my mate to mind, I also remind myself of her absence or, even if she were in my study with me, the irreducible difference between Jane and the me who says "Jane." All language--all naming, anyway--by substituting the word for the thing, establishes both a presence and an absence. One could say, therefore, that in all naming "desire speaks," that is, the naming announces both the desire and the absence of what is desired. Following Jacques Lacan, many French analysts of the last decade have drawn just this analogy between unconscious processes and language.

With Freud and Jones, however, it seems to me that their analogy of dream symbols to language takes them out of the isolation of the psychoanalytic dream symbol that Jones first posited. As Charles Rycroft says, "When Balzac likened clumsy men making love to gorillas playing violins, or when Queen Elizabeth I said 'If I had been born crested not cloven, your Lordships would not treat me so,' they must both have been fully aware of the implications of what they were saying and used images which do in fact frequently appear in dreams in contexts which permit their interpretation as sexual symbols" (1979, p. 75).

In other words, dream symbols are not confined to dreams; we use them in our everyday metaphors, as in Freud's own evidence for his symbolic interpretations. Those who use the interpretation of symbols heavily in literary criticism tend to use them as a fixed, fully demonstrated system. (I wrote that way myself for perhaps a dozen years.)

If, however, you turn back to the sections of The Interpretation of Dreams where Freud first broaches the psychoanalytic idea of symbolism and follow closely the dates of his various interpolations (provided in the Standard Edition), you will find a good deal of hesitation and qualification expressed. Thus there is another way of thinking, derivable from Freud himself, that treats symbols in a far more flexible way.

Freud's whole idea of symbolism came into being after the main ideas of The Interpretation of Dreams, as an adjunct to "typical" dreams, the dreams that many people share of examinations, nakedness, and so on. Symbols were not whole dreams but particular images that many people share. In particular, symbols tended to appear as a substitute for free associations as if the dreamer took them over ready-made (like grammatical and arithmetic constructions in dreams).

Always, however, Freud insisted, even as he added the reading of symbols to The Interpretation of Dreams, that it was a supplement to the primary method of psychoanalysis, free association (5:359-60). Sometimes even a symbol has to be interpreted "in its proper meaning and not symbolically" (5:352). Symbols can have several meanings, and "as with Chinese script"[88] one arrives at the "correct" interpretation from the context rather than through a dream-book decoding. Further, many of the original meanings given for symbols represent "the reckless interpretations of Stekel," who had the lamentable habit of making up his evidence (Jones, 1955, 2:134-37).

Freud puts aside all hesitations and decodes like the veriest dream-book, however, in his fullest and most vigorous statement of the theory of symbolism, Lecture X of the first series of Introductory Lectures (1916-17). Although these lectures precede a number of Freud's major discoveries, they are widely used as an introduction to psychoanalysis for students. Hence psychoanalytic symbols bulk large in most American undergraduates' idea of psychoanalysis (particularly psychoanalysis as applied to literature and the arts or at).

After 1916 the whole trend of psychoanalytic thinking about symbols opens up a less one-to-one, less doctrinaire theory. In particular, Melanie Klein (1930) and later Kleinians like Hanna Segal (1957) wrote important essays expanding the early psychoanalytic view of symbolism. Where Jones had maintained that "symbolism . . . constitutes a barrier to progress," Klein (working from a case history) claimed that symbolism made tolerable, structural, and creative identifications possible. Hence symbolism, by joining conscious and unconscious mental processes, made unconscious processes accessible and available for creativity or therapy. Klein thought that symbol-formation went along with growth within or without therapy. Segal used Klein's approach to explore the whole self-symbol-symbolized relationship.

This exploration took hold among all three of the English psychoanalytic schools. Notably, Marion Milner carried Klein's idea of identification farther, to fusion (1957a). Symbolism, she pointed out, involves a fusion of symbol and symbolized and symbolizer. (In the symbolizer's mind, the shoe is a vagina.) In play and artistic creativity, symbols take on a life of their own: the logic of the drawing or the toy dictates the outcome, instead of the will of the creator. Whatever the content, then, the mode of symbolic fusion recaptures the original fusion in infancy of mouth and breast or infant and mother. We temporarily give up the discriminating "objective" ego and undergo in favor of the symbol-other a temporary loss of self (that Ernst Kris described in the language of ego-psychology as "regression in the service of the ego"--1952). Thus art provides the adult a framed space, bounded time, and pliable medium with which one can at will, again and again, re-create part of the everyday experience of infancy (Milner, 1952, see also 1957b, 1969).

In a famous study of 1963, Werner and Kaplan related the child's learning of language to general psychoanalytic patterns. They too stressed the role of inner dynamics in symbolization as opposed to external similarities between[89] the symbol and its object (very like the linguist Saussure's insistence on the "arbitrariness" of the relation between signifier and signified). In early childhood, symbols gain their significance in an interpersonal context, as part of the relationship of mother and child. Only later do purely linguistic considerations take over.

The sociologists Weinstein and Platt developed a cultural aspect to the psychoanalytic theory of symbols (1973, pp. 69-89). Symbols represent ways that all structures of a personality, even the id, become socialized. The id is a purely wishing agency, yet to wish one must have symbols to wish in, and those symbols will have a social origin. The argument applies with even more force to ego and superego. In the same vein, George Klein had put forward the view that unconscious fantasies (triggered by anxiety, threats to self-esteem, and the like) include directives of approach and avoidance. Unconscious fantasies induce symbols which in turn become the basis for behavior (1970, pp. 397-404).

In France, Lacan carried the whole idea of symbolism much farther and wider into "the symbolic," the symbolic order, the law that underlies--is really--our culture. Clearly, however, this is a much larger concept of symbolism than Freud's or that of most non-French analysts. (A lucid and lapidary account of Lacan's view can be found in Laplanche and Pontalis's dictionary [1968, 1973], s.v. Symbolic [sb.])

Less abstractly theoretical but intensely commonsensical (in the best British tradition) are the writings of Charles Rycroft on symbolism, particularly symbolism in dreams. For example, he points out that Jones's "ideas of the self and the immediate blood relatives or of the phenomena of birth, love and death" need not be thought "the most primitive ideas and interests imaginable." His own phrase, "imagery related to biological destiny," suggests rather that dreams and psychoanalytic symbols deal with the most fundamental and perennial concerns of humankind (1979).

Rycroft interprets a number of familiar symbols, although he says that it is pointless to list standard meanings for, say, dogs or horses in dreams. Rather, we should ask, What imagery does this particular dreamer have available for constructing his dreams? and, What determines the aptness of that imagery to express these particular ideas? In other words, Rycroft restores dream symbols to their proper context, the statement of a dreamer in his personal and cultural setting (1979).

For Rycroft, dreams are simply the way the imagination functions during sleep, and there is no particular reason to assume that symbolism is a device by which dreamers deceive themselves. Rather, dreams are continuous with other forms of imaginative thought. To be sure, the statements made by dreams tend to be of a kind that the dreamer is reluctant to understand, and[90] this phenomenon amply justifies Freud's division of the human into conscious and unconscious parts (1974, 1975).

His is a considerable enlarging of the classical Freud-Jones view of symbolism. Even so, Freud himself gives hints of this larger view. He accepts Aristotle's view that a dream is thinking that persists into sleep. Although it is true that some waking thought differs sharply from dream thinking, metaphorical, imaginative thought does not. It provides a continuity between our dreaming selves and the most intelligent acts of our waking selves.

Freud himself suggests a very large idea of dreams in the famous last sentences of The Interpretation of Dreams: "By picturing our wishes as fulfilled, dreams are after all leading us into the future. But this future, which the dreamer pictures as the present, has been moulded by his indestructible wish into a perfect likeness of the past." Freud here refers to his basic definition of a wish as an impulse to bring back a perception linked to the situation of an earlier satisfaction. Dreams are wish fulfillments. Hence dreams and dream symbolism do always point us to a future, but one totally defined by the past wish.

When Freud wrote about creative writers and daydreamers, he developed this more open model:

The relation of a phantasy to time is in general very important. We may say that it hovers, as it were, between three . . . moments of time which our ideation involves. Mental work is linked to some current impression, some provoking occasion in the present which has been able to arouse one of the subject's major wishes. From there it harks back to a memory of an earlier experience (usually an infantile one) in which this wish was fulfilled; and it now creates a situation relating to the future which represents a fulfilment of the wish. . . . Thus past, present and future are strung together, as it were, on the thread of the wish that runs through them (1908e, 9:147-48).
Freud here is setting out in a very compressed form a model of mental activity he derived from the study of dreams which we can apply to conscious fantasying and imaginative activity in general. It has three phases that we could picture by a sort of large square-root sign.
Square root sign
[91] We have a conscious thought or perception. It sinks down, as it were, into "the unconscious," where it attracts to itself unconscious fantasies, repressed or from early childhood. We then transform that thought or perception (loaded now with some version of those fantasies) into conscious, sharable ideas.

Freud gives an analogy from business:

A daytime thought may very well play the part of entrepreneur for a dream; but the entrepreneur, who, as people say, has the idea and the initiative to carry it out, can do nothing without capital; he needs a capitalist who can afford the outlay, and the capitalist who provides the psychical outlay for the dream is invariably and indisputably, whatever may be the thoughts of the previous day, a wish from the unconscious (1900a, 5:561).

As settings change, different phases of this process weigh more and less in the whole. Presumably, in imaginative planning or daydreaming of the future, the future aspect will be stronger than in dreams. In problem-solving or reality-oriented thinking, the effect of childhood or unconscious wishes will be less visible. Different dreamings will draw more and less on what in the present we consciously know.

In bringing out the similarities between daydreaming, night dreaming, and imaginative, metaphorical thinking, Freud establishes a spectrum of related modes of thought instead of a sharp boundary (as Jones seemed to posit). What Freud has set out is a paradigm for what the preverbal, unconscious past does whenever we think. The process of symbolization takes place in all our thought.

Nor need we assume that symbolic thinking always points backward in time. As Paul Ricoeur says,

Are not dreams a compromise fluctuating between these two functions [regression and progression], according as the neurotic aspect inclines dreams toward repetition and archaism, or as they themselves are on the way to a therapeutic action exercised by the self upon itself? Inversely, are there any great symbols created by art or literature that are not rooted in the archaism of the conflicts and dramas of our individual or collective childhood? The most innovative figures that the artist, writer, or thinker can produce call forth ancient energies originally invested in archaic figures; but in activating these figures . . . the creator reveals man's most open and fundamental possibilities and erects them into new symbols of the suffering of self-consciousness (1970, p. 322).[92]
Indeed Freud himself in those last sentences of The Interpretation of Dreams shows how "dreams are after all leading us into the future," albeit a future which is "a perfect likeness of the past."

Jacques Lacan has expanded the early Freud-Jones concept of symbolism another way, into le nom du père, a rich phrase that leads outward equally to the "name," the "noun," and the "no" of the father. In the name or noun of the "father," Lacan intends, not your father or any particular father in a literal sense, but a more metaphorical father, a "function" who embodies the whole symbolic order we inherit from our forebears as we are born into culture, notably language. In this as in other concepts, Lacan, as Murray Schwartz puts it, is drawing on first-phase psychoanalysis for metaphors (fathers, phalluses), with which to express the third phase of psychoanalysis, the psychology of the self (see pp. 331-33), one reason for the notorious difficulty of his writing. Obviously, in a literal sense, mothers embody the symbolic order, and maybe more so.

That order constitutes a storehouse or a treasury from which we draw our symbolic vocabulary. It is like a "promptuary," a type of book in the sixteenth century that stored quotations and other structured verbal information from which one could copy to construct one's own book (like a modern clipping file).

A symbol or "sign" in Lacan's sense (which here goes well beyond Freud's sense of "symbol") includes a signifier, the sound of a word or the written representation of the sound, and the "signified," its meaning, understood with all we mean by meaning in the word "meaningful."

According to some linguists, notably Jakobson, this symbolic order is structured by a series of negatives: a symbol is what it is because it is not this, that, or the other. "Duck" does not have a long vowel, is not inanimate, not human, and so on, following certain linguistic pairings, some peculiar to English, others characteristic of all languages, that mark "duck" off from "shoe." (Twenty such pairings would suffice to distinguish all the words in English--the reason we can play Twenty Questions.) Hence the symbolic order is defined by the not or "no" of the father, the negatives of language and culture, but also the prohibitions, the Thou-shalt-nots, that Freud allocated to the father as disciplinarian. Again, "father" is metaphorical here. In the nonmetaphorical world, both fathers and mothers say no. Yet how very French is Lacan's view--to link authority and language and call them the central qualities of the human experience.

According to Lacan, as each of us grows toward the adult world, we acquire our own version of le nom du père. For example, in Lacan's interpretation of the Rat Man, he formulates "the neurotic's individual myth," "the original constellation that presided over the birth of the subject, over his[93] destiny, and . . . his prehistory, specifically the fundamental family relationships which structured his parents' union." For the Rat Man, this myth was: "the conflict rich woman/poor woman." Lacan says that all the confusion about whether Lanzer was to pay one lieutenant or the other or the woman at the post office constitutes "a scenario . . . a schema which, complementary at some points and supplementary in others, parallel in one way and inverted in another, is the equivalent of the original situation." In other words, Lacan uses something like an identity theme for the Rat Man, but he limits it to "the neurotic" and to the particular symptoms that brought Ernst Lanzer to Freud (Lacan, 1953, pp. 410, 415, 413, 415; Evans, 1979).

It seems to me that he thereby both overstates and understates the case. He understates it in that he limits his formula to the neurotic part of Lanzer, as if we could separate Ernst Lanzer into healthy and sick parts, like a bin of apples. Lacan does not seek a myth that would cover all of Lanzer's life, a style that permeates all of his particular symbolic network. He overstates the case in that he sees Lanzer's life determined by his father's debt and marriage in a manner almost astrological.

Possibly I am being American in response to Lacan's intense Frenchness, but I think even someone as upset as Ernst Lanzer has more autonomy than such a "scenario" suggests. It seems to me that each of us enters the symbolic order in our own way. We will use its symbols, its chains of signifiers and signifieds, for our own purposes, although (as Lacan correctly insists) this symbolic order exists before we arrive on the scene and after we are gone. Like a dictionary, it is open to all equally and equally binding on all. Yet, like a dictionary, we each use it differently. We may use the symbolic order very effectively or very badly, but it is we who do so by our own choices, not our parents' choices and not some pre-existing code.

If I look up "spider" in the "Index of Symbols" in the Standard Edition of Freud, I find, "the fear of spiders expresses dread of mother-incest and horror of the female genitals" (1933a, 22:24). Thus I can see the appropriateness of the metaphor in this passage about courtship: "It is assumed that the woman must wait, motionless, until she is wooed. Nay, she often does wait motionless. That is how the spider waits for the fly. But the spider spins her web. And if the fly . . . shows a strength that promises to extricate him, how swiftly does she abandon her pretence of passiveness, and openly fling coil after coil about him until he is secured forever."

The author is Shaw, writing in the preface to Man and Superman. Yet, in the play itself, when his hero realizes just exactly that, that he is being entrapped, he says, "Then--I am the bee, the spider, the marked down victim, the destined prey." He uses the symbol in a very Shavian way, opposite to its usual meaning. The spider is not the predator, but[94] the prey, although in the same context of a woman's body swallowing a man's.

In using the symbol first with its usual coding and then with the opposite, Shaw demonstrates that we have considerable power over the cultural and symbolic order we inherit. We do not simply acquire a code to which we can only add variations. We can turn the code inside out. Shaw can have one of his heroes call the heroine a "boa constrictor" even though a snake is supposedly a penis. Therefore we need a more subtle way of thinking out our relations to our cultural aquarium. Just knowing how the muscles of our legs work does not tell us how we use our legs to get where we want to be. Identity gives us a way to inquire into this larger sense of symbolism as a dialectic between past and future, regression and progression, conscious and unconscious, signifier and signified, self and other.

Freud's dream of the table d'hôte we can read as some function of his identity. When Shaw wrote in opposites, we can say that he did so as some function of his identity. What is that function? We can best explore it in a more circumscribed situation than the lives of such great symbolizers as Shaw or Freud. We can turn to

Sam and Sandra

were two students of literature who were reading short stories for me as part of a study of literary response (and whom I have described before [1975a]. I do not know a better tailored sample of the ways we adapt the symbols the world provides us to our individual needs.) Among the stories they read was Faulkner's "A Rose for Emily." The narrator, speaking for the townspeople, describes two of the main characters this way:
We had long thought of them as a tableau, Miss Emily a slender figure in white in the background, her father a spraddled silhouette in the foreground, his back to her and clutching a horsewhip, the two of them framed by the back-flung front door.
The passage is straightforward enough, and it has some "Freudian" symbols like the phallic horsewhip or the vaginal door and uterine house. As Sam and Sandra talked about the passage, however, I could see they had transformed it (by some such process as that diagrammed on p. 91) through their own unconscious desires.

Good-natured, easygoing, dapper Sam singled the tableau out as virtually the first thing he wanted to talk about in the story: "The father was very domineering. One of the most striking [sic] images in the book is that of the townsfolk looking through the door as her father stands there with a horsewhip in his hands, feet spread apart and between or through him you see a[95] picture of Emily standing in the background, and that pretty much sums up exactly the kind of relationship they had." Sam was stressing the father's dominance and, in doing so, was positioning the townspeople so that they could see Emily between her father's legs. Emily became a mere appendage of her father, the very appendage that made him masculine.

Yet this picture was part of what Sam found highly romantic about Emily. "The frailty and femininity that that evokes!" he sighed. "Just that one frail, 'slender figure in white,' . . . " Yet, almost at the same moment he was imagining this helpless Emily, he could say, "The word 'tableau' is important. While they [the townspeople] may be envious and while they may be angry at the way that these people act, they yet need it, it seems, they in a way like to have it, much as one is terrified at the power of a god and yet needing him so much, and, you know, sidling up to him and paying homage to him and in the same way I think Emily comes to function as this god symbol." A curious turnabout from frailty and femininity to a "him" of godlike power.

Sandra liked this story intensely, had read it several times, and had even, in her freshman year, written a term paper on it. Yet she recalled the tableau very hesitantly: "They said they always had this picture of him standing, you know, sitting in the door with a whip in his hand." As for Emily,

I see her as very young and dressed in white and standing up--I guess she's supposed to be standing up--behind her father, who would probably be looking very cross, say, if someone had come to call on her. No doubt, she would have a certain amount-- Possibly fearful, but probably more regretful because she's being, they even say, robbed of something at that point . . . There would be a great amount of strain on her face because of her inability to do anything except just watch.

In other words, in their different readings of the story, their resymbolizings of it, Sam and Sandra actively shape it into a function of their identities. Even when they see and say much the same thing, they give it a personal touch. Sandra's phrasing, "with a whip in his hand," for example, seems less forceful or brutal to me than Sam's "with a horsewhip in his hands."

In interpreting their readings, I have the benefit of knowing their responses to Rorschach, TAT, and COPE tests, and I also know their readings of ten other short stories as well as a great many incidental remarks that enable me to read their identities as themes and variations.

Sam, as I read his personality, hovered between being masculine and feminine, active and passive, strong and weak. He wanted to be helpless so as to come close to and identify with supplies of love and admiration that would confirm him as strongly and safely male. Conversely, he would flee a[96] male power that seemed too threatening. Thus, when he re-created the tableau for himself he composed a perspective in which, by identifying with the townspeople, he saw a frail, feminine Emily positioned under or between her domineering father's legs. At the same time, he thought both Grierson and Emily the possessors of godlike powers that the townsfolk needed to sidle up to. I think Sam is seeing the tableau through a preferred defense, that is, dealing with a source of threatening male power by identifying with it. He gratified this childlike wish to have power given to him by picturing the townspeople sidling up to the lordly Griersons. At the same time, however, he asserted his own safe separateness by putting the door between the viewers and Grierson's horsewhip.

In reading the tableau, Sam stressed the distinction between male and female. He searched out power, particularly male power. "The father was very domineering." The image was "striking." "One is terrified at the power of a god and yet needing him so much and . . . sidling up to him."

I saw Sam's need to confirm physical masculinity in what he said about Miss Emily's father at other points in the interview. "She lived under her father. Her father made the decisions. Her father kept things up. Her father was the leader, the member of the town council [not in Faulkner's text] . . . a very strong male figure in her life, and she was totally dominated by it (sic)." As if Grierson were not strong enough already, Sam elected him to the town council (Faulkner had not). He made him a dominating "it" and a "member," keeping things up. Emily is "under" him, and Sam makes her an emblem of "frailty and femininity." At other points in the interview he called her "the darling lady."

As I interpret Sandra from personality tests and from her remarks about stories and other things, she sought to avoid depriving situations and to find sources of nurture and strength with which she could exchange and fuse. In doing so, she used seeing or not seeing as her primary sensory mode (as opposed to Sam who used closeness and distance). Where Sam spoke of the tableau as an "image," Sandra said, "They said they always had this picture. . . . " She tried to bring Mr. Grierson down to manageable force by changing him from standing to sitting and "looking very cross," surely a dainty way of referring to a father who Faulkner says is "clutching a horsewhip." Emily would be "standing up," hence somewhat strong, rather than "fearful," but "she's being . . . robbed of something." Seeing is the mode of her distress and Sandra's: "There would be a great amount of strain on her face because of her inability to do anything except just watch." That would be the worst thing for someone with Sandra's character structure: to have to watch oneself being robbed of a source of strength or care.

Clearly, I think, Sam and Sandra are working more actively with the story than any notion of a fixed meaning or a symbolic code or a "right" reading[97] leaves room for. For example, Faulkner refers in this story to a Colonel Sartoris--"he who fathered the edict that no Negro woman should appear on the streets without an apron." Le nom du père with a vengeance. When I asked Sam about it, he said:

I hate, I hate to say. I like it. It's terrible. It's the worst thing. I mean, it's not terrible for me to like it. It's terrible that that kind of thing ever once existed. It's inhuman. It treats the Negro as animal. The Negro is not animal. But I can't help but be charmed by the naivete, by the total lack of concern that such a position of power places one in.

Later he admitted to a more sadistic pleasure: "I react with a kind of a smile that says, 'Oh, those were the good old days' type of thing, when I read that with the apron on the street." Interestingly, other readers of the story said similar things, ways (I think) of trying to mute the ugly cruelty. Nevertheless, that whipping or the edict or his wishes to be (in favorite phrases of his) "on the top," "in that kind of position," "such a position of power," all let Sam get a taste of the pleasures of masculine dominance in its crueler forms.

Sandra took a different tack. She reduced the edict. "I smiled a little bit, sardonically, I'm sure, because it's a great little touch of ironic humor. I think Faulkner meant it that way, and I think [of] the voice in the story as meaning it that way. Proclaiming it. 'He who fathered the edict.' Using these heroic terms to describe such a petty and obvious extension of bigotry." Having built up the narrator in other remarks as a wise and supportive figure, she said, "I smiled almost at the way he [the narrator] said it because it was such a perfect undercutting of the heroic Colonel Sartoris." (I think "undercutting" someone who "fathered" represents her weakening him specifically as a man.) "If this is--as he singles this out--his most important piece of legislature [sic], that's pretty good!" and Sandra laughed outright. Again she had reduced a threatening source of power and strength (the "heroic" Sartoris) to a safe size.

Sandra was an intensely visual person. I thought of her as constantly looking into an unknown world for a flow of strength or nurture that would equalize older and younger, stronger and weaker, or male and female. If she found such a source she wanted to see it more closely, touch it, even merge with it. If she found instead some extreme, a weakness or an overpowering, she wanted not to see it. If I were to word that theme in a single phrase, it would be: to see and approach more and more closely a source of power and nurture, but not to see its loss.

Sam tended either to flee into an isolated, boyish masculinity or to seek out mutual admiration leading to his passively identifying with a stronger male or asexual figure. If I were to state a theme for him, it would be to take[98] in or to get out so as to be male. That is, he wanted to get out of dangers to his maleness and to take into his body love and admiration.

In many ways, he was like Sandra. Both were much concerned with power, particularly masculine power, and they tended to deal with their fears of being overpowered by physical or visual avoidance. If I were using an earlier psychoanalytic language, a characterology based on fantasy and defense (see pp. 160-64), I would say they were both phallic types with a preferred defense of denial. Indeed their remarks on the story are full of old-fashioned phallic symbols. Yet they responded very differently, so much so that even a very good characterology (like fantasy and defense) does not account for the difference in their readings. Using these identity themes, however, we can articulate different aspects of their responses as functions of their identity.

Consider the way they each read the character Bugs in Hemingway's story "The Battler." One night a young man, bumming around the country, wanders up to a campfire near the railroad tracks. The fire belongs to Ad, a punchdrunk ex-champion fighter, and Bugs, a black hobo, who takes care of Ad on their ramblings, as a squire might take care of a knight. As Bugs is making ham and eggs, Ad becomes increasingly and irrationally angry at the young man, to the point where Bugs, to protect him, has to knock Ad out with a blackjack he keeps for just this situation. While the fighter is out, Bugs gives the young man a sandwich and urges him on his way before the fighter wakes up.

Confronted with this character who cooks eggs but coshes his friend, Sandra concentrated on his giving food. "He's probably the easiest character to like in the story," she said. "Big and gentle and helpful." "He's all gentleness. Even when he's hitting his friend over the head, he's being [as] gentle as he could possibly be." "One of those generous people that it's hard not to like." "You have the feeling that he could handle a whole roomful of people and give each enough attention, make each feel pretty special."

Sam, however, focused on the blackjacking and a tiny detail, that Bugs, to revive the fighter, "pulled his ears gently." Sam said, "Bugs came across as kind of the nightmare element . . . kind of the devil of the nightmare and the excruciating type of oriental torture where [in a Chinese accent] 'Oh, you very nice. I'm now going to rip your ears apart. . . .'" "A kind of Fu Manchu person, kind of polite mingling with merciless from whom you never know what to expect." "This kind of exquisite, terribly polite type of torture which is to me more frightening than brutality." "Where you're kind of helpless before the politeness of the enemy and yet there's something behind the politeness that you can't deal with until it comes to the front, till you see it, and you've got to react only to his front, which is politeness."[99]

In just these remarks I can hear Sam and Sandra positioning themselves and Bugs along two of the great axes of human experience. Sandra is setting him outside herself--"He"--but she is also taking aspects of Bugs into herself: "You have the feeling," meaning I, you, all of us. Similarly, in Sam's "Bugs came across as . . . ", I sense he is distancing Bugs by looking only at "his front" but also becoming Bugs by imitating a Chinese accent. Both Sam and Sandra are setting themselves vis-à-vis Bugs on the axis between self and other.

Sandra speaks of her liking Bugs "in the story," in its immediacy and sequence. In thinking of Bugs as "one of those generous people," however, she sets him in a more enduring framework of her beliefs and experiences. Sam, too, dealt with the immediate Bugs, the ear puller, but also put him in a more timeless setting, "this kind of . . . polite type of tor- ture," referring to fictions and movies about orientals. He, like Sandra, was placing Bugs between the immediate and the enduring.

These two continua, between inside and outside, between the timely and the timeless, provide angles for looking at the ways Sam and Sandra differently yet similarly symbolized Bugs. One can, I think, convert these axes of time and space, however, into more psychological modes of exploring Sam's and Sandra's different relations to Bugs.


First, Sam and Sandra brought to the Hemingway story certain expectations: about short stories, about Hemingway, about me, for whom they were reading this story, and about the kinds of thrills and satisfactions the story might yield. We all bring expectations to bear, most obviously from our cultural codes. Less obviously, we will seek--and find--in a literary work the kind of thing we as individuals characteristically wish and fear the most. Sam and Sandra's expectations about stories and Hemingway seem to have been more or less the same. Their expectations about another person, though, about Bugs, differed sharply. Sandra looked for something in terms of a source of nurture and power--and found it. Sam looked for something that would relate to his own masculinity--and he found that.

To be somewhat more precise, I can see them each approaching this new piece of the world, the story, expecting they could deal with it as they customarily dealt with other pieces of the world and other stories. To respond positively to something, a short story, for example, we need to be able to re-create from it our characteristic strategies for coping with reality, for achieving the pleasures we want from the world and defeating the dangers we fear.[100]

Here Sandra hoped for a source of nurture or power and found it in Bugs, the cooker of eggs. Her characteristic pleasure was to draw close to such a source, provided it was not too strong, and she was able to do that with this story. She measured Bugs and found him "gentle," "helpful," "hospitable," and "generous." She found him neither weak (as some critics have) nor overpowering as Sam did, but a balanced strength. "If anybody's really in command, it's Bugs, the one that can hit him on the back of the head and wake him up. He's really in control," but nevertheless, "[as] gentle as he could possibly be." In effect, she was able to find in Bugs her characteristic mode of defense or adaptation: matching and balancing a source of nurture and power to her own strength.

Once she had matched her defenses, she could infuse the story with a pleasurable fantasy. She could imagine a sort of headwaiter Bugs who could "handle a whole roomful of people and . . . make each feel pretty special," an emblem of "true hospitality . . . that few people really have." "A lot of people can bring you over and throw a lot of food at you. [But Bugs's hospitality is] very different from just entertaining people, and it carries over, I think, to all kinds of different relationships." In effect, having matched her expectations in the character, having re-created her defenses through him, she was able to project a fantasy of hospitality into Bugs.

Sam had a rather different experience. He came to the story expecting that Bugs would have something to do with his, Sam's, masculinity. Sam perceived that something as a threat. Sam thus ignored Bugs the provider of food and concentrated on Bugs the blackjacker, whom he found a torturer, "merciless," "frightening," and before whom he felt "helpless." (In other contexts, masculinity meant for Sam being active, erect, and on top of things.) Having perceived this threat, Sam tried to deal with it by his characteristic defenses. He got himself out by distancing Bugs into someone unreal, a "nightmare element," "the devil of the nightmare," not an American black, but a faraway and fictional "Fu Manchu." He seems to have needed more defense, however, and he began to take Bugs into himself by introducing the social front. (Incidentally, one of Sam's own ways of getting people to admire him was through manners and dress.) "Politeness," he said. "You've got to react only to his front which is politeness," but "there's something behind the politeness that you can't deal with. . . ." I hear Sam's defenses saying, "Take in part of Bugs and get yourself away from the rest that you can't manage", but his defense didn't quite work. He was left feeling, as he said "You never know what to expect." Finally, he took into himself both Fu Manchu and a victim and (complete with Chinese accent) imagined himself tearing the victim's ears off.

By contrast, Sandra had matched her defenses and adaptations. She had gotten Bugs to be just the right amount of power and nurture. Then she[101] could enlarge on the wishes or fantasy she brought to the story as a whole. "After the feeling of loneliness . . . that coming up to a light is sort of a feeling of, well, promise, perhaps, relief. But it's a strange feeling, too, because there's always the element of whatever is unknown, is-- You know, what kind of person is it? Is it somebody who's going to be happy to offer you some warmth with maybe some food? Or is it somebody . . . who would have turned on him right away?" Happily for Sandra, it was someone with food: "After you'd had the earlier feeling of loneliness, physical hurt, he [the young man] began to build up a sense of hunger. And it was really, really well described, like how he was watching him cook something so good, and just the way everything was described as he put together the sandwich and just how good everything tasted!" Sandra had found the source of balanced power and nurture she sought, and she was gratifying her wishes about that source.

Expectations being met (or not), defenses being matched (or not), fantasy being pleasurably enjoyed (or painfully avoided), a reader can bring a fourth principle into play. Sandra could "make sense" of the story. She used the defenses she had matched from the work to transform the fantasy she had projected into it toward some esthetic, ethical, intellectual, or social coherence.

Sandra said of the story as a whole: "Something like this in its very most isolated sense probably calls on any time that you've ever been . . . bullied or picked on when . . . the other person ran away or was out of range . . . for some reason, that, say the person was an authority. . . . Any time when you were put down and didn't have a chance to stand up and fight back on your own terms either verbally or physically." At an intellectual level, now, Sandra re-created her identity just as any of us makes a unity of a literary text, compares it to other works, associates to it, brings knowledge or expetise to bear, evaluates it, places it in a tradition, decodes it--in short, commits one or all of our respectable and meaningful. At an intellectual level, now, characteristic strategies for tranforming crude daydreams into something respectable and meaningful.

Sam was more explicitly intellectual: the story "shows [the young man] Nick moving away from home and moving out into the big, wide world." "It shows what Nick has to accomplish, the symbolism of staying on the railroad track with the swamp to both sides and the railroad track, clear and well-paved, which made easy walking. . . . Nick has to stay on it to reach the goal. . . . In the story he digresses and picks up a bit of education of sorts." Even in abstract statements of theme, I can hear in their wordings these readers' identities. Sam speaks of a young male's moves toward and away from threats. Sandra contrasts being overpowered with an equal exchange of strengths.[102]

A classic psychological study at the height of the Cold War showed how a group of ordinary men in Boston formed their opinions of Russia. "In each man," the authors wrote, "we find an effort to make the world congruent with or supportive of his way of life--within the limits imposed by the requirements of minimizing surprise, for contact with reality must remain the pied à terre" (Smith, Bruner & White, 1956). Each man symbolized Russia so that it would fit the economy of his particular personality. In effect, we are seeing Sam and Sandra do the same with these short stories (and there is less need to make stories fit reality than Russia). Sam and Sandra have resymbolized Faulkner's and Hemingway's stories so as to re-create their own identities.

In much recent European writing about texts (I am thinking of the philosopher Jacques Derrida and the semiotician Umberto Eco), it is customary to speak of the text as the active one in the combination of text and person, as though Russia alone shaped those Bostonians' opinions of Russia. This story, for example, might be said to "valorize," by the darkness of the scene, the strangeness of Ad and Bugs, and the danger of Ad, a general sense of education through "otherness." Yet one could equally say the story "subverts" this otherness through various hints that Nick's pose of toughness is simply a boyish version of Ad's fierceness.

I find this a jazzy set of metaphors with which to shorthand the way persons create texts and texts create persons. As for me, however, and this is very much a part of my identity, I feel uncomfortable with figures of speech that mask a human activity. I want to know what the people in any given transaction are doing. Hence, where others, seeing persons and texts mutually creating each other, talk about the text, I prefer to speak about the person. I also think, in all candor, that a model that makes the person active allows us to understand reading and listening and speaking more fully than a model that makes the text active. Hence I insist on the evidence of Sam and Sandra.

Sam and Sandra show how we can use these four aspects of their identities to explore the space between them and Hemingway's story: their characteristic expectations, defenses, fantasies, and transformations. If we shuffle those terms a bit we can get a convenient acronym: defense, expectation, fantasy, transformation, d--e--f--t, DEFT. DEFT is a way of exploring what people do in Winnicott's "potential space" or in the philosophers' "intersubjectivity." We can ask how Sam and Sandra DEFTed Hemingway's story.

Further, this DEFTing corresponds to two continua inherent in all human experience, the situating of an event between self and other and between the now and the beyond-now. One can think of Sam's "expectation" as a way of asking, How does Sam fit what he is reading into the immediate before-and-[103]after sequences of his experience, his now? Conversely, "transformation" enables me to ask, How is Sam fitting this story into themes that are meaningful beyond the immediate here and now or before and after? With "expectation" and "transformation," then, we are asking about the relation between Sam's experience within a time sequence and his efforts to transcend time. In the same way, "fantasy" and "defense" come into relation along the axis between self and other. "Fantasy" asks, What does Sandra project out into the world? "Defense" asks, What does Sandra admit in from the world? DEFT gives us a way of setting "intersubjectivity" and "potential space" themselves in time and space.

In a famous essay Lionel Trilling showed that Freud had made poetry indigenous to the human mind. "The devices of art--the most extreme devices of poetry, for example--are not particular to the mind of the artist but are characteristic of mind itself" (1950, p. 177). We could, I am suggesting by this theory of DEFTing, turn Trilling's insight around. The activities of mind are themselves artistic. The diagram on page 91, which began as a way of describing daydreaming, night dreaming, or creativity, really describes all human thinking. We are all engaged all the time in that process of regression from the present into the past and projection into the future, expectation, defense, fantasy, transformation, and expectation again.

When Jones and Freud limited symbols to "the" unconscious, they presupposed a conscious mind free of symbolism that could see right through the symbolic disguise. As Ricoeur, Barthes, Todorov, and other French thinkers have pointed out, this is to assume a "cognition degree zero," a mode of thought free of symbolism, from which one interprets symbols. Freud's larger view and the later analysts suggest the opposite: that symbolism is both conscious and unconscious, at all times a part of our relation to the world around us. Sam and Sandra demonstrate that, at least for the symbolic environment of short stories.

Symbolism, both personal and cultural, is the means by which we establish relations to an other. That is, in symbolizing, we fuse symbol and symbolized and thus selves and other, just as Sam and Sandra each in their own ways became Bugs precisely in order to recognize the otherness of Bugs, the essential Bugs-ness which was not Sam-ness and not Sandra-ness. Symbols are relations, and relations are symbolic. Through symbols and symbolization we can create a dialectic between inner and outer, an interplay of union and disunion.

In the act of symbolizing (the act of DEFTing, I would say, that being a more detailed term), we undo boundaries in order to discover them and in discovering boundaries we create them again. In imagining Bugs as a threat, Sam found a fantasy that worked for him, an image of an oriental torturer, and this idea set off a new set of already known possibilities and associations[104] (to, perhaps, Ming the Merciless of the Flash Gordon comic strips, "mingling with merciless"). In a way, the symbol had, on its own, answered back to Sam's discovering.

Living itself consists in this continued growth of a sense of duality, the me-ness of me and the other-ness of the other. Yet this two-ness itself grows out of recurring states of one-ness. To be two, we need to have felt at one with another, or at the very least, that that was a possibility.

We need to have felt one-ness in order to make the world a part of ourselves--one with us. We sift the heterogeneous world of symbolic experience around us, editing it, so as to create an inner world, more homogeneous, more congenial (con, according to, the genius, spirit), a world we can live in and act on. We sift the world of symbolic experiences so as to add to ourselves realities that are congenial, and in the act of sifting, our genius, our identity, is itself both challenged by, changed by and able to accept the new. Thus DEFTing can model the very act of human growth or adaptation.

The psychoanalyst Joseph Smith puts the idea elegantly:

Every naming, every instance of differentiating an aspect of the world, also names and differentiates the namer. For the most part the latter knowledge is tacitly organized while conscious attention remains focused on the objects of need, interest, or danger out there in the world. The fact that rupture of a prior unity has occurred is but dimly felt. The fact that a world and a self are being mutually constituted only occasionally flashes forth--as when, in recognizing water, Helen Keller recognized, became in some new way, Helen Keller (1978, p. xxvii).

Just as fish-ness consists partly of living in water, so a part of our humanness consists of our living in a world of symbols. The sea goes into the making of fish, yet fish also make the sea what it is. So we are created by the culture and codes around us, yet we create what we are created by. The four terms of DEFTing allow us to explore (through our own symbols) the symbols in which we swim and the way in which we create those symbols.

In effect, DEFTing suggests that we live in a psychological version of our basic adaptation as mammals. The body temperature of a fish simply equals the temperature of its water. We, however, have evolved beyond the cold-blooded animals. We create inside our bodies, in our blood, the warm, nourishing sea the fish finds outside. Physiologically, we create an inner environment. Psychologically, we create an inner environment of symbols. Then we live in a delicate tension between our inner and outer worlds of symbols.

We are active human beings precisely because there is a difference between the heterogeneous, constantly changing outer world and the more homogeneous, more constant world within. In acting we accept that differ-[105]ence, for example, the separateness and absence registered by language, but we try to DEFT it away. We try to make absence presence. We necessarily fail, and that is just as well, for without that absence or difference we would cease to act. We live in a feedback or deconstruction which requires difference for us to (try but not quite manage to) render difference into sameness.

Feedback, we shall see in the next chapters, is an important idea for understanding our relation to symbols. The furnace warms the living room, but it is also true that the living room turns off the furnace. To talk about either of these relations alone leaves out part of the transaction, for neither of these statements takes into accountthe thermostat or the person to suit hir particular taste. In the same way, culture creates us and we create culture but neither of these statements alone captures the dynamic relation between the two.

We create a common human and cultural environment of symbols inside ourselves, but we also create inside ourselves our special variation of that world, just as physiologically we create an inner sea in which our lungs and heart can do their work. Our inner seas are generically human, but varied to be specifically our own. To understand how that can be, we need to explore the way we bring symbols from outside into ourselves. We need to perceive perception.[106]

5  /  Perception

"It's a duck's foot," said Iiro to the first "squiggle" Winnicott drew. "This came as a complete surprise to me," wrote Winnicott, "and it was clear immediately that he wished to communicate with me on the subject of his disability," that is, his own webbed fingers and toes. And Iiro went on to use the duck he had perceived to symbolize not only his webbed fingers, but himself as a boy who could swim but could not play horns and flutes.

Consider a four-word dream, one of the shortest Freud reports, although that did not prevent him from mentioning it three times (1900a, 4:232, 1913h, 13:194, 1916-17, 15:94). "One morning at the height of summer, while I was staying at a mountain resort in the Tyrol, I woke up knowing I had had a dream that the Pope was dead."

He could not explain the dream, except that he could remember reading a short time before in a newspaper that the pope was slightly indisposed. During the forenoon, however, Freud's wife Marthe asked him, "Did you hear the dreadful noise the bells made early this morning?" Freud replied that he had not--and that, he realized, was the explanation of his dream. "It had been a reaction on the part of my need for sleep to the noise with which the pious Tyrolese had been trying to wake me. I had taken my revenge on them by drawing the inference which formed the content of the dream, and I had then continued my sleep without paying any more attention to the noise." Evidently he had heard the bells (since he dreamt about them), but his dream that Pius X was dead had enabled him to sleep contentedly on. His need controlled his perception. Indeed, according to Freud's own theories, dreams are the guardians of sleep. If they are successful, they always serve to ward off stimuli that might wake the dreamer up.

Iiro's perception would seem to prove the old adage, "People see what they want to see." Freud's sleeping through the bells adds, "People hear what they want to hear." Such adages have become more than folk psychology, though, because twentieth- century psychologists have proved over and over again that perceptions follow motivation.

Freud, however, had inherited an earlier idea of perception. In his theories, Freud used a nineteenth-century concept that one analyst has nicknamed "immaculate perception" (Schimek, 1975). That is, Freud de-[107]fined eye and ear into a system Pcpt. (perception), a part of the ego. The system Pcpt. delivers a faithful copy of the world to another system Cs. (consciousness), which might then and only then distort the original perception in response to unconscious pressures. "All perceptions which are received from without (sense-perceptions) and from within--what we call sensations and feelings--are Cs. from the start," Freud wrote (1923b, 19:19).

When he was writing clinically, however, he recognized that perception is not such a cut-and-dried affair. He discussed, for example, the way that dreams could control perceptions, as in his own dream that the pope was dead.

Transference, too, hinges on the patient's altered perceptions. Patients attach new editions of old impulses to their analysts, using the physician to represent some earlier person, often a loved or feared parent. For example, one balding analyst told me of a patient who kept speaking of his thin red hair as Samson's long brown tresses. Indeed, by having the patient lie on a couch with the analyst out of sight, psychoanalytic technique deliberately encourages the subversion of the patient's manifest idea of the analyst by latent needs and feelings. The analyst can then use the patient's changed seeing and hearing as a point from which to discover the deeper, less conscious feelings coloring and toning his perceptions.

Freud had known about transference ever since 1882 and his colleague Breuer's horrified discovery that patient Anna O. had fallen in love with him. Freud, however, did not extend the concept to perceptions. Characteristically, I think, he needed the feeling of being firmly grounded in reality. Perhaps that is why, in most of his theoretical statements (as opposed to his clinical ones), he insisted that our eyes and ears faithfully copy the real world into our minds. But not in all.

The projection outwards of internal perceptions is a primitive mechanism, to which, for instance, our sense perceptions are subject, and which therefore normally plays a very large part in determining the form taken by our external world. Under conditions whose nature has not yet been sufficiently established, internal perceptions of emotional and thought processes can be projected outwards in the same way as perceptions; they are thus employed for building up the external world . . . (1912-13, 13:64).
Freud, Murray Schwartz wrote me by way of comment, "clearly recognizes the active, constructive function of projection in creating the external world, which he usually calls 'reality,' but unfortunately, he also ends the last quoted sentence with the words, ' . . . though they should by rights remain part of the internal world.'" Freud, continued Schwartz, "is judging[108] ('should') or denying his own insight rather than building on it. Nevertheless the insight remains."

Similarly, in a remarkable passage in his late, "testamentary" essay, "Analysis Terminable and Interminable," Freud arrived at a powerful psychoanalytic idea about perception:

The psychical apparatus is intolerant of unpleasure; it has to fend it off at all costs, and if the perception of reality entails unpleasure, that perception--that is, the truth--must be sacrificed.
The mechanisms of defence serve the purpose of keeping off dangers. It cannot be disputed that they are successful in this; and it is doubtful whether the ego could do without them altogether during its development. But it is also certain that they may become dangers themselves. It sometimes turns out that the ego has paid too high a price for the services they render it.
These mechanisms are not relinquished after they have assisted the ego during the difficult years of its development. . . . They become regular modes of reaction of his [the individual's] character, which are repeated throughout his life whenever a situation occurs that is similar to the original one. . . . The adult's ego, with its increased strength, continues to defend itself against dangers which no longer exist in reality; indeed, it finds itself compelled to seek out those situations in reality which can serve as an approximate substitute for the original danger, so as to be able to justify, in relation to them, its maintaining its habitual modes of reaction (1937c, pp. 237-238, italics mine).
That last sentence seems to me to coincide with the DEFTings we have seen in the perceptions of Dr. Vincent and the rest. To cope with the world, we need to see the world as the kind of thing we can cope with. As a doctor in China, Vincent sought in medicine a defense of mastery against a dependency that the adult Vincent no longer needed to fear, some childhood relationship from far away and long ago. Similarly, Sam and Sandra shaped Faulkner's story through their characteristic defenses to avoid dangers when stories cannot pose any real danger.

In general, our defenses are part of our character. Our defenses are also therefore part of our perceptual and cognitive style, even though the real dangers that gave rise to those defenses are buried in our childhoods. To keep our defenses functional and our character intact, Freud is saying, we seek substitutes for the original dangers that gave rise to the defenses. That seems to me an astonishingly true description of some of our most painful self-defeating tendencies as human beings.[109]

It also seems to me that Freud is describing a theory of perception that one can easily develop into the one I shall urge in this chapter. Identity (characteristic defenses) governs perception. In doing so, he strikingly anticipates modern

Perceptual Psychology

Specifically he anticipates the "new look in perceptual theory" that emerged in the 1950s. When I looked at the scientific literature on perception, I found that one major group of psychologists has arrived in our century at virtual unanimity around the idea that people perceive constructively. There is considerable disagreement about the particulars of that transaction, but within that school all agree that we see as much with the brain as with the eye. That is, we bring concepts to bear on the inputs to our eyes and ears so they do not simply copy aspects of the real world onto our brains.

Think of the way we hear a language we do not know. The words seem to run together into an indistinguishable stream of syllables. When we listen to a language we do know, however, even if at a crowded cocktail party, we hear it as distinct and separate words. Perhaps the party is so noisy we cannot hear everything, perhaps not even coherent statements. Nevertheless we hear in words, not just sounds. We hear by means of constructs we already had (Bergson, 1896, p. 136).

Another example: if I am in a train station and, out of the corner of my eye, I see the train next to my train start up, I have a momentary and startling feeling that my own train is going backwards. I am interpreting what I see and hear through one of the basic guesses people supply to the perception of motion: the surroundings are at rest and the object surrounded is what is moving--my train instead of the other train that forms the background to my sight. I then correct my reading when my central vision shows that my train is not moving (Wallach, 1959, pp. 310-14).

Although they are very specialized perceptions, the study of so-called optical illusions provides the most telling examples of constructive perceptual theory. Transparent cubes seem to turn inside out spontaneously. Flights of stairs flip from right side up to upside down. A picture of a vase snaps into two profiles facing each other, then snaps back again. After staring at a red spot we look away at a gray wall and see a green spot. Flashing lightbulbs on the movie marquee seem to form moving arrows. Arrowheads at the ends of equal lines make one seem longer.

"When perceptual illusions were introduced as a topic of study in the 19th century," says one experimenter, Paul Kolers, "the prevailing attitude toward them was that they were . . . minor imperfections or errors in the working of man's perceptual apparatus." Hence they were called illusions,[110] but, "The notion of error . . . implied that there is some 'real world' faithfully reported by the senses. Few contemporary investigators take this view. Instead of thinking of illusions as errors in perceiving they regard them as genuine perceptions that do not stand up when their implications are tested" (1964, p. 316).

"It is as though," says Richard Gregory, another perceptual psychologist, "the brain entertains alternative hypotheses of what objects the eye's image may be presenting. When sensory data are inadequate . . . the brain never 'makes up its mind'" (1968, p. 241). Instead, the cubes seem to flip back and forth, colors appear and disappear, vases and profiles alternate. But it is not the objects which change--they are fixed, after all--it is the brain which restlessly tosses back and forth between inconsistent hypotheses about those objects.

You can test the brain's role if you can find the "Rubin figure" (often reproduced in psychology books), a line drawing that you can see as either a vase or two profiles, and your perception will ordinarily flip back and forth between those two interpretations. You can "beat" the illusion, however. Imagine two people pressing their noses up against a vase, and you will "see" just that. Because your brain has supplied a hypothesis that admits both, you can (most people can, anyway) see the figure as both a vase and two profiles.

The brain's hypotheses interact with sensory data in at least two directions, from outside to inside and from top to bottom. That is, we perceive when data from an object pass through our sense organs to our brains. We perceive from outside in. Yet we also perceive as we bring schemata from our brains to bear on those data. In that sense, we perceive from inside out.

At the same time, we have to fit those data into the various levels at which our minds operate. Occasionally, we operate at the level of pure sensation or with continuities, flows, or invariants; in the familiar metaphor, we perceive at a "low" level. One might look up close at a television screen, for example, to see the colored dots. Most of the time, however, our minds operate our processing of sensory data at a "high" level of completed objects or even complicated abstractions or impersonal ideals. We don't see tv as dots but as "can of beer" or "patrol car." Psychologists speak of perception "from the top down" or "hypothesis driven" perception in contrast to perceptions which are "stimulus-driven," which go from bottom to top.

In constructivist theory, the brain is perceiving at any given moment from outside in and inside out and from bottom up and from top down--much as we DEFT along axes between self and other and between the stream of experience and more enduring themes. Then, can we interrelate these directions of inside-outside and bottom-top? I think we can, using one of the great technological discoveries of the twentieth century,[111]


Gregory Bateson suggests (as many others have) that we can model this high-low, inside-outside interaction cybernetically, that is, through feedback systems. In general, "feedback" means a transaction in which someone or something tests some aspect of its environment and modifies itself as a result of what it finds. Bateson gives as his example, chopping down a tree. "Each stroke of the axe is modified or corrected, according to the shape of the cut face of the tree left by the previous stroke." We operate within a whole feedback loop, tree-eyes-brain-muscles- axe-stroke-tree (1972, pp. 317-18).

Herbert Simon, they say, is the only psychologist to win a Nobel Prize (by disguising himself as an economist). Simon uses feedback as his basic axiom: "Given a desired state of affairs, and an existing state of affairs, the task of an adaptive organism is to find the difference between these two states and then to find the correlating process that will erase the difference." With such a model, human living becomes information processing on the inside and on the outside a search through comparisons--"large combinatorial spaces" (1969, pp. 112 and 54).

Think of driving. If I see the highway ahead of me turn right, I turn the steering wheel of my car right. I do so in just such a way that I can continue to see the car's right front fender keep the same distance from the right side of the road that it has been. If I see the right fender get too close to the shoulder, I turn the steering wheel left. Inside my head I am processing information about the present and future position of the car's right front wheel, of the edge of the road, and of the steering wheel. In general, my behavior with the wheel controls my perception of the distance between the right of the car and the edge of the highway.

You could diagram such a feedback loop this way:

Feedback loop
[112] The loop has to have three things. First, a behavorial end that acts on the input--my hands that turn the steering wheel. Second, a comparator, like my eye and brain feeling out the distance between the right wheel and the shoulder. That comparison, my seeing, becomes the perceptual end that the behavioral end controls. Third, and this is the crucial item that is often left out, some standard or reference signal that is "above" and not controlled by that loop--like my need to keep three feet from the edge of the road. This is the "schema" I bring to bear when I drive. This is what determines whether the position of the car "feels right" to me.

Brain physiologists confirm this model and the importance of the standards even if they are "teleological."

Objective study of living organisms shows clearly that they do indeed have targets and take actions that ensure survival.

We now find that every organism contains systems that literally embody set points or reference standards. The control mechanisms operate to ensure that action is directed to maintaining these standards . . . (Young, 1978, p. 17).

The feedback picture also allows for various purposeful or random stimuli that disturb the equilibrium established by the loop of driver, car, and road. A traffic sign says "Right lane ends"--that would be a purposeful stimulus. More randomly, the road might take a turn or the wind blow across the road or a pothole twist the front wheels. The stimuli would be the same for everybody, but the responses will be unique (in speed of reaction, for example, or degree or shape of turn). These stimuli from outside the loop change what goes into the comparison, and I have to reset the car. I steer so that I get the right wheel back to a distance from the right edge of the road that feels right to me. Behavior is the control of perception (Powers, 1973a, see also 1973b, 1978).

Interestingly, that is exactly the way Freud defines a wish in the last, metapsychological chapter of The Interpretation of Dreams, as controlling perception. A wish seeks to re-create the perception of a satisfaction (1900a, 5:565-66). According to the "pleasure principle," all our waking behavior serves to gratify a wish--as opposed to our dreams, where we merely hallucinate our pleasures. Either way, however, a wish equals a wish for a certain perception. For Freud, just as for the computer scientist, behavior controls perception.

This feedback model, for all that it sounds electronic and high tech, is thus profoundly psychoanalytic. All his life, Freud used homeostasis, the organism's acting to bring itself back to a "normal" equilibrium, as a model for psychological processes, and homeostasis is simply a feedback through[113] enzymes and hormones and other biochemical processes. Indeed Freud's 1895 Project for a Scientific Psychology, his first big psychological effort, provides drawings of cross-correcting processes that look very much like feedback diagrams, although, of course, feedback was not yet available to model his ideas.

Jonathan Miller points out that the scientific and engineering discoveries of any given age are not only useful in their own right but valuable for the models they provide us for other processes. The fountains of the Renaissance served as models for the circulation of the blood, as today we use radar and automatic gun-turrets to model the movements of our limbs (1978, pp. 284 and 4-7). There is nothing odd, then, in our using feedback as a framework for the insights of psychoanalysis. Frankly, I think psychoanalysis would have made faster and greater scientific progress had metaphors of electrical amplification and feedback been available to Freud instead of the hydraulic models to which he had to resort.

Further, a feedback loop combining stimulus and response, the individual and the shared, serves as a useful metaphor in other kinds of psychology besides psychoanalysis. In particular, it lets us articulate the long line of research in the psychology of perception, cognition, and memory demonstrating that we see and hear and know and remember actively.

Many, many experiments in perception have led these psychologists to the conclusion that all human beings perceive by checking an event against internal constructs, perhaps memory traces determined partly by previously sensed objects, partly by the personality of the perceiver (Noton & Stark, 1971, p. 219). One perceives not only familiar phenomena such as verticality but optical illusions, skewed rooms, and all the perceptual psychologist's bag of tricks by generating and testing hypotheses about the reality before one (Witkin, 1959; Ittelson & Kilpatrick, 1951). So with propaganda and mass communication: the media are not guns that shoot magic bullets of content into their audience. Rather, senders put out signs, and receivers choose among them, making such use of them as they can and will (Schramm, 1973). Even by the age of nine months, infants have begun to think in this way, by forming hypotheses with which to assimilate the world around them (Kagan, 1972).

Things don't just impinge on us. We impose schemata on things to assimilate them to our minds. We reach out to see, hear, know, or remember, using our innate capacities to see, hear, know, or remember. Hence this feedback picture of a self using inborn ("hardwired") capabilities to feed back into an internal equilibrium corresponds quite usefully to what the cognitive psychologists tell us. Presumably inside out and outside in and bottom up and top down all work together in intricate feedback loops at different levels.[114]

Similarly, Jean Piaget's studies of children's intellectual and moral development led him to posit a short-term loop in which we take an external object through an internal schema (assimilation) and a long-term loop in which we slowly adjust that schema to meet the requirements of the object (accommodation). The first schemata are innate, but from them proceeds a cyclic interaction through all the stages of childhood into adult intelligence, morality, and the rest (1970). Chomsky goes farther, holding that for language at least and probably for other parts of our perceptual apparatus, our cognitive structures are inherited and their development preprogrammed, like our limbs or body organs (1975).

Whether we hold with Piaget or Chomsky or both, then, some of our perceptual schemata are innate--biologically defined. For example, we can smell only those molecules that the receptor cells in our noses are shaped to receive. The very shape of the cells acts as a perceptual schema. Images held motionless on the retina, not subjected to the three scanning motions of the eye, break up and disappear (Pritchard, 1961). Speech also disintegrates without the normal auditory input from one's own voice (Klein, 1970, p. 352). I do not feel my socks unless I think about them. Evidently, to feel, speak, or see we have to be able to do something so as to bring a schema to bear and so create a sensory-motor feedback loop (Held, 1965). That, at least, is part of our biology.

Another part of our perceptual schemata is cultural. You and I, for example, live in a world of right angles, while Zambians do not. We see a certain picture made up of straight lines as a perspective of a box. They see it as simply a flat design (Deregowski, 1968). Colin Turnbull reports that his BaMbuti guide, who lived in a forest where the greatest distance he could see was a couple of hundred feet, could not recognize buffalo grazing some miles away as buffalo. He called them insects and even tried to identify the species of insect. On driving closer and finding that they were in fact buffalo, he concluded that he had been deluded by witchcraft. In effect, he was demonstrating that even so basic a visual schema as the correlation of size with distance rests on cultural experience (1961; see also Munroe and Munroe, 1975, Segall et al., 1966, or Lloyd, 1972).

Finally, some part of our perceptual schemata is individual. Herman Witkin found that the perception of verticality expresses the personality of the perceiver, so much so that one can even infer clinical characteristics (1959). Similarly, the eye movements with which people scan objects and pictures vary widely. George Klein was able to relate the eye movements with which people scanned pictures to the flexibility or constrictedness of control in their total personalities (1970, p. 185). "Every person," conclude Noton and Stark, "has a characteristic way of looking at an object that is familiar to him" (1971, p. 218).[115]

In short, experimental psychologists have come to three conclusions about perception. First, we see and hear actively; we construe the world. Second, we mingle culture, personality, and physiology when we actively perceive. Third, the act of construction through which we perceive involves something like many nested feedback loops that link self and not-self and work up and down a hierarchy from sensory details to global aspects of personality, physiology, and culture.

To be able to say this much amounts to quite a bit. Alas, however, this account lacks one crucial element, our perception of language.

If you say to me, "Please pass the salt" or "Kiss me" or "What time is it?" and I respond by passing you the salt cellar, kissing you, or telling you what time it is, no theorist can say definitively what has gone on in my mind. Likewise no one can say what went through my mind so that I could say even the silliest sort of linguistic example, like "The bat is in the bin."

A half-century of research, say three major writers in the field, Fodor, Bever, and Garrett, "suggests that what subjects remember about a text is a complicated function of the literal text and their beliefs and values" (1974, p. 273). Except for some such general principle, however, we have no idea of the precise mechanisms by which we grasp meanings or make connections in language. "Almost every aspect of sentence recognition," note these psycholinguists, "remains unsettled despite the experimental attention the problem has recently received."

Nevertheless, there is some agreement, I believe, that we understand a sentence like "The bat is in the bin" by guessing what sounds and words are coming next and comparing what we then hear with our guess. In more technical language, a constructivist model for the way we understand speech would include a system in which we generate patterns inside ourselves to check against what we are hearing or seeing. "Patterns are generated internally in the analyzer according to a flexible or adaptable sequence of instructions until a best match with the input signal is obtained" (Halle and Stevens, 1964, p. 604). We keep trying patterns until we get a satisfying match with what our eyes or ears are taking in. We evidently have some sort of inner program for generating patterns according to a sequence that we can vary according to the context or the matches we do or don't get.

In other words, information-processing feedback. We understand (at least partly) by "an active internal synthesis of comparison signals" (idem). In interpreting sentences, we guess ahead, and we guess at the whole. We hear the same way you read a line in this book. Your eye skims over the various strokes and serifs of the letters and jumps from word to word forming a rough idea of what you are going to get which you do not correct unless I write something unexpekted.[116]

When we speak, we "coarticulate" consonants and vowels into bursts of sound. When we listen, we have to decode those bursts back into individual speech sounds. Furthermore, the same piece of sound carries information about two or more successive phonetic segments. For example, when we say "bat," we don't make three separate sounds, b-a-t. We say one syllable. If an engineer displays that syllable visually on an oscilloscope, at any given moment the wave may be carrying information about b or a or t or all of them. Measured physically, the b in bat is different from the b in boot. Yet we hear them both as b. We are able, somehow, to decode syllabic bursts of sound back into individual speech units (Liberman, 1973).

The basic unit of listening is the whole syllable. We perceive individual consonant and vowel sounds only after perceiving the whole syllable. At the same time, however, it is the individual speech sounds that yield meaning ("bat" and not "boot"). In effect, when we listen to speech, we put into operation a basic psychological principle that applies to all perception.

We tend to organize our perception of the world in terms of the highest level of organization. We see a house, not a heap of shingles and bricks. We see a truck rather than an assemblage of tires, bumper, flatbed, and so on. More exactly, we hypothesize at the highest level of organization, and we check that hypothesis against smaller details (Bever, 1973).

In this constructivist view, hypotheses from the higher levels of the brain activate the small physical movements of the eye. The thing perceived is neither just a figment of the mind nor a revelation of a reality that exists apart from the perceiver. Rather, "Object and percept are part and parcel of the same thing." "The thing perceived is an inseparable part of the function of perceiving, which in turn includes all aspects of the total process of living"--so Ittelson and Kilpatrick (1951). "Vision is a dynamic process, using a series of scans, but these are not rigidly determined as in a television raster. They are varied according to the nature of the scene itself and the previous experience of the individual." "Seeing is not really representational but interpretive." "The brain picks out features of the pattern that combine with its internal hypotheses to provide programs of action" (Young, 1978, pp. 119, 123).

So far, in discussing perception of language and perception in general, I have been emphasizing the constructivist school, which holds that we perceive against schemata or constructs. In recent years another strong group of theorists has emerged around the work of J. J. Gibson. This "ecological" school believes in direct perception both of language and of things. A ball coming at you or a face beginning to smile or a policeman's whistle carries information, not in the sense of what a telephone operator gives but information as engineers use the term: the result of having excluded alternatives. In[117] that sense, this page carries information even before anything is written on it because it is white, not black. A baseball has a certain speed and curve and not some other, hence information.

According to this "ecological" theory of perception, I perceive the page or the ball because I have systems in my brain and eyes for getting out the information, that is, telling white from black or determining the speed and direction of the baseball. As applied to "The bat is in the bin," you hear the b in "bat" or "bin" by distinguish- ing it from nonplosive voiced consonants like v and from non-voiced plosives like p. That is how you can hear both those bs as b although, if we were to put them through electronics to make their waveforms visible, they would appear quite different.

From this point of view, the one question, How do you understand a sentence? becomes two. What is the information a given sentence holds? and What systems do we have for processing it? (Gibson, 1977; Shaw & Bransford, 1977). The earlier form of the ecological approach, the "adaptive theory of perception," concentrated on the first of these questions. Information is simply "there" and need not be processed at all, only picked up. More recent "information processing" theories of perception concentrate on both of these issues, but more on the second. What are the systems we have for, as it were, reaching into the world of language or sensations and bringing out the information an organism like us finds useful?

Both versions of this "ecological" school set our perceptions of language and the world in a larger framework: How does this perception of information fit economically into what this organism does or needs at this moment? Not all information is useful at all times to all organisms. What a monkey can grasp may be of no use to a giraffe. We humans adapt our systems to process the information that is relevant to us (Verbrugge, 1977). When you hear me say, as a linguistic example, "The bat is in the bin," it is important that you hear the difference between an i as in in and an o as in on, but it doesn't matter, for linguistic purposes, whether the sentence is spoken fast or slow, loudly or softly, or by a woman's voice or a man's.

Within the constructivist or analysis-by-synthesis theory, we would treat our perceptions of language (or the world) as a feedback. You would recognize the b in bin by checking the plosiveness of the sound waves against your internal references for plosive sounds like b or p or t, and you would also check that b against your internal references for voiced consonants like v or d. You would hear the whole sentence, "The bat is in the bin," by checking it against your internal programs for fitting individual sounds together to form words and for fitting words together into sentences (Halle & Stevens, 1964).

Although the individual psychologists who hold these constructive or ecological or information-processing points of view tend to debate the[118] nuances fiercely, it seems to me they share an important fundamental. As I read them, no one of the three suggests that it is simply a stimulus--either a word or a sound--that causes a given response. All three involve some sort of feedback or dialectic between reality and something relatively fixed inside us (an information processing system, a reference signal, or an ecological need). To that extent, all find the perception of language and of things a two-way street, although the ecological theory relies more heavily than I would on a subject-object distinction. Even within the ecological school, however, one can acknowledge:

Knowing is itself the process which resides neither wholely within the subject . . . nor wholely within the world as a cause or stimulus; rather, as an ecological concept, it stands like the mythical Colossus of Rhodes, astride the physical and psychological domains, one foot planted firmly on either shore (Shaw and Bransford, 1977, p. 10).

The most fully developed theory of language perception, however, comes not from psychologists but from linguists and psycholinguists building on the thinking of Noam Chomsky. (An excellent introduction is Lyons's 1977 book.) Chomsky developed the idea of a grammar as a system of rules for generating the correct (and only the correct) sentences of a language, in the same sense that the rules of arithmetic will "generate" the correct and only the correct balance for a checkbook. To be sure, some person has to do the "generating," but the rules themselves are independent of that person. They are some complex mixture of linguistic universals common to all languages and the grammar of a particular language like English.

Given this idea of grammar as a system of rules that generate sentences, one can ask some of the very basic questions about "The bat is in the bin" with some hope of their being answered. How is it that I can produce as many new sentences as I please? How can I interpret an infinity of sentences I have never heard before? Indeed, children six years old or even younger can do that. How did they learn to do it so quickly? One cannot possibly answer such questions by a notion of language as behavorist stimulus-response or a list of what everybody has already said in English or a dictionary of signs with fixed meanings, signifiers linked to signified in the theories of Saussure, for example.

According to Chomsky's "standard theory," every sentence in a natural language like English has a deep structure and a surface structure. The surface is what you actually hear as "The bat is in the bin." The deep structure may involve things like "noun phrase," "verb phrase," or "prepositional phrase," which have no meaning ("semantic content") at all, no particular reference to a bat being in a bin.

Like Freud, Chomsky has continually revised and refined his theoretical[119] structures around a largely unchanging core of data, sometimes in ways that go beyond his followers. He has kept, however, one basic assumption. Sentences can be declared well-formed or ill-formed by linguistic criteria alone. Thus, Chomsky himself does not claim that his grammar for "generating" sentences corresponds to anything that actually goes on in our heads, any more than the laws of arithmetic correspond to the way I know immediately, without any calculating, that 5 plus 2 is 7.

Psychologists, however, and psycholinguists (by definition) would like to get from a purely formal analysis to what our minds do in speaking, hearing, reading, or writing. The evidence suggests that Chomsky's abstract, almost mathematical analyses correspond somewhat but only somewhat to our thought processes. There are "divergences between grammars and recognizers." Also, researchers have not been able to show that what language learners, like children, learn is a transformational grammar (Fodor, Bever, & Garrett, 1974). Chomsky's linguistic criteria seem to correspond to some psychological reality, but no one has yet shown just exactly what.

The strongest antithesis arising from Chomsky's work drives in at precisely this point. The cumbersomely named "generative semanticists" say that Chomsky was wrong to assume that one could talk coherently about syntax without taking meaning and use into account (Lakoff, 1974). Generative semanticists therefore try to arrive at ways to "generate" not just sentences but contexts and meanings, too. In so doing, they confront head on the basic problem linguists face. Most linguists would like to be able to discuss the principles that govern language only in relation to language. They would like to box off all consideration of the real world and consider questions about language as though they could be answered entirely in terms of formal principles governing language. Yet is that possible?

I like apples.
The sentence says nothing about eating, but all of us know it doesn't mean the speaker maintains friendships with pippins and Macintoshes.
The policeman held up his hand and stopped the car.
Strong policeman? No. We supply a driver who saw the policeman and stepped on the brake (my examples are from Schank and Abelson, 1977, p. 9). In general, we have to use what we know of the world to interpret sentences.

Further, our knowledge of the real world enters into relations that seem purely linguistic. For example, the sentence, "The patrol cars were ordered to stop honking after 2:00 ;sca.m.;es" is ambiguous, supposedly on linguistic grounds alone. The surface sentence could come from two different underlying structures: "[Somebody] ordered the patrol cars to stop [somebody[120] who was] honking after 2:00 ;sca.m.;es" "[Somebody] ordered the patrol cars to stop [the patrol cars who were] honking after 2:00 ;sca.m.;es"

However, "The elephants were ordered to stop honking after 2:00 ;sca.m.;es" is not ambiguous. There is a linguistic ambiguity. That is, the surface sentence about elephants could have come from two different underlying structures just like those for the patrol car sentence. But, since elephants cannot stop other beings from honking (a marginally useful bit of knowledge about the real world), I do not sense any ambiguity in the sentence about elephants. The ambiguity, then, is partly a matter of language and partly a matter of experience.

My lack of confusion involves another puzzle. The distinction between inanimate and animate beings is apparently a part of all grammars, a linguistic universal. In these two sentences, however, I accept the idea that inanimate patrol cars can hear orders but animate elephants cannot.

Consider this pair of sentences:

   1. Helen is Manny's widow.
*2. Manny is Helen's widower.
As Robin Lakoff has pointed out, to understand the unacceptability of the second, you need to know that sex roles dominate our society's typing of women but not of men (1973). In short, "It is impossible to tell where linguistic knowledge leaves off and extralinguistic knowledge takes over" (Jackendoff, 1972, p. 19).

Within this general framework, I see a certain consensus among linguists and psycholinguists in 1984. True, no one knows just what it is we do when we recognize a sentence, and no one has discovered a psychological reality that corresponds to grammatical structures. Most students of the way we perceive language do agree, however, that we generate something--maybe a sentence, maybe a sentence plus what we know about its subject, maybe just part of a sentence or a set of phrase markers (like "noun phrase," "verb phrase"), perhaps nothing more than the rhythm of what a certain kind of sentence feels like--but we generate something sentencelike to understand someone else's sentence. That something can include principles either of language or of the world, and those principles support our top-down interpretation of language.

On that basis, it was possible for Frank Smith in 1971 to write a coherent, convincing account of what we do at the sentence level when we read, and for George Dillon to develop the same principles still further in 1978. We "predict" our way through a book. We project alternatives ahead of our eyes and then we eliminate alternatives until we arrive at a single interpretation. We eliminate some quickly, in advance as it were, because English (or any other natural language) repeats information. "The bat is in the bin."[121] After in, as soon as I have seen or heard the I can guess a noun is not far behind. Indeed, once I have seen in I can be fairly sure an article and a noun will follow. Hence, by the time I get to bin I have already had two clues that tell me to expect a noun and specifically a noun that in will fit--bin or basket, but not, say binge or basketball. This redundancy of natural languages allows us to eliminate most alternatives very quickly. Then, as our eyes pass over the page, they acquire just enough visual information to eliminate the remaining alternatives--to separate bin from bun or ban.

In other words, having read "in the," I project ahead some such hypothesis as "noun that a thing can be in." I get back from the text "bin," and that fits my hypothesis. The sentence makes sense. Had I gotten back "binge," grammatically possible but senseless in terms of my real-world knowledge, my brain would have flicked my eyes back to the capital that began so bizarre a sentence and started over, like jumping back through a German sentence after getting the verb at the end. Projecting a hypothesis ahead and then comparing the return one gets on the hypothesis with what was wanted is, of course, a feedback process. In this respect, our perception of language is like our perception of anything else.

In such a two-stage, feedback processing of language, the brain plays a far larger part than the eye--at least with fluent readers. The eye gulps in blocks of information, but to do so, the brain must tell the eye when it has absorbed all the information it can process in short-term memory.

Short-term memory can deal with between five and seven items at a time, but (interestingly) these may be letters, syllables, sounds, words, or ideas, depending on the fluency of the reader. The capacity of your short-term memory depends not on the total amount of information you are storing, but on the number of units into which your brain can group the information (Bever, 1973). A fluent reader carries five ideas, while a beginning reader carries five words or letters. Thus, to achieve useful speed in reading, we must be able to make brain and eye work together, using prior knowledge and skill, to shape processable units out of the whole conglomeration that our eyes perceive. Those units can be maps, structures, or cognitive schemata--according to the various schools of perceptual psychology--but they are what we see by.

Once our short-term memory has reached capacity with five or six units, the brain greatly reduces the eye's seeing (like turning down the brightness on a television set). It sends the eye into a jump (or, technically, saccade) to the next place on the page where the brain thinks it useful to process information (Smith, 1971, pp. 82, 104). Presumably, brain and ear work together in somewhat the same way when we hear language, with the important difference that we cannot actively "listen ahead" into the flow of speech as far as we can look ahead on the printed page.[122]

According to Smith, the way we perceive a page--indeed, "the way in which we perceive the world depends on the manner in which we categorize the incoming information, not simply on the characteristics of the incoming information itself" (p. 74; Dillon, 1978). Conversely (says a group studying cognition), "Much more than the knowledge of words is required to understand a sentence: There must be general knowledge about the world as well." "To understand a sentence, we appear to combine general knowledge about the world with knowledge of the structure of language and the meaning of the parts of the sentence" (Norman et al., 1975, p. 5), for example, that elephants honk but do not understand or convey orders while patrol cars can be said to do both.

In general, then, understanding language involves a dual, active process, some sort of analysis-by-synthesis, so that, as Smith says, "Whatever meaning is, it must be defined with respect to a listener or reader" (p. 35). The psychologists--even this very hard-nosed experimental group--return us to the elusive I. In other words,

Perception, Symbolization, and Identity

all go together.

We can't be sure about the details of the way we understand language, but we can feel fairly sure about the broad outlines of the process. We perceive words, as we do the world, by an active process of analysis by synthesis. Perception involves both the low-level, nearly reflex actions of eye or ear and the highest processes of which the brain is capable. In the automatic actions of our senses, we use our physiology. In the higher processes of the brain, we use our cultural training. And somewhere in it all, we make a perception into a personal, individual experience.

Sam's and Sandra's perceptions of the Faulkner story worked the same way as their symbolizations. Sandra hesitated as to whether Emily's father was standing or sitting. (The word "spraddled" admits this ambiguity.) Sam placed Mr. Grierson in the door with Emily framed by his legs rather than the two of them framed by the "back-flung front door." In a class devoted to the exact reading of literary texts, a teacher might well want to settle whether Faulkner's "spraddled" meant sitting or standing and whether Emily and her father were framed by the door or the door-frame. Here, however, we can simply note that Sam's and Sandra's needs govern not only their symbolizings but their very perceptions. Sandra's hesitation between sitting and standing lets her make Emily's father a comfortable size. Sam's framing makes masculinity reassuringly dominant.

In the same way, Ernst Lanzer, caught in a thunderstorm with his lady, felt compelled to count between the lightning and the thunder. That is, he[123] perceived the time between thunder and lightning (not as I would, as a measure of the distance of the storm) but as a gap that could be dangerously penetrated. He had to prevent that. Similarly, he saw a stone in the road as a potential source of harm, therefore something to be moved out. Both of these perceptions work variations on his identity theme about controlling things coming in and going out.

Anna S. perceived her lover's behavior this way:

When he tells me something that was unpleasant to him . . . I hate the thing or person for it. I feel it displeased him and that makes it terrible. If he is very tired, fatigue takes hold of me. . . . When he laughs . . . I am filled with sheer glee.
Again, at the very level of perception, Anna construes the effect of events on her lover to achieve the same state of mind herself and so "become another's essence."

Hence the defense, the D, in DEFT refers to a very large idea: strategies for deciding what can come into the individual's reality. "People see what they want to see." More exactly, people see what they can cope with in order to cope with it, just as Anna sees fusion or Lanzer dangerous penetrations. Freud anticipated this principle, we have seen, not only in his 1937 "testamentary" essay, but even from the very beginning, in the unpublished Project of 1895 and The Interpretation of Dreams (1900).

The defense (in DEFT) with which one can analyze a person's perceiving is thus intimately involved with that individual's style. So also is the fantasy. The individual builds on what he has taken in to project into the event a characteristic wish-fulfilling fantasy. The fantasy does not lie latent in the event, only the materials for it. People fantasy, not stories and not events. By perceiving the lightning and thunder as a gap that he could fill, Ernst Lanzer gratified both sides of his ambivalence: his wish to penetrate and see his lady hurt; his counterwish to protect her. Only by singling out her lover's feelings as the important part of his behavior could Anna S. acquire those feelings herself and so gratify her deep wishes to be at one with a caring, loving person as once she had wanted to fuse with her mother. Sam perceives Emily as frail and fragile in order to remove threats to his own manliness and conversely to imagine a strong masculinity on top of things. Sandra sees a Bugs who does not threaten, indeed who exemplifies hospitality. All these moves constitute a DF combination, negative and positive, defensive and gratifying, within a DEFT perception or symbolization.

In the same way, experimental psychologists show, for example, that the differences in a person's tendency toward active coping or passive submission will enter into the purely perceptual task of separating an item from its context (Witkin et al., 1954, p. 489). A brain physiologist, J. Z. Young,[124] concludes: "What we see and hear is largely the result of our own programs of search" (1978, p. 69).

Getting an act of perception to fit one's characteristic defenses is quite delicate, for it involves considerable adjustment at the gateway between self and reality. The second phase--the projection of fantasy--is much more free precisely because the first has kept sources of anxiety to a manageable size. The third principle, transformation, has to do with the same general flow toward gratification as fantasy. We need to pack our intellectual baggage neatly, to keep on good terms with our cultural heritage and our intellectual paradigms, to bring events into tune with our personal view of the world--all to the degree that feels comfortable to us. Percepts that make sense do not evoke the anxieties that raw fantasies or wishes might. They help stabilize our coping. Hence political, ethical, and the most abstruse kinds of scientific or philosophical "making sense" all serve our personal needs. None are "objective," free of emotional investment.

Hence Sam reads the Hemingway story as being about a young male's education. Sandra reads it as a story about being overpowered. Anna S. (after her analysis) does not simply decide she wants to be at one with her mother. She generalizes. "Does real loving make one feel a part of another?" Or she develops a social theme: "Women should not have to work at a public job for money . . . . They should have a man to take care of that part, but they should . . . do all they can to make him happy."

Defense, fantasy, and transformation are ways each of us shapes an event to make it a tolerable mixture of what our ordinary psychic functioning can cope with and what feels strange or different. Hence, before any of those three came into play, there was a first and primary mode: expectation. Each of us approaches a new experience with a characteristic cluster of hopes, desires, fears, and needs. We trust that the new experience will let us act out this economy of expectancies to net us pleasure. Consciously, we expect satisfaction. Unconsciously, we expect to be able to DEFT the new experience in our particular way of DEFTing.

When we feel satisfied by an experience, as Sandra's perception of Bugs pleased her, that signal tells us we have succeeded in working through some system of matching defenses, projecting fantasies, and transforming into meaning characteristically. If we do not feel satisfied, if we respond to a perception with guilt or fear or pain, that signal of unpleasant affect tells us that so far our effort to fit the event into our psychic processes has failed. The event feels alien to us, like Lanzer's neurotic compulsions or Vincent's political coercions or Anna's sense that her real self was not involved.

Accordingly, we try to change that unpleasurable answer to our expectation. We do something about the event, or we do something about our relation to the event. We may even just blot it out--not perceive it--as[125] Freud managed not to hear the Tyrolean church bells or as Iiro was able to deny his inability to play the flute or do carpentry.

In any case, it is our emotions that monitor the moment by moment functioning of these four aspects of perception or symbolization. Whether or not our ego "can deal with" some feature of reality, we know only by the way we feel. The emotional signal we get depends in turn upon the whole circuit of expectation, defense, fantasy, and transformation.

Within that fourfold interaction, two aspects, transformation and expectation, relate to each other in a special way, around time. Meaningful minutes of experience build on instants of expectation. Each separate millisecond of eye scan which Iiro gave Winnicott's first squiggle was governed by his overall expectation (an unconscious E in a quite unconscious DEFT) that he could interpret this drawing as evidence that this new person would accept him as he was. The psychoanalytic theory of identity thus combines with the theory of perception that experimental psychologists have been developing.

The psychoanalyst looks at larger perceptions, more personal, more committed to entireties, but the shape of the psychoanalyst's theory has the same shape as the experimentalist's. What the experimentalists have shown in a great variety of contexts is: "Perception is never a sure thing, never an absolute revelation of 'what is.' Rather, what we see is a prediction--our own personal construction designed to give us the best possible bet for carrying out our purposes in action" (Ittelson and Kilpatrick, 1951, p. 179).

Thus, the particular studies of the perceptual psychologist and the larger theory of the psychoanalyst lead to a still more general philosophical position: a theory of knowledge (rather like that of Husserl). People are not blank sheets (Lockean tabulae rasae) on which the world writes its message, nor do we perceive the world through some kind of behaviorist conditioning. Rather, we confer meaning. You and I are freely emitting centers of meanings which others either share and so confirm or refuse to share and so deny (Poole, 1972). I remember the confirmations or disconfirmations and the feelings that went with them, and they form the basis for new expectations and new standards to govern the feedback loops of perception. Just as perceptual psychology describes the external confirmations of the meanings I confer, so psychoanalytic psychology describes the way they both express and create my identity.

Thus, when we perceive and when we symbolize, we are doing much the same sort of thing. To see a chaos of areas and colors as a bin, you and I need to be able to symbolize it as a bin. To symbolize it as a bin, we need to be able to see the lines and textures that make it up. The two processes are not contradictory, as Jones thought, perception faithfully copying the world and symbolization then distorting it. Neither does one come first. Rather, per-[126]ception and symbolization work together. We see and we sign in one continuous process of feedback and dialectic. "The brain," writes physiologist Granit, "does what no computer can imitate: in growing and developing, it creates the world it needs" (1977, p. 128). The brain's ability to make experiences into coherent symbols (of Jones's kind or any other) enters into the very workings of eyes and ears. Hence, I can ask of Iiro's discovery of a duck in Winnicott's squiggle, How does this relate to the rest of what I know of Iiro? Combining the modern idea of identity with the modern idea of perception points toward an account of mind itself.[127]

6  /  A Model of Mind

In answer to my craving to know about mind, a philosopher, Jerrold Katz, once suggested to me that different theories of something as elusive as mind or matter are like different architectural drawings of a house. The perspective sketch of the house on its site will look very different from the floor plan which in turn will differ from the wiring and plumbing diagrams. Each of these blueprints has its different use for which the others may well be useless or invalid. One cannot use the wiring diagram to landscape nor find the electric outlets in the overall perspective sketch.

Blueprints are not simply right or wrong; they are each right or wrong within their particular context. It makes no sense to say a blueprint is right or wrong in some context for which it was not designed. We can ask, however, that the wiring diagram not indicate rooms which do not appear on the floor plan nor the overall perspective show a storey which does not appear in the lumber specifications. In other words, we can ask that a blueprint, whatever its context, not be inconsistent with some blueprint which is valid in its context.

In the same way, in physics, we can ask that macroexplanations of falling bodies and colliding billiard balls not conflict with the microexplanations in particle physics of electrons and muons. To be sure, the two kinds of explanations will look very different and will not transfer into each other's contexts. Nevertheless, they should not contradict each other.

So with mind. We can--and should--ask that psychoanalytic and psychological explanations of large-scale behavior not contradict models of the brain from molecular biology, but we need not ask that they take the same form (Weatherick, 1980). It is in that spirit that I want to propose a model of mind that will be consistent with psychoanalytic theory, with new physiological discoveries about the brain, and with new psychological theories about perception and language.

Psychoanalysis began (in a sense) with Freud's effort to make a nerve-brain model of the neurotic processes he was observing (1895), and recently Karl Pribram has updated Freud's model using holography (1969). Heinz Hartmann continued Freud's bias toward the natural sciences but included biology in one of his models; the other, according to Roy Schafer, proceeds[128] from an analogy to government (1970). As long ago as 1959, Lawrence Kubie suggested using electronic models from communications engineering and computers for psychoanalytic theory and Emanuel Peterfreund has thoroughly developed such models for the therapeutic situation (1971).

For a model of mind in general, we can begin with the information-processing feedback loop that proved useful for thinking about symbolization and perception. Suppose you are expecting to read the letter b. You compare what you see to some internal idea of the features of b: consonant, lowercase, a unit round but with an ascender (hence not a q or a g) on the left (hence not a d). If the test checks out (provides satisfying feedback), you go on to the next letter. If the test does not, you bring another set of features up for comparison.

The loop has to have three features. One, a behavioral end acting on the input: my eyes scanning the print on the page. Two, some standard or reference signal that is outside and uncontrolled by the loop--like my ideas of what b and q and d are like. Three, a comparator like my brain, comparing the marks my eye sees to that standard. Some

Some psychology books treat feedback loops as though they had only two elements, an output and a comparator. That's wrong. There has to be that standard inside me but outside the loop sensing this b. Usually the way I know that standard is being met or not is by the way I feel. The car feels safely positioned--or it doesn't, and I feel anxious and reposition it so that I don't feel anxious.

Through that preestablished reference, I can link these loops into a hierarchy, by having one loop set the standard for another (as in fig. 2). In the earlier example of steering a car, I can imagine a smaller or "lower" loop concerned simply with physically positioning the car where I want it to be. Suppose I have decided that the car should be one meter from the shoulder. If a gust of wind causes it to drift to half a meter from the shoulder, I turn the

Two-loop feedback diagram

[129] wheel left to get it back to one meter. That standard of one meter has to come from outside the loop positioning the car.

It could come from another, larger loop. If the smaller loop is "positioning the car," we might imagine that larger loop as something like "thinking about road shoulders." Suppose I come to a section of highway under repair. A sign says "No Shoulder." I see that driving off the right edge would entail dropping six inches into some loose gravel. Comparing my one meter standard for a smooth shoulder with this new riskier shoulder, my ideas of safety say I should now stay a meter and a half from the edge of the road. I change the standard for the smaller loop. I then use the lower loop to turn the steering wheel to position the car according to this new standard from the upper loop.

Suppose, however, that a child runs out from the left side into this construction work. Drawing on a still higher level of abstraction--an injured child is worse than a broken axle--I drive on the edge of the no-shoulder road or over it to avoid hitting the child. I have changed the standard for the "thinking about shoulders," which in turn changed the standard for the "position the car" loop.

Memory also enters in. My previous experience with road shoulders feeds back through memory to become one of the things I (a rather high-level I) take into account in setting my standard at one meter.

Notice that a two-loop feedback picture allows me to take into account cultural ideas within my overall style. It might be the custom in my part of the country to drive one meter from the side of the road. I, a conservative driver, customarily drive 1.5 meters from the shoulder. My son customarily drives 0.5 meters from the edge. Suppose, however, we lived in Quebec and the custom were 0.5 meters from the shoulder. A similarly conservative and not-so-conservative Quebecois father and son might drive at .75 and .25 meters, respectively, from the shoulder.

In 1948, Clyde Kluckhohn and Henry A. Murray issued one of the grand gnomic statements of modern psychology: "Everyone is in certain respects (a) like all other men, (b) like some other men and (c) like no other men." They were stating the puzzle of personality and culture, for example, my writing the b in The bat is in the bin. The marks I see on the page are the same marks that probably every other human being will see there, but not everyone will experience them the same way I do. A French writer's idea of b will not correspond to mine because his culture teaches a different way of writing b. Finally the b I write will be like no one else's b.

Will a two-level feedback picture allow me to interrelate the ways in which a given experience has some elements that are the same for all people, some that are true for certain groups (like cultures), and some that are unique to[130] one person? Yes. This is just what William T. Powers suggests in order to generate a general model of mind.

A Linking of Loops

Powers is building on the achievement of Norbert Wiener, who worked out the mathematics of cybernetic systems and suggested as far back as 1943 that they could be used to describe living, adaptive organisms (Rosenblueth, Wiener, and Bigelow, 1943).

Powers states his basic thesis in the title of his book: Behavior: The Control of Perception (1973a). He intends a reversal of the usual stimulus-response view which would make perception the control of behavior. He goes on to model the mind as a whole by imagining a graded series of such feedback loops, a hierarchy, in which each loop corresponds to a certain level of perception and also provides the reference signal for the loop below it.

At the most basic level, Powers posits signals of muscular effort and signals of intensity, things like loudness, brightness, stiffness, or pungency. They are the most basic because one must perceive a certain minimal glimmering of light before one can see other visual qualities like color or an edge, a minimal faint hum before one can detect pitch, a certain minimal tactile pressure before flatness, and so on.

To develop the next step in the hierarchy, Powers seeks a set of control systems that is hierarchically above the first-order systems, but at the same time as close to the level of first-order systems as possible (p. 99). The inputs to the second-order control loops are combinations of such first-order intensities as pressure, light, sound, vibration, balance, or taste. Powers suggests thinking of the second order as controlling a sum (mathematically, a vector sum) of first-order intensities. He calls this sum a sensation, for example, of purple, chocolate, or G sharp.

In other words, our second-order controls organize first-order intensities into qualities loosely corresponding to the traditional five senses, although other groupings also occur. For example, when I hear the sound sss, I simply hear it. When I speak sss, however, I sense both sound and effort. Yet both times I "sense hissing." The sensation is nominally the same but actually involves different intensities, cutting across different "senses."

Powers comments on the "'philosophical fact' that emerges from this theory." I taste "chocolate," but no physical entity corresponds to my taste--only a combination of sugars and oils. "Perceptual signals depend on physical events, but what they represent does not necessarily have any physical significance" (p. 113). In general, Powers concludes,[131]

The brain may be full of many perceptual signals, but the relationships between those signals and the external reality on which they depend seems utterly arbitrary. At least we have no assurance that any given perception has significance outside of a human brain. . . . We may strongly suspect that there is a real universe out there, beginning a millimeter outside of our nervous systems, but our perceptions are not that universe. They depend on it, but the form of that dependence is determined in the brain, by the neural computers which create perceptual signals layer by layer by layer through transformations of one set of neural currents into another (p. 37)
In other words, Powers is describing at the perceptual level something very like the DEFTing we have seen at higher levels, the matching of expectations, defenses, and fantasies to form a top-down, bottom-up bridge from "in here" to "out there." He goes on to build a hierarchy that rises to the highest levels of mental functioning.

To think our way from a second to a third order, we need to imagine a system just above but as close as possible to the second-order system dealing with sensations. Powers suggests that we think of the third order as configurations, that is, an unchanging function of a set of sensations (second order) which are in turn sums of (first order) intensities.

For example, the retina's individual rods and cones generate first order signals of brightnesses. Still within the retina, computing networks of cells organize these signals of intensity into edges, areas, or gradients, that is, sensations (second order). Powers theorizes that the visual centers near the thalamus process these into configurations, the third order. He finds some confirmation in brain physiology for this arrangement of the three layers: Wilder Penfield's well-known experiments stimulating with electric probes the visual cortex of people undergoing brain surgery (1975, pp. 21-27, 55-56). The subject "sees" forms: a red star, a blue disk, a green ball, or a black wheel. Similarly, when Penfield stimulated the auditory cortex, his subjects heard ringing, humming, chirping, and other auditory configurations. One particularly important class of third-order perceptions, then, would be phonemes like the bs in "The bat is in the bin."

We may inherit some of these configurations. René Spitz showed that a three month old infant recognizes faces in one particular way: as forehead-eyes-nose. The baby will respond even to a Halloween mask having these features but not to a real human face in profile. It seems likely that forehead-eyes-nose is an inherited configuration for organizing experience, biologically a very useful one for a baby seeking to nurse and be nurtured by someone paying attention to him, as we say, "facing" him (pp. 86-96).

We adults have a similar ability. We can recognize faces from just a few[132] features and even though they may change drastically. Since last I saw friends from college or graduate school, their faces may have wrinkled or their hair grayed, yet I recognize them. Indeed, I can recognize a cartoon of Winston Churchill or Hitler that shows no more than a globular head and a cigar or a mustache and a shock of black hair.

The opposite of this ability is the disease prosopagnosia: the victim can no longer recognize faces. "What is remarkable about the disorder is its specificity," remarks the psychologist Norman Geschwind. One can know people by name, and one can describe faces verbally, but one cannot put the image of the face together with the name. "The lesions that cause prosopagnosia are as stereotyped as the disorder itself." "The implication is that some neural network within the region is specialized for the rapid and reliable recognition of human faces" (1979, p. 189). Possibly Powers's third-order control systems are this sort of thing.

The idea of a third level dealing with configurations also gives us a way of thinking about the familiar but puzzling psychological phenomenon called "object constancy." We see a chair as a chair no matter what angle we see it from or even if we see it upside down. We see it as comprising horizontal seat and vertical back even though it may never, in the real, physical world, actually present those directions. In a feedback model, we do not have to think of the brain going through some very complicated "compensation" to make constants out of the varying ways our sensations present a chair to us. We need only assume that the reference signal for the chair configuration remains constant as the chair rotates and pivots with respect to us so that we always have a constant to compare our sense data to. In effect, for certain larger purposes, we tune out certain kinds of changes.

Then how do we perceive differences as we assume different positions with respect to the chair? Because we perceive distances and size and orientations as third-order configurations, says Powers. "What we express in serial language as 'the big chair cattycorner on the far side of the room' is probably perceived at third order all in parallel: big and chair and cattycorner and far and side and a room" (p. 126). These configurations may account for some of the familiar optical illusions: the staircase that turns upside down into a cornice or the vase that turns into two faces if you think of it that way.

A fourth level of control will account for our ability to see "motion pictures," the rapidly repeated images of television or movies. "When a human being is presented with two related configurations in sequence, within a short enough span of time, he perceives a new entity of experience not present in either configuration alone: change" (p. 130). Too fast a transition, and we see simultaneity, for example, the blur of the blades of an electric fan rotating faster than ten revolutions per second. Too slow a transition, and we do not perceive change at all as, for example, the minute[133] hand of a clock, which moves at 1/3600 of a revolution per second. Between, we perceive change, like the turning of the second hand or a series of musical notes that seem to be a glissando or like the light bulbs on a theater marquee that make a moving arrow, although they are only stationary lights flashing on and off. We also control these transitions in any movement like raising an arm, creating a rising or falling tone of voice, or tracing out a path in space as when a child runs to catch a ball.

If the fourth level enables us to control movement, the fifth must enable us to choose one movement rather than another. That is, the fifth must involve the way we sense and control the sequence in which we sense lower-order perceptions. I decide how to move to get up out of my chair, go across the room, pick up theTimes, read the headlines, and say "Damn!"

The saying of "Damn!" I find particularly interesting. It would be at this fifth level, sequence, says Powers, that we recognize words, and he goes on to design circuits of neurones that would do this job. At the third level, the recognition of phonemes like ss, only sixty or so circuits would be needed. The fourth level would involve common transitions from one phoneme to another, like st or the ay in day. The fifth level would provide circuits for recognizing sequences of phonemes, one circuit per phoneme in a word. Hence, he argues, in order to recognize 150,000 words, averaging six phonemes apiece, no more than 900,000 neurones would be required. Such a "lexicon" (to borrow a word from the linguist) could fit into a space of 150 cubic millimeters, about half the size of an almond. In this circuitry, as in the various speeds involved, nothing that Powers proposes is physically impossible for our brains.

Even so, this whole description of the brain as operating through five levels of feedback is speculative. Levels beyond the fifth are even more so. Powers suggests for the sixth level relationships among fifth- and lower-order entities: intensities, sensations, configurations, changes, and sequences. The relationships might be: and's, or's, cause-and-effect, if-then, space relations or relations in time, probabilities, or just plain association. At the sixth level, we organize and interpret our experience. We experiment with it. We understand how the six of hearts relates to the deck or why pushing a doorbell makes a sound.

For a seventh level, Powers suggests the control of programs: looking for one's glasses; playing chess; baking a cake; making love; holding a conversation. The seventh level builds on the ability to make and recognize relationships--the sixth level--but goes beyond it in setting up a series of tests and decisions leading to different procedures. I am trying to do the crossword puzzle. I need my pen. I go into the bedroom, lift up a jacket, grab the shirt under it, and poke into its pocket. These into's, under's, and on top of's are sixth-level relationships, but I have organized them into a compli-[134]cated series of programs and subroutines. These strategies or programs or sequences of relationships plus control of sequences of relationships illustrate what Powers' proposes for a seventh level.

Powers raises the question of the role of language in this description of human rationality. There are two ways of thinking about language. In one, words are simply perceptions that evoke nonverbal perceptions--referents or "signifieds." I perceive the word bat and I think of a club you swing. Alternatively, I might think of the word bat as something that can be manipulated in a purely symbolic way without any reference to the bats of the real world (purely a "signifier"?). If the latter, we would have to assume a special level for recognizing and controlling verbal symbols as such, and Powers rejects this for several reasons, one of them parsimony. Either way, however, language fits into the model as the most powerful means we have for setting up and altering programs--although, as Powers notes, we also seem to have programs that are nonlinguistic, like tying one's shoes or touch typing.

For an eighth level, Powers posits a system for choosing among programs, setting up programs, or altering programs toward certain ends. He borrows the concept of "heuristic principles" from artificial intelligence: general ideas of how to win the game, sell the product, make it neat. Principles are facts confirmed by experience, such as strength in the center of the board is good to have, honesty is the best policy, or "I could show this psychologist up." Eighth order systems would operate in very complex situations that do not yield to exact procedures that can be set down ahead of time (like tying one's shoes--seventh level) or in situations with a lot of distracting detail where one needs to average out and select data. Eighth order systems are what computer programmers would call "executive programs." A psychoanalyst might think in terms of superego directives to ego.

Beyond the eighth level, Powers suggests a level that chooses one set of principles over another. He calls this level system concepts, a procedure for seeing some collection of moral, factual, or abstract principles as in fact a connected set. Such a loop would perceive things like the United States Army, the Democratic party, or "my family." With these entities all the physical things that embody them may change--everything that was the Democratic party during the Roosevelt years has disappeared--yet I recognize a thing I call the Democratic party above and beyond those sensory particulars.

This very high level would deal with the philosophical problem of "Locke's sock." A stocking may be darned and darned and darned until not the slightest bit of the original stocking remains, yet at no point (or at what point?) do we say this is no longer the original stocking. "System concepts such as Society or Culture are not to be found in the world represented by[135] physical models of the universe; they are elements of psychological models of the universe . . . effective exactly to the degree the individuals learn to perceive them and choose goals with respect to them and develop means of maintaining their ninth-order perceptions in those goal states" (p. 173). Powers gives the concept of "the system" itself (that is, the system you can't beat) as such a ninth-level perception.

In psychoanalytic terms, I would say the boundary between Powers's imaginary eighth and ninth levels corresponds to the boundary between secondary process, problem-solving thought and primary processes like dreaming or free association. In a classic study of opinions, the researchers concluded that an opinion was "a resultant or compromise between reality demands, social demands, and inner psychological demands. The three are inseparable" (Smith, Bruner, & White, 1956, p. 275). Their metaphor is "compromise" or "resultant" as "boundary" was mine. Powers's hierarchy of feedbacks offers us much more powerful metaphors for these complex psychological relations. Self guides cultural standards which guide biophysical standards.

Oddly, it seems likely that our feelings, our intuitions, our half- conscious "gut reactions," primary process thinking, in other words, directs our perceptions and movements, secondary processes, rather than the other way round. If we are working with reality we use our lower-level processes of sensation and perception, and secondary-process thinking directly guides these reality-oriented processes. Primary-process thinking however, connects to the feelings that tell us whether our secondary-process feedback through reality are working. If we are asleep, as Freud surmised, our primary processes can run free, unconstrained by the real world. When we wake, then our primary processes take control again. Hence primary processes guide secondary (see pp. 336-37).

Beyond the ninth level Powers bows out, offering only guesses. Perhaps instincts provide goals--and here, I think, Freud's notions of instinctive drives, combinations of love and aggression, might fit, for example, my own urgent need to understand that propels me through all this cerebration and lucubration. Alternatively, Powers suggests memory, surely an important element throughout the hierarchy since it allows for changing (or not changing) the personality embodied in all these reference levels. To do something differently or the same as before--at any level--one must have a memory of how it was done before. Memory is the way lower levels change upper ones.

For levels ten and beyond, says Powers, "Maybe our brothers from the East have something to tell us" (p. 174). I would suggest something less mystical, a theme-and-variations concept of identity. All lower-order processes, from notions like society and culture at the ninth level down to the[136] mere sense of brightness or loudness at the first, all serve to preserve the I, to make it feel "right." I need to know things. It is more than need. I hunger, I desire, I crave. I am an addict of ideas. Hence I feed on the authors I have been quoting (like Powers) and write the pages you have been reading, using lower-level systems, configurations or sensations, to do so.

I am in a continuing process of creating "I" through these various lower-order systems. Feelings tell me that at any given level my actions are yielding perceptions that suit the goals of my identity at that level. Memory at all these levels permits the accumulation of samenesses and differences in that "I" which you (or I) can read as themes.

One can imagine identity (in all three senses, agency, consequence, and representation) as the top level of the hierarchy. "All the programs of the brain constitute one single model or one single model or structured system," writes J. Z. Young, a brain physiologist. "The brain has many distinct parts but there is increasing evidence that they are interrelated to make one functioning whole, which gives a unique and characteristic direction to the pattern of life of that one individual" (1978, p. 265). I read him as legitimating the idea of a holistic identity governing a variety of programs, like Powers's hierarchy of feedbacks.

Having tried out Powers's richly imagined model of the brain, however, we now have to give it up--at least partly. A modern brain physiologist like Ragnar Granit agrees that feedback loops serve us well for a rough description of the self-correcting principle even in high-level processes, yet they can be no more than that. Our minds' circuits are nonlinear. Neurones do not simply switch on and off. These circuits work in incredibly complicated patterns in which one circuit excites or inhibits another in response to a comparison. While one cell in our retina responds to a color, the cell next to it suppresses its response in the interests of exactitude. Contexts enter into each on-off, and any one neurone may be the context for the next. These circuits, moreover, have a vast redundancy, nature's safety factor in the organ most important to our survival. All these circuits interact and are repeated so many times over that one cannot simply represent these three-dimensional structures by the two-dimensional diagrams of a computer scientist.

Perceptual psychologists of the ecological persuasion, for example, are showing that one needs to go beyond hierarchy. One needs to think in terms of heterarchies (in which control is passed back and forth from an upper loop to a lower loop) or even "coalitions" in which the system consists of part of the organism acting with its context in the environment. One must think of the "functional integrity" of the individual complementing an environment that also has "functional integrity" (Turvey, Shaw, and Mace, 1978)--dare I say wholeness, unity, identity?[137]

Finally, though, I have to admit that we cannot use Powers' loops even repeated many times and ranked in a hierarchy to discover the workings of our eyes or ears or memories. Nevertheless the feedback model has its uses. For example, a certain area of the brain moves the eyes in response to an electrical stimulus. Moving the eyes is a motor action, notes a team of researchers, but the same action also reprocesses visual information. It is part of the sensory system. "The lesson is that no line can be drawn between a sensory side and a motor side in the organization of the brain" (Nauta, 1979, p. 106). In other words, Powers's general principle, that behavior--motor action--is the control of perception, is sound. Our brains do operate like feedback systems. We cannot use this broad generalization to map the cell-by-cell workings of our brain and senses, but it will serve as a framework within which to think about less detailed principles like psychological laws or psychoanalysis' idea of individuality.

Thus, Powers's loops fit the descriptions in some well-known psychological analyses of planning and thinking, for example the Test-Operate-Test-Exit (TOTE) sequence suggested by Miller, Galanter, and Pribram (1960). One checks the image of things against a plan--that is the test. One then operates to bring the two together. Test again, and if the result is satisfactory, stop. Newell and Simon's "heuristics" are very similar. You set a goal, judge the separation of the present situation from the goal, take a step, see whether the separation is reduced, judge the difference between the present and the goal, take a step, and so on (Simon, 1957, see also Bruner, Goodnow, & Austin, 1956). These are feedback procedures, as are most "information processing" moves of artificial intelligence.

Granit notes the "hierarchic stratification" of the various controls in the brain, and this is another central feature of Powers's model with which most brain scientists would agree. "Here," writes one authority, "one finds remarkable agreement among scholars of widely different backgrounds, disciplines, and eras . . . that purposive movements are built on a base of reflex processes." "Reflexes and voluntary movements are not opposites," he continues, discussing the brain's control of movement. He gives the example of a Russian study of the champion pistol shots in the Red Army. As they aimed, their pistols remained immobile, even if their legs or arms or shoulders moved. In effect, reflex mechanisms stabilized the position of the marksman's hand in space. He ends by thinking only of his goal: hitting the target. "The actual events that underlie the achievement of the goal are built up from a variety of reflex processes" (Evarts, 1979, pp. 170 and 179).

Further, those processes are relatively isolated one from another. As Herbert Simon has shown, evolution would favor organisms built on hierarchical principles, because disruption of parallel hierarchies would cause only partial loss (one finger among five, color blindness but not total blindness),[138] whereas an undivided system would undergo total loss. Hierarchical systems thus have the important property of "near decomposability." That is, the connections within a subsystem are stronger than the connections between subsystems. Nature has separated the fast processes within a subsystem (seeing by an eye, say) from the slower processes linking subsystems (coordinating arm and eye in firing at a target). This partial isolation lets us think and use the systems as though they were separate. It corresponds to Powers's connecting his systems only by a reference signal.

When we speak or write, we start with an idea we want to express, and we then find ways to express it. In other words, we use grammar (whether we are thinking of a formal linguistic grammar or a psychological one) in the service of higher-level aims (Schank and Abelson, 1977, p. 7). Our grammar or "lexicon" functions as a separable circuit.

Granit, in his aptly titled book The Purposive Brain (1977), represents today's thinking, anchored in electrophysiological studies of the brain. The remarkable discoveries of recent years, he notes, allow scientists to list in detail "inhibitions and symptoms at synaptic loci," separate systems. Even so, scientists have to give those listings "teleological relevance" to make them more than merely lists (p. 175). The scientist has to take into account goals and purposes, and that is why behaviorism and the routine correlation of behaviors are bound to fail. We can understand the mechanics of voluntary motion by studying physical facts, but we can only understand the whole process if we set those mechanics in a hierarchy that involves "higher" mental levels. We need to be able to talk about goals. Hence we need to include in our physical model of the brain psychological concepts of purpose, will, or demand. Granit is performing an important service. He is placing alongside the "hard" research of brain physiology reasoning from such "soft" disciplines as psychology, psychoanalysis, or identity theory. We have to talk in terms of

Higher and Lower


If we accept this idea of a purposive hierarchy, a lower stratum in the hierarchy can never explain the raison d'être of the level above it. Rather, each level will require goals from higher levels to explain it--Powers's "reference signals" or "standards" from outside the several feedback loops.

By endorsing the principle of hierarchy and admitting considerations of purpose into physical accounts of the brain, a modern brain physiologist (like Granit) is continuing in today's thinking some of the classical experiments of Sir Charles Sherrington. Sherrington showed experimentally that reflex activities, such as a dog's scratching with his rear leg, persisted even though the cord to the brain had been cut so that there was no way the brain[139] could participate in the act. This is not to say the act is not changed, however.

In his lectures of 1937-38, Sherrington explains that any individual organism's motor behavior has two components. One is reflex, and since it predominates in the lower, earlier levels of the evolutionary scale, it seems basal. The other is a superstructure, and it is not reflex. In the higher animals the roof-brain supplies this superstructure, whether the behavior be instinctive or rational. As one proceeds up the evolutionary scale, activities that were controlled at their site--the movement of a lobster's antenna, for example--become controlled in mammals in the brain. The roof-brain component increases the finesse, skill, adaptability, and specificity of the motor act. As we go higher on the evolutionary scale, this superstructure becomes more prominent, and it is most prominent in humans.

If the motor act is deprived of that roof-brain component (as in Sherrington's cutting the spinal cords of dogs), if it is reduced to its reflex foundation, the motor act becomes, in the highest animals and most of all in humans, imprecise, inconsequent, and without "skill." In lower mammals, evidently, the reflex has normally a larger share in the behavior and the reflex foundation is in itself more capable and complete, for the motor act suffers less on withdrawal of the higher, the nonreflex, component. Even as between dog and man, the dog can stand and walk and run competently after exclusion of the roof-brain. It can direct itself visually. But man not so. A dog, however, after its spinal cord has been cut, can no longer adapt its reflex acts to a special purpose nor can they be given greater skill by training.

Evidently, sensory impulses get "long-circuited" to the roof-brain, which issues a "call" to reflexes to get them to cooperate. The call fits a reflex to this or that purpose of the moment which the animal's situation asks for. The dog not only walks, it walks to greet its master. "In a word the component from the roof-brain alters the character of the motor act from one of generality of purpose to one of narrowed and specific purpose fitting a specific occasion" (1963, pp. 182-83).

Sherrington's picture matches that of Powers. When Sherrington speaks of "long-circuiting" through the "roof-brain," his lower-level loops embody the dog's (or our) ability to walk or our ability (and perhaps the dog's) to recognize faces. These lower levels work within a hierarchy so that they serve higher needs of consciousness. Granit uses the term "cephalization" (which we could English as 'headening'). Von Bertalanffy speaks of three levels in the human. The spinal cord acts as a reflex apparatus. The ancient brain, the "palaencephalon," is "the organ of depth personality with its primeval instincts, emotions, and appetites." The cortex is the organ of the "day personality," that is, the conscious I (1952).

For all these theorists, mind is a gradual higher and lower. Aristotle[140] decided that the lower limit of life defies demarcation. Sherrington concluded the same of mind: "There seems no clear lower limit to mind. . . . Ultimately mind so traced [downwards along the scale of being] seems to fade to no mind." "A well-versed observer of the one-celled animal world has said that were an amoeba as big as a dog we should all acknowledge its mind" (pp. 208, 209). Granit adopts the same position: "From the evolutionary standpoint of modern biology, consciousness is an emergent novelty, probably still existing in an increasingly rudimentary form as one descends in the phylum" (1977, p. 72).

Michael Polanyi, the biologist and philosopher of science, puts this gamut more formally by speaking in terms of boundary conditions. Think of a vacuum cleaner. It works under two distinct sets of principles. One set is the general laws of electricity and mechanics. The other is the special ideas that went into the design of this particular Hoover. Polanyi suggests that we think of the particular ideas that adapted the general laws to the cleaning of rugs as boundary conditions. One can think of any useful or purposeful restriction of nature the same way: the organs inside a body, for example, plants and animals, ultimately even we humans. There is no way the laws of electricity, chemistry, or mechanics can provide their own boundary conditions. A more particular or "higher" embodiment of those principles has to do that. Hence, one has to think of a continuum of control running from the highest to the lowest forms of life and even below that, to vacuum cleaners (1968, p. 1311).

Trying to simulate human understanding in computers, the psychologists Schank and Abelson found it necessary to posit a similar hierarchy, parallel but in terms of understanding rather than brain function. If a computer is to understand such ordinary human activities as going to a restaurant, "understandability is a function of the place of a piece of information in context. A script is understandable as a particular realization of a plan. A plan is sensible only if it leads to some desired goal. And, a goal is understandable if it is part of a larger theme" (1977, p. 132).

Thus, intelligence itself answers to a hierarchy, and at the top of the hierarchy is what they call themes: "A life theme is no more than a collection of goals that in some sense 'go together' and a set of behavior patterns appropriate to attaining those goals." In the hesitation Schank and Abelson embody in those quotation marks around "go together," I recognize the experimentalist's unfamiliarity with what is commonplace to me as a literary critic: the organization of ideas from details to a center by means of themes and themes of themes. They are finding in intelligence the holistic unity in variety that I find in personality as a whole.

Polanyi and Schank and Abelson provide us with a model as general as Powers's hierarchy of feedbacks and very like it. In effect, the structure of[141] the internal and external feedback connections for Powers's organism constitute boundary conditions for Polanyi's "forces of inanimate nature." Just as Powers' single feedback network leads naturally to a hierarchy of networks, so Polanyi's higher levels provide boundary conditions for lower levels until we get down to unformed physical and chemical laws. So also Schank and Abelson's hierarchy provides successively more general contexts as one goes up the scale for events that are successively more specific as one goes down the scale.

A variety of scientific perspectives, then, I find converging on one general picture. The computer engineer (Powers), the brain physiologists (Granit, Sherrington, Young), the philosopher of science (Polanyi), the simulators of human intelligence (Schank and Abelson, Simon) all agree on two fundamental principles. First, one can symbolize mind at least partially as a hierarchy of processes. They may be feedback loops, reflex arcs, TOTE systems, or "scripts," but they all process information so as to correct themselves against a goal beyond the loop. Second, therefore, in such a model, higher processes provide reference levels, boundary conditions, purposes, goals, criteria, or understandability for lower processes. Such a hierarchy will span the "highest" functions of human individuality or intelligence down to one-celled animals or even inert matter. Even a cockroach can have a style.

The model of mind I am suggesting is one more version of this ladder of feedbacks. I see a hierarchy consisting of various forms of DEFT. There are ways of looping wishes or fantasies out into the external world (fantasies) and ways of controlling what shall be admitted from that external world into the self (defenses). There are ways of looping low-level expectations in the immediate flow of experience back into high-level transformations of experience into significance. Both at the highest level of that hierarchy and permeating all the loops, I see a personal style. I can represent that style as a theme-and-variations identity. It is the goal of goals, a unifying "life theme" in Schank and Abelson's sense. It is

The Top

I am suggesting for identity a role like that often given by students of the brain to consciousness: a giver of direction from the highest level, a unifier, a centerer. Sherrington described it as "the unifying by the mind of its experience of the moment," "an integration." He gave as an example a familiar "optical illusion": "What is being looked at as a set of steps suddenly without warning becomes an overhanging cornice. But it is always the one or the other wholly." The mind interprets its "now" into a situation with a single meaning. Sherrington calls this constant unifying "the principle of[142] convergence." "That unifying of the experience of the moment is an aspect of the unity of the 'I.'"

Some forty years after Sherrington, a physiologist like Granit can draw on many more achievements in unraveling the networks that make up that I. For Granit the key terms are "purpose" and "creativeness." At the lowest level, we have the homeostatic balances of molecular details, further up, the holistic workings of the organs and limbs, still further up, behaviors, and finally, the processes we call mind. At every level, the creative purpose of a higher level shapes the working of lower levels. The body uses lower processes for its higher ends so that there is "a reorientation of purpose from level to level." "A warp of creative purposiveness is woven into the fabric of biological hierarchies with consciousness at its top level." One has to take that purposefulness into account, Granit insists, in one's explanations as a scientist, even if such explanations "never will end in the differential equations that the physicist uses for his world of interpretation" (1977, p. 85).

Long before Granit, in 1932, Edward Tolman had issued his well-known maxim, "Behavior reeks with purpose." Long before Tolman, William James was explaining, "why we ought to continue to talk in psychology as if consciousness has causal efficacy" (1890, p. 138). Consciousness brings efficiency to a noninstinctive brain by "bringing a more or less constant pressure to bear in favor of those of its performances which make for the most permanent interests of the brain's owner" (p. 140). "The study . . . of the distribution of consciousness shows it to be exactly such as we might expect in an organ added for the sake of steering a nervous system grown too complex to regulate itself" (p. 144). His metaphor, steering, looks forward to guidance systems that would be described mathematically half a century later by Wiener and applied to just this problem by Powers and the rest.

For example, one recent psychologist, Roger Sperry, speaks of the brain, by means of its own consciousness, monitoring its own activities. Consciousness does not, of course, check on the actions of individual cells. It detects overall qualities of different activity patterns. The brain does not "intervene," altering the laws that govern the generation or transmission of nerve impulses. Rather consciousness "supervenes," fitting the brain's activity within an envelope of larger configurations, as the drops of water swirl in an eddy in a stream (Calder, 1970, p. 260). Sperry sounds very like James.

"The cerebral activities of every [person] are a unity," writes another student of the brain, J. Z. Young (1978, p. 134). I am suggesting that we think of the unifier that supervenes or steers at the highest level as "consciousness" or "feelings," yes, but more precisely, as a theme-and-variations concept of identity. We have defined it as agency, consequence,[143] but most importantly in this context as representation. In effect, identity allows us to "the" an "I."

The "interests" that James finds consciousness favoring may not be universal. Eating is a pleasure to most of us, but may not be to the dieter, the bulimic, or the anorexic. Not many people will "need to know" as hungrily as I do. A theme and variations concept of identity allows us to translate such general notions as consciousness or interests or pleasures into Shaw's interests, Freud's pleasures, Fitzgerald's purposes, Iiro's worries, or my own need to know things wholly and certainly.

Lichtenstein uses two different terms: primary identity and identity theme. Both refer to a theme that we infer, that runs through all we know of some mind's activities, and that we put into words. By "primary identity," he intends a style in the person. Created in the earliest relationship between the baby and its first caretaker, it is a way of being which is in, even is, that person. Formed before words, it is a preverbal thing that never can be put in words, "known" in that sense. Just as we can never know the mind of another person, so this primary identity, the essence of that mind, must remain mysterious.

By contrast, an identity theme (as I use the term) is my way of re-presenting a human being to myself. To be sure, when my inference of Bernard Shaw's identity theme "fits" or makes sense (feels right by my criteria), I can say it must approximate his primary identity. I can never be sure of that, however, nor need I be.

"Identity theme" is a paradoxical concept since it puts the essential me-ness of me somewhere between me and you. On the other hand, as we have seen, this between-ness in our interpretation of another fits what we know of the ways humans see the world or put it into words and other symbols. We mingle ourselves with the "out there," creating and re-creating it as we bring to bear on the world our characteristic ways for perceiving it.

In a purely psychological sense, a theme-and-variations identity governing a hierarchy of DEFT networks will provide a model for the processes of perception and symbolization by which we continue both our one-ness and our two-ness with the world. At the same time, however, we need to remember that even simple perceptions are a function of a person's


as well as of a person. The identity cum feedback picture allows us to portray the relationships among identity, biology, and culture.

Recall Colin Turnbull's BaMbuti guide who had grown up in a jungle where he saw nothing beyond a couple of hundred feet. He saw buffalo[144] at a distance of some miles and thought them insects. In effect, the guide demonstrated two things. First, culture can limit biology, removing some possibilities in even so basic a visual schema as the correlation of size and distance. Second, if so, culture must even more deeply and pervasively color the way we perceive family relations, politics, the meanings of words, beauty, funniness, or the feelings we have on hearing a story.

In our feedback metaphor, we can take the role of culture into account by compressing Powers's multitier hierarchy back into a two-tier diagram:

Two-loop feedback diagram

As before, the human being lives in the world by means of behavior that controls perception. That is the lower loop: a person acts on the physical world to change perceptions so as to feel better emotionally. An upper loop lets me visualize the way a "social factor" can limit or add to biology by setting the reference level for the lower loop. The upper loop uses the lower, just as the I uses the upper. The individual uses the culture he has internalized, which in turn uses physiology. Physiology both equips and limits culture, and culture both equips and limits the individual.

For example, my body is built for sitting. In that sense, my body makes something possible for me. I can't twist my waist enough to sit with my knees facing backwards, however. My body makes something impossible for me. In the same way culture makes things possible and impossible for me by means of physiology. I can spin my swivel chair around--my culture makes that a possibility for me--but by sitting in chairs I have atrophied the muscles for squatting that Africans and Asians find so useful.

In general, the culture we internalize changes what our physiology can do. Not vice versa, notice. As I internalize My culture it equips my body with a language. So provided, I can fashion the words you are reading. That language then sets limits to what I can physically hear or think or say--as our students keep demonstrating and even our most eloquent poets:[145]

                    . . . every attempt
Is a wholly new start, and a different kind of failure
Because one has only learnt to get the better of words
For the thing one no longer has to say, or the way in which
One is no longer disposed to say it.
Culture provides words for T. S. Eliot as for the most inarticulate of us. Physiology provides a tongue and vocal cords with which to speak, ears with which to hear, and (so Noam Chomsky might say) part of my brain wired for language. Culture also confines us to what Frederick Jameson has called the prisonhouse of language. We can say only what we already know how to say.

My body, like everyone else's body, is built for language. But I speak English and not Italian. That means I can't even hear the difference between "r" and "double-r" in Italian, and even though I speak some French I have trouble not hearing "nom," n-o-m, and "non," n-o-n, differently (which I have to avoid if I am to read Lacan). Further, my English is not quite the same as yours. I speak with a certain drawl and pitch. I write using sentence modifiers like "nevertheless" and "however" more than most people do. In linguistic jargon, I write and speak an idiolect and a dialect and a language. My identity sets the standards for the particular brand of language my culture provides me, English, which in turn tunes my body's speech equipment to certain sounds and not others.

Language, tools, cooking, family structure all provide standards for eyes, ears, tongues, and fingers. The ground of our lives is physical and biological, yet culture can limit our physiology, as it did with the pygmy guide.

The two-tier diagram provides a metaphor for him. Quite unconsciously, you or I would "set" our reference levels for judging distances. We would program ourselves with data about the angles, the hues, or the sizes and silence that indicate a buffalo is two miles away. We do so by means of cultural resources, but the pygmy's culture has never let him learn about a buffalo two miles away. In effect his social loop has never set this reference level for the biophysical loop, and a buffalo two miles away simply can't exist for him.

The culture we internalize both enlarges and limits biology. Finally the individual uses or works with his ability to see buffalo at a distance or, indeed, works with his inability (says they are bewitched). Language functions the same way, as a tool, as when we speak of the "linguistic resources" of different cultures (Hymes, 1973).

Identity as a hierarchy of feedbacks also allows us to model expression and communication in a world of private minds. I express myself according to my[146] identity. To do so, I use cultural resources like language, chapter divisions, the apparatus of scholarly reference, or the system of university presses. Through them, I create a physical object: the book you are holding.

You, in turn, see my book. You perceive it physically, testing it against your schemata for seeing lines, curves, or serifs as letters and words. You measure my syntax and grammar and spelling against your knowledge of English, your vocabulary about minds or Shaw, and my book (I hope) passes these as well. You can then test my writing against your knowledge and experience of people, your belief or disbelief in feedback, say, or unity, and your idea of what constitutes evidence and truth.

I wrote by using a system--English--and you understand by using a system--English. But I write and you understand by using that same English on behalf of two different identities. Your private feelings and beliefs will guide and finally govern your perception of my English words, but it was the expression of my private feelings and beliefs through an English we both share that made communication possible in the first place. Your and my private feelings make communication personal, the English we share makes it transpersonal, and neither mode makes the other impossible.

This picture of reading and writing as two similar feedback loops acting on a single text provides a model for our situation in a culture. We share loops with our fellow-citizens, but each of us uses those shared loops for personal purposes.

This is a very different picture from the usual metaphors about "social pressure" or "political forces"or "economic pressures" or "the impact of society." To be sure, societies enforce a consensus as to what is real, true, and valued. No doubt people often experience this enforcement as a "pressure." Even so, it is a confusing metaphor. Culture does not exert "pressure" or "force" or "impact" so much as it either gives us means or takes them away. It gives me a Toyota but denies me a Rolls Royce. It gave me English but withheld Homeric Greek. Society may make one or another choice very costly or painful, the loan company may even take my Toyota away, but society cannot exert a physical force the way chains or walls or drugs do.

What culture does do is give us resources. Within those resources, some things are possible for an individual to choose and others not. Some cultures deny some individuals some choices (being able to see buffalo at two miles), but that is as much as culture can physically "coerce".

So far as freedom is concerned, the individual is to culture as culture is to physiology in our two-loop picture. The most rigid of totalitarian regimes can do no more than take over the lower two loops, the way Orwell's 1984 society takes over language and history. No matter how coercive the society,[147] it can only reach the private, individual choice through those lower processes. Even a slave is free in his thoughts.

In this feedback picture, as I use the language or the technology or the symbolisms of my culture, they act back on me, as a hammer both enables and limits a carpenter or as words let Eliot start anew but doom him to failure. The realities we use change us.

They change, even, the very structures of id, ego, and superego. From the first, Freud and later analysts understood that intrapsychic structures develop out of a person's adaptations to reality. The id manifests the body's biological drives psychologically. The superego internalizes our acceptance of our parents as parents and ourselves as children with all the limitations and delays that acceptance implies. The ego sophisticates the pleasure principle: we continue to seek pleasure (or the absence of pain) but with a canny eye for possible advantages by delay or compromise with the demands of reality or the laws of logic or cause and effect.

The ego also embodied, however, the individual's seemingly inborn ability to cope with an "average expectable environment." Hence the ego must have existed from birth, and psychoanalysts of the 1950s tended to think of functions of the id, ego, and superego as absolutes, like functions of heart, lungs, or liver. Most psychoanalytic theorists of that time assumed that id, ego, and superego, as structures, would be much the same for any human, independent of cultural or historical setting.

The later analysts were close to Freud's own thinking because at the time Freud created psychoanalysis reality meant an ordered universe of eternal natural laws which set boundaries within which humans had to work out their lives in a similarly ordered society. Freud himself believed in the reality of reality with especial intensity. If reality is as constant as human biology, then it doesn't matter whether psychic structures are inborn or adaptations to reality.

The turmoil of the 1960s turned that fixed reality upside down. We have seen technological and social change accelerate to the point where very little in our environment is "average" or "expectable." Today, when the very dimensions of space and time have stretched beyond all comprehension, reality may attack more than support one's sense of a meaningful and continuous self.

If so, then, as Heinz Lichtenstein points out in his analysis of the crises of the late sixties, the functions of id, ego, and superego may change (1973, 1977). The id and the pleasure principle may become ways of affirming one's own reality, and an ego whirling in a spiral of change may need to seek stability in mystical states of union. The superego may cease to be the incorporated voice of parents and culture, for they are lost in a distantly[148] fading past, and instead become a way of claiming for oneself the authority the elders once had.

Societies exist because they permit a consensus as to what is real, a consensus that finally they have to coerce. A revolution marks the replacement of one consensus about reality by a new one, and the various revolutions of the 1960s (colonial, sexual, civil rights, drugs, musical) all did or tried to do just that. As these movements changed realities, the psyche that reality feeds back into also changed. In a very real sense, the ego and the superego of the 1970s and 1980s are different from the structures of the 1950s and early 1960s--at least in the patients American psychoanalysts see. The instability of "objective" reality (in the nineteenth century sense) lets us realize once again what reality "really" is--the sharing of experience with meaningful others through processes of shared feedback.

As Erikson said, there is no person without society. Theme-and-variation identities are social and political as well as individual. More precisely, theme-and-variation identities are the way people are social and political. Still more precisely, theme-and-variation identities are a way to describe the way people are social and political. As in the famous 1956 study of the Bostonians' views of Russia,

A man's opinions inevitably bear his personal stamp. His capacities for abstract or practical thinking, for intense feeling, for forthright action set limits on his response to public issues, and indeed, on what he makes of any significant event that impinges on him. His intellectual and temperamental qualities, general features of his behavior stabilized in the complex interaction of constitution and personal history, give distinctive form to his opinions about Russia no less than they do to his copings with a psychologist's ink blots. . . . If the trivial ink blot evokes a valid sample of his personal style, so does a question about Russia. The tasks are more similar than might appear at first glance (Smith, M., Bruner, & White, p. 259).

What this book does is make terms like "capacities" and "qualities" more precise: a theme and variations concept of identity governing a hierarchy of personal, cultural, and physiological feedbacks. The individual sets the limits within which culture sets the limits within which he puts limits on the physical world so as to control his perceptions to yield satisfactions. This being feedback, physiology enables and limits culture, and culture enables and limits the individual. The individual manipulates culture and culture manipulates physiology.

Metaphors of social "forces" or "determinism" do not convey that curious mixture of limitation and freedom. I think it more accurate to say the[149] individual uses what cultural resources are available to him which in turn use what physical and biological resources are available. I like John Dewey's term "transact." We transact cultural and physical realities, much the way we carry on affairs, both of the heart and the briefcase. When it comes to culture or language, we are lovers or dealers, artisans or even artists.

There is a way, then, that we can combine thinking about limitation, freedom, beauty and laughter and love, semantic codes and transformations, neurones and synapses and feedback, all these levels, by closely examining the lower levels, even in a two-loop simplification. Identity is the reference level governing the hierarchy which is mind. The lower loop has to do with the feedbacks through which we move and sense the physical world. The upper loop has to do with the cultural systems and codes we have internalized. Although they may feel to us as though they have a "force" or "impact," they do not coerce us like a wall. We should define them dialectically. Identity represents choice and freedom, whatever is a function of the person, and it is the thesis. The lower loops must be the antithesis: whatever can not be chosen.

What cannot be chosen in a cultural sense is what any normal member of the culture could not find otherwise. "Furiously sleep ideas green colorless"--there is no way I can find that sentence grammatical. At the movies, the sound track carries a woman's voice, and the camera looks over the left shoulder of a woman to a man's face, then cuts and looks over his right shoulder to the face of a speaking woman. There is no way I can read that as her not talking to him. The camera shows a basketful of what look like potatoes, then cuts to a close-up of a potato. There is no way I can read the second picture as other than an example of the potatoes in the first picture. Someone else, however, who knew no movies or no English would simply find these questions incomprehensible--out of their cultural loop (like the African natives who, on being shown a United Nations agricultural film with close-ups of potatoes in it, ruefully said, "But our potatoes aren't that big").

The second loop, in short, makes use of codes and syntaxes that admit no options--unless one steps out of the culture. If, however, normal moviegoing adults could differ--some people, for example, might see the woman as angry, others as alarmed, some might see the potatoes as pale, others as dark--then we are in a higher loop where identity governs cultural perceptions.

That kind of cultural or linguistic or semiotic code differs fundamentally from another kind, from the conventions we adopt, for example, as interpreters when we accept a shared idea of what is relevant, what will count as data, or what qualifies as confirmation. These are rules of a game that we adopt only from and toward other players who are doing the same kind of feedback as we. They are rules we get from our "interpretive community" (Fish, 1980).[150]

A semiotic code, it seems to me, has to be defined as "No normal member of this culture could read this otherwise." Interpretive communities are more elastic. They provide us with something like ready-to-wear hypotheses to wear while going into the world, but they leave us choices between ready-to-wear and tailor-made and even between various styles and colors of ready-to-wear.

In other words, we should include in our feedback picture one of Chomsky's most useful concepts, degrees of grammaticality. The loops that govern eyes and ears are physiological, almost as limiting as physical laws. In our picture, they are the hard-wired connections from behavioral output to perceptual input. The loops that govern the way we see letters and words are cultural, not physical, but to be understood as "No normal member of this culture could see this otherwise." They too need to be understood as hard-wired connections. In our diagram, a code like that functions almost like the hard-wiring or environment of the feedback loop.

By contrast, the interpretive communities to which we hold temporary allegiances provide us with ready-made conventions, expectations, or gambits, for example, the whole idea of thinking of a person as a theme and variations. These are simply hypotheses--outputs--which we feed into the world by means of hard-wired connections and to which we get either positive or negative feedback. They would appear in our diagram as behavioral outputs from an individual, the content of the hard-wired semiotic codes rather than the connections themselves. This kind of convention does not limit or enlarge us in the willy-nilly way that physical or (I would say) cultural codes do. They open up possibilities, to be sure, but we are quite free to choose otherwise.

We defined identity as agent, consequence, and representation. Our abbreviated feedback diagram gives us a picture of identity as agent. Identity as agent sets the reference level for our cultural resources which in turn sets the reference level for our physical abilities. Psychoanalysis and brain physiology are, to be sure, very different, but psychoanalysis can provide blueprints of mental processes that accord with the blueprints the neurologist draws. With this simplified picture, we have, in effect, arrived at a metaphor that fits different but consistent blueprints of mind, as Jerrold Katz prescribed. This is surely what Freud, when he was an ambitious young neurologist just beginning to study the mind, always wanted: a

Model and Definition

of mind.

The house of mind is built with small functional networks. As a first approximation, we can describe them as feedback loops, although "feedback" alone and even "information processing" are too simple for the complicated, nonlinear patterns of inhibition, excitation, mutual regulation,[151] and redundancy with which our nerve and brain cells combine. DEFT provides another first approximation for sorting out the trends that run from inside to outside and from "low" to "high" functioning.

As those terms "low" and "high" presuppose, we can think of mind as a hierarchy of functional networks. Higher levels use lower levels to do their bidding, controlling them by reference signals, much as an artisan controls a lathe or an angler the line. Lower levels contribute the information that higher levels need to function, the tug on the fishing line or the sound of the bowl turning on the lathe. Thus there is an interaction all up and down the hierarchy. At every level, an upper level pipes down into a lower one some personal standard--identity.

One can inquire into and explain much of human behavior by assuming for the top of this hierarchy, identity: a final, pontifical, unifying I--like the relentless analyzer and knower who has hauled you through these theoretical chapters. That I creates itself as a consistent I by means of satisfactions and reassurances (for me, knowings) that fit the I's nature up to that moment.

Yet this I is also between observed and observer. It is a way for one human being to represent another. Your idea of your identity will not coincide with someone else's idea or mine. While at the lowest levels everyone will find in you much the same circuits of seeing and hearing and reading as in anyone else, at the identity-finding level, we are engaged in a human interaction that is as unpredictable, as "different" as any human relation.

In Kluckhohn and Murray's gnomic statement, each of us is partly like everybody else, partly like somebody else, and partly like nobody else (1948). The two-level feedback picture allows me to think about those relations, to think about an earlier example, writing the b in bat. The marks I see on the page are the same marks every other human being will see there, but not everyone will experience them the same way I do. In France, they teach a different way of writing b. Hence, a French person's perception of b will not correspond to mine. Finally no one else writes b with the same handwriting as I do. b. In general, the picture of an identity governing a hierarchy of information processing feedback loops allows me to understand how I am unique, cultural, and human all at once.

This picture of an identity governing feedbacks also allows us to define "mind": that which adapts less-than-mind toward identity. Less obliquely, mind (as opposed to body or matter) consists of the reference levels (in Powers's sense) or the teleological elements (in Granit's), the boundary conditions (Polanyi), the "life themes" (Schank and Abelson), or the purposes and goals (James and Sherrington) which guide lower functions toward higher aims. In the three elements of the feedback network (the first two being behavioral output as compared to perceptual input), mind must[152] be the third: the comparison and reference level. Mind reveals itself, however, only as the goal supplied to a lower network by a higher one. In a simplified picture, the reference level for one loop--its mind--is simply the next higher loop.

Mind is therefore future-oriented, mind is more abstract, but what is more important, mind is relative to the level we are talking about. What is mind to the individual cell of the retina is less-than-mind to the process of seeing a face. What is mind for seeing a face is less-than-mind to the process of recognizing Winston Churchill.

In these various levels of mind, we can use a higher level to look down into the levels below, but we cannot use a lower level to look up into the levels above. The higher comprehends the workings of the lower and thus forms the meaning of the lower. Hence, as we ascend a hierarchy of boundaries, we reach to even higher levels of meaning. Our understanding of the whole hierarchic edifice keeps deepening as we move upward from stage to stage. We can generally descend to the components of a lower level by analyzing a higher level, but the opposite process involves an integration of the principles of the lower level, and this integration may be beyond our powers. In other words, we can always get knowledge of the lower processes in the brain, but we may not be able to have knowledge of the highest levels of mental functioning, those we customarily call mind and to which I would add identity.

This limitation on our knowledge offers a sort of answer to our century's strong philosophical critique of the very idea of "mind." Gilbert Ryle and, following him, Stuart Hampshire and many others have criticized our common usages of the term as, for example, a location: "He was out of his mind." "The first thing that comes to my mind . . ." Or as a manikin: "My mind refuses to accept the idea that . . ." "He has a healthy mind" (Ryle, 1949, Hampshire, 1962a, b).

They extend their attack to such ideas as motive or intention. If I decide to buy a bag of popcorn, why should I assume that there is some personlike mover of that action prior in time and somehow inside or behind the scenes that "intended" to buy popcorn? Thinking that way gets us into an infinite regress: I made up my mind to buy popcorn, I made up my mind to make up my mind to buy popcorn, I made up my mind to make up my mind to make up my mind to buy popcorn, and so on. If one action requires a preceding intellectual move, why not all?

The psychoanalyst Roy Schafer has extended the argument to much of psychoanalytic language: ego, id, superego, impulse, defense, libido, and even everyday words like love, guilt, aggression, or anxiety. All entangle, he says, actions with pre-actions and pre-pre-actions and pre-pre-pre-actions (1976). Even the word "I" becomes an elusive fiction from this point of[153] view, a phantom of the opera loose in the mental works.

Yet the brain physiologists do not shy away from the infinite regress as the philosophers do. If the brain works by a hierarchy of functional networks, and if the lower networks require reference signals from the higher, then we are dealing with a regression. I can look down through any given level, but I may not be able to look up from a low level into a higher one, just as I may not be able to look through the words of a poem toward a meaning for it or a purpose or an intention.

"Mind" and "I" will be more than my mind when I am trying to think of "mind" or "I." They will be, in a philosopher's phrase, "systematically elusive." Yet it is possible to explore and to infer things about them.

Powers, using his electrochemical model of feedback networks, can infer how long before an action its intention must be, how long before the intention the higher levels that give it the necessary information have to be set, therefore at what frequency they must cycle. While we cannot say those "higher" levels are in fact "higher" or "within" we can say they are prior in time--and by how many microseconds. Herbert Simon, similarly, concludes that the speed within loops greatly exceeds the speed of the loops controlling them, guaranteeing the "near decomposability" of the various loops. Up to a point, it does seem as though we can interpret concepts like "purpose" or "goal" or phrases like "I intend" as referring to physical entities (1969, p. 106).

We can and should purge psychoanalytic language of fictions like "I" or "ego" if they conflict with other blueprints that have equal claims on our belief, such as the work of the brain physiologists. We should cling, though, to the blueprints that accord with radically different cross-sections of the house of mind like the brain physiologists' account of mind as a hierarchy. So far as "mind" is concerned, identity theory provides just such a summing of different blueprints within the general system of psychoanalysis. Mind is 1) the body's hierarchy of 2) feedback (or information-processing) networks whose reference levels are a function of 3) identity.

This identity at the top--and therefore permeating the system--is threefold. It is an agency, the "subject" of the various loops that initiate our seeing or walking or talking or reading. It is also the consequence of those acts as they feed back into the being who initiated them, the I created by seeing and walking and talking and reading. That identity--that I--uses culture as culture uses physiology. Culture both limits and makes possible the aspirations of an I, as physiology both limits and makes possible culture, as language limits and makes possible the realization of the poet's vision. Finally, identity is a representation, necessarily partial or elusive, by one I or another of this whole. This last--this between-ness--is the mind modeling mind that permeates its own modeling yet cannot ever be represented in it.[154] Identity theory yields not only a model but a necessary paradox, an elusiveness intrinsic to a thinking, changing being, the I as it grows in time, the I with a history.

Back to the cover page?