The Message in the Bottle and Lost in the Cosmos

by Walker Percy

  A Tertium Quid:

  The Lady Novelist?

  Tolstoy once said that a talented lady novelist could spend five minutes looking through the window of a barracks and know all she needed to know about soldiering.

  If she can see so much in five minutes, how much more must the talented therapist see after, say, a hundred hours with his patient?

  So here is the real question, or rather the main specter which haunts every inquiry into language as behavior. Granted the shortcomings of the two major methodological approaches to the talking patient—the analytic-psychical and the organismic-behavioristic—is not the sole remaining alternative the novelistic? Instead of “novelistic” we could say phenomenological, for the novelist must first and last be a good phenomenologist, and to most behavioral scientists phenomenologists are closer to novelists than to scientists. But is it not the case that when all is said and done and all theories aside, what happens is that the therapist gets to know his patient pretty well, understands him, intuits him, can talk with him and about him—and that behavioral theory can never say much about it?

  Let us at least articulate our unhappiness. Unhappiness changes. We are no longer miserable about the old quarrel between classical behaviorism and classical psychoanalysis or about the more intricate quarrels and rapprochements of their followers. For it has become more and more evident that our main emotion when confronted by both Freud and Skinner, say, is not partisan feelings—for both are “right” in their way—but rather epistemological embarrassment. Both men put forward dyadic models, one for organisms interacting in an environment, the other for invisible “forces” interacting within a psyche. The question now is not which approach is right but how both can be right at the same time. To us now, Freud’s and Skinner’s models stand to each other like the two worlds on each side of Alice’s looking-glass. Both worlds are demonstrably right and useful in their way, but how do you get from one to the other?

  Is the lady novelist the only tertium quid?

  But first, what does the lady novelist see if we put her down, not outside a barracks window, but on the other side of a viewing mirror through which she can see therapist and patient who were talking about the rat behind the arras and related oedipal feelings? She notices first off, let us say, that the patient does get excited. But far from its being the case that he is upset and is “resisting” the disclosure of unpleasant unconscious contents, she has the distinct impression that the patient is delighted. Moreover, being a good novelist and well attuned to the intellectual fashions of the day, she has the distinct impression that the patient’s pleasure has something to do with the fact that he has produced a kind of behavior which measures up to, or fits in with, the very theory to which he and his analyst subscribe. Perhaps it also occurs to her that the patient is in a sorry fix indeed if his chief claim to happiness is that occasion when he manages to be sick in the right way.

  Suppose that the lady novelist is right. Is she then the tertium quid? Is her way the only way to get at what is going on? And if it is, has not all the fun gone out of the game of behavioral science and the scientific method itself lost its splendid rigor?

  Have we not in fact come back to George Miller’s original misgiving, which haunts all behavioral scientists when the subject of words and meanings is raised? Must we not then let it go at that, surrender the field to Tolstoy’s lady novelist, or to Husserl, which is to say the same thing?

  Perhaps. But Charles Peirce did propose a radical theory of signs which undertook to give an account of those transactions in which symbols are used to name things and to assert sentences about things. In view of the heroic and generally unavailing attempts during the past fifty years to give such an account through one or another dyadic theory, it might be worthwhile for once to approach triadic behavior with a genuine triadic theory.

  Such a theory might bestow order and system upon the phenomenologizing which to the behavioral scientist must seem closer to novel writing than to a science of behavior.

  For example, the oedipal patient’s agitation may be given some such preliminary reading as follows:

  The patient’s agitation is not dyadic misery—resistance to the disclosure of unacceptable unconscious contents—but triadic delight. This delight, moreover, is quite as fundamental a trait of triadic behavior as organismic “need-satisfaction” is in dyadic behavior. It is a naming delight which derives from the patient’s discovery that his own behavior, which until now he had taken to be the unformulable, literally unspeakable, vagary of one’s self, has turned out not merely to be formulable, that is to say, namable by a theory to which both patient and therapist subscribe, but to be namable with a name which is above all names: oedipal!

  As such, the patient’s delight has good and bad, authentic and inauthentic components, which must be traced out and identified within an adequate triadic theory. Thus, the patient’s sentence It’s oedipal! must be investigated for Platonic and even magical components in its mode of coupling as well as for its valid intersubjective celebration of an important discovery. Perhaps the patient’s sentence can be paraphrased in some such terms as: “At last I have succeeded! At last I have produced a proper, even a classical, piece of psychopathology!”

  Accordingly, the patient’s behavior with its strong normative components must be evaluated on a normative scale which is in turn an integral part of the triadic theory in question. It is impossible in other words to avoid the subject of the patient’s impoverishment and loss of sovereignty.

  In his astounding achievement of applying the scientific method to the irrational contents of the unconscious, Freud did not have time to consider what goes on between doctor and patient, nor how a technique itself can loom large as part of the intellectual furniture of a later age, much less how it could come to pass that one can fall prey to the very technique one seeks help from.

  But that does not excuse us from investigating these matters.

  * Actually the dyads should be segmented in some such order as O =f(S), in which O = the organic variables and S = the stimulus variables; Ib = f(Ia), in which I = the intervening neurophysiological variables within the organism; and R = f(O), in which R = response variables, or measurement of behavior properties.

  * “Ouch” is a learned response. A German wouldn’t say “Ouch” but perhaps “Aie,” a Yiddish speaker “Oy.”

  * Or the NP-VP division of transformational linguists. Or Strawson’s division of a sentence into what you are talking about and what you are saying about it.

  * Nor are language couplings the only kind of couplings which occur. There are other kinds of symbols and other kinds of sentences, e.g., the coupling of a map with the territory, the coupling of van Gogh’s painting The Cypresses with what is symbolized (which is not merely the cypresses but forms of feeling as well). But here we are concerned primarily with language sentences.

  † In Chapter 9 I describe symbolusing behavior as characterized by a tetradic structure. Thus, if one were to observe an utterance of a symbol—or, as I would say here, of a sentence—one would notice that there is not only an utterer and a coupling of sentence elements, but also a listener or receiver of the sentence. “The second person is required as an element not merely in the genetic event of learning language but as the indispensable and enduring condition of all symbolic behavior. The very act of symbolic formulation, whether it be language, logic, art, or even thinking, is of its very nature a formulation for a someone else. Even Robinson Crusoe, writing in his journal after twenty years on the island, is nevertheless performing a through-and-through social and intersubjective act.”

  Today, ten years later, I would broaden the notion of coupling “symbol” and “object” to the utterance of sentences in general, whether symbol and object, naming sentences, or traditional declarative sentences with subject and predicate.

  This “tetradic behavior,” involving an utterer, a receiver, symbol and object, is contrasted with the “semiotic triangle” of Ogden a
nd Richards, involving a sign which affects an interpreter which in turn responds with behavior relevant to an object or referent.

  I find it convenient here, however, to observe Peirce’s distinction between dyadic relations and triadic relations. It will be seen that no substantial change has been made. What matters is the difference in “valence” between the semiotic relations encountered in symbol use and those in signal use, whether the difference is between triads and tetrads or dyads and triads.

  Thus, the “semiotic triangle of Ogden and Richards with its “causal” relations between sign and interpreter and between interpreter and referent is clearly, in Peirce’s scheme of things, a pair of dyads.

  The tetrad I proposed can, if one wishes to deal with atomic rather than molecular events, be split apart along its interface between utterer and receiver of a sentence, yielding a coupling of sentence elements by utterer and a subsequent coupling by receiver. The tetradic model, I see now, is appropriate only in successful communication, i.e., those transactions in which the same elements are coupled by both utterer and receiver and in the same mode of coupling. Unfortunately this is not always the case.

  In short, in Chapter 9 I deal with the “molecular” structure of the communication process, whereas I am here dealing with the “atomic” structure.

  * It is this transformation of symbols and their subsequent confusion with things that Count Korzybski used to rage against. “Whatever you choose to say about this object,” he would say, holding a pencil aloft, “don’t say ‘this is a pencil.’ “ “Whatever you say the object ‘is,’ well it is not” (p. 35).

  In point of fact, I have never seen anyone mistake a word for a thing or try to write with the word pencil, though the magic use of words undoubtedly occurs in primitive societies and perhaps an analogous misuse in modern technological societies.

  Korzybski tended to treat the peculiar features of symbol use as misbehavior to be gotten rid of by a therapeutic semantics which was almost an ethical science.

  In a triadic theory of meaning it is to be hoped that symbolic transformations and sentence couplings with the verb is will not be put down as instances of bad behavior or human stupidity but rather will be regarded as a fundamental property of sentence utterance.

  What needs to be explored is not human perversity as such but rather a parameter variable of symbol use. All sentences entail couplings. The mode of coupling is a normative dimension in which couplings may be used truly or falsely in propositions, well or badly in poetry, as a transparent vehicle of meaning or as an opaque simulacrum which distorts meaning.

  † Werner and Kaplan note that the word chair is not merely a sign or label for chairs: “…the material, phonemically unique sequence, ch-ai-r, is articulated into a production whose expressive features parallel those ingredients in the percept ‘chair.’…Only when the vocable has become imbedded in an organismic matrix, regulated and directed by an activity of schematizing or form-building, does it enter into a semantic correspondence with the object (referent) and does it become transformed from the status of a sign to that of symbolic vehicle.”

  * Cf. Braine: He and others have noted that an early stage of language acquisition in children features two-word utterances comprising a “pivot” word and an “open” word. Thus a child using the “pivot” word there might combine it with any number of “open” words and say there ball, there man, there doggie, etc. Then in a few months a second stage is reached in which the child combines two “open” words. Thus instead of saying there car or there man, the child might say man car, meaning “A man is in the car.”

  Braine noted a pause or juncture between the two “open” words. Thus baby chair or baby book, uttered without a juncture, is presumably a pivot-open construction meaning “(There is) a little chair” or “(There is) a little book.” Whereas the utterance baby#chair, uttered in a certain context, is reliably understood by the mother to mean “The baby is in his chair.” The symbol # represents a juncture or pause.

  This open-open construction is a very large class and represents, to my way of thinking, nothing less than the child’s graduation from the naming sentence (there ball) to the syntactical, “subject-predicate” sentence.

  Let us agree with Chomsky that a child’s linguistic behavior cannot possibly be accounted for by traditional learning theory with its notions of “stimulus control,” “conditioning,” “generalization and analogy,” “patterns,” “habit structures,” or “dispositions to respond.”

  The question, however, is whether the sole alternative to learning theory is Chomsky’s “innate ideas and innate principles,” specifically in this case a “language acquisition device,” a kind of magic black box interposed between input and output which contains not only the principles of universal grammar but the capacity of generating the grammar of one’s own language.

  I wonder whether Chomsky’s LAD (language acquisition device) is nothing more nor less than the unique human ability to couple sentence elements, to couple symbols with things, symbols with symbols, which couplings may be understood to mean whatever context allows them to mean.

  Indeed, may not grammar itself be defined as the primitive coupling plus whatever inflection, particles, and patterns may be required to supplant the diminishing context and the intuitive grasp by the mother of the child’s couplings? Thus the child’s sentence baby#chair may be understood infallibly by the mother to mean The baby is now in his chair. But as the intimate mother-child relationship declines and as it becomes necessary for people to talk to strangers over telephones about

  babies and chairs which at least one party cannot see, it becomes necessary to add such words as the, is, in, his, etc.

  If one must speak of a universal grammar, it is surely impossible to avoid the basic phenomenon of the sentence as a coupling and the basic division of couplings into two sorts, whether the language be English or Algonquin: (1) an object beheld by both speaker and hearer and pointed at and understood as one of a class of like objects and named by a sound which is understood as a class of like sounds—thus the pointing at and the utterance of the single-word sentence by father to son: balloon. (2) the coupling of symbol and symbol, e.g., baby#chair to signify vvhatever world relation or event is beheld in common by speaker and hearer.

  * According to Veatch, mathematical logicians habitually confuse logical relations with “real” relations—here we would say sentence relations with world relations. Veatch calls the sentence coupling “an intentional relation of identity.” Thus the relation of John to Bill asserted in the sentence John is larger than Bill is a world relation which can be expressed by the isomorphic form xRy. Mathematical logicians persist in setting forth the sentence in the form xRy, whereas in truth the sentence relation is of the form S is P.

  Lord Russell and the early Wittgenstein of the Tractatus believed that the sentence must be in some sense isomorphic with the fact asserted by the sentence. The later Wittgenstein changed his mind and came to believe that sentences were plays in a language game and could mean whatever they were used to mean.

  * Transactions between analyst and patient are especially open to sudden shifts of context, missing referring words, uncued worlds, since the rules of this language game require the patient to say “what comes to mind.”

  † Here again, the uncritical use of analogical terms has impeded inquiry into distinctively human modes of meaning. Thus, when instrumentalists like Dewey describe scientific research as socially useful activity like farming and marketing, they state a not very interesting similarity at the expense of a much more interesting difference. What concerns us here is how the farmer sees himself vis-à-vis the world, and how the scientist sees himself. The two are not necessarily the same.

  More interesting still is how the layman sees himself vis-à-vis the world of science. Is it possible, for example, for a layman to benefit in one sense from the goods and

  services of scientific technology while in another sense falling prey to
them, e.g., coming to see himself as a consumer of these same goods and services as a passive beneficiary of a more or less esoteric, not to say magic, enterprise? “They will soon come up with a cure for cancer,” one hears. The question is, Who is “they,” and how does the speaker see himself in relation to “them”?

  * Here I am making the case that sentence utterances are triadic events about dyadic events. My utterance Sodium reacts with water is a triadic event about a dyadic event.

  It is also true, of course, that a sentence utterance, a triadic event, can be about another sentence utterance, also a triadic event.

  Thus, a coupling can be about another coupling. A therapist makes an analysis of a patient’s dream, to which the patient replies, “That’s a lie!” The patient is making a coupling about the therapist’s coupling. Note that the patient’s sentence addresses itself to a normative dimension of the analyst’s sentence. Sentences about other sentences tend characteristically to be judgments about the norms of the latter. E.g.: “That’s a lousy painting,” “Nixon’s speech last night was not his best,” “Kennedy wowed them in Berlin,” “Stalin lied,” “That’s a bad metaphor,” “So that’s a sparrow. So what?”

  The only point is that a sentence coupling, being what it is, can be about anything whatever. Since the coupling China is larger than Japan is wholly unlike the relationship of China and Japan, it can assert that relationship. Note that a map cannot. A map is isomorphic but it asserts nothing, unless some assertory claim is appended, e.g., the signature of the cartographer.

  Note that those mathematical logicians who believe that propositions are isomorphic with the reality they refer to have found it necessary to invent another mark which shows that the propositional relation is asserted, e.g., Frege’s assertion mark.