If most psychiatrists were asked why they don’t pay much attention to the linguistic behavior, considered as such, of their patients, they might give two sorts of answers, both reasonable enough. One runs as follows: “Well, after all, I have to be more interested in what the patient is saying than in the words and syntax with which he says it. “ And if, like most of us, he has been exposed to the standard academic behavioral sciences, he might add, again reasonably enough: “Well, of course we know that conversation is a series of learned responses, but these are very subtle events, occurring mostly inside the head, and so there is not much we can say about them in the present state of knowledge.”
Both explanations are familiar, reasonable, and dispiriting. But what is chiefly remarkable about them is that they are contradictory. No one has ever explained how a psychiatrist can be said to be “responding” to a patient when he, the psychiatrist, listens to the patient tell a dream, understands what is said, and a year later writes a paper about it. To describe the psychiatrist’s behavior as a response is to use words loosely.
Charles Peirce was an unlucky man. His two most important ideas ran counter to the intellectual currents of his day, were embraced by his friends — and turned into something else. William James took one idea and turned it into a pragmatism which, whatever its value, is not the same thing as Peirce’s pragmaticism. Peirce’s triadic theory has been duly saluted by latter-day semioticists — and turned into a trivial instance of learning theory. Freud was lucky. The times were ready for him and he had good enemies. It is our friends we should beware of.
What follows does not pretend to offer the psychiatrist an adequate theory of language sprung whole and entire like Minerva from Jove’s head. It is offered as no more than a sample of another way of looking at things. I hope that it might either stimulate or irritate behavioral scientists toward the end that they will devise operational means of confirming or disconfirming these statements — or perhaps even launch more fruitful studies than this very tentative investigation. What follows is adapted freely from Peirce, with all credit to Peirce, and space will not be taken to set down what was originally Peirce and what are the adaptations. Here again Peirce was unlucky, in that his views on language were put forward as part of a metaphysic, i.e., a theory of reality, and in a language uncongenial to modern behavioral attitudes. To say so is not to put down Peirce’s metaphysic. But the problem here is to disentangle from the metaphysic those insights which are germane to a view of language as behavior.
First I shall give a brief statement of what I take to be Peirce’s theory of language considered as a natural phenomenon, i.e., not as a logic or a formal structure but as overt behavior open to scientific inquiry. There shall follow a loose list of postulates which I take to be implied by Peirce’s triadic theory of signs. These “postulates,” unlike the arbitrary postulates of a mathematical system, are empirical statements which are more or less self-evident. From them certain other statements can be deduced. Their value will depend both on the degree to which the postulates are open to confirmation and the usefulness of the deduced statements to such enterprises as the psychiatrist’s understanding of his own transactions with his patients.
Peirce believed that there are two kinds of natural phenomena. First there are those events which involve “dyadic relations,” such as obtain in the “physical forces…between pairs of particles.” The other kind of event entails “triadic relations”:
All dynamical action, or action of brute force, physical or psychical, either takes place between two subjects…or at any rate is a resultant of such action between pairs. But by “semiosis” I mean, on the contrary, an action, or influence, which is, or involves, a cooperation of three subjects, such as a sign, its object, and its interpretant, this tri-relative influence not being in any way resolvable into actions between pairs.
If A throws B away and B hits C in the eye, this event may be understood in terms of two dyadic relations, one between A and B, the other between B and C. But if A gives B to C, a genuine triadic relation exists. “Every genuine triadic relation involves meaning.” An index sign is part of a dyadic relation. An index refers to the object it denotes by virtue of really being affected by that object. Examples of indexes: a low barometer as an index of rain, the cry of warning of a driver to a pedestrian. A symbol, however, is something which stands to somebody for something in some respect or capacity. “The index is physically connected with its object…but the symbol is connected with its object by virtue of…the symbol-using mind.”
Dyadic events are, presumably, those energy exchanges conventionally studied by the natural sciences: subatomic particles colliding, chemical reactions, actions of force-fields on bodies, physical and chemical transactions across biological membranes, neuron discharges, etc.
Triadic events, on the other hand, characteristically involve symbols and symbol users. Moreover, a genuine triadic relation cannot be reduced to a series of dyadic relations. Peirce seems to be saying that when a symbol user receives a symbol as “meaning” such and such an object, we may not understand this event as a sequence of dyadic events or energy exchanges even though dyadic events and energy exchanges are involved: sound waves in air, excitation of sensory end-organ, afferent nerve impulse, electro colloidal synaptic event, efferent nerve impulse, muscle contraction, or glandular secretion.
Peirce’s distinction between dyadic and triadic behavior has been noted before, but so pervasive has been the influence of what might be called dyadic behaviorism that Peirce’s “triadic relation” has been recognized only to the degree that it can be set forth as a congeries of dyads. Morris, for example, interprets Peirce’s triad as implying that in addition to response and stimulus there is a third factor, a “reinforcing” state of affairs. This is like saying that Einstein’s special theory will be accepted only to the degree that it can be verified by Newtonian mechanics. Like Newtonian mechanics, dyadic theory can account for perhaps 98 per cent of natural phenomena. Unfortunately the phenomenon of talking-and-listening falls in the remaining 2 per cent.
What would happen if we took Peirce seriously? That is to say, if we retain the posture of behavioral science which interests itself only in the overt behavior of other organisms, what are we to make of observable behavior which cannot be understood as a series of dyadic energy transactions? What has happened in the past is that we have admitted of course that there is such a thing as symbol-mongering, as naming things, as uttering sentences which are true or false, as “rules” by which names are assigned and sentences formed. We have admitted that such activity is a natural phenomenon and as such is open to scientific investigation. But what kind of scientific investigation? We have gotten around the difficulty by treating the products of symbol-mongering formally, by what Carnap calls the formal sciences (logic, mathematics, syntax), while assigning the activity itself to a factual science, in this case learning theory, which has not, however, been able to give an account of it. It is no secret that learning theorists will have no truck with symbols and meaning. Most textbooks of psychology do not list the word symbol in their indexes. Indeed, how can learning theory, as we know it, give an account of symbolic activity? If we are to believe Peirce, it cannot. For the empirical laws of learning theory are formulations of dyadic events of the form R =f(O), in which R = response variables and O = stimulus variables.*