In self-control the alternative courses of action are specifiable in advance, and the issue is resolved before control is exerted. The techniques of control can be efficiently designed to achieve a particular state of affairs. There are instances of the manipulation of one’s own behavior, however, in which the outcome cannot be predicted. Some sort of “self-determination” is involved, for example, in deciding which of two courses of action is to be followed. The task is not simply to make a selected course of action probable but to decide an issue. The individual sometimes does this by manipulating some of the variables of which his behavior is a function. The techniques are more limited than in self-control because the outcome cannot be specified in advance.
In making a decision, as in self-control, the manipulated variables are often private events within the organism. As such they present a special problem, to which we shall return in Chapter XVII. Familiar instances in which the variables are accessible to everyone will suffice here. The processes appear to be the same whether the variables are public or private. “Making a decision” also resembles self-control in that some of the techniques are used in essentially the same way in controlling the behavior of others. This is not true when we persuade someone to behave in a given way since our variables operate in favor of a single alternative, and no decision is involved. When we attempt to help someone “make up his mind” without prejudice to any course of action, however, we employ the techniques which the individual may use upon himself in reaching a decision.
Although variables in the field of motivation and conditioning are used in making a decision, they are less specific and their effect is often delayed. For more direct results we resort to the manipulation of stimuli. If all relevant courses of action show some strength before we decide among them, our techniques consist of finding supplementary sources of strength which, when applied to the behavior of others, would be classified as prompting or probing (Chapter XIV). In deciding whether to spend our vacation in the mountains or at the seashore, for example, we may pore over travel magazines and vacation booklets, find out where our friends are going and what weather is predicted for each place, and so on. This material may, if we are unlucky, simply maintain the balance between the two courses of action, but it is more likely to lead to the prepotent emergence of one of them. “Deciding,” as the term will be used here, is not the execution of the act decided upon but the preliminary behavior responsible for it.
The process of deciding may come to an end before the act is executed when some relatively irrevocable step is taken—for example, we may decide about the vacation by making a down payment to hold a reservation. A common conclusion is simply to announce our decision. By saying that we are going to the seashore, we insure aversive consequences if this prediction of our future behavior is not fulfilled. The new variable may prevent the reinstatement of any conflict and hence of any further behavior of deciding. Deciding is also brought to an end when the techniques begin to be applied toward a single outcome—when we throw away the pamphlets describing the seashore and continue to work to strengthen the behavior of going to the mountains. We are then behaving as if we had been told to go to the mountains for our health and were simply accumulating material which made it possible to carry out the order (perhaps in competition with aversive variables which strengthened staying home or going elsewhere).
The individual manipulates relevant variables in making a decision because the behavior of doing so has certain reinforcing consequences. One of these is simply escape from indecision. Conflicting alternatives lead to an oscillation between incomplete forms of response which, by occupying a good deal of the individual’s time, may be strongly aversive. Any behavior which brings this conflict to an end will be positively reinforced. What we may distinguish as “due deliberation” has other consequences. When we look a situation over carefully in the course of making a decision, we presumably increase the probability that the response eventually made will achieve maximal reinforcement. In the long run the net gain may be enough to maintain the strength of the behavior of looking over the situation.
Escape from indecision or the net advantage of a deliberated response may seem inadequate to explain the origin and maintenance of the behavior of deciding. They are certainly defective reinforcers, for they may be long delayed and their connection with a response may be obscure. We may readily admit these deficiencies, however, for the behavior of making decisions is also usually deficient. It is not present in any degree in the behavior of lower organisms or of many people. When present, it is usually the result of special reinforcements applied by the community. Though the individual may accidentally hit upon various ways of deciding, it is more likely that he will be taught relevant techniques. We teach a child to “stop and think” and to “consider all the consequences” by supplying additional, and to some extent irrelevant or spurious, reinforcements (Chapter XXVI). Even these may not be successful. The child may still find it difficult to “make up his mind” and may occasionally experience the pathological condition of folie du doute or some version of the plight of Buridan’s ass.
In making a decision the alternative courses of action can be specified in advance, even though the outcome cannot be foreseen. Are there circumstances under which an individual manipulates variables to affect a response which he cannot identify until it is emitted? At first glance this may seem not only improbable but impossible. Nevertheless it is done—and done extensively. Let us suppose that we have forgotten the name of a man we must shortly introduce to someone. Since the response cannot be specified in advance, the usual techniques of self-determination may seem not to apply. There is, indeed, nothing we can do unless we have a lead of some sort. But not being able to identify a response does not mean that we cannot make other statements about it or manipulate conditions relevant to it. We may be able to say, for example, that it is a name we once knew, that it is a name which will be correct in introducing a particular person, that we shall probably recognize it at once as correct, or that it is the name of a man whom we met on a particular occasion and with whom we discussed a particular subject. With these extra specifications it is not impossible to work upon oneself in order to strengthen the response. The available techniques should be classified as self-probes (Chapter XIV). (A self-prompt would presuppose that we could identify the response.)
The techniques are familiar. We use thematic probes when we review a conversation we had with the man in question, when we describe the circumstances under which we were introduced to him, or when we review thematic classifications (was it a German name, an Irish name, an unusual name, and so on?). We use formal probes when we try various stress patterns—ta-da-ta-dada—or recite the alphabet repetitively in a form of verbal summation. We may even set up an aversive condition from which we can escape only by emitting the name. This is done in rehearsing a formal introduction—“I’d like to have you meet Mr . . . . ”—or by embarking upon the actual introduction, counting upon the powerful pressure which will arise when the appropriate point is reached to produce the name. If, as the result of any of these procedures, the name “suddenly pops into our head,” then a response has been strengthened which could not be specified in advance.
In recalling a name it is assumed that the response exists in some strength and that other information is available as a source of supplementary stimulation. These are the essential features of a broader and generally more complex activity commonly called “problem-solving,” “thinking,” or “reasoning.” The analysis of recalling a name thus serves as a preface to a much more important field of human behavior.
The language in which problem-solving is usually discussed does not differ much from the layman’s vocabulary. The rigorous concepts and methods developed in other areas of human behavior are commonly abandoned when this field is reached. It is easy to give an example of a problem, but it is difficult to define the term rigorously. There appears to be no problem for the organism which is not in a state of deprivation or aversive stimulation, but something more is involved. The hungry organism eating ravenously is perhaps disposing of a problem, but only in a trivial sense. In the true “problem situation” the organism has no behavior immediately available which will reduce the deprivation or provide escape from aversive stimulation. This condition may be expressed more generally. We need not specify the deprivation or aversive condition if we can demonstrate that a response exists in strength which cannot be emitted. Discriminative stimulation may be needed to determine the form or direction of the response (the golfer cannot shoot for the green until he finds the green); or the response may require external support or instrumentation which is lacking (the golfer cannot shoot for the green until he finds the ball). We may demonstrate the strength of the response in several ways but usually by showing that it occurs as soon as the occasion is suitable.
A locked drawer presents a problem if behavior requiring an open drawer is strong and if the individual does not have the key or other means of opening it. The strength of the behavior is inferred from the presence of responses which have previously opened the drawer or from the appearance of the behavior as soon as the drawer has been opened. We can say that a stalled car presents a problem if no behavior which succeeds in starting it is immediately available and if behavior which has previously succeeded in starting it is strong or if we have other evidence that behavior which depends upon a started car is strong. Interlocked wire rings are a problem if the behavior of demonstrating them apart is strong and no available response makes this possible. A murder mystery presents a problem if we are strongly inclined to name the murderer—to show that one name fits all statements in the story consistently—and cannot do so. Buying wallpaper for a room is a problem if we cannot say how many rolls we need; it is another type of problem if we have measured the room but have not converted our measurements into rolls of paper. Mathematics is rich in problems, but the motivation of the mathematician is often obscure. The deprivation or aversive stimulation responsible for the strength of writing a formula which always generates a prime number or of proving that a given formula never fails to generate a prime number is by no means clear.
In any case, the solution to a problem is simply a response which alters the situation so that the strong response can be emitted. Finding the key to the locked drawer, putting gasoline into the car, twisting the wire rings in a certain way, emitting a name which fits all the statements in the murder mystery, and writing a formula which always generates a prime number are solutions in this sense. Once the solution has occurred, the problem vanishes simply because the essential condition has been eliminated. (The same problem is not likely to recur since the situation will no longer be novel. Henceforth, the response which appeared as a solution will occur because it has been reinforced under similar circumstances.)
Simply emitting a solution, however, is not solving a problem. We are concerned here with the process of “finding the solution.” Problem-solving may be defined as any behavior which, through the manipulation of variables, makes the appearance of a solution more probable. This definition seems to embrace the activities most commonly described as problem-solving, and it permits a fairly rigorous analysis of procedures or techniques. We may solve the problems of other people in this way, but we shall limit the discussion here to the case in which the individual solves his own problems.
The appearance of a solution does not guarantee that problem-solving has taken place. An accidental change in the environment often brings about a similar result—the key may be found or the car suddenly responds to the starter. A more subtle example, which has already been mentioned, is Descartes’s explanation of the behavior of the living organism. The problem arose from a strong disposition to emit explanatory remarks concerning the operation of the living body. We must assume the strength of such behavior even though at this late date we cannot account for it. The explanation was a metaphor; a response based upon certain fountain figures which were constructed to resemble living organisms was simply extended through stimulus induction to the living organism itself. We need not suppose that at the moment this occurred, Descartes was engaged in solving the problem in any active sense. The availability of the information about the fountain figures may have been wholly accidental. We need not, therefore, treat any particular part of Descartes’s behavior as problem-solving. It was simply “hitting upon a solution.”
For the same reason, so-called trial-and-error learning is not problem-solving. The state of deprivation or aversive stimulation required by a problem implies the high probability of many responses. Some of these may be emitted because the situation resembles other situations in which they have been reinforced. It is possible that one of these will be a solution—that it will solve the problem by disposing of the essential condition. But this requires no special treatment. Another kind of behavior likely to be observed is random exploration. In the presence of a problem the organism is simply active. Here again the solution may follow by accident.
An example of problem-solving in the sense of finding a solution appears in connection with trial-and-error learning when the organism “learns how to try.” It emits responses in great numbers because of previous success and perhaps according to certain features of the problem. Suppose we challenge an individual to identify a word selected from a list. Our challenge provides aversive stimulation, and our statement that we have chosen a word from a particular list provides a discriminative stimulus increasing the probability of a corresponding set of responses. The individual’s only recourse is to emit words on the list until he hits upon the effective response. He may have discovered ways of ordering his behavior, however, to avoid repetition, to avoid omissions, and so on. He may progress rapidly toward a solution if we reinforce him with descriptive categories. He may then run through the alphabet for the initial letter (“Is it a word beginning with . . . ?”), then for the second letter, and so on. A formal prompt will soon be generated which will strengthen responses having a reasonable chance of success. Or he may guess thematic or grammatical categories—animal or vegetable, noun or verb, and so on. The approach to the solution may be very skillful when profitable categories have once been reinforced. But in spite of the fact that one learns to use such a technique and in spite of the apparent direction of the process, the behavior is scarcely more than a trial-and-error performance. We can account for the emergence of each trial response in terms of the current occasion and the past history of the individual. There is a minimum of “self-determination.”
One way to encourage the emission of a response which may prove to be a solution is to manipulate stimuli. A simple example is a survey of the problem situation. This is often the effect of random exploratory behavior and is therefore grouped loosely with trial-and-error learning. But the effect is not to emit a response which will prove to be a solution but to hit upon stimuli which may control such a response. Improving or amplifying available stimulation is especially effective; we increase the chances of a solution when we look a problem over carefully, when we get all the facts, or when we point up relevant stimuli by stating a problem in its clearest terms. A further step is to arrange or rearrange stimuli. In the game of anagrams, for example, the problem is to compose words from a miscellaneous assortment of letters; the solution is simply spelling out an acceptable word. It is helpful to rearrange the available letters since some arrangements may resemble parts of words in the individual’s repertoire and hence serve as formal prompts. The experienced anagram player learns to group letters efficiently, especially in certain subgroups which enable him to make profitable larger groups. He learns to put “q” and “u” together, to try various combinations of “sl,” “sh,” “sp,” “th,” and so on.
The logical syllogism is a way of arranging stimuli. The logician possesses a verbal repertoire in which certain conclusions are likely to be made upon the statement of certain premises, but a particular problem may not present itself in the required order. Solving the problem consists of arranging the materials in syllogistic form. If the solution is obtained wholly by applying a formula (Barbara celarent . . . ), the arrangement does not merely facilitate a response but actually determines it, and the process is not problem-solving as here defined. But there are less mechanical cases in which the arrangement is made primarily to encourage the appearance of a response which has other sources of strength. In the same way the mathematician is trained to transpose, factor, clear fractions, and so on, until an expression appears in a form which suggests a solution. Much of this may be relatively mechanical, but in true problem-solving the procedures are used to encourage the appearance of a novel response which has other sources of strength.
Scientific knowledge often advances as the result of the arrangement of stimuli. The Linnaean classification of species was an arrangement of data which led, among other results, to Darwin’s solution of the problem of the origin of species. Mendelyeev’s table of the elements was an arrangement of the data of chemistry which necessarily preceded modern atomic theory. The marshalling of relevant information is now so obvious a step in the solution of any problem that it is a matter of routine where problems are to be solved by groups and where the different functions of problem-solving are delegated to different people. The “fact researcher” is a familiar figure in the organized problem-solving of science and industry.
Another technique of problem-solving consists essentially of the self-probe. Tentative solutions, perhaps assembled for this purpose, are systematically reviewed. There are also certain practices which are not to be overlooked even though they are not directed toward specific solutions and hence are not ordinarily included in problem-solving. An example is a certain type of self-probe which is so general that it must be used repetitively in the manner of the verbal summator. Repetition is, of course, helpful in increasing the effect of more specific techniques, as when we repeatedly survey relevant material or restate a problem again and again. But something like a formal probe which has no specific reference to a given solution appears to be exemplified by people who can “think better” in a noisy or otherwise apparently distracting environment. Features of the noisy background appear to operate like speech patterns of the verbal summator to contribute strength to solutions. Visual materials in the form of ink blots, “doodlings,” or the ambiguous stimulation of a crystal ball contribute to some kinds of solutions.
The person who is skilled in “how to think” often manipulates his levels of deprivation. He may know how to generate interests relevant to a problem. He may generate an adequate energy level by arranging a satisfactory program of sleep or rest. He may arrange aversive schedules which keep his behavior at an efficient pace. He may follow a rigid routine to achieve the same result. Solving a problem may also be facilitated by eliminating responses which conflict with the solution. The techniques for doing this do not, of course, depend upon a particular solution. In recalling a name, for example, a wrong name may seem to stand in the way of the right one. Here the response to be controlled, the intruding response, can be identified, and any of the devices employed in weakening behavior described in Chapter XV may be used.
The “difficulty” of a problem is the availability of the response which constitutes the solution. We may not need to increase the strength very much. This is the case when the problem closely resembles an earlier one: the wire ring puzzle is like one which has previously been solved, the murder mystery uses a standard plot, and the scientific problem parallels a problem in another field. As the similarity with earlier instances increases, and with it the availability of an adequate response, a point is reached at which it is idle to speak of problem-solving at all. At the other extreme there may be little or nothing in the present situation which strengthens appropriate responses, and in this case the individual must industriously manipulate the variables of which his behavior is a function. If no behavior at all is available, no matter what is done by way of changing the variables, the problem is insoluble so far as he is concerned.
The result of solving a problem is the appearance of a solution in the form of a response. The response alters the situation so that the problem disappears. The relation between the preliminary behavior and the appearance of the solution is simply the relation between the manipulation of variables and the emission of a response. Until the functional relations in behavior had been analyzed, this could not be clearly understood; and meanwhile a great many fictional processes were invented. Conspicuous examples are the “thought processes” called thinking and reasoning. A functional analysis removes much of the mystery which surrounds these terms. We need not ask, for example, “where a solution comes from.” A solution is a response which exists in some strength in the repertoire of the individual, if the problem is soluble by him. The appearance of the response in his behavior is no more surprising than the appearance of any response in the behavior of any organism. It is either meaningless or idle to ask where the response resides until it summons strength enough to spring out into the open. We may also easily represent the activities by virtue of which the thinker gets an idea—at least so long as the behavior is overt. Special problems undoubtedly arise when it is not, but they are not peculiar to the analysis of thinking.
Instances have been described in which a mathematician abandons a problem after working on it for a long time, only to have the solution “pop into his head” quite unexpectedly at a later date. It is tempting to suppose that he has continued to work on the problem “unconsciously” and that his solution follows immediately upon some successful manipulation of variables. But variables will change automatically during a period of time. Variables which have interfered with a solution may grow weak, and supporting variables may turn up. We need not, therefore, suppose that any problem-solving occurred after overt work on the problem was dropped. The fact that the solution comes as a surprise to the individual himself does not alter this conclusion, We shall see in Chapter XVIII that genuine problem-solving may well take place when the individual himself cannot observe it, but many instances of “unconscious thought” can be accounted for simply as changes leading to a solution which ensue with the passage of time.
It is not only in problem-solving that one “suddenly has an idea” in the sense of emitting a response. In a metaphor, for example, we have seen that a response is evoked by a stimulus which shares only certain tenuous properties with the stimulus originally in control. One suddenly “sees the similarity” between repeated misfortune and the repeated assault of waves against a rocky coast in the sense that a response appropriate to the one is now made to the other. This may occur with or without external aid. The metaphor may “come to us” as we are speaking or writing, or we may “see the point” when someone else emits the transferred response. On a broader scale we “get new ideas from a book” in the sense that we acquire many responses to a situation which we did not possess before reading it. In this sense the book may “clear up our thinking” about a given situation.
We often manipulate materials in the world about us to generate “new ideas” when no well-defined problem is present. A child of six, playing with a badminton bird and a white rubber ball, put the ball in the feathered end of the bird. This “gave her an idea.” She began to lick the ball as if the whole assemblage were an ice cream cone and immediately spoke of it as such. There is nothing mysterious about this “act of thought.” The manipulative and verbal responses appropriate to an ice cream cone were brought out by similar geometrical features of the bird and ball. There was no significant problem; an idle manipulation of nature simply generated a novel pattern which, through stimulus induction, evoked a response characteristically in some strength in a child of six.
The artist may manipulate a medium simply to generate ideas in much the same way. It is true that he may mix or place colors on a palette or canvas to solve a specific problem—for example, that of producing a likeness. The trained artist has already solved some of the subsidiary problems and possesses a repertoire, similar to those discussed in Chapter VII, which generates patterns resembling the properties of the object to be copied. There may also be certain novel features in the object which call for the preliminary behavior which we should here designate as problem-solving. The artistic exploration of a medium may, however, proceed in the absence of any explicit problem. This behavior is most obvious when the task is delegated to mechanical devices. The artist may generate novel geometrical designs by following an arbitrary formula, such as that of “dynamic symmetry,” or by “doodling.” In the same way the writer may generate novel plots by manipulating stock characters in stock situations, just as the composer may generate new melodies or rhythms by changing the settings on a mechanical device or by manipulating symbols on paper or by allowing his cat to walk across the keyboard. All this may be done, not to solve a specific problem, but to enlarge an artistic repertoire. The general problem is simply to come up with something new.
We saw that self-control rests ultimately with the environmental variables which generate controlling behavior and, therefore, originates outside the organism. There is a parallel issue in the field of ideas. Is an idea ever original?
We do not call original that response which is obviously imitative or controlled by explicit verbal stimuli, as in following spoken or written instructions. We are not wholly inclined to call a response original, even though it has never been made before, when it is the result of some established procedure of manipulating variables—as in routine mathematical operations or the use of syllogistic formulae. When a pattern of manipulation has never been applied to a particular case before, the result is, in a sense, new. For example, the individual learns to count as the result of explicit educational reinforcement, but he may be original in what he counts. The observation that a cube has six faces must at one time have been an original idea.
We reserve the term “original” for those ideas which result from manipulations of variables which have not followed a rigid formula and in which the ideas have other sources of strength. A given procedure in problem-solving may never have been used in precisely the same way before or in connection with the same material, and it does not lead to the conclusion by itself. Some additional strength is supplied by stimulus induction from similar situations. This induction, however, is also the result of a particular personal history and of well-defined behavioral processes. We may, therefore, acknowledge the emergence of novel ideas, in the sense of responses never made before under the same circumstances, without implying any element of originality in the individuals who “have” them.
Man is now in much better control of the world than were his ancestors, and this suggests a progress in discovery and invention in which there appears to be a strong element of originality. But we could express this fact just as well by saying that the environment is now in better control of man. Reinforcing contingencies shape the behavior of the individual, and novel contingencies generate novel forms of behavior. Here, if anywhere, originality is to be found. As time passes men react to more and more subtle features of the world about them and in more and more effective ways. The accumulation of behavior is made possible by the growth of a social environment which forces modern man to respond to differences which only very slowly gained control of the behavior of his ancestors (Chapters XIX and XXVIII). Educational agencies established by the group provide for the transmission of the results of environmental contingencies from one individual to another, and it becomes possible for the individual to acquire effective behavior on a vast scale.
We cannot rigorously account for the origin of important ideas in the history of science because many relevant facts have long since become unavailable. The question of originality can be disposed of, however, by providing plausible accounts of the way in which a given idea might have occurred. The study of the history of science has made this task somewhat more feasible than it once seemed, since it has tended to minimize the contribution made by any one man. It is much easier to account for Harvey’s discovery that the blood passes from the right to the left ventricle by way of the lungs and not through the septum when we learn that the view had already been proposed that some of the blood passed this way. James Watt’s invention of the steam engine seems much less miraculous when we have once learned about the earlier forms of the engine upon which his contribution was based.
A formulation of creative thinking within the framework of a natural science may be offensive to those who prize their conception of the individual in control of the world about him (Chapter XXIX), but the formulation may have compensating advantages. So long as originality is identified with spontaneity or an absence of lawfulness in behavior, it appears to be a hopeless task to teach a man to be original or to influence his process of thinking in any important way. The present analysis should lead to an improvement in educational practices. If our account of thinking is essentially correct, there is no reason why we cannot teach a man how to think. There is also no reason why we cannot greatly improve methods of thinking to utilize the full potentialities of the thinking organism—whether this is the individual or the organized group or, indeed, the highly complex mechanical device.