by Morton Hunt
Out-of-Reach Bananas and Other Problems
Sultan, a male chimpanzee living in an anthropoid research center, has had nothing to eat all morning and he is hungry. His keeper lets him into a room where a bunch of bananas is hanging from the ceiling, out of reach. Sultan jumps toward the bananas a few times but comes nowhere near them. He then prowls around the cage, making discontented sounds. Some distance from where the bananas are hanging he comes upon a short stick and a large wooden box. He picks up the stick and tries to knock the bananas down, but they are too high. For a while, he bounces around, upset and angry; then, he suddenly rushes to the box, pulls it under the bananas, climbs up on it, and with a little jump seizes his prize.
Days later: The situation is the same except that now the bananas are hanging considerably higher. This time there is no stick, but there are two boxes, one larger than the other. Sultan knows what to do, or thinks he does. He drags the larger box under the bananas, climbs up, and crouches, as if to jump. But after looking up, he does not; the bananas are well beyond his range. He leaps down, seizes the smaller box, and, pulling it behind him, gallops around the room, shrieking in anger and kicking the walls. Plainly, he seized the second box not with the thought of putting it on the first one but merely to help vent his anger.
But all at once he stops shrieking, pulls the smaller box over to the other one, with some difficulty hoists it on top of the larger one, and climbs up. He has solved the problem. Wolfgang Köhler, watching and making copious notes, is deeply pleased. 17
Köhler conducted a series of studies of chimpanzee mentality between 1914 and 1920 that became almost as famous as Pavlov’s experiments with the salivating dog and Watson’s with little Albert. Not only were Köhler’s findings valuable in themselves but they led directly to similar studies of human problem solving by Gestalt psychologists that produced a number of significant discoveries.
The nature of the thinking involved in problem solving had interested philosophers and psychologists for much of the previous twenty-four centuries, but in Germany the subject had been out of fashion for some time. Like all higher-level mental processes, it lay outside the boundaries of scientific psychology as defined by the physiological psychologists and the Wundtians. In America, although William James and John Dewey had written about problem solving, Thorndike’s puzzle box experiments with cats had led many psychologists to regard it as the result of trial-and-error activity, even in human beings, rather than of conscious planning and problem solving.
Wertheimer, who in his formative years had read and admired Spinoza, took a different view: he believed in the power of the thinking mind. He was impressed, too, by the statements of Galileo and other great discoverers indicating that their breakthroughs often came from a new view of the problem that produced a sudden insight.
To illustrate how such a perception can produce a solution, Wertheimer liked to tell a little anecdote about Karl Gauss, the famous mathematician. When Gauss was six years old, his teacher asked the class who could first give the total of 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 10. In seconds young Gauss raised his hand. “How did you get it so quickly?” the teacher asked. Gauss said, “If I had to add one and two and three, and so on, it would have taken a long time, but one and ten are eleven, two and nine are eleven, three and eight are eleven, and so on— five elevens in all. The answer is fifty-five.” He had seen a structure that led instantly to a solution of the problem.18
Wertheimer was interested all his life in reasoning and problem solving, and in his last years wrote Productive Thinking (1945), a general discussion of the subject as seen by Gestalt psychology. But other Gestaltists, Köhler leading the way, did the bulk of the experimental work on the matter.
Köhler, after working with Wertheimer on the motion-illusion experiments, had stayed on at Frankfurt for another three years; then, at twenty-six, he was offered the post of director of the Prussian Academy of Sciences’ anthropoid research station on Tenerife, one of the Canary Islands, a Spanish possession off the coast of northwest Africa. Köhler shipped out in 1913, never imagining that a world war and chaotic post-war conditions in Germany would trap him there for over six years.
But he put the time to good use. He had been profoundly impressed by Wertheimer’s ideas and, as he later recalled, “had a feeling that his work might transform psychology, which was hardly a fascinating affair at the time, into a most lively study of basic human issues.”19 During his years on Tenerife these ideas were often on his mind, and his primate studies, although not formally couched in Gestalt terms, strikingly confirmed Gestalt theory as applied to problem solving. He pursued the investigation, with numerous variations and replications, for a number of years. Several British intelligence agents were convinced that he was a German spy, since no scientist would have spent so much time studying how apes get hard-to-reach bananas.20
(Ronald Ley, a psychologist at the State University of New York at Albany, recently spent nearly fifteen years trying to determine whether Köhler had been a spy. He gathered gossip and rumors from elderly Tenerifans, but neither on the island nor in Germany did he find any hard evidence bearing on the matter. Ley thinks Köhler may well have been a spy; other scholars doubt it.)
Köhler created a number of different problems for his apes to solve. The simplest were detour problems, in which the chimpanzees had to get to the bananas by a roundabout route; that gave them no trouble. More complicated were problems in which the chimpanzees had to use “tools” to reach bananas hanging out of reach—sticks with which they could knock them down, ladders they could lean against a wall (they never did figure out how to prop the ladders securely but always stood them sideways to the wall), and boxes.
Some of the chimpanzees took a long time to see that the boxes could be used to reach the bananas, and they never did use them well. Some would do ineffective things like piling up boxes where they happened to be rather than under the bananas, or stacking them so poorly that they toppled over when the chimpanzees tried to climb on them. Others, clearly smarter, did better, learning to stack boxes in a more secure fashion even when it took more than two boxes for them to reach the bananas. Grande, a female, actually was able, albeit with difficulty, to build a stack of four boxes when necessary.
Time and again, an ape would seem to suddenly see a solution at some juncture; Köhler interpreted this as a restructuring of the ape’s view of the situation. He called the sudden discovery “insight,” and defined it as “the appearance of a complete solution with reference to the whole layout of the problem,”21 obviously quite a different process from the trial-and-error learning of Thorndike’s cats.
Köhler thought the cats might have exhibited insight in a different kind of situation, but the puzzle box was a problem they could not solve through intelligence because it contained mechanical elements they could not see. But he did determine that insight thinking does not take place in simpler animals. He set up a fence at right angles to the wall of a house, with a segment at a right angle to its outer end, making an L. When he put a chicken inside the L and food outside, the chicken rushed back and forth along the fence, unable to recognize that by momentarily turning away from the food it could get around the end of the barrier. A dog, however, quickly sized up the situation and ran around to the food. A one-year-old girl, put inside the L and seeing her favorite doll on the other side, first tried to push through the fence but then laughed joyfully and toddled around the corner to it.22
With chimpanzees, some of the most dramatic instances of insight were elicited by another problem. Köhler would put an ape in a cage and a bunch of bananas outside, out of reach. In the cage would be several sticks; a chimpanzee might not realize for some time that it could reach the bananas with a stick but then all at once see that it could. A female chimpanzee, Tschego, first tried to reach the bananas with her hands and after half an hour got discouraged and lay down. But when a few other chimpanzees came into view outside the cage, she leaped to her feet, seized
a stick, and deftly pulled the bananas within reach. Apparently the sight of other apes nearing the food served as motivation and produced the click of insight.
In another stick problem, the moment of illumination was even more dramatic. Köhler’s account of it:
Sultan cannot reach the fruit, which lies outside, by means of his only available short stick. A longer stick is deposited outside the bars. [It] cannot be grasped with the hand, but it can be pulled within reach by means of the small stick. Sultan tries to reach the fruit with the smaller of the two sticks. Not succeeding, he tears at a piece of wire that projects from the netting of his cage, but that too is in vain. Then he gazes about him (there are always in the course of these tests some long pauses, during which the animals scrutinize the whole visible area). He suddenly picks up the little stick once more, goes up to the bars directly opposite the long stick, scratches it toward him with the “auxiliary,” seizes it, and goes with it to the point opposite the objective, which he secures.23
In an even more complicated problem, the bananas lay beyond reach with either of two available sticks; one of them, however, was thinner than the other and could be pushed into the thick one to combine their lengths. Even clever Sultan did not quickly see this solution. He spent about an hour trying to reach the fruit, to no avail; Köhler gave him a hint by sticking one of his own fingers into the end of a stick, but Sultan did not get the idea. Then:
Sultan squats indifferently on a box, which has been left standing a little back from the bars; then he gets up, picks up the two sticks, sits down again on the box, and plays carelessly with them. While doing this it happens that he finds himself holding one rod in either hand in such a way that they lie in a straight line. He pushes the thinner one a little way into the opening of the thicker, jumps up and is already on a run toward the bars, to which he had up to now half turned his back, and begins to draw a banana toward him with the double stick.24
One of Köhler’s most important findings, with sweeping implications for the psychology of learning, was that insight learning does not depend on rewards, as did the stimulus-response learning of Thorndike’s cats. The chimpanzees were, of course, seeking a reward, but their learning was not brought about by the reward; they solved the problem before eating the fruit.25
Another important finding was that when animals achieved an insight, they learned more than the solution to that particular problem; they were able to generalize and apply the solution in modified form to different problems.26 In psychological terms, insight learning is capable of “positive transfer”; in lay terms, the chimpanzees became test-wise.
Köhler reported his findings in a monograph in 1917 and in a book, The Mentality of Apes, in 1921. Both monograph and book made a considerable impression in the world of psychology, and not only as a study of animal problem solving; Köhler’s observations prepared the way for Gestaltist studies, using the same techniques, of human problem solving.
In 1928, a psychologist at Teachers College, Columbia University, used Köhler-type situations with children ranging in age from a year and a half to four years. Instead of bananas, the desirable objects were toys, which she placed out of reach, either outside the bars of a playpen or on a shelf. Sticks were available in the playpen experiment, and a chair and a box for the shelf. Sometimes the children showed immediate insight, and sometimes saw the solution only after a certain amount of fumbling around. The process was remarkably similar to what had taken place in the apes’ minds, although, not surprisingly, even these immature human beings were more insightful than the mature chimpanzees.27
Similar experiments with eight still younger children, ranging in age from eight to thirteen months, were conducted a little later by a young German psychologist, Karl Duncker, who had studied with both Wertheimer and Köhler at Berlin. He used a simple problem situation. The children sat at a table on which an attractive toy lay beyond reach; a stick was at hand. Only two children had insight almost at once; five others played with the stick until they either deliberately or accidentally moved it close to the toy, at which point they abruptly perceived the stick as a retrieval implement. The youngest child never solved the problem.28
Duncker’s more important work involved a series of problem-solving studies conducted between 1926 and 1935 with adult subjects. One of his research methods was to present a problem and ask the subject to think out loud as he tried to solve it; Duncker recorded what was said and then analyzed his “protocol” or written record to see how the subject formulated the problem and searched for a solution. This was one of the two problems:
Given a human being with an inoperable stomach tumor, and rays which destroy organic tissue at sufficient intensity, by what procedure can one free him of the tumor by these rays and at the same time avoid destroying the healthy tissue which surrounds it?29
A typical subject’s protocol (excerpted and much abbreviated here) read like this:
Send rays through the esophagus.
Expose the tumor by operating.
One ought to decrease the intensity of the rays on their way; for example—would this work?—turn the rays on at full strength only after the tumor has been reached.
Either the rays must enter the body or the tumor must come out.Perhaps one could alter the location of the tumor—but how?
Through pressure? No.
The intensity ought to be variable.
Adapt the healthy tissues by previous weak application of the rays.
I see no more than two possibilities: either to protect the body or to make the rays harmless.
(Experimenter: How could one decrease the intensity of the rays en route [as you suggested earlier]?)
Somehow divert… diffuse rays… disperse… stop! Send a broad and weak bundle of rays through a lens in such a way that the tumor lies at the focal point and thus receives intensive radiation. 30
This protocol and others showed that when faced with such a problem, people use a number of different heuristic (exploratory) techniques. Most often they start with mechanical or routine heuristics, such as trying random possibilities based on the most immediate or obvious characteristics of the problem; such heuristics usually yield poor solutions or none. In the above protocol, sending rays through the esophagus or exposing the tumor by operating are efforts of this kind.
Eventually, after reaching a number of dead ends, many subjects get around to more productive “functional” heuristics (a few others use them from the outset), such as trying to identify the essential properties of the problem. They ask themselves, for instance, what the fundamental goal is, and only then do they look for a specific solution. In the above protocol, the subject began thinking this way when he said, “One ought to decrease the intensity of the rays on their way.” He then reverted to the first kind of thinking (“Perhaps one could alter the location of the tumor”), but after the experimenter reminded him of his more basic heuristic he suddenly had his dramatic insight into a feasible solution. The mechanical heuristics are analogous to the chicken’s running back and forth along the fence, the functional heuristics to looking at the situation in broad perspective and seeing a less direct but effective way of reaching the goal.
Duncker’s other principal research method was to bring a subject into a room where a jumbled array of objects and materials lay on a table and then ask him to perform a task for which none of the objects seemed suitable. The goal was to see under what conditions the subject would weigh other possible uses of one or more of the available objects and under what conditions such restructuring was inhibited.
In one situation, for instance, the subject was asked to mount three small candles on the door at eye level, ostensibly for “visual experiments.” On the table were some candles, a few tacks, paper clips, pieces of paper, string, pencils, and some other objects, including the crucial ones: three small empty cardboard boxes. After fumbling around, every subject eventually restructured his view of the things at hand and saw that the boxes could be tacked to
the door and used as little platforms to mount the candles on.
But in another version of the problem, the three boxes were filled, one with little candles, the second with tacks, and the third with matches. This time, fewer than half of Duncker’s subjects solved the problem. They had seen the boxes being used for a specific purpose, and that made it harder to see them as usable in an unboxlike way.31
Duncker called this common and serious impediment to problem solving “functional fixedness”; when a problem solver sees an object as having a specific function, it is far more difficult for him to see it as serving any other function.
This was a noteworthy discovery. It explains why so often the very people who know most about any subject are the least likely to find a good solution to a new problem in their field. Education creates expertise but also functional fixedness. An expert sees the tools he has at hand in terms of the functions he knows they serve; a neophyte may, while coming up with uninformed and even absurd suggestions, see them more creatively. It is no accident that scientists generally make their most original and important contributions early in life.
Duncker, thought by many to have been the most brilliant of the Gestalt group in the 1930s, might have gone much farther with his investigation of problem solving had he not died tragically early. A political liberal, he fled from Germany in 1935 and went first to England and, in 1938, to the United States to teach at Swarthmore. In 1940, at thirty-seven, deeply depressed by the outbreak of war, he committed suicide.
The studies of problem solving by Köhler, Duncker, and other Gestalt psychologists look relatively simple but their implications were profound. They demonstrated that problem solving in human beings (and to some extent in animals) is not limited to trial and error and to conditioned responses but often involves certain kinds of higher-level thinking that produce new vision, thoughts, and solutions. The studies of problem solving were one of the most important ways in which the Gestaltists restored mind as the central concern of psychology.
