CHAPTER 1
Introduction
DISCOVERY VERSUS UNDERSTANDING
Science promises discovery. Great industries and great nations spend huge sums of money hoping that further discoveries will make them richer and more powerful. Publishers and broadcasters retail news to audiences eager to share in the adventure and the profits of scientific discovery. Philosophy promises understanding. But, what makes us say that we understand this or that idea? What makes us say that students understand? Is it that they can repeat back what they have heard or read, that they can pass examinations, that they say they understand? Those who promise adventure, profit, power, must have something concrete to offer: mushroom-shaped clouds, moon rocks, chess-playing machines. Where philosophy can be lofty and noble, science must be crass and materialistic.
Scientists must work with practical possibilities rather than ideals. All experiments are flawed, but they are not all equally flawed. Some are much better than others. There are significant degrees of difference between disciplined, publicly verifiable observation and ordinary, subjective experience. Several hundred years of scientific history have demonstrated that each level of improvement is well worth the effort. The whole structure rests on operational definitions.
BELIEF VERSUS EVIDENCE
Early in the 20th century, Robert H. Goddard patented the basic rocket design that remains in use to this day. In a 1919 report, Goddard described the principles of rocket travel in outer space, and in 1920 the Smithsonian Institution, which had financed some of Goddard’s work, summarized the gist of this article in a press release. Journalists were skeptical. An editorial in The New York Times reminded the public that it would be impossible for a rocket to function in a vacuum, and dismissed Goddard with, “Of course, he only seems to lack the knowledge ladled out daily in high schools.” To its credit the Times eventually printed a retraction, but not until 1969 as U.S. astronauts prepared for the first moon landing (Wohleber, 1996).
Before he wrote this article, Goddard first proved that rocket engines could operate in a vacuum by testing an actual engine in a vacuum chamber. This early model not only worked in a vacuum, but it worked 20% better in the vacuum than in air. Goddard was well aware of the traditional belief that engines cannot work in a vacuum, but he settled the question with experimental evidence rather than plausible arguments. The chief objective of this book is to offer tools for distinguishing between belief and evidence.
OPERATIONAL DEFINITION
Goddard did much more than assert that his rocket engine could work in a vacuum. He described his experiments in detail: how he created a vacuum in his experimental chamber, how he measured the vacuum, how he measured the output of his engine, and so on. He described his experimental operations in such detail that other scientists could evaluate his evidence and repeat his experiments. They could reach their own conclusions. Operational definitions tie theories and explanations to observation and experiment.
Facial Vision
Can blind human beings locate objects at a distance? For hundreds, perhaps thousands, of years natural philosophers debated this question. Many cultures believe that losses such as blindness stimuluate compensation. To skeptics, compensation seems like a sentimental denial of the harshness of nature and the principle of survival through fitness. Believers retold dramatic stories about the remarkable abilities of blind people. Skeptics found it easy to discount such informal, unverified anecdotes. Without direct experimental tests, the debate could have continued for hundreds of years more.
The dispute ended in 1944 when Dallenbach and his associates placed a 4-ft wide by 7-ft high Masonite screen at random places across the path of blind subjects. The path ran lengthwise through a 60 × 20 foot corridor. The blind subjects were asked to signal by raising one arm when they first detected the screen, and to stop walking and raise the other arm when they were about to collide with it. There were also catch trials without any obstacle. Blind subjects were quite good at detecting and locating the screen. Next, normal-sighted student volunteers attempted the same task while blindfolded. They failed at first, but within about 20 trials they, also, could detect and locate the obstacle and their performance approached the accuracy of the blind subjects after a few hours of practice (Supa, Cotzin, & Dallenbach, 1944).
The experimenters asked both the blind and the sighted subjects to explain how they detected the screen. About half of the blind and the sighted subjects were convinced that they felt it on the skin of their faces—their foreheads or their cheeks. This agreed with the consensus of blind people in centuries of introspective inquiry into this phenomenon. In fact, the phenomenon had come to be known as “facial vision.”
Further experiments showed that blind and blindfolded people could do just as well when their heads were covered with hoods, so long as their ears were exposed. When their faces were bare, but their ears were stopped, they failed. Both blind and blindfolded subjects also failed when the corridor was covered with a thick carpet and they had to walk without shoes. Eventually, the experimenters devised an apparatus that emitted artificial tones and picked up echoes with a microphone. They suspended this contraption from an overhead track so that it could be moved through the corridor by remote control. Under these conditions, subjects could control the movements of the apparatus and locate the obstacles from another room by listening with headphones to the sounds received by the microphone (Cotzin & Dallenbach, 1950).
Dependent Variable
Notice the difference between armchair argument about whether or not a blind person can locate objects at a distance, and an operational test. Dallenbach and his associates devised a way of observing—and better still, measuring—the alleged abilities of blind human beings. It is always the most critical, and often the most difficult, step in any experiment.
Clearly, the responses of the subjects—raising one arm when they first detected the screen, stopping and raising the other arm when they were about to collide with it—depended on their ability to locate an object at a distance without sight. In general, the dependent variable in a behavioral experiment is some form of observable response.
Independent Variable
An independent variable is a variable that the dependent variable depends on. The terms mean just what they seem to mean. About half of the blind and blindfolded subjects claimed that they felt the screen on the skin of their cheeks or foreheads. When the experimenters eliminated that source of information with thick hoods, subjects could still locate the screen as long as their ears remained outside of the hoods. This eliminated skin sensation as an independent variable. When the experimenters eliminated hearing with ear stoppers, the subjects failed to detect the obstacles, even when their faces were completely exposed. This identifies hearing as the sense that the ability depends on, and further eliminates facial sensations. The subjects also failed when they had to remove their shoes and walk over a carpeted floor. This identifies the sound of footsteps as the sound that the ability depends on.
Experimental Control
To isolate the effect of an independent variable, everything else must be held constant. In the case of hearing, the experimenters deafened the subjects with ear stoppers while keeping all other aspects of the task constant. By controlling for all other independent variables the experimenters isolated hearing as an independent variable.
Experimental control defines the independent variable. Control conditions ask the questions that advertising slogans want you to forget. When the ad says that a brand of soap flakes “washes clothes whiter,” you are not supposed to ask, “Whiter than what?” Suppose the ad only means whiter than no soap at all, or whiter than coal dust. The control group defines the “what” in whiter than what, faster than what, and so on.
Introspection
Both the blind and the sighted university students learned to locate obstacles in the dark by listening to the echoes of their own footsteps. It is hard to believe that they accomplished this demanding task without any awareness of which sense organ they were using. Blind people down through the ages solved the same problem for themselves outside of the laboratory, most of them also unaware of which sense organ they were using. They could do it even when their mental imagery of facial vision was quite vivid. One of the blind subjects, for example, described faint shadows on his face that became clear and sharp as he approached an obstacle. That so many agreed in their reports of facial images, and that they were so firmly convinced that they used facial images, is a very interesting phenomenon. And yet, neither the numbers of those who agreed nor the strength of their convictions are evidence for the validity or even the relevance of their subjective reports.
Intervening Variable
Usually, the independent variable in expermental psychology is some sort of stimulus input arranged by the experimenter, and the dependent variable is some sort of response output of the subject. In an experiment in learning, the independent variable is often a series of trials that can be read on the horizontal axis of a graph as in Fig. 1.1, and the dependent variable, which rises or falls as a function of the independent variable, usually appears on the vertical axis of a graph as also shown in Fig. 1.1.
FIG. 1.1. A curve of percent response plotted against trials in a typical graph of a learning experiment. Copyright © 1997 by R. Allen Gardner.
Intervening variables are relationships between independent and dependent variables. Suppose that Fig. 1.1 plotted the results of an experiment in classical conditioning (chap. 2). On each trial the experimenter lighted a light and then gave the dog some food. The experimenter measured the number of drops of saliva collected in a test tube on each trial, after lighting the light, but before presenting the food. Figure 1.1 would then show how the number of drops of saliva increased from trial to trial. This is called a learning curve.
Where is the learning in this curve? Obviously, there is no learning in the saliva in the test tube. Equally obviously, there is no learning in the light or in the food. A single trial cannot show any evidence of learning. The only evidence of learning is in the relationship between trials. We can see and measure the saliva, and we can see and measure the light and the food and the time between them, but we can only infer the relationship from trial to trial by looking at the record and making inferences.
Unseen and unseeable relationships are commonplace in all sciences. Suppose you hold a pencil out at arm’s reach and then release it. The pencil usually drops in this experiment. Why does it drop instead of drifting gently up to the sky, or just floating there in the air? The phenomenon is called gravity, of course. Where is the gravity in this experiment? You cannot see gravity any more than you can see learning. The gravity is not in the pencil, and it is not in the floor. There is gravity in the relationship between the pencil and the earth, and all that anyone can see or measure is the rate of fall of the pencil. Once again, a single observation of the position of the pencil at any point in time or space cannot show evidence of gravity.
All sciences must deal with phenomena that can only be inferred from a series of observations. Intervening variables involve additional problems of operational definition over and above the definitions of directly observable dependent and independent variables. These problems of operational definition are straightforward, but the rules of operational definition must be obeyed.
Figure 1.1 is a typical learning curve. The vertical axis represents some measure of response, usually related to probability. The horizontal axis represents some time series, usually trials. Notice that the level of response is greater than zero at the start of learning, even before the first food on the first trial. Anybody who is at all familiar with dogs knows that they are practically always salivating. Pavlov never had to teach any dog to salivate. All he could do was to increase the rate of salivation at certain times. This is typical of all learning situations. The to-be-learned response is always something that the subject does in the learning situation before the start of learning. Otherwise, it could never be affected by experience.
Notice that the curve approaches 100% but never quite reaches it. This is also typical of learning curves. As well as a habit may be learned, behavior is never perfectly stereotyped. Some variation always remains. Without this variability, nothing new could ever be learned and all habits would be permanent.
The curve in Fig. 1.1 is also typical in that it rises rapidly at first and then rises more and more slowly as it approaches 100%. This is called negative acceleration.
Experiment Versus Correlation
Since the 19th century, experts have advised parents against picking up a crying baby. They argue that picking up babies and cuddling them only rewards crying. The baby will learn to get affection and attention by crying and become a whimpering, cranky child. This is very difficult advice to take. Most parents cannot resist picking up a crying baby, even when they believe the traditional advice. A baby’s cry seems to evoke cuddling and comforting.
Ainsworth and her associates (Ainsworth & Bell, 1977) produced the first actual evidence on this subject by observing parents in their homes and recording how quickly they picked up newborn infants. These investigators found that the babies who were picked up sooner actually cried less at the end of their first year than the children who were picked up later. Not only did the babies who were picked up sooner cry less, but they started to speak earlier. Many parents who picked up their crying infants in spite of the conventional advice have felt vindicated by this result. But, does this study really prove that it is better to pick up a crying baby?
Is it likely that the kind of mother or father who can let their baby cry piteously without responding is the same sort of person as the mother or father that rushes in with comfort at the first sign of distress? It seems very likely that we have two different human types here, one that is more nurturing than the other. If that is true, then it is also likely that nurturance is a genetic trait passed on from parent to child. Nurturing parents might very well have nurturing children who calm down sooner and are more sensitive, responsive, and communicative than the offspring of nonnurturing parents. Perhaps, if nurturing parents could only control themselves and refrain from picking up their squalling infants for a few years, their offspring would be even calmer and more communicative.
This argument can only be settled with an experiment. Correlational studies can never answer questions of this kind. To eliminate (control for) genetic factors, experimenters have to manipulate conditions. One possibility is to switch half of the babies at birth and have half the nurturing parents raise the children of nonnurturing parents, while half of the nonnurturing parents raise the children of nurturing parents. The other half of each group would raise their own children. In that way we could see if picking up the crying infants made them calmer and more communicative whether or not they had nurturing genes—and vice versa, if letting them lie crying in their cribs made them more cranky and less communicative whether or not they had nonnurturing genes. This experimental design is called cross-fostering. The object of cross-fostering is to distinguish between two independent variables—in this case inheritable nurturing traits of the parents as distinct from the nurturing behavior of the parents.
Alas, it seems very unlikely that we could get a significant sample of parents to cooperate in this fascinating experiment in child rearing. The experiment is also flawed with respect to the central question of whether or not to pick up a crying baby. Very likely nurturing parents have a whole pattern of nurturing behaviors, and picking up the crying infant is only a part of their nurturing parental style. The cross-fostering experiment is too general to answer a question as specific as the one we have here. To test the specific question, the experimenter should let nurturing parents treat their babies as they wish in every way except that half must pick them up as soon as they start crying and half must let them cry for some specified interval, say 15 minutes, before picking them up. The nonnurturing parents would be divided in half in the same way.
Despite its logical advantages as an operational definition of the independent variable, the more specific experiment may be quite as impractical as the cross-fostering experiment. Often in a natural science, particularly a biological science, the necessary experimental conditions are so impractical that critical questions remain unanswered. Practical difficulties have a way of resolving themselves with time, however. Questions about what is on the other side of the moon, for example, were once thought to be unanswerable. Also, bright young minds often discover practical solutions to problems that baffled their elders for generations.
Adaptive Versus Maladaptive
The staggering variety of life forms on our planet is plainly related to specialized adaptations for different ecological niches. The hovering flight of the hummingbird and the treetop feeding of the giraffe are familiar examples. Increasingly specialized wings, hooves, necks, and so on, appear clearly in the fossil record. The geological record also tells a tale of flux and change. Inevitably, the environment changes, and those creatures that are most finely tuned to the environment of their ancestors perish when conditions change. The paths of specialization are also paths to extinction. The ancestors of the giant panda lived in a time when bamboo shoots were plentiful. In that place, and in those times it was highly adaptive to specialize in that one kind of food. Now that bamboo is scarce, the panda is rapidly approaching extinction.
Human beings are remarkably unspecialized. We live in the Arctic and at the Equator, in the mountains and by the sea, in deserts and in marshlands, and we eat just about everything. We are the least specialized of species and that is probably why we are overrunning the earth. Our ability to learn to alter our habits with changing conditions is a vitally adaptive asset. Nevertheless, anyone old enough to read this book probably has already acquired a long list of bad habits. This baggage of maladaptive habits that most adult human beings carry with them is quite as well learned as their adaptive habits—often, maddeningly better learned. Clearly, the mechanisms of learning fail to protect us from acquiring maladaptive habits. Any general theory of learning must account for maladaptive habits as well as for adaptive habits.
ETHOLOGICAL EXPERIMENTS
Color Vision
The experience of color is plainly subjective. Isaac Newton, who discovered the relationship between the physics of light and the experience of color, said it best: “the Rays to speak properly are not coloured” (1730/1952, p. 124). Color vision must be somewhere in the beholder. Although about 2 in 25 European men have deficient color vision, color blindness was unknown to the scientific world until 1798 when John Dalton described his own anomalous experiences. Moreover, that brilliant scientist was in his mid-20s before he realized that his visual world was severely abnormal.
How about other animals? Do they also have color vision? Early naturalists attempted to answer this question by looking at the habits and neurology of other species to see whether they were similar enough to the human case to suppose that this or that nonhuman being might see the world in color, as humans see it. The habits of honeybees, for example, seem to indicate that they distinguish flowers on the basis of color, but the anatomy and neurology of their eyes and brains are far from human. Some argued that honeybees only seem to be distinguishing the color of the flowers and that simpler explanations could account for their behavior. Others argued that the apparent color vision of the honeybee is only another case of anthropomorphism, of projecting human traits into other animals. Still others argued for the simplicity of color vision itself, that it might not require organs as complex as the human eye and brain. There the matter stood, perennial fuel for debate, until the rigorous and precise experiments of von Frisch early in this century.
Von Frisch showed that bees not only discriminate colors, but that they sort them into groups of primary colors the way human beings do, although they may group them more crudely than humans. He also showed that the spectrum of colors visible to the honeybee is shifted away from the red and into the ultraviolet; they confuse red with black and see ultraviolet as a distinct color. Plainly, the cleverest philosophers could never have achieved these discoveries by deep thought on the subject. Von Frisch had to devise measures of color vision for bees, by operational definition (Lindauer, 1961; von Frisch, 1950).
Before von Frisch, experimenters had found that bees without previous experience find some colors much more attractive than others. If colored patches are laid out in a field in a random array, many more bees investigate patches with colors near the middle of their visible spectrum and hardly any investigate colors near the ends of their visible spectrum. This only proves that the honeybees have species-specific color preferences. Scientists of the early 20th century concluded that true color vision was obviously beyond the capacity of such a tiny brain. Of course, human beings also have species-specific color preferences and this hardly proves that they are color-blind.
Von Frisch made history by laying out squares of colored paper on a table in a meadow. He placed a small dish on each square. All of the dishes were empty except for one dish that he baited with a small amount of syrup. Honeybees soon found the syrup and returned to the same dish time after time. When von Frisch randomly rearranged the squares on the table, the bees returned to the same colored square, even if he washed out all of the dishes so that the only remaining cue was the color of the square (von Frisch, 1950). Von Frisch had devised an operational definition of color vision in honeybees.
Von Frisch demonstrated that he could manipulate the color preferences of bees arbitrarily. That is, he could teach them to approach any visible color that he chose. A stimulus that is arbitrarily interchangeable with other stimuli in this sense is an arbitrary stimulus, a useful term that plays a central role in the descriptions of this book.
Honeybees can remember more than the color of flowers; they can also remember some flower shapes (Gould, 1985; Plowright, 1997). Thus, operational definitions derived from careful experiments permit us to speak with some confidence about the perception of a creature as alien to human beings a honeybee.
Further experiments show that when they return to the hive, these simple creatures with their tiny brains routinely communicate to other bees about distant sources of food. Rather than pinpointing precise targets, scouts only indicate a promising area to forage. Followers are free to take advantage of other targets that may present themselves, as when a change of wind across the flight path carries the scent of richer or closer forage (Gould, 1975; Gould, Henerey, & MacLeod, 1970).
When bees swarm, they form a mass clinging to a convenient branch while scouts search out a favorable site for the new hive. Returning scouts dance on the surface of the mass to communicate their findings. Sometimes their reports disagree. “Prime location, 100 meters due south,” dances one scout. “Spectacular site 75 meters north by northwest,” dances another. Each scout recruits followers who reconnoiter the proposed sites for themselves and return to join the debate. The swarm remains until there is consensus, with disastrous results if the debate goes on too long (Lindauer, 1961, pp. 34–58).
Arbitrary and Obligatory
Honeybees investigate patches of color. It is a prominent response that generations of human beings observed long before von Frisch arrived on the scene. Bees are more likely to investigate some colors than others and most bees have approximately the same color preferences. No one had to teach them that. The arbitrary effects of experience, however dramatic, are always superimposed on obligatory species-specific patterns of response.
Human beings in widely separated communities all over the world have the same repertoire of facial expressions. The same expressions of anger, disgust, happiness, and sadness are associated with the same types of affective stimuli, but the particular stimuli that evoke these responses, the stimulus events as well as the language in which the events are described, depend on cultural experience. Ekman (1973) demonstrated this by trekking to a remote village in New Guinea where all of the inhabitants had formerly been so isolated that they had never met any foreigner nor had they met anyone else who had ever met a foreigner. Ekman asked these people to make the facial expressions for four basic emotions and photographed their responses. He also showed them photographs of American college students who had been asked to make the same facial expressions. The New Guinea villagers named the facial expressions in the American photographs easily and accurately and, later, American college students named the facial expressions in the New Guinea photographs, also easily and accurately. Ekman’s most difficult technical problem (apart from finding a sufficiently isolated village and getting there) was translating the descriptions of basic emotional experience from one language to another.
Not only the pairing of events and emotions, but also the social conditions in which some responses are suppressed and others counterfeited, depend on social learning. Facial expressions of anger and disgust seemed to be absent in native Japanese and, indeed, these expressions fail to appear in ordinary social situations. They appeared quite readily, however, when Ekman (1973) secretly photographed the same Japanese while they were alone in the dark watching films of emotionally charged material, such as surgery.
Along the same lines, young chimpanzees, born in American laboratories and cross-fostered by human beings, have the same repertoire of emotional vocalization observed in young chimpanzees reared by their own mothers under wild conditions in Africa. Positive events and situations evoke positive vocalizations such as pants and grunts, while negative events evoke negative vocalizations, such as whimpers, whether the infant chimpanzees are reared by human psychologists in a U.S. suburb or by wild chimpanzee mothers in an African forest. But, for the cross-fostered infants, the cookies or ice cream that evoked pleasure grunts and the bath preparations that evoked whimpers can only have acquired their affective values through experience (R. A. Gardner, B. T. Gardner, & Drumm, 1989).
Whether we are interested in species identity or gender identity, eating habits or mating habits, survival skills or drug addictions, we must study the obligatory consequences of species membership as well as the arbitrary effects of individual experience.
Obligatory Patterns
The traditional view acknowledges the existence of obligatory species-specific responses while claiming that there is also a contrasting class of arbitrary behaviors that are “malleable” or “plastic” so that they can be “modified” or “shaped” by consequences, in Skinner’s words “as the sculptor shapes his figure from a lump of clay” (1953, p. 93). Pressing a lever, riding a bicycle, playing a violin all look quite arbitrary, but only if we treat them as acts defined by their effects on particular human artifacts and ignore the patterns of behavior that go into the acts.
Popular animal shows amaze audiences with bears riding bicycles and porpoises walking backward on their tails, but knowledgeable ethologists point out how skillful trainers systematically capitalize on the species-specific patterns of behavior that go into the amazing acts (cf. Hediger, 1955). Male bears fight by striking with their forepaws while standing errect on their hind legs. They must balance on two legs to engage in this species-specific pattern. Circus trainers claim that bears transfer this two-legged pattern to “dancing” to music or riding a bycicle. Backward tail walking appears spontaneously in untrained porpoises (Pryor, Haag, & O’Reilly, 1969). The dancing of the famous Lipizaner stallions is their species-specific mating display set to music. The arbitrary effect of experience lies in the connection of obligatory patterns of response to arbitrary patterns of stimulation.
Observation and Experiment
By the end of the 19th century, field naturalists formed a prominent group within biology. In the 20th century, behavioral biologists led by Tinbergen, Lorenz, and von Frisch and their followers carried on this tradition as ethology. They studied the stream of ongoing behavior under natural conditions in great detail.
When a bird is sitting on its eggs you know that it is incubating. When it delivers heavy wing-blows at another bird you know that it is fighting. But when two gulls, standing opposite each other, are pecking into the soil, or are tearing at grass tufts, or are walking around each other with the neck drawn in and the bill pointing upward, you have not the slightest idea what they are doing, at least at first. (Tinbergen, 1953a, p. 51)
Tinbergen and his students found that male gulls attack and retreat at the borders of disputed territory but, at the same time, they engage in apparently irrelevant behavior such as pulling up clumps of grass and preening their feathers. A prospective female mate approaches the borders to be met with a mixture of courtship and defense. These are all patterns that ethologists must observe directly and describe in detail.
Field studies continue to enrich science and culture. Goodall’s (1986) monumental study of the chimpanzees of East Africa is an inspiring recent example. Field studies emphasize obligatory behavior unfolding under natural conditions.
Species-specific behavior, the domain of field naturalists, used to be called instinctive or inborn as opposed to learned by experience. For example, the mating songs of many birds are so species-specific that an ornithologist can identify a bird from its song without ever seeing a feather or a beak. English chaffinches are a typical example. Normally, chaffinches hatch in the spring and migrate south for the winter. During that first spring the immature males never sing, but they hear the normal songs of adult males. Adult males sing in the spring and summer. The young males begin to sing the chaffinch mating song the following spring when they return to the breeding grounds and become sexually mature. Thorpe (1961) gathered chaffinch eggs in the field before they hatched and raised the young birds in his laboratory in isolated cages. Deprived of experience with adult male songs, the isolated males failed to sing normal chaffinch songs the next spring when they became sexually mature. A whole field of research grew out of Thorpe’s discoveries (King & West, 1989; Slater & Williams, 1994; West, King, & Freeberg, 1997). Many birds have to learn vital elements of their species-specific songs. Rather than asking whether species-specific behavior is instinctive or learned, modern experimenters ask how much is learned, and when and how is it learned.
Ethologists led by Tinbergen and von Frisch introduced elegant and rigorous experiments in the field and in the laboratory under natural conditions. In order to demonstrate that bees have color vision, von Frisch had to introduce experience with particular colors under experimental conditions. At the time that von Frisch reported these experiments, of course, it was as revolutionary to claim that a tiny bee’s brain could learn anything particular about a flower as to claim that a bee’s brain could tell one color from another.
Many animals mark their territories with distinctive scents, but many others have to remember the arbitrary landscape of trees, bushes, and rocks in their territories. In many species, bonding with one mate and caring for one’s own young depends on learning to recognize the sights and sounds of individual mates and young. Obligatory and species-specific behavior such as homing, foraging, mating, and caring for young often depends on arbitrary learning. Ethologists must study the arbitrary effects of learning on obligatory, species-specific behavior.
Experimental Psychology
Experimental psychology also has roots in the late 19th century. From its beginning the experimental psychology of learning concentrated on arbitrary learning. In early experiments humans learned arbitrary lists of words and meaningless strings of letters, and both humans and other animals learned arbitrary paths through artificial mazes. Skinner invented the operant conditioning chamber or Skinner box in the 1930s to serve as an artificial learning environment. In the traditional view, learning to press a lever or peck a key is completely arbitrary and independent of the evolutionary history of rats and pigeons.
There are two, closely related, reasons why experimental psychologists looked for arbitrary forms of learning. First, the advance of modern science depends heavily on the discovery of general, widely applicable principles. Physicists look for general laws that apply to all forms of matter and all forms of energy. Physiologists look for general laws that apply, for example, to the blood and circulation of all animals. If mazes and Skinner boxes represent arbitrary forms of learning in artificial situations, then the principles of learning in these situations should be general rather than species-specific. In the traditional view, learning in mazes and Skinner boxes should reveal the basic principles of all animal learning, including human learning.
The second reason for studying arbitrary forms of learning in artificial environments has to do with the traditional and popular view that human behavior transcends human biology and evolution. In this view, human beings live in totally artificial environments of their own recent creation. Most of their learning about speaking and writing, driving cars and flying airplanes, building cities and paying taxes is arbitrary and independent of their evolutionary history. If this is true, then the principles of learning discovered in artificial environments like mazes and Skinner boxes should be the principles that are most relevant to human learning.
Ethology of the Skinner Box
Moving away from the detailed descriptions of the ethologists, the founders of the experimental psychology of learning, such as Pavlov, Hull, and Skinner, insisted on single, arbitrary indexes of learning. They counted drops of saliva or number of lever-presses to the exclusion of all other observations. Chapter 2 discusses a sample of Zener’s full description in 1937 of the rich and varied behavior of a dog during a typical trial of salivary conditioning in Pavlov’s laboratory in Russia. But, Zener was exceptional. He parted with Pavlov and his colleagues both in methods and conclusions. To the end of his career, Skinner urged readers to study the marks of the pen on the recorder rather than the animal in the box (1988, p. 466). In his view, watching the animals inside the box only diverted the experimenter from the arbitrary principles of learning revealed by the pen writing on the recorder.
Fortunately, in recent decades more and more experimenters have been curious enough to look for themselves to see just what else the animals do in the conditioning chambers besides pressing levers and pecking keys. Relatively inexpensive videotape recording has encouraged this trend. With videotape, observers can describe varied ethological patterns of behavior at leisure and without disturbing the experimental subjects. At the same time, experimenters can compare the descriptions of independent pairs of observers to see whether they agree with each other. Independent agreement between observers is called reliability. Descriptions of what the animals are doing besides pressing levers and pecking keys reveal a rich ethological flux of behavior in the conditioning chamber (Timberlake & Silva, 1994).
Despite the fears of traditional experimental psychologists, ethological descriptions of species-specific behavior lead to general laws. By discovering general principles relevant to human welfare, the founders of ethology, Tinbergen, Lorenz, and von Frisch, earned the Nobel Prize for Medicine in 1973. Meanwhile, the International Society of Human Ethology, founded in 1972 to recognize the ethology of human behavior, has flourished and grown. In the traditional view, obligatory, species-specific behaviors were troublesome artifacts to be eliminated from experiments and forgotten in theories. Recent discoveries show that, on the contrary, ethological patterns are at the heart of all animal learning.
As artificial and arbitrary as the conditioning chamber may appear, rats and pigeons bring their species-specific ethology with them into the chamber. Human beings also bring their species-specific ethology into the artificial and arbitrary environments of the modern world. Careful experiments have given us a large body of evidence about the way animals learn in artificial environments. This book looks for fresh insights into the general laws of learning by examining this body of evidence from an ethological and ecological point of view.
This book also looks for fresh insights by examining laboratory evidence from the point of view of experimental operations. Readers who can master the principles of operational definition can test propositions for themseleves and come to their own conclusions about each topic in this book. They can apply the same principles to evaluate theories and evidence that fall outside the limited scope of this book as well as theories and evidence that appear in the future. They should find that the same rules of evidence apply to other fields of animal behavior and psychology.
PLAN OF THIS BOOK
This chapter introduces the main themes of this book: operational definition and an ethological, ecological, and evolutionary perspective. The next two chapters introduce the descriptive terms of most experiments since the 1930s that divide experiments into two procedures: classical conditioning and instrumental conditioning. Later chapters raise questions about this distinction, but the terms in chapters 2 and 3 remain useful for describing experiments.
Chapter 4 introduces a framework for describing basic mechanisms that have emerged as candidates for the fundamental unit of classical conditioning: S-S, S-S*, S-R, and S-R-S*. Chapter 4 also introduces the method of testing the operational implications of theories against experimental findings.
The remaining chapters apply the method of chapter 4 to a selection of central problems in human and nonhuman animal learning. They trace basic themes from their roots early in the 20th century to recent developments in animal behavior and robotics. Concrete operational definitions and modern developments in computer science challenge traditional views. The object of each discussion is to raise questions that remain unanswered or only partially answered at this writing.
The last two chapters discuss research on infant chimpanzees that were cross-fostered by human adults. These chapters apply the ethological principles of learning to a situation in which a free-living, well-fed, infant chimpanzee has to learn something much more advanced than lever-pressing or key-pecking. The results open fundamental questions about learning, intelligence, and language. Here we included some of the texture and detail of a personal adventure in ethology and experimental psychology.