Chapter 14 Chapter 9 Behaviorist-2

After a slow start, behaviorism soon gained the favor of many psychologists in the 1920s, especially in the United States.It quickly became the dominant view and, before long, almost the only acceptable view, at least in academic circles. The main reason for its popularity is that it claims to be the first truly scientific psychology.Until the nineteenth century, psychology remained mostly philosophical rather than scientific.In the 19th century, the successors of the new psychology tried to turn psychology into a natural science, but they stopped at explaining some simple reflexes and perceptions in physiological terms, and even this This is also achieved by relying on unverifiable introspection methods.

Behaviorists, by contrast, say they can build a psychology entirely on visible, measurable phenomena—these mutually causal stimulus-response phenomena, they say, are the sum total of animal and human behavior. the basic unit of .Such a psychology would be based on specific and unchanging responses analogous to those in chemistry or physics, and would allow psychologists, in Watson's words, to "predict the expected response given the stimulus— Or, where the response is seen, specify the stimulus that elicited the response." Another reason so many psychologists find behaviorism so appealing is that since they only need to focus on a few observable behaviors, they can ignore what philosophers and psychologists have been working on over the past twenty-four centuries. Thinking, untraceable issues.Behaviorists say that not only is it impossible to know something going on in our minds, but we also don't need to know it for the purposes of explaining behavior.They often liken the mind to a black box with unknown circuits inside.It doesn't matter what's inside if we know that when a button on the box is pressed, the box will emit a particular signal or action.What is in the mind should not be discussed at all, because all talk about mental processes is tantamount to believing that there is some invisible thing that controls the brain machinery-"ghosts in the machine", British behaviorist philosopher Gill Sir Burt Ryle sneered. (One anti-behaviorist's jibe is equally notable: "Mere mentioning the word 'mentalism' offends behaviorist sensibilities as much as mentioning 'masturbation' in the presence of polite people Just as repulsive.")

Besides, there are profound social and cultural reasons for the success of behaviorism.It appeals to some 20th-century figures, especially in the United States, because it is practical; it does not seek fundamental explanations, but rather common sense that can be put to work. At least one behaviorist historian has linked its rise to urbanization and industrialization in the United States, and that is David Bacon.These social developments, he argued, gave rise to an urgent need to know the unknowable and disturbing strangers around us, which behaviorism promised people could address. Bacon also mentions two other social reasons for the popularity of behaviorism.One was that World War I aroused hostility to German psychology, and behaviorism became just the fashionable and available alternative;It states that ignorance of some of the subtleties of mentalist psychology is all right, on the basis that mental phenomena are either illusory or unknowable, and not worth the time and effort to study them.

From the 1920s to the 1960s, behaviorism (or its more sophisticated counterpart, neo-behaviourism) was the dominant force and paradigm in American psychology, and it spread to the rest of the psychological world.Some psychologists still cling to earlier ideas, while others, including Freudians, developers of psychometric tests, child developmental psychologists, and Gestalt psychologists, focus on mental processes, In most universities, however, these individuals have had to adapt their work and language to the behaviorist paradigm.Behaviorist historian Gregory Kimble puts it somewhat hyperbolicly: "In American psychology in the 1950s, publishing on mind, consciousness, will, or even energy risked being squeezed out of the profession." Because Use of these terms indicates that someone is a spiritualist who believes in outdated, subjective and mystical concepts.

As a result, much of the research done in the 1920s and 1960s dealt with extremely fine-grained, admittedly objective, but not very illuminating subjects.Some representative titles drawn from the 1935 contents of the Journal of Psychology and the American Journal of Psychology are as follows: "Effect of starvation on chicken pecking response" "Comparison of the first and second explorations of a mouse in a maze" "Use of maze-trained mice to study the effects of morphine and related substances on the central nervous system" "Misidentification in the Animal Maze"

"Development of a circuit for electrophysiological skin resistance meter" Even when humans are the subject of experiments, some dissertation topics and methods are constrained by behaviorist dogma. Some typical theses from the American Journal of Psychology in 1935 are as follows: "Reliability of pH in human mixed saliva as an indicator of accompanying behavioral physiological changes" "Comparison of Muscle Conditioning with Different Degrees of Voluntary Control" "Experimental elimination of higher order reactions" "The galvanic skin reflex associated with explicit emotional expression"

"Overcompensation of temporal dependencies in finger lateral motion" The authors of these articles, and those conducting similar studies, were not really interested in pecking behavior in chickens or the pH in human saliva, but in learning—acquiring behavioral responses to different stimuli.Learning was a central issue in American psychology during the era of behaviorism, which assumed that nearly all behavior could be explained by the principle of stimulus-response learning.It is also equally important to assume that these principles are as true for all sentient animals as the principles of valence are for all the elements in compounds.What people know from chickens, cats, dogs and especially mice applies to humans as well.

Mice are a favorite type of laboratory animal to use because they are relatively inexpensive, small, mobile, and mature quickly.Countless tens of millions of mice have been used in this research enterprise by navigating mazes, manipulating levers or pressing buttons to get food, jumping over doors of different colors, pressing a stick to disconnect their claws The pulsating electric current, and many other tasks.There is nothing trivial about these experiments; they all aim to discover important universal laws of behavior.Here are a few examples: - Put a mouse at the entrance of a simple maze with 6 choice points (each choice is a T-shape, one branch is a dead end and the other branch can continue), and there is a goal box at the end .The mouse began to sniff and explore, and then ran in small steps; it entered a dead end, turned around, and ran to the other side; after three times of right and wrong, it reached the target box - after being brought out for a short rest Put it back in place.By the seventh time, it found the feed at the target.The rat sniffs the feed, then swallows it in one gulp.Another mouse underwent the same training, but without any food reward, and there was no reward for running to the end.

Two mice underwent the same training every day for a week.By the end of the week, the first mouse had completely mastered the route and went straight through the maze without making any mistakes; the second mouse made the same mistakes as before.Eventually, though, the second mouse also got the food at the end of the track, and then, surprisingly, tried again without making any mistakes.It learned in a day what the first mouse had learned in a week.This experiment demonstrates two principles at work: Reward produces learning, as can be seen in the first mouse, and potential learning in the absence of reward, as can be seen in the behavior of the second mouse. (In a sense, learning occurs without a reward but is motivated as soon as a reward is associated with the "correct" behavior.)

What does this have to do with human behavior?Any teacher will tell you that.A child learning to draw or any other skill may make some progress until the teacher says something encouraging or praises; then suddenly the child shows improvement.Similarly, a novice pilot may have to stumble a dozen times to land, and finally "get lucky" semi-accidentally, get praise from the instructor, and every time since then, he seems to "have a lot of experience" when he lands. ". - Several mice were placed into the starter box of a simple T-maze, one at a time.At the end of the right-hand branch is a white door with a bit of cheese behind it; at the end of the left-hand branch is a black door with a metal grate behind it that gives a very light but uncomfortable electric shock to the rat's paw .But after they learned it, the experimenters changed the situation.Now, the white door and food go in the left branch, and the black door and electric grid go in the right branch.The rat turned right, was shocked, and immediately learned to turn left.

Once again, these diabolical experimenters turned everything upside down, but now, the mice knew it right away.They already know to associate rewards and penalties with the color of the door, not the direction.The experiment again demonstrated Pavlov's law of discrimination, which states that given two cues, they would be taken from the rewarded cues. Does this apply to humans?certainly.A novice gardener may end up with a small pile of tomatoes, while his neighbor grows a different variety, gets more sun, and has a bountiful harvest.The novice tried to plant neighbor's seeds in the second year, but luck was still bad.Realizing that daylight hours must be the key factor, he sawed down some trees to let in more sunlight, and this time he succeeded. — Another type of T-maze in which the rat learns to turn to the right.This time, there is no penalty for choosing the left branch, just no reward.Some mice were luckier and were rewarded every time they picked the right side.Others were less lucky, finding food only once in four.The unlucky rats learned to choose the right much slower than the lucky ones.Experiments have shown that partial reinforcement in learning is much better than continuous reinforcement. However, the experimenters turned everything upside down.Both groups of super rats received no reward in either branch.What will happen?Oddly, some of the previously lucky mice quickly forgot the previous conditioning and changed their choice, while those previously rewarded only every fourth time continued to choose the right over a long period of time.Experiments have shown some reinforcement effects: the more an animal expects, the harder it is hit by changes in the situation; if the expectations are low, their learned behavior is more stable when changes occur. Human analogy: An employee who does a pretty good job gets a sizable salary increase every year; one year, the company earns poorly, he gets only a small salary, loses motivation, eats lunch longer and longer, afternoon I leave work on time at 5 o'clock and call in sick from time to time.A less outstanding employee only occasionally gets a slight increase in salary due to the rise in living standards, and when the benefits are poor, he can only get a Coke.His commitment to work was unaffected because he had low expectations, and he believed that the reduction in rewards was not a change in the system as a whole. As shown in the experiments above, behaviorists expanded behaviorist doctrine and methodology well beyond Watson's set of concepts.He described behavior in simplistic terms as "variations of muscles and glands, with and without striations, following some given stimulus," a view that later became known as the "psychology of muscle twitches."For a while, his disciples held to this view; as one of them, Walter Hunt, wrote in 1928: "All behavior seems to be a combination of relatively simple muscular and glandular activities of varying complexity." However, to say something meaningful about complex forms of behavior, it must be observed in its undamaged form, as some behavior with character and meaning.A nesting bird is not just an organism with several reflexes to several stimuli, it is also a nesting bird—a complex behavior with a purpose.As the behaviorist Edwin Holt put it in 1931, behavior is "what the organism is doing"—foraging for food, making love, etc.—and it is an organic whole, not just a series of The reflection used to form its properties is not just an "arithmetic and, not only related to the relationship of addition and subtraction". Holt, however, is reluctant to ascribe a purpose to the animal itself; that would imply the influence of the mind, which anticipates a goal forward and then strives to achieve it.Conversely, he attributed the purposefulness of complex behaviors to the process of combination of stimulus-response units: the search and avoidance in each step of the animal constitutes the collective stimulus-response unit in such a way that this combination of behavior It looks like it was done on purpose.This statement is vague and unconvincing, but it goes as far as any orthodox behaviorist can go. Clarke L. Hull (1884-1952) of Yale University made a more important effort as a new behaviorist who promoted behaviorism as a rigorous science of quantitative analysis on the model of Newtonian physics.Hull initially planned to be a mining engineer, but his lower limbs were disabled due to polio.He turned to psychology, a science that was unlikely to involve too much physical activity, but, having completed his engineering training, he went so far as to develop a calculus for behaviorism.As he puts it in his autobiography: Around 1930, (I) came to the fairly certain conclusion that psychology was a true natural science; that its fundamental laws were quantitatively describeable by a few ordinary equations; will be generalized in accordance with the secondary laws derived from (1) these primary laws and (2) the situation in which the action takes place; are derived from the same original equation as the quantitative law. Hull's central idea is a familiar one: Behavior consists of a series or string of connected habits, each a stimulus-response connection, that arise as a result of reinforcement.He is a replica of Thorndike's law of effect.What is relatively new in Hull's work is that he postulates a series of factors, each of which he believes reinforces, limits, or inhibits the formation of such habits, and that he formulates some equations according to which people The precise effect of each of these factors can be calculated. They include the level of animal drive (a hungry mouse has a much greater drive to seek food than an already fed-up mouse); the force of reinforcement (expressed in terms such as "5 grams of standard food") ; the number of times the stimulus was followed by reinforcement; the degree of "demand reduction" obtained after each reinforcement; the degree of "drive reduction" due to fatigue and the length of time between one attempt and the next (drive fueled by demand), etc.It was, as Edwin Pauline later mercifully put it, in a most tactfully phrased, a very "dumb" doctrine. Here is an example.One can calculate the degree to which any given number of repetitions of a reinforced action can increase the strength of a learned habit by the following equation: NsHR=M-Me-iN What this equation means is that the strength of a learned habit depends on the number of reinforcement attempts (N), the relationship between the input and output nerve impulses in a specific action (sHr), the maximum physiological strength of that particular habit (M) minus, Well, it goes on like this. Hull's greatest original intention was to establish a new behaviorist psychology according to natural science, so that it could gain an intellectual status.His study of calculus, which seemed fragmented in the 1930s, took a systematic form in his Principles of Behavior (1943), when the method was revered and influential.In the late 1940s and 1950s, thousands of master's theses and doctoral dissertations were based on one or more of his postulates.He became the most cited psychologist in the psychology research literature and among leaders in learning psychology. But by the 1960s, the bulkiness of Hull's theory and the decline of behaviorism's status made his name and work fade rapidly.By 1970 he was rarely cited, and today almost no research is based on his theories.When Hull died in 1952, he must have thought that he had achieved scientific immortality.Now, he is only a slightly noticeable figure in historical research, and his name is rarely known to young psychologists and psychological circles. B. F. Skinner (1904-1990), another major neo-behaviorist, had a very different fate.He became and remained, until his death at the age of 86, the most famous psychologist in the world.His ideas are widely used in psychological research, education, and psychotherapy today. So he must have been one of the biggest contributors to the human quest to find self-understanding, right? Far from it. Human self-understanding, at least the kind of self-understanding that philosophers and psychologists have sought for so many centuries, was not at all one of Skinner's goals or contributions.Throughout his long life, Skinner adhered to his behaviorist views, in which he believed that "subjective beings" such as consciousness, thought, memory, and reasoning did not exist at all, but were only "some linguistic constructs and human beings in language Some grammatical traps that unfortunately fell into the process of development", "are some annotative things that cannot be explained by themselves".Skinner's goal was not to understand the human psyche, but to determine how behavior is triggered by external causes.He was so convinced of his views that he wrote in a short autobiography—he also has a three-volume edition: “[Behaviorism] may need clarification, but it no longer needs to be debated.” Nor does Skinner add anything to the theory of psychology; he does not think a theory of learning is necessary, and declares that he does not have one.A theory he did believe, which can be summed up in one sentence, is that everything we do and ourselves are determined by a history of rewards and punishments; Constructed of principles such as the reinforcement effect, his theory is concerned with situations that elicit the need for behavior and those that cause the extinction of the behavior. So, what made him so famous? Like Watson, Skinner was a natural mobster, a demagogue and a brilliant ad man.In his first television appearance, he posed a dilemma originally posed by Montaigne—"If you had to choose, would you burn your children? or the books?"—then he I would personally rather burn my own child, because he would contribute more to the future through his work than through his genes.Predictably, he sparked outrage - and was invited to make further appearances. At other times he seems to enjoy poking fun at words that thoughtful people use to talk about and understand human behavior: Behavior... still comes down to human nature, and a generalized "psychology of individual differences" where people compare and describe humans in terms of character traits, abilities, and potential.Almost everyone who cares about the human condition... talks about human behavior in this pre-scientific way. Skinner always scoffed at the effort to understand the human mind: We don't need to try to discover what personality, mental state, feeling, character trait, plan, purpose, intention, or some other prerogative of what an automatic person really is, in order to catch up with a truly scientific analysis of behavior... Thought It is action.The mistake is to assign behavior to consciousness. What we need or can know, he says, are the external causes of behavior and the observable consequences of that behavior.This would lead to "a complete picture of the organism as a behavioral system". Consistent with this view, he is a strict determinist: "We are what we are today because of our history. We love to believe that we can choose, we can act... (But) I don't believe that a person can be free or responsible." The "automatic" person is an illusion; a good person is conditioned to be that way, and a good society will be based on "behavioral engineering" Yes, that is, the scientific control of behavior through positive reinforcement. Skinner was a deft showman who made himself popular.He spoke clearly, was self-centered without blushing, and was attractive.To demonstrate his own conditioning skills, he taught a pigeon to play a tune on a toy piano, and a pair of pigeons to play a type of tennis ball in which two pigeons roll a ball around in their beaks.Millions of people who saw him on TV documentaries thought Skinner was a great guy, or at least an animal expert.He expressed his vision of an ideal, scientifically controlled society in the form of a utopian novel Walden II (1948), in which he presented a picture of a small society, from birth , the children are rigorously conditioned through rewards (positive reinforcement) to make them cooperative and social; all behavior is controlled, but for the good and well-being of all.Despite the bland dialogue and contrived plot, the novel has become a highly regarded book read by college students for years and has sold more than 2 million copies. Yet Skinner's reputation with the public was greater than with his peers.As the admirer psychologist Norman Gutman wrote in The American Psychologist a few years ago: (Skinner) is a famous figure in mythology...the scientist-hero, the Promethean fire-seeker, the skilled technologist...the iconoclast, the man who defies authority, who set us free thinking, thus breaking away from the limitations of ancient times. Skinner was born in a railroad town in Pennsylvania in 1904, where his father was a local lawyer.As a child, he had a penchant for making intricate gadgets.Later, as a psychologist, he also invented and built many extremely effective devices for animal experiments.During school and college he aspired to be a writer, and after college he spent a year practicing writing, mostly in Greenwich Village.Although he carefully observed all kinds of strange human behaviors around him, after a while he found that he had nothing to say about what he saw. After being extremely discouraged, he decided to give up this plan. However, he soon found a way, which seemed to him a more practical way of understanding human behavior.In his reading he came across the work of Watson and Pavlov, among others, read the books written by them, and then decided to dedicate his future to the search for a scientific method for understanding human behavior, especially It is the study of condition formation. "I'm obsessed with my failure in literature," he told an interviewer in 1977, "and I'm convinced that writers never really understand anything. It's for this reason that I turned to psychology." .” He came to Harvard.It was a place of introspective psychology, but he was no longer interested in what he called "insider information," and he turned sharply to behaviorist research with rats.In his autobiography, Skinner recalls with glee that he was thankful he had been something of a bad boy: "They may have thought that something in psychology was watching me, but the truth is , I can do whatever I want, whatever I want.” Skinner disobeyed the teachings of his professor and became more and more thorough behaviorist. When defending his doctoral dissertation, he was asked to list some objections to behaviorism. , but he couldn't think of a single one. Skinner used his dexterous mechanical abilities to build a maze that was much improved on Thorndike's appearance; since then, this kind of maze has been widely used and known as the Skinner box.In its basic form - and it comes in many varieties - it is a cage big enough for a white rat to stay comfortably inside, with a horizontal bar on one wall that fits in a small food tray and water spray. above the mouth.The mouse crawls around in the cage, and when it happens to rest its front paws on the crossbar and press it down, a grain of food will automatically fall into the food dish.Some equipment attached to the outside of the cage draws a line showing the total number of bar depressions minute by minute, automatically recording the rat's behavior.This was more efficient than Thorndike's maze box method, and it was easier to collect data, because the experimenter did not need to watch the mouse or deliver the feed when it pressed down on the bar, but just looked at the records. The box also yielded far more objective data on the acquisition or disappearance of behavior than was collected by anyone at the time.The mouse will decide, and it alone, how long the interval between pressing the bar this time and the next time will be.Skinner could base his discoveries about the principle of learning on the "response frequency," that is, the frequency at which an animal's behavior changes in response to reinforcement, undisturbed by the actions of the experimenter. In addition, Skinner could tune the box to mimic in various ways the many real-world situations in which behavior is reinforced or not.For example, he could study how animals learn to respond when they are regularly rewarded; how learned responses disappear when the reward is suddenly interrupted; How does that affect learning and response elimination; what happens when rewards are delivered sporadically; what happens when pressing a bar yields mixed outcomes (such as a reward followed by a shock), etc.In each case, the data-derived curves show the frequency of acquisition and elimination of the behavior under these various conditions. From these curves Skinner developed several principles that improved our understanding of mouse behavior, as well as human behavior.In one instance, he found that partial reinforcement had an important change effect.Food is delivered occasionally or irregularly. After the planned training, the rats will continue to press the bar, even if the feed delivery device has been completely turned off.The behaviors they learned were less likely to be eliminated than some mice trained with intermittently delivered reinforcement.Someone compares this to the behavior of gamblers playing slot machines in a casino: Neither the rat nor the gambler can predict when the next reinforcement will come, but, because they are used to getting some rewards occasionally, they will keep trying in order to expect Win a prize on your next attempt. Yet Skinner's most important contribution was his "operant conditioning," which alone deserves a permanent seat in psychology's hall of fame. In "classical" (Pavlovian) conditioning, an animal's unconditioned response to food (salivation) is transformed into a conditioned response to a previously neutral stimulus (the sound of a metronome or a bell); A key factor in behavior change is novel stimuli. In "instrumental" (Thorndike-style) conditioning, the key factor in behavioral change is the response, not the stimulus.The neutral response—happening to step on the pedal during random access to food—was rewarded with food and became a learned behavior that served a purpose that hadn't been done before. Skinner's operant conditioning is an important development of instrumental conditioning.Any random activity performed by an animal for whatever purpose can be seen as a "manipulation" of the environment in some way, and thus, in Skinner's terms, can also be a "manipulative action" that rewards the activity for Generate operating condition formation.By rewarding a series of small random actions one after the other, the experimenter can "set" the animal's behavior until it takes an action that is not inherent or part of its natural repertoire. Here's an example of how Skinner stereotyped the behavior of a pigeon. In a Skinner box, flush with the side of the box, was a small colored plastic disc, which he asked the pigeon to peck at. plate: We first feed the pigeon when it turns its body slightly in the direction of this point (the plate) from any point inside the box.This increases the frequency of this behavior.Then, we don't reinforce the rewards until there's a slight pivot in that direction.This is again an ordinary assignment that changes the behavior without spawning a new unit.We then proceeded to reinforce it when it was getting closer and closer to the point, then only when its head moved slightly towards the point, and finally, rewarded only when its head actually touched the point. In this way we can establish complex maneuvers which would otherwise never have been possible in the repertoire of this organism.By intensifying a series of successive close actions, we can derive in a short period of time a rare response to a very high probability... turning from any point on the box to this point, approaching it, looking up Come, and the repertoire of pecking toward this point may appear to be a functional unit of behavioral development, but it is built up in indiscriminate behavior by a continuous process of discriminative reinforcement. (Other experimenters have used Skinner's technique to create far more exotic behaviors. One taught a rabbit to pick up a coin, put it in its mouth, and drop it into a piggy money box. Others taught A pig named Priscilla turns on the TV, picks up dirty laundry and throws it in a basket, and vacuums the floor.) Skinner likened the handling training of his pigeons to the adult behavior of a child learning to talk, sing, play games and everything in due time.In his view, everything is a long series of behaviors formed by using operant conditioning, connecting the small connection points of some simple behaviors.One might as well call it the erectile repertoire of human (Homo erectus?) behavior—that is, a mindless robot composed of operant conditions drawn from a myriad of meaningless little units. Skinner was somewhat neglected by the psychological establishment for quite a long time, but slowly won some dedicated minds - enough results to eventually lead to the publication of four journals of Skinner's behaviorist research and his teachings , and set up a Skinner-type Research Department (Part 25: Behavioral Experimental Analysis Department) in the American Psychological Association, which currently has more than 1,000 members, or about 1.5% of the total members.Skinner boxes and operant conditioning techniques have been used by experimental psychologists ever since.In recent years, Skinner's name and work have been cited hundreds of times in social science publications each year (though only one-seventh of Freud's). However, Skinner's main influence remained outside mainstream psychology. When Skinner visited his daughter's school in 1953, it occurred to him that some operational skills similar to those he taught pigeons to play the piano might be more effective teaching methods than traditional methods.Complex topics can be broken down into simple steps in a logical order; students are asked questions and immediately told whether the answers are correct or not.There may be two principles at work here: the knowledge that students have answered correctly is a very powerful behavioral reinforcement (reward), and immediate reinforcement will be better than delayed reinforcement.The result is what is known as "controlled pedagogy." However, because it is impossible for a single teacher to provide reinforcement to a classroom of students at the same time, new textbooks had to be written that contained pairs of questions and answers, each one a small step toward an overall grasp of the topic. Furthermore, students can immediately reward themselves by closing answers in their textbooks.Skinner also developed a teaching machine for operant self-teaching by comparable means.这种机械模型当时红火了一阵子，然后没有人使用了，可是，今天，以计算机为基础，带有立即强化的自我教学法又在快速发展了。 有一些年，控制学习运动对教学法产生了很大的影响，设计用来通过操作性条件形成进行教学的课程和备课材料广泛使用，美国相当多的中学和大学，以及十几个国家的许多学校都使用这种方法。可最终，教育者们认识到，控制教学的细分法只提供了人类所需知识的一部分，他们还需要完整和有层次的思想结构。而且，以后的研究显示，在人类当中，延迟的强化经常比立即强化有更好的结果，思考别人的反应可能会导致比立即反应和得到答案更多些的学习。最后，观察别人的行为，尽管对猫不一定是有效的，可对人却是一种非常有效的学习形式，而且它不牵涉到立即强化。不管怎样，斯金纳关于立即强化的教条证明是有用的，而且为大多数教师所熟悉，并被溶入许多教程和中学教科书中去了。 斯金纳对一些精神和情感疾病的治疗还起过不小的作用。他曾想到，通过对病人从病态行为向正常行为些微转化的奖励，说不定可以使病人的行为重新定型。他和两位研究生从40年代开始进行了首次实验尝试，后来叫做行为修正法。他们在波士顿附近的州立医院搭了一些按压横杆台，如果病人按照有顺序的方式操作机器，他们就会得到一些糖果或者香烟奖励。一旦实现这个目标，治疗师就给他们一些象征物，以奖励精神病人的合适行为，比如自愿进食，自我修整，并协助进行房间整理工作等。这些象征物可以换糖果、香烟，也可以换一些特权，如选择进餐的隔间，与医生交谈，或者看电视。 在深度偏执的精神病人当中奖励期望得到的行为经常能够奏效。一位压抑的妇女不愿吃饭，还有饿死的危险，可是，她喜欢探访者，还喜欢电视机、收音机、书籍和杂志以及鲜花摆在她房间里。治疗师把她移到一间没有这些东西的病房里，并把一份便餐放在她面前；她只要吃下任何一点点东西，其中的一些享受品便临时给她恢复过来。治疗师慢慢保留奖励，直到她吃得越来越多。她进餐的情况好转了，增加了一些体重，并在两个星期的时间内出院了。18个月后随访，发现她过着正常人的生活。 行为修正法运动传播到了好些精神病院和感化院。精神病学家和心理学家们现在认为，它对一些病情严重的精神病人有作用，是他们自己的疗法当中有用的补充，不过，就时间和员工精力来说，这是一种很昂贵的办法。行为修正法还被许多心理治疗者们用来治疗一些不那么严重的精神病人，如吸烟、肥胖、害羞、抽搐和语言障碍。它是行为治疗法领域中的的个专门技巧，大部分是以巴甫洛夫型条件形成理论，而不是斯金纳的行为修正法为基础的。 斯金纳最有名的作品，《沃尔登第二》并没有重塑美国社会，甚至一部分也没有，不过，它无疑对成百上千万读者的思想和社会概念产生过影响。只有一次，有人的确是想按照《沃尔登第二》实际地创建一个乌托邦：这就是弗吉尼亚路易斯萨市的“双橡公社”，由8个人在1967年建立的一个社区。经过许多年风雨之后，它的人口已经增长到81个。虽然仍然是按《沃尔登第二》的模式进行管理，可公社的社员们早些年以前就已经放弃了定义理想行为的努力，他们不再通过斯金纳强化法来给彼此的行为定型。 斯金纳有时候对自己在世界上的影响进行自我贬低。“总起来看，”他曾说，“我对别人的影响远不如我对老鼠和鸽子——或者作为研究对象的人的影响重要。”这话可能不是当真的。斯金纳真正的意思是下面这段话：“我从没有在任何时候对（我的工作的）重要性产生过怀疑。”而且，他还带着很有特色的乖张口气说：“当它开始吸引注意力的时候，我对其影响是忧虑多于高兴啊。我的档案里有很多笔记说到这个事实，即对于这些所谓的荣誉，我感到很是害怕或者深为不快。我常常放弃会占用我的工作时间，或者过度强化其具体方面的一些荣誉。” 行为主义研究在积累势力的途中，除了这门学说最执着的追求者外，人人都很清楚，老鼠和其它实验动物经常以这门学说无法解释的方式行动。 一方面，它们的行为经常不符合所谓的万用条件形成原理。“鸽子，老鼠，猴子，哪个是哪个呢？可这并不重要”，斯金纳曾写道，可这的确很重要。研究者们可以很容易地教会一只鸽子去啄一块圆片或者开启食物门的钥匙，可是，他们发现几乎不可能让这种鸟类扇动翅膀来取食。他们可以轻松地教一只老鼠压下挡杆取食，可是，却要花天大的劲才能让一只猫也这样做。给一只老鼠喝发酸的海水，然后喝一种恶心的药水，他会避开发酸的水，却愿意喝海水；对鹌鹑进行类似试验，它会避开海水，却愿意喝发酸的水。这些比较得来的成果迫使行为主义者承认，每个物种都有其自己内部的电路原理，使它很容易地通过本能来学习一些东西，另外一些却有困难，而还有一些一点也不会。学习的原理远远本是放之四海而皆准的。 行为主义心理学更为严重的一个错误是，实验动物经常以一些不能够按照很清晰明白的反应曲线率来行动。例如，许多研究者都曾发现，在一次反应消除尝试的开始，一种动物会以更大的精力来对刺激作出反应，不似它们在长时期的强化训练中那样。一只一直通过按动横杆取一颗饲料的老鼠，如果发现没有饲料，它会一次又一次更用力地按动横杆，而按照严格的行为主义学说，奖励的缺失会使反应强度减弱，而不是增强。 但是当然，人类亦是如此。当一台自动售货机不再发货时，客户会更用力地推拉几下，甚或敲打或者踢几脚，要么是发泄，要么是以为哪个地方卡住了，需要再踢一脚。行为主义学说对此内部的过程没有什么解释，特别是对问题的思考不予解释，可是，若干行为主义者注意到，他们的老鼠有时候的行为看起来好像是在做些基础的目标型思考。 一位著名的研究人员爱德华·查斯·托尔曼（1886-1959）了解这种情况，他是赫尔同时代的著名人士，也是30和40年代著名的新行为主义者。他观察到，一只老鼠跑过几次迷宫之后，它会在某个地方停下来做决定，左边看看，右边看看，往前走一点，也许再往回跑一点，都是在作出决定并继续进行之前发生的。他在1938年就任美国心理学协会主席时的致辞中说，很清楚，老鼠是在脑袋里面进行“替代性的试误法”。“从人类的角度来看，”他说，“老鼠似乎是在进行'三思而后行'之类的把戏。” 这是托尔曼在老鼠的许多行为当中找出的一部分例子，他认为，这些行为只能解释成老鼠的头部有某种过程在进行中。几年以前，他和一位同事曾制作过一只简单的迷宫箱，里面有三条通向目标盒的路径。最短的那条是从启始处直通目标盒的；第二条稍长一些，向左弯了一下然后在中途接入最短的直路，距目标盒尚有一半的路程；第三条最长，向右转了很长的弯，然后才在靠近目标盒的地方接入最短的那条直路。经过一系列试验后，老鼠按行为主义的理论所预测的那样三条路都跑过，然后学会了选最短的那条直路，因为这是最容易建立起来的习惯。 然后，托尔曼在直路的中途设了一道障碍，这样的话，老鼠只能通过最长的那条路才能取到食物。按照行为主义理论，当老鼠顺着直路跑下去然后发现障碍时，它应该绕回头来试下一个最容易建立起习惯的路径——即中等长度的那条——可它立即就选了最长的那条。对托尔曼来说，这意味着，老鼠已经建立起了这整个迷宫的某种思维全图，并“意识到”障碍物挡住所有的路径，只除开最长的哪条。 托尔曼进行过许多类似的实验，其中大部分的实验都要复杂得多，可所有的实验都支持他的观点，即，“老鼠的大脑里已经建立起了这个环境的某种类似场图的东西。”他说，标准的行为主义理论只提供了迷宫学习的部分解释：“我们同意……穿迷宫的老鼠经受着刺激，作为这些刺激的结果，它最终导向实际发生的反应。可是，我们感觉到，其中的大脑活动更为复杂，更有模式，而且，从实用主义的角度来看，比刺激-反应论心理学家的自主能力更大。” 这些研究导致托尔曼推敲出一种他叫做“目标性的行为主义”的学说。它的基本意思是，老鼠并非作为一种自动机而产生行为，它们并非完全按照自己所体验的刺激的次数和种类而形成习惯，而是，就好像，它们还受到自己的期盼、它们认为某种东西在某种情况下会导致某种结果的知识、它们的目标和其它一些内部的过程或者状态的影响。如一位正统的行为主义者所嘲笑的那样，托尔曼的老鼠已经“陷入了沉思”。 托尔曼把这种内部的因素称作“干扰变量”（它们会干扰刺激-反应过程），并坚持认为，它们与行为主义是相兼容的，并行不悖的。“对于行为主义者来说，”他写道，“'精神过程'应该被承认而且按照它们所导向的行为的术语来定义。（它们）是看不见的，可都是推断出来的、行为的决定因素……行为和这些推断出来的决定因素都是客观的，是有定义的存在类型。”这是在尽力维护行为主义的理论，可是，托尔曼不管愿意不愿意都还是在行为主义的大堤上掏了一个细缝，导入了一小滴思维。到时候，它会变成一场洪水。 如果奖励和重复只能部分地解释老鼠的行为，则它们对人类行为的决定因素和工作机理的解释就会更有限。拿记忆作比。行为主义者以纯粹数学的术语来描述它：尝试和强化的次数越多，奖励越多，刺激和反应的时间就越接近，刺激产生反应的可能性就越大。如果刺激是诸如“5之后是什么”？这个问题，反应就是“6”。如果刺激是“你的电话号码是多少？”答案就是一串7位的数字（包括区号在内有10位）。第一位数字是对总是的反应，可也是能够产生第二位数字这个反应的刺激，等等，其方式是一串联想联系。 可是，就算在行为主义时代的高涨时期，心理学家们也都知道，人类的记忆比这个复杂得多。一方面，我们“整块”记忆某些信息：比如，我们把区号作为一个单元来记忆，而不是作为一系列互有连接的反应。另外一方面，我们有不同种类的记忆：我们可以暂记住某些电话号码，我们查出号码，暂时记住，直到拨号，然后立即把它忘记掉，可同时，我们还会记忆“长效记忆”的东西（我们把认为需要的东西当作知识长期堆积在仓库里）。某些东西需要无数次的重复和奖励才能固定在记忆里面（很多人好像记不住自己的社会安全代码，曾经看过几十次也记不住）；其它一些东西（在某家饭馆进餐付的一次特别贵的账，孩子说的第一句话）只需要经历一次就能在记忆里保持很长的时间。人类记忆的这些特点和许多其它的特点，并不能用行为主义狭隘和古板的公式来解释。 在整个行为主义时代，一些心理学家持续不断地以更为宽泛，更为深刻的方法来探索人类的记忆，不仅如此，他们还探索行为主义曾经没有注意到的一些心理学现象，如感觉、动机、性格特征、推理、解决问题、创造力、儿童发展、遗传倾向和经验之间的内部作用还有人际关系。慢慢地，关于这些话题而收集起来的数据，和这些数据提出来的一些行为主义不能够回答的问题，为托马斯·库恩在他对科学革命的著名分析中表达的新理论铺平了道路，这就是他叫做“范式转移”的东西——向一种新学说相对突然的转移，它囊括积累起来的大量数据，并使这些按照目前的学说很难解释的数据产生意义，如果这些数据是有意义的话。 同时，其它领域里进行的一些研究也开始对思维的工作机制带来新的曙光了。从人类学里，得出了有文字以前的人类如何思想的研究；从心理语言学中，得出了人类如何获取并使用语言的知识；从计算机科学中，人们知道了一种全新的设想思维的办法——像信息处理一样，一步一步地往前进，就像一道计算机程序。 到60年代，所有这些影响开始汇集成一种有关思维和行为的观点，称作“认知科学”——没有超自然存在并建立在实验方法上的一种认识论，通过它，人们可以对精神过程进行合理的推断。 随着认知科学的到来，行为主义很快失去了它在心理学中颐指气使的地位，它不再是像它宣称的那样一种足以解释所有行为的方法。公爵大学的格列高利·金布尔总结了心理学家们对行为主义幻想的灭失： 虽然古典的学说以简单的学习理论形成并经过了检验，但是，在一切的背后，总是有一种假定，即这些理论可以应用到所有的行为之中去……（而且，）学习的大部分基本法则已经被寻找出来，而留下来的只是一些无足轻重的小问题，这些解决系统分类的小问题只不过是把一些主要的理论家们区分开来而已……（然而，）到本世纪中期，事情已经很清楚了，古典的学习学说在范围上受到局限，我们的科学知识已经上升到的高度还在伽利略以前，而不是在牛顿以后，如赫尔和其它一些人曾思考过的一样。 奇怪的是，只有当行为主义已经日薄西山、日渐式微的时候，其子属支流行为疗法才被广泛使用，并合乎情理地成为对范围有限的一些心理疾病，特别是恐惧症非常成功的疗法。 行为疗法正确的地方——它有用，但是用途有限——与其附属的行为主义理论是类似的。它不是一种谬误的学说，而是说，它只能解释最基本的一些行为形式，而这些行为只是老鼠心理的一部分，也是人类心理学的极小的一个部分。对它的成就的一个极端评论见于1968年一位心理学家尼西米亚·约但的说法： 现代美国科学心理学是在一片不毛之地上白费精力的不孕专家。数年辛勤劳动和数百名教授及数千名学生不懈的努力得出的成果几乎是一个零……自那个“划时代的”时刻（沃森1913年的宣言）以来，53年的时光已然过去，可不可以指出对于人类知识的增长有一点积极意义的任何贡献来？一点都找不出来。 尽管失望的很多人并非都如此激烈地批评行为主义，可是，在60年代，越来越多的心理学研究和教学中心都从行为主义向认知理论转移了。到70年代，行为主义在很大程度上已经被挤了出来，被新的、更为多样化的心理学所替代，这些新的流派研究更为深刻和更为广泛的内容，是本书余下部分要讲到的话题。