The New Behaviorism

Foundations of Behavioral Science

John Staddon

• Publisher: Routledge
• Year: 2021
• ISBN:
  9781000389685
• Edition: 3
• Topic: Filosofia
• Subtopic: Mente e corpo in filosofia

This ground-breaking book presents a brief history of behaviorism, along with a critical analysis of radical behaviorism, its philosophy and its applications to social issues.

This third edition is much expanded and includes a new chapter on experimental method as well as longer sections on the philosophy of behaviorism. It offers experimental and theoretical examples of a new approach to behavioral science. It provides an alternative philosophical and empirical foundation for a psychology that has rather lost its way.

The mission of the book is to help steer experimental psychology away from its current undisciplined indulgence in "mental life" toward the core of science, which is an economical description of nature: parsimony, explain much with little. The elementary philosophical distinction between private and public events, even biology, evolution and animal psychology are all ignored by much contemporary cognitive psychology. The failings of radical behaviorism as well as a philosophically defective cognitive psychology point to the need for a new theoretical behaviorism, which can deal with problems such as "consciousness" that have been either ignored, evaded or muddled by existing approaches.

This new behaviorism provides a unified framework for the science of behavior that can be applied both to the laboratory and to broader practical issues such as law and punishment, the health-care system, and teaching.

Part I
History

1 The Psychology of the "Other One"

Behaviorism was once a dominant force in American psychology. It led to great advances in our understanding of reward and punishment, especially in animals. It drove vigorous movements in education, therapy, and social policy; it is still a force as applied behavior analysis, and many of its ideas have been absorbed into cognitive psychology. Writing on the centennial of behaviorist icon B. F. Skinner’s birth, a leading cognitive psychologist put it this way: “Behaviorism is alive and I am a behaviorist.”¹ But as a separate entity, behaviorism is much diminished within psychology. I explore the reasons and propose an enlarged version of behaviorism in the rest of this book.

I look first at where behaviorism came from, how it dominated psychology during the early part of the twentieth century, and how philosophical flaws and an unexamined ideology in the original formulation prevented its advance—and left psychology and neuroscience prey to covert mentalism and naïve reductionism. I describe a new template for psychology that is behaviorist in spirit but also respects the logic of historical systems laid out by computer science pioneers like John McCarthy and Marvin Minsky. I believe this theoretical behaviorism can provide a basis for scientific psychology. With a clearer philosophical base and proper empirical rigor, perhaps it can help restore coherence to a psychology that has lost its moorings and finds itself adrift in a sea of inconclusive or uninterpretable research. With a clearer view of what is possible, what is true and what is not, what is testable theory and what is not, what can be measured and what cannot, perhaps psychology can move again in the direction of real science.

Psychology has always had its critics. But at a time when it occupies a substantial and secure place in university curricula, when the opinions of psychologists carry political weight, and when the numbers of research papers in psychology have reached the hundreds of thousands, the number of critics has not diminished, nor have their arguments been successfully refuted.² The emperor is if not unclothed, at least suffering a wardrobe malfunction. More on these problems in Chapter 8.

Psychologies

Many eminent psychologists, beginning with philosopher and proto-psychologist William James (Figure 1.1), have tried to sort out the divisions within psychology. James’s suggestion, before the advent of behaviorism, was to divide psychologists into “tough-” and “tender-minded.” He would have called behaviorists tough-minded and many cognitive, clinical, and personality psychologists tender-minded. James might also have mentioned the division between practice (clinical psychology, behavior analysis) and basic research (experimental psychology), between “social-science” and “natural-science,” or between “structural” and “functional” approaches—not to mention the split between mentalists and realists. ³ But the main division is still between those who think mind is the proper subject for psychology and those who choose behavior. There is also some contention about exactly what should be meant by behavior, as we will see.

Figure 1.1 William James (1842–1910), philosopher, pioneer psychologist, and promoter of the philosophy of pragmatism, at Harvard during the late 1890s. There were two James brothers. One wrote delicate convoluted prose; the writing of the other was vivid and direct. Surprisingly, it was the psychologist, William, who was the easier writer.

The nineteenth-century ancestor of the mentalists is Gustav Fechner (1801–1887), the “father” of psychophysics. Fechner was the co-discoverer of the Weber–Fechner law, which relates the variability of a sensory judgment to the magnitude of the physical stimulus⁴ (Box 1.1). Weber–Fechner is one of the few real laws in psychology. For very many sensory dimensions, variability of judgment, measured as standard deviation or “just-noticeable difference” (JND), is proportional to the mean. For example, suppose that someone is able to correctly detect that weights of 100 g and 105 g are different just 50% of the time—so the JND is 5%, then the Weber–Fechner relation says that he will also be able to tell the difference between 1,000 g and 1,050 g just 50% of the time.

Box 1.1 Weber-Fechner and Expected Utility

The Weber–Fechner Law

The negatively accelerated logarithmic curve (y = klog x, 0 < k) relates a physical measure of stimulus intensity to a hypothetical measure of sensation. The basis for the y-axis (JNDs) is the experiments on discriminability that show, for example, that the stimulus difference between 19 and 33 (ΔI) in the graph can be judged correctly (i.e., 33 judged as higher than 19) 50% of the time (the JND) and similarly for the difference between 33 and 57. Equal discriminability is then equated to equal mental effect. Hence, successive increments in the physical stimulus are judged to have smaller and smaller mental effects.

Marginal Utility

A similar curve is used by economists in functional theories of economic behavior. The curve relates the amount of a good, something like pizza, cash, or real estate, to the value it has for the owner, its utility. The fact that the curve is negatively accelerated means that constant increments in the good yield smaller and smaller amounts of additional utility (the equivalent to ΔI in the Weber–Fechner law), that is, diminishing marginal utility. Like the Weber–Fechner curve, the curve is derived by inference from empirical results. Many (but far from all!) equilibrium (stable) choice patterns in real markets can be explained by assuming that people choose so as to equate marginal utilities—as this usually maximizes their total utility. Given that marginal utility is assumed to decline with amount, this assumption can explain, for example, why someone who has much milk but no bread will be willing to exchange some milk for some bread with someone who likewise has a bread surplus. Diminishing marginal utility is one way to explain why people, unlike koala bears, are omnivores rather than single-foodies.

The Weber–Fechner relation is a matter of fact. But Fechner went beyond fact to propose that there is a mental realm with concepts and measurements quite separate from biology and physics—hence his term for this new domain: psychophysics. Many contemporary cognitive psychologists agree with him.

On the other side are biological and physiological psychologists. Biological psychologists believe that the fact of Darwinian evolution means that the behavior of people and nonhuman animals has common roots and must therefore share important properties. Physiological psychologists—now usually called neuroscientists—believe that because behavior depends on the brain, and because the brain is made up of nerve cells, behavior can be understood through the study of nerve cells and their interactions.

Behaviorism is in the middle of the mental–biological/physiological division, neither mentalistic nor, at its core, physiological. ⁵ Behaviorism denies mentalism but embraces evolutionary continuity. It also accepts the role of the brain, but behaviorists are happy to leave its study to neuroscientists. As we will see, behaviorists seek to make science at a different level, the level of the whole organism.

Since its advent in the early part of the twentieth century, behaviorism has been a reference point for doctrinal debates in psychology. Until relatively recently, most research papers in cognitive psychology were sure to include a ritual paragraph of behaviorist-bashing—pointing out how this or that version of behaviorism is completely unable to handle this or that experimental finding or property that “mind must possess.”⁶ Much cognitive theory is perfectly behavioristic, but the “cognitive” banner is waved, nevertheless. No cognitivist wants to be mistaken for a behaviorist!

Behaviorism has been on the retreat until very recently. In 1989, a leading neo-behaviorist (see Chapter 2 for more on these distinctions) could write,

I have . . . used a parliamentary metaphor to characterize the confrontation . . . between those who have taken a stimulus-response [S-R] behavioristic approach and those who favor a cognitive approach . . . and I like to point out that the S-R psychologists, who at one time formed the government, are now in the loyal opposition . . .⁷

In a critical comment on English biologist J. S. Kennedy’s behavioristically oriented book The New Anthropomorphism,⁸ one reviewer—a distinguished biologist who usually knows better—assumed that behaviorism is dead:

If anthropomorphism produces results that are, in the normal manner of science, valuable, it will be persisted with; if it does not, it will be abandoned. It was, after all abandoned once before. The only danger . . . is that scientists can be enduringly obstinate in their investigations of blind alleys. Just think how long behaviourism lasted in psychology.⁹

As recently as 1999, a philosopher blessed with the gift of prophecy could write “Not till the 1980s was it finally proved beyond doubt that although a clockwork toy [i.e., computer] may emulate a worm or do a fair imitation of an ant, it could never match a pig or a chimpanzee.”¹⁰ Evidently, all those artificial intelligence nerds toil in vain, and behavioristic attempts to reduce human behavior to mechanical laws are foredoomed to failure. And, of course, this particular critic wrote before an IBM computer beat the champions at Jeopardy and DeepMind’s AlphaGo became GO world champion!¹¹ On the other hand, to be fair, computers aren’t, in fact, doing all that well with worms and ants either (see comments on C. elegans later). But computer-based models cannot yet be counted out as ways to explain and understand animal behavior.

Behaviorism is frequently declared dead. But although services are held regul...