Chapter 1 Summary

In the 1600s, philosophers began to wonder whether humans are machines that operate according to scientific laws. Descartes’s distinction between mind and body held that only the body is controlled by such laws. Specifically, the human body (and all animal behavior) is controlled by reflex action.
Later philosophers suggested that even the mind is governed by scientific laws. The British Empiricists (e.g., Locke and Hume) argued that the mind is a tabula rasa at first, with knowledge being written on it by experience. Complex ideas are built up from simple associations, following several laws of association. Rationalists (e.g., Kant) differed from the Empiricists in supposing that the mind is not initially empty, but starts with certain a priori assumptions with which it actively molds experience.
Meanwhile, in the 1800s, biologists were beginning to learn more about the physiology of reflexes and reflex action. According to thinkers like Sechenov, even human thoughts could be understood as reflexes of the brain. By the turn of the 20th century, all this set the stage for Pavlov’s pioneering work on learned “psychic” reflexes. Processes of major significance could now be studied with conditioning experiments.
In the mid-1800s, Darwin’s theory of evolution emphasized that humans and animals are alike. The early comparative psychologists began to study one of its most astonishing implications: that even the human mind has evolved. To do so, they studied the behavior of animals in an attempt to identify the cognitive processes that they possess. Ultimately, parsimonious principles won out. Thorndike’s experiments on cat intelligence led to the conclusion that learning could generally be understood by knowing how reinforcers stamp in S-R associations. Thorndike’s work also encouraged interest in conditioning experiments.
Watson rescued psychology from the morass of introspection by proposing behavior as its subject matter. The main advantage of studying behavior is that everyone can see it; the facts therefore do not merely depend on what the most powerful people believe or introspect. Watson was also empiricistic and saw a central role for learning. Like others before and after him, he also saw the reflex as an abstract thing so that learned reflexes in animal conditioning experiments were directly relevant to the reflexes he saw in humans in the real world.
At least two forms of behaviorism emerged after Watson. Skinner’s radical behaviorism set out to study the empirical relationships between observable events, such as stimuli and responses. This approach identified two types of behavior: respondents, which are behaviors elicited by events that precede them, and operants, which are behaviors that are controlled by their consequences.
In contrast, Tolman’s operational behaviorism uses unobservable theoretical constructs (or “intervening variables”) to help explain behavior. These constructs are useful provided they are carefully anchored to things that can be manipulated and measured objectively. The main idea of operational behaviorism—that unobservable constructs are useful and scientifically valid if they are systematically linked to behavioral output—is accepted today by most parts of scientific psychology.
After World War II, psychologists began using the computer as a metaphor for human nature, which led to the information processing approach. In the 1980s, the connectionist approach began to use networks of neurons in the brain as its inspiration and metaphor. Both approaches are accepted and used today by modern students of learning in animals.
Modern Learning Theory accepts an overarching framework that can be used to analyze any example of human or animal behavior. Behaviors (R) typically occur in the presence of stimuli (S) and precede significant events or outcomes (O), like reinforcers. Several possible relations can be learned between S, R, and O, and each can occur and play a powerful role. S-O learning is studied with the methods of classical conditioning, which, as we will see in later chapters, has many surprising features and consequences. R-O learning is studied with the methods of operant conditioning and is also fundamentally important. S may also signal the R-O relation, may be connected directly with R, or may motivate behavior based on the R-O relation. In this book, we will consider what we know about each of these kinds of learning and their interrelationships.