Foreword

It was in Keller and Schoenfeld’s pioneering introductory Psychology lab at Columbia that I first comprehended the feasibility of an experimental analysis of behavior. Still an undergraduate, I had taken introductory lab courses in Physics, Biology, inorganic and organic Chemistry, and Mineralogy, and I was already disillusioned. Except for Mineralogy, in which we did not really do experiments anyway, the laboratory exercises rarely came out the way we had been led to anticipate. Lab sessions were unsuccessful, uninformative, and dull.

And then, in the behavior lab, lo and behold – the experiments worked. Our rat (each pair of students had one) did for us exactly what its cousins had done for B.F. Skinner ten years before. I had failed in the other labs to observe relations that everyone knew existed between physical variables and other events in the external or internal environments, events which nobody doubted were nevertheless strictly determined. Now I was confronted with a living animal, whose behavior was the subject matter of no deterministic science I had ever heard of, who nevertheless was acting in a thoroughly predictable manner when exposed to specified environmental contingencies.

I needed no long waiting periods before data could be calculated and transformed, no ‘fudge factors’ to cancel out the effects of uncontrolled variables, and no statistics to tease out significant differences. I could see the animal, I could see its behavior, I could observe the behavioral changes while they were happening, and I could often shift the animal’s behavior back and forth almost instantaneously just by making small changes in its environment. Here, finally, was the excitement that involvement in science was supposed to generate.

Nearly 45 years later, I still search for, and continue to find lawfulness in the behavior of individuals. New kinds of orderliness continue to emerge in behavior of ever greater complexity, in subjects with more and more extensive behavioral histories, and in environments which control behavior in important ways that we understand much better than we used to. What has come to be called the Experimental Analysis of Behavior remains viable and vibrant. Its solidity and fruitfulness are evident throughout the chapters of this book.

Research methods must, of course, be appropriate to the questions one is asking, and as we ask new questions, we will need new methodological and technical approaches. In Behavior Analysis, formidable methodological issues still remain to be resolved before we will be able to make substantial progress toward the solution of some important behavioral puzzles. The reader will find many such issues discussed in the chapters of this volume. And yet, despite such problems – which are inevitable in any science – the methodology of Behavior Analysis does possess certain features that give it a unique flavor, features that are responsible for much of its progress and underlie the enthusiasm of its investigators and practitioners. Many of those characteristics, too, are shared with other sciences.

For example, the Experimental Analysis of Behavior largely eschews the statistical evaluation of differences between groups of subjects. Instead, it favors direct experimental control over the behavior of individual subjects. Most experimentation in Psychology, and much of today’s medical research, first sets up control groups that are not exposed to the experimental variable. The investigator then determines whether other groups that have been exposed to the variable of interest differ significantly – statistically speaking – from the control group. Statistical manipulation is supposed to cancel out the effects of uncontrolled variables.

In contrast, before investigating the effects of a variable, the behavior analyst is likely first to establish a stable behavioral baseline in an individual subject. Then, the effects of the variable being studied will be evaluated by observing whether and by how much the variable causes the individual’s behavior to depart from its baseline. The generality of the observed relation between variable and behavior is evaluated not by statistics but by replication – with the same subject, with different subjects, and with the same or different subjects under different experimental conditions. This is control in its best scientific sense; it permits behavior analysts not just to observe differences in central tendency between heterogeneous groups, but to identify the variables and processes which underlie those differences.

In Psychology, except for investigations of the senses in Psychophysics, neither experimental control nor the study of individual subjects has ever been the classical approach. The discovery of controllable behavioral baselines of ever increasing complexity has been a continuing feature of Behavior Analysis, and has enabled behavior analysts to encroach on turf claimed traditionally by various psychological specialties. Behavior analysts have contributed substantially to more different research areas than have investigators in any other behavioral discipline. In other areas in which individual behavior is a datum of interest either as a dependent or independent variable – Pharmacology, Physiology, Neurology, Genetics, Endocrinology, Immunology, Ecology and others – investigators will find the methods and procedures described in this volume to be sources of potentially fruitful application and collaboration of which they may have been unaware.

An exciting consequence of these extensions of behavior analytic interest into other fields is the emerging recognition that the relations can go in both directions; behavioral variables can influence processes that define the subject matter of those fields. For example, differential reinforcement can sharpen sensory thresholds; reinforcement schedules can modulate drug effects; immunologic reactions can be behaviorally conditioned; programmed instruction can ameliorate genetic and developmental deficiencies; coercive control can stimulate endocrine activity; behavioral units can mediate the evolutionary development both of species and societies. Thus, Behavior Analysis has not only developed its own body of data, principles, and investigative procedures, but has made itself available to other disciplines that want to use behavior as a tool for investigating their own subject matters.

Much of the unique flavor of Behavior Analysis comes from its very definition of behavior. Behavior is not simply what an individual does. It is defined, rather, as a relation between an individual’s actions and events in the environment – events subsequent to, concurrent with, and prior to the actions themselves. Thus, the basic units of Behavior Analysis include both acts and their environment. In analyzing behavior, an individual’s actions are defined by their consequences, and the environment, in turn, is relevant insofar as it exerts some influence on the individual’s actions.

It is the inclusion of the environment in the definition of behavior that has led behavior analysts to investigate stimulus control, the development and maintenance of relational structures in which conduct and the environment are tied together into analytical units. This view of the environment as a necessary element in the definition of behavioral units has stimulated behavior analytic interest in many related fields: for example, Psychophysics, in which controlling properties of the environment are related to the sensory apparatus; Pharmacology, in which defining stimuli arise from chemical and other processes within the internal environment; Neurology and Neurochemistry, in which the central nervous system mediates relations between action and environment; and Cognition, in which knowledge is itself defined as behavioral units that are under the control of relations among environmental elements.

Experimenting with behavior for its own sake, or because one wants to use behavior to measure events in other systems or as a marker to make the presence of hidden events and processes visible, is not easy. Readers unfamiliar with Behavior Analysis can learn much about behavior analytic techniques from this volume. There is, in addition, laboratory lore, which one gains by experience or through collaboration with experienced investigators. For example, every species of subject is characterized by its own set of parameters. And so, although the principle of reinforcement will always be found applicable, the particular reinforcers that will be effective, the kinds of deprivation that will be required, the most efficient amount of reinforcement etc. have to be determined empirically for each species. Every laboratory has to work out a myriad of technical considerations about the experimental space in which the subject’s behavior is to be observed, the kinds of behavior to be observed, the techniques and units of measurement, the timing and frequency of experimental observations, the frequency and schedule of reinforcement, and many other details.

But Behavior Analysis is not just a set of techniques and procedures. It is also a set of data and a body of principles derived from the data. As in any experimental science, techniques are not only sharpened by practice but are guided by the general principles of the science. To seek technical proficiency, unaccompanied by a...