Finding Jack Michael’s Footprints in the Sand

Abstract

Jack Michael left a legacy of fine discriminations among technical terms, a careful parsing of theoretical concepts, and a variety of behavioral interpretations of complex behavioral phenomena. When pushing into relatively unfamiliar territory in areas such as “memory,” automatic reinforcement, multiple control, and additivity of response strength, I have been surprised and delighted to find evidence that Jack had once explored the same topics, in the same way, and had come to the same conclusions toward which I was groping.

I was not a formal student of Jack Michael’s. Indeed, for the entire decade of the 1970s, in which I was immersed in the Skinner canon, I was not a formal student of anyone, but behaviorism was my life and in my dreams, so it was inevitable that sooner or later I would meet Jack. By 1980, I had weaseled my way into graduate school, and together with John Donahoe, my advisor and lifelong collaborator, I was drawn to the thorny problems that lie at the edge of our understanding of human behavior: verbal behavior, “grammar,” problem solving, and “memory.” This was Jack’s turf, and I eventually met him face to face in 1987, at a Verbal Behavior Special Interest Group meeting. I found, to my surprise, that there was no social awkwardness, no need for formal pleasantries, no discussion of the weather, and above all, no need to grovel. Jack would talk about behaviorism to a lamppost. He had seen a draft of my paper on memory (Palmer, 1991), and we fell immediately into a discussion of our respective views on the subject, views that to my secret gratification were wholly compatible. From a behavior-analytic viewpoint, “memory” is not a coherent subject matter. The endurance of stimulus control does not, by itself, explain recall; for the latter, problem solving must be invoked.

That was the first of many occasions on which I thought I had been pushing into relatively new territory, only to discover Jack’s footprints in the sand. In graduate school I had made a pest of myself in linguistics classes, waving the operant banner. An important but ignored concept in language acquisition, I had insisted, was automatic reinforcement. A child can learn to whistle or wail like a siren without social reinforcement and can likewise learn to sound out familiar words and intraverbal chains without being coached. Grammar is just a step further: Autoclitic frames are intraverbals with variable terms sprinkled in, and they account for most grammatical regularities (Donahoe & Palmer, 1994/2019; Palmer, 1996, 1998). Skinner’s Verbal Behavior (1957) is peppered with references to the concept of automatic reinforcement, but when I turned to the subsequent literature, I found just one paper explicitly devoted to the topic, a paper titled, appropriately enough, “Automatic Reinforcement: An Important but Ignored Concept” (Vaughan & Michael, 1982). I subsequently learned that, under Jack’s supervision, Mark Sundberg had done some exploratory empirical work on the topic as early as 1977, work that eventually led to a flurry of research on the stimulus–stimulus pairing procedure (e.g., Esch et al., 2005; Petursdottir et al., 2011; Sundberg et al., 1996; Yoon & Bennett, 2000), as well as related work on other conditioned reinforcement procedures (e.g., Holth et al., 2009; Vandbakk et al., 2019; see Cló & Dounavi, 2020, and Petursdottir & Lepper, 2015, for reviews).

Automatic reinforcement can work in several ways: By blowing on a blade of grass, one can make a dreadful squawking noise. The noise itself is aversive, and we would quickly turn off a radio that made such a noise, but a child will work for many minutes to acquire the technique, and successive approximations are reinforcing. The example reminds us that “success,” or “controlling the world,” or “matching what others can do,” although sometimes difficult to operationalize, is usually a powerful reinforcer even if the stimulus properties of the thing produced are not. In language acquisition, pronouncing a word correctly is reinforcing, even if the same word pronounced by another may be neutral or even aversive (Palmer, 1996). The importance of automatic reinforcement in language acquisition, in all of its manifestations, cannot be overstated, and Jack was perhaps the first to realize the power and implications of Skinner’s frequent but offhand appeals to the concept.

The concept of multiple control is especially important in verbal behavior and can be overlooked by those who assume that the task of a behavioral interpretation is to identify examples of elementary verbal operant classes. On this topic, Skinner was both explicit and expansive. In the second half of Verbal Behavior (1957), he treated multiple control both as a subject to be analyzed and as a tool for interpreting other behavioral phenomena. I had made much use of the concept of multiple control before I realized how extensive Jack’s work was on the topic. One of Jack’s greatest strengths was his tendency to sharpen and refine the analytical tools of our field. His careful parsing of the various dimensions of motivation is one example (Laraway et al., 2003; Michael, 1982a, 1993, 2007), his reconsideration of the concepts of positive and negative reinforcement is another (Michael, 1975), his introduction of the concepts of duplic and codic is a third (Michael, 1982b, 1988, 1993), and his distinction between topography-based verbal behavior and selection-based verbal behavior is a fourth (Michael, 1985). In his many classes on verbal behavior, in workshops, and finally in published papers (Hubner et al., 2005; Michael et al., 2011), Jack expanded and refined Skinner’s concept of multiple control. He introduced three new terms, two of which have caught on and are in general use—namely, convergent multiple control and divergent multiple control. The third term, carrier source (on the analogy of a carrier frequency in radio transmission), was Jack’s term for the modality of a stimulus that exerts divergent multiple control over various responses, primarily in verbal humor. The modality of a verbal stimulus is often of only minor interest, but in giving it a name, Jack was resisting the temptation to think of words as abstractions, not responses with physical stimulus properties. In short, I discovered that Jack had clarified and extended a concept that I thought I knew thoroughly.

I discovered another unexpected footprint in the sand only recently, when it dawned on me to start reading footnotes. Jack is rightly credited with being one of the founders of applied behavior analysis, in part because his advisees included Ted Allyon and Montrose Wolf, and in part because his paper with Allyon was the first published example of analyzing applied problems explicitly in behavioral terms (Allyon & Michael, 1959). Wolf is mainly known for his considerable contributions to applied behavior analysis: The names Baer, Wolf, and Risley are engraved in the foundation stone of the field. But my own interest in Wolf’s work has an obscure origin. I have made abundant use of Skinner’s claim that “it is a well-established principle in nonverbal behavior that separate sources of strength are additive” (1948, p. 76; 1957, p. 228). In my various interpretive exercises, I have found the concept of additivity of stimulus control to be indispensable. Any behavior under multiple control—that is to say, most behavior—exemplifies additivity. Problem solving can be understood as the explicit manipulation of variables to exploit additivity: We incrementally bring supplementary stimuli to bear until the target response becomes the strongest response and is emitted. Of course, the concept of additivity is implicit in any prompting procedure as well. But given the importance of the phenomenon, I was surprised to find so little direct experimental evidence in the operant literature. There was ample evidence in the domain of reflexes (e.g., Sherrington, 1906), and Stanley Weiss has demonstrated additivity of stimulus control in a variety of experimental preparations (e.g., Weiss, 1964, 1967, 1977), but the first report that I have found in the operant literature is that of Wolf (1963), published 15 years after Skinner appealed to the concept. Wolf’s paper was extracted from his master’s thesis, in which he demonstrated the additivity of discriminative stimulus control in rats. But it is a rare master’s student with a broad-enough perspective to choose an important empirical topic without guidance, and Jack was Wolf’s mentor and advisor. Jack had been promoting Skinner’s analysis of verbal behavior from the first time he laid his hands on a copy of the William James Lectures in the mid-1950s. Might it have been Jack who suggested the topic to Wolf? A definitive answer to that question is perhaps beyond our reach, but one clue lies in a footnote in Wolf’s paper:

I wish to express my gratitude to the members of my committee, Drs. I. Goldiamond, A. W. Staats, and especially to Dr. J. L. Michael, the committee chairman, for their interest and suggestions concerning the research and preparation of the manscript. (Wolf, 1963, p. 343)

This could be merely a formulaic courtesy, but I suspect that Jack saw the far-reaching implications of that study from the start.

Like many others, I shared Jack’s goal of extending the basic principles of behavior to the whole landscape of human behavior. I have followed the paths that Skinner had blazed, but I have frequently encountered vistas, thickets, and quagmires that remained to be explored, and it was perhaps inevitable that I would see traces of Jack’s passage as I stepped into them. Over the past several years, I have had opportunities to look through Jack’s teaching materials for his verbal behavior class, and I have been astounded at how detailed they were, how many original examples he provided, and how carefully he analyzed and extended the finest of Skinner’s points. It was as if I had discovered the field notes of Lewis and Clark through the human behavior wilderness. I have profited much from my friendship with Jack Michael, but I realize now—wistfully—how much more there was to learn.

Declarations

Conflict of Interest

I have no conflict of interest.