Technologies of Behavior in Higher Education

Teaching Machines: The History of Personalized Learning by Audrey Watters. Cambridge: The MIT Press, 2021.

In 1958, in the report from a national conference dedicated to discussions of (and advocacy for) the use of “teaching machines” in the US educational system, psychologist Eugene Galanter asserted, “These machines are a theory of teaching.” And that, in a nutshell, is why anyone who cares about higher education and its future should read Audrey Watters’s Teaching Machines. Far more than simply instruments with which to “deliver instruction” efficiently and effectively to students, educational technology itself constitutes a distinct pedagogical ideology. The signal contribution Watters makes with this book is to explicate how that ideology has manifested itself in repeated iterations of educational “reform” and practice, thus reminding us of the stakes involved in seemingly benign questions of how to “leverage technology,” “find efficiencies,” or “personalize learning” for our students. Few subjects are more urgent for us to consider as we enter this not-quite post-COVID-19 era in higher education.

It’s tempting to say that after several years of pandemic pedagogy, Watters’s book reads differently from how it might have earlier, but to emphasize this difference risks overlooking a central piece of her argument: the animating impulses behind teaching machines in the mid-twentieth century were not substantially different from the present-day tropes of “ed-tech” discourse. The quest to make instruction automatic, to “personalize learning” through shaping individual student behaviors, guided psychologists Sidney Pressey and B. F. Skinner as much as it does Sal Khan, the founder of the online tutorial platform Khan Academy, and Coursera, the massive online open course provider. This vision of automated and adaptive learning, mediated through a machine, rests upon a very specific set of conceptions about what teaching and learning are (and—as important—what they are not). And this is where Teaching Machines brims with insight; it unpacks the guiding principles—the telos—that shaped the teaching-machine movement during its pre-internet heyday. But Watters does far more than simply proclaim “everything old is new again.” Rather, her decision to end the book’s narrative on the eve of the personal computer and internet revolutions of the 1980s underscores the persistence of those principles, that ideological framework, regardless of medium or level of mechanical sophistication, among those who have sought to “individualize” education—a quest which, according to Watters, rests on an “underlying desire to collect data and influence people.”

This emphasis on influencing others was most directly reflected in the conception of teaching machines primarily as what Skinner (who looms large in this history) called “a technology of behavior.” The movement to personalize learning did not have to be predicated on machine technology to instill behaviorist principles, but the fact that it was is telling. Watters’s narrative begins during the post–World War I era, with the initial efforts to find a solution to the perceived inefficiencies of mass public education. Devotees of the gospel of “efficiency” trumpeted by figures like industrial engineering pioneer Frederick Winslow Taylor, educational reformers of that era sought mechanical solutions to the problems wrought by the logistics of a rapidly expanding public educational system. Improving learning—measured through scores on the growing number of commercially available standardized tests—was the desired outcome, and the earliest proponents of teaching machines argued that the only way to achieve that goal was through eliminating wasted time and effort in the delivery of instructional content. One immediately obvious target was the time it took to grade the increasing number of tests students were taking to, as the reformers believed, demonstrate their learning. Indeed, the earliest “teaching machines” were essentially grading machines, a reality that, Watters observes, “highlights how closely [associated] the practices of testing and teaching had become—even in the 1930s and with the introduction of this new educational machinery.”

The conflation of testing with teaching underscores just how powerful the behaviorist impulses undergirding this branch of educational reform were. Much of Watters’s narrative—compellingly written and grounded in a wealth of archival research—describes how Skinner (the originator of behaviorism) sought to create an actual field of operations for “teaching machines,” including his decades-long quest to put his own iterations of this behaviorist technology on the market. One strand of the narrative traces Skinner’s tortuous and frustrated path through a partnership with the Rheem Manufacturing Company, which he hoped would lead to the mass production of a teaching machine designed with the goal of conditioning student behavior through exposure to discrete “frames” of content and then immediate testing (and possible rewards) to see if the student would be allowed to continue through the sequence. Not only were the engineering demands for such a device complex, but implementing the content material (which could potentially contain literally thousands of these “frames”) also proved thorny and difficult.

While it is tempting to scoff at the absurdity of expecting a student to move through fifteen thousand frames of elementary reading instruction, for example, it’s important to discern the larger behaviorist principles animating these designs. Watters points out that simply decrying Skinner’s machines as brute-force operant conditioning—that is, behavior modification through rewards and punishment—ignores the fact that, if you change “conditioning” to the current Silicon Valley parlance of “nudging,” it becomes clear how much of Skinner’s ideas “have made their way into the classroom via our contemporary computing devices.” Even though Skinner’s teaching machines were never the commercial success he dreamed of (and Watters lets us peek into his correspondence and see how this fact bothered Skinner), both they and the prototypes developed by others in this increasingly crowded field created a paradigm in which “their ideas about programmed instruction have become ‘hard-coded’ into all sorts of educational technologies and pedagogical practices.”

Teaching Machines is also a book about capitalism and how its imperatives shaped the development of teaching machines and personalized learning. Skinner’s behaviorism was the dominant paradigm not only because of the scientific clout he possessed but also because of the ways Skinner accumulated that clout—his Harvard affiliation, the income he made from that academic post and the numerous “side gigs” he leveraged using both his own and Harvard’s name, and his racial and gender identity that matched the era’s archetype of the “objective scientist.” This cultural capital was eminently spendable, and Skinner spent it—sometimes assiduously cultivating what would today be called his “brand.” On the other side of the equation, Watters’s discussion of the potential (or lack thereof) that various corporations (Rheem, IBM, and others) saw in teaching machines, as well as the friction between the design and fabrication of the actual prototypes, reveals how any technologically dependent reform initiative will eventually have to reckon with market imperatives.

While it may seem counterintuitive to look to rudimentary Cold War–era technology for a deeper understanding of the current ed-tech ecosphere, that’s exactly what Teaching Machines offers. And its denouement is chillingly prescient. After the eventual failure to effectively mass-market his teaching machines, Skinner split from his contractual arrangement with the Rheem Manufacturing Company. A few years later, in 1971, he published Beyond Freedom and Dignity, which was both the apotheosis of his behaviorist approach and a chilling glimpse into its totalitarian implications. It turns out that Skinner’s obsession with technological solutions to the problems of human behavior was remarkably amenable to what can only be described as protofascism. Watters observes that “his technology of behavior—and that included the teaching machine— was not interested in or committed to freedom.”

The slogan of my main professional organization—the American Historical Association—asserts that “everything has a history.” In other words, even what seem to be esoteric and quotidian subjects have deeply significant stories behind them. As Audrey Watters demonstrates in this gracefully written and deeply researched book, knowing, understanding, and reckoning with the history of teaching machines from the 1930s to the 1970s are far more than an exercise in antiquarianism: they are vital for our current moment. It’s important to note, as Watters repeatedly does, that we shouldn’t conflate understanding with inevitability. The story of teaching machines isn’t simply the march of technology eventually triumphing over all else—it is the story of historical contingency, choices, and even resistance. Skinner and his fellow tech-evangelists may have positioned their machines as the self-evident paradigm for the future of education (as have latter-day figures like Sal Khan and Sebastian Thrun, cofounder of the for-profit vocational-education platform Udacity), but teachers and students alike did not always play their expected roles in the behaviorists’ march of progress. In fact, they often “protested attempts to engineer them, either into enlightenment or submission.” It’s this sense of agency—deployed with the understanding of the larger historical stakes Watters has so compellingly delineated—that we’ll need in order to protect and sustain what’s most important in teaching and learning.

Kevin Gannon is professor of history at Queens University of Charlotte, where he is the director of the Center for the Advancement of Faculty Excellence. His email address is [email protected].