On account of the pandemic, there has been a seismic shift to remote or hybrid instruction. However, long before COVID-19, forces to harness instruction to technology were at play within the American school system. In Teaching Machines: The History of Personalized Learning (MIT Press, 2021), Audrey Watters masterfully explores this story and explains the consequences technology has had on the nature and architecture of American schooling.
As many policy analysts know, Watters has been writing incisively about this important topic since 2010 on her blog, Hack Education: The History of the Future of Education Technology. With Teaching Machines, she cements a decade of lucid, riveting commentary.
In this NCSPE excerpt, in particular, Watters establishes the foundation for her analysis with a depiction of the first efforts of the Harvard psychology professor B.F. Skinner to mechanize learning. Skinner would go on to develop several teaching machines during the Sputnik era and beyond. Watters explains how he incorporated his work as a behaviorist into the design of these learning devices.
Skinner, along with other progenitors of teaching machines such as education psychologist Sidney Pressey, aimed to pioneer the automation of pedagogy, “freeing the teacher from the drudgeries of her work so that she may do more real teaching, giving to the pupil more adequate guidance in his learning,” Watters writes. In doing so, they prioritized the interests of private entities looking to engineer systems of learning at the expense of teachers and school leaders who aimed to engender more democratic modes of education.
The posture of automated learning presumes that the work of evaluating student responses and guiding them to new levels of understanding can simply be outsourced to a programmed device and does not require the nuanced touch of a seasoned practitioner with deep content knowledge. Yet the word “assessment” itself derives from the Latin assidēre, meaning “to sit down to.” The role of the teacher to sit beside children and listen deeply and intently, not only to learn how students are approaching a particular task, question, or problem, but also to hear from them about what piques their curiosity about the work at hand and motivates them to persevere. Watters deftly details how even the most well-designed or well-intentioned teaching machines fail to achieve this. She moreover describes how critics of Skinner such as Paul Goodman raised these concerns as they saw these devices dehumanizing the educational process.
“Who, then, will watch the puzzlement on a child’s face and suddenly guess what it is that he really doesn’t understand, that has apparently nothing to do with the present problem, nor even the present subject matter?” Watters quotes approvingly from Goodman’s 1960 book, Growing Up Absurd. “And who will notice the light in his eyes and seize the opportunity to spread the glorious clarity over the whole range of knowledge; for instance, the nature of succession and series, or what grammar really is: the insightful moments that are worth years of ordinary teaching.”
Even with advanced programming and interactive computer displays, personalized teaching machines or programs may not be able to elucidate nuanced understandings of difficult concepts with struggling learners. Independent work with these programs is often unsupervised, and students may receive unauthorized assistance to particular questions instead of actually supplying their own authentic response. A recurring issue with struggling learners is also the motivation to complete the tasks themselves. Extended independent assignments often, in fact, result in fatigue and non-completion for students who are still building task stamina.
Watters also writes about the very challenges of implementing such programs, where private demands for technocratic control over the levers of schooling have clashed with the needs of actual practitioners and students. As we see in contemporary education settings, Watters documents that programs were often rolled out in a hasty and haphazard fashion, unsupported by research evidence demonstrating their effectiveness or appropriateness for students and without adequate levels of teaching training or adoption.
Programmed instruction in the form of teaching machines as well as the modern incarnation of computerized learning engines, Watters likewise makes clear, represent a highly systematized and standardized form of education that collides with more progressive, constructivist, and student-led pedagogical methods. They also reify practices and norms within school systems that promote a highly functionalist model of education, where students are fed bits of information as they are trained to complete discrete tasks serving little more than the informational needs of private companies.
While programmed learning systems and algorithms aim to provide individualization and personalized learning, Watters demonstrates how they can conversely serve to stifle creativity and individual expression, on the student, teacher, school, and system level. “These technologies foreclose rather than foster possibilities,” Watters writes.
For longtime followers of Watters’s blog, which is now on hiatus, Learning Machines will fulfill all expectations. For those who haven’t read Watters’s blog, this excerpt should pave the way to reading the book. Agree or not with Watters, readers will be glued and challenged.
Steven J. Koutsavlis
Research Associate, NCSPE
November 22, 2021