As I was reading Audrey Watters’s Teaching Machines: The History of Personalized Learning, recently published by MIT Press, the word “landmark” kept occurring to me.
This is a landmark book. It is a landmark book in both senses of the word: one, a marker by which we can establish a present location, and two, a turning point after which we can see a clear change in trajectory.
At least I hope that’s going to be the case, because Watters’s history of personalized learning reflects Faulkner’s aphorism from “Requiem for a Nun” -- “The past isn’t dead. It isn’t even past.” Better understanding this presence may make for clearer decision making for how we integrate technology into teaching and learning.
Reading Teaching Machines is like donning a pair of glasses that suddenly makes much of the present more explicable. This is why I want to urge people to read this book with all possible haste.
It isn’t even a matter of history repeating itself so much as the forces that have governed the pursuit of a teaching machine being ever-present in education atmosphere where humans and their variable behaviors and specific needs are treated as defects that need remedying. The vision of the creators of these teaching machines suggests that if we could just get everyone with the program (pun intended), we could save ourselves a lot of trouble.
Teaching Machines is several books rolled up into one. It is a history of the teaching machines themselves, including chapters on Sidney Pressey’s “Automatic Teacher” of the late 1920s and early ’30s, B. F. Skinner’s multiple attempts through the 1950s, and the multiple space race-related efforts of the ’60s, as education technologists promised that a device straight out of the Jetsons was always right around the corner.
These teaching machines were by their nature, design and philosophies essentially behavioralist, not at all removed from Skinner’s operant conditioning work with pigeons, the bird that has served as a symbol of Watters’s long-standing work at her Hack Education blog.
The hubris attached to each attempt at a teaching machine is mind-boggling, particularly as Watters stacks up the examples chapter by chapter. Men like Pressey and Skinner were convinced that they’d already cracked the “science of learning” and it was just a matter of creating a machine that reflected this discovery. It just so happened that the science they discovered perfectly corresponded with their existing worldviews. They found great fault with the ways schools and classrooms were organized and promised a brighter future for students and teachers alike who would be liberated from the existing tedium of the classroom to their highest purposes -- interacting with machines, apparently.
The sarcasm I’m injecting into my summary is not present in Watters’s original, though the evidence she marshals from an exhaustive dive into the historic materials -- including reams of correspondence from the major players -- paints an inescapable portrait of confidence reaching toward delusion. These men were certain society itself was broken and technology was here to fix it.
These men have always been with us and remain with us today.
Reading Watters’s history had me reflecting on things like Knewton founder Jose Ferreira declaring that he had created a “mind-reading robo tutor in the sky.” Or how about AltSchool, a Mark Zuckerberg-backed venture that was going to “reinvent” school that managed to flush Zuck’s $175 million investment down the tubes.
Remember the One Laptop Per Child initiative that was going to transform global education, but was instead a “spectacular failure”?
The bottomless faith in technology to improve lives is shot through every era, and is particularly pervasive in the post-Sputnik, space race period when the threat of nuclear annihilation combined with the potential for rapid technological change meant we had to throw every ounce of effort at scaling up the use of teaching machines.
My jaw dropped at Watters’s recounting of the vision of Simon Ramo from a 1957 essay on the role of technology in education, because it was simultaneously so unbelievable and so familiar.
Ramo envisioned a totalizing approach in which students would be “registered” into the system, after which a personalized, data-determined path would be followed.
Ramo wrote, “When the registration is complete and the course of study suitable for that individual has been determined, the student receives a specially stamped small plate about the size of a ‘chargaplate’ which identifies both him and his program.”
Ramo continued, (as quoted by Watters), “When this plate is introduced at any time into an appropriate large data and analysis machine near the principal’s office, and if the right levers are pulled by its operator, the entire record and progress of this student will immediately be made available. As a matter of fact, after completing his registration, the student introduces his plate into one machine on the way out, which quickly prints some tailored information so that he knows where he should go at various times of the day and anything else that is expected of him.”
That student’s day will involve some time with teachers but will primarily be in “automated” “push button” classrooms, working with teaching machines optimized for learning subjects like trigonometry.
Ramo believed a student would be less likely to daydream with a machine than with a human teacher. The machines would be “adaptive.” Teachers would benefit, according to Ramo, by having the drudgery handled by machines.
All of this would result in a virtuous circle -- data collected about students would lead to breakthroughs in knowledge about instruction, which would result in more and better teaching machines, an entire industry dedicated to continuously solving the problem of teaching.
Reading Watters’s recounting of Ramo and others who shared his vision brought to mind the Rocketship charter schools, where students spend hours a day on computers engaging with programs. It reminded me of the story of Amplify, a company that promised to map learning for all students to measure their progress against a library of necessary knowledge, only to have its founder remark years down the road, “The map doesn’t exist, the measurement is impossible and we have, collectively, built only 5 percent of the library.”
It also had me thinking also of Theranos, the blood-testing company that promised the impossible -- being able to do hundreds of diagnostic tests on a single pinprick of blood. Former Theranos CEO Elizabeth Holmes is currently on trial for fraud. Her description of what her nonexistent machine could do is no less fantastical than what we hear about the promises of education technology.
While Watters’s book is ostensibly about teaching machines, at its heart it is an exploration of the labor of teaching and learning. For teachers, the teaching machines are always framed as labor-saving devices, but in reality they impose additional burdens of tracking and assessing. Consider the scourge of learning management systems and how many instructor hours are wasted wrestling these systems into submission.
Some of that so-called drudgework is, in reality, a necessary component of effective teaching. For example, injecting automated grading of student writing into the process separates the instructor from the data that are most important when it comes figuring out what kind of assistance a student needs. As I say in Why They Can’t Write: Killing the Five-Paragraph Essay and Other Necessities, asking a teacher to work with students where you know the grade but haven’t read the writing is like asking a football coach to coach a team where he’s given the score but hasn’t watched the game.
It is literally nonsensical if we view the work of the teacher as something vital to the learning process, but this is the story of education technology. The teacher is a problem, not a solution.
If we were to see the actual work of teaching as central to the process of learning -- and how could it not be? -- most of these teaching machines would be viewed as absolutely absurd. And yet the entire history of these machines involves the diminishment and erasing of this labor, which was, not at all coincidentally, primarily done by women.
For students, Watters reveals the ways freedoms are constricted by a system focused on making sure students are prepared to be “productive” members of society. This is the language of Ramo’s 1957 memo when he argues that students must be more technologically advanced to combat the threat of the Soviet Union. It is the same language as “A Nation at Risk,” produced by the Reagan administration, which warned of a “rising tide of mediocrity” among American students that represented a national security threat.
We see it today as students are nudged toward STEM careers with very limited evidence that this results in widespread positive outcomes. Watters argues that a system that embraces teaching machines makes students not individuals with the freedom to carve out their own destinies, but objects to be controlled their behaviors to be shaped. They are simply another problem to be solved, a defect in the system.
It is fair to ask where this has gotten us, a system where teachers make 20 percent less than others with similar education; where students are increasingly stressed and anxious, often about school; where those who pursue postsecondary credentialling are likely to find themselves saddled with debt they may never pay off.
Watters provides an indispensable lens through which to consider these questions.
For me, I wonder if we’ll ever learn the lessons of the past (and present).
 It is both myth and illusion that a reviewer must strive for objectivity when we have the superior tools of accuracy, fairness and transparency to fall back on, so I’ll declare from the outset that I am a fan of Audrey Watters and an admirer of her work, which has been tremendously influential on my own thinking. Take these disclosures however you like, but trust that what I have to say about Teaching Machines is 100 percent accurate, or if not, consider how fun it would be to read the book for yourself and take me down a peg for my bias.
 This is a precursor to the credit card and consists of an embossed metal plate about the size of a business card.
 In an aside Ramo says, “If this proves too burdensome for the student, who will be required to have the plate with him most of the time, then we may spend a little more money on the installation and go directly to the fingerprint system.”
 True fact, Rupert Murdoch was a significant investor in both Amplify and Theranos.