Professor-Robots: The Worst 19th Century Idea of the 21st Century

There was a time before TSA and 9/11  when crazy people wandered freely around the nation’s airports. I was heading to the Eastern Airlines gates at O’Hare when I was stopped by  a guy in a suit wearing a sign that said there was now mathematical proof that the country was going to Hell in a hand-basket. I slowed a beat but it was enough for him to shove a magazine in my hand. “You should read this!” he said. “They don’t want you to know about it!.”  I glanced at the cover. It was Fidelio, Lyndon Larouche’s magazine of culture and science. “Great,” I thought. “NCLC propaganda.”

I was ready to bolt for the Eastern passenger lounge, when the guy tugged the magazine from my hand and flipped it open to a Larouche rant about Georg Cantor and transfinite numbers. My next mistake was to say something like, “Well, this is a load of crap.” I had taught set theory and for a brief instant I imagined that I could  waste a little time toying with the guy before pounding him to intellectual  pulp. It was the opening he was looking for.

I was hopelessly over matched. Never mind that I knew that he had no idea what he was talking about. Norbert Weiner, the Reimann Hypothesis, Plato, and  negative entropy and were all smashed together, mixed and reshaped as a serious critique of western political economy. He was getting louder and more aggressive, so I grabbed my bags and headed down the concourse at full speed with this guy and his sign chasing after me yelling in a vaguely threatening way.

I can’t stay away from encounters like that. I have had other run-ins with zealots who are shameless about misquoting, misapplying, and misappropriating stuff that I happen to know something about. I want to unmask them in public. Show them for the frauds they are. I tell you this story to prepare you for my comments about the worst abuse of a deep and very important idea that I’ve seen in a long time. It may end badly, but I can’t stay away.

My attention was drawn to a recent  CCAP post about unnecessary cost escalation in higher ed by Andrew Gillen who is inspired by Charles Babbage to suggest that there be a “division of labor” in which the job of lecturing is separated from the job of grading.  The result according to Mr. Gillen would be a 30% reduction in costs that can be passed along to students in the form of tuition reductions.

I am a computer scientist, so Charles Babbage is intellectually speaking a friend of mine, the originator of the very concept of a computer.  Nobody knew how to mechanize better than Charles Babbage. His book, On the Economy of Machinery and Manufactures, is aimed at the upwardly mobile assembly line workers of 19th century England. It is an explanation of manufacturing economics for the common man.

In 1850 London’s mills and factories were houses of horror:

Any man who has stood at twelve o’clock at the single narrow door-way, which serves as the place of exit for the hands employed in the great cotton-mills, must acknowledge, that an uglier set of men and women, of boys and girls, taking them in the mass, it would be impossible to congregate in a smaller compass.[P. Gaskell, The Manufacturing Population of England. London, 1833, pp.161-162]

I won’t recount all of Babbage’s arguments beyond saying that the mechanization of, say, the production of sewing needles–in which the dangerous and expensive alignment of parcels of metal so that the ends can be sharpened before they are separated can be done more cheaply by a primitive robot–has virtually nothing to do with the craft of being a university professor.  Yet here is what Mr. Gillen takes away from the Babbage text:

In higher ed, we need high [sic] skilled people for some aspects of the job (designing courses, creating assessments, mentoring, etc.) but many of the other tasks they perform don’t require as much skill, and could be performed much more cheaply by lower paid workers (routine grading, administrative tasks, most office hour questions, finding and dealing with cheating, even some of the teaching). And yet all tasks are performed by the most highly paid people.

Mr. Gillen is right that de-skilling is important in controlling costs.  It is a problem to be solved in healthcare, government and education, but not in the way that Mr. Gillen suggests. I’ve seen this argument enough in the last few weeks to know that there might be trouble brewing, so let me restate some points that are well-known to readers of Innovate.EDU:

1. There is simply not that much expense tied up in grading exams. It may be a pain in the ass for professors to plough through mountains of blue books, but the productivity gains of separating lecturing and grading are negligible.  How do we know that?  The experiment is being conducted every day in departments that make heavy use of teaching assistants as graders.  If there were great productivity improvements to be had, they would have been noticed already.  They have not.
2. Mr. Gillen’s  separation of labor concept actually makes matters worse for learning. An example of how it might be done is Salman Khan’s brilliant idea of inverting lectures and homework. Unlike Gillen’s proposal, this invests more professor time in one-one interaction with students, not less.
3. Universities are not factories. Most of the problems facing higher education today can be traced to this false analogy.
4. There is no conceivable educational benefit of decoupling the feedback that a professor gets from personal interactions with students.  Even if you are committed to a time-honored systems of lectures and exams,  simple questions like whether  the material is getting through cannot be answered by shipping students off to a specialized team of graders and routine question-anwerers.

In fact, Gillen has it exactly backwards.  The De-Skilling Argument is this:  let’s take the least value-laden part of teaching out of the hands of high-priced professors. The most valuable part of teaching lies in the personal interaction of students and mentors and in the peer-to-peer interactions of learning communities.  Why would anyone want to depersonalize that?

The issues raised by this notion are both legion and obvious.  How would you know in advance which office hour questions were routine? Why should delegating routine administrative tasks — whatever those might be — to unskilled labor lead to savings anyway? Don’t learning management systems/course management systems purport to do the same thing already? Why would this proposal not lead to increased costs as new, unskilled workers are added to the university workforce?

It’s not a fair fight, but the CCAP proposals are sometimes the only ones out there — and they are listened to — so someone needs to shout “Well this is a load of crap!”

Now, where’s that Eastern Airlines lounge?