To Jacques Barzun’s dictum that “All systematic devices for generating good writing are a mistake” we may now add a caution by Harvard’s President Drew Gilpin Faust about the related topic of machine-graded writing. She is justifiably worried that grading software will miss “irony and elegance” or, for that matter, anything “it hasn’t been programmed to see.” The critical approach of MIT’s Les Perelman has been to have satirical fun with grading software. These eminent cautions matter because we seem to be moving towards a swarmteaching paradigm that will require the adoption of mechanized and mediated inspection of students’ knowledge. What if it doesn’t work very well?
That even Harvard MOOCs are not immune to the pressure to impoverish discourse and evaluation in favor of mass grading is shown by the effort of Professor Gregory Nagy to produce his MOOC CB22x on the ancient Greek hero. How will he grade thousands of students? By assigning multiple-choice quizzes instead of papers. Two difficulties, which I have treated in the past, present themselves. One is that the grade for the sort of qualitative understanding needed for such a course will turn on fine distinctions that cannot be justified. At some point the difference between Excellent (A) and Good (B) may come down to one answer on one multiple-choice question. That is absurd.
The other problem is that the questions themselves remove material from the purview of maieutic discourse where some of it belongs, placing it squarely (and incorrectly) in the domain of simple recognition. An example of this possibility is Professor Nagy’s question about whether Zeus’s will is to cause the Trojan War or the Iliad. This would be an interesting question for the chance it would give a class to discuss the problem of agency in the ancient world, using as examples Zeus and the lesser gods. Students might note that Homer himself invoked not Zeus but Calliope to “cause” his poem by inspiration, and wonder how her agency and Zeus’s collaborated or collided. They might shape these ideas in a well-written, polished essay after the discussion, drawing on their knowledge of the Iliad. But not in CB22x, where the answer is simply that Zeus willed to cause the Iliad. Professor Nagy’s claim that multiple-choice questions are almost as good as essays and discussion appears to be disproven by his own example.
The second component of machine grading is that students will be required to “read” on-line texts, and that their “reading” will be “checked” by seeing how they “annotate” it. Recall for a moment Jack Torrance in The Shining, typing away at his manuscript in the Overlook Hotel. His wife hears the typing and thinks he is working. In fact, he is typing again and again, “All work and no play makes Jack a dull boy.” Students who have adopted Jack’s philosophy on a MOOC could tag & paste this saying at random on their “reading” to show the nanny machine that they are “doing” it. Who is going to know otherwise in a blizzard of thousands of “notes”? Not Professor Nagy. By contrast, two good questions asked in a live class could reveal sloppy or undone reading. I know: I have used them to do so. It is harder for a student to decide to bamboozle a live teacher than a machine nanny. One morning in my Modern Poetry course in college, Professor Koch sniffed out massive neglect of a homework assignment. He went around the class asking students questions that they couldn’t answer. Finally the usually genial Koch burst out angrily with “Why don’t you read the goddamned poems?” After that, we did. We did.
Third will be “online discussion boards.” Just what we need: classroom discourse like “Comments” columns, home of what Shakespeare called the reply churlish. (There are exceptions that prove the rule.) But churlishness is not the only problem: Kevin McGrath, a coordinator of CB22x, sputters, “You have a group who are—they talk about Christ, or about pride. They haven’t really engaged with what’s going on.” In a classroom the reply churlish could be checked, and the pronouncement pompous punctured. In a MOOC, who will rout the rubbish?
An answer may be in kinds of hybrid courses being tried out at various places and examined in the New Yorker article linked above[1]. The examination must be very careful and ready to detect wishful thinking and tendentious reporting. Dr. Faust’s worries are justified, and perhaps the answer will turn out to be the one Amherst has given: a No to MOOCs with Harvard. That should be no surprise. Robert Frost, who taught poetry at Amherst for a number of years, thought talking about poetry and marking for the understanding of it a very subtle thing. One of his students reported a class where a classmate gave a stupendous reading of a poem they were studying. Frost looked at him and said, “You get an A forever.” No MOOC will ever say that.