I know I’ve written before about the pedagogy that a friend of mine wanted to adopt in teaching introductory biology to college students. He wanted to recapitulate the history of biological science from the Enlightenment forward. Students would look in a simple microscope with no reference point or explanation of what they were seeing, as if they were Leeuwenhoek and Swammerdam. They’d look through mid-18th Century naturalists’ drawings and try to imagine taxonomy as if for the first time. They’d do a dissection as if anatomy was a novel concept. They’d do Mendel’s pea experiments and try to think it through.
He obviously didn’t do a course with quite that level of ambition—you can’t grow that many peas in just a couple of weeks, among other things. (I suppose if the professor planted the peas enough in advance…) You can teach some of the idea of Mendel’s experiments with some organism that is much easier to grow in a lab in a time-relevant way, but that’s the problem with this course idea: to show the principles derived from the history of biology in a way that fits a conventional semester, you have to use contemporaneous scientific knowledge and laboratory skills. That exposes the deeper problem with the idea, which is that there’s got to be a radical compression of experiences and knowledge-making that took centuries into 14 weeks, at which point the luxurious thought of re-experiencing the making of knowledge as if you were doing it for the first time stops making much sense.
I was sort of in love with the idea anyway because it’s really a historian’s idea. Almost no scientists think that if you re-ran the sequence of scientific inquiry, you’d discover an unexplored branch of research. The contingencies that they might agree exist would mostly be about efficiency, about getting to the proper conclusions earlier. So many scientists today wouldn’t see the value of trying to reclimb the shoulders of giants by re-enacting the experience of becoming a colossus. I think for my friend the value was not in re-thinking science itself; it was a pedagogical conceit that students would be more motivated to do science if they discovered it for themselves, from first principles.
I got to thinking about this ambition again while reading Columbia University undergraduate Owen Kitchizo Terry’s essay “I’m a Student. You Have No Idea How Much We’re Using ChatGPT”. Terry provides a clear and intelligent outline of how to use an LLM-based AI to simplify the labor of writing a simple undergraduate essay. He sagely observes that asking GPT to actually write the essay is a dumb move: the result will be a mediocre, generic, dull piece of writing riddled with dangerous factual errors that are likely to be detected by an alert reader who knows the content. What he suggests instead is using GPT as a kind of writing assistant: ask it to generate a good argumentative premise, then ask it to generate a good structure, then ask it for a good quotation or example of the points in that structure (which the student writer can check for accuracy).
Terry suggests that the worry that universities should have about this usage is not academic dishonesty in any sense, but simply that students will not have to think for themselves in completing writing that is meant to push them to do so.
Terry is persuasive, at any rate, that there can be a skilled usage of GPT in student work that is indetectable by any conceivable means. That isn’t likely to soothe the panicky mood among faculty worldwide (check out this story about a Texas A&M Commerce professor making a truly dumb mistake in trying to police GPT usage).
Me? I’m not at all worried about the usage that Terry describes. I might even endorse it. First off, what he describes is not that different from a long history of similar aids for writers, especially creative writers. I’ve got a set of cards from The Story Engine upstairs that are designed to function more or less like GPT in Terry’s scenario; the 1920s book Plotto was a much earlier work designed to do something similar. If a student came to see me after being assigned the prompt that Terry uses in the essay, I’d probably help them find an argument in the same way GPT does, only it would take longer and require more effort from the student.
It’s the effort part that I know many faculty are focused on. But a lot of those dialogues with students about what their argument should be in an upcoming essay feel like Sir Bevidere instructing the villagers on methods of witch-detection. The student knows the professor is hiding a bunch of acceptable arguments inside himself, but they also know they can’t just hit the professor like a pinata and have them all spill out. So we do the little dance: can you not also build bridges out of stone? what also floats on water? And if she weighs the same as a duck? At the end of it, most students are no wiser about why the argument the dialogue has produced is a good one than they would be if they just whacked the GPT pinata directly and got the goodies right away.
This is where my friend’s idea about teaching introductory biology comes in. In four years, what we’re trying to teach college students is how to do knowledge, not just to hear about what is known. So we have to catch them up to the edge of the present while also trying to inform them about how we got there, because the work of knowledge-making in the present incorporates knowledge-making in the past. Sometimes we very expressly reference and narrate that history (thus making it easier for a newcomer to absorb) but more often it is implicit. You need to know it to be knowledgeable, but we’re not going to step you through it as it happened, because we want you to be here right now, trained to take the next steps.
No wonder in that sense that almost all students, whatever their abilities and experience, hesitate to think when asked to read The Iliad and write a prompted essay about it. I don’t care how much work a professor puts into the discussion, how cunning their close readings are, how much good literary criticism you assign, there is no way a student in a few weeks—maybe a semester—is going to find a way to think about the Iliad that passes muster as contemporary knowledge production. We know it, the students know it. So most of them realize that they have only three hopes to flourish on that essay: say something that passes for literary criticism that’s been said a zillion times before, only somehow in their own words and in a way that seems to be convincingly arrived at through the student’s own thought; say something that is naive as literary criticism but has the virtue of being stylistically expressive and intellectually contemporaneous; say something that passes as baby’s-first-steps knowledge production by virtue of choosing a new object to place in relationship to the Iliad, however inexpertly so.
So maybe a student in the first case tries to grapple with the strangeness of Achilles’ new generosity to Agamemnon in Book 23 and his cruel mistreatment of Hector’s corpse in Book 24. Success! The student has demonstrated thoughtful reading. But if the student is really thoughtful, they also know they cannot possibly be the first person to write about that observation. They’re just demonstrating capacity—it’s a training exercise in an apprenticeship, an emulation of knowledge production. In the second case, maybe the student has some kind of direct experience of war and cannily uses the epic as a framing device for a personal essay or for some kind of expressively personal philosophical engagement with war. Success! But that’s not really the cutting edge of knowledge production in literary criticism. (Well, it can be, but we tend to limit that kind of work to people who’ve first produced an acceptable piece of scholarship of a more ordinary kind.) In the third case, maybe the student says “I’m going to compare the Iliad to the videos and frontline journalism coming out of Ukraine”. Big success! Brilliant! If they do it well, but a lot of the time that paper is just going to reveal that the student understands that the Iliad is about war and maybe knows enough to surface a few things about prophecy, hubris and anger.
No wonder students just want us to tell them what to write, a lot of the time. They’re sensing that when we ask them to think about a much-thought thing as if they’re thinking for themselves and of themselves, there’s something intrinsically a bit phony about it all. Imagine trying to binge-watch Game of Thrones for the first time right now while you’re surrounded by people who lived through its first airing. If you mention what you’re doing, even if you scream ‘no spoilers!’ at everybody around you, you’re going to still sense that they’re past that, that you’re heading for some preordained opinion that has moved on into other experiences.
Our students are as eager as we are to catch up so they can be part of the conversation, to do the work of the now. The problem is that this impatience lurks behind every learning experience we offer. If I take students on a tour of social theory from Descartes to Deleuze, it’s going to be remarkably difficult for me to not constantly render that trip in terms of its final destination, to turn that sequence into nothing more than a kind of literate precondition of knowledge making. The student who goes on that tour and wants to go off and read more Saint-Simon instead of Marx in order to derive a new radical theory of class parasitism applicable to 21st Century life is doing something really interesting but I can’t possibly stop the tour and go with her. At best I will have to tell her “hang on to that thought, it’s pretty cool—do something with it later on”. The foreknowledge of where knowledge has been and is (apparently) going kills most of the moments where students might in fact be really thinking something interesting and productive until or unless they have that thought at the preordained point on the scaffold where something can be done with it. Up to that point, I’m just trying to find the person wise in the ways of science who can shout “A duck!” at the right moment, a student who will present a thought as if it is novel that the scholarship has already thought. In trying to be the duck-sayer, many will tentatively mumble, “Churches? Apples? Very small rocks?” knowing already that those are wrong, but perhaps hoping by chance to find one of the answers inside my pinata-brain.
The goal of getting students to the point where they can make knowledge that advances a particular field of inquiry on a particular subject is an absolutely worthy one. And if we are going to make that a goal, given the costs of instruction, we need to do it with some degree of efficiency. We reason, with some validity, that in a lifetime of practice, we will be able to go back and dredge up all the readings and thoughts we had to put aside, we will think of many other things that float on water that Sir Bevidere wasn’t even considering.
What I wonder sometimes, though, in the spirit of my friend’s wished-for class, is whether we might not try, now and again, to allow our students to think as if nothing had ever been thought before. To start with some really basic questions about the world and ourselves as if we didn’t already know about four thousand years of attempts to answer them. Why is there something instead of nothing? War, what is it good for? Why do bad things happen to good people? What is the sun made of? What are those things I’m seeing in the microscope? In some fields of human endeavor, we’ve realized that good things happen when untrained people tackle something that we have elaborate systems of training, or when trained people work to see something naively. I feel as if we need a protected way to do something like that in college pedagogy: to teach and study occasionally as if we’re in no hurry to catch up to a preordained conclusion, or to maneuver a jet to the runway in a crowded airport.
I'm with you most of the way here. I agree that the standard is not "have an original thought", at least not until a student has committed to being a fully professionalized scholar who has to make some claim to pushing the state of knowledge forward. But the point is that our pedagogy often is considering student work in relationship to the thoughts that have been thought--sometimes accidentally, sometimes on purpose.
So yes, sometimes we really are able to have a kind of clean room where we all read a significant work where all thoughts are welcome and it doesn't matter if they've been thought before. I had a marvelous experience this last fall with a number of discussions like that--some students who'd never read Genesis before and had no real familiarity with Christianity who had some really interesting initial responses to what they were reading, for example. Since we weren't reading it in a class *on* Christianity or scripture, I didn't need to hurry students along towards knowledge about theological knowledge.
But a lot of the time, we're eventually going to be trying to synchronize what students write and think with what has been written and thought. Imagine trying to go into a 'clean room' class to think from first principles about kinship, where all thoughts were welcome and none were burdened by needing to be either faithfully referenced to existing knowledge or original arguments about existing knowledge. Where would we start? First, perhaps, from personal experience. We could share testimonies: this is what family and kin relation means to us. This is what we hear other people saying. We could begin to work outside of that from source material first: kinship in literature, in religious scripture, in laws and policies, in popular culture. But at some point, if the goal at the end of 14 weeks is for students in the class to be conversant with the way kinship and family have meaning in history, sociology and anthropology, I'm going to have to cross to the scholarship, and at that point I'm going to be producing a history of what the terms have meant and how they've been worked and an account of how that history has led to present projects and interests. Knowing that this is where I'm going produces a pedagogical pressure on me that will intrude on the open-endedness of the beginning of the course and a foreknowledge on the part of the students that there are going to be some "right" answers in the back half of the course that the first half has to lead towards, which is going to make them reluctant to think from first principles in an unguarded way. It may also make everyone end up feeling: well, why did we muck around with thinking this through as if there is not already a lot of thinking going on? I think it would be really valuable to do so--that's my friend's proposition about teaching biology--but I think it's hard for many people to embrace doing it.
Does it matter if we think of something already thought before, so long as we thought of it and it’s a good or an interesting thought? Trying to be absolutely original in a world with thousands of years of thinking (including all the people who thought but wouldn’t or couldn’t write down what they thought) is not a viable endeavor. What I wanted my students to do was to experience the joy of thinking—not to my destination but to theirs, whatever that might be. How does AI help them to make that effort? Not convinced.