A Simple Test

I just asked Microsoft’s Copilot to write me a 1000-word essay about the normative implications of Quine’s naturalized epistemology, giving it a prompt of less than 20 words. It immediately complied and, within a few seconds, generated a coherent essay that could easily earn a decent grade in an undergraduate philosophy course. What I mean is that if a student had written the same essay under closed-book conditions, given three hours at the end of a course, it would clearly have demonstrated a familiarity with the texts studied (Copilot was able to correctly cite the two key texts that I would have) and an understanding of the issues involved. The exact grade would of course depend on the level of the course and the standards of the teacher, but the student would certainly have had to attend the class and at least skimmed the readings to pull it off.

I mention this, not to counter those who still insist that AI is not capable of doing their assignments, but to answer those that would have us abandon as meaningless any assignment that an AI can easily do. Keep in mind that my little test used a very low-level model (Copilot is available for free to all staff and students at CBS) and my prompt consisted of a one-sentence question along with the instruction to generate a 1000-word essay (it went over by about 100 words). A sophisticated student, faced with a 5000-word term paper at the end of a course they had not followed very closely, would be able to provide a better model with the course syllabus, learning objectives, and even the actual readings. Given a few hours, and assuming an above-average intelligence, they could no doubt cobble something together that would be quite impressive by pre-2022 standards. This ability to fake a semester’s worth of learning over a weekend is the problem we have to have face, I think.

In the future, I think universities will have to make students sit for written exams, on-site and off-line, more often. A degree that does not require at least half of a student’s total grades to come from such performances cannot be taken seriously. In fact, transcripts should make it very clear which grades were earned through homework (where AI support should be presumed) and which were earned through invigilated examination. That is, it should be clear whether the graduate of a given program is capable of writing coherently about their subject themselves. Their future employers can use that information as they please.

The simple test that I propose, then, is a 20-word question with no further context than the course that the student has taken. The student is given three hours and up to 1000 words to demonstrate what they have learned by answering the question to the best of their ability. Understanding the question (and its significance) is itself part of the competence being examined. Under these conditions, I am convinced that the instructor who designed and taught the course can easily determine whether the learning objectives have been met, just as a music teacher can evaluate a student’s ability simply by giving them some sheet music and an instrument to play it on, or a drafting teacher can evaluate drawing ability by giving a student a piece of paper and an object as model. The fact that an AI can also do these things does not make it less impressive when a student can muster their flesh and bones, their brain and their heart, to do it. An education, after all, consists of disciplining the body so as to liberate the mind. It’s important that we show our students what they are capable of.

On Holding Beliefs

Quine suggested that we think of our knowledge as an ever-changing “web of belief”. These beliefs have ontological implications, which is to say that they commit us to the existence of “things” of various kinds, such as furniture, corporations, and even numbers. Some of our beliefs we hold very lightly, others more firmly, and we keep our commitments accordingly. We are not always very explicit about either our beliefs or our ontological commitments — indeed, we may sometimes be entirely unaware of them — but they can often be gleaned from our words and actions even by complete strangers. Granted, there will always be some “indeterminacy” about exactly what we believe and what we think reality consists of. But our peers, at least, usually have a pretty good sense how we parse our experience into objects of belief. After all, they live in the same world that we do and, for the most part, parse it like we would.

An education is both a revision and a disciplining of our beliefs. We not only come to believe things we hadn’t before believed, and stop believing things we once thought were true, we also learn to hold our beliefs more, let’s say, intelligently. Educated people are, ideally, less afraid of being wrong about something they believe. They have experienced it often enough. Having come to believe something through deliberate effort, they know what sort of doubts may be raised. They sometimes face those doubts very explicitly through the criticism they receive from their peers. And they are not afraid of this criticism either. Just as they are familiar with the experience of being wrong themselves, they are familiar with the errors and misconceptions of others. They hold their beliefs in the face of doubt and criticism, at least until it becomes too much. Then they willingly discard the discredited notion.

I said that we may hold our beliefs firmly or lightly. But it is important to remember that while the strength of a belief may be continuous, its attitude is discrete. We believe something or not. We think that something is true or we do not. We may believe something only for a moment, and hold the belief so lightly that even the slightest doubt removes it altogether, but, while we believed, we believed that something was the case. A belief is a “propositional attitude”; it is a particular take on the reality in which we live.

One of the most important lessons of higher education consists in appreciating the contingency of our beliefs. We believe any one thing only because we believe many other things. And that means that revising one belief will often require us to revise others. Holding beliefs intelligently, then, means being careful about our revisions, always considering the implications. There is no shame in refusing to believe something that requires too radical a change in our existing web of belief, even if the evidence for the proposition is, and even in our own eyes, rather strong. In fact, sometimes we are put in the uncomfortable position of thinking something is true that we can’t quite bring ourselves to believe. The problem lies elsewhere in the web, and it will take us a little while to make all the necessary adjustments, to make room for the new among the old. Give it time. And give your peers time to do likewise when it happens to them.

No one has ever been right about everything. That much is probably obvious. What may be less obvious is that the main purpose of an education is not to make you right about as many things as possible. It is to teach you how to be wrong.

Discipline and Discourse

No man ever knows enough about any art. I have seen young men with most brilliant endowment who have failed to consider the length of the journey. (Ezra Pound)

In my last post, I suggested that logic lets us improve the consistency of our beliefs and that such improvement is what science is really all about. In this post I want to emphasize that this is an ongoing, collective enterprise. We do not discover the truth, the whole truth, and nothing but the truth, and certainly not once and for all. Rather we are constantly revising our system of beliefs, adding truths, removing falsehoods, and, no doubt, introducing new errors as we go. That’s true of everyone from the individual student to the whole of science. We manage the process through discipline and discourse, governed by disciplines and discourses.

That’s not just a clever play on words. A “discipline” can be understood as a community of researchers working a range of problems, framed by accepted theories and guided by trusted methods, and the problems, theories, and methods may change over time, but only as a result of the steady and careful (“disciplined”) work of the members of the community. Moreover, they can’t just make discoveries, report them, and expect to be believed. They must persuade their peers that their results are both significant and reliable. That is, they must discuss their results with each other and be open to criticism. This conversation is itself disciplined by what we call “discourse,” which are the not merely logical but also rhetorical conditions under which scholars revise each other’s beliefs about the world.

Students experience this in the microcosms of their minds and classes. Their own personal web of belief is constantly revised as they read, write, and talk about their subjects, and the conversation among them (for they can see each other as “peers” too) also changes as their progress through the curriculum. They are not merely accumulating “truths” one at a time, by addition only. Rather, they are learning and unlearning and learning things again and again, always trying to maintain some semblance of consistency among their own beliefs (again, personally, in their own minds) and the beliefs of their peers (in their classes). Though they sometimes fail to notice it, they are also building craft skills: they are becoming better readers and writers and talkers about their subject, they are disciplining their minds and bodies, fashioning themselves as scholars of their subjects.

There is no end to this process. Science never reaches the final truth on all things and the student never stops learning, even when she makes full professor. And all the while, as science progresses and careers develop, new generations of students are being invited into the conversation, encouraged to change their minds again and again, belief by belief, always (I hope) supported in their efforts to maintain consistency and therefore rejecting some of the things we are trying to teach them until they’ve made the necessary adjustments elsewhere in their web. (I put this hope in careful parentheses because lately I’ve been getting the impression that students and teachers are getting impatient with each other. Some teachers don’t want to be questioned and some students just want to be told what to be believe for the exam. Hopefully, there are still some patient people in academia who understand that forming a belief “correctly” can take a little time, sometimes longer than a semester.)

All research is embedded in an ongoing conversation. It has been going on as long as there has been language, I would think, but since the advent of writing, it has been increasingly disciplined. Modern research is beholden to often very specific discourses that determine the meaning of words and the significance of statements. It even determines limits to what can be said. Fortunately, none of the conventions that govern discourse are absolute and they are likely to change under the pressure of the need to acknowledge particular truths. The good researcher is simply the person who applies that pressure, carefully and consistently, over the course of their career.

Consistency of Belief

I’m reading Wilfred Hodges’ Logic (1977) and I find it resonates nicely with my reading of Quine. Hodges suggests that we define “logic” as the consistency of our beliefs. That is, our a belief is “logical” if it is consistent with other beliefs. It is “illogical” to believe contradictory things, e.g., two things that can’t both be true. But he is careful to point out, right from the outset that beliefs are expressed in sentences and sentences can be ambiguous. We do not learn logic by studying beliefs directly. We learn it by studying the sentences we use to talk about our beliefs. And once we’ve learned the rules of logic, we are in a better position to maintain what Quine called our “web of belief”; we’re better able to trace the consequences of discovering the (inevitable) inconsistencies in our thinking and correcting our beliefs accordingly. That’s how we learn.

It’s also how we do science. Indeed, almost two-hundred years ago, Bernard Bolzano argued that logic constitutes a “theory of science” and should guide us in our composition of scientific “treatises”. Today, we might satisfy ourselves with the principles for composing journal articles, as long as we include the “logic” of reviewing and critiquing them. (Bolzano image of science was much slower and much more stable than ours is today so he felt that, in principle, “the sum of human knowledge” could be gather in a “single book.” It would be a very, very big book, the size of a library no doubt, but it could be written according to a consistent set of rules, namely, those of logic.

Today, we have abandoned all hope that the totality of science can be made consistent with itself, even within individual disciplines. We expect scientists to disagree with each other, sometimes even on very fundamental matters, and for these disagreements to remain open even as science progresses. Scientists are free to believe what they want, to draw their own conclusions from the evidence of their senses. What we hope, however, is that they are committed to an overarching ideal, namely, that their beliefs should be consistent. And we might add to this that, within any particular discipline or paradigm, scientists should be in broad agreement about what to believe. That is, not only is a scientist committed to personally striving for consistency of belief (logical coherence), it is assumed that since the web of belief involved is largely shared within that scientist’s community, the individual beliefs of individual scientists are formed in attempt to be consistent also with the totality of what is known about the subject.

This is a normative principle, not an empirical claim about science. Scientists should strive for consistency of belief, within themselves and among each other, but they will never perfectly achieve it. That’s what discourse is for: to achieve only a “more perfect union” of beliefs. It is the social practice through which we compare our beliefs and hold them to a certain standard. We can talk here of scholarly standards of citation, methodological standards of data collection, even stylistic and grammatical standards. But, at the end of the day, all these standards are realized in particular beliefs about the status of texts, quality of evidence, and clarity of thinking on particular subjects by particular scholars. These beliefs are then brought together and compared under a higher standard, namely, logic.

For my part, when I mention logic to students and scholars (working mainly in the social sciences) I am not talking about the sort of formal logic that Hodges’ book is a such a good introduction to. It is my sense that most scientific disciplines do not explicitly strive for that kind of logical “consistency”. Rather, in science, what is operative is something like everyday reason, even common sense. We expect each other to be “reasonable” about our beliefs and, when claiming that something is the case, to respect both the other beliefs we hold and the beliefs that are broadly held in the community. If we do challenge some of those beliefs (by insisting on others) we must acknowledge the contradiction, i.e., the inconsistency between the beliefs we hold and those that are generally held. “The civil status of a contradiction, or it status in civil life,” said Wittgenstein: “there is the philosophical problem” (PI§125). It is resolved by carefully taking apart and reweaving our web of belief to regain consistence in discourse with our peers.

Why Write?

Recent advances in artificial intelligence might make this seem like a rhetorical question. What, indeed, is the point of spending hours writing your own sentences and paragraphs when a large language model can do it for you in seconds? I suspect, however, that writing now seems pointless to students, and even some scholars, because they were writing for the wrong reasons before the new technology arrived. It is quite common these days for writing instructors to try to shift the focus from the product to the process, and in a certain sense that’s also what I’d like to do, but I think the problem is mainly what we think the effect of writing is. What are we trying to bring about with our writing? What is the change we seek when we write?

One very natural way to answer this question is to think of the reader. We write in order to somehow affect the mind of our readers, whether to inform, enlighten, provoke, or entertain them. The change we seek is in the readership, i.e., the population of people who read our texts. At a broader level, we might say that we’re trying to influence “the conversation”; we’re trying to change the way a topic is talked about, shift the agenda, raise new questions, reframe the issues. After our text is published, we imagine, people will have accommodate its rhetoric. We may not yet have changed their minds, but something will at some point have to give. We have made a contribution.

A much more cynical answer is to focus on the name we make for ourselves when we write. We write in order to occupy a position in the discourse, to become a recognized “author” on the subject. In this category of answers are also things like writing to sell books and writing to build your list of publications. Writing to be cited would also count in this category, as would the students quest for a good grade. To have have such reasons is not in itself shameful, but it doesn’t point you toward a solution to the problem of writing well. There are too many “tricks of the trade” that will reach these kinds of goals but have little to do with the quality of your writing.

Importantly (for this post, anyway) is that both sorts of reasons are also increasingly going to be reasons to let AI do your writing for you. Language models are only going to get better, and since they are trained on the very discourse you are either trying to affect or impress, their competence to reach your goals may often be greater than your own. If these are your main reasons to write, then, you will find them more and more useful.

The obvious alternative that I’m heading towards is to seek reasons to write within yourself, rather than in your environment. Write for the clarity it brings or the pleasure it affords. Write because it improves your mind, not the minds of your readers. In the future, as most of the prose we need to get by (the prose that stores and transmits useful information) is produced by machines, we will write for the same reason that we swim, rides bikes, jog, go to the gym. It will be something that we enjoy doing (most of the time) and makes us better able to accomplish (and enjoy) our other activities. It will keep us mentally — indeed, spiritually — healthy. A serious scholar (and a serious student) will attend to the reading and writing much as serious athlete attends to their diet and exercise.

It will also open our thinking to criticism from qualified peers. But that is something I harp on about often enough. Today, let me just remind you that learning that you’re wrong is still learning. Enjoy that too.