The Bookshelf: How Do We Read? Why Do We Read?
I don’t have a strong memory of not knowing how to read, or of learning how. When I entered first grade in 1965, I could read already because my mother had taught me. (This was my first schooling—pre-school and kindergarten were not universal in those days.) I have been more or less compulsively reading ever since. Not just books in my home and in school, but shampoo bottles in the shower, cereal boxes at breakfast, road signs and billboards on the highway, magazines in waiting rooms—and now, inevitably, text on screens.
I have written here before about the different experiences of reading physical books on one hand and e-books on the other—that informally surveying people I know has confirmed my own experience: what one reads in the former is more firmly retained in memory than what one reads in the latter. But screens—the computer on which I’m writing this, the tablet I carry around the house, the smartphone I carry everywhere else—bring to my eyes a colossal proportion of what I read aside from books. Daily news and opinion journalism, online scholarship, email, and social media are all delivered straight to my eyeballs. From the serious to the silly, there it is, all the time, if I will but look.
Our ubiquitous devices facilitate all manner of addictions—to porn, podcasts, videos, games, gambling, gossip, conspiracy theories, bigotry, rage, or just the enervating ennui of acedia as we doomscroll the hours away. But even if one avoids all these occasions of sin, the internet floods the zone of our available reading content with a constant freshet of brand-new distractions, many of them actually very interesting. It’s like being a dog in a forest full of constantly multiplying squirrels. The chase seems exhilarating at first, but is exhausting.
Commentators such as Jonathan Haidt, Clare Morell, Antón Barba-Kay, and Christine Rosen have discussed this “always connected” phenomenon, its ill effects, and what to do about it (especially what parents and educators can do). But I won’t say more about their books, because I’ve been online too much to have read them.
In all seriousness, as someone old enough to remember life before the personal computer revolution, let alone the internet or the smartphone, I don’t need Haidt and company telling me how much my perceptual world has been changed by these developments. I know I have to resist the phone and tablet, setting them aside and picking up (real) books for substantial parts of my day. If it’s hard for me, I can only imagine how hard it is for young people who cannot remember life before the smartphone. It is no surprise to hear it reported that teachers are requiring less and less reading of students at all levels of education. It’s a paradox of plenty: our devices have delivered to our eyes an unprecedented quantity of reading material, and our attention is now too shattered by a million small inputs to concentrate on one long thing and see it through.
Next month I will be back in a college classroom to teach for the first time in four years, and the latest worry is the “artificial intelligence” revolution. (I insist on the scare quotes because only the adjective and not the noun is accurate.) The AI revolution threatens our capacity for reading well, as Karen Swallow Prior would put it, because it seems to promise relief from the necessity of reading at all. The shortcut is there: our questions, AI’s answers. The CEO of one of the leading firms in the field says its latest version “feels like talking to an expert in any topic, like a Ph.D.-level expert.” Welcome to obsolescence, I say to myself—until I remember that “feels like” is not the same as “is.”
Our AIs are more precisely called LLMs, “large language model” programs hoovering up everything they can sink their bytes into, processing it, and producing consumable outputs we can see or hear. In quantity of memory, access to information, and speed of compilation and (apparent) analysis, the LLMs outstrip us already. Do we still have an edge as AI’s creator? I think we do. But we can keep it only by exercising our minds in the encounter with other minds, not by asking the LLM—which is not a mind and does not know other minds—to do it for us.
As we are made in God’s image, so the LLM is made in ours. The image, of course, is not the same as the thing of which it is an image, but something inferior to it. And God’s self-knowledge is perfect; he could make us with as much or as little understanding of ourselves as pleased him to provide. Our self-knowledge is imperfect; we do not fully understand ourselves, least of all our mind and consciousness. The human mind made the computer, but is not itself a computer, and we do not know how to make a mind. The LLMs are trained on enormous amounts of information already produced by human beings, including the written record of our moral judgments, principles, and sentiments; I have seen no evidence that AI can fashion any such artifacts of the human experience that are altogether its own. AI is our creature, and if it ever endangers us, the failure or fault will be ours, not its.
But this is all a prologue to saying what AI cannot do for us, as teachers, as students—as persons. It cannot understand for us. As my friend Charles T. Rubin recently wrote, of the sudden onset in the last few years of AI-reliant cheating among students: “No student would buy a robot, take it to the gym, put it on an exercise machine, and believe that he or she was going to become fit. Why turn your schoolwork over to an AI and expect to become an educated human being . . . ?”
How do we read? Why do we read? The how of it cannot be, “Instead of reading Hamlet and discussing it with my classmates, I asked AI to tell me about all the commentaries written on the motives and mental state of Shakespeare’s protagonist, in 2,500 words, so I could turn in my paper on time.” Even if the resulting essay earned an A, and even if the student had read the play and was merely using AI, with progressive queries, as a shortcut to find and refine the analysis most appealing to him, he will not have done the work that leads to understanding the play and the character himself.
The why of reading is in order to understand. Anyone who’s taught for a while has had students come to him and say, in appealing for a higher grade, “I’ve worked so hard, shouldn’t that count for something?” No, we say to such students, the result is what matters in the end, not the work performed to get the result. Now, in the age of AI, I expect we will soon hear (perhaps this is already happening), “my results are impressive, so why does it matter how little work I did?” The answer must be that the results are not one’s own without the work that generates understanding.
I don’t have any easy solution for the pedagogical challenge posed by AI. Some combination of small classes, tech-free classrooms, seminar-style discussion, oral presentations, handwritten exams (the bluebook returns!): I know all these things are being tried. But students are not in the classroom most of the time. Will they read Mansfield Park for English lit? Democracy in America for their poli sci class? Only if we can make the case that they will benefit by it, that understanding what such books tell us about our condition is a good to be attained for its own sake, and not just for a grade, a sheepskin, a credential.
And it may help students if we talk about how we read in mundane practical ways. Where are we sitting when we read? Are we at a desk, the book on its surface, a strong light on it? Or in an easy chair, a rocker, stretched on a sofa, lying prone under a tree on a lovely day? Do we have music on, or read in silence? Do we highlight passages in the text, underline in pencil, scribble with ink in the margins—and in what colors? Do we leave sticky notes in special places we want to return to? (Or these “book darts” that someone recently gave me, to my everlasting wonderment at never having heard of them before?) Do we take notes the first time through, or only the second time from our marked-up copy—or none at all? As we prepare to write, do we build an outline first, or start writing by hand before turning to the keyboard, or just plunge right in? It’s been said—an LLM could tell me by whom, if I bothered to ask—that one doesn’t really know what one thinks until one starts trying to set it down in writing. Or, I would add, in speech, as one attempts to explain one’s thinking to another person. Are we thinking about writing when we are reading, or does that come later?
Are there certain kinds of reading we do in different places, or at different times? Work for class in the morning, research in the afternoon, leisure reading in the evening, or what? When we read, where is our phone or tablet or computer? Within reach, separated from us, (deep breath now) turned off? If still on and within reach, how firm are we about not looking at it for extended periods? What do we mean by extended period? (Must check X after every chapter or I’ll go crazy. Chapter? Every page!)
Every age has its distractions, and its temptations for a shortcut. (When I began teaching, it was TV, and pamphlets called CliffsNotes.) These are our distractions, and our temptations. But if we are to understand ourselves and our predicament, we have the same resources our predecessors had. We have books to read. We will grow or shrink as human beings according to our willingness to read them.
Image licensed via Adobe Stock.