Can it be True?

That was my first, partly panicky, partly skeptical, response to Joshua Rothman’s recent New Yorker article, entitled, on my smart phone’s app, “Is Reading Dead?”

Somehow I hadn’t been aware that reading is sick, let alone dying. Now I’m discovering that it may be on its last legs!

But once I clicked on the article, I found out that Rothman isn’t really worried that reading may be dead, or even dying. His article’s app title was just to grab my attention (which it succeeded in doing). His actual title, obscured by the app title, turned out to be: “What’s Happening with Reading?”

Aha! This sub-title reveals what he’s really wondering about. It’s a much calmer question, and one that the author even gives a kind of answer to in his lead-in: “For many people, AI may be bringing the age of the traditional text to an end.” So things aren’t so bad with reading after all? The real problem is with the demise of the “traditional text”? But why mention “reading” if the “traditional text,” whatever that is, is your actual subject?

Undaunted by these gloomily blurry forecasts, I read on (yes, still reading, despite wondering if it was still possible to do so). I finally uncover Rothman’s true concern: not reading itself or “traditional texts,” but the effect of AI on our reading habits. Although some people, Rothman believes, still like to read, “for others, however, the old-fashioned, ideal sort of reading—intense, extended, beginning-to-end encounters with carefully crafted texts—has become almost anachronistic.”

This, he says, is because AI, capable of shrinking lengthy, complex texts into consumable bite-sized pieces, is just easier to deal with. No more squinting at page after page of fine print, no more slogging through libraries looking up actual paper books and lugging one heavy volume after another to a carrel to poke through and compile endless data. Life is short. Time is money. And you need the facts now. So reading isn’t dead. It’s very much alive. Thanks to AI, the “traditional text” is just being processed in a far more efficient way.

Yet despite its deference to the wonders of AI, there lingers in Rothman’s essay a kind of melancholy about the presumed passing of that ancient, stodgy thing, the “traditional text”—that is to say, of the habits associated with “intense, extended, beginning-to-end encounters with carefully crafted texts.” But Rothman doesn’t make clear why the loss of those old habits should be lamented. “It’s reasonable to argue,” he says, “that some kinds of writing shouldn’t, or perhaps can’t, be summarized.” Rothman doesn’t argue that “reasonable” point himself, though, except to say that if you do go in for summarizing, “you cheat yourself.” Cheated of what, I wonder? He also adds, hopefully, “And surely readers will continue to value the authentic voices of their fellow human beings,” while giving no clue about why we should be sure of such a thing at all.

That’s a shame. For if Rothman had paused to explore why he thinks readers “will continue to value the authentic voices of their fellow human beings,” he could have done a real service, by inspiring all of us, maybe even AI enthusiasts, actually to return to those “encounters with carefully crafted texts” to rediscover what had attracted us to them in the first place.

He could have begun by clarifying what he means by “reading.”

What he seems to have in mind—without sufficiently distinguishing them—are two broad categories: what I’ll call utilitarian reading and close reading.

Utilitarian reading considers words as symbols referring only to other symbols like them. Words are therefore interchangeable, fungible—place holders for each other. AI obligingly sorts the symbols so as to condense and summarize them. Utilitarian reading values summarization because summarization delivers what it seeks, a short cut to a well-defined end. A process that is painless, transparent, and quick. A question asked is one immediately answered.

By contrast, close reading offers not an answer but a presence. Words are not fungible tokens but embodiments of a speaker’s voice. In utilitarian reading we see the words. In close reading we hear them, and behind those spoken words someone speaking them.

Robert Frost put the difference this way:

…the ear is the only true writer and the only true reader. I have known people who could read without hearing and they were the fastest readers. Eye readers we call them. They get the meaning by glances. But they are bad readers because they miss the best part of what a good writer puts into his work.

Utilitarian readers are eye readers. Or, in the newer parlance, AI readers. They “get their meaning by glances,” that is, in AI compacted bundles. They get something useful, true, but they also “miss the best part of what a writer puts into his work.”

An example is in order. Let’s consider a short poem, Emily Dickinson’s “Tell the truth but tell it slant…” (BTW this poem inspired the name of the press for which I am writing this post.)

Tell all the truth but tell it slant —

Success in Circuit lies

Too bright for our infirm Delight

The Truth’s superb surprise

As Lightning to the Children eased

With explanation kind

The Truth must dazzle gradually

Or every man be blind —

If I squeeze this poem through my eye (or AI), I get something like: “Don’t tell the whole truth because people can’t handle it.”

Plausible advice, but not something I haven’t heard before—which is, after all, the point of AI: to collect all that’s been said on this or that topic and compact it. But at least I’m finished. I’ve “processed” Dickinson’s poem and can move on. Life, as I said before, is short.

But no, I’m not done. Now I’m going to close read, or “ear read” the poem. That is, I’m going to get the meaning by listening to the words. And when I do so… things slow down… I begin to hear the words’ rhythm, the alternating lines, first with four iambic stresses, then three, with rhymes on each pair of shorter lines. Altogether like a ballad, a hymn, even a lullaby.

The voice of someone singing these lines begins to form in my ear.

Not singing to a child, though, as I first thought. I’m beginning to feel the force of diction. “Tell all the truth” sounds like a parent admonishing a child, but the word “slant” in “tell it slant” is not a word I’d use with a child, not there, not with “truth.” Even with an adult, “slant” is tricky. In the negative sense of “slanted”? Probably not, but even if “slant” is to be understood as “oblique,” there’s the question of how “all” the truth can still be truth if it is off-center, at an angle. How much off-center? How much of an angle can it be and still be truth?

My reading slows down even further…

Or perhaps a better way to say it is that my close reading expands. I’m still, after many decades, discovering nuances or “slants” in Dickinson’s poem, or maybe they’re slants in my way of reading it. I haven’t reached a final answer about whether I’m reading the poem or the poem is reading me. But at least I feel I’m getting closer and closer to the truth’s “superb surprise.” And that feeling is its own reward.

After getting his PhD in English literature, George Dardess taught close reading to his own students until his retirement. Since then he has been ordained a Deacon in the Roman Catholic Church and written several books on Muslim-Christian relations. He has also created the graphic novel Foreign Exchange.