LOOKING FOR JOY IN ALL THE WRONG PLACES

by Roman Sympos

 

Back in January, in a “View” entitled “AI and Anger Management” (https://www.romansympos.com/anger-management-1) I addressed the issue of generative AI and our tendency to fetishize it, to think of it and treat it as though it were a real human being, with purposes, intentions, desires, and feelings. I got around to examining the fatal flaws in the assumptions underlying Alan Turing’s famous “Test” of whether computers can think (they can’t), and concluded with several questions about AI, the answers to which, I still believe, are all “No.” For example, “Does AI ever have to ‘manage’ its anger? Does it have any anger to manage?” and “Does AI ever get bored?” This month’s “View” is a spin-off of January’s. Its sequel, next month, will include a nod to the monthly essay on Aeolian harps that appeared in the inaugural issue of Sympos, published in July of last year, “‘Perform thy task untouched, alone’: Aeolian harps as home entertainment”(https://www.romansympos.com/perform-thy-task-untouched).  Both items are available on the “Archives” page. And if my description of Star Island intrigues you, you might enjoy our current Poem of the Month (https://www.romansympos.com/poem-of-the-month).

 

 

Part 1: “ALEXA, WHAT’S THE MEANING OF MY EXISTENCE?”

 

A few weeks ago I attended a series of lectures on generative AI delivered by the theme speaker at a week-long family conference on Star Island, one of the Isles of Shoales situated five miles off the coast of Portsmouth NH and owned, jointly, by the Unitarian Universalist Association and the United Church of Christ. The island is the site of the Oceanic Hotel, a large, rambling, three-story Victorian structure built in 1875, which serves as Star’s main conference center and living quarters. The Oceanic faces the ocean (as you might guess) across a broad, green lawn, with a veranda as wide as the passenger deck of the QE2 running around two and a half sides and several “cottages” attached, all subdivided into cubby holes called “rooms.”

 

Once the summer refuge of Boston Brahmins fleeing the city’s stifling heat, the facilities have been updated over the years with little detriment to their grandeur, historic value, and picturesque beauty, to the point where Star Island is now a showplace of innovative green technology. (Engineers take the hour-long trip out just to admire the array of solar panels out back.)

 

The lectures took place in a small auditorium next to the hotel dining room.

 

Our speaker was Nick Zufelt, Instructor in Mathematics, Statistics, and Computer Science at Phillips Academy in Andover, MA, who holds a PhD in Mathematics from the University of Texas at Austin. His title was “Purpose in the Age of Artificial Intelligence” and at first, because the weather that week was looking to be exceptionally fine, I was tempted to skip the talks. What did I need to know about AI that I didn’t know already? It was inevitable and a menace to civilization, and I intended to ignore it as long as possible by relaxing with a good book in a rocking chair on the veranda.  But my spouse reminded me that, despite our spotty track record over the half century we’ve been coming here, we always gave the theme speaker at least one chance to catch our interest.

 

We went. I was hooked.

 

Nick’s lectures were entertaining, engaging, and profound. Also hilarious. If you can imagine Robin Williams giving a TED talk, you’ve got the right idea. It didn’t take long for conferees to realize that the most intriguing computational machine in the room was Nick’s brain.

 

Among the dozens of counter-intuitive items flung out in the course of that week, like sparks from a fuse, was this: AI makes mistakes and it’s a good thing it does.  Most of us knew the first part of this statement was true. Few of us agreed with the second part. Do we really need accidental misinformation clogging up the arteries of mass communication when there’s already enough deliberate misinformation to do the job?

 

By “mistakes” Nick meant what he called “hallucinations.” If you’ve fallen into the habit of treating generative AI as a person, mistakes like these can give you the impression it’s dropped acid. For example, asked to complete the participial phrase, “foraging in its native ________,” Chatbot GPT might (because it makes its selections based on probabilities) choose “forests” even when the subject is “octopus”: “Foraging in its native forests, the octopus. . .”

 

BEEP!

 

We’d all encountered such errors when using generative AI, and that was Nick’s point: we’d learned to be skeptical. The biggest danger posed by AI was its ability to supplant rather than support thinking, and the most potent weapon we had to defeat or at least mitigate that threat was skepticism, applying what we already knew to our assessment of what AI was telling us, which is to say, thinking twice: once when posing a question, and again when evaluating the answer.

 

That’s why it’s important, said Nick, to ask AI questions about things within what he called the “proximate” range of what you already know. Reach too far and you run the risk of sharing AI’s hallucinations. (If you don’t know that octopuses live in the sea, you’re in trouble.)  But if you don’t reach at all, you’ll learn nothing important.

 

He gave another example of AI hallucination to make sure we got the point about proximate knowledge.

 

Nick has always wanted to play the guitar and recently took some steps in that direction. He bought a guitar and spent a few hours strumming it and trying to learn some chord progressions. At one point, he asked AI a question about the D minor chord. AI came up with an answer, but it included an erroneous statement about which scale the D minor chord was built on. Nick had enough proximate knowledge of music theory (in college he’d started out in music education, majoring in voice) to recognize the error.

 

But I was struck by something else. The last thing I’d do when learning to play a musical instrument was ask AI a question about music.

 

Why was that? I wondered.

 

The simple answer was that you learn to play a musical instrument by getting someone to tell you how to play it and then doing what they tell you. Over and over. Until you get it right.

 

I can imagine asking AI how to play the trumpet, which I learned in fourth grade from a teacher at a local music academy who visited my elementary school one afternoon and offered us free introductory lessons on any wind instrument of our choosing. His name was Mr. Samuelson, and he had a really cool soul patch. I chose the trumpet because it was shiny and loud and, when I picked it up, I found I could already make a sound resembling music. Also, I could play it using only three fingers.

 

I can imagine AI doing everything Mr. Samuelson did when he gave me my first lessons: showing me, animatronically perhaps, the right fingerings for certain notes, telling me how to produce one note or another by tightening and spreading my lips (my “embouchure”) on the mouthpiece, how to tongue the note, how to breathe, how to read musical notation. I’m sure AI could even provide useful feedback at every step, offering a detailed list of do’s and don’ts tailored to my shortcomings and level of musical development. In short, like Mr. Samuelson, AI could show and tell me how to play the trumpet.

 

But knowing how to do something is not the same as acquiring the know-how to do it.

 

Knowing is something we do with our brains—conceiving, imagining, remembering, calculating. For example, knowing how to boil water is “knowing that you boil water by putting water in a pot and heating it to 212 degrees Fahrenheit.”

 

Know-how is entirely different.

 

When I was a Boy Scout I desperately wanted to earn my Second Class badge, which required swimming the length of the pool at our local high school unassisted. I could tread water and dog paddle for a minute or two, at most. However, I’d read all about how to swim. (I was an avid reader.) When it came time for my scout master to examine me, the pool was closed for repairs, so he just asked if I knew how to swim. I said, “Yes,” and explained how it was done. He nodded and checked it off the list.

 

“Knowing how” is a kind of thinking, and the test of clear thinking is explaining, not doing.

 

But as any musician, outfielder, tap-dancer, painter, photographer, writer, platform-diver, or short-order cook can tell you, real know-how is not a type of knowing or thinking at all, and you can’t acquire it simply by being shown and told how to do something, no matter how good your retention rate may be.  Real know-how is a skill, what the Greeks called technê, as distinguished from intellectual kinds of knowledge or epistēmē.

 

You don’t acquire a skill by thinking about it. Whether you are learning how to improvise on “Straight, No Chaser,” or how to field a line drive, or tap out a cramp roll, or use a painter’s knife, or crop a photo, or find exactly the right word, or nail a reverse 4.5 somersault in the pike position, or flip a pancake in mid-air, thinking, more often than not, only gets in the way.

 

You acquire a skill by training your body to think for you. Learning a foreign language is a good analogy. You want to reach the point where you can focus on what you want to say instead of how to say it, and you don’t become fluent until you can stop thinking about grammar and start focusing on whether you’d like red wine or white. You’ve trained your lips and tongue to handle the vocalization part.

 

Language learning can help us understand what Nick meant by “voice” when he announced, on one exceptionally fine morning when I could swear I heard a rocking chair on the porch calling my name, “You should use Generative AI to support your voice, not supplant it.”  He was referring, proximately, to students using AI to write their essays or their algebraic proofs for them, but a few minutes of discussion revealed he had wider vistas in mind. “Voice” involved not just writing, but thinking, because the two are mutually dependent, and it encompassed not just committing words to paper, but verbalization in general, because preliterate societies could clearly think, and in quite sophisticated ways.

 

I could see where we were headed, but we hadn’t time to get there before the hour ended. By “voice” Nick meant language, and languages in general, mathematical and body languages as well as verbal and—why not?—musical and pictorial and gustatory (cuisines and recipes) and olfactory (perfumes and scented candles) and architectural languages and, as he put it at one point, “everything you do.”

 

Expanded to its fullest extent, what Nick said is this: “You should let Generative AI support the meaning of what you do in the world, not supplant it.” His idea of “voice” captures something that I’ve felt, intuitively, since the moment I picked up a trumpet and blew that first, tentative note: that the meaning of our existence is embodied and not intellectual, something rooted in what we do and say, not in what we think, and certainly not in what anyone or anything else, AI included, can do or say or think for us.

 

There’s a coda to my trumpet tale that illustrates the dangers of mistaking knowing-how, which AI is very good at explaining, for know-how, which AI is utterly incapable of imparting.

 

I stuck with the trumpet after fourth grade and got better at playing it, to the point where, like Nick, I ended up as an education major at a midwestern university school of music.

 

I was enrolled there because my high school band leader encouraged me to apply and playing the trumpet brought me joy. I had reached a pretty advanced stage of proficiency and was ready to devote my life to it.

 

My sojourn lasted one semester. Piano was my Waterloo. Music education requires proficiency in several instruments, not just one. (Mr. Samuelson’s primary instrument was the clarinet.) I hated piano lessons. They brought me no joy. Spoiled by my mastery of the trumpet, I found them tedious and practicing scales sheer drudgery.

 

But I hadn’t forgotten what I’d learned from my swimming test. Unfortunately.

 

Instead of training my fingers to play the assigned exercises, I memorized the fingering patterns for all the scales and chords and practiced only to the point where I could work my way through the exercises by stopping to think whenever I wasn’t sure what came next. Had generative AI been available in 1967, I might have asked it to show me the D minor chord.

 

The barely passing grade I received for the course was a gift of amazing grace that I didn’t deserve, and that I paid forward by leaving music school at the end of the semester to enroll in the school of Literature, Science, and the Arts. There I found joy in mastering other skills than trumpet playing and, eventually, made a career of them.

 

Now retired, I’ve put them aside for the most part.

 

But I still play the trumpet, and it still gives me joy.

 

(Next month, Part 2: “Alexa, Give Me Joy.”)

 

 

Part 2: “ALEXA, Give Me Joy”

​

In Part 1 of this essay, “Alexa, What’s the Meaning of My Existence?” (https://www.romansympos.com/view-from-the-precipice) we examined the difference between “knowing how” and “know-how,” or what we might call brain-thinking and body-thinking. The first is cognitive, the second corporeal. We acquire the first by committing to memory what we’ve read or heard about the way to do something, and the second by doing it repeatedly, “according to the book,” until our bodies have been trained to do it without our having to think about it.

 

But isn’t “not-thinking-about-it” how accidents happen?

 

Not at all. When you stop thinking about what you’re doing (or rather, what you are letting your trained body do), you don’t necessarily stop paying attention to it. Paying attention isn’t the same as thinking about what you’re doing. It’s living completely in the moment and at the precise spot your body’s doing something. It’s watching to see that the task at hand is going as intended. In our analogy to learning a foreign language, it’s paying attention to the person you’re speaking with and listening to their response to see if you’re making sense, rather than thinking about the subjunctive.

 

At the same time, you are anticipating how to reply based on what’s being asked or implied, or where you want the conversation to go. You are inhabiting the moment’s full potential, or rather, letting its full potential inhabit you, and having the patience to let it show you where it wants you to go next.

 

Thinking as a disembodied form of intellection is your mind inhabiting some place and time other than the here and now, like back when you were just learning how to do what you’re doing and had to look at an illustration, or jumping to the future, when you’ll have to stop what you’re doing and start on a different task, like preparing dinner. That’s when you get “absent-minded” and prone to accidents—you miss the nail and hit your thumb or lose the thread of a conversation.

 

Know-how is trusting your body to show or re-mind you of what comes next. Or what might come next, depending on what you want to “say” at that moment.  It might not be what you planned to say before that moment arrived. It might even come as a surprise to you.

 

In short, know-how is immersing yourself in the moment of creation, or what the late University of Chicago psychologist Mihali Csikszentmihalyi called “flow.”

 

“Flow,” writes Csikszentmihalyi, is an “order in consciousness” that occurs whenever a person “must concentrate attention on the task at hand and momentarily forget everything else” (6). “Flow” relieves the self of the “social controls” and instinctive desires of everyday life— love, food, sex, status—which are serial, diffuse, distracting, and in the end, unfulfilling. Such desires create “psychic entropy” (39), or disorder, and feelings of anxiety or boredom because their aim is passive gratification, mere “pleasure” rather than the satisfaction of accomplishment. Flow arises from “autotelic” or self-directed activities that have no goal other than their own achievement.

 

If you’re looking for joy, flow is a good place to start.

 

The “elements of enjoyment” that characterize flow, according to Csikszentmihalyi, include “a challenging activity that requires skills” (49), “the merging of action and awareness” (53), “concentration on the task at hand” (58), “the sense of exercising control in difficult situations” (i.e., a sense of mastery) (61), and especially “the loss of self-consciousness” (62)—not blacking out, but an unwavering focus of attention on what needs to be done rather than on the person doing it or the motives of the persons hindering it or the tools necessary to accomplish it, which seem, in any case, almost to move under their own power.

 

AI can tell you how to do something. It can even do it for you, saving you the time and trouble of learning how to do it at all. But it can’t provide you with know-how. For that reason, it can’t give you joy.

 

And, on some level, we are aware of this because, every day and in countless ways, AI is taking away our sense of mastery and the joy that comes with it.

 

Here’s an example, courtesy of Nick Zufelt, whom you met in last month’s “View from the Precipice.” He was the theme speaker on the topic of “Purpose in the Age of Artificial Intelligence” at a conference I attended in August. The article from which Nick drew his example was written by James Walsh for The Intelligencer and is entitled “Rampant AI Cheating is Ruining Education Alarmingly Fast.”

 

“I really like writing,” she said, sounding strangely nostalgic for her high-school English class—the last time she wrote an essay unassisted. “Honestly,” she continued, “I think there is beauty in trying to plan your essay. You learn a lot. You have to think, Oh what can I write in this paragraph? Or What should my thesis be?” But she’d rather get good grades. “An essay with ChatGPT, it’s like it just gives you straight up what you have to follow. You just don’t really have to think that much.”

 

It would be easy to focus on the word “think” in this passage. Writing your own essay, says this young woman, “you have to think.” When you let AI do the writing for you, “you just don’t really have to think.” Isn’t this what parents and teachers fear most about introducing Large Language Models and generative AI into the classroom? That they are destroying our kids’ ability to think, or in Nick’s terms, “supplanting” rather than “supporting” it?

 

Well, yes. But that’s true of nearly every new technology going all the way back to the invention of writing, which Plato, in Phaedrus, indicted for, in effect, “supplanting” our power to remember. Why bother to memorize all 30,000 lines of Homer’s epic poems when you can read them whenever you want? (Or have your literate slave read them to you?)

 

I’d rather concentrate on the word “beauty.” “There’s beauty in having to plan your essay,” says Walsh’s student. She’s sensitive, in other words, to the form of what she creates—its overall shape, its rhythms, patterns, balances and counter-balances, its arc and order of presentation—and misses the joy of mastery that once enabled her to impose some measure of control over these beautiful things, to make them her own creation, while motivating her to look for challenges that, ultimately, would improve that mastery. “You learn a lot,” she says, when writing your own essay. You learn how to create beauty.

 

This is what the joy of learning looks like when we stop confusing know-how with merely knowing how.

 

But why indict “Rampant AI Cheating” as the master criminal when, to judge by this young woman’s lament, rampant grading has been ruining education at least since the invention of the SAT in 1926, and probably long before? Students like Walsh’s are only following the script for advancement they’ve been handed by their teachers and parents and employers from their first day of kindergarten, and in some cases, pre-school.

 

Why do teachers grade their students’ work when the net effect is to deny them the joy of learning? Ask any half-dozen educators picked at random and they’ll tell you they hate grading as much as their students. Press a little harder and they’ll admit they do it only because they’re paid to. And why are they paid to?  Because our schools are charged, from Kindergarten on up through graduate school, with the task of sorting the desirables from the undesirables and ranking everyone in between according to their perceived aptitude for “success,” which mostly means, success at doing their jobs, or even finding any. School is modern society’s Human Resources Division.

 

This is how learning loses its autotelic joy, the gladness and excitement and sense of accomplishment that comes with mastering a skill and looking for challenges that will make you better at it just for the sake of doing it. Grading forces the innate joy of learning to give way to learning as a means to an end—“getting a good job,” “getting ahead in life,” “making money,” “building a career,” even “attracting a good marriage partner.”  Grading instrumentalizes learning.

 

But let’s back up: how is learning a bodily activity? It involves, proportionally, a great deal more mental activity than fielding a line drive or nailing a perfect ten from the diving platform. In fact, learning is often entirely a process of thinking, isn’t it?  Memorizing, understanding, imagining. . . . what do you have to train your body to do in order to learn that the square root of 2 is pi? Surely, when Czikszentmihalyi talks about “a challenging activity that requires skills,” he’s not talking only about stuff like woodworking or making omelets. Or, if you're Nick Zufelt, playing the guitar. If learning is a skill—and it is—then thinking about what you are doing is far from an impediment to achieving “flow.” It’s nearly all that learning, as we generally conceive it, consists of.

 

The most succinct answer to the question, “How is learning a bodily activity?” is “It involves a part of the body called the brain.” The question itself reminds us that when we say a physical skill requires “training the body to think for us,” the body in question includes the bundle or network of neurons in the brain that controls movement, as well as the nerves connecting that network to the muscle groups we’re training. That integrated neural network is the part of the body we’re training to think on its own, to the point where we no longer have to think about it.

 

Skills can be mental, then, as well as physical, and both, when mastered, can bring joy. But there’s no clear dividing line between the two and neither of them is an all-or-nothing proposition.

 

Memorization, for instance, is training a part of your brain to master the skill of recalling numbers, speeches, and events by repeatedly going over a mnemonic routine, like associating the initial letters of the items you want to remember with an acronym or, as the classical orators did, mentally placing the items in the rooms of an imaginary house. These techniques in particular involve visualization, an imaginary form of sensory—that is, corporeal—perception. You practice the mnemonic until you don’t have to labor so hard to call these items to mind.

 

But memorization may involve the body more explicitly and employ other senses besides the visual, as in a rhyme—“Thirty days hath September,” and so on—or a song, whose chiming patterns and repeating melodies make it easier to commit to memory. Often, especially at an early age, you begin by singing the song at the outset, along with a little dance: “Head . . . shoulders, knees and toes . . .”

 

The most common way the body is involved in mastering mental skills is, of course, manually, by writing down an idea in words or numbers or by drawing diagrams or illustrations to help in mental visualization, a crucial skill not only in memorizing, but also in understanding abstract relationships. Writing more than a sentence or two (I’m not speaking of copying) often requires revision and re-arrangement, “sculpting," as it were, your words and sentences and phrases as they appear on the page and smoothing out the seams between them the way a good mason would point a brick fireplace. I assume that Edith Wharton experienced many moments of joy when she was hard at work writing and revising and blocking out The House of Mirth, however difficult the challenges—indeed, precisely because those challenges offered her opportunities to experience the joy of finding just the right word or conceiving exactly how the next scene would play out. I’m just as certain that James Walsh’s student, if given the chance to do so without the Sword of Grading hanging over her head, would find joy in writing about Wharton’s book.

 

If you are lucky in life, the skills you are most eager to learn when you are young, whether predominantly mental or corporeal or balanced somewhere in between, eventually add up to a career or a profession that brings you joy, regardless of the income you may earn. But most of us are unlucky, or we make the wrong choices. When I was a teacher and advising undergraduates, nothing gave me a deeper feeling of dread than a student who came to my office and said they chose to major in something because it would give them more money instead of more joy. This is what comes of the instrumentalization of joy that begins with grading: a life of tedium, if not immiseration.

 

In any case, and in every case, the joy of mastering a skill is directly proportional to the difficulty of the challenges it presents at every stage of advancement and the opportunities it opens up, at every incremental step along the way, for more complex—more beautiful—forms of creative self-expression. Whether you are a pin-ball wizard or a chess grandmaster, a platform diver or a quantum physicist, a luthier or a basement bird-house builder, joy awaits you, but only if you refuse to settle for knowing-how and work at acquiring, and continuing to acquire, the know-how.

 

And Alexa can’t do that for you.

 

CODA:

 

Our current infatuation with Generative AI and the wonders it can perform fits perfectly into our modern obsession with efficiency. There’s nothing we love better than saving time and trouble, by which we mean, generally, doing things ourselves. Why shop at a bricks-and-mortar store when you can shop online without the trouble of getting in the car and fighting traffic and finding a parking spot and discovering the store doesn’t have what you want after all? It takes much less time to stay home and “let your fingers do the walking,” as they used to say in the Yellow Pages ads. (My apologies to Gen Z: the ads were discontinued while most of you were in diapers, so you’ll have to look it up. You know where.)

 

Our love of efficiency in general originated in the early stages of the Industrial Revolution as a growing attachment to and dependency on “labor-saving devices.” In the factory, this meant replacing wage laborers with machines that required fewer workers to operate them and turning those workers into parts of the machines they tended, doing the same elementary and repetitive tasks all day, every day, week in and week out.  “Labor” in the abstract, as a cost of doing business, was thus “saved” at the expense of the real laborers who were either put out of a job or reduced to automatons.  

 

In the home, labor-saving devices like the washing machine and the vacuum cleaner saved time and trouble for working class females (whether mothers or daughters) and money for upper class females. Servants could work more efficiently, which meant you could pay fewer of them to do the same work or, eventually, hire a cleaning service to stop by once a week.

 

Somewhere along the line, “labor” came to include the idea of doing anything yourself. We became consumers, not doers. We began searching for gratification instead of mastery, pleasure instead of joy.

 

In the first issue of Sympos, published in July of last year, I included an essay on the Aeolian harp, an oblong wind harp popular in England from the late 1700’s to the mid-nineteenth century, the period when the Industrial Revolution was beginning to pick up steam  (https://www.romansympos.com/perform-thy-task-untouched). Unlike the mechanical music boxes that appeared at about the same time, the Aeolian harp required no winding and its melodies were unique, albeit limited by a severely restricted sonic palette. Just place the instrument in a sash window on a breezy day and let Nature entertain you.

 

Until the advent of such devices, there was no way to hear music of any kind unless a human being played or sang it within your range of hearing. People made their own music. From kings and princes all the way down the social ladder to servants, peasants, and children, they acquired the skills to do so and found joy, and pride, in mastering and using them to entertain themselves and others.

 

Today, music education, along with arts education generally, is moribund in school districts throughout the nation, if not already dead. Hearing live music is a rare treat and seldom offered gratis. Even subway buskers leave their instrument cases gaping expectantly. And when was the last time you were asked to bring your viola da gamba to dinner?

 

None of us can pinpoint the day the music began to die. But the music box and the Aeolian harp were the first streaks of its bleak and dismal dawn.