woc 2: more miscellany

April 2024

Seeing and Being

Ernest Hemingway, known for bluster, wrote that the experience of war was a great advantage to a writer and that writers who had not been to war and who belittled the subject of war were “always very jealous” for not having had “something quite irreplaceable that they had missed.” The passage is found in The Green Hills of Africa and is little more than a passing comment on the work of Tolstoy. It has always raised a distasteful envy in me, as I was one who escaped the Vietnam War draft by just a year or two and joined in with the Peaceniks, but felt—because of Hemingway’s passage—that some crucial honing of my writing had been missed. To be sure, I have never belittled the subject of war, and even as one who has never seen it, I portray it in several of my works, published (and unpublished).

Hemingway saw more than a fair share of war, but never as a soldier. Famously, he was wounded while handing out candy bars on the front lines near the Piave River in Italy as a volunteer ambulance driver in World War I. A mortar shell exploded near him, riddling him with hundreds of shards. “I died, then,” he would later say. “I felt my soul or something coming right out of my body, like you’d pull a silk handkerchief out of a pocket by one corner.” As a correspondent, he reported from the front lines of the Spanish Civil War and World War II. He even took his yacht out on patrols to look for German subs off of the Florida coast.

I suspect, though, it was his skill as an empathetic observer more than his experiences, per se, that gave him the “something quite irreplaceable” that allowed him to write convincingly about men and women at war, and the healing they sought after the war.

Many extraordinary war novels have been written by those who experienced it directly: Erich Maria Remarque, Kurt Vonnegut, Tim O’Brien—just to name three. But convincing war novels have also been written by authors who have had no direct experience with war: Stephen Crane, Chimamanda Ngozi Adichie, Aleksander Hemon for example. These writers have not observed battle, though Adichie’s and Hemon’s families were gravely affected by war.

Experience and observation are important for writers, but these alone are not enough to write convincingly. The crucial “something quite irreplaceable” for writers of any subject is a sensitive and sensible use of the imagination. “Imagination is more important than knowledge,” Albert Einstein is quoted as saying in a 1929 The Saturday Evening Post article. The full quote is even more inspiring, for the great scientist connects science and art: “I am enough of an artist to draw freely upon my imagination. Imagination is more important than knowledge. For knowledge is limited, whereas imagination encircles the world.”

Though he may have been wrong about quantum entanglement, calling it “spooky action at a distance,” Einstein is right in emphasizing the use of imagination in the creation of anything— scientific, literary, social or personal. But even here I insist on subtitles. The use of the imagination by itself, without the tempering complements of knowledge and empathy, is what leads to conspiracy theories, demagogueries, inequities and other insanities. Experience, observation, imagination and empathy are all critical skills for writers. And now I bluster: No significant art or science is made without the complement of all of these, and to paraphrase First Corinthians, the greatest of these is empathy.

Among the easiest skills for writers to learn are the technical and critical ones. These are usually taught in workshops and are developed through practice. Like physique building or sports training, the practice must be continual. Life-long. Observation, to some degree, is a teachable skill, but imagination and empathy are more difficult to teach. Yet they can be learned. They are learned from the inside out—and they are honed by technical, critical and observation skills. As the elements of good writing are interrelated and synergetic, so are the skills of a good writer; and like the hard skills of technique, the soft sills of observation, imagination and empathy must be practiced.

One of my favorite writing texts is not about writing at all, but travel. It is The Art of Travel by the philosopher Alain de Botton. Among the chapters about anticipation, exoticism, and curiosity as they relate to travel, are discussions on observation, beauty and the use of imagination. In particular, De Botton introduces a satirical travelogue called A Journey around My Room written by the Eighteenth century aristocrat Xavier de Maistre. De Button proposes the exercise of "traveling” around one’s bedroom as an experiment designed to revitalize observation. “Home…finds us more settled in our expectations,” De Botton reminds us. Travel on the other hand, is receptive. “We approach new places with humility. We carry with us no rigid ideas about what is or is not interesting.” Carrying this receptive mind-set into a familiar setting, then, defamiliarizes the setting and allows refreshed engagement with the details of the setting and the emotions that come with them. De Botton recounts that de Maistre rediscovered the qualities of his sofa, and “remembers the pleasant hours he spent cradled in its cushions, dreaming of love.”

I enjoy this exercise because it practices the use of imagination while sharpening the powers of observation. I often uncover details in the room—typically not my bedroom—that I wouldn’t have noticed in my daily bustle. The guise of the traveler allows me to imagine things of my ordinary experience as exotic, the way I might see the green cross of a European pharmacy, or hear the sing-song of a French ambulance. Sometimes, I take on an alien persona, pretending I am an ET exploring Earth culture, or perhaps I am a historical figure—I like Ben Franklin—suddenly transported to the future.

Whatever the persona, the exercise is to observe, to look closely and to discover. Such observation is done with all the senses—smell, taste, sound—are as important as touch and sight. Also try a synesthetic approach—imagine the color of the sound of the air conditioner, or a taste for the softness of a pillow. Have fun.

Crucial to this exercise is that sixth sense—that is emotion. Consider how you feel about the details you encounter. Likely the feelings are associated with memories, as de Maistre remembered the dreams of love he had on the couch. For the sensitive writer, observation is always entangled with emotion. Whether or not the emotion makes it to the page depends on the purpose and focus of the writing. Good editing is paramount.

As de Botton discovered, he needed more space than just his bedroom. Without seeming too crazy, you might take this exercise to the street, imagining that you are seeing your town as a visitor. You might also extend the exercise to one of empathy, and imagine how the people you pass as you stroll along might see you. Imagine how their day might be going, whether they slept well or are well fed.

The selective rendering of details and its emotional coloring enriches the setting. Believe me, a little goes a long way. I have just read a novel set in the streets near my house. It simply name drops the streets without details or coloring. In this novel Decatur Street in Atlanta, could also be Decatur Street in New Orleans, or Brooklyn. In effect, the “something quite irreplaceable” is missing because there is no engagement with a rendering of imagination through observed detail and emotional coloring.

This essay originally appeared at the Georgia Writers Association website.

March 2024

A Glitch in the Glitch in the Matrix

A YouTube video of the October 2023 annular solar eclipse shows a small square opening up on the surface of the moon, and for a millisecond, sunlight streams through as if the moon had been punctured. True believers theorize that this is proof that reality is a simulation and that we are living in a matrix. The hole in the moon, they say, is a glitch in an otherwise mostly seamless deception generated by nefarious technology. For most of us, this purported deception isn’t too awful—we have clean water, good food and warm homes—and if the believers are right, on the other side of the illusion is a planet destroyed by nuclear war, cast into perpetual darkness and ruled by tentacle-dangling robots. I’ll take the blue pill, Morpheus, thank you very much.

On this side of the Matrix, we already have enough trouble with machines—mostly of the damn-thing-isn’t-working variety—than to have to worry about a war-of-the-worlds with dictatorial robots. Among our troubles are rising concerns about artificial intelligence or AI, computer systems which are designed to carry out functions that typically require the perception and decision making of a person. The worry, it seems, is a kind of replacement conspiracy theory, in which people fear that we will be replaced by machines. Machine replacement of people has been going on since the advent of the cotton gin, and like the cotton gin, sometimes has profound implications for human societies. In these cases, it is often not so much the machines we worry about, but the people who control them.

As I understand it, AI works by recognizing patterns in very large data sets, which then allows it to make predictions about any new pattern it encounters. For example, a computer program, or neural network, which has studied millions of faces may then be used to identify your face as you rush to get on your plane at the airport. But AI is only as smart as the information it is trained on, and already we have seen that its learning reflects limitations of the data given to it and the biases of its programmers. I do not belittle concerns about the way AI is used to make decisions, for example, in our criminal justice system or in our hospitals. I applaud recent efforts by the Biden administration of set up guidelines for the use of AI; however, it is not artificial intelligence, per se, that is my worry, but human ignorance.

Educators and writers, too, have had special worries about how AI is used. Teachers fear that students will turn to AI programs to write their term papers. This shouldn’t be a surprise, since students have been buying term papers in one way or another since the days of Aristotle. Teachers can devise ways to circumvent this cheating. Writing teachers have long implemented teaching methods like multiple drafting and in-class writing that circumvent the cheating. It is shameful that school administrators and legislators have not fully supported such methods, as they push to increase class sizes, pay poverty-level wages to adjunct faculty, and in other ways devalue humanities professionals.

Creative writers, too, have expressed worries about the use of AI. This year, at least three different groups of authors, including the likes of John Grisham, George R.R. Martin and Steven King, have sued OpenAI, the research company, over its use of their books for the training of ChatGPT and other generative AI systems. The authors assert that OpenAI used their materials as a source of AI training data without regard to copyright protections. The case is not as straightforward as it may seem, since the question impinges on fair use—How is AI training different from a writing class in which students study an author’s style? Is a story generated in the style of an author plagiarism, or fair use akin to fan fiction? Perhaps the answer to these questions lie in the extent to which AI companies and their programs are considered to have the same rights as people. Since the 19th Century, the U. S. Supreme Court has well-established a trend toward supporting cooperate personhood. Most recently in the Citizens United and Burwell cases, it established that corporations have free speech and religious rights. It could be that AI will be seen to have the same fair use rights as you or I.

“It is imperative that we stop this theft in its tracks or we will destroy our incredible literary culture, which feeds many other creative industries in the U.S.,” Authors Guild Chief Executive Mary Rasenberger, who joined the authors in one of the lawsuits, said in a Los Angeles Times article. “Great books are generally written by those who spend their careers and, indeed, their lives, learning and perfecting their crafts. To preserve our literature, authors must have the ability to control if and how their works are used by generative AI.”

On the surface, Rasenberger’s statement should be a rallying cry for writers, and it rankles me to think that AI could destroy literary culture. But another view argues that culture is far too complex to be “controlled” by authors and it is mimicry and alteration that revitalizes culture and generates new ideas. For me, the current copyright law finds a balance—giving author’s limited control over our works, but eventually releasing them to public domain.

The goal of the lawsuits, however, might not be so lofty as to save literary culture from our robot overlords, as Esquire magazine writer Josh Rosenburg pointed out about the Authors Guild case: “As you read about the case, however, it feels like less of a plan to slay AI for good—and more of a fight to make sure writers just get their money. ChatGPT is crossing a lot of bridges on its path without paying any tolls, and writers simply want their due.”

And yes! Since apparently there is money to be made, why shouldn’t we authors have our due?

AI, though, is only a tool, and the real danger of theft comes not from programmers but from corporate publishers who may use AI as a cheap way to produce blockbusters without authors. I do not believe AI is quite this good yet, but take note that Amazon has already changed its policy to restrict the proliferation of AI generated books on its Kindle Direct platform.

No matter what the Supreme Court rules, a corporation is not a person, and no matter how engaging an AI book is, it is not written by a person—and so ultimately does not embody the expression and originality of a person. No matter how predictive AI is, it is simply recognizing patterns—it is predicting how readers will engage with those patterns—it has none of the ingenuity of a human soul and therefore can only provide a mimicry of humanness to the reader. It might look like the real thing, but it will never satisfy the sensitive reader.

Still, it might be good enough for most readers, and that would be a worry for many writers, especially writers of popular stories which depend heavily on formula. In an interview in The New York Times Magazine, literary agent Andrew Wylie, known as “the Jackal” for his business savvy, declared that the literary writers he represented—a who-is-who among prominent writers like Martin Amis, Raymond Carver and Salmon Rushie—were in no danger of being replicated by AI. “But take the best-seller list. That’s a little susceptible to artificial intelligence because the books on it are written without any particular gift in the nature of their expression. Stephen King is susceptible to artificial intelligence. Danielle Steel is even more susceptible to artificial intelligence. The worse the writing, the more susceptible it is to artificial intelligence.”

The assessment of King and Steel is Wylie’s, not mine. It is fair to say that the more a style or genre adheres to a predictable literary pattern, the more AI will find examples of those patterns, learn and mimic them. The more original an author’s style or subject matter, the fewer the examples of the pattern and the less it is likely to be learned by the machine.

Let this be a warning to writers regardless of genre: It is incumbent on us to be as boldly original as we can imagine. Stretch the boundaries of genre patterns, create new expressions of style, see the world as no one else sees it. A tall order, and I can only offer this advice about how to achieve it. It comes from Dorthea Brande’s 1934 tome, Becoming a Writer, (full text is easily found online) in which she asserts that the only way to be original is to be honest.

Here is another supposition—a hard pill for writers—what if AI is capable of producing engaging literary works—works that readers enjoy? Would that be so bad? Writers will still write—though likely not publish through commercial houses—and readers would enjoy a plethora of stories—perhaps stories produced on their laptops in moments: You are at the beach, and you forgot to bring a book. ”Alexa,” you say, “write me a novel about a lovesick man on a beach,” and moments later, a 400-page beach read appears on your laptop. No author. No publisher. No distributor. No bookstore.

No way! At least not yet—I don’t discount the machine miracles of the future. However, my experiments with generative AI programs suggest that they are not yet capable of producing any creative writing of any level of sophistication. Keep in mind that I’ve only played around with a few free online programs, and I grant there are likely more advanced AIs that might deliver a more satisfying result. I amused myself with my AI concubine, Alexa, asking her to produce a poem. Her poem was a confidently delivered piece of doggerel. When I asked for a sonnet, she replied, “Sorry, I don’t know that one.” Her story about a man who collected shells was better, but never advanced in detail or character development beyond the anecdotal.

I instructed online programs to compose about the themes of some of my favorite poems. (I didn’t ask for stories for this part of the experiment.) For example, I instructed it to write about perseverance using the symbol of a house in a storm. I got several stanzas—too many in fact—AI could use a good editor—that addressed the subject, but in a predicable way in its use of imagery, sound and rhythm.

As thunder roars and lightning streaks the sky,
The house stands tall, determined not to shy,
A testament to perseverance and grace,
Withstanding the storm's relentless chase.

I had in mind Ted Hughes’s “Wind”:

This house has been far out at sea all night,
The woods crashing through darkness, the booming hills,
Winds stampeding the fields under the window
Floundering black astride and blinding wet

The Hughes poem is well observed, and the observations are rendered through human sensibilities—beginning with the denotation of the house—not any house, but “this” house having been out at sea all night—a connotation of isolation. Note also the onomatopoeia, particularly “booming,” which conflates the sound with an image of the hills. At its root, the language is illogical as hills cannot boom, but human sensibility describes the sound of the wind in the dunes, and makes perfect sense to the reader.

Other experiments with AI brought similar results. I prompted it to write a sonnet about the variety and beauty of God's nature. It provided a Shakespearean sonnet that ended surprisingly well considering the vague stanzas that preceded the final couplet:

Oh, God, in all Your wondrous works we see,
The varied beauty of Your vast decree.

But it does not compare in innovation and feeling to Gerard Manley Hopkins’s ending for “Pied Beauty”:

He fathers-forth whose beauty is past change:
Praise Him.

Note that Hopkins famously breaks the sonnet form, inventing what he called the “curtal” or “curtailed” sonnet. Hopkins’s break in pattern deepens the emotion of the poem, transitioning from high lyricism to a simple declaration.

Even though AI spoke of the variety of nature’s colors and creatures it never named a color or creature, creating a series of vague images:

A symphony of colors, sights, and sounds,
In every leaf, in every creature made.

These lines are drab compared to Hopkins’s:

For rose-moles all in stipple upon trout that swim;
Fresh-firecoal chestnut-falls; finches' wings;

AI is still very much at the greeting card level of poetry writing. It produces faithfully about the subject but lacks the innovations that come with human experiences and emotions.

Perhaps this is humorous, perhaps not. When I asked the infamous ChatGPT to write a poem about God’s grandeur, it first wrote that my topic was interesting. Then it spilled out “God’s Grandeur” by Hopkins but before it got to the last verse, as if recognizing the plagiarism, the poem vanished and this message came up, “My mistake, I can’t give a response to that right now. Let’s try a different topic.”

I also asked one of the AIs to write a generic poem—a wedding poem—and it produced many sentimental rhymed quatrains about a couple’s future happiness. It occurred to me that it was just the kind of poem my relatives might enjoy—lots of generalities and rhymes, but no specificity or challenge. The next time one of my relatives gets married and asks me for a poem, I won’t sweat it.

This essay originally appeared at the Georgia Writers Association website.