The Whole and Every Whit

 This week brought us news of another AI-related controversy in the literary world. Apparently a number of readers are speculating that this year's winner of a major short story prize was actually created by (or with significant help from) a chatbot. 

This seems to be part of the misery we all live with these days—and it wounds in both directions. 

For those of us who write our own stuff and aren't winning literary prizes for it—we are jealous at the thought that other people are getting ahead and winning laurels by cheating. 

But at the same time, it must be no less miserable to be in the opposite position: to win something through your own effort, and then have legions of envious people on the internet cast doubt on your achievement by speculating you didn't really write it yourself. 

Suppose this author actually did create their own story—how infuriating it would be to have one's authorship questioned!

This seems at first like a problem unique to the age of AI—like a whole new injustice to layer on top of the already (and eternally) precarious lot of the inhabitants of Grub Street. 

As if the writer's life weren't hard enough already—now we have generative AI both elevating the unworthy to new heights and leading to unjust accusations of cheating. 

Oh the "spurns/ That patient merit of the unworthy takes"!

I took some comfort this week, however, from reading Robert Merton's whimsical classic of tongue-in-cheek pedantry: On the Shoulders of Giants: A Shandean Postscript. 

The book reveals that little about our present-day authorial plight is actually new. 

He notes there that—long before generative AI was so much as dreamt of—writers already had to defend themselves from the charge of merely pillaging from other human authors. 

Merton observes that writers even in earlier centuries "were the targets and often the entirely innocent targets for calumnious charges of plagiarism." 

John Bunyan—of Pilgrim's Progress fame—felt the sting of this so keenly that he penned some defensive lines warding off the charges, which Merton quotes.  

The book, Bunyan insists, came from my own heart, so to my head

And thence into my fingers tickled; 

Then to my pen, from whence immediately 

On paper I did dripple it daintily. 

Manner and matter too was all mine own; 

Nor was it unto any mortal known,

Till I had done it. Nor did any then, 

By books, by wit, by tongue, or hand, or pen,

Add five words to it, or wrote half a line

Thereof; the whole and every whit is mine. 

I'm tempted to staple this over my door or to the masthead of this blog, now that we live in an age when all of us who publish live under the threat at any moment of having our own achievements misattributed to an LLM. 

"The whole and every whit" of this blog "is mine," I say! Freshly fermented in my own heart and brain, and therefrom directly "drippled" by my own fingers. 

Except when I'm quoting from other authors (like Merton and Bunyan) with acknowledgement, of course. 

Which brings me to another sense in which the misery of our writerly lot under the sign of AI is not actually so new as it appears. 

One of the constant sources of complaint about generative AI, after all, is that it cannot actually do anything original. All it can do is statistically model whatever human writers have done before. 

As one humanities student recently told the New York Times—for an article about why so many college grads this year are booing their commencement speakers at the mention of AI: the technology is "by definition unable to create a real-life story that hasn’t been told."

But this problem of originality was likewise one writers had to entertain centuries before ChatGPT. 

Even in the eighteenth century, they already had to contend with the question of whether what appeared at first to be originality was really just a rag stitched together from quotes and half-remembered cryptomnesiac borrowings from their earlier reading. 

Merton cites the classic lines from Laurence Sterne: "Tell me, ye learned, shall we for ever be adding so much to the bulk—so little to the stock? Shall we for ever make new books, as apothecaries make new mixtures, by pouring only out of one vessel into another? Are we for ever to be twisting, and untwisting the same rope? for ever in the same track—for ever at the same pace?"

And as Merton goes on to reveal—the ultimate Shandean irony of this passage is that it is itself made up of gleanings and odd-and-ends filched from an earlier author: in this case, Burton.

Generative AI may create prose through stealing from all the human authors who came before it. But even in the days when it did not exist, authors still had to contend with the related problem: suppose that's all human creators are doing too. 

Creativity is—as John Livingston Lowes shows us, in his classic book on Coleridge—a fundamentally agglutinative process. 

Very little proceeds ex vacuo in the creative life. Our brains—even in producing "original" work—are often just rearranging elements of what we have found elsewhere into new combinations—"pouring one vessel into another," as Tristram Shandy puts it. 

Long before the days of Claude and LLMs, then, we already had to contend with the fear that all our apparent artistry and originality was really mere "mechanical joinery," as Lowes puts it. Just "twisting and untwisting the same rope," as other authors had done before. 

Oh, say it ain't so, Yorick! "Not untwist — slack they may be — these last strands of man / In me," to filch a line from G.M. Hopkins. 

Merton's entire book suggests the only possible answer to all this: Go with it. Enjoy the comedy of pedantry, whereby even our dearest thoughts will turn out somehow to have been anticipated by someone else, for there is nothing new under the sun. 

As Burton (as quoted by Merton) states the case for all time: "We can say nothing but what hath been said, the composition and method is ours only."