Indulge me, for a moment, in a thought experiment based on a famous thought experiment:
You may, through your various literary or philosophical exploits (or perhaps, a dalliance with the work of Douglas Adams), have come across something known as the Infinite Monkey Theorem. There are all sorts of mathematical proofs and statistics and theories of probability accompanying the Theorem, but the core of the idea is really quite simple: an infinite number of monkeys randomly hitting keys on typewriters over an infinite amount of time will eventually produce the complete works of William Shakespeare.
This is all to illustrate that, given enough time, there is a nonzero chance that any outcome that could occur will occur, including a monkey randomly typing the entire text of Hamlet. But let’s leave the math behind for a moment and move on to another question: Does the monkey-produced version of Hamlet have, in any meaningful sense, value beyond that of the original (identical) text written by the Bard?
Now the question might seem silly but let us not forget how fond Americans, at least, are of animals and their antics. Just this week, in fact, the Twittersphere went wild for a photograph of an enormous Australian cow named Knickers. Artistic works produced by animals are no different––in 2005 someone paid $25,000 for three paintings by Congo, a chimpanzee. And while in that case the paintings weren’t identical to any other existing works of art, it seemed (with all due respect to Congo) to be the novelty of the production of the work itself rather than the actual final artistic product that gave the paintings their value.
It does not seem too courageous a leap, then, to say that even if the text was exactly identical to Shakespeare’s original work, Hamlet written by a monkey would be of some interest and could be seen to have some extra-literary societal value.
But let’s for a moment consider if the monkey did not type up the text of Hamlet, which––having been written a quite few centuries ago––is very much in the public domain. Let’s say instead that it was another famous work, say Harry Potter and the Philosopher’s Stone by celebrated British author J.K. Rowling, who has famously been involved in numerous legal disputes over alleged copyright infringements of her book and its sequels. Would there be any way that the monkey’s identical text of Harry Potter could escape a claim of copyright infringement if sold at auction or made commercially available? (We know, of course, from the infamous Ninth Circuit monkey-selfie case Naruto v. Slater that the monkey itself could not file for copyright protection for its work, but a litigious Rowling could perhaps attempt to bring an infringement claim against, if not the monkey itself, whomever aided the monkey in seeking profit for its work.) Naruto v. Slater, 888 F.3d 418 (9th Cir. 2018).
The question would seem to hinge on whether the context of the work’s production and the unusual simian scribe make it sufficiently different, even though the text is exactly the same.
Argentinian writer Jorge Luis Borges’s 1939 short story entitled “Pierre Menard, Author of the Quixote” uses an amusing meta-fictive device to explore this concept. Borges’s story takes the form of a literary review by an unnamed author who sings the praises of fictional novelist Pierre Menard, whose greatest work, the reviewer asserts, is a word-for-word reproduction of the ninth and thirty-eighth chapters of Don Quixote (and a fragment of the twenty-second chapter as well, for good measure). Menard, the reviewer states, “did not propose to copy” the Quixote, rather, “his admirable intention was to organically produce a few pages which would coincide––word for word and line for line––with those of Miguel de Cervantes.” Echoing the premise of the Infinite Monkey Theorem, the reviewer quotes Menard as saying that his endeavor is not difficult, as he “should only have to be immortal to carry it out.” (That is to say: spend an infinite amount of time at the typewriter.) The humor of Borges’s story derives from the reviewer’s assertion that Menard’s reproduction of the Quixote is “perhaps the most significant work of our time,” far more complex than the original (identical) Quixote because while Cervantes’s words are merely archaic, the “anachronism” in Menard’s work, while “verbally identical,” makes it “infinitely richer.”
Though the absurdity of the reviewer’s proclamations is plainly funny, and while I’m not suggesting that a monkey writing Harry Potter enhances the actual story in any way, in both cases the simple fact of the matter is that the revelation of the author does cast a different light on the work. (We only need look to Mary Anne Evans, better known as George Eliot, for another example.)
So can an exact reproduction have elements of originality? The legacy of American conceptual artist Sherrie Levine probably supports the view that it can. Her most famous work was a series of photographs from the 1970s entitled After Walker Evans, which were photographs of historic photographs taken by photojournalist Walker Evans during the Great Depression. Though the series was made up of almost direct reproductions, it became emblematic of the postmodern movement and one of the defining works of Appropriation Art, calling into focus issues of authorship, gender, and reproduction. The Estate of Walker Evans, however, did not seem to appreciate the ingenuity of Levine’s work, holding it to be copyright infringement and eventually prohibiting the sale of the images. However in 1994, the same year that Levine ceded her works to the Evans estate, there was a significant development in US copyright law: the Supreme Court ruled on the Campbell v. Acuff-Rose Music case, establishing the doctrine of transformative use as a part of the traditional fair use defense against copyright infringement claims.
A fair use defense traditionally relied on four factors: the purpose and character of the use, the nature of the copyrighted work, the amount and substantiality of the use, and the effect on the potential market of the copyrighted work. In Campbell, the court elaborated on the first factor, holding that the more “transformative” a work––“altering the original with new expression, meaning, or message”––the less the other factors would weigh against fair use. Campbell v. Acuff-Rose Music, Inc., 114 S. Ct. 1164, 1166 (1994).
The Court did note that the fact of “whether a substantial portion of the infringing work was copied verbatim,” was relevant especially when the work could act as a replacement for the original in the market, harming the original work’s potential commercial value (the fourth fair use factor). 114 S. Ct. at 1175–76. However, as in Campbell, where the allegedly infringing work was a parody of a copyrighted work, it is likely that a monkey-made text of Harry Potter would serve a “different market function” than Rowling’s work. Id. at 1167. Those hoping to immerse themselves in the wizarding world would probably opt to buy the original, while those delighting in primates and probability theorems would likely be buying the typewritten text based on the novelty of the author, rather than to indulge in Harry’s adventures at Hogwarts.
Seeing as finding an infinite number of monkey typists willing to tap away for an infinite period of time is just about as probable as the endeavor resulting in Hamlet or Harry Potter, why should it matter whether the fruits of the monkeys’ labor could make them liable for copyright infringement? Well, let’s go back to the original experiment:
Let’s say, instead of a monkey, we have a machine (perhaps something similar to Roald Dahl’s Great Automatic Grammatizator). Maybe this machine is simply running an algorithm that strings random words together at great speed, producing results not unlike a bunch of monkeys at a bunch of typewriters. Or maybe the machine is even more complex, maybe it isn’t entirely random at all––maybe the machine is thinking, and, much like Menard, spontaneously composes some famous text or other. If you’ve kept your eyes on news about artificial intelligence and machine learning, you’ll know that this is already the world we live in. While some are willing to shell out $25K for a chimpanzee’s oeuvre, others are willing to shell out just as much for historic examples of technological innovation, and the first machine-authored texts may very well be among them.
Perhaps it is time to start thinking seriously about these questions of man versus nature and man versus machine. The experiments, and their legal implications, won’t be mere thought for much longer.