Poets have accumulated many labels throughout history. Some are not so flattering—Plato saw them as liars and demanded their exile from his precious utopia. Others glimpse, in the role of the poet, something closer to the sacred: Emerson called them “liberating gods,” while Stevens cast them as “priests of the invisible.” Some view poets as playing important, functional roles in society—Shelley famously championed them as “the unacknowledged legislators of the world,” while Lorca branded them “revolutionaries,” and Brodsky enshrined them as “custodians of language.” Yet others see the poet in a realm neither human nor exactly divine—they’re “bees” in Rilke, “antennae” according to Pound, and “dead” according to the think pieces that appear regularly in today’s newspapers and magazines.
Ryan Ruby, a novelist, critic, and translator, has a surprising characterization to add, a vision of the poet as not only thoroughly alive but extraordinarily relevant: poets do R&D. Ruby advances this argument in an audacious book-length poem, Context Collapse: A Poem Containing a History of Poetry, which is unlike anything I’ve ever read. How to classify it? I suppose the subtitle gives some clues. It is indeed a poem—it’s written in verse (pentameter). And it is indeed a history—leading us from The Odyssey to The Xenotext (a recent project by Canadian poet Christian Bök, which involved implanting a poem, encoded as DNA, into a bacterium). But it is also—and perhaps most of all—a major contribution to media theory. It positions poetry as a vital site of innovation and experimentation, where language is continually renewed through its collisions with new media technologies, ultimately transforming not only the impact of these technologies on society but the very enterprise of meaning-making itself.
Ryan and I spoke about Context Collapse over Zoom in October. Our conversation has been condensed and edited for clarity.
—Zoë Hitzig
Zoë Hitzig: I need to know about the origins of this wild project. Did you start with the thesis in mind and then go through the history of poetry to prove it? Or did you begin with a desire to understand what poetry is by going back to its roots, walking through its history, and then arriving at your argument?
Ryan Ruby: Part of the argument had been forming for quite some time—since around 2012, when I wrote an article about Occupy and the People’s Mic. In that article, I was really interested in the curious coexistence of social media platforms like Twitter, SMS, and Facebook as organizational tools alongside, at Zuccotti Park itself, a form of technological amplification through the People’s Mic. Due to restrictions on the use of amplified sound at Zuccotti during talks or speeches, everyone would repeat the speaker’s words collectively and signal assent or dissent through various hand gestures. It struck me as very interesting that the most advanced forms of communication technology were coexisting with some of the most archaic forms.
To write that article, and to think through that peculiar paradox, I became interested in media theory. I started first with McLuhan and moved onto figures like Eric Havelock and Walter Ong. As I began developing an argument—that the best way to understand our post-literate culture was to look at pre-literate Greece—I turned to German system theorists like Niklas Luhmann and media theorists like Friedrich Kittler. That was one part of the process.
At every significant change in the medium, poets seemed to be the ones doing the first experiments. Poets were doing R&D.
Another part was more personal, involving my changing relationship to poetry. As I reached my mid-thirties, I found I had lost my taste for contemporary poetry. I just wasn’t reading it––I didn’t know who the contemporary poets were. Whenever I read poetry, I returned only to what I had loved in my early twenties. Dissatisfied by this, I started casting around for poets to read . . . but I really just couldn’t make heads or tails of the landscape. So I decided to start at the beginning in hopes that the long view would help me understand the poetry being written today. I reread Homer, Dante, Pope, the modernists, the Language poets, and the New York School—working my way back up to the present.
These two projects—my exploration of communication technologies and my journey through the history of poetry—began to converge. I started looking at the poems through the lens of media theory. A pattern emerged: at every significant change in the medium—whether in storage, dissemination, or reproduction of language—poets seemed to be the ones doing the first experiments. Poets were doing R&D. Poetry was the site where new media technologies were first tested and explored.
And so I thought, this would be one way to understand Western poetry as a transhistorical phenomenon. The thread that defines poetry over thousands of years, the thing that makes poetry poetry, is its particular relationship to media—technologies of storage and dissemination. This seemed like the key to understanding what today’s poetry has in common with poetry that existed even before writing.
ZH: Speaking of poetry that predates writing—can we talk a bit about the difference between poetry and song? Is one a subset of the other? Is a pop song poetry?
RR: Let’s historicize and ask ourselves: When did this question become askable? When did it start to make sense as something we might disagree about and debate? Based on my research, it really only becomes meaningful in the mid-nineteenth century at the earliest, and it doesn’t fully come into focus until the early twentieth century. In other words, it’s only recently that it’s even occurred to us to ask this question. Why?
From Homer’s era—before the alphabetic writing—all the way up to the mid-nineteenth century, poetry was primarily received orally. Homeric verse, for instance, was accompanied by a lyre or a staff beating out rhythms, and the word lyric itself, as you of course know, comes from poetry’s relationship to music. Even though alphabetic writing existed in Europe as early as the seventh century BC, and printing began in the mid-fifteenth century, poetry retained its primarily oral nature until the nineteenth century.
In mid-nineteenth-century England or the United States, you would typically hear a poem recited aloud, and memorizing and reciting poetry was still a common practice. You’d encounter poetry in contexts similar to those of folk songs, tunes, or hymns––poetry and song were not as clearly differentiated then as they are now.
This situation begins to change in the mid-nineteenth century with figures like Whitman in the United States and Baudelaire in France. Earlier experiments had occurred, such as seventeenth-century shape poetry, where the arrangement of words on the page (as a Christmas tree, chalice, cross, or pyramid) conveyed meaning visually. Still, despite all these developments, I think the person who really marks the definitive split between poetry and song is Mallarmé.
ZH: Oh yeah––I remember the importance of his poem “Un coup de dés jamais n’abolira le hasard” from the book. I hadn’t read it before but spent some time with it after its mention in Context Collapse.
RR: What Mallarmé does—building on elements that were already incipient in shape poems, Baudelaire’s and Rimbaud’s prose poetry, and Whitman’s long lines—is scatter the poem’s words across the page so that the piece can’t be fully experienced orally or sonically. It has to be read; it has to be seen; it’s a form of visual art, an art object. Decoding its meaning involves making interpretive decisions about how the lines are structured and how they break across the page.
Now, in the twenty-first century, we take this sort of visual arrangement for granted, but in Mallarmé’s time it was extremely radical. That’s the moment when poetry truly begins to become a visual rather than a sonic art, and we become invested in the materiality of the sign. We might think of it as a genealogical break: poetry moves through the twentieth century toward the visual art, while popular song continues the older oral tradition of poetry accompanied by music.
So, in fact, if we’re thinking about the history of poetry and music, the real aberration isn’t pop music—it’s the book of poetry. In some ways, the printed, silent poem is the newer form, even though it might not seem that way, because of the kinds of instrumentation, amplification, and recording technologies that are necessary to produce and consume it. The tradition of what we now call pop music is actually older than the tradition of poetry as we encounter it now, in slim volumes in the bookstore.
ZH: How did the concept of authorship and ownership evolve along this trajectory? Was the page––that Mallarmé moment––an important turning point for establishing the idea of singular authorship, or a poem as the intellectual property of the author, or is it more complicated than that?
RR: This is a huge question. If we take the oldest possible case—Homer—we know he wasn’t actually a single individual, but rather a composite figure representing a class of roving musicians. There’s a long and fascinating story about how the figure of Homer was constructed by a group known as the Homeridae, and how this related to political and cultural factions under the dictatorship of Peisistratus in the late seventh and early sixth centuries BC. What’s important is that the notion of authorship as we understand it today simply wasn’t applicable in Homer’s time. There was no concept of a single “author,” and no way to authenticate the correctness of a line until writing on papyrus provided a means of recording and stabilizing what had been purely oral and attributing it to a single person.
The notion of the author has gone in and out of fashion throughout history. Foucault, for example, points out that in the medieval period, literary authorship mattered less. Scientific or philosophical authorship, on the other hand, was crucial: people would attribute their own theories to someone like Aristotle to give them more authority. Over the modern period, this situation reversed. We now consider it very important to identify the literary author, while scientific authorship is less tied to an individual’s authority and more to a collective pursuit of objective truth.
It’s intriguing to ask why we allow certain people to produce certain combinations of words and have those words regarded as “theirs.”
The history of copyright and intellectual property is no less fascinating. In England, the landmark moment is the Statute of Anne in 1710, which granted authors the right to control the printing of their works—essentially, the first modern copyright law. Before that, authors typically relied exclusively on patronage. You would sell your work to a patron who then owned it and could print it if they wished. As this patronage system broke down in the eighteenth century, a new economic model emerged. Now authors (or their publishers) could reprint and sell works directly to a mass public. Suddenly, it made economic sense to create works that appealed to a broad, anonymous readership rather than just a small circle personally known to the writer. This shift fostered the idea that the author “owned” the text, and our understanding of the meaning of the text changed as a result.
Of course, the law lagged behind technology. Piracy, forgery, and authorial hoaxes were rampant. Books were often reproduced without permission. Over time, legal frameworks developed to curb these abuses, helping to solidify the modern concept of the author as a unique originator of the text and the owner of intellectual property.
So, yes, there’s a whole process by which the modern concept of the author, which we now take for granted, is constructed. Along with that, there’s the economic side—it’s a chicken-and-egg question, really—but the ideology of the author as a unique, original source of meaning is actually very recent. It probably doesn’t solidify until the late seventeenth or early eighteenth century at the earliest. Then, of course, this idea gets projected backward onto earlier figures. Take Shakespeare, for example. The notion of Shakespeare as an “author” in our modern sense is something we retroactively impose––it didn’t exist in quite the same way in his own time.
ZH: This chicken-and-egg problem, as you call it, really makes me wonder how things could be different! If the economic system hadn’t become so fixated on property, we’d probably have a different notion of authorship. But also, if we had insisted on a more collective understanding of authorship—one that acknowledges that no work is truly made in isolation—maybe we would have been steered away from our obsession with property in the first place.
RR: Let’s jump forward to the present. There’s a good version of that and a bad version. Think about the class-action lawsuit by authors against the AI companies building large language models. The authors are suing OpenAI, Microsoft, and Google and probably others for using tons of existing works—many of which are owned by individual authors or their publishers—to train AI.
If you want to play devil’s advocate (literally!), one positive way to understand what OpenAI is doing is that it’s realizing that possibility of the communal, propertyless, authorless world of collective production dreamed of by the most starry-eyed Marxist aesthetician. When AI generates something from a prompt, it is not an individual author but the collective voice that emerges from the history of literature, scraped and stored by the machine. In principle, AI has done something that left-wing authors have been interested in doing for centuries: collective artistic production. But these companies have done so in a way that looks more like wide-scale property theft. It’s not merely plagiarism, per se, it’s a kind of primitive capitalist accumulation of intellectual property because the value created by producers is not redistributed to them; in fact, it’s being used to put them out of business. For all these reasons, the present moment is a watershed in the history of IP and copyright.
ZH: Totally––these new language models definitely require new thinking about intellectual property, and maybe new ideas about authorship too.
RR: Another aspect of all this that I find philosophically interesting is that this very issue arises with language itself. Language is inherently collective, right? None of us invent the tools we use to create our work: the alphabet, the lexicon, the rules of grammar. For our work to be effective, these things have to be shared; otherwise, it wouldn’t be understood. In the book, I describe authorship as enclosing your field of common discourse. It’s intriguing to ask why we allow certain people to produce certain combinations of words and have those words regarded as “theirs.”
On one hand, we do associate certain strings of words with particular individuals, who we think of as their origin points. Two issues emerge: property and origin. Many people are motivated by the idea that they’ve discovered a particular “secret code” within the common discourse. When put into the world at the right time, it creates certain effects, and they want their effort and identity associated with that code.
But in truth, what’s happening with large language models, for example, is already implicit in the alphabet. Once you can code speech in a finite set of symbols, everything becomes combinatorial. Everything is, in principle, already there. Writing just selects a particular combination of letters. Borges talks about this in “The Library of Babel.” The Catalan troubadour and alchemist Raymond Llull was talking about this in the early fourteenth century.
In a strange way, what we see with LLMs, or with something like the chess-playing machine Deep Blue, is the realization of an ancient fantasy: actualizing all possible combinations. We’re living through a time when these dreams are becoming real, and we’re finding the experience more unsettling than we’d perhaps anticipated. We don’t exactly like how it looks when these possibilities are laid out before us, but that has more to do, I’d argue, with the economic relations that go into making an LLM than with the technology itself.
ZH: Yeah, that’s a really nice way of putting it. It does make me wonder: What would be a way of encountering these technologies that would feel better for those of us who care deeply about language?
RR: Well, first let’s just set the ecological issues aside for a moment––obviously the astonishing energy intensiveness of AI and the ecological impact of the data centers that power it is a whole other issue that makes it impossible for any sane person to endorse their existence with a good conscience.
One of the things you seem to be suggesting here is that the author may be too deeply invested in the notion of property with respect to their work, and this is alienating us from the true nature of intellectual labor. Yet the same logic ought to apply to the tech companies on a larger scale. I think the conversation around LLMs would be very different if, instead of a private corporation running these technologies, these technologies were treated as a public utility. That was part of the utopian vision people had in the early days of the internet. If an LLM was collectively owned and maintained, with equal access for anyone to contribute their work and benefit from the collective whole, then people might feel a little more comfortable with it in the long term. Collective ownership, I think, is the bare minimum requirement for us to think of this technology as a positive development.
We can get lost in the technological or philosophical details, but the reality is that what we’re now describing—treating technologies like LLMs, or social media for that matter, as public utilities, collectively owned—would require a complete overhaul of the society in which they’re embedded. I don’t know about you, but I’m pretty pessimistic that we’ll reach that point anytime soon.
ZH: Did you see Holly Herndon and Mat Dryhurst’s project that launched earlier this month? They released a public dataset of around twelve million public domain images, alongside a platform for collectively governing the dataset and its uses in training new machine learning image models. Seeing that kind of artist-led approach to developing a training corpus gave me a bit of hope!
RR: Artists—poets in particular—have been interested in these issues for quite some time. In the last section of Context Collapse, I give a sort of prehistory to this moment. Poets have been working on these questions since at least the early 1960s, experimenting with text-generating processes that anticipate what we’re seeing now.
I think it’s easy to imagine these machines producing literature that we’d consider good by our aesthetic standards. That’s not far-fetched. What I’m less sure about—and maybe this is because I don’t fully understand it—is whether these machines can produce work in a discrete historical context that makes literature that is actually meaningful.
We could theoretically live aesthetically in the way only dreamed about by the Dadaists in Zurich. It’s pretty wild. It’s pretty dystopian.
When you look at a text, you’re not just looking at words on a page. There’s the book form, the codex, which has a whole tradition behind it: the way it’s published, the historical moment it’s published in, how it intersects with other media forms, the fact that it’s produced by a known or unknown individual, etc. All these elements form a broader constellation that interacts with the world as it exists, all of this is acting in parallel to the work’s appearance. Over time, as the world changes, the meaning and reception of that text also change. Our problem when talking about AI art or poetry is that we are primarily evaluating them as art or literary objects, as we’ve been taught to do by a century of erroneous theory, when these things, as I argue in the book, are really social relations. That, I think, is where most of the confusion arises.
As for the economic aspect, do you know Mark McGurl, the literary sociologist? I love his most recent book, Everything and Less: The Novel in the Age of Amazon. He examines the “genre turn” in literature by looking at huge amounts of self-published genre fiction—often very demotic, “subliterary” forms, with strange genre mutations. He shows how creativity in genre fiction isn’t at the sentence level but at the level of genre variation. A romance novel turned into a vampire romance set in corporate America where everyone wears diapers during sex. That’s the kind of creativity at play, not a novel way of writing a sentence.
But what McGurl points out––and it’s a fascinating and troubling point––is that, because of the sheer volume of production of this subliterary prose, the result is a realization of a great modernist dream of the purely uncommodifiable word, of language freed from capital. This astronomical supply of language means demand can’t keep up, so the “price” of language drops to nothing. The vast majority of genre writers—and increasingly other kinds of writers too, especially poets—do not get paid for their work. The modernist dream of decommodifying language has come true, oddly enough, in the form of hundreds of thousands of people reading, say, “adult diaper baby romance,” formally the least modernist kind of writing you could possibly imagine.
ZH: Wow, amazing. I need to read that.
RR: I think it is a phenomenal book, a must read to understand what’s going on now. It is just another one of these weird moments in which we have the thing that we thought we wanted––the thing we thought might free us from the economy. A purely linguistic economy has in fact been achieved collectively by all of us, and the great historical irony is that it produces something absolutely awful aesthetically. It is sub‑aesthetic, and that is a pointed irony about the present.
ZH: It does make me a little confused about what we are doing as writers in this bizarre present!
RR: Well, for better or worse, it’s no longer possible to say anymore who “we” is. But let’s talk about you and me. I was in Paris recently, having a debate about whether or not avant‑garde was a meaningful term to retain. I was actually taking the position of the old fuddy‑duddy who was saying it’s important that we retain this term. The avant-garde ideals of novelty and originality—of “make it new”—are devalued concepts for a whole host of reasons, but it has become so common to devalue them that we forget what they are for. If the thing already exists, in my view, there is no need to do it again. There is actually no need to iterate writing. We can read the thing that already exists, or we can permute it in such a way that it does not matter whether a human being is doing the permuting.
One of the symbolic resonances of the idea of novelty is that there is a future world in which the thing produced now has yet to be appreciated. With this idea, we project ourselves into a future in which we can imagine the new thing that exists now finally coming to be taken up and appreciated. That is not just our personal vanity about immortality and surviving our own deaths. It is also a token of hope about the continuity of our culture and the species. To lose that––to lose that fantasy structure about the future—would, in turn, do our understanding of the present—what we’re doing now—quite a bit of violence. And in fact by foreclosing the possibility of novelty and by relegating the avant-garde to the past we are in fact doing violence to the present.
ZH: Thank you for giving me this defense of the avant-garde. I don’t have my own, but I like this, so I’m going to use it.
RR: Yeah, in our own ways, I think of both of us as defenders of the “tradition” of the avant-garde. Your work even more obviously fits in that category––both of your books make gestures toward the future, formally, and much more noticeably in Not Us Now. My book is, in a strange way, backward‑looking formally. You’re carrying out the program of making it new, and the program I’m carrying out is recovering something that’s been lost or forgotten, though these programs are working in tandem. Both of us are working on similar problems and issues and using the same medium to do it, but our strategies are quite different. That is why I think we shouldn’t get hung up on individual books. The proper unit of analysis is the sum total of literary production.
ZH: When I think about the totality of literary production, I’m not even sure what to count. We’re always putting language into the world in so many ways. Lately, I’ve become more conscious of my digital communication and feel a strange pressure, a good kind of pressure, I think, to resist what each medium wants me to do—rush in a text, singsong a phone call, follow email’s corporate tone. I’m unsettled by the idea that all this language is recorded somewhere, that it may last for a hundred years, and I don’t know where it will end up. If my emails and texts might someday be fed into a model that represents all of human knowledge and the English language, I feel compelled to use language more deliberately. It’s like I have to be a poet all the time—fighting clichés, refusing predictability.
RR: I think that’s noble, personally. In fact it relates to another great modernist ideal––the breakdown between art and life. Not just to write poetry, but to be poetry, to have all of your communications carry the same level of linguistic charge as when one deliberately sits down to create an object to be consumed.
Now here’s another perverse irony. Today, we have this great collective Boswell in the surveillance state. In a weird way we are all performing, much as the Greeks imagined they were performing, for the gods, except our gods are the state and corporations, which are now almost totally integrated. There is a sense in which the technological possibility of total surveillance makes it possible for us to conceive of ourselves as living an aesthetic life, since everything we do is recorded and stored. Every second of our lives has an audience. We could theoretically live aesthetically in the way only dreamed about by the Dadaists at the Cabaret Voltaire. It’s pretty wild. It’s pretty dystopian. Because—I don’t know about you—the state and corporations are not exactly the audiences I’d like me or the data I produce to have.
On top of that, it is temporary, because of the hard ecological limit imposed on this kind of media, which I mentioned earlier. Imagine a future where all this data is recorded, but a century from now, no one’s around to engage with it, or the incentives to engage with it have disappeared, or no one knows how to use or maintain the machines. Picture a desert with old servers rusting into the sand beneath the sun like the statue of Ozymandias. Eventually, maybe, the technological cycle will be reset. If the species survives the century, there will be some kind of renewal. Maybe a civilizational collapse like the one that occurred in the Bronze Age. And we’ll just go right back to talking to each other face-to-face, and recording our collective history in epic poems, just like we did in Homer’s time.