The Conversation Trap

More conversation can be a weapon of mass obstruction.

The editorial board for the student newspaper of Middlebury College recently published an opinion that leads with the righteous headline “AI Should not Have a Place in the Liberal Arts.”

I agree. The piece articulates a number of points that I was cheered to hear being made by student journalists, and it adds to a mounting campus movement across the country that includes students refusing to cede the direction of their education to the technology industry and the ill-informed decisions made by college administrators to do their bidding. This piece by student journalist Naishu Vempali from my own institution UMass Amherst is a powerful example. There are many others.

I do think that a few sentences in the Middlebury piece call out for some more elaboration, not because they’re unique, but because they are a recent and emblematic line of argumentation. They write,

The Board was unable to arrive at an agreement on how AI should be addressed on campus, a testament to its many intricacies and its uniqueness. However, there was a general consensus that there should be more substantive conversations about AI and an attempt to create a more cohesive and collective ethos among Middlebury students and faculty. The 2025 Clifford Symposium explored questions about AI’s impact on the world, while touching on the role of AI in the liberal arts. Middlebury requires more of these open forums where people can express their opinions and work towards developing a solution. 

Such open forums and other venues for discussion and debate are vital, not simply to campus life but democracy more generally. But these conversations, and calls for further debate and exploration and consideration and deliberation: all of this is for nought if nothing structural changes around them while they are ongoing. In other words, if a university has an enterprise software agreement that offers AI to every student and member of staff and faculty, if a university has partnerships with tech companies that are dictating AI use or lending millions of dollars of support to educational initiatives that involve AI, if a university has mandated courses in AI as a requirement for graduation…then all these conversations do is provide the semblance of negotiation while AI gets ever more entrenched in the administrative and educational activities of a university.

I applaud the editorial board of the paper. More of this, please! My comments are intended to support them for writing it. I am adding here that what they are advocating can only have an impact, can only produce material change and not simply provide a placebo, if universities relinquish their FOMO-driven zeal to radically reshape the educational mission according to the will of an industry that is moving fast and breaking learning. In other words, break those contracts, not succumb to the overtures of OpenAI.

Then, we can have a conversation. Then, educators and students and staff and everyone else who makes up a university’s community can determine the nature and degree of adoption or rejection. Then that community can sit at a table with industry and determine the nature of the relationship.

This sounds quixotic. I know this. But what I am trying to do here is show how we are past the phase of adopting industry-supplied hype (“AI is inevitable”) and are now in the phase of industry-supplied resistance. Industry loves a conversation. So long as nothing changes or pauses while we’re having it.

The notorious Pepsi ad showing a protest with people holding signs that say “join the conversation.”

AI Is a Book Ban

When access to books is actually a book ban.

The student-run Amnesty International Club at Williams College has invited me to participate in a conversation with them this weekend about technology, AI, and information access as they relate to censorship and book bans (a slightly belated Banned Books Week event). So I thought I would use this space to work out some of the things I want to float for conversation: above all that too much information access can be a form of censorship.

 One of the most important things I consider when thinking about AI in its relationship to the humanities and the study of history is access. The term "access" has an almost totemic charge in our culture because of its definitional appeal to openness and inclusion. When we use the term "access" with respect to information, we often think of "gatekeeping" as its opposite.

Calls to universal information access are coextensive with the life of the commercial internet, and if you read the (by now, hopefully universally?, notorious) Magna Carta for the Knowledge Age you'll see throughout it demands for the vaults of knowledge to be unlocked everywhere so that a vast market of open information can be availed to anyone to improve her condition. Of course—and this is all is well-trodden territory—none of this was what materialized. When power is distributed inequitably, then the empowering potentials of information access are distributed unevenly too and those with the capacity to both profit off free information and then use it to reinscribe inequitable power relations will do just that.

 The point is that one of the normative assumptions governing American culture is that universal access to information is an unqualified good, and it's the responsibility of the producers and stewards of information to make it available. This sounds nice.

 What does any of this have to do with book bans?

 It has been almost 25 years since the launch of Google Books. On the surface...what an amazing boon. When I want to verify texts I've quoted in anything that I am publishing and when I don’t have the actual book to hand, I can just go into Google Books and quote-search those lines of text to ensure I didn't mis-copy anything. When I want to see if a book interests me, I look up an excerpt from Google Books. This is great. But the scanning that availed those books to me while I sit on my couch at home, the scanning that's been carried out by countless other institutions and corporations, the drive to get every library and every special collection, every museum, every archive to mount their holdings online, to make them "universally accessible" (and whether access is actually universal when you can't even get reliable internet in many parts of the US and the world is another issue) has brought us to our current moment in which information is alarmingly inaccessible. We have stuff, we have content, we have images, we have all the things. And yet it has become increasingly difficult to access information as such.

For one, the ease now with which information is availed to people by summary means that there is a screen between people and the actual sources of information they seek. That's the obvious part, and that's a kind of low-hanging-fruit argument about how generative AI has mounted barriers between people and information.

 But I want to stoke a conversation that takes seriously how

 (a) the welter of stuff available online is there because commercial interests drove us towards a political economy that makes disputing the value of universal information access a non-starter (and in which stewards of information and intellectual laborers including academic researchers, librarians, curators, and archivists are often vilified as "gatekeepers" of it)

 (b) we are witnessing a manifestation of the weaponization of the backlog, an insistence that the most important thing is just to digitize and process everything at lightning speed and sort shit ("shit" being the quality data and metadata, not to mention further discursive framing) out later, but also often never, so that material is just brazenly being miscatalogued by OCR and Automated Text Recognition (this is like putting a library book on the wrong shelf; you can technically still access it, but lord help you trying to find it until the next shelf audit gets done in five years)     

 (c) every information ecosystem is a nuclear waste dump not simply of summary and slop but just plain unnavigability and un-useability because bots are busy scraping a library's every bit of data, and sorry, the servers are overwhelmed now

 (d) AI has provided the permission structure for firing or not just not hiring the desperately needed caretakers of information and producers of knowledge because, well, you can just ask chat, and chat has access to all the things

 (e) the illusory nature of universal access to information obscures not only how partial that information is, but how also stripped of context it becomes through every form of digital mediation and remediation one can imagine; access in this respect entails making context inaccessible

 (f) the sheer volume of things being produced by putting all of this information in the AI blender and then lobbing it at people to process and verify—I'm thinking in particular of the Math Dump here—means that there a real people who have to devote their time to wading through this tsunami and not doing their actual work

 And so, while we are right to focus on the evils of literal book bans, of demanding that libraries purge their stacks of volumes deemed by one group of people to be objectionable, I want, at this event, to invite us to think about how addition can also be subtraction. 

The End of Trust

We need peer review for lettuce and for letters.

It’s been a bad summer.

Today I read that jalapeños have been recalled. I checked FDA dot gov and saw it's the thirteenth recall in a week. During the same period in 2022, only two products were recalled, both of them pharmaceutical products. What's the average number of food recalls a year? Is this normal? I don't know. I have never needed to know.

An article on Salon reports that RFK Jr.'s program as Secretary of Health and Human Services has, in effect, made us all amateur food safety inspectors now. This is what happens when government turns on itself, fires and defunds agencies to fecklessness and hollows out it own functions. It is the Republican playbook: rescind all of the support needed for a government to function, wait for the disaster to unfold, and then point to it as evidence of governmental failure. Lather, rinse, repeat.

I also read a few days ago that a (supposedly) peer reviewed journal is awash in AI-generated articles. Authors attest that they are submitting their own, original research to this same journal, but frankly who among us now can discern the real from the slop? Who has time for this audit? Do we really need to play out Sokal on hyperfuckingdrive? Surely this is what editors and peer reviewers are meant to do, but when you're under a mountain of shit it's hard not to just give up and live with the cyclospora. Do we just read articles authored by scholars whose names we know and who know we can trust now? Seems bad.

I've been publishing for years now (and I hasten to add, I always hasten to add, before the advent of ChatGPT because the writing has been on the wall long before we were panicking about the so-called death of the college essay and pervert glasses and all the other nonsense that is just today's hell), but I've been writing for a while on how the purpose of the tech industry is unanimous with the purpose of our authoritarian regime: namely, to force citizens to abandon trust in expertise so that we can buy a subscription to platforms that will make decisions for us. This is the thesis of Shosana Zuboff's Surveillance Capitalism, most prominently, but it is intimated in so many others, a number of which I list here.

Food recalls and an inability to know whether AI is forging the news, scholarship, and literature are not unrelated phenomena. I do not know how we can emerge from this. But this fall I'll be teaching a course called, "Extracting the Past: Democratic Futures without AI," hopefully working with young folks who want to learn and want to converse with me and with each other while I try to get the book done. That is what I'm hanging on to right now.

Tonight we’re having jalapeños in dinner. We got them from a farm we trust down the road. I hope they’re safe.

"Employers Want Applicants with AI Skills": How Hearsay Manufactures Reality

I recently had an informative and valuable conversation with someone who works in career services. The conversation was wide-ranging but among the more mutually illuminating parts of our conversation involved the differences between what we’re being told and what we see.

My interlocutor conveyed how relentless is the message coming from seemingly innumerable channels that “employers want applicants with AI skills!” But as we got further into our conversation it emerged that—at least in the fields for which this person is responsible—job ads aren’t actually requesting “AI skills.” Nevertheless, a general sense of urgency is impressed upon folks like them, and educators like me, to create AI-ready students, whatever that means.

I got curious and went on to the websites for a museum and an auction house—two places where my own students in the History of Art & Architecture go to look for jobs. Of all the ads for positions I clicked, none said a single thing about “AI.” This is a minuscule sample. Of course. But it’s an experiment I’d like to throw out to anyone reading this. Go to the website for an institution or company in your field or industry, so long as that industry is not itself related to software development, computer engineering, and cognate fields. See how many positions are requiring applicants have “AI skills” or something similar. Some do, for sure. And some probably ask about it in the interview.

My guess is it is not as ubiquitous as the nightly news, legislators, and the frantic calls from university administrators make it seem. And as the shine is coming off, at least as employers realize that AI slop is inundating their applicant pool, it seems doubly important to get out in front of a narrative that is provoking a kind of panic.

As I was quoted saying in this article, here,

Urgency is meant to flood the mind of the customer so they can’t take a pause to consider whether the thing that’s being sold really needs to be bought,” she said. “I see no need for urgency. I understand that the phrase ‘But we’re going to fall behind’ can be very convincing. But no one is really asking, ‘Fall behind what?’

The more we succumb to frenzied demands, the less capacity we have to shape the future in a way that is hospitable to life. Better we should add the friction that comes with asking questions, demanding to see the proof that justifies rebuilding our world in someone else’s image.

In short, we need to get better at seeing discrepancies between rhetoric and reality and understand how the former can end up manufacturing the latter.

You Do Not Need an AI Policy

You do not need an AI policy.

Forget the crisis in plagiarism and cheating; that’s yesterday’s news. It is becoming increasingly common for my colleagues, both within my own university and elsewhere, to wonder aloud what it is they are supposed to do when confronted with an email ghostwritten-by-prompt.

A seemingly natural inclination is to come up with a set of guidelines. That’s certainly something that I, as graduate program director of my department, have been urged to do at my institution. Guidelines offer what appears to be a contract, a system, an outsourcing of messy arbitration to a decision tree: “if —> then.”

But ultimately such guidelines serve the entrenchment of a set of products that are being leveraged to enervate interpersonal relations, negotiation, and shared governance. Such guidelines give the product that we are calling AI a futurity, a legitimacy, a standing so seemingly permanent and inevitable that they merit a place in the manuals of our places of work.

So in instituting an AI policy, no matter how well intentioned, no matter how seemingly opposed to the eviscerating and immiserating effect of these products on educational institutions and society, we are inadvertently doing the industry’s work. We are explicitly saying, this product has so guaranteed a space in our institutions that we are mapping out sanctioned uses for it.

And by implementing an AI policy, we are strengthening the punitive element of education. We are creating a code of conduct that, if we are true to our word about committing to this code, mandates enforcement. A policy requires policing.

That is the impact of instituting an AI policy.

This is all well and good, and leaving students and colleagues to their own devices sounds very utopian. But it does nothing to solve the problem of these products being used in ways that educators know are detrimental to the process and mission of education.

That mission was already compromised when our institutions decided to pay for products and partner with tech firms with the left side of their mouths, while insisting that we, the educational laborers, preserve the integrity of education with the right side of their mouths.

Devising an AI policy means more work for me to think of it, and more work for me to enforce it.

There is a solution. And the solution is to concede that we are confronting a social, economic, and political problem; not a technological one.

So when someone chooses to subtract themselves from a social encounter, whatever means that self willing subtraction takes, the most appropriate response is to meet that self willing subtraction in kind. Meet absence with absence, meet apathy with apathy. It is not my job to compensate for a person’s erasure of me by making myself more present.

What does that look like in practice? In practice that means, for me, meeting with students and colleagues who ask for meetings to speak with me, that means responding to emails that are grounded in our relationship with emails that reflect our relations. That means being a decent human when someone shows me some goddamned human decency.

And if someone wants to complain about my non-responsiveness when those conditions aren’t met, then I welcome them to involve their whole selves in the process of interpersonal relations that complaining entails.

One Cool Trick to Increase Your Workload: AI "Literacy" for Faculty

A refrain among the profiteers of automation is that the technology they’re selling will decrease everyone’s workload. But it has been shown time and again that automation intensifies labor: whether it means that employers require laborers to fill the other six hours of the day with more work now that they can supposedly do their old jobs in two or that the expectations for what one should be able to produce in eight hours increases their work to twelve, it’s definitively true that automation intensifies and immiserates work.

But rarely do employers say as much explicitly.

I promised a colleague that I wouldn’t put them in an awkward spot but I am told that this week someone at the teaching center at their university sent out an email recommending faculty apportion 90 minutes a week to informing themselves about AI. The recommendation is cited as coming from Marc Watkins, writing for higher ed administrators’ trade rag, the Chronicle of Higher Ed.

It’s a very boring piece with all of the same slogans and platitudes about the inevitability of AI becoming a vital part of modern life, so important that teachers have an obligation to teach their students how to use it responsibly (if I hear one more time the comparison to Sex Education as if natural biological impulses were the same thing as using software, I will scream). The inevitability here is a foregone conclusion (begging the question of in what ways AI will be a part of life, what kind of AI???), but let’s take Watkins on his own terms.

How can educators do this without being informed themselves, he asks?

Set aside 90 minutes a week to educate ourselves. That’s the magic bullet. It’s a simple three-step process requiring 30 minutes of reading the curated list of industry shills he handily provides. Then spend another 30 minutes using AI from his handpicked list of platforms. Finally, spend 30 minutes “reflecting.” He goes on to state:

At this point, you may be thinking: I don’t have a spare 90 minutes a week to read, explore, and reflect on generative AI. But really, that is very likely the minimum amount of time that it will take you to become informed enough about AI to offer the type of practical advice that students desperately search for about their own AI usage.

This is an unserious piece that collects all of the mantras and canards so easily debunked by Bender and Hanna in the The AI Con. And there is, indeed, so much more out there to read, and I encourage anyone who takes a professional interest to do so. Here’s one list of books that I provided for Art in America, in which I gathered scholarship under the headings, bias, extraction, augmentation, operation, the so-called, categorization, and power.

But the key here is that I teach and research this material, have been publishing on it since 2020 and am writing a book about it (snapshot of the gist of it in Artforum here). As I suggested when interviewed by Kathryn Palmer for Inside Higher Ed, none of my labor on this issue has been made compulsory, and yet locking educators into a compulsory regime of technologized education is precisely what profiteers and opportunists are attempting to do.

None of this will ease the labor of education. It will only intensify the worst parts of it.

So I’m grateful to Marc Watkins and to his mouthpieces sending out emails informed by his recommendation . I’m grateful to have someone articulate so clearly how AI will not in any way shape or form save educators time or exertion. In fact, all AI will do is require that we work more, harder, and on the parts of our job that shouldn’t even be our job.

Starting with his one cool trick of just 90 minutes a week.

On Exactitude in Science; Or, No, Full Digitization of Every Archive Is Not the Point

In 1946, Jorge Luis Borges and Adolfo Bioy Casares published a short story called “Del Rigor en la Ciencia” (“On Exactitude in Science”) The story, which you can read in the original magazine context in which it was published on page 53 here, goes as follows

In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast Map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography.

(Suarez Miranda,Viajes devarones prudentes, Libro IV,Cap. XLV, Lerida, 1658)*

This short story is gold for any educator who wants to get across to students the idea that maps are representations and that to think that a map is in any way an objective projection of the territory that it represents is pure folly. To attempt for such “obectivity” would require a duplication of the territory itself, rendering the map useless. Or, as we in art history teach our students day-in-day-out: a representation of a thing is not the thing.

Recently, historians have once again had to remind a new phalanx of tech enthusiasts that, no, not all of the world’s knowledge (which they always seem to equate with written things and not other kinds of artifacts, but that’s another battle) has been digitized.

There’s a wonderful piece from 2019 by Marc Reyes in Contingent Magazine that not only explains this fact but also addresses why physically going to libraries, archives, and rare books collections is important to the research process. I was reminded of this essay by a Bluesky thread by Erin Bartram, historian and editor of Contingent Magazine, which fleshes out some of its discussion. If you haven’t read either, click the links and do so.

These discussions might tempt the casual reader, and maybe even some historians, to think that the “solution” (there is no problem here, or if there is, the problem isn’t digitization; the problem is the underemployment of archivists and librarians and curators…it’s always a labor problem), but anyway, the “solution” that might occur to some people is to renew digitization campaigns so that AI can then be unleashed on these newly digitized materials, and we can OCR our way towards Complete Knowledge of Everything. Certainly this is what OpenAI is promising in their new partnerships with libraries and archives from Howard University to the Bodleian Library.

There is an end-game here, an end game that entails an arms race to hoover up more data for ever more model refinement, so I would argue that partnering with OpenAI is a devil’s bargain. But let’s leave that aside for the purposes of this post.

If you go back to Marc Reyes’s essay, to Erin Bartram’s Bluesky post, and to all of the many, many quote posts and responses, what you’ll find is something far more systemic impeding the utility of “just digitizing” (it’s never a matter of just when it comes to digitizing, for which read Bridget Whearty’s excellent Digital Codicology on that kind of labor). And that something is this:

the sheer recalcitrance of diverse materials to the homogenizing process required both by digitization and processing by means of AI.

None of the materials that we are talking about here are born-digital objects. Our archives comprise rag paper, birchbark, nose grease on parchment, worm holes, water stains, creases and pleats, linen curtains, sixteenth-century secretary where every letter ‘h’ took a left turn at the corner of drunk and dizzy; they are everything idiosyncratic whose essence is resistant to the need for flattening that AI products demand.

It is not simply that you cannot CTRL-F the archives for a doctoral dissertation; it’s that CTRL-F-ing would only grasp in a partial and deeply impoverished way the bare lexical content of things in the world that we can only see as such when expert librarians, archivists, curators, cataloguers, and historians have deployed their expertise in making them visible. This is necessarily an inefficient process—it is the WORK OF HISTORY. And in doing it, we adduce more information, more context, more knowledge to the artifacts. Digitization is just one part of it.

None of what I have gestured to is the work of duplication that any program of digitization and automated processing could achieve. If it did, all we would have would be a second earth of digitized objects and data centers littering the wasted landscape they reiterated. And no one would be wiser for it. We would, in fact, be smothered to death by the ecocide.

The digitization is not the archive.

Here’s a screenshot of a manuscript on which I published many years ago. It contains a damaged fifteenth-century leaf , and underneath it you see the image that Eliza Dennis Denyer (d.1824) restored to it. Next to it is the nonsense of what OCR made of the page.

*translation of “Del Rigor en la Ciencia” here

How to See the Past Seeing the Past

When I ask the question, "why can't computer vision see the past?" the narrow answer is because computer vision is linear algebra, and as an art historian I reject the premise that math is sight.

But the big answer is that the data extraction and brokerage industry is leveraging computer vision and a host of algorithmic models to erase the incalculable parts of history to engineer a profoundly anti-human future.

These lines above are the concluding sentences to a talk I’ve been giving all around my region this fall. (Here’s a great write-up by student journalist Rohina Mahadik.) They encapsulate the conviction I have that understanding what the AI industry is doing to education isn’t, at core, about education; it is an assault on the structure of civic society, and educational institutions are just one of the pillars that are being knocked down to dismantle the whole edifice.

***.

Yesterday I visited Sing a New Song: The Psalms in Medieval Art and Life, an exhibition of Psalters—one of the most consistently produced types of manuscript during the Middle Ages, which contains the text of all 150 Psalms. This was arguably the most successful exhibition of medieval art I’ve ever attended: the thematic organization of the galleries, the discursive wall texts and object labels, the sub-themes in each vitrine, the blockbuster objects shown…perfection. But this isn’t a review. This is an opportunity to talk about the value of human memory and of embodied perception.

One of the things that I value about exhibitions is that they offer me a chance to exercise my eyes. As a professional art historian I believe that it’s fundamental to our discipline to be able to have a well-trained sensibility for form not so that we can make judgments about quality but so that we have a robust mental stock of images and objects that we use to assist us in contextualizing any object we might be researching. When I visit an exhibition I first look at the work and register my initial impression: when and where do I think it was made? What can I identify in the image, what perplexes me, etc? I then turn to the label, check to see if I was right, in the ballpark, or wrong, and then I revisit the work to see what it was that led me astray and what are the features that are more “right” to the period and place when and where the work was produced. Then I store the work in the correct spot in my memory for the next time I see something similar. It’s easy to see how someone who doesn’t practice this regularly might compare it all to the way that computer vision works. It’s tempting to refer to the repository of works of art and architecture in my brain as a “database.” But it is none of that because my memories derive from physical experience and my manner of accessing them does too. Objects are not pixels or weighted numerical values attached to them; neither is my recollection of them. Anyway, so that’s all the preliminary stuff I do before I really begin thinking about the works on display, their place in the exhibition, and the story they’re being used by the curators to tell.

So I’m at the Morgan, and I’m eye-drunk on the splendor. But I get to this leaf, and it stops me in my tracks.

I’d never seen it in person before.

This is a leaf that was once dismembered from the famous Eadwine Psalter (aka the Canterbury Psalter). And I know this, but still. I do the thing. Because I’ve never before experienced it as a thing-in-the-world, as opposed to a reproduction, and I want to pop this badboy into my peepers. So I approach it, and for a moment I’m perplexed. This is, no-question, mid-12th-century England; this is the Eadwine Psalter. I’m not going to bother enumerating the minutiae, Morelli- or Berenson-style to say why I know it. I just do after having spent 25 years studying manuscripts from medieval England. Here’s another leaf that the Morgan has, which was also dismembered from this manuscript, but this is a scan of it, which, as you can see, flattens and deadens it. So I see it, and I know what it is, but there is something off about it. It doesn’t feel “right.” Now we know that the illuminators who made this manuscript were looking back at earlier models. The Eadwine Psalter is famous for that. One thing they were probably looking back at is this, the St. Augustine Gospels, a gospel book that was possibly brought from Italy to Canterbury in 596/7 (but certainly some time in the 7th century as part of the Roman mission to to the British Isles), and it remains just 100 miles away in Cambridge to this day. It’s fucking awesome.

The unusual gridded format might clue you into the fact that the 12th-century artist is looking to earlier books. But what defies enumeration, what emanates from the leaf as you perceive it with your own eyes and body is how purposefully archaizing it is. Yes, the artist is working in a 12th-century idiom with 12-century pigments, but he was also attempting to capture something of the older style, the forms from half a millennium prior. He was painting retro.

None of this makes for the kind of empirically authoritative work that gets valued in the academy. To a non-art-historian it probably sounds like I’m vibing. I get it. I know. And if I could program an algorithm to detect the degree of purposeful archaism in a work of art, I’d get grant money and esteem. But I cannot. There is no way to calculate how someone from the past saw the past. But to the extent that I’m an expert who has been researching this particular past for more than half my life, you’re going to have to trust me.

And ultimately, as I have said elsewhere, we cannot function as a society without trust, certainly not without trust in expertise. Especially when it comes to images.

Omnopoly Money

This morning I woke up to an Amazon box on my doorstep. This was weird. We haven't ordered anything lately from Amazon; we try to avoid it as much as we can (not impossible, but challenging when you live 90 miles from the nearest city). When I opened the box I found a copy of Luke Healy's Americana in it, a book my husband ordered from a used book seller on eBay. After a minute of confusion, the penny dropped.

Amazon Fulfillment must have captured another market. Of course it had. A quick search online confirmed this.

My husband and I traded laments. "This sucks," he said. "It's so fucking dystopian," I said.

"It's omnopolistic."

"Get that in print," he said. He's an economist, so I trust him on these things.

One of the few television series I've watched and rewatched is the fated space western Firefly, doomed to a single season after a few bad decisions by whichever rocket scientists were in charge of Fox in 2002. A ubiquitous and looming presence throughout its fourteen episodes is the logo of Blue Sun, and while the show was never given enough time to develop the story line that involved the corporation, it was clear that it was (a) everywhere and (b) Up To No Good.

 Whenever I see those sludgy blue Amazon Prime vans sluicing through my neighborhood I think of Blue Sun. The plump curving arrows on their sides promise easy cheer, but we hear too many stories of the relentless pace demanded of laborers by Amazon's ruthless logistics to believe it.

 We have words to describe Amazon and businesses of its ilk. Monopolies. Concerns. Multinationals. Oligopolies. Cartels. I'm sure there are others.

 But none of these words seems to capture the sheer dystopian creepiness of entities like Amazon (or Alphabet for that matter). Even megacorporation, which was taken up early on in science fiction, seems incapable of encompassing the unrelenting voracity, the all-consuming extractive will, the omnivorousness that characterizes their activities.

Omnopoly feels closer to conveying the perversions of what is already perverse. From “omni,” the Latin for “all,” It describes a kind of capitalistic agency that exceeds the engine of competition to such a degree that a ravenous all-encompassing dominance drives it in search of ever more markets. Scrambling “monopoly,” “omnopoly” is a deranged force, close enough to monopoly not to deny that it is indeed continuous with the system of capitalism. It is capitalism ad absurdum, what metastasized monopolies become.

 Amazon ships what the USPS used to deliver to my door.

 Amazon hosts universities' web-based services.

 Amazon produces your movies and music now.

 Amazon sells you your groceries.

 Amazon monitors your home.

Can You Become "Fluent" in "AI"?

No. But the Ohio State University says that its students are required to pursue this goal if they want to receive their degree.

Time was when engineers, linguists, and psychologists wanted to create programs modeled on human cognition and speech. In fact, three quarters of a century later, that time is still now, as thousands, if not millions, of poorly paid contract laborers across the globe are being exploited--arguably abused--to train large language models for commercial firms. The idea, to put it broadly, is to create a program that could speak human fluently. 

Which all makes the Ohio State University’s new initiative in "AI Fluency" so strange. I am not a linguist or a cognitive psychologist, so I'm unable here to speak to the concept of "fluency" itself, a complex and no doubt deeply contested one. What I'm interested in here is the use of "fluency" in this marketing campaign and the perverse use of this metaphor to sell this now compulsory program at a public, state university.

According to their website

 Ohio State is leading a bold, groundbreaking initiative to integrate artificial intelligence into the undergraduate educational experience. The initiative will ensure that every Ohio State student, beginning with the class of 2029, will graduate being AI fluent — fluent in their field of study, and fluent in the application of AI in that field.

 Ravi V. Bellamkonda Executive Vice President and Provost of OSU, is quoted on the same webpage as saying that all OSU students will "become 'bilingual,'" frankly a galling use of the term given that bilingualism itself has been nearly criminalized, as Americans are being targeted by ICE just because they're heard speaking Spanish...but I digress.

I wrote the term "perverse" above to describe how the metaphor of fluency is being used here. I find it perverse for two reasons. First, because it commits to an anthromorphizing fallacy: one does not "speak" or become "fluent" in AI any more than one can speak and become fluent in refrigeration. (And all of this begs the question of why someone needs to spend a semester learning how to use commercial software that is designed to be easy enough for a crocodile to use.) Second, the use of fluency here folds back on itself the very goal that the engineers of "AI" are attempting to achieve, which is to make algebra speak human.

 As with literally everything in life, Star Trek TNG provides great wisdom here.

In the episode, "Hero Worship," we meet Timothy, a traumatized young boy who has lost his whole family in a cataclysm suffered by their ship. Over the course of the episode we see Timothy suppressing his trauma by adopting Data's mannerisms and deciding that he wants to be an Android, just like Data. Because this is television, the episode takes us neatly through the arc of discovering what happened to Timothy's ship and the acute cause of his trauma (he thought he was responsible for the destruction after he accidentally knocked a computer panel), and eventually, though the help of Data's and Troi's counseling, he decides he wants to be a boy again.

 American universities have been beset by rolling waves of blunt force trauma over the last four decades. I think we can take a cue from TNG and decide that wanting to become fluent in a "neural net" modeled on human cognition that itself remains incompletely understood is folly. And maybe if more Americans watched TNG, they'd know how foolish this is.

"The Printing Press Democratized Knowledge": When Slogans Masquerade as History

The phrase is said so frequently it seems, like the mechanism it celebrates, to mechanically replicate itself. 

It's become a favorite catchphrase among tech boosters of any sort (see my post on Apple's "Crush" commercial), but especially now among the hustlers of "AI." The argument goes: just as Gutenberg's invention freed the minds of fettered and unlettered medieval Europeans, launching them into modernity, so "artificial intelligence" technology will elevate humanity to a higher plane of being. Forget about the ecological costs, forget about the expropriated labor, forget about the surveillance, forget about the reduction of social interaction and negotiation to calculation and prediction from the lowest common denominator of number crunched text. What matters is that history is a series of stages of revolutions, in which a sequence of technologies lead humans toward exceeding our own limitations. One day we will be Multivac.

 It's not difficult to dispute this lazy, analogic use of history. Lemme see if I can excavate the formula from the primordial part of my brain that retains stuff I learned for the SAT

printing press: democratized knowledge  :: "artificial intelligence" : ???

 The consequence on the other side of "AI" is rarely stated in concrete terms. Usually it's enough to say that because "the printing press democratized knowledge," AI will have a similarly beneficial effect. We see Sam Altman saying this and anyone acting as his surrogate now that they've bought it: whether it's Randy Weingarten justifying selling out public school teachers and students or start-ups crowing that they’re about to democratize music production.

 This phrase is everywhere. But while those opposed to the insertion of "AI" into every sector of life dispute that a text extrusion machine and racist pile of linear algebra can achieve anything so wonderful as its proponents use this analogy to proclaim, I have yet to see anyone in this context questioning the first formula in that analogy.

 So. To the phrase. "The printing press democratized knowledge."

Here's how I teach slogans that masquerade as history.

What, exactly are we talking about here? Is it the mechanical process of replicating by means of impressing an image or letter on to a surface? Because if that's the case, then the cylinder seals of ancient Mesopotamia "democratized knowledge." Forget about woodblock printing from Asia in the 8th century, humanity has had printed words for literally thousands of years.

So if it's not the mechanical process, is it the machine, the technology of the printing press and movable type itself? Well for that to be the case, then we would have to start talking about a whole host of other things because printed words on paper don't do anything sitting in a room by themselves unless you believe in magic. How about the production of paper, for one? But the words printed on those sheets of paper, they would have to go somewhere to be sold, so we probably want to talk about transportation infrastructure--roads, waterways, and all the vehicles and conveyances that trundled and floated over them. Once those words pressed on to paper ended up at their destinations, it wouldn't be very good at all if they didn't have literate people to read them. And it would have been real weird for some German businessmen to sink an absurd amount of money into creating a product for which there was no market. They had to learn to read from somewhere, and it wasn't a hulking machine in Mainz that immaculately birthed a literate customer base.

OK, OK, one says, but you know what I mean. The printing press MADE MORE BOOKS. It gave us SCALE!

But is that "democratization"? Democracy sounds so nice. I love it! But is that the appropriate term here? Democratization describes a political process of expanding the franchise and availing political representation to more and more people (the "demos"). We still need more of it here in the United States. We would do well to assist the process of democratization to the residents of Puerto Rico and Washington DC and to the formerly incarcerated and to those removed from voter rolls due to clerical error and “clerical error.” Clearly this literal definition is not what anyone means when they claim that the printing press achieved it, because the printing press has nothing to do with representative government.

What they mean here is that the printing press produced more units of certain class of commodity (words on paper) and made that commodity more affordable. As a result more people could own this commodity. Is this commodity "knowledge"? We'll get there in a sec. The issue is whether allowing more people to buy something is a democratizing process. I'm going to let people fill in a blank here because I will bet that you can think of things--some very terrible, violent things--that, when owned by more and more people actually thwart democracy.

Finally, to knowledge. It's ridiculous that I should try to dispatch the complexity of what "knowledge" is in three sentences, which is about the length of a paragraph that people are willing to stick with on a screen. I will simply say here that equating words pressed on to paper with knowledge both restricts what knowledge itself is and the vast number of channels by which humanity has acquired it over millennia (I, for one, didn't learn how to rollerblade by being taught from words pressed on paper), at the same time that it oversells what words on paper are (sure, everything is "knowledge," but when people use that word, they're generally dressing up whatever it is they're calling "knowledge" and I can assure you that a lot of what the printing press printed at scale was bullshit).  

No, I don't hate the printing press. What I object to is the use of slogans based on fairy tales. But "the printing press democratized knowledge" is not just a fairy tale; when you actually think about the words being used next to one another, they don't make the sense they are purported to make.

Yes, it’s important to dispute on its own terms that “AI” will achieve all the glorious things its boosters promise. But it’s also important to understand that the very historical example that is cited as proof by means of the precedent it set is itself a lie, largely because it makes no damned sense.

MOOCs and the Myth of "Access"

The recent piece by Vauhini Vara, “How Chatbots and AI Are Already Transforming Kid’s Classrooms” is important reading for the week, and there’s much in it to infuriate. It’s good.

But there’s one paragraph, arguably featuring the least enraging material, that I paused over because it’s a testament to how successful the tech industry has been in ensuring that the favorable language they use to frame and retail their products is the language we all use to describe them.

Vara writes

The outreach from Silicon Valley seems to follow a playbook developed during a decades-long attempt to turn public education into a market for private products. That effort has seen some high-profile failures: for instance, massive open online courses, which were supposed to expand access through web-based classes, and personalized learning regimes, which were supposed to use software to perfectly match students with material at their level. It has also included the enormously successful diffusion of screen-based education platforms such as Google Classroom. One lesson learned along the way is that school technologies rarely catch on without buy-in from teachers.

It’s deeply gratifying to read a journalist recalling the most recent colossal failure of the tech industry to “disrupt” education. That’s excellent. But it’s that bit I put in bold that we need to put pressure on, and only then can we start to become better advocates for ourselves as educators and (for those who aren’t educators) citizens who have a stake in all this, because our future depends upon the competencies and compassions of today’s students. And the tech industry is doing everything in its power to diminish that.

And so. That phrase I bolded: MOOCs "were supposed to expand access through web-based classes.” Sure, that’s one way of putting it. That’s certainly how they were sold. Another would be to say that their value proposition was for universities looking for new revenue streams. In that case, MOOCs weren’t about access. They were about expanding the market for education past what the physical structure of a university campus could accommodate and what the university was willing to pay for increasing the number of faculty to do the education. More students without the need for more classrooms, without the need for more professors.

No university went wild for MOOCs because they were excited about bringing ArtHist 101: Introduction to Art History to students in faraway places. They went wild because MOOCs were supposed to make them money, and for a fraction of the overheads.

That’s not “expanding access.” That’s creating and capturing a new market.

When vendors feed you language, it’s in the hopes that you’ll regurgitate it to sell their product for them. Like using an LLM to write for you, it’s easier than thinking for yourself.

Vara’s piece is good journalism and she’s obviously thinking for herself. My point is that all of us do this, all of us slip into canned phrases from time to time. The best we can do is audit those phrases, try to avoid them, and stay alert to the work that they do.

Will the Real Hildegard von Bingen Please Stand up? It's Replicas All the Way Down

A year ago, I received a fascinating email from a stranger. The stranger was a man who had recently traveled with his wife to visit family in Germany and had decided to visit Bingen with the expectation of seeing the original paintings made by Hildegard herself in the twelfth century. However, the man had found Hildegard to be "quite elusive!"

The paintings were nowhere to be found. There were backlit, enlarged reproductions of her paintings. But the original, twelfth-century works of art themselves were just..not there. He went on to write,

I asked the one attendant in the museum where are the originals? Her English was not very good, and my German is worse. In response to my repeated questioning, she kept pointing to the area of these reproductions saying "there, there, there they are." But those were not them.

The man left the museum with questions and searched the internet with no luck for answers. Hence, his email to me. Having found my name associated with Hildegard, as this was shortly after I published this article on the most famous manuscript of her magnum opus, Scivias, the man emailed me to see if I might help sort this matter out and offer some information to his question: "where are originals of Hildegard's paintings?"

I wrote back with what was sure to be a disappointing response:

 Dear ________,

 Although I have not visited the Museum am Strom, I suspect the paintings to which you're referring are those from the famous Scivias manuscript, which was not painted by Hildegard herself but rather is believed to have been produced under her supervision some time in the 1170s at the Rupertsberg Abbey.

 I'm sorry to say that this manuscript, which had been in the Wiesbaden Landesbibliothek until WWII, was placed in a bank vault in Dresden for safekeeping during the war. When an attempt was made to recover it, it had disappeared from that vault. Only one manuscript believed to have been produced under Hildegard's supervision (which lacks illustration) that was also placed in that vault was recovered. You can read the exciting story of its recovery here [in a great article by Jennifer Bain]:

 https://theconversation.com/how-two-women-pulled-off-a-medieval-manuscript-heist-in-post-war-germany-130179

 Regarding the reproductions you saw: a number of photographs of the original Scivias manuscript were produced in the early 20th century, and it is likely that you were seeing prints of those, which have been widely reproduced. Alternatively, if the reproductions you saw were in color, then they were taken from a hand-made replica (a facsimile) of the original manuscript that was made by nuns of Eibingen Abbey between 1927 and 1933. That facsimile is the subject of my Modern Language Quarterly article, the full text of which I attach here so you may learn more about its production.

 I am sorry to report the bad news about the disappearance of the original manuscript!

 Not long thereafter, the man responded to my email with a kind and engaging note that thanked me for my explanation and expressed the deep frustrations he experienced in his efforts just to be in the presence of some important works of art, works that he genuinely thought he could see. The internet, he said, is awash in confusing information about these artifacts. This is no doubt unaided by tourism boards that are rather coy about how many of them are replicas.

 In the year since this email exchange transpired I've returned to it in my mind to think about the stakes of replicas, of information and histories that become untethered from the reproductions that make up so much of our exposure to art and the artifacts of the Middle Ages that I study.

Rule One in the Intro to Art History class that I teach is: these things on the slides here are not the works of art that you're studying; they're digital copies of them. Incidentally, the most gratifying and annoying thing my students do after that is ask me for the dimensions of works of art when I've forgotten to include them on my slides. Gratifying because clearly my admonition was heeded; annoying because goddammit labeling things is a pain in the ass, and my neck already hurts from spending hours on my Powerpoints. Anyway, it's genuinely gratifying. But also my neck hurts.

I've never thought that interpretation or even enjoyment of art in reproduction can't be done. That would be ridiculous. I'd not be in this profession. But I do think that it's important to know and to teach people to develop an understanding of how knowledge itself gets made and lost when we come into contact with, when we look at reproductions.

The internet hasn't caused us to fall from a state of auratic grace. It's just accreted one more medium to the media, channeling media, channeling media that have always serialized originals (and yes, this precedes the printing press; for the love of God, please don't think this has anything unique to do with the printing press).

I love that this person reached out to me. I'm so glad he was appreciative of the conversation we had. A year after he contacted me, I'm still appreciating it too.

 Here are two replicas of replicas of replicas.

The Sympathetic Magic of Education Reform: Tech woo-woo

I enjoyed reading this week Neil Kraus’s The Fantasy Economy: Neoliberalism, Inequality, and the Education Reform Movement (2023). The book’s core claim is that corporations and elites who subscribe to neoliberal tenets have promoted a narrative that has succeeded in becoming the conventional framework for addressing education reform. This framework he calls the “fantasy economy,” one in which there are abundant jobs but not sufficiently skilled and technologically educated graduates to fill them, as opposed to the real economy in which there is wage stagnation and rife underemployment. According to Kraus, “The fantasy economy's framing of economic inequality… focuses exclusively on education and intentionally deflects attention away from decades of public policies and changing business practices that have directly and indirectly contributed to stagnating wages for most while simultaneously producing extraordinary income and wealth gains for a small minority.” In short, corporate interests and the foundations that business elites have established to disseminate their ideology have promoted education reform as the solution to economic problems and inequality (which are not within the power of educational institutions to solve).

One thing came to mind as I was reading his book, and it’s the idea of sympathetic magic.

[Let’s leave The Golden Bough out of this if for no reason other than that I’ve never actually read it, but my second-hand knowledge of it is that as a work of 19th-century anthropology it’s full of the standard Victorian pretenses to rationalism that were part and parcel of the colonialist nous. I’m just thinking of sympathetic magic as a concept with a more common, colloquial currency, the belief that like causes like and, similarly, the remedy to an ailment resembles that ailment. If anyone has a better term for what this characterizes, let’s go for that. Woo-woo? Maybe that’s all this is. Tech woo-woo.]

For the last forty years we have been told that there is a high-tech global economy out there just pumping out new and exciting jobs, which American students simply cannot get because of failing secondary schools and a weak system of higher education. This claim is fallacious, those jobs are not there; instead we have underemployment and wage stagnation. Nevertheless, the solution proposed is always to introduce more technology into classrooms, to optimize, digitize, and technologize what we as educators do and the materials from which students learn. Despite the absolute failure of technology to have any positive impact on either educational or jobs outcomes, this has been the unchanging refrain for years.

Teaching with more digital tools does not exciting jobs in a high-tech global economy make. Nor would it do anything to prepare students for those jobs were they there. The most comical thing about promoting “tech readiness” among students is that it’s a legacy line from c1995 when it actually did take about half a day to get your Gateway 2000 up and running. It’s 2025, a caterpillar can figure out how to use apps.

Suggesting that there’s an app for education is not “innovation.”

It’s just sympathetic magic.

Tech woo-woo.

A Dictionary of "AI"

Even if you spend just a small amount of time listening to people who believe that any kind of digital technology (whether that’s “artificial intelligence” or not) is an unquestioned good, you’ll hear a set of terms used to frame the conversation. Most of these terms have become a common part of public discourse, and they comprise a standard vocabulary of accepted values. Often, when they are invoked, no one feels the need to follow up by asking what is being promised. A now-classic example: say something is “democratizing.” When someone says that the internet is “democratizing,” few will ask what that actually means and how the internet can achieve it.

Four decades of scholarship, from Langdon Winner’s The Whale and the Reactor: A Search for Limits in an Age of High Technology (1986) to David Golumbia’s Cyberlibertarianism: The Right-wing Politics of Digital Technology (2024) have shown how powerful rhetoric is in paving the way for certain technological regimes to determine the shape of our lives: not technology itself, but the language used to frame how people think and talk about technology.

It seems like we could do with a small dictionary that defines how this common vocabulary is being used by the proponents of “artificial intelligence.” So here it is.

*Update: with thanks to Joseph Potts and Peter Tarras for their contributions, I’ve added “empowering” and “tool” to the dictionary (August 27, 2025).

**Update: with thanks to Brendan Greeley for his contribution of “regulatory uncertainty” and to Emily Bender for her post on Bluesky reminding us all of the canard about technology moving fast. And while I’m here, if you want more in your anti-AI arsenal, read Emily M. Bender and Alex Hanna’s excellent The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want (2025) (August 28, 2025).

How to Read an "AI" Press Release

Every so often someone like Mark Zuckerberg or Sam Altman will dribble out some unadorned text, announcing with stentorian certitude the advent of a new world that their latest product will avail. Zuck seems to love dressing up his thought bubbles in Times New Roman for the purposes of LARPing intellect, which I find funny and tragic.

I take an interest in this genre of bad copy not only because of the failure of journalists to report on its inanity, bullshit, and empty promises. I take an interest in it also because I’m writing a book on the relationship between the tech industry and the discipline of history: in particular how the stagist, teleological framing of the past that one finds in Dad History books (you know what I’m talking about: ones with titles like, How Thomas Edison Made the Modern World, or The Swerve) has served the argument that the next “stage” in human development will be led by a given technology invented by or pushed to the extreme of its potential by singular Great Men, in this case, “AI” in all its weird bastard forms. Bad history is the basis for the vision of the future being written by tech CEOS to sell their products.

Anyway, every time I read one of these very unserious sequences of predictable PR, I mentally annotate the nonsense, since journalists routinely fail to do it for all of us, which should be their job. But I had an hour to spare and a big cup of coffee, so, to exorcise this from my system, herewith: How to Read an “AI” Press Release. The title is deliberately ambiguous. Was this digested and regurgitated by a LLM from the massive cache of such press releases disguised as the Deep Thoughts of Visionary Bros? Who can tell? Original text that I’ve annotated is here.

That's Numberwang! On Quant Methods and Medieval Manuscripts.

When I lived in the UK I watched a lot of Mitchell and Webb. A British comedy duo who should be more famous among Americans than they are, they made sketches just the side of absurd where—unlike Monty Python—things never devolved into the annoyingly silly (or racist). They did a recurring sketch which, signature Mitchell and Webb, sent up both the subject of the sketch as well as the other comedy shows that sent up the topic in the past: the network television game show.

This is Numberwang, and it is one of my favorite things. 

The sketch features two contestants who alternate calling out arbitrary numbers until the host—Robert Webb as Chuck Woolery on uppers—calls out “That’s Numberwang!” before a new round of random number-calling begins. The comedy of the sketch relies on the two contestants, fully committing with studied gravity to the nonexistent logic of the game. Robert Webb’s sunny excitement as he rewards the correct number is the punctuation that pitches its absurdity into the comic.

Numberwang came to my mind when I read the most recent medieval manuscripts study making ripples online because of its promotion in Nature—this is a semi-regular event when medieval studies gets invited to the beach by science and enjoys a moment in the sun. 

The study ensues from a simple question. It asks: “What was the quantitative contribution of female scribes [to the production of manuscripts in Europe between 800 and 1600 CE] based on available sources? How large was the fraction of the manuscripts copied by women?” After citing some of the  historical research that has examined the work of female scribes in the Middle Ages the authors refer to the target of their research (how many were there?) as a “gap in the knowledge.” Strictly speaking, yes this is a gap in our knowledge: nobody knows know how many female scribes there were working in Europe between these years, nor do we do how many manuscripts they produced. But we don’t know how many male scribes were working during this period either, and when you put it like that it seems like a strange question to ask. 

This isn’t to say that quantitative analysis shouldn’t be done or can’t be useful. Looking at a narrow window of time in England and using financial records detailing the amount of wax that chancery scribes paid for as the basis for extrapolating how many sealed documents issued from the office of chancery per day: this strikes me as a really clever use of quantitative analysis based on pretty firm foundations (and it’s why Michael Clanchy had the well-deserved vaunted reputation he had).

But historical research is fun like this, and we can ask whatever questions we want even if skeptics like me can’t see the value of the question asked. If I had my druthers, I’d follow the lead of Linda Nochlin: I wouldn’t try to collect a full deck of female scribes to prove their place in history, inevitably coming up short; I’d inquire into the institutional conditions that impeded their access to the opportunities enjoyed by men, or I’d wonder about the nature of their training and how it differed from the educations afforded to their male counterparts; I’d ask about the criteria for success, who got to dictate that criteria, and whether elements of it were framed specifically to place the goalposts beyond what women’s training could reach.

Nevertheless, the authors are in search of numbers. So they approach an answer to their question by using what they call “bibliometric analysis” based on the data provided from colophons as contained in the multivolume Colophons de manuscrits occidentaux des origines au XVI e siècle compiled by the Benedictines of Bouveret between 1965 and 1982.

By definition, a colophon is a statement made within a book which contains details about the production of the book in which it’s found (this information could run the gamut from containing the names of everyone involved in the book’s production to the date and place of the book’s completion, alongside complaints by the scribe about what a long boring task it was; other colophons might only offer a date or a place or a single name). Colophons are notoriously cagey forms of historical evidence: sometimes scribes copied them verbatim from their exemplars (the books they were copying), and this even occurred when scribes copied by hand printed books. It’s very fun finding a manuscript that announces in  handwritten ink where it was printed. Moreover, the impetus to include a colophon was often not present, and swathes of manuscripts produced in the Middle Ages (most of them, in fact) contain no such information at all. So they’re an imperfect guide to the subject of scribes’ identities and their labor output to begin with.

But as the authors of this paper themselves acknowledge—and they are scrupulous about outlining many of the weaknesses of this dataset—the six volumes of colophons compiled by the Benedictines of Bouveret are even less reliable than the colophons they record (and there are, further, unknown historical variables that resulted in a disproportionately higher survival rate of manuscripts produced by men). Reviews of the volumes consistently refer to their inaccuracies, their gaps, and in the words of one scholar who reviewed the final volume and summarized how the entire set should be approached, “caveat lector.”

So this is the dataset from which the authors proceed, and after counting up all the 254 colophons that identify a female scribe (1.1% of the 23,774 colophons in the catalogue), they extrapolate this percentage to some other extrapolation of all the manuscripts that might have been produced between 800 and 1600, based on an estmation of loss rates (a notoriously uncertain conjecture, and, again, the authors note this). They calculate 110,000, of “which at least 8000 should still exist.” 

The authors refer to their work as a “first step,” but the questions to which they point don’t ensue from their conclusions. They suggest that scholars might attempt to identify as-yet unknown communities of female scribes from historical records or they might consider”socio-political and socio-economic links to literacy, throwing light on when, why and how women worked as scribes during specific time periods.” These are questions that historians have already, for a long while, been asking. The “big picture” provided by the present study, built on a foundation of extremely partial and often incorrect data, is not actually a first step towards any of this.

I admit I’m not sure what the aim of this paper was given that the guiding question is not answerable based on the data or methods used to approach it. The project was supported by the Trond Mohn Research Foundation, which awards grants “to support research at the interface between basic research and clinical research at the University of Bergen and Haukeland University Hospital.” This is all very much above board: a foundation has money to offer research that involves scholars in clinical research, and that’s exactly what the authors of this paper did. I wonder, though, whether the existence of the grant was the impetus here. And if that’s the case, then molding a project to satisfy its conditions engendered the results the usefulness of which the authors do not articulate. 

But now we have a piece circulating the number that 1.1% of manuscripts were produced by female scribes, which might have once amounted to 110,000 books, 8,000 of which might still be around.

That’s Numberwang!

From Jewish Badges in 1218 to Technofascism in 2025

A short book I’m writing with the somewhat forbidding title, Book History Is Not the History of the Book: Four Essays from Fifteenth-Century England, is founded on the simple premise that the traditional, linear formulas that have structured the field of book history (from orality to literacy, from roll to codex, from manuscript to print, and from print to digital) represent a technologically deterministic and teleological view that belies the complexities of the history of the book.

The introduction to this book takes as its starting point the moment when the idea that “the book” has a history came to be codified, politicized, and racialized in thirteenth-century England. It is, as I argue, no coincidence that in this moment, Europe’s first mandates to visibly tag a minority population was first enacted by law and that the form of that badge was what was seen as an obsolete technology—namely, the tablets of the Ten Commandments.

Publication pipelines are long (and I’m trying to finish another book first!), and scholarship demands patience that is often ill-suited to the urgencies of the moment. So I post here the introduction and the conclusion to the essay that opens my book.

INTRODUCTION TO BOOK HISTORY IS NOT THE HISTORY OF BOOKS

In May of 2024, Apple aired a commercial that aspired to showcase the revolutionary power of their latest gadget by crushing within the gargantuan plates of an industrial press the implements and artifacts of creative culture: a trumpet, an arcade game, cans of paint, a piano with sheet music, a metronome, a sculpted bust in clay, cameras, a television, a guitar, notebooks, and bizarrely a toy emoji. And when that press releases its clench, all that remains on its bed is an antiseptic tablet, vapid in its gleaming sterility. It is a violent declaration of techno-utopian independence, convinced of the creativity of its own destruction. It is no wonder the ad was unanimously reviled and retracted; no wonder, because it lays bare in celebratory tones the technochauvinist vision that is a hallmark of our age. This is the legacy of uncritical arguments about the liberating novelty of the press to supersede the bounding conditions of history, the springboard required to launch humanity out of the Middle Ages into modernity.

While arguments against the role assigned to the printing press in birthing western liberal democracy have been voiced for a long time now, these criticisms tend not to take into account the broader historiographical model on which this idea is premised: specifically, a Christological paradigm of book history. This is a supersessionist one with Pauline origins that periodizes the history of the book into four phases: from orality to literacy; from scroll to codex; from manuscript to print; from print to digital. In what follows, I want to hone in on one important moment in this history when arguments about the obsolescence of certain textual technologies were yoked to Jews, forever racializing the history of the book. In her 2019 monograph, Race after Technology Ruha Benjamin implored us to scrutinize the kinds of assumptions that underwrote Apple's ad, asking "what is hidden by an idealistic vision of technology. How is technology already raced?” For Benjamin that means interrogating the racist assumptions written into modern computer technology. But long before those assumptions were inextricably woven into code, it was the codex itself that encoded hegemonic ideals. The case that I'll be making is that a landmark instance of the racing of the book as technology occurred in England between 1218 and 1290 and that the impact of this racing continues to be felt today.

[…. my essay then recounts the history of the Jewish badges, why the shape of the Ten Commandments was chosen, and the ideological legacies of this history]

CONCLUSION

Supersession, replacement, a progressive movement away from the material: these ideas are so entrenched in conventional book history, are so naturalized that it has been impossible both to see its Christological origins and orientation and to comprehend the narratives that it continues to underwrite. I have often told people that I am a medievalist for the same reason that I read dystopian novels: nothing makes me think better than to see my own world reflected back at me with enough difference to make me question that which I take for granted. In Gary Shteyngart's 2010 novel Super Sad True Love Story our protagonists inhabit a near future in which America, on the brink of economic collapse, has only three main industries: media, credit, and retail. The social media website that dominates public discourse has promoted the idea that paper books are, in a word, gross. In one text message to her friend, our narrator Eunice Park describes her budding love interest, the nebbish Lenny Abramov,

Anyway, what kind of freaked me out was that I saw Len reading a book. (No, it didn't SMELL. He uses Pine-Sol on them.) And I don't mean scanning a text like we did in Euro Classics with that Chatterhouse of Parma I mean seriously READING. He had this ruler out and he was moving it down the page very slowly and just like whispering little things to himself, like trying to understand every little part of it.... I thought Ben was really brain-smart because I saw him streaming Chronicles of Narnia in that cafe in Rome, but this Tolstoy was a thousand pages long BOOK, not a stream, and Lenny was on page 930, almost finished.

 Later in the same message to her friend Eunice describes going to Tompkins Square Park where she sees

all these Low Net Worth Individuals there and they're camped out with all their dirty things... and they have all these really old computers they try to boot up but really they have no Images or streams. So after Sally left I ran home and got all my old äppäräti and I gave them out... They were so happy to have my old ghetto stuff that it made me sad because this is what their lives have become and just last year some of them worked in Credit or were engineers.

I was reminded of this novel when I was re-reading Sara Lipton's Images of Intolerance. Speaking more generally of the trope of the Old Law's obsolescence, Lipton points to an aesthetic attached to this idea, writing "In text after text, the Law's obsolescence is attributed not to its honorable expiration or to an orderly passing of the baton but to its own inherent aridity, deformity, and ugliness." There is a violence to casting the obsolete off the page, not unlike the brute force committed by the industrial press crushing the notebooks of the past, all to make way for the shiny, the new, ever striving to reach the immaterial. Long ago Kathleen Biddick argued that "the purported 'secularization' of modernity... has never overtaken this core Christian conception of supersession." What I hope to have shown in capsule form that is state-imposed signs of Jewish identity, mapped on to books and bodies reveal technology's place in this history. If the funding of far right wing figures like J.D. Vance by tech billionaires like Peter Thiel seems perplexing, it is because we have completely forgotten the supersessionism that makes such an unholy alliance ideologically coherent: it is not surprising that when Vance followed Thiel's advice and converted to Catholicism he chose the committed supersessionist Augustine as his patron saint; Thiel's less-subtle approach was to tout Rene Girard's Catholic anthropology and the socially redemptive role of the scapegoat.

The Colophon in the Rüdesheimer Codex of Hildegard von Bingen's Scivias

In the December 2023 issue of Modern Language Quarterly, I published an article that is the first to examine the circumstances surrounding and production of a renowned manuscript facsimile of the Rupertsberg Scivias, the most celebrated copy of the visionary text by Hildegard von Bingen. That manuscript facsimile was copied by three nuns of Eibingen Abbey and illuminated by the artist Josepha Knips between 1927 and 1933. You can access a copy of the article here. Sonja Drimmer, “Hildegard’s Scivias in Weimar Germany: Media Theory by Hand,” Old ‘New Media’ and Literary Studies before 1960, ed. András Kiséry and David Nee, special issue Modern Language Quarterly 84.4 (2023): 443–464.“

My main concern in that article was to take seriously the the 20th-century scribes’ and artist’s labor as articulating a theory about reproduction and mediation by comparing it to an influential debate about the role of replicas and facsimiles in public education (known as the “Hamburg Facsimile Debate”), which took place precisely as these women were producing their facsimile. Despite a paucity of textual material outlining the manuscript producers’ ideas about these subjects, the facsimile itself, I argue, reveals not simply their methods but also the theory guiding these methods—that is, a methodology.

However, there is one extensive textual document that does tell us something alongside the traces of the scribes’ and artist’s work. That document is the colophon inscribed on a final leaf in the facsimile. I discuss this colophon in my article and include a picture of it, but due to the word limit, I was unable to publish my full translation. To my knowledge this colophon has only been reproduced in subsequent facsimiles of the Rüdesheimer facsimile, but nowhere else.

My translation follows immediately below, and I include my transcription of the original Latin below that. If you would like to use it for your work or teaching, cite it as follows: “The Colophon in the Rüdesmeimer Codex of Hildegard von Bingen's Scivias,” translation by Sonja Drimmer. https://sonjadrimmer.com/blog-1/2024/1/19/the-colophon-in-the-rdesmeimer-codex-of-hildegard-von-bingens-scivias. January 19, 2024.

My thanks to Joy Elisabeth Reeber for reading and offering corrections to my translation.

Colophon. Rüdesheimer Codex. Rüdesheim am Rhein, Abtei St.Hildegard, Hs.1.

TRANSLATION.

This codex contains the most accurate, most true likeness, in both word and picture, of that glorious book, Scivias, which the very Abbess of Rupertsberg at Bingen St Hildegard herself composed and wrote.

This most precious original work held in the sacred building of St-Rupert at Bingen, which was devastasted and destroyed by Swedes in 1632 AD, was deposited in Eibingen monastery. Then after the first building of St-Hildegard of Eibingen was desecrated by secular use in 1804, it was joined to the public library “Aquas Mattiaca” [the Wiesbaden Kurhaus, which bears those words on its entablature over the entrance] where it is kept to this day.

The reverend mistress Regintrudis Sauter, who occupies the thirty-second position as abbess of the monastery at Eibingen after St. Hildegard wanted this excellent work of her own holy fore-mother restored in the first and truest form of the pictures/tablets and letters according to the  exemplum by her very own skilled daughters.

His Majesty Ferdinand, Emperor of Bulgaria, most August Lord, gave the parchment and pigments necessary for accomplishing/completing this work as a gift of the greatest princely munificence and nobility, out of the closeness of his connection to the Reverend Lady Maria Benedicta, a sister of this convent, who was born a duchess of Württemberg and was [then] called/dedicated [to the same].

The ingeniously designed and ornamented binding of this book was made with elegance by hand by craftsmen at the Abbey of Maria Laach, completed by means of the singular, true method in the office of the Herder Verlag in Freibourg.

This most serious work (i.e. copy?) of the original as though [it had been] created and fashioned anew finished within the span of a year and six months to the 25th anniversary of Reverend Mistress Regintrudis Sauter’s Abbacy on the 8th day of September in 1933, the jubilee year of our redemption, celebrated with the highest joy and with most intimate devotion.

This divine work is a gift of God, thanks be to God.

Colophon. Rüdesheimer Codex. Rüdesheim am Rhein, Abtei St.Hildegard, Hs.1.

original text.

Hic codex continet accuratissimam et in verborum contextu et in miniaturis picturisque verissimam imaginem incluti illius libri, quem ipsa Abbatissa monasterii in monte Sancti Ruperti ad Bingum quondam siti Sancta Hildegardis scripsit et inscripsit Scivias.

Pretiosissimum huius operis archetypum sacris aedibus S. Ruperti ad Bingium anno Domini 1632 a Suecis pervastatis atque destructis in monasterium Eibingense delatum est. Unde postquam prius aedificium Sanctae Hildegardis Eibingense ad usum saecularem anno Domini 1804 profanatum est, bibliothecae ad Aquas Mattiacas publice institutae adjiunctum est, ubi adhuc asservatur.

Domna Reverenda Regintrudis Sauter, prima redintegrati monasterii Eibingensis Abbatissa, quae trigesimum secundum post sanctum Hildegardem huius coenobii obtinet locum abbatialem, hoc eximium Sanctae Proavia suae opus in prima verissimaque tabularum litterarumque forma ad exemplum reddi voluit per suas ipsius filias artifices.

Membrana et colores ad opus perficiendum necessarios dono dedit summa cum principis munificentia ac liberalitate Sua Maiestas Ferdinandus, Bulgagorum Imperator, Augustissimus Dominus propinquitate coniunctus cum Reverenda Domna Maria Benedicta, huius monasterii moniale, quae ducessa de Württemberg nata est et nuncupata.

 Huius libri tegumentum ingeniose compositum atque ornamentum elegantiaque manu fictum est ab artificibus Abbatiae B.M.V. ad Lacum, singulari vero arte confectum est in officinis Editionis Herderianae Friburgensis.

Gravissimum hoc opus archetypi quasi denuo creandi atque formandi moniales Eibingensis intra sex mensium et anni spatium ad vigesimumquintum anniversarium Reverendae Dominae Reginstrudis Sauter Abbatissae die 8 Septembris anni iubilaei redemptionis nostrae MCMXXXIII celebratum summo cum gaudio intimaquae devotione perfecerunt.

 Est Donum Dei hoc opus divinum Deo gratias.