[This is part 4 in a series of posts building off of James Scott’s Seeing Like a State; it references topics from Part 1, Part 2, and Part 3.]
I saved this idea for last, because I wanted to tinker around with it a bit. Scott closes his book, as I’ve mentioned, by discussing metis and techne as complimentary forms of knowledge, and the way that large-scale social engineering tends to privilege the latter at the expense of the former. But Scott makes it very clear that the point of his book isn’t to argue against the many positives that have resulted from these schemes, nor is it to critique the type of knowledge required to produce them. Instead, he “want[s} to make a case for institutions that are instead multifunctional, plastic, diverse, and adaptable—in other words, institutions that are powerfully shaped by mētis.” While this sounds pretty idealistic, Scott does provide one example1 of such an institution, and that’s language.
Finally, that most characteristic of human institutions, language, is the best model: a structure of meaning and continuity that is never still and ever open to the improvisations of all its speakers.
Not that I needed a great deal of convincing that Seeing like a State was relevant to my interests as a rhetorician, but if I did, this closing line would have sealed the deal. And that’s why I wanted to spend the time with it that I have thus far.
Let’s Get Cooking
This may seem like a roundabout path, but I’m going to spend the rest of this episode arriving at a metaphor that occurred to me as I was thinking about Scott’s book as well as Raghavan and Schneier’s update of it. In order to get there, I need to spin out an extended analogy, one that’s rooted in my own experience as someone who’s thought closely about writing and language for a living. I’m going to write about the resemblance between writing and cooking.
The funny thing about this is that one of the earliest texts about rhetoric is Plato’s dialogue Gorgias, where he has Socrates compare rhetoric to “cookery,” and not in a flattering way. Rhetoric, he explains, is to justice as cookery is to medicine. Doctors attend to our health in “real” ways, while cookery is effectively distraction.
What does cookery offer, then, if not health? Why would anyone seek the services of a cook instead of those of a doctor? Quite simply, Plato says, cooking aims to flatter. The cook offers pleasant feelings, but feelings which masquerade as health – or which are so pleasant that health comes to seem unimportant. So of course anyone who doesn’t know any better will be attracted to the offerings of the cook. Cooking involves no genuine knowledge of bodily health, but it acts as if it does – Plato says it ‘impersonates medicine’ – and it is treated as if it does by those who do not know better. (Heldke)
For Plato, the medicine/cookery opposition (attending to the body) maps neatly across justice/rhetoric (attending to the soul). This attitude towards rhetoric—that it’s empty, flattering, manipulative, deficient—has persisted throughout millennia, so I’m not really interested in refuting it here. (It makes sense from a certain philosophical perspective, one with which I generally disagree, as you might guess.)
Neither do I want to spend too much time laying out the groundwork for this resemblance: the ways that both writing and cooking take a set of ingredients, apply procedures and technologies, involve genre and style, can provoke joy, satisfaction, discomfort, confusion. For a variety of reasons, I’ve always found the two activities to be deeply resonant2. I love cooking shows and especially food competition shows: Iron Chef, Top Chef, Chopped, Beat Bobby Flay, the Gordon Ramsay family of products—I have seen and will watch them all. In the latest season of Top Chef, a couple of episodes ago, host Kristen Kish remarked, “I feel like storytelling is the thing that separates good food from exceptional food.” (The frequent mention of storytelling as a culinary virtue on this season of TC was one of the reasons for my own ruminations). The connection may seem flimsy to some, but I feel it deep in my bones.
I’d also be remiss if I didn’t mention a connection that I wrote about a year ago, Robin Sloan3’s comparison of internet apps to home-cooked meals, where he describes himself as “the programming equivalent of a home cook.” He offers a different take on the massive cultural wave of folks telling us to “learn to code” (for economic reasons), asking instead what if we treated coding like cooking:
The list of reasons to “learn to cook” overflows, and only a handful have anything to do with the marketplace. Cooking reaches beyond buying and selling to touch nearly all of human experience. It connects to domesticity and curiosity; to history and culture; to care and love.
When I read this, my mind substitutes writing for cooking, which we do for most of the same reasons and connections. And I want to capture that same spirit for this analogy—it’s not that writing and cooking are somehow the same, but there’s something fundamental to each activity that resonates, enough that thinking about food metaphorically might tell us something about rhetoric or language or writing.
The Tropes that Eat like a Meal
One place where writing and cooking diverge is that the latter has a foundational unit of analysis, the meal, that is a lot more consistent than anything we might suggest for writing. In part, that has to do with physical needs; it’s a lot easier to go one or more days without reading than it is to go without eating. This is one of the things that makes this analogy a little looser than perhaps I’d like. A book feels conceptually more substantial than a single meal, but anything shorter (an essay, a newsletter, blog, or social media post) doesn’t seem sufficient. There’s also the fact that meals are often more immediately communal; it’s not that we don’t share language, especially if we’re speaking face-to-face, but we tend to experience it more individually, even when we know the person on the other side of a given text. But there are echoes, too: when I publish posts here, I have the same intuitive sense about whether or not they represent a “full plate.” Even when I write about wildly different topics, I tend to use the same sorts of discursive “spices” in my treatment of them. And I have certain “ingredients” that I gravitate towards. I could probably pack a whole bunch of additional cooking metaphors into this discussion, but I want to push forward with the idea of the meal.
That idea of the home-cooked meal carries with it a great deal of tradition. Not only does it summon up a certain sense of domesticity, but it holds connotations for family, for holiday celebrations, for religion in various ways, and for our ability to get along with one another on a fundamental level. While I’m conscious of the fact that some of those features and traditions can be (and have been) politicized (treating the nuclear family as the norm, gender roles associated with “traditional” domesticity, etc.), my own goal here is simply to think of the home-cooked meal as the innermost in a series of concentric circles.
One step out from that would be something like the potluck, which doesn’t stray too far from the associations we have with home-cooked meals. Potlucks are social, cooperative, and play a role in community-building. When I was growing up, our family hosted an annual 4th of July potluck/party for friends, neighbors, and relatives, which I always think about this time of year. There’s something faintly transactional about the potluck, but it’s more of a barter network: I’ll grill you a hamburger in exchange for some of that fruit salad from your carved-out watermelon or one of those home-made brownies. Each of us contributes our share; everyone gets a full meal out of it.
The next step outward becomes a bit more transactional, though not purely so. Going out to eat at a restaurant can be celebratory, a labor trade-off, or even a means of supporting a community or culture, if it’s locally owned. Restaurants are not typically thought of as profit centers—you don’t need to spend much time on the Food Network to know that the vast majority of restaurants struggle to survive over the long term. While they wouldn’t technically (obviously) provide home-cooked meals, local restaurants may still emerge from a community, providing jobs, sponsoring events, featuring local or regional ingredients, and expressing the point of view of their chefs.
Casting our view wider still takes us to the place where Scott’s notion of large-scale engineering enters the picture. I think of this stage primarily in terms of the franchise, where the personal component of cooking tends to recede. Whether we’re talking fast food, casual dining chains, or the “fast casual” hybrid, this stage operates with economies of scale. The art of cooking is proceduralized, with the kitchens themselves optimized for the quick and efficient production of a specific set menu of items, conveniently and consistently made available to customers who expect no more (and no less) than what they already know about it. This is the realm of French sociologist Marc Augé’s “non-places”: walking into a McDonald’s or an Olive Garden, ordering from the menu, and having a meal will be roughly the same experience regardless of where in the country you do so. (It’s hard not to sound a bit judgey, but I genuinely don’t mean to—the person at Applebee’s preparing one’s food doesn’t necessarily care any less than the local restaurant chef, but the system they’re embedded in requires them to care in different ways and about other things.)
Last summer, I compared the internet to a shopping mall food court (I called it the Sbarro-fication of new media), so that stage of things deserves at least a mention. Food courts are kind of weird fast food potlucks, positioned to take advantage of relatively captive audiences. On the one hand, they’re sites of direct competition, which implies the purest version of actual marketplace economics. And yet, the more successful any one station might get, the longer the line becomes, and the more likely it will be that footsore shoppers will take the quality/wait trade-off (should I wait 20 minutes in line for B- food, or eat the C food immediately?). The different “restaurants” are primarily concerned with appearing distinct from one another, more than anything, and of course, turning around customers as quickly as possible. In an odd way, though, they depend upon each other to a degree, and upon an implicit agreement to a certain standard of bare efficiency (and, I’d argue, mediocrity, if I weren’t trying to avoid being judgey).
Spooky Food
For most of my life, if I asked myself, “what am I going to do for dinner?” those were my options, ranging from fixing something for myself at home to going out, with various levels of commitment to quality or convenience. There are some other trends I might observe—the explosion of cookbook sections in the bookstore, the rise of foodie culture and celebrity chefs, the effect of social media, dinner clubs and pop-up restaurants—but I want to focus more recently. Raghavan and Schneier, in their extension of Scott’s ideas to contemporary data structures, spend some time talking about ghost kitchens4.
No longer did the restaurant a customer was ordering from actually have to exist. All that mattered was that the online menu catered to customer desires. Once ordered, the food had to somehow get sourced, cooked, and packaged, sight unseen, and be delivered to the customer’s doorstep.
As they note, this process was “more like Amazon’s supply chain than a local diner of yore,” and as with Amazon, we allow(ed) ourselves to be priced (and gouged) according to a traditional value model while they reap the profits generated by the scale of their operation. If you’ve watched restaurant makeover shows (as I do), you’ll know that of the price that we pay for a restaurant meal, only about 25 to 30% of that is food cost, the actual cost of the ingredients. The rest of that price pays for overhead costs, salaries, labor/expertise, etc. Just as Amazon charges regular book prices for its e-books, whose cost to reproduce is nominal, ghost kitchens charge us full restaurant prices as though they hadn’t radically reduced all of the other costs entailed. And then food delivery services exploit the process even further—these services take somewhere in the neighborhood of 25-30% off the top of every sale (like the app store for your phone). Although there was a great deal of talk about helping “rescue” the restaurant industry during the pandemic, it was debatable at best that that’s what these services actually did. (There’s an ironic edge to the latest Uber Eats campaign about how remembering that “Uber Eats delivers X” requires us to forget other things.)
Many of the traditional values I named above have been collapsed into “traffic,” in ways strikingly similar to how social media turned towards clickbait. Restaurant “concepts5” are algorithmically generated and marketed to us through search engines and grubchoweatsdash apps, we scroll through AI-powered images of instagrammable meals, and then ghost kitchens assemble approximations of these images whose ideal is not differing too far from the image lest they provoke demands for refunds. If this sounds vaguely dystopian when compared to the model of the home-cooked meal, that’s not accidental on my part. It’s not a scam per se, but it’s an impoverished approach to food service that exploits the value of restaurant quality food, simulates it as cheaply as possible, and profits off of our unwillingness to look closer at the entire system. Opening a restaurant in the best of times was never a simple or an easy proposition, but the ghost kitchen/delivery app ecosystem takes the resources that were once part of that process and channels them elsewhere.
Large Language Models are the Ghost Kitchens of Writing
And now we’ve arrived (finally!) to the point that I wanted to make all the way back at the beginning of this post. I don’t know that there’s an analogous straight line to be drawn from personal literacy to the large language models behind so-called AI, but I think that the comparison holds. In many ways, including the devaluation of the people involved and including profit optimization, LLMs are basically ghost kitchens for language.
These corporations have set machine learning to the task of converting as much of language itself over to epistemic knowledge, to techne, with the goal of “disrupting” all of the many professions that rely upon the local, tactical expertise of human beings. This process has already been underway for some time. Just as there are enough folks willing to trade quality for convenience when it comes to their meals, enough to support massive numbers of fast food restaurants, I expect that the trend of giving over writing jobs to custom LLMs, capable of churning out “good enough” prose by the ream, will continue to rip through all sorts of industries. I’m sure that I’ll be “fooled” by students who are more interested in a last-minute visit to UberWrites than they are in taking advantage of the education that their families are paying for.
Whether it’s seeing like a state, a corporation, or a data structure, it strikes me that maybe the most important point about these optimized, large-scale epistemologies is one that Scott makes with the help of 17th century French mathematician and philosopher Blaise Pascal. In the second-to-last chapter of Seeing like a State, Scott explains:
As Pascal wrote, the great failure of rationalism is “not its recognition of technical knowledge, but its failure to recognize any other.”
The problem with LLMs, ghost kitchens, and all of these focused, optimizing, tunnel vision attempts to engineer our lives isn’t that they don’t yield some benefit to us, it’s that they actively prevent us from seeing in any other way, or accounting for quality of life. Don’t forget: love doesn’t scale.
I think that’s about all I have to say about Scott’s work for the moment. I’ve already shared this book with multiple people, so I feel like my job is done. For the first time in at least a couple of weeks, I’m not sure what I’ll be writing about next, just that I’ll be writing. Stay tuned.
Scott also suggests that a healthy democracy functions this way.
I flirted seriously with the idea of attending culinary school in my 20s before ultimately deciding to return to graduate school to get my PhD.
Robin’s latest novel Moonbound just arrived in my mailbox this past week, and I had to stop myself from reading it all in one go, in case you’re wondering whether I liked it.
I wrote about ghost kitchens early in the life of my Substack, in a discussion about how our newspapers and magazines were undergoing a similar, private-equity-inspired “disruption.”
The idea of the “restaurant concept” has entered food competitions in a substantial way the past few years. Top Chef has always had their “Restaurant Wars” episode about halfway through the season, but the “concept,” which typically involves a name, theme, and 2-4 signature dishes, has become a pretty standard challenge. I feel about the concept the way that many writers, artists, musicians do about “content” as an umbrella term for their work, even if I’m not a food person myself.
A. I wish I'd have had (and been able to absorb) the Socratic dialogue herein (or that you'd been sitting in the next row of desks over) when I was in Classic Pol Phil lectures back in the last century; and,
B. More recently, when I see a line of cars and pickups in the 'drive-through' line at any mass food franchise (loved the more honest advert of the golden arches sign proffered), with their 'drivers' all, every man-jack one of them, 'engaged' with their little black mirror devices, whilst waiting, I think "I saw this all in a dystopian science fiction film ten (20 - 30) years ago", except there was more random violence - oh, wait...; But then,
C. I just listened to an hour-long discussion on Huxley's "Brave New World" by a brave Old-World Anglo-Irish writer who's coming to the same thing on the matter of 'scale' and life; and,
D. Can no longer take in fast casual (or fine) dining without remembering Terry Gilliam's sterling silver tray of greenish glop, delivered as 'elegant dining'; and,
E. I'm looking for a set of Collier's Encyclopedia, ca. ~ 1960 so as to have reference at hand that's more grounded in actual editorial debate and conclusion, prepared by knowledgeable people endeavoring to be responsible to their readers (and themselves).