The Divergence Machine

Introducing the 2026 Book Club theme

The Contraptions Book Club 2026 is underway. The theme is the divergence machine. I introduced the idea with this diagram a couple of weeks ago, in The Modernity Machine III, in relation to the modernity machine we explored last year.

In January, we are reading Voltaire’s Candide (Theo Cuffe translation recommended) and any related essays we can get our hands on. In February, we’ll read Henry Farrell’s The Underground Empire. In March, you’ll get to choose between a Leibniz-Spinoza book and an Adam-Smith-David-Hume book. Beyond March, the menu is still under construction. Our focal period this year is 1600-2000, with particular emphasis on the early and middle parts.

This essay is intended to set up the dev-environment for the book club so to speak, laying out some initial frames, themes and prejudices. The setup may seem a little elaborate, but our book club isn’t just one damn book after another. It’s more a cunningly contrived contraption designed to enable systematic study of an idea-space.

We’re not picking books because they’re necessarily “good” or fun to read (though I hope most will be both), but because they help assemble a view of the world from a certain opinionated perspective (largely mine, but shaped a lot more by others this year than last year).

There’s an overarching logic and vaulting conceit to what we read, why, and how that I’ll introduce in this essay.

For those who came in late

If you’re just joining our book club shennanigans, it runs on a grand theory. In our grand theory, we are concerned with what we call world machines: contraptions that embody the logic of how the entire world works for a period of time. Our book club studies these world machines through a selection of books from, and about, the construction period of each machine.

The modernity machine that we studied in 2025 was constructed 1200-1600 and operated at a steady plateau of capability 1600-2000. It is now undergoing rapid, partially scheduled disassembly. The divergence machine was constructed 1600-2000 and has been operating in fully deployed mode for about 25 years so far.

By our grand theory, at any given time, one world machine is in operation, another is under construction, and a third may be undergoing (usually rapid) decline/dismantlement/destruction (aka rapid, partially scheduled disassembly, to adapt a term of art from rocketry). So at any given time, you have to understand the logics of two, possibly three world machines in tension to understand how the world works.

The meta-logic is derived from the Gramsci Gap and the idea of worlds being born and dying, with monsters appearing in the passages between, though the mapping is not perfect (the “world being born” is actually two worlds — a completed new world entering full production mode, and the seeds of a future world being planted).

Only two world machines are illustrated above though. The declining medieval machine that enjoyed a plateau of stable operations 1200-1600 and collapsed rapidly after, is not shown. Neither is the as-yet-unnamed machine that is is beginning to be constructed today, and destined for full deployment well past our lifetimes (whose internal logic will likely be shaped by AI crossed with End of History conditions). Juggling two world machines in our heads is hard enough I think.

In our color-by-numbers potted history template, each world machine takes about 400 years to build and turn on, operates for another 400, and declines in another 100-200 years, making for a lifespan of 900-1000 years.

Historiographic Hygiene Rules

Don’t take these numbers or underlying machinic world models too seriously. They are intended more as mnemonic devices and intuition pumps than rigorous theories of history. The point of the models is to approach our book selection and reading with a particular sensibility, paying attention to particular themes and questions (in brief: those that center machinic phenomenology over either organic/“natural” or humanistic).

Historical time, of course, can and does dilate or contract according to both the raw pace of historical events, and the ontological status of “history” as such, and whether or not one can meaningfully posit the existence of a gestalt world process even deserving of the name. The Paleolithic world machine probably had a lifespan of 100,000+ years. The Neolithic world machine had a lifespan of at least 15,000 years. The Bronze Age world machine probably lasted about 4000 years. Over the really longue duree, world-machine lifespans appear to be falling.

But even more importantly, world machine temporalities appear to be — complicating. Later historical machines feature more complex multi-temporalities, suffused with more kinds of atemporalities in their interstices, than earlier ones. To quote the tenth Doctor Who, they are bigger balls of “wibbly wobbly, timey wimey stuff.” The temporal vorticity of history increases with chronological time, and is systematically higher with each new world machine. This is one reason the future always looks more chaotic from perspectives rooted in the past.

Beyond helping book-keep cause and effect, and correlate human events with natural ones, raw chronological time is not as useful for studying history as you might think. Periodizing into “world machine” epochs, while it might seem like a caricaturing move, is surprisingly helpful in organizing our understandings.

This is why there is no problem with continuing to study history after the end of history. Fukuyama’s End of History argument, a staple of my thinking and this book club, rests on a particular ontological of history (his intellectual ancestor, Alexander Kojeve, is on the shortlist for our book club). It is possible to study history in “end of history” conditions, making use of Fukuyama’s models, without being exclusively committed to his ontology of history.

Thanks to the acceleration of history and the shortening of world-machine lifespans, the unnamed AI-powered world machine now being seeded might very well have a lifespan of only 300 years total perhaps, rather than 900-1000, and cut short the reign of the divergence machine now entering production. Given that it accelerates the rerun-heavy nature of post-End-of-History temporalities, those 300 years might feel a great deal more atemporal. And a lot more wibbly-wobbly, timey-wimey too.

Keep these hygiene notes in mind as you read.

Divergence in Plan View

The first graph with its intersecting S-curve portrait might tempt you into a too-reductive view of the succession of world machines as some sort of “disruption” process. Resist the temptation. World machines embody richer phenomenology.

Here is another view of both machines, this time in historical plan view. Treat this as your main conceptual map for the book club this year.

On the y-axis, instead of machine complexity portrayed as a classic pair of S-curves, with one displacing the other, we see the fundamental dynamic of each machine, measured by span of non-canoncity.

The modernity machine manufactures and projects convergent canonicity across the planet. You can think of this as a kind of homogenizing, legibilizing fiat reality imposed from a few sources of centralized power (James Scott Seeing Like a State is not on the menu for the book club, but would be a useful prerequisite read if you haven’t already read it — most people I am expecting will join the book club have either read it or good summaries of it).

Those who don’t participate in the exercise of that centralized power are defined by their responses to it. But this doesn’t make those responses “divergent.”

Principle: In general, responses to convergent canonicity tend to become part of that convergent canonicity.

Ie, responses to the modernity machine should be considered part of that machine, because they’re attentionally and informationally entangled with it.

In the path-dependent history that played out on this planet, convergent canonicity can be allegorically represented as a braid of strands winding around a European history centerline.

Excursions away from that centerline, particularly those that stray beyond a recognizably European band, or even play out entirely outside it, are a measure of non-canonicity. These excursions are anomalous from the perspective of the modernity machine; its bugspace so to speak. But non-canonicity, the bugspace of the modernity machine, is the feature space of the divergence machine.

The divergence machine manufactures, or rather, spawns variety. Specifically, variety in a state of expansion and mutual retreat, something like the cosmological red shift playing out in civilizational configuration space.

The divergence machine doesn’t contest political space organized by the modernity machine, based on principles of canonicity. In fact it rests on and relies on, what the modernity machine has already done, without attempting to either perpetuate it or destroy it.

Principle: The divergence machine is not an anti-modernity machine, or in any other particular systematic relationship to it; it has its own internal logic not derived from that of the modernity machine. It assumes the persistence of some of the major historical effects of the modernity machine, but not the perpetuation of the machine itself.

Specifically, the divergence machine doesn’t try to subvert the modernity machine critically (the way “postmodernity” did). Rather, it renders canonicity irrelevant, by creating and organizing civilizational space that is past the reach of the logic of the modernity machine entirely.

In divergence-machine regimes (which, remember, are only 25 years old in their complete form, even if they have been 400 years in the making) the space previously organized by convergent canonicity (the “center” so to speak) or responses to it (which orient towards that “center”), gets phenomenologically bracketed. This is not an intellectual move but a natural and emergent side-effect of divergent historical processes. It is not some sort of partially-self-conscious decentering theorized and narrated by a tribe of intellectuals on a mission.

I’m still working out the nuances of my account of the world after modernity, and how it resembles or differs from accounts that rejoice under names like postmodernity, late modernity, and metamodernity, but it is important to have rough-and-ready understandings of these idea-spaces as we read, because a lot of what we will read will have elements of all these world processes and associated intellectual currents.

So it will be useful to ask questions like is this book late modern, postmodern, metamodern, or divergentist?

Here’s my current cheat sheet. Some of this may change/evolve.

Late Modernity

Late modernity is localized, slowly unraveling, zombie persistence of the modernity machine, including both natural persistence, and conscious political projects to perpetuate it while that’s still an option, or restore it to a pristine state once it is clearly entering a state of dereliction.

There is not much more to be said about late modernity (the associated intellectual currents are fairly weak — Zygmunt Bauman and a few others come to mind), but there is a lot to see of it. Almost everything you see around you is late modern. Almost everything in the news headlines is late modern. We won’t see the last of late modernity in our lifetimes.

Arguably, the bulk of the energy of the world will continue to flow through late-modern pathways for at least our lifetimes, even if very little of the evolutionary intelligence of the world flows through those pathways (hence “zombie” or perhaps “energetic zombie” would be a better term, like the ones in the Korean movie, Train to Busan).

China as it is being imagined and conjured by the CCP today, if not its people, is primarily a late modern world process. But China, the larger, more nebulous civilizational unit and world sub-process, is a more complex beast, not reducible to late-modern dynamics and phenomenology being supervised by the CCP.

Late modernity is not a focus for us, but I liked Adam Curtis’ fevered account of it as “hypernormalcy” in his 2016 documentary. All the energy of the Trumpist planetary turn, right up to the Maduro kidnapping/rendition this weekend, is late-modernity in action. Trumpism and related reactionary turns worldwide are obviously late modern, at least doctrinally, even if they are tactically more open (including borrowing from postmodernity and meta-modernity, which we discus next) in their attempts to shape the fate of the world.

We will not read much about late modernity, or make much use of late-modern perspectives (our readings from last year, suitably extrapolated, should be sufficient to make late modern dynamics sufficiently intelligible for our purposes), but we’ll keep our eyes open for late-modern dynamics.

Our go-to move as we read will be to try and artificially eclipse out the bright dying star of modernity that is late modernism, so we can look at more interesting things. To the extent our book club takes note of headline-grade current events, we will make use of late-modern perspectives to make sense of them, but look for divergence-machine metabolic processes that also respond to them in more long-term consequential ways.

Postmodernity

Postmodernity is both a class of world processes that fall short of defining a world-machine proper (certain strains of post-colonial nationalism for instance), and an attempt to theorize and construct the world entirely in terms of adversarial responses to modernity. As such, postmodernity encompasses both natural modes of alterity (a postmodern term of art) and intellectual-political projects that attempt to make those modes legible and rugged, while simultaneously making the dynamics of the modernity machine fragile and vulnerable to attack.

For our book club, we will use “postmodern” in the broad, loose sense used in popular discourse, covering everything from the original french theorists to more recent American Marxist flavors. We will also use the term for historical processes (such as say independence movements, language/culture revitalization movements, various feminisms and late-stage manifestations in the environmental and social justice movements) that are usefully described by postmodern intellectual perspectives primarily because they constitute themselves with reference to those perspectives. It is worth noting that postmodern intellectual perspectives have been much more constitutive of world-processes in the last century than late modern ones, which have largely emerged as post-hoc narratives of decline processes already underway.

Postmodernity features proportions of energy and intelligence that are the opposite of those exhibited by late modernity. It is an intelligent ghost rather than an energetic zombie. It organizes very little of the energy of world processes today (though it lays claim to a great deal through energetic labeling and map-making), but the intellectual currents associated with it are still extraordinarily strong, decades past their peak. To the point that if you presume to think about world processes at all without using their preferred terms of reference (or worse, using and abusing them partially where useful), they will send representatives to knock on your door in the dead of the night and lecture you.

Unlike divergence, postmodernity does contest the civilizational space organized by modernity (it can conceive of no other), and is therefore in a zero/negative-sum relationship to late modernity. Postmodernity tries to override the logics that late modernity tries to perpetuate.

Principle: The postmodern project, I believe, is essentially complete and has been largely successful as an analytical and political project.

There is much to be learned by studying its discoveries and history. But there’s not much point to continuing the postmodern project, either through constituting world processes by its logic, or keeping the associated intellectual currents going. They’ve made all the discoveries they are going to. The paradigm is exhausted.

Equally, to the extent there are things to dislike about postmodernity, there is no point fighting it, as late modernity likes to do, because on most consequential matters where postmodernity pursued clear objectives, it has already won in ways that cannot really be undone or reversed.

One entailment of this position: the still-ongoing battle between reactionary politics and wokism is something like a cage match between an energetic zombie and an intelligent ghost in the mental model we’ll be adopting for the book club.

To the extent the concerns of either side remain live and consequential ones, we will look elsewhere for meaningful phenomenology to think about. Our assumption will be: If the divergence machine “solves” for social justice or environmental stewardship for example, we should not expect the mechanics to look anything like the ones postmodernity as a constitutive force briefly powered. Equally, if the divergence machine “solves” for some recognizable continuation of things like ethnic or racial identity and nation-state-based culture and traditions, it will look nothing like the solutions of the modernity machine in its late-modern perpetuations and life-extensions.

Our go-to move will be to treat the current war between Late Modernism and Postmodernism as noise to be filtered out as we attempt to decipher the workings of the divergence machine. We might retain problems posed by those perspectives, but likely not any proposed solutions. We will treat both as spent perspectives, as far as their creative constitutive capacities go.

Our attitude towards the intellectual legacy of postmodernity (in the narrow sense of a set of twentieth-century intellectual currents) will be cannibalistic: An occasionally useful source of frames and terms, and a historically consequential set of world processes through part of the twentieth century. Think raw material, not authority or influence. Ghosts can’t defend their corpses after all, or they wouldn’t be ghosts.

Metamodernity

Metamodernity could perhaps be clubbed with Late Modernity, but it is useful to keep it distinct.

I define it as attempts to resurrect patterns of modernity in piecemeal forms that might be viable for contemporary circumstances. (A friend of mine, Rob Knight, evocatively called it “modernism in drag”).

This project, I believe, is ill-conceived, unnecessary, and doomed. I react poorly to things with “meta” in their name and I intend to impose this prejudice on the book club and its activities :)

Perhaps the most important element of metamodernity is world processes and intellectual currents that can be understood as responses to unrecoverable localized psyche failures of modernity. Metamodernity is what you dream up when modernity fails completely enough that there is nothing to revivify or perpetuate, but you still want what it used to deliver reliably, especially inside your head. You go meta when it is too late to be merely late.

Continuing our monstrous taxonomy, if late modernism is an energetic zombie, and postmodernity is an intelligent ghost, metamodernity is ennervated necromancy.

The metamodern project is not reactionary (for a long time, I was convinced it was, but I’ve changed my mind). Metamodernists are typically neither deluded enough, nor chauvinistic enough, nor have enough raw material to work with, to be reactionary about the things they care about.

The so-called “meaning crisis” and ideas like “re-enchantment” could be classified as metamodern turns in intellectual currents. Sincerity, authenticity, and irony are particular concerns of metamodernity, but unlike late modernity, metamodernity does not treat irony as the evil manufacture of postmodernity, constituting a casus belli for culture warring. Rather, in metamodern accounts, irony appears as an emergent consequence of historical processes, resulting in a set of problems to be solved, rather than a set of crimes to be prosecuted.

So metamodern responses take the form of rather doleful adaptation and sentimental creativity, rather than culture warring. The so-called “trad turn” strikes me as more metamodern than neoreactionary (but I’m not attached to this reading). Philosophers in the neo-Heideggerian tradition like Byung-Chul Han strike me as pursuing metamodern projects.

So far, metamodernity exists only as a few weak intellectual currents and perhaps a few attempts at post-ironic art, especially in screen media. There are no meaningful world processes I would classify as metamodern. There are no high-energy phenomena like ethnonationalist political movements, or DEI-ESG wokeism, that we can associate with metamodernism. So far, metamodernists seem to have contented themselves with writing and making art with rather gloomy, tortured gravitas.

There is however, a negative space we can attach to metamodernity that helps define it — the class of ennervation phenomena generally referred to as involution. This spans hikkikomori, “laying flat,” “quiet quitting” and so on. Territories defined by the failure to meaningfully address what only the maps of metamodernity even attempt to organize and attend to.

Our go-to move in relation to metamodernism will be to pay attention to the negative spaces it points to, without doing anything about them.

Rewinding 400 Years

I’ve written elsewhere about my philosophy of Divergentism, which has informed, but only partially determined what I’m calling the divergence machine and the agenda of the 2026 book club.

Our go-to move in relation to divergence is to assume everything important about it started as much as 400 years ago.

If something we read looks like it belongs in an account of the divergence machine, we will try the following moves:

1. Check if it can be traced back to seeds planted ~1600 or thereabouts

2. Test to see if it’s better understood as an zombie late-modern thing

3. Test to see if it’s better understood as an intelligent ghost of postmodernity

4. Test to see if it’s better understood as ongoing metamodern ennervated necromancy

Phenomenology that gets past this four-stage filter of negative definition can then be explored with “divergence” questions and probes, such as:

• Is there plurality in the mechanics of whatever is happening?

• Does it involve people understanding each other less, but getting along better?

• Does it smell like Darwinian evolution?

• Does it relativize or bracket things that seem canonical?

• Is there generative variety emerging from it?

• Are there elements of absurdity or humor to it?

• And perhaps most importantly, is it alive?

Our goal will be to look for, analyze, and place machinic constructions on, things that have the energy intensity of late modernity without significant convergent canonicity in associated intelligence processes; the irony, obliquity, and indirection of postmodernity without its negative-sum attention-entanglements with late modernity; and finally the expansive and reality-based concerns of meta-modernity without its interiority or fundamentally gloomy suffering-and-healing sensibilities.

Above all, our true-north question will be does this embody new forms of liveness being newly and generatively turned on in the world?

Our exploration of the divergence machine is a search for emerging liveness, so to speak.

Let’s have at it. Post your thoughts on Candide in this chat thread.