Last month, in the book club, we read Inventing Nature by Andrea Wulf, about the life and work of Alexander von Humboldt. This month, I ended up also reading Wulf’s earlier book on the Jena set, Magnificent Rebels, which is on our side quests list. Alexander von Humboldt, along with his brother William, were both part of this set, though the former was arguably on the margins of it rather than the core, in part because he was gallivanting around South America during the crucial period, and in part because he was not humanist-reactionary enough to belong. The Jena set arguably invented the modern (essentialized and rather narcissistic) idea of “human.”
This month’s main pick was Revolution in Time by David Landes, which I’ve owned for 15 years (bought and scanned when I was writing Tempo) but hadn’t actually read until this month. I’m almost done with it and now wish I’d read it earlier. Evolution in time-keeping through the period we’re studying right now (1600-2000) is a critical subplot but really hard to appreciate in conventional accounts of it.
I’m just starting to read the June pick, The Business of Enlightenment by Robert Darnton, which covers the diffusion of Enlightenment ideas after 1770, through the medium of the later editions of Denis Diderot’s Encyclopédie (the first edition of which was completed between 1749-1772), which was as epochal an event in publishing history as in intellectual history.
The ideas we’re juggling in World Machine theory are starting to get quite complex, so I’m overdue for some synthesis/integration effort. This essay is a trial assembly of the gear-shift mechanism between the Modernity Machine and the Divergence Machine. It probably won’t make a whole lot of sense if you’re coming in cold to this series. I recommend catching up by browsing my previous World Machines writings, or better yet, pointing your LLM at them, and getting tldr-ed up.
NOTE: I’m writing this essay as much for the in-development AI agent at the World Machines project (WMP), as for the human readers of this newsletter. Both the WMP and this book club are now being hosted by the SIGPSY group (Special Interest Group in Psychohistory; no we’re not kidding) that has just kicked off in the Protocol Institute discord. Future book club chats will be held in the group’s #psychohistory channel of the Discord — details and invite link on the book club page.
A complicated but elegant picture is taking shape now, of how the Modernity Machine began giving way to the Divergence Machine through a full-stack set of revolutions, from rarefied and intellectual to bloody and violent, which drove the gear shift in the political, cultural, and economic infrastructures of the world, starting with Europe.
Enlightenment: From Idea to Infrastructure
The Darnton book, which might otherwise seem like a very oddly specialized and nerdy pick for our book club, is interesting precisely because it helps complete a picture of the gearshift dynamics in our world machines theory.
The book is not about the ideas of the Enlightenment itself (talk to ChatGPT about that if you’re participating in the book club), or even about the Encyclopédie itself, which was a late-stage synthesis of Enlightenment thinking. It is about the structural diffusion of Enlightenment thinking through the social fabric, transforming it from a subculture of marginal heretical ideas to civilizational infrastructure, through the best technological medium available at the time — print. The “installation” of the Encyclopédie completed the Modernity Machine, right on the eve of its obsolescence, and the beginning of its replacement by the Divergence Machine.
The story of Diderot and d’Alembert’s Encyclopédie was something like a second-order sequel to the first-order installation of print culture in the 15th century (which we read about last year in the The Printing Revolution in Early Modern Europe by Elizabeth Eisenstein). It was also very much like the installation of internet culture in our own time. As an encyclopedia, Diderot’s was an ancestor of Wikipedia, and like it, an expression of an infrastructural maturation, not just of an intellectual milieu.
The relationship of the publishing ecology around the Encyclopédie to the big names of the Enlightenment, like Newton, Bacon, and Locke, was something like the relationship of the internet in our time to names like Vannevar Bush and J. C. R. Licklider. The Encyclopedists, as the group contributing to, and publishing it came to be known, were something like the first wave of internet entrepreneurs in our time.
The Darnton book also puts the other history we’ve been exploring into perspective — the work of David Hume and Adam Smith (which coincided with the work of the Encyclopedists), Voltaire’s role as a thought leader (he was directly associated with the Encyclopedists), and the subtle influence of changing temporalities being driven by the maturation of time-keeping technology through the era.
I want to try and connect all these threads of development and paint a rough picture of how the transition between the Modernity Machine and Divergence Machines actually happened.
Let’s start with a timeline. It’s easy to get very confused by the complexity of various streams of events (I briefly badly confused myself by mixing up Roger Bacon (13th century) and Francis Bacon (17th century).
The Timeline of the Shift
Here is a rough view of the timeline, which is something of a Doctor Who style ball of wibbly-wobbly, timey-wimey stuff.
1600: The intellectual phase of the Modernity Machine essentially ended around 1600, with the burning of Giordano Bruno at the stake in 1600 serving as a useful and macabre marker. Galileo, tried in 1633, is something of a transitional figure, playing a role in both WMs. So roughly between 1600 and 1640, the Modernity Machine entered production mode as completed infrastructure, and the seeds of the Divergence Machine were planted. It is worth noting that despite the name. the MM was firmly traditionalist, in the sense of being an operating system designed by and for the traditional ruling classes, monarchs, and religious authorities. The arrival of the MM was also a convergence to a kind of civilizational-infrastructural consensus that Europe was just starting to export to the rest of the world.
1620-1690: The ideas of the Enlightenment, in the form synthesized later by the Encyclopedists, took shape roughly between 1620 and 1690. Three works are foundational: Newton’s Principia (1687), Locke’s Essay Concerning Human Understanding (1690) and Francis Bacon’s Novum Organum (1620). It is worth noting that this period also corresponds to the early settlement of what would become the United States, which had already begun to shape the psyche of Europe (starting with tobacco, ending with revolutionary catalysis).
1637-77 (Descartes and Spinoza): Two works have more complex relationships with the Enlightenment. Descartes’ Discourse on the Method (the “cogito ergo sum” book), 1637, was a prequel that the Enlightenment built on but superseded, while Spinoza’s Ethics (1677) was too radical for the Enlightenment proper to absorb, but sort of haunted it like a scary ghost in the Enlightenment infrastructure. Descartes doesn’t go far enough to be part of the Enlightenment, while Spinoza went too far. Leibniz appears in this sideshow tent of related figures too, but as marginal rather than structurally relevant, and something of a lolcow, thanks to Voltaire’s Pangloss parody. He does briefly re-emerge into relevance a couple of centuries later via Mach, Bergson etc. Interestingly, Leibniz has suffered a devaluation in status, similar to Bruno, through the reframings of our book club. But unlike Bruno, who I now think of as a largely irredeemable crackpot, Leibniz still retains critical value in the mathematics and computing storyline, if not in the philosophy storyline.
1749-1789: The Enlightenment, as an institution, as opposed to a set of abstract ideas, was essentially an institutional compromise between radical and traditional thought brokered by the Encyclopédie in the decade before the French Revolution; between objectivity (Newton), empiricism (Bacon), and a natural conception of self (Locke) on the one hand, and ecclesiastical authority, divine monarchial authority, and the individual self as a sort of expression of the will of the Christian God. So the Enlightenment represented a cautious and pragmatic rupture from tradition that had just enough institutional support, in an era where it was struggling to survive. The Encyclopédie threaded that needle, through a mix of covert and ironic subversion and some compromise. It survived through its first edition years despite (somewhat nominal) official censorship, but escaped Inquisition grade active suppression/elimination efforts. Too many people in the establishment were sympathetic to the Encyclopedists for it to be seriously suppressed. But after 1770 and up to the French Revolution, it basically installed the Enlightenment as institutionalized social reality.
1789-1799: The French Revolution, which to some extent drew inspiration from the American Revolution (which was culturally simpler, even if in other ways more profoundly consequential), marked the transition to the post-Enlightenment era. Immanuel Kant was the hinge figure (I’ve picked up this use of the word hinge from ChatGPT — delving into AI is good for your vocabulary), attempting to synthesize empiricism and idealism, subjective and objective, and personal and religious notions of self. A Critique of Pure Reason (1781) appears after the Encyclopedists, but before the Romantics.
1790-1807: Following this arc of Enlightenment, from ideas to institutionalization (pirates to navy?), German romanticism appears in some ways as a reactionary cultural movement that reacted to the decentering of the human effected by the Enlightenment with what we could call Humanism 1.0. The official position of this newsletter is that all humanisms are reactionary. Some are just confused and call themselves progressive, a pattern that started in Jena. I’d heard of some of the key figures (Goethe, Alexander von Humboldt, Schiller) but not of others (Fichte, the Schlegel brothers and their wives, Schelling, Humboldt’s brother William). It was an oddly schizophrenic movement that seemed to believe that the Enlightenment simultaneously went too far, and not far enough. Schizophrenic, but consequential. German romanticism created the romantic idea of the self that is still the default idea we’re enculturated into, around the world, by the liberal middle class.
1890s, 1910s, 2020s (Humanist spasms): Jena romanticism was a short-lived but intense phenomenon — just a decade or so, coinciding with the rise and fall of Napoleon (the romantics broadly supported both the French Revolution and Napoleon, which is sort of revealing in the same way people pivoting from Bernie to Trump is revealing). I think this is characteristic of humanist spasms between major technologically determined world machine eras. when humanist delusions of agency and significance are at a peak, along with anxieties about potential terminal insignificance. We see similar dynamics around the Bloomsbury group in the 1910-30 period (ironically associated with “modernist” literature). And we’re witnessing a similar period now, in anxious efforts to reclaim a human center for an AI age. The Pope’s recent encyclical on AI is notable more for clearly flagging the nature of humanist tendencies in any era than for things it says about AI. Modern trads, Progressive anti-AI types, Singularitarian AI doomers, AGI theologists, metamodernists, re-enchantment types, and the Catholic Church all share a loose humanism comprising a variety of flavors of neo-romanticism. Which to first order is just techlash+poignant poetry.
1848-89: The period of the Encyclopédie’s brief reign as the high-water-mark of civilization (roughly 1770-1789) is uncannily like the reign of the the early internet era, (roughly 1969-1993) and the neoliberal ideological tendency that accompanied it. Both were terminated by seismic geopolitical events (the American and French revolutions; the end of the Cold War and 9/11) and followed by a second wave of smaller revolutions (the revolutions of 1848, known as the springtime of nations, and the Arab Spring through Trumpism in our time). Modern nation-states may have been conceptually started with the Peace of Westphalia, but became a practical reality starting around 1848 (there are multiple books about this year; one is in our side quests list — Revolutionary Spring by Christopher Clark).
Trump as a Farcical Napoleon: There are uncanny but twisted similarities between the careers and historical roles of Napoleon and Trump, a case of history repeating itself, first as tragedy then as farce. Both had similar relationships with the prevailing revolutionary tendencies in their times, and similarly weird relationships with cultural elites. Curious learning: Napoleon was apparently much more attached to a self-image as a scholar than as a conquering general and emperor. He traveled with a personal librarian and campaign library while on the warpath and signed documents with his title as a memory of the French Academy of Sciences. I’d really like to read a comparative biography in about a decade.
End of History, 1806 vs. 1991: Hegel marks the completion of the philosophical transition away from the Enlightenment to the post-Enlightenment era, ending the brief reign of the Romantics. His is a complex legacy. While on the one hand he replaced the Enlightenment’s universalist pretensions with a historically contingent (and therefore structurally divergent) understanding of reality, the specific understanding he argued for was teleologically convergent towards an “end of history.” That’s always been one of my favorite ideas, in the form that emerged in our time, via Kojeve and Fukuyama, but I’ve always wondered why Hegel himself proclaimed the end to have occured at the Battle of Jena in 1806, when Napoleon steamrolled through Prussia via Jena. That always seemed oddly arbitrary to me. But now, in the context of Jena romanticism, it is somewhat clearer, and I realize I was unfairly thinking of Hegel as a small-minded creature of his own times. Hegel briefly overlapped with the Jena set in Jena, and had to leave in a hurry when Napoleon invaded (just barely saving the only draft of Phenomenology of Self — weird to think of a time when making backups was actually hard and losing valuable work was not attributable to sheer carelessness). But his choice of 1806 is at least as defensible as Fukuyama’s choice of 1991 (which I think is actually the correct date implied by the model).
The Gear Shift
According to World Machine theory, the Divergence Machine began to emerge around 1600, and was completed and put into production in 2000. So the 1750/1850 period is likely where the S-curves cross, so to speak; the rising curve of the DM intersecting the plateau of the MM and begining to disrupt it. Viewed in this light, the events in that period lend themselves to a specific interpretation.
First, the Enlightenment was divergent in content, but convergent in intent. The intellectual content was pluralist, as suggested by the fact that it took an encyclopedia to synthesize it, rather than a single authoritative interpretation. Its natural tendency was to spark a sort of Cambrian explosion of divergent thought, which did in fact happen, in the form of Romanticism and in the historicist-contingent Hegelian eras that followed. But on its own terms, the Enlightenment was convergent. It attempted to construct a monolithic understanding of the world and the place of humans within it, to directly compete with the similarly monolithic understandings of received tradition. By this account, we can think of the Enlightenment as a late-stage infrastructure project of the Modernity Machine. The Toyota Prius phase between IC and EV automobiles, so to speak.
But centrifugal forces overcame centripetal ones, and it was the post-Kant inheritors of the legacy of the Enlightenment who actually ported its logic to it’s natural home in the Divergence Machine. The idea of the self inaugurated by John Locke was taken to its natural conclusion by Fichte, who laid the foundations for thinkers like Freud who came a century later. The logic of the universe as first perceived by Newton, which led to a reductionist understanding of it, was engineered into the logic of divergence by Humboldt, who foreshadowed Darwin’s completion of the task of conceptualizing nature in divergentist terms.
Divergence dynamics fundamentally yield to, rather than resist, centrifugal forces, allowing the monolithic to give way to the pluralistic; objective consensus to subjective dissensus; and perhaps most importantly, the synchronized to the asynchronous.
This last is the counterintuitive lesson of the evolution of time-keeping: Clocks drove divergence as they improved, not convergence.
Clocks and Asynchronicity
Technology is generally not considered part of the Enlightenment story, which is generally considered a story about science and philosophy. But it should be part of the story. Particularly a technology that was the computing of its time — time-keeping.
The most significant developments in time-keeping unfolded over exactly the same period that the events on our timeline unfold. Galileo’s pendulum discovery around 1637 begins the story, and John Harrison’s H4 chronometer, which finally claimed the Longitude prize in 1761, concludes it. Over that long century, clocks grew smaller, cheaper, and far more accurate. Accurate enough to help disrupt one world machine and power its successor.
A naive view of the history of the clock is that it led to convergence and synchronization of civilization. As it turns out, this is the opposite of the actual story. I’ve been sort of clumsily reconstructing the actual story since around 2018 (when I gave a talk about it), and I wished I’d actually read Landes earlier, because it makes the story clear, and I didn’t have to reinvent the wheel.
The big lesson of the book is that between the 13th century, when large mechanical clocks began to be built, and our era, when we finally shed our quartz wristwatches in favor of ubiquitous GPS-driven time displays on all our screens, two sets of changes unfolded in tension with each other: Time-keeping simultaneously got more precise (due to fundamental scientific-technical advances) and more decentralized (due to becoming smaller and cheaper, via a Moore’s Law type dynamic).
To put it crudely, in the Modernity Machine, time was inaccurate and centralized, under the authoritarian control of the owners and keepers of monumental water clocks and mechanical turret clocks in clock towers of the 14th century. For 300 years there was a steady but mostly futile push towards both accuracy and decentralization. Small, personal-scale mechanical timepieces (comparable to modern wristwatches) were being made as early as the 14th century. The problem was, though they were very clever mechanically, they were extremely inaccurate compared to larger clocks, which were themselves pretty bad and had to be constantly reset to match solar time. At the smallest scale, the value of mechanical clocks lay more in their ability to drive complicated clockwork toys (popular with nobility around the world) than tell time.
The 17th century changed that. Galileo’s pendulum made large clocks radically more accurate, and the development of the balance spring made small, personal scale clocks and watches more accurate than the clock towers of previous centuries. Externally imposed (by monarchs and priests) time authority gave way to internally maintained time discipline. External locus of control gave way to internal locus of control. The modern self was born, with an internally clocked psyche.
Basically clocks grew far more decentralized than they grew usefully accurate (beyond a point, accuracy gains had low marginal value for pre-digital humans), and drove devolution of control over time to the smallest scales. You could now organize your personal life by your personal watch, and gain all the benefits of accurate time-keeping, without subjecting yourself to time-keeping authority. You could coordinate with personal friends and networks without relying on centralized time.
Fichte’s Ich philosophy could not have been conceived without the personalization of time. The French Revolution was arguably in part a response to the pressures created by disruptive time-keeping technologies.
This story largely played out over precisely the period that our revolutionary tale and the gear shift from MM to DM happened.
Here is one way to cash out the difference: The MM ran on centrally controlled turret clocks, the DM ran on personal-scale spring-driven watches and clocks. It was a shift comparable to the evolution from mainframe computing to iPhones, except unfolding over a century instead of half a century, and preceded by 400 years of “mainframe clock” time instead of 20 years.
The development of a usable marine chronometer allowed planetary integration to finally go from dangerous exploratory activity to routine infrastructural activity. In a way, the chronometer did to the 19th century what AI is doing to our time. A fun learning from the Landes book — John Harrison gets the credit for winning the Longitude prize, but his clock was the equivalent of IBM’s Watson AI winning Jeopardy and Deep Blue beating Kasporov — impressive and technically a legal solution to the underlying challenge, but fundamentally a dead-end and not the path technical evolution actually took later.
Marine chronometry in the form that actually powered the colonial globalization era developed from a parallel and more practical and divergent French tradition that got transplanted to England, and was arguably also the genesis of interchangeable parts manufacturing. The French tradition emphasized robust and simple designs that could be easily copied and manufactured along industrial lines, and not coincidentally, France of the same era was also the point of origin of the Système Gribeauval which eventually influenced and found its fullest expression in the American system of interchangeable parts manufacturing (see my old blog post on Hall’s Law). I haven’t yet traced the direct connection between the Système Gribeauval and the chronometry story, but I’m convinced it’s there to be found. Both also curiously foreshadow the worse-is-better principle in computer programming from our era.
That’s just a taste. There’s a lot more insight to be found in the history of time-keeping for the future of computing and AI.
I want to conclude with a broader point. The mature clock, at cheap-and-accurate wrist-watch level, was a pure divergence driver, it desynchronized civilization that had previously been kept inefficiently synchronized by large turret clocks calibrated to solar time.
The clock is also divergent in a deeper way, as a new class of artifact that sustained seemingly endless variety. The technology of mechanical clocks existed in a dizzyingly pluralistic and varied design space of dozens of different types of escapements, hundreds of clever mechanical engineering tricks, and astoundingly complex mechanism powered mathematical calculations. An early genre of clocks was “equation” clocks, designed to keep clock time synchronized with Sun time. By the 18th century, mechanical clocks had gotten too accurate to be calibrated by the Sun, and could be used to actually track and measure variations in solar time. But since tradition (and inaccurate old clocks) were bound to solar time, for a transitional period, people needed to translate. Hence equation clocks to translate. Eventually, solar time was abandoned and mechanical clock time became the standard. Before then, clocks showed varying day/night hours to match a “day” defined by sunrise-to-sunset rather than a fixed 12 hours. After, sunrise and sunset times were allowed to vary on the mechanical clock.
Clocks then, weren’t just like computers in our time. They were computers. Rigidly specialized mechanical computers by our standards, but radically flexible and programmable by the standards of 18th century technology’s familiar technologies like swords or cannon. The clock was the first technology that could compute, be “programmed,” and inventively embodied by a dizzying and growing array of specific designs (which should be analogized to software rather than computer hardware). Designs that could not just keep time and translate among times, but also drive a near-biological ecology of clockwork devices. Steampunk is less about steam power than clockwork mechanical governance of devices.
This topic obviously bleeds into my book project (which I’ve refactored significantly and will be doing an update on soon), so I’ll save more thoughts for that.
But the tldr of this preliminary synthesis is that the cutover from the Modernity Machine to the Divergence Machine happened somewhere in middle century of 1750-1850, culturally marked by the culmination of the Enlightenment project, and the beginning of divergent post-Enlightenment projects that inherited its divergent soul. This transition was marked by revolutions at all levels from bloody to bloodless.
Gold medal for the second paragraph in “The Gear Shift”