David Landes’ classic account of the Industrial Revolution, The Unbound Prometheus, is one of those books that makes you want to give up writing. It’s unimprovable.
From the start to the end, it’s fantastically generative—almost every page triggers a new idea, particularly in today’s context of anxiety about runaway technological progress.
For Landes, an ‘industrial revolution’ is “that complex of technological innovations which, by substituting machines for human skill and inanimate power for human and animal force, brings about a shift from handicraft to manufacture and, so doing, gives birth to a modern economy.”
This is the lowercase form of the phrase, applicable not just to the original Industrial Revolution—the one that took hold in late 18th century England, shifting the economy from agrarian handicraft to industrial machine manufacturing—but to any such process, wherever and whenever it takes place.
This abstract form of the noun immediately raises a challenge: what criteria should we use to identify an industrial revolution? If we use the term too loosely, we will end up finding industrial revolutions everywhere we look, and ultimately we’ll split the whole of history into a series of sequential revolutions, diluting the whole idea.
If we use the term too tightly, on the other hand, we’ll end up saying there’s only ever been one such event, the capitalised Industrial Revolution, and then we’ll have lost the analytical usefulness of the term.
This debate might seem pedantic but it’s resurfaced in the last few years with the publication of Brynjolfsson and McAfee’s book The Second Machine Age. Economic historians like Carlota Perez have argued that Brynjolfsson and McAfee are too tight in their approach, leading them into the mistaken view that the information revolution is only the second such event to have taken place in modern times.
As I’ve written before, I find Perez’s view persuasive. It seems to me that books like The Second Machine commit a kind of present-exceptionalism, the perennially alluring idea that we’re living through extraordinary times.
I call TSM’s particular version of present-exceptionalism the ‘sliced bread fallacy’, namely, the idea that some contemporary phenomenon is the most extraordinary such phenomenon of all-but-one such phenomena in human history.
Claims of this type are common. Just think how familiar it is to read a statement with the form ‘X is the most Y since Z’, in which Z is the most Y thing in history. The temptation in these statements lies, it seems to me, in that they combine present-exceptionalism with a certain false modesty; the writer avoids the childish enthusiasm of claiming something is the most extraordinary thing ever, but they do so by claiming it’s the second most extraordinary thing.
When it comes to historical analysis, falling prey to this temptation doesn’t just lead to untruth, it also leads to an unnecessary loss of insight because your historical sample is always one. Perez makes this point well in her analysis of Brynjolfsson and McAfee, when she argues that we’re living through the fourth or fifth technological revolution since the first such technological revolution, the rise of steam power in England around 1770. This might all seem like an academic distraction but it matters for this reason: the more accurately we can situate our current moment in economic history, the more successfully we will deal with the challenges it presents.
My own preference is to be more precise in our use of the term industrial revolution.
It seems to me that, properly defined, there has only ever been one true Industrial Revolution and there’s a good case for seeing this as the most consequential moment in human history—unique in the genuine sense that it started the dynamic of self-sustaining progress itself (more on this later).
The digital revolution clearly does not tick this box—it is simply the latest unfolding of the dynamic that was started when the Industrial Revolution took place.
Having said this, since 1770 there do seem to have been several developments that can safely be called technological revolutions. These do not compare to the original ignition of industrial capitalism but they are historic events in their own right. Properly defined, these moments are paradigm shifts, dramatic in their own right, each marking a qualitative break with what had come before, and each changing, eventually, pretty much everything—from the way we structure and organise human activity, to cultural norms and aesthetics, to the role and shape of the state.
This, for me, is a box the digital revolution does tick, just as steam, steel, and electricity ticked it before. The advent of digital technologies, and the information/data economy they enable, is, in these terms, the fourth technological revolution since 1770.
Why was the Industrial Revolution so special?
To clarify this distinction, it’s useful to look back to Landes’ account of the original Industrial Revolution, which underlines just why this moment was unique.
The Industrial Revolution took place, Landes argues, when three technological breakthroughs came together:
- Mechanical devices were invented that could replace human dexterity, massively simplifying the work required from human hands in the manufacturing process
- Steam power was harnessed to replace human and animal power in manufacturing, allowing factories to be bigger and faster and to run for longer
- New techniques were developed to obtain raw materials, particularly metals and chemicals, so that machines and the products they made were stronger and more durable
Individually, these breakthroughs did not change the world, but what did change the world, explosively and irreversibly, was the combination of the three breakthroughs in the petri dish of profit-maximising capitalism.
(It’s intriguing that Landes does not include the role culture played in enabling and even triggering the Industrial Revolution. Later, he recognises culture as a key explanation for why the Industrial Revolution took place when and where it did, although his argument on this is not as well-developed as later writers such as Mokyr.)
It is Landes’ vivid account of the interaction between these elements that makes him so pertinent today in the context of concerns about runaway technological progress. Re-reading him now, it’s hard not to shift around in your seat at the uncomfortable feeling that you’re watching the opening seconds of an unstoppable chemical reaction.
Let’s take a specific example: John Kay’s invention, in 1733, of the flying shuttle. Kay’s invention replaced the human dexterity required in weaving with a much simpler series of movements that was therefore dramatically faster. By speeding up weaving in this way, the flying shuttle increased demand for yarn—the weavers were weaving so damn fast that they kept running out. In 1764, James Hargreaves then invented the spinning jenny, allowing one worker to operate many yarn-spinning spools at once, dramatically speeding up the process. Richard Arkwright’s water frame then introduced a powered version of the same process, speeding up yarn-making even more. The weavers now had all the thread they needed and the overall system of fabric production jumped forward.
Examples like this give us glimpse into the glowing heart of the furnace just after it was lit. Productivity-enhancing inventions were prompting other productivity-enhancing inventions, feeding back in a continuous, self-amplifying loop.
And things didn’t stop here. Before long, the combined effect of these many technological breakthroughs burst beyond the bounds of technology to wider society. Landes notes, for example, how new techniques created pressure for new ways of arranging human activity that could better coordinate the industrial processes now being undertaken. This prompted shops and home workrooms to be replaced by larger mills and factories, organising human activity into bigger clusters that coordinated the efforts of more people, in part by delineating their individual roles more clearly—with all this newly co-ordinated activity aimed at the clear and simple goal of maximising profit, i.e. the different between these factories’ inputs and outputs.
This process created, in short order, many of the essential proteins of our modern economy. The employer was born, playing the role of hiring labour, marketing products, and supplying machinery. So was the modern worker, giving their labour in exchange for a wage. Other, more complex proteins, such as the company, were to follow.
These individual developments mattered. But what remained essential was, as Landes explains, not the quantitative progress they represented, but the qualitative shift that had taken place: economic growth had become self-reinforcing.
Thus, these changes fed back into, and further accelerated, technological progress itself. In Landes’ words, developments in organisational forms held within themselves “the seeds of further technological advance”. A more tightly defined manufacturing process, built around increasingly specialised jobs, lent itself even better to optimisation—and to the cumulative and compounding effects of that knowledge over time.
So, knowledge built upon knowledge, and technique upon technique, meaning that a breakthrough in one area made other breakthroughs possible. And, more than this, because the whole thing ticked to the rhythm of the machine—and a machine now powered by steam or water—a breakthrough in one part of the process did even more than this: it actually incentivised breakthroughs elsewhere, creating bottlenecks that needed to be solved—or, more precisely, that promised vast profits to the person who could solve them, drawing human effort and energy, and finance, to where it would add most value.
It’s in this account that you get the sense of a runaway reaction. And it seems to me that even this doesn’t go far enough. When you step back, you see how progress in the private economy then prompted progress of a similar kind in entirely different spheres of human life, from culture to aesthetics to the role and shape of the state. This is the road to later thinkers, from Polanyi to Galbraith, who lamented capitalism’s totalising effects.
As Landes tells it, however, there is beauty in this early part of the story. At every level you see the same dynamic of self-accelerating progress play out. A specific technological innovation trigger another innovation. The growing body of technical innovations then trigger new and more productive ways of organising human activity. These new forms of organisation then lead to wider innovations in government and culture. In Landes’ words: “In all of this diversity of technological improvement, the unity of the movement is apparent: change begat change.” Like a fractal, the system looks the same however much you zoom in or out.
This is what was so special about the Industrial Revolution: it started the cascading dynamic of progress that continues to this day. In doing so, it made possible—or even inevitable—the several subsequent technological revolutions: steel, electricity, and now the pervasive, paradigm-shifting effects of digital tech and ubiquitous information.
All of this prompts two counter-veiling thoughts.
One, it makes you feel that recent fears about runaway technological progress and AI are endearingly late to the party. Yes, we should worry about unstoppable technological change. But if the part we’re worried about is AI or technologies themselves, we’re mistaken a symptom for a cause. The unstoppable part of technological change isn’t the technology, it’s the self-sustaining change, and that started 250 years ago.
Two, more reassuringly, seen another way, the period we’re now living through is not actually all that special. It is not in the same category of the original, game-changing, Industrial Revolution, and more like the other chapters of the story that has unfolded since. And that means we can learn lots from reading previous chapters—by exploring how progress unfolded in previous technological revolutions, and how these moments themselves fed through from technology to technology, into organisational forms, and, ultimately, into the state.