Collections: Why No Roman Industrial Revolution?

This week we are taking a look at the latest winner of the ACOUP Senate poll, which posed the question “Why didn’t the Roman Empire have an industrial revolution?” To answer that, we need to get into some detail on what the industrial revolution itself was and the preconditions that produced it, as well as generally sketching the outlines of what the Roman economy looked like. This certainly won’t be a comprehensive description of either so much as merely nailing down the definition of the industrial revolution and the basic outlines of the ancient Roman economy to see what elements of the former were missing from the latter.

As always, if you like what you are reading here, please share it; if you really like it, you can support me on Patreon; members at the Patres et Matres Conscripti level get to vote on the topics for post-series like this one! And if you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings.

The Question

That said this is a question that is not absurd a priori. As we’ll see, the Roman Empire was never close to an industrial revolution – a great many of the preconditions were missing – but the idea that it might have been on the cusp of being something like a modern economy did once have its day in the scholarship. As I’ve mentioned before, the dominant feature of the historical debate among scholars about the shape of the Roman economy is between ‘modernists’ who argue the Roman economy is relatively more like a modern economy (meaning both that it was relatively more prosperous than other ancient economies but also that the Romans themselves maintained a more modern, familiar outlook towards money, investment and production) and ‘primitivsts’ who argue that actually the Roman economy was quite primitive, less prosperous and with the Romans themselves holding attitudes about the economy quite alien to our own. But here we need to get into a bit more specificity because beneath that quick description it is necessary to separate what we might call the ‘old modernists’ and the ‘new modernists.’

The ‘old modernist’ view of the Roman economy ran essentially from the 1910s to the publication of Moses Finley’s The Ancient Economy (1973) which fairly decisively put an end to this view of the ancient world; the major ‘old modernist’ scholars were Teney Frank and Michael Rostovtzeff. In many ways Frank and Rostovtzeff were blazing a trail, some of the first to attempt really systematic study of the Roman economy drawing on the complete range of sources (which the increasingly interconnected world of the 1800s and 1900s made possible). The late 1800s and early 1900s are full of studies like this: the first efforts to pull together everything and then comprehensively assess topics often in massive and magisterial multi-volume works (e.g. Beloch on ancient demography, Mommsen on the structure of the Roman state and law). Those grand magisterial studies then go on to form the foundation for later scholarship, though this is often in the form of later scholars pointing out all of the ways that those grand magisterial studies were wrong.

Frank and Rostovtzeff’s works form the necessary foundation for everything that followed but they both fell into versions of essentially the same mistake of assuming that Roman social and cultural systems largely mirrored their own. This is an easy mistake to make generally but it must have been an especially easy mistake to make in societies that were only starting to industrialize and so still had a lot of old agrarian social structures; great landholders in their manors with their tenants (think Downton Abbey) and household servants must have seemed very much like the Roman elite indeed. And the anxiety of those great landholders facing an upwardly mobile class whose wealth didn’t come from land must also have seemed very familiar too; Rostovtzeff especially reads this sort of proto-capitalist orientation into the Roman equites.1 And so both Frank and Rostovtzeff assumed the Romans thought about money, profit, finance, wealth and even progress mostly the way they did. The picture that emerged from those assumptions was exactly the sort that prompts this question: Rome as a highly advanced agrarian economy, growing in wealth and even potentially in the early stages of that wild capitalist2 economic takeoff – except that it never quite got there. Needless to say that vision enhanced the apparent bitterness of Rome’s decline.

Finley’s The Ancient Economy (1973) essentially detonated a bomb in that sort of scholarship and created a clear break between those ‘old modernists’ and the ‘new modernists’ who are ascendant in the scholarship today. Finley3 sought to demonstrate that the ancient economy was not ‘proto-capitalist’ in its orientation but rather a decidedly alien economy where economic relations were structured by status, legally enforced class and slavery more than money or profit. While Finley’s school of thought (the ‘primitivists’) have largely lost the argument post-1990 or so, the ‘new modernists’ that won it would hardly contend that the Roman economy was very much like the Europe of the 1600s and the 1700s, verging into modernity. Instead they stress that Rome, while it was a complex agrarian economy, was nevertheless still a fundamentally agrarian economy, which in turn demanded different mindsets, risk calculations and so on.4

The thing is, being a particularly complex or efficient agrarian economy doesn’t seem to have been the most important thing for producing the industrial revolution or it almost certainly would have happened earlier and probably not in Europe.

So let’s look at what the industrial revolution was.

The Industrial Revolution

The first key is understanding that the industrial revolution was more than simply an increase in economic production. Modest increases in economic production are, after all, possible in agrarian economies. Instead, the industrial revolution was about accessing entirely new sources of energy for broad use in the economy, thus drastically increasing the amount of power available for human use. The industrial revolution thus represents not merely a change in quantity, but a change in kind from what we might call an ‘organic’ economy to a ‘mineral’ economy.5 Consequently, I’d argue, the industrial revolution represents probably just the second time in human history that as a species we’ve undergone a radical change in our production; the first being the development of agriculture in the Neolithic period.

However, unlike farming which developed independently in many places at different times, the industrial revolution happened largely in one place, once and then spread out from there, largely because the world of the 1700s AD was much more interconnected than the world of c. 12,000BP (‘before present,’ a marker we sometimes use for the very deep past). Consequently while we have many examples of the emergence of farming and from there the development of complex agrarian economies, we really only have one ‘pristine’ example of an industrial revolution. It’s possible that it could have occurred with different technologies and resources, though I have to admit I haven’t seen a plausible alternative development that doesn’t just take the same technologies and systems and put them somewhere else.

Now we can’t cover the entire industrial revolution with all of its complex moving parts but we can briefly go over the core of it to get a sense of the key ingredients. Fundamentally this is a story about coal, steam engines, textile manufacture and above all the harnessing of a new source of energy in the economy. That’s not the whole story, by any means, but it is one of the most important through-lines and will serve to demonstrate the point.

The specificity matters here because each innovation in the chain required not merely the discovery of the principle, but also the design and an economically viable use-case to all line up in order to have impact. The steam engine is an excellent example of this problem. Early tinkering with the idea of using heat to create steam to power rotary motion – the core function of a steam-engine – go all the way back to Vitruvius (c. 80 BC -15 AD) and Heron of Alexandria (c. 10-70 AD). With the benefit of hindsight we can see they were tinkering with an importance principle but the devices they actually produced – the aeolipile – had no practical use – it’s fearsomely fuel inefficient, produces little power and has to be refilled with water (that then has to be heated again from room temperature to enable operation).

Via Wikipedia, an illustration of the ancient aeolipile, an early use of steam to create reciprocal motion. Apart from the use of steam pressure, the aeolipile shares very little in common with practical steam engine designs and the need to continually refill and heat the water reservoir would have limited its utility in any case.

So what was needed was not merely the idea of using steam, but also a design which could actually function in a specific use case. In practice that meant both a design that was far more efficient (though still wildly inefficient) and a use case that could tolerate the inevitable inadequacies of the 1.0 version of the device. The first design to actually square this circle was Thomas Newcomen’s atmospheric steam engine (1712). The basic principle here is that you have a boiler at the bottom connected to a cylinder with a piston; a valve opens which admits steam from the boiler into the piston; the steam is at very low pressure, so a weight on the arm helps pull the piston up. When the piston reaches the top, another valve opens and sprays water into the cylinder, condensing the steam (by cooling it) and creating a low pressure zone so that atmospheric pressure pushes the cylinder down; in this basic design it is the down-stroke that is the ‘power stroke.’ Though a substantial improvement on previous efforts, the Newcomen engine had all sorts of limitations: the power it could produce was limited to atmospheric pressure, the motion it created was jerky rather than smooth and most importantly it was hideously fuel inefficient.

Via Wikipedia, a diagram of Newcomen’s atmospheric steam engine. The boiler (A) produces steam at relatively low pressures which push up the piston (D). When the cylinder (B) is full, cold water from a reservoir (L) is sprayed into the cylinder, condensing the steam in the cylinder and creating a partial vacuum. This causes atmospheric pressure to push down the piston, which then pulls down on the fulcrum (F-H), creating reciprocal motion.

Now that design would be iterated on subsequently to produce smoother, more powerful and more efficient engines, but for that iteration to happen someone needs to be using it, meaning there needs to be a use-case for repetitive motion at modest-but-significant power in an environment where fuel is extremely cheap so that the inefficiency of the engine didn’t make it a worse option than simply having a whole bunch of burly fellows (or draft animals) do the job. As we’ll see, this was a use-case that didn’t really exist in the ancient world and indeed existed almost nowhere but Britain even in the period where it worked.

But fortunately for Newcomen the use case did exist at that moment: pumping water out of coal mines. Of course a mine that runs below the local water-table (as most do) is going to naturally fill with water which has to be pumped out to enable further mining. Traditionally this was done with muscle power, but as mines get deeper the power needed to pump out the water increases (because you need enough power to lift all of the water in the pump system in each movement); cheaper and more effective pumping mechanisms were thus very desirable for mining. But the incentive here can’t just be any sort of mining, it has to be coal mining because of the inefficiency problem: coal (a fuel you can run the engine on) is of course going to be very cheap and abundant directly above the mine where it is being produced and for the atmospheric engine to make sense as an investment the fuel must be very cheap indeed. It would not have made economic sense to use an atmospheric steam engine over simply adding more muscle if you were mining, say, iron or gold and had to ship the fuel in; transportation costs for bulk goods in the pre-railroad world were high. And of course trying to run your atmospheric engine off of local timber would only work for a very little while before the trees you needed were quite far away.

But that in turn requires you to have large coal mines, mining lots of coal deep under ground. Which in turn demands that your society has some sort of bulk use for coal. But just as the Newcomen Engine needed to out-compete ‘more muscle’ to get a foothold, coal has its own competitor: wood and charcoal. There is scattered evidence for limited use of coal as a fuel from the ancient period in many places in the world, but there needs to be a lot of demand to push mines deep to create the demand for pumping. In this regard, the situation on Great Britain (the island, specifically) was almost ideal: most of Great Britain’s forests seem to have been cleared for agriculture in antiquity; by 1000 only about 15% of England (as a geographic sub-unit of the island) was forested, a figure which continued to decline rapidly in the centuries that followed (down to a low of around 5%). Consequently wood as a heat fuel was scarce and so beginning in the 16th century we see a marked shift over to coal as a heating fuel for things like cooking and home heating. Fortunately for the residents of Great Britain there were surface coal seems in abundance making the transition relatively easy; once these were exhausted deep mining followed which at last by the late 1600s created the demand for coal-powered pumps finally answered effectively in 1712 by Newcomen: a demand for engines to power pumps in an environment where fuel efficiency mattered little.6

With a use-case in place, these early steam engines continue to be refined to make them more powerful, more fuel efficient and capable of producing smooth rotational motion out of their initially jerky reciprocal motions, culminating in James Watt’s steam engine in 1776. But so far all we’ve done is gotten very good and pumping out coal mines – that has in turn created steam engines that are now fuel efficient enough to be set up in places that are not coal mines, but we still need something for those engines to do to encourage further development. In particular we need a part of the economy where getting a lot of rotational motion is the major production bottleneck.

Via Wikipedia, a late version of Watt’s final steam engine design. Watt made a number of improvements to the Newcomen engine, adding a separate condenser to allow the cylinder itself to remain hot, including a vacuum jacket around it to limit the energy loss from heat loss and eventually introducing a double-action where the piston was pushed by steam pressure in both directions, enabling a stronger and smoother stroke, along with gearing that allowed the reciprocal motion of the piston to be translated into the rotational motion necessary for most tasks.

You may be thinking that agriculture and milling grain is the answer here but with watermills and windmills, the bottleneck on grain production is farming, not milling; a single miller with a decent mill can mill all of the grain from many farmers, after all. That’s not to say mechanized grain milling couldn’t realize gains, just that they were slight. No, it is the other half of the traditional agrarian economy: textiles. You will recall that the major production bottleneck, consuming 80% or more of the time intensity of textile production (not including fiber production), is spinning the fibers into thread – a process which relies on lots of rotational motion (as the name implies). And indeed, in the 1700s, further improvements in looms (the flying shuttle) had intensified this bottleneck by making weaving progressively more efficient.

And yet again we have serendipity because Great Britain was the major center of textile production for much of the world. Through the Middle Ages, the movement of wool textiles was one of the most important trade systems in Europe: wool produced in Scotland and Wales was moved to England where it was turned into thread and then cloth and then sent to the Low Countries to be dyed before using Europe’s river systems to reach consumers all over the place. European imperialism had only intensified this system because British conquests in India had directed massive amounts of cotton into this same system alongside the wool.

But there is another key step necessary here: the steam engine produces rotational motion and the spinning process requires rotational motion but you also need a machine capable of turning lots of rotational motion into real efficiency gains for spinning. Prior to the 1760s, no such machine really existed. Since the Middle Ages you had the spinning wheel, but applying a lot of energy to a spinning wheel isn’t going to help – the spinner is still only managing a single thread. Still the pressure to produce spinning technology that could match the efficiency gains of the flying shuttle was on and in 1765 it resulted in the spinning jenny, developed by James Hargreaves. The spinning jenny allowed a single spinner to manage multiple spools at once using a hand-crank. Unlike the spinning wheel, which could be a household tool (and thus before 1765, most spinning was still literally ‘cottage’ industry, farmed out to many, many spinners each working in their homes), the spinning jenny was primarily suited for commercial production in a centralized location (where the expensive and not at all portable spinning jennies were). The main limit on the design was the power that a human could provide with the hand-crank.

Via Wikipedia, a diagram of a spinning jenny, with a hand crank (B) that can be used to turn multiple spools at once, multiplying the efficiency of the spinner.

And now, at last, the pieces in place the revolution in production arrives. There a machine (the spinning jenny) which needs more power in rotational motion and already encourages the machines to be centralized into a single location; the design is such that in theory one could put an infinite number of spools in a line if you had sufficient rotational energy to spin them all. Realizing this, textile manufacturers (we’re talking about factory owners, at this point) first use watermills, but there are only so many places in Great Britain suitable for a watermill and a windmill won’t do – the power needs to be steady and regular, things which the wind is not. But the developments of increasingly efficient steam engines used in the coal mines now collide with the developments in textiles: a sophisticated steam engine like the Watt engine could provide steady, smooth rotational motion in arbitrary, effectively infinite amounts (just keep adding engines!) to run an equally arbitrary, effectively infinite amount of mechanical spinning jennies, managed now by a workforce a fraction of a size of what would have once been necessary.

Via Wikipedia Commons, a photo from the Library of Congress showing how many spinning ‘mules’ could be connected via an overhead shaft to a single large source of rotational motion, like a steam engine, allowing truly massive amounts of thread to be spun at once by a much smaller work force (in this case smaller in more than one way; this picture was taken as part of the records of the National Child Labor Committee, an example of boys (center) working in textile mills. The work could be dangerous as there were few safety systems in place).

The tremendous economic opportunity this created in turn incentivized the production of better steam engines and the application of those engines to other kinds of production; there is a whole additional story of how the development of the steam engine interacts with the development of new artillery-production methods (both relying on the production of strong, standardized pressure-containing cylinders). All of those use-cases push steam engines to become smaller, more fuel efficient and more powerful, which in turn increases the number of tasks they can be put to. Eventually in the 1800s, these engines get small enough and fuel efficient enough to be able to move their own fuel over water or rails, collapsing the prohibitive transportation costs that defined pre-industrial economies and in the process breaking the tyranny of the wagon equation, decisively transforming warfare in ways that would not be fully appreciated until 1914.

But the technology could not jump straight to railroads and steam ships because the first steam engines were nowhere near that powerful or efficient: creating steam engines that could drive trains and ships (and thus could move themselves) requires decades of development where existing technology and economic needs created very valuable niches for the technology at each stage. It is particularly remarkable here how much of these conditions are unique to Britain: it has to be coal, coal has to have massive economic demand (to create the demand for pumping water out of coal mines) and then there needs to be massive demand for spinning (so you need a huge textile export industry fueled both by domestic wool production and the cotton spoils of empire) and a device to manage the conversion of rotational energy into spun thread. I’ve left this bit out for space, but you also need a major incentive for the design of pressure-cylinders (which, in the event, was the demand for better siege cannon) because of how that dovetails with developing better cylinders for steam engines.

Why Not in Rome?

Putting it that way, understanding why these processes did not happen in the Roman world is actually quite easy: none of these precursors were in place. The Romans made some use of mineral coal as a heating element or fuel, but it was decidedly secondary to their use of wood and where necessary charcoal. The Romans used rotational energy via watermills to mill grain, but not to spin thread. Even if they had the spinning wheel (and they didn’t; they’re still spinning with drop spindles), the standard Mediterranean period loom, the warp-weighted loom, was roughly an order of magnitude less efficient than the flying shuttle loom, so the Roman economy couldn’t have handled all of the thread the spinning wheel could produce.

And of course the Romans had put functionally no effort into figuring out how to make efficient pressure-cylinders, because they had absolutely no use for them. Remember that by the time Newcomen is designing his steam engine, the kings and parliaments of Europe have been effectively obsessed with who could build the best pressure-cylinder (and then plug it at one end, making a cannon) for three centuries because success in war depended in part on having the best cannon. If you had given the Romans the designs for a Newcomen steam engine, they couldn’t have built it without developing whole new technologies for the purpose (or casting every part in bronze, which introduces its own problems) and then wouldn’t have had any profitable use to put it to.

All of which is why simple graphs of things like ‘global historical GDP’ can be a bit deceptive: there’s a lot of particularity beneath the basic statistics of production because technologies are contingent and path dependent. Now all of that said I want to reiterate that the industrial revolution only happened once in one place so may well could have happened somewhere else in a different way with different preconditions; we’ll never really know because our one industrial revolution spread over the whole globe before any other industrial revolutions happened. But we can still note that the required precursors for the one sample we have didn’t exist in the Roman economy.

But then that raises, I think, another question with how we think about economies in the past: if it wasn’t on the cusp of a revolution, what made the Roman economy unusual?

The Nature of the Roman Economy

Broadly speaking we can think about human production as fitting into three major types: non-agrarian hunter-gatherer societies, agrarian and pastoral societies, and finally industrial societies. The first merely harvests what the environment already provides, while agrarian and pastoral societies actively reshape local ecology to make it provide more. Yet both are ‘organic’ economies in that nearly all of the energy they use (with a few, largely marginal exceptions) is provided by muscle power which in turn derives from food consumption which in turn derives, ultimately, from solar energy and photosynthesis. Industrial economies, by contrast, derive the majority of the energy they use from sources other than muscle power – initially chemical reactions (burning coal and other fossil fuels) and later nuclear power, solar, etc.

The point is that these systems are not merely different in degrees, but different in kind, functioning on a very different basis with different potential avenues for growth in production. In an industrial economy, there are many new potential sources of energy which can be harnessed for production, whereas in an organic economy inputs are functionally limited to the land. New land can be brought under cultivation (or cultivated more intensively) of course, but marginal gains decrease rapidly (because the best land is cultivated first and because adding more labor to already cultivated land, while it can increase harvests, is less efficient than cultivating new land) and there is a fairly hard ceiling on total production of this sort that was, for the most part, fairly low.

Nevertheless there is room in an organic economy for small sorts of efficiencies that can collectively add up to greater economic output, albeit not on anything like the scale of increasing output in an industrial economy. We’ve actually talked about some of this before. The summarize, it is broadly supposed by historians that the Roman economy (particularly c. 100 BC – 200 AD or so) was remarkably productive for an organic economy enabling a relatively high general standard of living for an organic economy. How does that happen? We think there are a few favors that led the Roman economy to perform better.

First, the Roman Empire as a result of its conquests created a linguistic, customs and monetary union over the whole Mediterranean, which was kept relatively free of things like pirates and bandits. Each of these changes made markets more reliable and efficient, which in turn could mean that a larger proportion of farmers could specialize their farming output (with elite estates probably leading the way), resulting in higher total output as the market supplanted safety-through-diversity farming strategies common in agricultural economies with low degrees of farming integration. That greater output then enables the economy to support more specialized workers with high productivity making non-agricultural goods which thus become more common and eventually affordable by the farmers. We can see these processes only imperfectly, but archaeological evidence in the early empire seems increasingly to indicate there was meaningful regional specialization (most visible with olive oil because of how it was transported), suggesting these processes were at work. Likewise, we see specialized non-agricultural goods showing up in non-elite contexts at greater rates, suggesting that even the lower (non-slave, an important caveat for this period) classes have greater access to these things.

Shipwrecks (and by implication, shipping volume) in the Mediterranean 1500 BC to 1500 AD. As the volume of shipping grew, it would have been easier for farmers to trust markets to provide essentials from other parts of the Mediterranean and thus for some regional specialization to emerge, resulting in overall gains in productivity. We see this clearest in olive oil production, where the wares of production centers in North Africa and Spain are detectable (because they traveled in ceramic vessels that survive) as far away as Britain.
Graph after Fig. 2.5 from A. Wilson, “Developments in Mediterranean shipping and maritime trade from the Hellenistic period to AD 1000” in Maritime Archaeology and Ancient Trade in the Mediterranean (2011).

Second, the interconnectedness the Roman Empire created also encouraged the spread of innovations in production, both agricultural and non-agricultural, things like watermills for the grinding of grain, new more effective presses for olives, higher quality metal-working and so on. We can map these innovations only imperfectly, but once again archaeology is slowly filling in a picture where the movement of these ideas was significant. Once again I want to note these technologies were not revolutionary but evolutionary and often what was changing was not their existence but their distribution: ideas that had been ‘stuck’ in one corner or other of the empire can suddenly spread out over those more interconnected lines of trade.

Finally we also have evidence (albeit somewhat more tricky) that the period also sees the accumulation of productive capital, plausibly encouraged by the relative stability and peace the Roman Empire created with in its borders. Diet indicators and midden remains indicate that there’s more meat being eaten, indicates a greater availability of animals which may include draft animals (for pulling plows) and must necessarily include manure, both products of animal ‘capital’ which can improve farming outputs. Of course many of the innovations above feed into this: stability makes it more sensible to invest in things like new mills or presses which need to be used for a while for the small efficiency gains to outweigh the cost of putting them up, but once up the labor savings result in more overall production.

But the key here is that none of these processes inches this system closer to the key sets of conditions that formed the foundation of the industrial revolution. Instead, they are all about wringing efficiencies out the same set of organic energy sources with small admixtures of hydro- (watermills) or wind-power (sailing ships); mostly wringing more production out of the same set of energy inputs rather than adding new energy inputs. It is a more efficient organic economy, but still an organic economy, no closer to being an industrial economy for its efficiency, much like how realizing design efficiencies in an (unmotorized) bicycle does not bring it any closer to being a motorcycle; you are still stuck with the limits of the energy that can be applied by two legs.

As a result, the ingredients for the ‘take-off’ of the industrial revolution (which involves adding more energy to the economy on a per capita basis) aren’t there. While the Romans are coming up with clever ways to drain deep mines (mostly mining for precious metals; deep shaft mining for tool metals mostly seems like it wasn’t done. Probably it wasn’t generally economical), they aren’t doing this at coal mines (because they don’t use much coal, though they do use some), which means they don’t have the neat coincidence of abundant fuel in a place that needs pumping which gave rise to the first practical steam engines. They also lack the metallurgical capacity to easily build such engines and even if they had them they lack very many industries prepared to be revolutionized by cheap reciprocal or rotational energy (remember, they’re only in the slow beginnings of the process of switching to watermills from muscle-powered mills).

Instead the Roman economy essentially moved from a ‘low equilibrium’ organic economy (that is stable at low efficiency, with little specialized farming production and very limited agricultural capital being used) to a ‘high equilibrium’ organic economy (that is stable at higher efficiency due to markets encouraging specialized production and more agricultural capital). We cannot be sure exactly the scale of production growth that movement entailed; some estimates put it around 25% but these are very speculative; in practice we really don’t know. Just as the political conditions provided the ‘push’ to move the economy from one stable position to the other, political collapse seems to have moved the Roman economy right back down. Roman economic growth, because it was a product of institutions instead of technologies, was not durable in the face of those institutions collapsing.

In my view the key takeaway here is just how contingent the industrial revolution was: the industrial revolution that occured required a number of very specific pre-conditions which were really on true on Great Britain in that period. It is not clear to me that there is a plausible and equally viable alternative path from an organic economy to an industrial one that doesn’t initially use coal (much easier to gather in large quantities and process for use than other fossil fuels) and which does not gain traction by transforming textile production (which, as we’ve discussed, was a huge portion of non-agricultural production in organic economies), though equally I cannot rule such alternatives out.

Much of history ends up this way. As much as we might want to imagine that the greater currents push historical events largely on a predetermined path with but minor variations from what must always have been, in practice events are tremendously contingent on unpredictable variables. If Spain or Portugal, for instance, rather than Britain, had ended up controlling India, would the flow of cotton have been diverted to places where coal usage was not common, cheap and abundant, thereby separating the early steam-powered mine pumps both from the industry they could first revolutionize and also from the vast wealth necessary to support that process (much less if no European power had ever come to dominate the Indian subcontinent)? This question, like so many counter-factuals, is fundamentally unanswerable but useful for illustrating the deeply contingent nature of historical events in a way that data (like the charts of global GDP over centuries) can sometimes fail to capture.

Next week is going to be a gap week because I will be giving a talk at PDXCON2022. I know there were plans to record that part of the event; if that recording is made available (I hope it is) I will be sure to share it here.

  1. Historians these days generally reject this pattern of thought which sought to reframe historical systems in the context of modern systems. Perhaps unsurprisingly the scholarship of the early and mid-1900s on labor and the economy in the ancient world was heavily inflected by Marxist historical and economic theory, either by historians who were themselves Marxists (Finley but also G.E.M. de Ste. Croix (fl. 1954-1981) or who were reacting violently against Marxist thought – Rostovtzeff, who fled his native Russia after the Russian revolution, being the obvious example here. These days historians tend to be quite skeptic of ‘grand narratives’ of this sort, generally contending that the particulars of time and place overwhelm the superficial commonalities upon which such grand narratives rely.
  2. Emphasized more by Frank and Rostovtzeff than industrial production.
  3. Drawing himself on the work of Max Weber and Karl Polanyi
  4. In particular a great many of the essential ‘bricks’ of Finley’s argument have collapsed. Finley argued against trade in bulk staples, a position sustainable in 1973 but not today due to archaeological evidence. Likewise, Finley’s argument that the Romans couldn’t do sophisticated accounting has crumbled as more evidence for Roman finance has emerged. In many cases it has become clear that what Finley viewed as a general ancient aversion to profit-making, careful accounting and market interactions was in fact just the snobbery of our sources (landed elites for whom a more sophisticated economy created as much competition as it did opportunity) adopting a pose of disdainful moralism; as our evidence has improved it seems increasingly clear that this was a facade even among the elite.
  5. Terminology here borrowed from E.A. Wrigley, Continuity, Chance and Change (1988).
  6. Likewise the earliest steam engine locomotives were first used in coalmines for two decades (1804 – 1825 or so) before they became efficient enough for general use.

398 thoughts on “Collections: Why No Roman Industrial Revolution?

      1. Also the first sentence ends with “Why did the Roman Empire have an industrial revolution?”. I assume there is a missing “not” in the question.

        1. Although I would have definitely enjoyed reading Bret’s essay on “Why did the Roman Empire have an industrial revolution, and why was all the evidence for it destroyed?”

          1. Same reason that all evidence of the Industrial Revolution in Arthurian England spearheaded by the mysterious Sir Hank was covered up – the Church intervened. The Pontifex Maximum thought that all the foundries & windmills were causing cultural decay and loss of character-building that the Romans of old knew.

      1. Head is an archaic measure of pressure (“static head”), a measure of how high a column of water would have equivalent pressure (p0 = ½ρv² + ρgh)

        1. In metric units, the bar (which is customary, the actual SI being pascal or megapascal) has about the same function: it is 100,000 Pa, very close to athmospheric pressure. Because metric units are surprisingly nifty, the bar also measures rather closely (with 1 % error) the static pressure of a water column in decameters, and millibar gives the same height in centimeters.

    1. > the industrial revolution that occured required a number of very specific pre-conditions which were really on true on Great Britain in that period.

      on true on ?

    2. Normally a fan but this is a disappointing post. A post on this topic that follows the modern literature on economic growth would discuss Roman Society: culture, institutions, scientific innovation. Acemoglu, Mokyr, Robinson, McCloskey, and many others have passed by a strict “coal and steam engines” understanding, and even those who stick with such an opinion have to reckon with why such inventions (and the preceding scientific knowledge) were innovated.

      Also there’s a pretty big factual error in this post.

      Not only was coal not as critical to the IR as Pomeranz and Wrigley assumed, but the substitution cost of coal was not significant. Early efficiency innovations were made using mechanical innovations without involving coal (Things like the spinning jenny, which you mentioned). In addition, the total cost to England of switching from coal to firewood would’ve been something like 5% of English GDP, a high amount but not super high. A society without coal could’ve gotten through the initial stages of industrialization. Source: https://doi.org/10.1017/S1361491606001870

      In general, the specific technological innovations that drove the IR mattered less than the factors that led to them: culture (McCloskey 2016), institutions (pick an Acemoglu article out of a hat), and human capital (https://doi.org/10.1146/annurev-economics-080213-041042, https://doi.org/10.1017/S0022050705000112 other such Mokyr articles).

      1. “Points still under contention in the scholarship” is not the same as “factual error.” I presented what, as far as I can tell, remains broadly the consensus view among historians.

        1. Reply well crafted, sir. I have to ask, and in doing so risk the revelation of my enormous ignorance, what constitutes an industrial revolution? I would have thought that the Romans brought the waging of war to an industrialized level through the manufacture of swords and armor, all requiring the specialized and decidedly scaled-up processing of ore, metal, blacksmithing etc. to unprecedented numbers. Was not the gladius and its numbers produced at an industrial level in order to equip the legions? Didn’t the financing, manufacture and distribution of armor and “uniform” of a professional standing army constitute a revolution in said supply, delivery and usage? And then there is the question of amphorae and the ships that carried them. Weren’t the numbers attested to in the chart of shipwrecks in and around the turn of the BC to CE timeline testament to another, earlier revolution?

          Truly, I am asking sincerely what would qualify as an industrial revolution, the above being just a couple of examples of observations from a poorly (Hollywood?YouTube) (mis)informed fan of all things historical and such.

          1. You might be able to make a case that Roman enjoyed some sort of economic if not industrial revolution: the co-dependent shift to a maritime market economy and specialization of agricultural labor, whatever organization of workshops enabled equipping large armies in f*cking *chainmail* (rather labor intensive to make, especially if you don’t even have wire drawing), etc.

            OTOH, the “first industrial revolution” of the 1700s…, well I don’t know that much in detail, but you’ve got better machinery: flying shuttle loom, spinning frame; better iron smelting (use of coked coal instead of charcoal, better furnaces and techniques to deal with the impurities; I would guess better tooling and linkages for exploiting and transmitting water power.

            And then after that you have the steam engine, and a huge increase in the amount of work that can be done, fed by coal and then any other heat source. As I said in another comment, being able to efficiently turn heat differences into work (and vice versa) is a huge physical advance.

            Plus everything else going on: scientific chemistry (plastic,explosives, fertilizer), better farm equipment (that better machinery again, even before tractor engines)…

          2. Our host lays out the three basic economic frameworks: hunter-gatherer, agricultural, and mineral. The Roman Republic/Empires were firmly agricultural for their entire existence. They had water- and windmills and burned some amount of coal for heating, but the vast majority of their energy was organic in nature. The IR is a paradigm shift to using mineral energy.

        2. The consensus according to who? This article’s only citation that seems to be on the early industrial revolution specifically is dated back to 1988. That’s a dark time in economics, back before the econometric revolution. Our body of economic data concerning the early industrial revolution has expanded greatly since then and thinking from that era has aged extremely poorly.

      2. Switching from coal to wood for heating, at a collective price of 5% of national GDP, is not an insignificant cost! This is still an economy with a lot of subsistence farmers, remember. If “GDP” in such an economy represents total economic output, then most of that output is still food grown for subsistence, homespun made to be worn by the family that made it or a close loved one, and so on. Only a fraction- and typically not a majority- of GDP is available to be diverted to any purpose other than the absolute minimum survival needs of the specific individuals performing the labor.

        5% of such a country’s GDP represents a very sizeable slice of the entire economic surplus. Which means either a massive decline in the availability of specialist labor (which in turn reduces the efficiency of subsistence activities because fewer people are making tools and containers and sturdy buildings and so on), or that a significant slice of the general populace takes a significant (but proportionately smaller) hit to their survival needs.

        In this case, the hit to survival needs expresses itself as a massive increase in heating fuel costs. I don’t know what percentage of the English GDP was spent on heating fuel historically, but I doubt it could plausibly have been more than 10-20%… In which case increasing fuel expenditure by 5% of GDP would increase fuel costs by at least 25-50%. More likely it would increase prices by more, because the higher prices would price some people out of the market for fuel entirely, and the remaining fuel would be chased by fewer customers individually paying more per ‘BTU,’ so to speak.

        The problem is that these are people living in poorly insulated shelters, who have to live through British winters (soggy and rather cold). A significant number of them will die every winter if they don’t have the means to heat their home. And indeed, affording fuel for the home was a nontrivial problem even up into the Victorian age for the lower classes.

        This is more than enough to incentivize a search for nonstandard (read: non-wood) sources of heating fuel… which, in Britain, meant coal mining. There might have been some hypothetical path to industrialization where “so, how do we get the water out of our deep coal mines” wasn’t a necessary step on the path, but it sounds like Britain in particular did need an answer to that question.

        1. “Switching from coal to wood for heating, at a collective price of 5% of national GDP, is not an insignificant cost! This is still an economy with a lot of subsistence farmers, remember.”

          I imagine it probably sounds like a heavy burden to someone unfamiliar with the size of the British national debt back then or how quickly it grew. The British economy was well capable of sustaining a deficit extracting far more then 5% of the economy diverted into war or foreign subsidies and had been for quite a while before the industrial revolution.

          1. 1) Insofar as “the British economy” was capable of doing this, these diversions of economic resources often involved a lot of extraction and a large fraction of the population being driven into the most brutal imaginable kinds of poverty, including deaths caused by sheer poverty and privation from malnutrition and exposure. The fact that the kingdom doesn’t suddenly cease to exist or sink beneath the waves if they try to convert from coal back to wood doesn’t mean that it’s in any way a reasonable thing to suggest that they do so.

            2) When you say this, are you referring to “the British economy,” or to “the British government’s budget?” Economy and GDP counted how? We are rapidly bumping up here against the practical limits of what we can say without a deep dive into the details of how an economy works.

            I’m afraid I don’t have the journal access to follow the link ‘Winter’ originally provided in the post I’m responding to, but you seem fairly confident you know your stuff. Would you mind providing more detail?

          2. “involved a lot of extraction and a large fraction of the population being driven into the most brutal imaginable kinds of poverty”

            Britain was notable for it’s high standard of living at the time compared to economies far worse at extracting surplus from their population…

            You keep trying to pound a square peg into a round hole. This is an intensely studied case with it’s own peculiarities.

            “When you say this, are you referring to “the British economy,”

            Budget deficits and account deficits are closely interrelated in economics. An economy that is capable of spending a century running up deficits sending money off to continental allies as subsidies and sending goods off to support colonial ventures which send back far less in exchange is an economy that has goods to spare for exports to pay for those things. In more modern times then the early 1800s exports might take the term of foreign direct investment if the supply of such investment opportunities is rising quickly enough but Britain was doing this back in the 1700s where that was not the case.

      3. Correct me if I am wrong but my understanding was that by the beginning of the 17th century in England diminution of forests, increased population and industries demanding more energy had led to a “wood famine” wherein people were paying for the right to pick sticks up off the ground in groves for fuel. The sheer unavailability of sufficient wood would seem to make abandoning coal impossible.

        1. Importantly, this kind of dynamic means that the equilbrium price for firewood does not accurately reflect how much firewood would cost if the British Isles had stopped burning coal or tried to do without it in the 1700s. In the absence of coal, demand for firewood would skyrocket, necessitating clear-cutting of the last remaining patches of forest in the UK, and scarcity would drive up prices rapidly.

          So just comparing “this is how much enough coal to keep your house warm would cost in 1775” and “this is how much enough firewood to keep your house warm would cost in 1775” and saying “no, the British could totally have converted to wood burning only, the increased expenses would be about 5% of GDP” misses a lot of what’s going on.

          From a purely materialistic analysis, no, they couldn’t, because it would be unsustainable. The firewood cutting would strip the island of every last tree in short order, and importing firewood for an island populated by millions from overseas is impractical in the extreme. People would have to use a lot less heating fuel, which would mean a sudden crash of key industries (e.g. metalworking) and a lot of people freezing to death.

          From a more ‘money’ oriented economic analysis, the problem is that without the substitution effect provided by coal, the price per cord of firewood would skyrocket, but the demand for firewood wouldn’t decrease significantly. Firewood would just get vastly more expensive, the same way any other commodity does when its demand suddenly spikes and supplies can’t change much.

          1. Many people used fuels other than wood. The advent of coal saw the land dedicated to those fuels vanish. So it was wood or nothing.

            Also, woodland was now used for lumber so back to wood would make that vanish too

      4. I’ve read a fair bit of Acemoglu and Mokyr. Their work on institutions and culture usually draws on very broad-brush history, and often also not the most recent. Human capital – in the sense of skills and knowledge, certainly. But it helps to remember Keyne’s dictum that ‘we can afford anything we can actually do” – the issue is what you can do, not what you can afford. Textiles made a lot of money (the initial returns to mechanisation were ridiculously high – 200% pa), but they were not the industrial revolution. To consider a counter-factual: the combination of skills, maritime dominance (and hence uninterrupted access to cotton and dyes) and water-power might have given Britain a near monopoly on textiles, and made it very, very rich. A sort of cloth petro-state. But that would not have led to railways or steamships or mass production of steel. It was those that ushered in the industrial age. So coal and steam are at the heart of the story.

  1. Interesting topic, but I’m really not sure that I agree with the characterization of hunter-gatherer societies “merely harvesting what the environment already provides”. Hunter-gatherer societies engaged in massive alteration of the environment to increase food yields, such as controlled burns to clear forest and scrubland. Sure agricultural societies are orders of magnitude more productive but hunting and gathering still involves managing the environment.

    1. I’m not sure that the controlled burns of hunter-gather societies do would count as a “massive alteration of the environment” ie. burning the scrub to get turtles and eggs, some controlled burns to time the flowering on the veldt and some use of fire in hunting. It is not nothing but is it significantly different than what the average lightening strike would do? Or even a herd of elephants passing through?

      That is quite different than say agriculture which actively changes the landscape in my view.

      I also wonder where in this complex semi-nomadic groups fit in folks in the style of the “Sioux” who can’t really be considered hunter gatherer but neither are they agricultural once they leave their sedentary lifestyle.

      1. There is stuff like niche cultivation, where the landscape is essentially “gardened” over generations, encouraging and discouraging various plants and animals. The Jomon (Japan) turned out to be big nut arborealists, weeding the forests to encourage nut orchards; many Amazonian tribes, rather than leaving the forest pristine wilderness, cultivated groves and stands of different vines and trees, either for plant products or to encourage animals. The Dark Emu hypothesis, for which evidence is more debatable than the Amazonian forest garden, suggests that such niche cultivation was possible over large grassland. The archaeology in Starr Carr in Yorkshire (UK) shows intensive wetland cultivation, with selective reed burning (encourages deer), earthworks for fish weirs, selective forest clearing or replanting, and this is Mesolithic, with a very high (and diverse) yield thousands of years before arrival of farming. The high yields of wetlands niche cultivation causing a population boom in mesopotamia is a hypothesis for why people turned to grasses (because skeletons show that farmers had poorer lives).

          1. I think the point is that calling them “intermediary stages” paints a weird kind of teleology: Just like the roman economy wasn’t developing into an industrial economy the australian economy wasn’t developing into an agricultural one: It was a heavily sophisticated hunter-gatherer economy involving quite a lot of reshaping of nature.

          2. To answer Arilou’s point, one could substitute “intermediate points” for “intermediary stages,” although there may not be that many historical examples of cultures that remained at such intermediate points for more than a century.

          3. One “intermediate” is “complex hunter-gatherers” like the Pacific Northwest (PNW), where you had not just abundant food but some level of food preservation (dried/smoked salmon), going with sedentary lifestyles, social stratification, and slavery. AFAIK it lasted far longer than a century and showed no signs of “progressing” toward agriculture.

            Any similarly rich locations in the Old World would have been taken over by invading farmers millennia ago.

      1. Except basically all observed hunter-gatherer societies, from Native Americans to Australian Aboriginals, did this sort of thing.

        1. I combine this with the observation that nevertheless, there does seem to be a real qualitative difference. Even a crude agricultural society like ancient Sumeria starts to look quite unlike even a highly developed hunter-gatherer society.

          The combination suggests that as with the Industrial Revolution (though to a lesser degree) there is a specific and relatively narrow “keyhole” here. A hunter-gatherer society that happens to pass through the keyhole undergoes a very dramatic transition into a nearly unrecognizable condition of things.

          In this case, the “killer app” sounds like it’s “Find a high-efficiency food crop that lets you get more mileage out of your existing experience with altering the landscape and ecology in your favor.” Historically, this is usually some kind of high yield starch such as grain or potatoes.

          Having ordinary edible wilderness plants such as can be found anywhere in the tropics and temperate zones throughout the world lets you develop a highly sophisticated hunter-gatherer economy in most places that are reasonably fertile.

          Having one of the handful of really good edible plants (often created or improved by selective breeding) means that within a few millennia, your highly sophisticated hunter-gatherers are building temples the size of hills and amassing armies that shatter entire coalitions of the same kind of people you used to be before you went through the First Agricultural Revolution.

          1. I don’t think agricultural societies are *always* going to develop complex states though. Papua New Guinea never did, even though they had agriculture for a very long time (and had sweet potatoes starting maybe after 1000 AD or so). I’m guessing the forbidding topography of New Guinea probably has something to do with that.

          2. There are two different books titled _Against the Grain_ arguing that states needed grain to exist. Grain harvests mature at once, making it easier to collect as taxes, and can be stored for a long time. Also dried grain is a lot denser than water-rich root/tuber crops, easier to collect and move.

            The Andes/Incas might seem an exception, but they did grow maize (also pseudocereals like quinoa and amaranth), and also the climate allowed for freeze-drying potatoes, solving at least two of those problems (density and preservation)

          3. There are two different books titled _Against the Grain_ arguing that states needed grain to exist. Grain harvests mature at once, making it easier to collect as taxes, and can be stored for a long time. Also dried grain is a lot denser than water-rich root/tuber crops, easier to collect and move.

            You did get a few state-level societies that existed in areas that depended on roots, tubers or other non-cereal starchy staples like bananas. The Kongo Kingdom, for example, or the Polynesian proto-states like Tonga. The Kongo eventually did get corn via the Portuguese, but that’s well after they had formed as a state.

            Coastal West Africa also depended largely on root and tuber crops but they did have at least one native cereal domesticate (African rice).

    2. The problem I have is more the clear divide between “just harvests what the environment provides” and “alters the environment to increase food yields”. Farming was not a singular stroke of genius, but a line we draw in the sand between “just gathering” and “real agriculture”.

      1. Real life is messy and defies easy classification. And yet we need to do classification to study it, else there’s no way to do wide-ranging studies that can reveal similarities and deeper truths.

    3. You could say “hunter gatherer adapts and harvests what is there, agriculture creates an ecology”, since gardening and adjusting doesn’t come close to replacing everything with a wheat field, or bringing a giant herd of probably foreign animals onto land, plus breeding will get more intense. How to fit a quick two sentence version into the blog post I’m not sure how to do.

        1. There is still a gap between humans directly doing almost all the work of controlling plants vs. burning and than letting whatever grows afterwards grow.

          There’s obviously a transition where management shades into outright farming, but the same is true of partly industrializing societies, and using machinery for things like mills and spinning wheels goes a good bit back into mostly agricultural societies, so there’s always some overlap.

      1. the word you are looking for might be “monoculture,” not ecology. a forest or marsh, with diverse species filling various niches that runs without human intervention, are ecologies. a food forest as seen in the pre-colonized Pacific Northwest, where humans encouraged the propagation of preferred plants and animals but not in monoculture, might be a managed ecology.

        a wheat field is a monoculture; a vegetable garden is a collection of small monocultures.

    4. It’s hard to draw a line between hunter-gathering and agricultural in what they do, but it’s more obvious in the population density that can be supported. In areas that could be used for either, agricultural societies (once they’ve full converted) have much higher populations, order of magnitude difference.

  2. In the first sentence…

    This week we are taking a look at the latest winner of the ACOUP Senate poll, which posed the question “Why did the Roman Empire have an industrial revolution?”

    Shouldn’t the quote be “Why didn’t …”?

  3. Very nice. Probably one of my favorite ACOUP posts in a while!

    It’s interesting to read this after having just read What We Owe The Future, in which Will MacAskill argues that today’s civilization might not recover post collapse, because we might have mined out all the coal by then, and easy-to-access coal was critical to the industrial revolution.

    Quite interesting, because Deveraux argues here that it was actually *difficult-to-access coal*, not surface coal, which drove the demand for early steam engines. But I guess a post-collapse industrialist would still have plans for higher-tech steam engines (and use-cases for them), so would probably be able to skip past the point of needing deep coal to catalyze it. Still, very interesting to see the contingency of the industrial revolution.

    1. Easy access to combustibles will still be an issue. I remember Mote in God’s Eye, (spoiler for a 50-year old novel) where a civilization collapses regularly due to population pressure. They build “museums” to safeguard critical technology for the next cycle, but this has been going on for so long that they’ve used up all the hydrocarbons and even fissile materials around. One character mentions the difficulty of leapfrogging from firewood to fusion power.

      1. That book also says that the Moties have stable colonies in outer space at various points around their solar system, which don’t collapse when the onworld society does.

        This would obviously remove any need for museums or technological rediscovery.

        1. The Moties are also very competitive and clan/lineage orientated, much more so than humans. Building museums helps ensure the descendants of the planetary civilisation don’t get conquered and/or wiped out by the descendants of the space going Moties.

      2. So how does magic, especially ubiquitous magic common in most High Fantasy and D&D settings, change the equation?

        Even a baby wizard or local seminarian brings a nigh-infinite supply of energy in just their cantrips, while more advanced spellcasters rival nuclear reactions will a flick of their fingertips. Magic also provides levels of nigh instantaneous information sharing via sending, crystal balls, teleportation and telepathy that rivals radio, TV and arguably the Internet. Yet most classical fantasy presupposes a High Middle Ages tech tree for millennia – the Forgotten Realms is a classic example of a relatively high magic setting that hasn’t developed.

        Is that even feasible? Or is it almost inevitable that a high or ubiquitous magic society would rapidly industrialize and develop an industrial/knowledge economy fuelled by magic?

        1. Depends how you write it. I write ‘magic is common’ stories where the constraint is that the land does not like fire or major disruptions. So magic for lighting and heating and harvesting and communication, but medieval plus for fighting because earthquakes are inconvenient. Lawrence Watt-Evans’ Ethshar did something similar (different constraints)

        2. Depends on the rules. D&D magic is combat oriented. Perhaps the gods, including the god of magic, granted it for that purpose.

          Of course, the nominally medieval world often has features that would require high yield agriculture, vaccines, and antibiotics. The survival rate of children is too high for actual medieval times.

          1. Because the D&D is combat oriented, they didn’t put a huge amount of thought into non-combat applications of magic. Some of this stuff has world-changing effects that they didn’t think of.

            Looking at the 5th edition Player’s Handbook, and looking only at cantrips (which even a beginner spellcaster can cast an unlimited number of times per day):
            Acid Splash and Poison Spray: Create some acid or poisonous gas out of nothing. Intended use is throwing the acid at an enemy, but also pretty useful for industrial chemistry.
            Fire Bolt and Produce Flame: Create fire from nothing. Intended use is throwing it at an enemy, but you could also use it to operate a forge or steam engine with no fuel.
            Mending: With one minute of work, perfectly repair any damage less than a foot long. This would dramatically increase the useful life of a great many things.
            Prestidigitation: The spell does several different things, including instantly cleaning an object (saving a great deal of labor, especially on laundry) or flavoring food (potentially killing the demand for spices).
            Ray of Frost: Reduce something’s temperature. Intended use is freezing an enemy, but could also be used for refrigeration, and I figure there are probably chemical and metallurgical applications too.

          2. The chemicals don’t stick around. And all require a target that is not an object.

            It’s really combat-oriented. And there are literally no references to using stuff out of combat.

        3. Yes — you can just define your magic to not work for magitek applications. “NPC wizards use cigarette-lighting spells to light cigarettes; PC wizards use cigarette-lighting spells to start arbitrary kinds of fires” — but this is only because the author/DM/whoever decides that the spell creates “ordinary”, physical fire. But if you want to prevent the Magitek Revolution, you can have a wizard cast “light small fire” at a dry haybale with the result that a small fire burns steadily for a defined time and then goes out, having consumed only a small chunk of the haybale.

          (Amusingly, basic spark-gap radios are stunningly simple to build — a limited amount of wire, glass/pottery for capacitors, a two-metal-sandwich pile to power it — easily within the ability of the Romans. Telegraphs, with their immense lengths of wire, are industrially much more demanding! Too bad that, as it turns out, it was scientific study of electromagnetism that gated radios.)

          What I personally see as the less acceptable part is the continued existence of megafauna (dangerous or not). Cavemen have driven a whole bunch of species into extinction. And if you have an agricultural society — with high population density, transforming/controlling a large fraction of the land for production — then the peasants with torches and pitchforks will steamroll most things that eat their crops and/or sheep. (IIRC bears and wolves were driven extinct in the large lowland areas of Western Europe before gunpowder.)

          Personally, I’m in a sense pessimistic and I also extend this problem to intelligent species sharing a subsistence system. If everyone from hobbits to dwarves is a cereal farmer, and they have their respective polities, over *centuries* someone is eventually going to disappear. If you make them subordinate under one empire, presumably the process is slower (tenants have some traditional protections against being replaced by hobbits) but the balance has to be surprisingly even to not get any movement at all (and the empire unreasonably stable).

          The obvious solution is to create ecological niches with corresponding subsistence systems for each species. A temperate region too wet for intensive cereal agriculture (as I understand, half of Ireland goes into this bucket) could have a nerfed potato (nerfed so that its cultivation doesn’t invade drier climes) or, more fancifully, forest gardening. A temperate region too dry gets horse nomads. (You can make the milk-equivalent be literally poisonous to the cereal-farmer species, see lactase persistence.) If you don’t mind coming off heavy-handed (“reality is unrealistic!”), you can also write that e.g. tropical diseases kill horses (good luck with logistics and plow-pulling).

          Alternatively, you can decrease the driving differences by making the “species” actually be human ethnicities. There are still events like the Indo-European and Bantu expansions, but you can pick your era “around” them.

    2. I think it was the combination of both: You need enough “easy” coal that the economy can start to depend on it. And once you’ve used up all of it, there’s vastly more under the ground, so it’s economically viable to start exploiting those resources at a higher cost. And an even higher amount once you’re digging deeper. If you plot the amount of available coal versus the effort you’re able to put into mining it, the curve should steadily rise, providing enough energy to support an industrializing economy for the decades it takes to figure out the use of other sources of fossil energy. Or nuclear energy.

      That certainly is no longer given in Britain, but if the collapse brings sufficient climate change, maybe there’ll be another coal-fuelled Industrial Revolution in Antarctica.

    3. Well, there are degrees of difficulty.

      When your main implement for getting coal out of the ground is “burly man with a pickaxe,” coal on the surface of the Earth is “easy” to get at, and coal 100 feet down is “hard” to get at. Because it involves battering through 100 feet of solid rock armed with nothing but a pointy bit of metal and a lot of cuss words, and fueled by the cheese sandwich the man brought from home.

      When your main implements for getting coal out of the ground are “thousand-horsepower diesel engines,” coal 100 feet down is “easy” to get at, and coal 10000 feet down is “hard” to get at. The work is now armed with excavators the size of office buildings that can rip the top off a literal mountain and casually dump it over the side into the nearby river valley, and fueled by the accumulated power of burning millions of years of dead bioaccumulated marine algae and occasional dead dinosaurs. As opposed to, y’know, a cheese sandwich.

      When someone says “the Industrial Revolution was motivated by the desire to get at hard-to-access coal seams, they are thinking in terms of “burly man with pickaxe” and “cheese sandwich” difficulty levels.

      When someone says “a second Industrial Revolution would be unlikely because all the coal is too hard to get at now,” they are thinking in terms of “thousand horsepower diesel engine” difficulty levels.

      Coal which is challenging to access even with giant mechanical excavators will be effectively impossible to access armed only with a pickaxe and that cheese sandwich.

    4. While I enjoy fiction, like The Mote in God’s Eye, it seems relevant that the “World War Four fought with sticks” has NEVER HAPPENED. There is a subset of sci-fi that loves this idea of “all technology lost,” we only have historical examples of narrow technological expertise being lost. And even in those limited cases (most examples I know of are the loss of metal working knowledge) the technology was recovered eventually. The “loss of technology” typically wasn’t a complete loss of knowledge, but a loss of specific expertise and/or institutional support.

      We also have extensive documentation ina number of more permanent formats. This isn’t trivial, as a number of modern formats (most digital formats) are actually less stable than older formats (books) but we also have archivists working on the problem. There aren’t a lot of such people, but we don’t need all our records to last forever, just enough to help future generations.

      I think there is value in thinking longer term, but stuff like leaving coal to start a second industrial Revolution makes no sense when you could just leave some metal plates describing how to build a variety of power sources. It would take a lot of such metal plates, but we can easily do it.

      1. Since we haven’t seen World War Three yet, it seems a little early to say that WW4 has never happened.

        On losing technology, the one usually referenced is the Late Bronze Age collapse around 1200-1150 BCE. It was bad! While the actual technological expertise might not have been permanently lost, for a couple of centuries everyday life for nearly everybody in the area had reverted back to a lower equilibrium agriculture or worse. So yeah we do have a historical example.

        There are also different expectations in the 21st C about how bad World War 4 could be. I’m old enough to remember the 1970s when the USA and Soviet Union each had many tens of thousands of nuclear weapons. The scale of destruction that could be expected was a lot worse than today when the number of nukes is a couple of thousand each.

        As for the metal plates, where are they? New Zealand is a popular choice in science fiction for the recovery of technological civilisation after the apocalypse. I know there’s a Long Now Foundation which includes archivists as you describe, but they’re in San Francisco.

        1. How bad the Late Bronze Age collapse was depends on where you were. Egypt had instability and lost the outer edges of its empire, but there’s still a continuity of culture. Greece went from a prosperous bronze age merchant/pirate empire to stone age farming, and even lost their entire language! I’d say that counts of “expertise… permanently lost”.

          1. Lost their entire language? They lost writing, because writing had only been used to serve the state, but they didn’t forget how to talk, for crying out loud. Also they didn’t lose metalworking; they used iron, which became common at about the same time as the collapse.

      2. Yeah, I doubt the post-post-apocalyptic types are going to forget something as basic and as useful as germ theory, or at least the idea that washing hands and cleaning wounds can lower risk of infections. And literacy is so widespread it’s hard to see how it can be lost across the entire world.

  4. Couple of pedantic nit-picks: hunter-gatherers shape the ecology too, often in quite drastic ways – they eliminate megafauna, use fire-stick farming, and encourage food plants in ways that push evolution towards human needs (eg larger seed heads).

    Second is that Britain was not the largest producer of textiles in the world. That was India, which had a lead in weaving and especially in printing and dying. Over the course of the 18th century Britain set out to supplant India in this trade (probably the single largest global trade sector), using tariffs, subsidies and control of maritime traffic. The application of steam power to spinning (and then weaving) allowed it to do so. So you can add an actively dirigiste state to your list of prerequisites.

  5. But why did it take the English to develop the spinning jenny and flying shuttle? Why couldn’t the Romans have done the same?

    1. The spinning jenny is an improvement on the spinning wheel, which was not invented until sometime between 1000 and 1200 CE. The spinning wheel is one of those commonplace but not at all obvious inventions that, once they were invented, everyone adopted them almost immediately, but someone had to think of the concept first. Like the horse collar or the stirrup or the wheel, the world had to wait until someone thought of the idea and actually made one.

      1. Spinning wheels actually took a long time to become unbiquitous. In some regions of France, for instance, they weren’t common until the 18th century! Similarly, while used for silk production in Italy quite early, they weren’t as commonly used for linen or wool for a couple of centuries.

        In part this was because of capital cost and the need for merchants who were willing to buy and then loan/rent them to their spinners or for the spinners to save up for them, but also because it took time for weaving technology to need such a fast rate of spinning that only mass use of spinning wheels could keep up with it.

      2. Or knitting, of all things: a cheap and easy way of making useful clothing out of yarn. Think knitting goes back to time immemorial? On the contrary, we have no evidence of it before a sort of proto-knitting called naalebinding dating to third century CE Egypt.

  6. I was struck reading this by my memory of the 18th Century Scottish thinker Lord Kames, living amid the earliest days of the Industrial Revolution, who postulated one of the first systematic economic-age based theories of human history. Kames’ model (in his 1758 book *Historical Law Tracts*) he laid out a four-stage model of human history, starting with hunting and fishing, then moving to pastoralism, then to agriculture.

    So far, so similar with the general model you shared. But in these earliest days of steam, Kames didn’t identify “industry” as a qualitatively different mode of human society. Instead, his fourth and (until that point) final stage was “commercial society,” born of trade, monetization, and new forms of contract law. For Kames and his followers (which included, ultimately, Gibbon), 18th Century England was a commercial society, but so were some of the most advanced classical economies.

    Obviously our level of historical analysis is much more sophisticated than Kames and his ilk (with the advantage, among others, of being able to build on them) but it’s an interesting prod to think of a different model that recognizes the Roman economy, while still pre-industrial, as nonetheless different in kind and more merely degree from less sophisticated agrarian societies.

    1. The features which distinguish a “commercial” society are primarily in the area of legal and social structures. Actual productive activity is not that different from what occurred in feudal societies. So these societies are really a subset of agricultural societies generally.

      What I find really useful in Bret’s article is its point that there was only ever one industrial revolution. Other “commercial” or advanced agricultural societies existed (Imperial Rome? Ming China? Byzantium? The Ottoman Empire?), but none of them had an industrial revolution. So the question is not, “What was wrong with Imperial Rome?” but “What was unique about Hanoverian England?” If it wasn’t coal, what was it?

      1. Despite the fact that limited examples elsewhere can be pointed to, movable-type printing only really took off in Europe where it sparked a “knowledge revolution” that as I alluded to earlier formed i.m.o. the indispensable base the industrial revolution required. Paul Kennedy in “The Rise And Fall Of The Great Powers” attributed this to the geographic and political diversity of Europe preventing the suppression of innovation by status quo actors. Now within Europe itself a good case can be made that England was the most favorably positioned for steam power to be implemented, but that’s comparing England to the rest of Europe, not the whole planet.

      2. What’s unique about Britian seems to be deposits of *really good coal* (anthracite, a naturally-low sodium deposit of coal. Most coal deposits in continental Europe are too high in sodium to use without coking them), a lack of other heating options (forests having been cut down for lumber, clear-cut for farming, etc), and an industry that required a lot of rotational energy (spinning jennys).

  7. Is it intentional that the question posed in the first paragraph is “Why did the Roman Empire **have** an industrial revolution”, rather than “Why did the Roman Empire **not** have an industrial revolution”? Seems like maybe a typo.

  8. I believe one other factor for Rome’s organic efficiency was the climate: From around the Punic Wars to the reign of Marcus Aurelius was warmer, wetter, and mostly stable than when scholars were looking at the problem. This made agriculture generally easier on harsher terrain that would later become borderline or non-viable. There was sources talking of vineyards creeping up mountainsides that today you’d have no joy with. This Roman Warm Period began to destabilize around the third century, before being truly gone by the time the empire split in two.

  9. Part of the story I have heard around the Industrial Revolution also deals with the availability of labor supply. There’s a contrast between 17th century Britain and the Roman Empire in this regard: Labor (in the form of slaves) is cheap and plentiful, whereas Britain is labor-constrained in many ways. I have heard this is related to the Black Plague reducing population density all across Europe, and most specifically in Britain where the plague struck many times, but I wonder how much of a reach that is.

    So how much does the relative supply of labor play into this?

    1. You briefly mentioned metallurgy, but i think this deserves more attention. Finding new alloys in an era before modern chemistry and crystallography was a matter of chance and expensive experimentation, and the power of heat engines depends nearly linearly (in the low-temperature early-development regime) on the maximum temperature* your engine materials can endure. The cast iron used in Watt’s engines has about 100C advantage in operating temperature over bronze, and even more so over the copper used in Newcomen’s engine, and that advantage only rose as known-but-expensive steel alloys came in economic reach.

      In an economy like Rome’s, where such strong and heat-resistant materials cannot be had for love or money, the steam engine is stuck at about 1750 levels of efficiency until the scattershot experimentation of pre-modern metallurgy comes upon better steels.

      * Technically on the difference between said maximum temperature and the environment

      1. I do think that one thing we start to, haphazardly, get in the 1600’s and 1700’s is kind of governmentally-directed R&D. Mostly for military stuff, but enough other stuff as well. It’s not just people experimenting to get new cannons, it’s the governent trying to systematically incentivize people to come up with new cannons. (at least when they can afford it, have the time, etc.)

        1. One institutional thing that the Englishmen had going for them was the absurd wealth of the rich landowners. Even the poorest peers of the realm were richer than almost any French magnates, amd the dukes and earls commanded wealth comparable to minor German states.

          Because they didn’t, unlike German princes, need to worry about fighting for their capital by military means, they could use it for different leisurely pursuits. They were, thus, able to undertake projects that usually only countries could do. You can’t spend it all on women, booze and neoclassical architecture, but every now and then, somebody pays for economically useful undertakings.

          1. One thing they did do was spin off their relatives into parsonages, where the combination of good education and a very light work-load led to many engaging in intellectual hobbies. The number of clergymen who made significant technical or scientific contributions is quite astounding.

    2. If Britain in c1700 were that labour-constrained, you might have expected it to import slaves, as so many of its American colonies did.

      Put another way: it seems unlikely that Britain invented the Industrial Revolution because it lacked access to slaves, at a time when it was buying and shipping to the Americas tens of thousands of slaves per year.

      1. Adam Smith, writing shortly before the American Revolution, said that wages were higher in the American colonies than in Britain. I believe this was the main reason so many British moved to America in that era.

        I’ve also read that more efficient farming equipment in Britain put a lot of farmers out of work, driving them to the cities, and that this glut of urban labor helped drive the Industrial Revolution.

      2. The usual argument is the reverse: The agricultural revolution had freed up agricultural labour (and also increased food yields) which meant there was a lot of cheap labour around that wasn’t doing anything/was willing to work horrible conditions for some kind of wage.

        It should be noted that the “mineral economy” did not immediately *supplant* the organic economy, but rather give new uses for it: The number of horses and oxen engaged in industrial work increased drastically for most of the 19th century, IIRC, and it wasn’t really until the modern automobile that they started getting phased out.

    3. Wages were high in England relative to other similar European countries. But the cause is unclear. Britain was open to immigration, and could draw on low-wage Scotland and Ireland. More likely it was that it was already an economy reliant on skilled trades (eg all the maritime industries) and a politics that mitigated the ability of the elites to suppress wage demands

    4. Don’t forget the enclosures, when the nobility forced many small farmers off the land to accommodate sheep, thus creating a whole class of landless unemployed. There were enough of these people that poor laws were passed against them throughout England. These people were available to become industrial workers.

  10. Ruth Goodman in Domestic Revolution does a marvelous job talking about the use of coal in homes.

    Such details as the developments in metalworking to let you deal, which were also a factor.

  11. This does make me wonder, if the Western Roman Empire had hung around as long as the Eastern one, would *Roman* Britain perhaps have met the requirements for industrialisation a few hundred years sooner than an independent Britain did? It doesn’t make sense for a pre-industrial society to ship firewood or charcoal across the Channel, I wouldn’t think, so Roman Britain in the mid-second millennium would still have a strong economic incentive to mine coal, creating the same pressures to delve greedily and too deep that led to the atmospheric engine.

    I’m also inclined to think that once the engine gets to a certain level of sophistication, people will inevitably start to wonder what else they can do with it. The flying shuttle and spinning jenny probably explain why industrialisation took off *so quickly* once the steam engine was invented, but I think as soon as you have a machine that can create rotational energy (which is desireable for a smooth pumping action even if all you’re doing is sucking water out of flooded mines) people are going to wonder if they can use it to push a cart. It just might have been a slower revolution than ours was without a more immediate use for the technology.

    1. I think pushing a ship with steam would have been more likely (after all, we see it happen). And it would have been useful. Ships carry a LOT of stuff compared to carts, and being able to ship more stuff per unit time (either by increasing size of the ship or increasing the speed of it) would more quickly provide return on investment. The application of rotary motion to this problem would be obvious to anyone who’s had to row–something every Roman sailor (merchant or navy) would be very familiar with.

      Rome offers another option for the application of rotary power: War. They had a sort of oversized repeating crossbow (I forget the name) that relied on a hand crank; speeding that up would have obvious benefits. One could also speculate about hurling war darts via a machine akin to an automated pitching machine. If the machine could be made small enough, the rotary motion would help a battering ram. Pour boiling water onto a steam engine and it doesn’t have quite the effect that boiling water on people does, after all.

      1. Not quite, because you need a lot of progress on the steam engine for any of these things to become viable. The actual history of marine propulsion shows us that, because a steam-powered ship has to carry its coal (unlike a sailing ship), its steam-only range is severely limited by thermodynamic efficiency. But spending a great deal of construction expense and carrying capacity for auxiliary propulsion is a non-starter. Thus the first steam-powered vessels were mostly harbor tugs. (Small steam engines were also used on the last sailing ships to replace human muscle at the task of pulling on the various lines that control the sails.)

        Steam-powered siege engines are, I think, not a possibility even in alt-history. Once you “solved logistics” with ships+railways, you win wars, period. By the time you have steam engines with the power-to-weight ratio to be worth hauling around for siege engines, you more or less have traction engines (“road locomotives”), and definitely have railways. (Well, assuming you aren’t specifying an out-of-step where coke-fueled blast furnaces and/or reducing steelmaking haven’t been invented yet.) This is on top of mechanical reliability issues, always a bane for field applications.
        (Minor detail: why bother with rotation for a ram? Just have the piston impact the wall — a steam hammer laid horizontally.)

        To OP: Some of the “progress studies” people have written about how, apparently, the idea of technological progress was not at all obvious to people who lived before (roughly) the Industrial Revolution. Part of the reason was that, before …science, basically, applications were not engineered from an understanding of principles — thus almost nothing was known about whether alternative configurations would work, or how efficient the current arrangement was. (Also, obviously, progress became common and observable enough that people realized it was a thing. This appears not to be a natural state of mind for humans — see all the people who are, in the 2000s, complaining about stagnation because they do not see progress, despite living in the fricking 2000s.) Thus it may have taken quite a long time until somebody went “hey, this small-building-sized machine could not possibly move itself, but attach a cable drum to it and it can pull carts along”. It took the Romans a few centuries to figure out that the rotation of a waterwheel can be transformed into reciprocating motion for sawing things (timber and marble).

  12. A fascinating look at the situation! I’ve always heard that the industrial revolution couldn’t happen in Rome because bronze wasn’t sufficient for steam engines, but it’s great to see further elucidation on the environment necessary.

    Are there any pre-mineral industries besides fabric that could have made use of early, inefficient steam engines?

    1. Any industry where lots of energy is needed at a single location* might do it, depending on the cost of labor or other materials, and whether the industry is big enough to have lots of demand. Possibly (based on my limited knowledge)

      -Metalworking, replacing strikers, bellows, and a few other roles. Was an actual use of these technologies in real life history, though I don’t know how fast they caught on. Might come in handy equipping lots of soldiers.
      -Maybe some shipbuilding tasks (it is all done in one place, though I don’t know much much of ancient shipbuilding lends itself to mass production)
      -Operating big ports possibly, if early cargo crane equivalents or such are worth it compared to lots of people doing the unloading or loading. Big issue is enough big ports existed for someone to think up something like this.

      *(Rotational or up and down movement isn’t particular important, once someone comes up with a good way to convert these types of movement, they can probably convert to any type of movement needed)

      1. The essential requirements of life are food, clothing and shelter. The Agricultural Revolution was built on the first of these. The Manufacturing Revolution was built on the second, the only one of the requirements of life to be a manufactured good. Few people have their house delivered from a factory.

        So I would think that in every alternate history, that Revolution was built around textiles and clothing.

        (Although if access to cotton really was that essential, I might have expected the thing to have started closer to some actual fields of cotton plants.)

        1. > Few people have their house delivered from a factory.

          These are called mobile homes. And even homes built on site are made from factory-made parts. (Lumber from an industrial sawmill, nails from a nail factory, etc.)

          1. And we’re just starting to see 3D-printed houses. Sure they look pretty ugly and I wouldn’t want to live in one at the moment, but maybe in a few decades we’ll all think weird that houses used to be hand-built one by one instead of sending over the printing machine and setting up a a hundred per week or whatever.

          2. I said few people, not no people, lived in mobile homes. Tens of millions of people is 1% of the worldwide population. A hundred times as many people wear clothes. Much larger market.

            (And how many people lived in a mobile home in 1700?)

      2. I do note that metalworking is one of those things that also requires a lot of coal (at least once you start getting to the 18th century) so they’re another industry that combines the two use cases, since you’re already bringing a ton of coal there, using it to make the process more efficient makes more sense than trying to apply it to something unrelated.

      3. I agree that having the industry that provides the need for power is important, but not with the specific examples you suggest:

        * I really don’t know much about roman metalworking, but I think the cost of transporting raw minerals and firewood in an ancient economy basically would have meant that they tried to build the processing plants near the mine itself, and use local wood as much as possible, and that, I guess, would have meant that metalworking operations were bound to be quite small-scale (at least relatively to what happened soon after the start of the industrial revolution), so an expensive and inefficient steam engine prototype might have been too costly for an individual metalworking operation, basically. But then, there were at least big-ish mines in the ancient world, so I don’t know.

        * I don’t know much about shipbuilding either, but you can’t use basic rotational motion to assemble a ship, and if you use that power to saw the wood for the ship, you’d better have an easy way to have an abundant supply of wood, again with ancient world transportation costs. And the main bottleneck was there, not on actually sawing that wood.

        * I… don’t know a lot about port infrastructure either, but a steam engine basically provides a rotational movement, and regulating precisely the amount of fuel that goes in to control the amount of power, or the means to direct that power with precision, would have been very hard to achieve. Basically, you don’t want to plug Watt’s engine to a port crane as long as you don’t have any control over the speed of the motion and you can’t stop it and start it again very easily and quickly.

  13. A thought that occurred to me while reading your description of the Industrial Revolution was that this was also a time of rapid scientific progress in thermodynamics, which started out as the science of making better engines. (Oddly, the most important names that come to mind, e.g. Boltzmann and Gibbs, were not Englishmen.) This work required an infrastructure of mathematical technique (algebra, calculus, etc) that certainly didn’t exist in Ancient Rome.

    That leads me to wonder about the causal relationship between thermodynamics and the Industrial Revolution. Clearly there is an arrow from the IR to thermo, but was thermo at all important in designing better engines in the early days? Nowadays it is, of course — any engineer who designs practical engines needs to understand thermodynamics.

    1. Thermodynamics wasn’t important to early industrialization, since the field developed over the 1800’s, but scientific developments from the 1600’s and 1700’s almost certainly are important, such as gas laws and force and motion physics.

      1. Well, then, those antecedents might then be seen as another of the reasons contributing to the impossibility (or unlikelihood) of an industrial revolution in Rome.

    2. No, thermodynamics was not an important contributor to the early Industrial Revolution. The reason the most important minds that come to mind in thermodynamics are not British (with the exception of James Joule) is that thermodynamics was invented as a science by Continentals trying to figure out how the British engines worked (Frenchman Sadi Carnot should be added to the list, too). British folks like James Watt worked based on an imperfect understanding of heat combined with practical observation and trial and error.

        1. Yes, he was, but thermodynamics was firmly established as a science by the time Gibbs made his contributions to the field. Note that Gibbs spent time in France and Germany attending lectures and interacting with scientists after completing his PhD in the USA.

      1. But, fun fact, thermodynamics were incerdebly important for developing more efficient oil, and gas lamps. The whole concept of a storm lantern was dowstream from understanding how this “fire” thing actually worked.

    3. Like folks above have noted, thermodynamics was not important for early industrial revolution. It was based on work and intuition of practical engineers. Scientific engineering was used only in building bridges, where you can work with mechanics, and even there, to limited extent.

      The scientific engineering didn’t have much role in England, because the local education system stressed humanities, and engineering was a middle-class pursuit. In the Continent, especially in Germany, things went the other way: scientific engineering became a respectable pursuit, and major industrial companies were (and very often, still are), managed by engineers all the way to the top. In England, “engineer” is still a job with some lower class aspects attached. In Germany and the Nordic Countries, it carries the sense of profession: a graduate engineer is socially on par with physicians and lawyers.

      This became important when you started the second phase of the industrial revolution: a practical engineer can design steam engines and even locomotives, but when you are doing serious metallurgy, steam turbines, generators and electrical power networks, you need scientifically educated people in charge.

      1. Many of which were pioneered in Britain. The ‘two nations’ thesis is overblown – Britain was a world leader in metallurgy, electronics and aviation (and much else) into the 60s.

  14. What is your opinion of Robert Marks’ “The Origins of the Modern World”? That was basically the only “grand narrative” we read when I was studying history last year

  15. Another reason why the Roman Empire probably could not have had an industrial revolution is that lacked a source of demand for mass-produced products. Industrialization is an uneven process. In its early stages, a few producers (e.g. spinners and weavers) create far, far more of a product than they could possibly consume themselves. This means that they must find a wide market for their wares. In modern times, this market is usually rich consumers in developed countries (though these likely buyers consider themselves middle class or even poorer, relatively speaking, they are rich indeed).

    In early industrial societies, other people are still producing goods at preindustrial levels, so they have limited capacity to absorb the industrially-produced excess. Fortunately for the spinners and weavers of Eighteenth-century Britain, brave sailors had been discovering consumers for over two hundred years. Much of the tropical and subtropical world was a market for British cloth. My favourite illustration of this demand is a drawing by Tupaia, a Tahitian navigator and artist who joined Captain Cook on his voyages. Tupaia depicts a Maori man and a British sailor (or possibly Joseph Banks). The Maori is exchanging a crayfish for some British cloth. The two men had just discovered each other, and already the Maori sees the value in some quality British threads. (The drawing can be seen here: https://en.wikipedia.org/wiki/Tupaia_(navigator)#/media/File:A_Maori_man_and_Joseph_Banks_exchanging_a_crayfish_for_a_piece_of_cloth,_c._1769.jpg).

    Where would a Roman industrialist find similar demand for his or her products? Could an empire of 45 million have absorbed manufactures on an industrial scale (compare this population to the 11 million in the possessions of the King of England in 1700, with access to a much larger world market)? Without such demand, what would be the incentive for industrialization.

    The connection between demand and innovation is contentious. Allan (2009) says that demand for an invention can lead to its creation. Mokyr (2009) says that invention on the scale of the Industrial Revolution requires a cultural shift (e.g. the Enlightenment) that is unconnected with material demands. This leaves the question: did the Romans not make lots of thread because they lacked spinning jennies, or did they lack spinning jennies because they did not need that much thread?

    1. Interesting, but I’m not sure I entirely buy this explanation. We see the Romans, for instance, mass batch-firing household ceramics for commercial production and distribution over large areas. The bottleneck was productivity, not consumers.

      1. Pottery breaks easily, making it hard to transport. It also tends to be cheap. Pottery in the Roman Empire was therefore generally produced and consumed locally, meaning that potters had little incentive to expand and produce at industrial scales.

        One of the few exceptions to this rule was the ceramic production centre at La Graufesenque, in what is now southern France. Pottery was mass-produced in kilns here and exported across the Roman Empire. However, it is likely that this pottery production was piggybacking on demand for wine produced in Southern Gaul. Empire-wide demand for wine led to great local demand by vintners for containers in which to ship their product.

        My source is Lewit, Tamara, “The mysterious case of La Graufesenque? Stimuli to large-scale fine pottery production and trade in the Roman Empire” in Fulford, M. and Durham, E. (eds) “Seeing Red: New economic and social perspectives on Gallo-Roman terra sigilata”, London: University of London Press (2013), pp. 111-120.

        The case of ceramics supports the hypothesis that Roman industrialization was limited by consumer demand, not productivity. The technology for mass production existed, but was only used in places were demand was idiosyncratically high. If demand was high for pottery for personal use (as opposed to use to carry some other trade good), it would have made sense to mass produce all pottery. Note that, though pottery is fragile, long-distance trade was still possible in pre-Industrial times, as demonstrated by the kilns of Jingdezhen, China.

        1. > Pottery breaks easily, making it hard to transport.

          Transporting pottery is easy enough that the Romans used pottery to transport other goods.

          Also, China exported a great deal of the stuff we now call china.

          1. Pottery requires specialized transportation (i.e. it generally must be moved over water, so your kiln needs to be near a river or the sea). Cheap pottery is worth transporting if you are using it to move a more valuable product across a sea, but not if you are transporting it for end use.

            Chinese porcelain was a luxury good and therefore worth transporting over long distances by sea. Real luxury goods are generally not amenable to industrial production because, if they are produced in large quantities, they stop being luxury goods. (I am making a distinction from mass-produced products that benefit from a brand associated with fine craftsmanship).

    2. To be able to sell products, you need people who will use the goods, who are willing to do some sort of exchange. A high population Roman Empire provides the people, they clearly would want textiles as otherwise lots of time wouldn’t be spent making them. some sort of exchange or equivalent is the big question, but if Roman farmers were doing at least some trade in farming products, olive oil, etc., than including mass produced textile products in that trade almost certainly happens if they are available.

      1. The question is whether you have a large enough consumer population that you can mass-produce a good in sufficient quantities that the start-up and transportation costs are lower than the costs of producing locally. Keep in mind that, in the case of the Roman Empire, a fabric that would be comfortable on the banks of the Nile might cause one to freeze in the Scythian frosts. The size of a factory’s market might be smaller than the whole Empire.

        1. Oh good grief. Many people in the ancient and medieval worlds lived in places where the temperature changed considerably from season to season, or even between day and night for example Scythia. They were quite familiar with the concept of layering: if you have thin fabric wear one shirt when it’s warm; an undershirt, shirt, and overcoat when it’s cold.

          As observed by Jared Diamond in “Guns, Germs, and Steel” and others Eurasia (which for Romans includes Africa north of the Sahara) has bands of similar climate that stretch very long distances horizontally. The market size for a particular weight of fabric might be smaller than the whole empire, but still very large.

          1. The point of the focus on steam as a source of power (and all the precursor technologies that enabled steam) is that it’s not enough to have a general demand. You have to have the means to meet it. Not just at the point of production, but also in transport, packaging and so on. Roman shipping was a very long way from 18th century European (or European mid-late medieval) shipping, Roman metallurgy was a long way back, timber cutting was a long way back and so on. Cotton goods from India were much sought after in Rome, but it was a luxury trade given shipping costs. Britain or France or the Dutch could ship cotton cloth at a price that ordinary consumers could meet.

          2. You raise a valid point. I was thinking in my reply of differences in relative demand for wool and cotton cloth in different regions during the Industrial Revolution. I wanted to work in a reference to the “I don’t want to be Caesar please…” poem, hence my example. In hindsight, it wasn’t the best.

            The point I should have made is that demand for a given product is not the same across regions, for reasons related not only to climate, but also to local customs and other considerations. What washighly desired in one part of the Roman Empire might not have been in high demand elsewhere.

          3. I love Jared Diamond in general, but like a lot of Big Thinkers he makes some very big mistakes as well. (I’m trained as a biologist, not a historian, and I picked up some glaringly obvious mistakes in the biology, both in GGS and in his more recent book The World Until Yesterday, which I also loved).

            The piece about latitudinal belts just made me chuckle: Diamond surely knows that his sample size is way too small to draw any such conclusions.

          4. @Hector_St_Clare, “latitudinal belts” is my no doubt poor re-phrasing of Jared Diamond, not a direct quote.

            The “sample size” in Guns Germs and Steels appears, again to me if not you, to be Eurasia, Africa, and the Americas over the past ten thousand years or so. That seems to be a pretty big sample size!

            I’ve also seen ecological maps of the world in a couple of books about evolution and the spread of mammals, before the Neolithic revolution, and those also stated that Eurasia saw faster dispersion of new species from being “horizontal” rather than “vertical”.

            If there’s an alternate explanation as to why domestic crops and domesticated animals spread so fast in Eurasia and slower on other continents, I’d be interested. Can you point me to sources suitable for a reasonably scientifically literate but not expert reader?

    3. Early industrial societies would also create markets at gunpoint. The British, for example, destroyed the looms of Indian weavers and imposed large taxes on Indian textiles to crush the local industry. With greatly curtailed domestic supply of textiles, Indians needed to buy cloth from their conquerors.

    4. I don’t think the IR was driven by selling cloth to Tahitians and such… There’s already high demand for domestic cloth, thus all the effort, as mentioned. Effort which people would gladly displace if you can provide cloth cheaper. Also we’ve see what happened as cloth became cheaper: people *used more cloth*: more flowing clothes, more outfits.

      (There’s also a military side: cloth armor, which I suspect was not particularly cheap: the equivalent of 15 or 30 shirts packed together is a lot. But if cloth is cheaper…)

      1. Tahiti on its own was unimportant, of course. The overall influence of long-distance trade on industrialization is a live subject of debate, with Mokyr and McCloskey arguing that industrialization was all about ideas leading to innovation leading to cheaper goods. Clark, O’Rourke and Taylor emphasize the importance of foreign markets to industrialization. They conclude a 2008 article noting, “In Smithian terms, in the nineteenth-century global ‘division of labor,’ it was the ‘power of exchanging’ that gave occasion’ to the Industrial Revolution. The highly specialized British economy was extremely dependent on foreign trade by the 1850s.” (see “Made in America? The New World, the Old and the Industrial Revolution”, American Economic Review: Papers & Proceedings, 2008, 98:2, 523-528 at 527).

        I am inclined to believe in the “exports are important” thesis because, over the history of industrialization, few countries have industrialized successfully without an export market. Numerous countries have tried “import-substitution industrialization”, in which tariffs are placed on foreign manufactures in the hope that domestic consumers will buy locally-made goods and kickstart industrial takeoff. This strategy has tended to fail because domestic demand is not high enough to make manufacturing profitable.* Other countries have pursued “Export-oriented industrialization”, in which companies are encouraged to seek opportunties in export markets (through incentives like subsidies and protection in home markets for companies that succeed abroad). This strategy has succeeded in Japan, Taiwan, South Korea and has been successful in China so far.

        *I understand that the United States may be an exception to this rule, industrializing based on domestic demand protected by high tariffs. However, the United States was relatively rich per capita during its period of industrialization.

        1. Britain used a combination of tariffs and subsidies to capture key manufacturing sectors which then were a base for exports (textiles is the best example, but raw materials processing from its colonies is another). France did much the same, but with more emphasis on its (larger) home market. Russia and Germany followed suit. A key part of the issue was getting enough foreign exchange to buy the technologies and skilled labour without those foreign earnings being diverted into luxury goods for the elites (the Latin American problem). As always, it’s less a technical than a political issue.

      2. An interesting cultural development was that once cloth became cheap enough, elaborate outfits eventually ceased to be the status symbols they had been when cloth was expensive, being replaced with emphasis on quality (fine tailoring, prestigious name brands) rather than sheer volume of cloth.

        1. Have clothes now ceased to be a status symbol at all? (At least for men.) Bill Gates and Marc Zuckerberg don’t dress any differently than an office messenger, in contrast to twentieth century captains of industry, whose clothes may have seemed drab compared to Louis XIV, but whose suits were made of fine wool, their shirts of quality cotton, and their ties of imported silk.

          1. I think rich people who aren’t at the very top maybe still wear expensive clothes. Zuckerberg doesn’t need to tell anybody that he’s rich; if you see his face, you know who he is, and everyone knows he’s rich.

          2. Gates and Zuck are in tech, which has long had its own particular disdain for conventional clothes signalling.

            It’s not like office clothing has lost all status. If you walk in somewhere and ask to use the bathroom, you probably have better odds in a suit than in some other outfits.

            Many people seem to still care about brand name clothing or accessories, or what Hollywood stars are wearing… Beyond that, I’m a techie myself, so my own impressions of “clothes don’t matter” are suspect as a generalization.

          3. “[R]ich people who aren’t at the very top maybe still wear expensive clothes.”–Not at my mid-tier law firm. The lawyers all make between $200K (our starting salary) and a few million, so rich but not at the top, but essentially they wear jeans or khakis (or sometimes athleisure if they are women). Nothing that a secretary or a messenger wouldn’t wear.

          4. Mark Zuckerberg’s hoodies are cashmere and cost thousands of dollars. I believe his t-shirts are custom-made for him by an Italian luxury brand as well.

        2. In previous periods, when the lower classes became able to afford better clothes for some reason (e.g. post-Black Death) the result was the institution of sumptuary laws. I think the difference is that this time around, society had been sufficiently commercialized for long enough that the English Civil War and “the victory of the middle class” had already taken place before the IR. Today, ridiculously-dressing dictators can still be found in a few poorly-commercialized places.

          1. And sumptuary laws generally made it easier to show off. Look I can afford the fine, too!

    1. This is quite true. It is an example often mentioned when discussing the concept “High Level Equilibrium Trap.” See Mark Elvin, The Pattern of the Chinese Past, C. 17

  16. (Reposting top-level)

    You briefly mentioned metallurgy, but i think this deserves more attention. Finding new alloys in an era before modern chemistry and crystallography was a matter of chance and expensive experimentation, and the power of heat engines depends nearly linearly (in the low-temperature early-development regime) on the maximum temperature* your engine materials can endure. The cast iron used in Watt’s engines has about 100C advantage in operating temperature over bronze, and even more so over the copper used in Newcomen’s engine, and that advantage only rose as known-but-expensive steel alloys came in economic reach.

    In an economy like Rome’s, where such strong and heat-resistant materials cannot be had for love or money, the steam engine is stuck at about 1750 levels of efficiency until the scattershot experimentation of pre-modern metallurgy comes upon better steels.

    * Technically on the difference between said maximum temperature and the environment

  17. I wonder if the pressure vessels necessary for liquid fire siphons, and the oil extraction necessary to produce the liquid fire itself, might not have provided an alternate route toward industrialization had not the unending pressures of constant existential warfare precluded the Medieval Roman Empire from prioritizing survival over other investment paths.

    1. Bah, that last bit I switched part way through. Read it either as “precluded [them] from prioritizing other investment paths over survival” or as “required [them] to prioritized survival over…”

    2. Probably not. That kind of early flamethrower was always a niche weapon for the Romans/Byzantines. They’re good on wooden ships, but you don’t need them to win a naval war if you’ve got good sailors and a prepared fleet of reasonable size. They’re kind of useless in land warfare, as a rule.

      To make them much more effective than they historically were, you’d need to upgrade to a mechanical pump capable of projecting a stream of liquid accurately for a distance of at least a hundred meters (read: outside enemy archers’ effective range). At that point you’re talking about the kind of performance you get out of a literal fire hose in modern times. Those use pumps driven by something like a 500 horsepower engine. They are far beyond the performance you can plausibly get out of an early steam engine installation, especially,/i> one that has to fit on a mobile platform such as a ship or inside a mobile siege engine.

  18. For alternatives to British textiles as a breakout application outside of coal mines: what about water pumping outside of coal mines?

    Specifically, I’m thinking of East Asian rice cultivation and its need (in certain strains) for regularly changing water levels. An engine on the efficiency of Watt’s could make this doable in a wider variety of terrains and, through labor savings, make more marginal land economically viable to farm.

    1. Water pumping is also useful for production of salt, although I don’t know if salt was ever valuable enough to justify the fuel cost (in the pre-modern world) of running the water pumps.

      1. Salt was VERY expensive in some pre-modern periods; not sure if the numbers work out, but still.

        1. Salt either came from mines [Salzburg, Austria had a Roman antecedant] or from evaperative collection from coastal salt pans improved with low dykes

          1. I was thinking of the salt pan method specifically: I’ve seen modern-day salt production (in developing country) involving pumping seawater up to saltpans above the high tide line and allowing it to evaporate. Presumably an early industrializing society could have used steam pumps to move the seawater, but I’m not sure the economics works out.

          2. As noted, an operation like this isn’t entirely out of the question, but you have a few problems.

            1) Your basic business model is “use Newcomen engine or whatever to pump seawater into elevated salt pans, then evaporate.” The question you have to ask yourself is, why not just build the pans surrounded by dikes that rise above the high tide line, and let the tides bring the water into the salt pans for you? The Newcomen engine is a very complex and expensive device requiring a constant stream of maintenance and fuel input, while “wait for high tide” is the easiest thing in the world and costs you nothing. If engines are already cheap and widely available, this changes the equation, but if engines are rare, expensive, and usually useless unless you need them literally on top of the place that extracts their fuel, you’re out of luck. Unless someone builds an engine that burns salt, of course.

            2) The seashore is usually not home to massive old-growth forests. If it was, opportunistic loggers would have come by in ships and cut down all the trees for masts centuries ago. The available firewood near your salt operation will be exhausted in short order, and then your business model collapses because you can’t function without the engine, whereas your neighboring competitor half a mile down the beach (who relies on tides to bring water into his lower-lying salt pans) is doing fine.

            3) The seaside and constantly handling large amounts of salt water will make for a fairly unforgiving operating environment for your engine. It’s gonna break a lot. Marine steam engines (especially once we step away from fresh-water riverboats) were much slower to become reliable than stationary engines on land or even on rails. There’s probably a reason for that.

      2. I believe “a man worth his salt” dates from the Roman Legions. Marching armies require salt. At least some Legionnaires were either paid in salt or given a salt ration called a “salarium”, from which we derive “salary’.

        (Failed to come up with a decent pun involving salt on celery.)

        1. I think it comes more from food storage . You use a lot of salt to store the food for the voyage or for the winter, and that salt is expensive, but only a part of the cost of the worker’s upkeep paid in natura. Being “worth one’s salt” means that you have at least worked to recuperate the part of your pay going for storing the food. So, I would tend to take it is an absurd exaggeration, like many folk idioms (like “halfpennyworth of brains”)

      3. In the times of the most famous Polish king, Casimir the Great(Kazimierz Wielki), a single salt mine in Wieliczka provided 1/3 of the state income.

        1. Ooh I’ve been to that one! Absolutely stunning place. Not many places in the world that have a cathedral carved into a mine, complete with the Poland-ubiquitous statue of Pope John Paul II.

          Seeing the scale of it, I can see how it could have supported a significant chunk of a state (although I expect it was smaller back then).

    2. Two points:
      1. As Devereaux pointed out, pumping water out of coal mines wasn’t just an ideal use case for early steam engines because it was pumping water, but because fuel was abundant in a way it wouldn’t be most other places.
      2. I don’t get the impression that rice cultivation needs water levels to be changed rapidly, just precisely. If nothing else, I doubt anyone in the premodern could breed a strain of rice that couldn’t easily be cultivated by muscle power alone.

      1. Oh I meant not as a substitute for that first application (coal mines), but as a substitute for the second application (textiles).

        On point 2: traditional rice cultivation doesn’t move water with a lot of power, because the only use for a pump (moving water *up*) gets so very power-intensive so very fast. Most water pumps in traditional paddies are lifting water a meter or two at the point of use, from an unpumped shared water source.

        Modern rice farming can use much more land and is much less vulnerable to variable water levels because water can be pumped up much higher terrain. Even a Newcomen engine could, over a not-ridiculous period of time, irrigate a terraced hill without an uphill water source or an aqueduct leading to one; in this sense, the capital expense and fuel are competing with either losing the land entirely, or with the enormous capital and maintenance cost of an aqueduct.

    3. The problem is that you are proposing to solve the problem of “farm some additional marginal land” (the kind of thing a bunch of desperately poor peasant farmers on the margins of the cultivated zone do) with “buy and maintain and operate a steam engine.” The steam engine requires an extreme investment of metal (very expensive per unit mass in this era) and metalworking skill to create and maintain, and a constant stream of fuel that must be somehow extracted by mostly-manual labor. By the time you factor in the amount of work required to run the steam engine, extract its fuel, transport its fuel, mine the ores that go into the metal, extract the fuel to smelt those ores and work them into the engine…

      …It is probably considerably more efficient to just say “ah, to heck with it” and have all those laborers just haul the water up the side of the hill with shadufs or something. Irrigating any remotely practical area of rice paddies just isn’t going to justify the expense.

      Furthermore, the engine represents a single point of failure for the entire agricultural operation; if the machine breaks down and cannot be repaired in a timely manner, the whole village starves. There is no comparable failure mode for “a bunch of guys manning shadufs,” because they are much easier to repair and any one breaking down is not a disaster.

      What you need is a situation where the engine’s power output can be directly converted into large amounts of cash by selling its products. Because access to cash means reliable access to specialists who can repair the thing when it breaks down, and access to cash savings that will keep you from going out of business while that happens.

      Which is why we see steam engines first applied in capital-intensive businesses (mines, textile mills, railroads along heavily traveled freight routes, ferryboats) and in applications where a breakdown didn’t stop you from getting the job done (auxiliary propulsion on ships with a sailing rig).

      1. If the land is only marginal because of lack of water (as a lot of land was!) and you have large landowners with lots of capital producing for the market (as many were in ancient Rome and China!) then this is very much doable.

        All the other prerequisites in terms of market development and metallurgy still need to be met, this only handles the issue of a second market beyond coal mines for funding further development of steam engines. If you don’t have a well-developed trade system you’re still boned.

        1. Something’s being left out here. Remember, the question of “what will the second market for steam engines be” isn’t just “what can a steam engine do?” It’s “what can a steam engine do more economically than a bunch of guys with hand tools?

          This is where applications like mine pumps, locomotives, and steamships really shine. Human beings and draft animals are not a scalable alternative for pumping out mines when they get deep enough. No reasonable number of humans or animals can consistently haul an umpty-ton load along a railway at five miles an hour all day. No number of human rowers can consistently propel a cargo ship upwind all day; you can build a ship powered by oars for the purpose but by the time it has enough, you have used up most of your cargo space and have effectively turned the ship into a vehicle for delivering the rowers and not much else.

          Textile mills were a viable choice for this in England, a country with nearly unique advantages in terms of “fiber to be made into textiles comes to us cheaply,” to the point where farming it out to thousands and thousands of women with spinning wheels became less productive than farming it out to a steam engine running a bunch of spinning jennies supervised by several dozen workers.

          But pumping water for irrigation is not an example of this kind of application, until engine technology becomes far more mature. We can tell because historically, combustion engines were not used for irrigation pumps until much, much later in the Industrial Revolution, long after they had been made more efficient and relatively cheaper to operate, so as to power ships, locomotives, and where appropriate factories.

          If it is 1775 and the state of the art is a slightly refined Newcomen engine, there are simply not going to be any landowners, large or small, who will be interested in buying a Newcomen engine to irrigate their rice paddies. They will instead buy the services of a couple of oxen to turn a capstan, or just enlist the aid of a bunch of tenant farmers and day laborers to form a bucket brigade or something. The engine’s power of extremely steady mechanical nonstop work doesn’t make it pay, not at that level of technology.

          1. Let me restate.

            The Industrial Revolution can, loosely speaking, be characterized by several “generations” of steam engine advancement, each of which made it economical to use steam engines to mechanize certain industrial tasks that were hitherto uneconomical.

            The following is a crude summary and outline, which may not be perfectly accurate, but hopefully at least vaguely maps to reality

            The first generation of engines (e.g. the Newcomen engine) was uneconomical to use in almost any application that wasn’t sitting literally right on top of a high-density fuel source, and was thus mostly limited to pumping out coal mines, or other mines that happened to be close to a coal mine, or maybe hauling heavy objects up a steep incline using a winch or something.

            The second generation of engines (e.g. Watt’s 1775 engine) were more efficient and reflected experience and iteration on the designs of the first generation. These were useful in a wider variety of applications, though in practice they were still too inefficient to power a mobile platform to a useful standard of performance. But here, you start to see steam engines appear as stationary engines powering various mills and foundries.

            It takes further significant improvement (various inventors, c. 1800) before you start seeing engines that are reasonably practical for mounting in a mobile platform such as a locomotive or steamship. Loosely, we might call this the third generation of engines.

            Advancement doesn’t stop there, and a more qualified person might list many generations of engines, or subdivide my own categorization more finely. But this basic dynamic is what we need to look at.

            I do not think that irrigation pumps are a good alternative for “so, what would be the source of demand driving development of second and third generation engines?” Because that demand already existed, and historically did not drive such demand. Other options were available for irrigation pumping, and remained more cost-effective than engine-powered pumps until engines were many generations more advanced than the early ones we’re discussing.

            Now, pumps for canal locks, that’s a different story- because you need to move much greater quantities of bulk water for a canal lock, and speed matters because the ability to pump water uphill into the lock is a key determinant of the throughput and capacity of the canal itself. For irrigation, you’re working with less water and less of a hurry, as a rule… Or you’re irrigating such huge areas at once that “just have a bunch of peasants and oxen do it” is a perfectly cost-effective solution.

    4. You know how the Netherlands has a lot of windmills? Most of them were not for grinding grain but for lifting water — for drainage rather than irrigation, but same difference. (Using Archimedean screws, because those have no piston seal than needs precise manufacturing.)

      1. Given the intermittency of wind, it seems useful for non-urgent things, like pumping water, over a dike or up to a reservoir. Be rather awkward if a miller has a line of people waiting to grind their grain but no wind, but pumping could be averaged out over days.

        Hmm, in the hypothetical Electricity Revolution, maybe it’s less for hydropower (already used efficiently mechanically, though wires are easier to string around than rods and belts) but as a way of handling wind. Though that requires batteries better than wind-pumping water to a reservoir, and storing electricity is a hard problem for *us*. Though they would have lower quantities to deal with.

        1. Grain keeps very well, so as a miller, it’s really not a problem if you have a line of people with grain on a day with no wind. You just smile, nod, and take their grain. Then you write down the amounts, stick the baskets of grain in the storage area in back, wait for a windy day and start working while telling little Timmy to go run and get So-and-So, their flour’s ready.

          It’d be a problem if they wanted their grain ground into flour today, but that would be silly, if they want their grain ground right now they can damn well do it themselves on a hand quern.

          1. (Clarification: I am obviously not a medieval miller; I’m trying to adopt the viewpoint of one. Importantly, medieval millers were not harassed retail clerks. They had standing within the community and were usually in a good position to say “my way or the highway.” Furthermore, they lived in an era with a very different sense of expectations regarding time than we have.)

          2. And were notoriously corrupt. There’s many a case where the lord forced peasants to send their grain to the mill when they had and preferred hand querne

          3. Even so!

            And when the local lord is forcing all the villagers to go to the designated and approved miller, while prohibiting the villagers from organizing their own alternatives or contesting control of the mill along more populist lines…

            Well, the miller has even less reason to worry that he can’t hurry up and grind the villagers’ grain right now on account of the lack of wind.

          4. Did millers have to return a particular batch of grain back to the farmer as flour, or could they just keep a small stock of flour on hand and swap out?

      2. Speaking of windmills and scientific knowledge underpinning industrial advancement: I find it remarkable that the classic Dutch windmill or the steel windmills later used on farms in the USA used simple slat blades with no apparent knowledge of airfoil design- even though nature provided examples like maple seeds that were far more aerodynamic and would have produced better power.

        1. Possible factors:

          1) Ease of maintenance. The slat style is relatively easy to build and easy to replace; it does not require wood to be warped into a precise geometry in a molding loft in order to work- which was a very real problem with the design of historical wood-and-canvas wings for aircraft!

          2) Technological conservatism. Building a windmill is a lot of work, and building new blades is a lot of work. If you screw up the blade geometry it may be useless or at any rate less productive than a slat-blade windmill.

          3) Lack of mathematical grounding. You can’t do aerodynamics without calculus, so any attempt to predict which blade shapes would work best would involve very closely mimicking something like a maple seed. Except that your windmill operates under design constraints and serves functions that don’t precisely mirror those of a maple seed’s gliding wings. So a direct copy may prove very counterproductive, as per (2).

  19. Why didn’t the Romans use coal in any sort of extensive way? I remember your series on iron production, and that you needed roughly 100 tons of wood to eventually produce 1 ton of iron products. That involves clearing huge amounts of forests, transporting all that wood, possibly overland, to where you’re doing your smelting, (and in practice, that means doing your smelting in the middle of forests, which consequently limits where iron mining can be done), and seems like a massive hassle. I don’t know much about comparative fuel efficiency for producing heat, but I would have thought that just mining coal and burning it is way more efficient from a given weight of fuel than chopping down trees. And the Romans definitely had incentives to produce more iron. So why did coal lose out so badly to wood and wood derived charcoal?

    1. For very high heat applications like forging or smelting, mineral coal won’t do the job unless it is ‘coked’ – raw mineral coal has too many impurities and so doesn’t typically burn hot enough. The related problem for metal-working is sulfur content; even fairly limited amounts of sulfur in coal used for metal work will utterly ruin the iron as sulfur causes all sorts of problems in an iron or iron-carbon (read: steel) alloy.

      Finally, mining is hard and labor intensive, more so than cutting down trees (which conveniently grow everywhere whereas coal deposits only occur in some places). Thus we see at most limited use of coal.

      1. Why is it that Britain became wood-poor and Italy did not? Was Italy just more heavily forested to begin with, or is it the mildness of winter? Certainly it can’t have been until the 17th/18th century at least before Britain rivaled Italy for population density.

        1. Problem is, Italy does basically no coal production in the modern era, and I’m not aware of evidence that there were ancient seams of coal that were exploited in antiquity, though I am hardly an expert like the good Dr. on that. (They do seem to have found coal in Roman Britain, and possibly the Rhineland/Saar deposits as well? But not in Italy itself.)

          Something like 90% of all coal that has ever or will ever exist comes from one particular geologic 60 million year period (roughly the Carboniferous) between plants evolving lignin and bacteria and fungi and other decomposers figuring out how to process it. If you don’t have rocks from that particular era that are easy to get to, coal is not easy to get to.

          1. Side note: the entire reason we call it the Carboniferous era is that it is literally the “coal-bearing” era of Earth’s geologic history.

            Early British geologists called certain types of rocks “Carboniferous” because they were the kind of rocks you found coal in, long before there were even coherent theories about how old those rocks were. Once it was realized that all these rocks dated back to roughly the same period of the Earth’s history, that period became known as the “Carboniferous period.”

            So the geologic age is named after the coal, its most distinctive associated geological feature.

        2. It’s likely that Italy became wood-poor in the early period Roman Empire, or possibly even before. Classical civilisations used timber for everything the way we use plastic.

      2. Additionally, while charcoal or even wood doesn’t produce enough smoke to foul food (and what it does produce often enhances the flavor), smoke from raw coal would make food absolutely inedible. Coking allowed coal to be used for energy intensive food processing like brewing and distillation.

    2. Coking. Coal usually has high sulphur, which makes very brittle and poor quality steel. Coking (anaerobic decomposition like charcoaling), reduces the sulphur content.
      Coppicing can also regeneratively create lots of wood rapidly without deforestation.
      Also not all coal is created equal.
      Anthracite (the og black gold), the highest quality coal (almost pure carbon, highest calorific value) has very low sulphur and is sometimes even preferred in high alloy steelmaking.
      Coal in British isles is mostly anthracite. A lot of the rest of Europe however is lignite or bituminous coal, which are very smoky, oily and high sulphur. (Most polluting, least energy, opposite of anthracite).

  20. One nit, you don’t HAVE to pump the water to the surface of a mine in a single stroke. It’s certainly desirable though

    Multi stage pumping was known and used though clearly it is more complex and requires running pumps inside the mine

    1. Of course that line was a bit of a simplification. So each pump needs to move all of the water in its system in each movement. Now you can – and they absolutely did – create multiple pumps where each one feeds into a pump system higher up, but if you are relying on muscle power to work multiple vertically stacked pumps you now have the problem of getting that muscle power to the pumps, either by moving the muscle into the mine (not great for a space with limited air supply) or with a transmission system that introduces its own problems.

  21. The only other place I can think of with similar potential conditions to the British Empire is ancient and medieval China. Archaeology is showing significantly more coal use than previously believed, but on the other hand they didn’t deforest to the extent as the UK because prior to the arrival of the industrial revolution, abundant surface seams were sufficient to supplement wood fuel and later bamboo charcoal (and bamboo is fast growing and naturally invasive, so you will not run out of charcoal easily).

    The other key factor which is mitigated is textile production. While silk is not the only textile produced in China, sericulture is significantly less bottle-necked at the thread production stage – base fibers don’t need to spun. So you have at least somewhat less pressure on textile production, perhaps just enough to make a difference?

    Someone mentioned irrigation of rice fields as a potential application of steam pumps, but it doesn’t follow. Rice cultivation doesn’t typically require continuous alteration of water levels – it’s an annual or semi-annual flood and drain cycle that doesn’t need to be done all that rapidly, so gravity and mill systems can and do suffice.

      1. SRI is ultimately a mono-culture approach (and one which strongly encourages the use of proprietary seeds…), and like you say, I’m not sure it is overall more beneficial than rice-and-fish coculture, even if I accept Cornell’s claims about yields (which I don’t – their methodologies are questionable and there’s a dearth of peer-reviewed scholarship and independent corroboration).

        But also, the conditions that led to developing SRI don’t exist in the same place where e.g. abundant coal exists, so the precondition would still be missing.

        https://link.springer.com/article/10.1007/s13280-022-01711-5

    1. Wasn’t silk just for rich people, though? I don’t know what ordinary Chinese people wore, but I don’t think it was silk.

      1. Yes, only the aristocracy could wear silk until the late medieval/early modern period, although they were happy enough to export significant quantities on the silk road – if it hadn’t been a “cash crop”, it may well have been used more widely instead. Silk was also used by just about everyone in non-textile contexts as well. Which is why my mulling includes that silk may have *lessened* the pressure, not eliminated it.

    2. Beat me to it. Why Britain and not Ming or Qing China is actually the more difficult question than why Britain and not Rome, for some of the reasons you just cited. While I favor environmental answers by training, I suspect here that politics might have played a bigger role, with the Qing dynasty (Manchus who stuck themselves with ruling a really big Chinese population) enforcing a different set of priorities than did the British government (expanding colonialism in the New World).

  22. > there needs to be massive demand for spinning (so you need a huge textile export industry fueled both by domestic wool production and the cotton spoils of empire)

    Agrarian societies commonly put enormous effort into obtaining spun fiber. The demand was always there. The practical ability to export industrial quantities of cloth and import the raw fiber was not, it happened under more or less the same conditions as a large empire, and Rome qualified. If Rome had somehow developed industrial spinning, the textile trade would have followed.

    The industry need not be in the seat of imperial power, even though it was historically. It only needs bulk sea trade with the imperial market.

  23. It’s possible that it could have occurred with different technologies and resources, though I have to admit I haven’t seen a plausible alternative development that doesn’t just take the same technologies and systems and put them somewhere else.

    I’ve seen plenty of arguments that a slower industrial revolution could be achieved with non-fossil-fuel sources of energy, particularly hydropower (and occasionally geothermal or crude nuclear). This is generally in the context of a post-post-apocalyptic world where industrial technology was lost or alien worlds which can’t access fossil fuels for some reason, not alternate history; any nascent hydro-industrial revolution would almost certainly be outcompeted by OTL-industrial technologies using mineral coal.

    1. If you want to see a hydropower industrial revolution, look at early-industrial Britain. They used hydropower for *everything* — coal-powered factories didn’t really take off until every stream and river was clogged with water wheels.

      Thing is, there’s only so much hydropower available. Once you’ve dammed all the worthwhile water sources, you’re stuck: you can’t increase your economy’s energy budget until you develop some other source of energy, and the other non-fossil-fuel sources of energy are generally quite high-tech. Geothermal requires the ability to drill extremely deep wells, photovoltaic requires absurdly pure silicon, and nuclear requires a massive industrial base. You might be able to get somewhere with solar thermal, but that’s not a technology that’s seen much development, so it’s hard to say how effective it would be.

      1. I almost thought you would be talking about https://en.wikipedia.org/wiki/Hydraulic_power_network

        Nuclear doesn’t necessarily require a huge industrial base. Unenriched uranium and appropriate carbon-bearing compounds (not just graphite) can produce a chain reaction, the only problem with this setup is that it’s particularly susceptible to misbehavior (see Windscale). Going from the discovery that uranium is fissile in 1938, or that plutonium is fissile and can be produced from uranium in 1940, to chain reaction in 1942, to boom in 1945 is what takes industry.

    2. Any re-industrialization in the wake of an apocalyptic collapse would be aided greatly by people remembering or reverse-engineering basic facts about how machinery and electricity work, and having access to large quantities of material from the fallen civilization (especially metal) that could be reworked.

      A giant steel machine may be nothing but a rusted mess 100 years later, but from another perspective, that rusted mess is a “natural” outcrop of high-grade iron ore.

    3. Beat me to it. A quick search says that Vitruvius definitely had access to hydropower, and Hero mostly used wind for his machines. (Not to mention that a sail-cart does seem rather more obvious than a steam train.) Though even crude nuclear would only be discovered by accident, and I’m not sure how much you could get out of ancient solar-thermal.

      1. Hmm. How old is passive thermal hot water, i.e. having a dark water tank on your roof?

        Of course that requires having a way to get water up to the tank, whether by aqueduct or water pump (if you’re near a river) or wind pump or muscle power.

      2. Second thought: Nuclear and solar-thermal both run into a lot of the same problems as in the article, namely that they’re both steam engines with a heat source other than coal. So even if you can manage to concentrate sunlight or refine fuel somehow, that’s a limit on what you can do with them.

        1. Assuming that A. you’re only looking for steam engines and B. there is no situation where a steam engine powered by nuclear/solar would be helpful. Nuclear should be self-explanatory, since you need to mine for uranium; primitive photovoltaic cells have existed since the early 19th century, don’t seem to rely on post-industrial technologies, and are handy with all sorts of electric devices (some of which were experimented with even earlier).

          Come to think of it, a photovoltaic industrial path, while barely plausible, seems like it would be fun to speculate about. Starting with (weak but precise-ish) electric currents before (crude but powerful) angular motion would lead to a way different revolution!

          1. Are we seriously talking about a pre-industrial civilization somehow implementing nuclear (!) energy or solar-electric energy rather than burning coal in a first-generation steam engine? Regarding primitive photovoltaics, if you’re talking about things like selenium photocells, my understanding was that those could modulate an electric current by their variable conductivity, but did not actively generate voltage. Until the very first semiconductor solar cells became available, space pioneers were still planning to implement solar-thermal generators that boiled mercury up to the late 1950s.

          2. 1. You don’t need refined uranium for nuclear power, you just need radioactive material and some way to use the heat it gives off.
            2. I dunno about selenium photocells, but either a couple gold plates immersed in “an acid, neutral, or alkaline solution” can generate an electric current when exposed unevenly to sunlight or Edmond Becquerel is a liar. (Electrochemistry is not my strong suit.)
            3. Those options were identified as rarer, less plausible suggestions upthread. The post you’re responding to was specifically a response to the idea that they have the exact same requirements as coal-powered steam engines, which they do not. Context matters, my dude.

  24. You seem to be mangling the early industrial history. Textile mills were powered by water for 15 years before they were powered by steam. You don’t need to invoke “serendipity” to justify coal being used in textiles, it’s simply a matter of local availability. Some parts of England were both sheep pasturing and coal mining country but didn’t have good water power, meaning that there was a niche for steam mills to compete as an alternative to the water mills. In the US the coal producing regions didn’t produce wool or cotton so textiles remained pretty much entirely water powered for almost a century. There’s a reason why Lowell, the first town built specifically for a textile mill, was located in an ideal location for water power and nowhere close to coal country, even though steam engines were quite well known at that point.

    There is a reoccurring fallacy that crops up in alternative energy discussions: people tend to forget that while chemical energy is often extremely abundant, it’s expensive and inefficient to translate into what you actually want. If you need kinetic energy, starting with kinetic energy will be both far more efficient and far less capital intensive then translating chemical energy into heat then translating heat into kinetic energy. Coal is great for steel because there you are looking for a new form of chemical energy (reducing oxidized iron to make it pure for steel) but it’s not important for textiles and it’s not strictly necessary for transport.

    (Regarding transport, early railroads were horse powered, the B+O railway was only 24 years after the steam locomotive was invented on another continent and it was not remotely connected to coal regions and the canal boom predated the Newcomen engine).

    We only have one example as you pointed out and in the one example we have coal was neither a necessary nor a sufficient condition. Hydropowered machinery was the necessary and sufficient condition. Roman cities had access to quite a bit of water energy in the form of aquaducts but they didn’t use it at all. They didn’t have the spinning wheel let alone the flying shuttle. It seems they didn’t even have good wagons based on what you’ve written in other articles. If the challenge was moving past industrializing textiles to other things, coal would start to matter but the Roman’s weren’t even close to industrializing textiles so coal is entirely besides the point.

    1. Bret not only knows about the textile watermills, he incorporates them into his explanation.

      Realizing this, textile manufacturers (we’re talking about factory owners, at this point) first use watermills, but there are only so many places in Great Britain suitable for a watermill and a windmill won’t do – the power needs to be steady and regular, things which the wind is not. But the developments of increasingly efficient steam engines used in the coal mines now collide with the developments in textiles: a sophisticated steam engine like the Watt engine could provide steady, smooth rotational motion in arbitrary, effectively infinite amounts (just keep adding engines!) to run an equally arbitrary, effectively infinite amount of mechanical spinning jennies, managed now by a workforce a fraction of a size of what would have once been necessary.

      To borrow the structure of Devereaux’s argument: The watermills encouraged the development of spinning machines that could be turned by non-human rotational power, which could be combined with extant steam engines into a product that was needed in places that didn’t have convenient rivers. This encouraged further steam engine refinement, etc.

      There is a reoccurring fallacy that crops up in alternative energy discussions: people tend to forget that while chemical energy is often extremely abundant, it’s expensive and inefficient to translate into what you actually want. If you need kinetic energy, starting with kinetic energy will be both far more efficient and far less capital intensive then translating chemical energy into heat then translating heat into kinetic energy. Coal is great for steel because there you are looking for a new form of chemical energy (reducing oxidized iron to make it pure for steel) but it’s not important for textiles and it’s not strictly necessary for transport.

      That’s not how efficiency or chemistry works. Chemical energy isn’t better at making chemical changes; plus, since the chemical energy is being converted into thermal energy which is converted into chemical energy (in the form of de-oxidized iron and oxygen), steel-production doesn’t involve any more conversion steps than a steam engine (chemical —> thermal —> kinetic).

      More important than that lack of understanding in how coal power produces steel, kinetic-to-kinetic translation is also inefficient. Entropy doesn’t give a shit how we classify different kinds of energy! Depending on the mechanical systems involved, kinetic-to-kinetic translation can be near-100% efficient or horrifically inefficient. What matters isn’t the form energy takes, but the process that turns it into useful work.

      TL;DR: You don’t seem to understand how coal power, hydropower, or even energy work, and you don’t seem to have even clearly read the article you’re responding to. I would kindly ask you to make sure you understand the topics involved in an argument before responding to it.

      1. “Chemical energy isn’t better at making chemical changes”
        “Entropy doesn’t give a shit how we classify different kinds of energy”

        Entropy is not the only means by which energy is classified. While I was studying this stuff in college for instance there was quite a bit of dreary time spent on reactions that were not prohibited by entropy but were prevented by enthalpy. If you make even the most cursory glance at modern efforts to reduce energy usage you will see a major focus is on electrification because electric processes can often displace processes that are reliant on the chemical bonds in fossil fuels at several times greater efficiency, reducing overall energy demand.

        You should maybe take a good hard look at your own advice.

        “that was needed in places that didn’t have convenient rivers”

        Production wasn’t displaced from the initial mill towns on the water until the 20th century. Once again if the question was “why was there not a followup to the invention of industrial textiles” this problem would be relevant.

        1. I still have no idea what your argument is.

          Like, let’s start with the first part. It’s wrong to say “entropy isn’t the only means by which energy is classified,” because entropy isn’t a way energy is classified. It’s a word used to refer to a wide variety of effects which prevent any process from being 100% efficient, because some energy is always lost to heat (if nothing else).

          Or let’s look at what you say about enthalpy. “We spent a lot of time on dreary reactions that were prevented by enthalpy.” I’m going to guess you’re referring to standard enthalpy of reaction, which can be described as either the difference in chemical energy tied up in the reagents and products of a reaction, or the energy released/absorbed by a reaction. Obviously, if the product of a reaction is more energetic than the reagents, you need to add energy to make the reaction happen. I’m still not sure what that has to do with efficiency of different kinds of energy!

          And perhaps most importantly, these points don’t seem to have anything to do with each other, other than using questionably-relevant scientific terminology.

          I could go on. I could question who’s reducing energy usage by electrification, for starters. (There are plenty of places where electrification is being used to reduce greenhouse gas emissions, either through green energy or by simply burning the fuel at larger, more efficient generators, but that’s obviously not the same thing.) Or I could point out that even in mill towns, watermills are useless in the 90% of the town that isn’t actually next to the river, and that there’s a variety of reasons both economic and physical you can’t simply cram more waterwheels into the available river. But I’m not convinced that my arguments would be met by anything logically sound, let alone anything with a coherent thesis beyond “You’re wrong, dumbass!”

    2. Kinetic energy can be more efficient than chemical energy, but it is NOT transportable. That’s why we use chemical energy in the form of firewood, coal, petrol, … (And batteries probably count too.)

      Hydropower, a watermill or today the turbine(s) in a dam, provides wonderful kinetic energy for someone right next to it. Anyone even a couple of hundred metres away is out of luck if you’re relying on mechanical linkages. Your own example of railways shows this: they are horse powered at first, stay horse powered until steam engines are small and efficient enough to move themselves thanks to a portable chemical energy source coal. Today we do have hydro powered electrical trains, but only because we’ve developed electrical generation and transmission technology. Cars, aircraft, and (maybe soon) ships will use batteries.

      It is possible to imagine an alternate history where electricity developed before the steam engine, but it seems unlikely. We human beings have been burning stuff to do work for a very long time, and there were plenty of precursor technologies that involved finding more energy dense fuel (charcoal) and pumping air and fluids around (bellows). Burning coal to generate steam was the easier “tech track”.

      1. Actually, “electricity before the steam engine” might well be one of the most plausible ways to get an industrial revolution without easy access to fossil fuels. That sounds like it might be worth exploring.

        1. Yep. In that case, the route would be: small desk-top demonstration toys -> electrical generators on water mills powering motors working winches on upslopes of horse-drawn railways and hammers at steel mills. That would be the use case transferring the use of electricity from university labs to practical uses in metal industry. (Assuming no spinning jennies around.) It would require 1860’s technology level, though.

      2. I do agree with your argument, but I would like to point out that kinetic/mechanical energy IS transportable, at least for shorter distances by using a flatrod system (https://en.wikipedia.org/wiki/Flatrod_system). This was used in mining at least in Sweden and Germany before steam power was invented (and in some cases for quite a while longer). The efficiency was pretty bad, and the maintenance needs were huge, but at least you could transport the power from a watermill a couple of kilometres to the mine.

        It doesn’t change the fact that energy is far more moveable in chemical or electrical form, but it’s a fascinating piece of obsolete technology. Then again, I find all obsolete technology fascinating…

        1. The flatrod system is just a complex series of linkages. It’s neat and cool to look at, but it’s not fundamentally different.

    3. >(Regarding transport, early railroads were horse
      >powered, the B+O railway was only 24 years after
      >the steam locomotive was invented on another
      >continent and it was not remotely connected to
      >coal regions and the canal boom predated the
      >Newcomen engine).

      Actually, the Baltimore and Ohio Railroad was explicitly planned to connect Baltimore (a major port city and potential industrial hub) to the Ohio river valley. Along the way, it would be passing through some of the largest coal fields then known in North America.

      https://upload.wikimedia.org/wikipedia/commons/b/b7/Cumberland_coal_trade.jpg

      By 1842, roughly 14 years after work on the line began (and most of that time was spent building the line), the B&O was already moving coal from the Cumberland area. Up to that point, it wasn’t as close to coal as might be desired, but the availability of coal was definitely a major factor in the thought process of its investors.

      Moreover, early railroads adopted steam engines after- and only after- James Watt and others had dramatically improved their efficiency. The first steam locomotives emerged roughly 90-100 years after Newcomen, and roughly 30-40 years after Watt’s work. The railroads represent one of several groups that opportunistically seized on the steam engine’s potential to do work for industrial applications after the technology was already available and, if not “mature,” at least “half-grown.”

      1. That was not a major coal mining region in 1828. The coal mining followed the transportation network, not the other way around.

        1. So what? The coal deposits were already known to exist, even if they couldn’t be exploited economically for lack of transportation. This was the 1820s, and western Maryland and Pennsylvania were, while not strictly on the frontier, still sparsely populated regions undergoing intense migration and economic development.

          The people who invested in the B&O were not immediately expecting to profit from access to cheap coal. But they knew cheap coal would be in the railroad’s future long before it reached all the way from Baltimore to the Ohio River, so it was factored into the railroad’s long-term economic outlook.

          In the meantime, of course, the B&O would have had to rely on other sources of fuel. But this wasn’t as much of a problem, because Baltimore was a major port, making importation of coal fairly practical, and also because the Eastern Seaboard was at this time far more heavily forested than Great Britain, making wood fuel for steam engines more practical. By 1828, engines had become significantly more efficient, meaning that they could be economically used even if fuel wasn’t maximally cheap. And, again, the project was begun with a long-term outlook, as was relatively common in those days.

  25. When I was (very strongly) considering being a historian, this was what I wanted to look at. I didn’t at the time, but I would now phrase the question that dominated my first two decades like this:

    In 1790, when George Washington was President, future president John Tyler was born. Because he remarried late in life and one of the sons from that marriage also remarried late in life, one of John’s grandsons is still alive (or at least was in 2020). In 1790 almost all information traveled at the speed of either muscle or wind (1 bit messages could travel faster, with sufficient state capacity to maintain, e.g. https://www.youtube.com/watch?v=i6LGJ7evrAg but that’s very rare). So I was fascinated by the question of how, in the course of just three human lifetimes, we go from all communication is by muscle or wind power… to walking on the moon and TikTok.

  26. Two thoughts, not necessarily in conflict with this article.

    1. An abundance of slaves supplied adequate energy to power the Roman agricultural economy.

    2. The European scientific revolution of the 17th century was a necessary precondition to the technological advances of the industrial revolution.

    1. As noted in another comment above, the British did, in fact, have access to slave labor. It’s just that they only practiced slavery in the colonies, evidently seeing no use for it at home.

      1. Slavery of on an industrial (heh) scale is largely down to *expensive* labour costs, not cheap ones: The opposite of 18th century britain (where the agricultural revolution and rapid population growth had created a lot of laborers that *could now be fed* but weren’t actually useful in agriculture, and were willing to go do horrible jobs in industry in order to get a meal for the day)

        Meanwhile in the colonies the constant problem was lack of labour: When these things started in the 16th and 17th centuries you weren’t getting mass immigration, the natives were dying, so importing slaves made sense, despite the cost (because you could produce massively in-demand cash crops)

        1. Perhaps it must be factored in that fair-skinned people from an island as far north as Labrador do not tolerate manual labor in the Caribbean sun very well.

          1. It’s not just the climate; it’s the malarial mosquitoes. West Africans have some degree of genetic resistance to malaria. Englishmen do not- to the point where malaria (“ague”) was a serious problem even in swampy parts of England, despite it not being a climate we normally think of as conducive to the flourishing of mosquito populations.

            You can trace the line between “areas where slave-plantation agriculture thrived to the point where slave-plantation owners could credibly threaten secession during the American Civil War” and “areas where that wasn’t a thing.” It’s a pretty accurate match for the border of the areas where the nastiest varieties of malaria-carrying mosquito were found in the American colonies circa 1800.

            South of the line, you had endemic malaria and slave plantations. North of the line… well, there was still slavery, but the economy didn’t wind up revolving around it as much.

            While we’re at it, to some extent “enslaved West Africans are better at doing hard manual labor in the Caribbean sun than Europeans” might be a perception created by ruthlessness. The owner of a Caribbean sugar plantation could and did work enslaved West Africans to death so fast that the average life expectancy was only four to seven years. Doing the same thing to white indentured servants would have gotten the owners in a lot more trouble

          2. On the contrary it was notorious that they were much crueler to white indentured servants than to black slaves. The limits of the investment were the driving factor to be sure, but race did not restrain them

          3. “Perhaps it must be factored in that fair-skinned people from an island as far north as Labrador do not tolerate manual labor in the Caribbean sun very well.”

            Is this even, really, true?

            Some African peoples certainly have some resistance to tropical diseases due to having coevolved with them (Europeans have better resistance to some other diseases, like i think plague), but I don’t think that has anything to do with skin color.

            And given the rate that slaves in Haiti died, I don’t think they were “tolerating” the slave labor very well either.

          4. That’s part of it (well, malaria resistance more than anything else), but it’s also that when the slave economies is beign constructed europe was in a population crunch: The 1500’s was still recovering from the 1300’s-1400’s collapse, and recovery was slow. (partially due to the effects of the little ice age) the vast hordes of european migrants of the 19th century simply didn’t exist yet.

          5. Direct effects of fair skin under tropical sun:

            At a guess, constant sunburn might impair the effectiveness of your workers. But that’s just a guess.

            I assume sugar island slaves don’t live long enough for skin cancer to be an issue.

            Dark skin both absorbs and radiates heat more than light skin; I don’t know the net effect on heatstroke. Maybe radiation matters more, since you’re generating more heat by working hard?

            Correlates of dark skin in this case:

            Being more used/acclimated to hot and humid conditions to begin with, than someone from Little Ice Age England.

            And yes, disease resistance. Not just genetic, but selection bias: most people taken as adults from west/central Africa would have already survived malaria and yellow fever, and thus reached accommodation or outright immunity.

          6. Doesn’t really matter, does it? If mass immigration would have been attractive, they could have attracted volountary African workers.

          7. @Mary

            Just to make sure we’re comparing apples to apples, when we talk about white indentured labor and black slave labor, we are specifically talking about Caribbean sugar plantations in both cases, correct? Because I was speaking within that context when I made the remark.

            I’m not saying you’re wrong, I just want to make sure we don’t accidentally a fact true of, say, Virginia tobacco planters c. 1650 and generalize it to Haitian sugar planters c. 1775.

      2. On the contrary, they saw the use for it alright. But they had a very early version of the Lincolnian anathema for slaves displacing working freemen. The courts said that Englishmen (here meaning geographically not ethnically) were meant to be free and had been so for many generations. In the famous “Somerset’s Case” the judge famously took care to point out that this principle did apply in other geographies, pointedly meaning Britain’s Caribbean sugar-plantation colonies.

  27. What about electricity? In parallel with the development and commercialization of the steam engine, Enlightenment-era scientists were discovering the basis of electromagnetism. Could there have been an alternate industrial revolution with no steam, but with electric motors providing the rotational energy needed to revolutionize production? I don’t know if it would work in an alternate history where Britain was missing one of one of the prerequisites for our own industrial revolution. Historically, most electricity was produced using coal and steam engines, and maybe the watermills and windmills that could have produced electricity instead would not have been enough.

    I think that, even if my electricity-only industrial revolution could have worked, it would still face the same barriers as our own history’s coal-and-steam industrial revolution: prerequisite technologies and economic conditions need to be in place to incentivize each step. If an ancient Roman Faraday invented the electric motor and electric generator (perhaps after a century of ancient Roman scientists tinkering with electricity and magnetism), would there be an economic use case that would make the electric motor a way to start industrialization? Would there be such a use case in more recent times, if the Enlightenment occurred on schedule but Britain was missing a prerequisite for the invention/commercialization of the steam engine?

    1. Electricity has to be generated. Initially that meant using a steam engine to turn a dynamo to send an electrical current down a wire; electricity was thus often simply a means of taking power from a steam engine in one place and using it in another place.

      Indeed, most of our power production still works this way, just using more sophisticated fuels than coal.

      1. AFAIK most early power plants were hydroelectric, so it may have happened without coal. And Italy is a pretty good place for hydroelectric plants. Also, electric generators and electric motors are more or less the same thing, which would have helped developing uses for electricity.

        However, while this is one important component of being able to move energy between two places, the other component for doing that with any kind of efficiency is being able to make long wires, and I’m not sure whether a proto-industrial society would have been able to make those in a way that the effort was worth the gain.
        After all, if hydro power is widely available, one can just use it directly, which is what people in Italy did widely at least in the middle ages and modern era.

        Other early applications of electricity, such as light, also require materials (a clear glass container that can sustain a vacuum, carbon filaments) that were available in an industrial society, but I’m not sure how easy they would have been for the Romans, and whether they could have been considered useful enough to develop the technology itself on their own.

          1. Yes, but the wire in a coat of mail is wire that’s going to be snipped up into bitty little pieces and twisted into rings. If any single ring snaps, it’s a mild inconvenience, not certain death, because a weapon is being kept out by the surrounding rings and the underlying padding too.

            So the wire in a mail coat doesn’t have to be extremely consistent in thickness. And you don’t have to make mile-long strands of it with confidence that it won’t break under its own (supported by poles) weight anywhere along the line. Not necessarily suitable for electrical applications.

            It’s sort of like how the Romans could make pipes out of ceramics or lead, and probably out of bronze if they cared to, or even iron if for some bizarre reason you could get a blacksmith to do that… But that doesn’t mean they could make sturdy pressure tubing with the constant diameter needed for a steam engine’s cylinders.

            The idea of an “electricity without steam” industrial revolution is actually interesting- in that the desire for more versatile electric generating apparatus would probably motivate the desire for steam engines. But it would probably be a slower road than the historical path of “steam then electricity.”

          2. Romans made mail by hammering round bar-stock down or by cutting plate into strips and hammering them round. Tedious, and does not produce uniform wire. Wire-drawing is using mechanical power (can be animal, but often water) to pull the bar through successively smaller holes in iron plates. Earliest known use of the technique is 8th century.

          3. @Simon_Jester you don’t need uniform thickness wire to create a usable electrical motor/generator system. You just need the thinnest section to be thicker than the minimum required to carry the current, which can quite easily be discovered by trial and error. Remember, we don’t need to equal the efficiency of modern electric power to be viable. We need to equal the efficiency of the hideously inefficient early steam engine (plus, you know, all of the other prerequisites…).

            You also don’t need aerial lines at all, they’re just a modern method of routing electrical power without going to the expense of digging trenches to lay the wire in.

      2. I mentioned in my comment that in our history, electricity was mainly generated with coal and steam engines.

  28. Typo note:
    > very specific pre-conditions which were really on true on Great Britain in that period.

    “Only true in Great Britain” is the version you wanted, I think.

    Otherwise, a very solid essay as usual.

  29. This is definitely a interesting topic. There are other societies that also reached the edge of modernity but without tipping into industrialization. China had periods where all sorts of machines were invented and used, but never quite scaled the wall. Edo era Japan was extremely modern in a lot of ways, even having the world’s biggest city, but it took the military threat of the west to force industrialization.

    I think there is something interesting about agriculture. The snowball effect of urbanization and scientific advancement seem to me to be contingent on advancements in mechanization of agriculture, though this seems to follow rather than lead to some degree. But even the industrial revolution followed the renaissance, enlightenment, and colonialism (with the columbian exchange of crops); perhaps there has to be a base of conditions before you get to even the idea of mass machine labor.

    As to the extreme particulars, what you said about industrialization only happening once creates a certain sort of idea that it could only happen that way, which I think is a bit of a stretch. But of course that’s one of those things we can never know. Perhaps the efficiency of rice agriculture is why it didn’t happen the the east in the first place – China, India, and Japan had high urban populations at various times, and a pretty high knowledge base (Edo Japan was very literate by the standards of the time), but perhaps because the didn’t run into the manpower and environmental restrictions of Britain didn’t have the need to industrialize. Japan actually did have forestry issues but the government stepped in and prevented the sort of deforestation that happened in Britain. Ironic to think this may have prevented industrialization.

    1. Yea when I read this piece I was thinking specifically of Japan: I remember reading somewhere that in some ways they were quite “developed” on the eve of the industrial revolution, in terms of literacy rates, nutrition levels, etc.. I wonder if the industrial revolution hadn’t started in Britain, could it have eventually got started in Japan if you waited another century or two?

  30. Nice, something I’ve been reading a good bit on lately! There’s one book in particular, Andreas Malm’s 2016 book “Fossil Capital,” which covers most of what I’ll write here, which is well worth anyone’s time who’s interested in the Industrial Revolution.

    To the question of whether any other place had all the ingredients for an industrial transformation, there is one place: Northeastern China.

    There, as in England, there was a population concentration that had outstripped the wood supply, and had turned to the abundant coalfields to supply heating needs. The state of development of coal mining was at a high degree, and the technical arts around it, mining and metallurgy, were in a state roughly on par, and surpassing in some cases, the state in Europe in the 17th century. (Including in cannon.) Also, crucially, the textile trade was at a high state of development, with raw cotton fiber being shipped in great bulk from the northeast, in Henan, Hebei, and Shandong down to Jiangnan (area south of the Yangtze, around Shanghai, Hangzhou, many other large cities) for spinning and weaving. For a European equivalent, think taking the whole wool output of England, shipping it to Hamburg, Bremen, Lübeck, even Konigsberg, all those big Hanseatic League ports, and only then spinning and weaving into cloth (then shipping it all the way back for a good chunk of the cloth!).

    (I just found this nice summary paper about Ming / Qing – era cotton trade – https://www.lse.ac.uk/Economic-History/Assets/Documents/Research/GEHN/GEHNConferences/conf8/PUNEZurndorfer.pdf – if anyone is interested in reading more)

    So, all the *technical* pieces are in place. But we had *The* Industrial Revolution, not two of them. What was different? Let’s just pick a year as a kind of snapshot to get a vague idea of the era preceding the Industrial Revolution. Choosing 1650 – in England, there was a rather weak republic – the Revolution (also called the Civil Wars) were being wrapped up with the Parliamentarians on top, the king executed the preceding year. In China, the Qing state had taken the imperial throne six years earlier, and were slowly mopping up Ming loyalists for the next twelve years. So both places had large political changes through very significant military action going on. That isn’t a difference. What is different is the nature of the states which had been taken over by new masters at this point. There is a world of difference between what the Chinese imperial state could accomplish and what the English state could do. The English state was barely capable of keeping an army paid, but the Qing had inherited from the Ming (even after the damages of war) a civil service capable of maintaining tribute and taxation over a region with a similar population to all of Europe. (Though the Commonwealth in England would be short – lived, the size of the state apparatus would not wildly change with the Restoration and subsequent Glorious Revolution of 1688, especially compared to the Chinese imperial state.)

    This is important for one crucial reason – in England, the state was alternately powerless and disinterested in maintaining the bulk of the population on the land, as a peasantry. By contrast, this social stability was an explicit goal of the Ming and following Qing, even as their internal dynamics became more and more commercial, with merchants rising to very high status. (From an already high status, well above anything seen to this point in Europe. We are talking about the place where banknotes were first issued, as early as the 11th century! To regard China as somehow culturally foreign to merchant activity on a grand scale while portraying any European locale as inherently mercantile is to ignore world history before the past few hundred years.) To summarize – in England in the broad period both sides of 1650, there formed a large and desperate working class shorn from any personal means of support, while in China the state managed a similar (but not identical) set of contradictions, keeping the old social systems in place with gradual changes. The work done by English laborers, of whom a large part, by my example year of 1650, were properly working-class in the modern sense, shorn of their own productive property, is being done in China by people largely stable in their social situation, possessing by custom or ownership that which they work on, accompanied by a small (and intentionally being shrunk as part of Qing policy) underclass.

    We get a final confirmation of the importance of social decisions to the question of an industrial revolution from Andreas Malm’s work that I mentioned at the beginning. In what I think is the most thoroughly argued part of the work, he demonstrates that at *no point* during the transition from water power to (coal-fired) steam was energy from the fossil source cheaper per unit energy. Rather, the more expensive form of power won out due to the fact that this expensive form of power could be moved directly to the concentrations of population. Why is this important? We return to the social dynamics – water-power, abundant in Britain (especially in the north of England, Scotland, and Wales), is distributed widely across the landscape. When the workers are distributed to match, each factory became its own little village, quite far from other population centers. This is not a problem for moving inputs and product (they’re on a waterway after all, it’s fantastically cheap to move the goods) but it is a huge problem for trying to squeeze more work out of the workers. If the workers organize for any purpose, the boss has a very hard time replacing them, as the surplus potential workers are all quite far off. So, the advantages that English capitalists possessed in having that large and desperate working class were being negated by the requirements of their abundant and cheap energy source. Over the very broad century around 1750-1850, the factory owners bit the bullet, paid for expensive steam power, and moved their operations into city centers, where they could find an easily disciplined surplus of the poor to be their workers. The increased reliability of the labor supply (not the power supply!) and the slow technical improvements to the engines eventually give steam dominance.

    By 1839, when Britain would force opium into Chinese ports by naval gunfire, steam had become technically well-developed enough that the gunboats that fought for Britain fought under steam power. (Though they still sailed for transit.) The Industrial Revolution had properly happened in Britain, but not in China, despite all the *technical* ingredients being present for it to happen. Northeastern China had cheap coal, well-developed mining industry, high technical proficiency in metallurgy, a booming textile trade in cotton (not to mention silk, which had been a global export commodity for two millennia from this region!), and plenty of hilly and mountainous places which could provide water-power. It did not, though, have the Industrial Revolution. The decisions of the people in each place, with the hand that they had been dealt by history to that point, made all the difference. The backwater English kingdom, pressed on all sides by rivals flush with cash from colonial looting, had only a weak state with which to preserve its social order, and developed a great mass of poor and dispossessed people, fertile material for textile manufacture along capitalist lines. (Which slotted neatly into England’s already prominent place in pre-modern world of textile trade.) The Ming and later Qing, having much more ability to chose, chose to preserve their social systems, in turn loosening the pressures to develop the technologies we are talking about now, by avoiding the social situation where they would develop in Britain.

    (As a last technical note on the main bottleneck in textile production, spinning – it seems to me that spinning wheels were at a higher state of technical development in China than in Europe, maybe all the way through their replacement by the spinning jenny. I saw references that Chinese spinners were working on wheels with multiple spindles for cotton from the 14th century onward. This may have reduced the pressure to develop the jenny or some equivalent, on top of the different social situation.)

    1. Were that true, the capitalist who turned to steam would have been outcompeted by those who stuck to water

      1. I believe what I have here mostly agrees with what you’re saying, but there’s a bit more nuance.

        That is: “the capitalist who turned to steam would have been outcompeted by those who stuck to water,” because, for a good part of a century, they did. Enormous water-power textile mills were being installed (and were often the state of the art, driving the general prices for goods they made due to their superior economics) into the 1850s in Britain (and the US. The Lowell, Massachusetts mill in 1850 generated 10,000 horsepower all from water, with 10,000 workers to match!). They truly did out-compete steam, when they worked. Those last three words are the key – “when they worked,” and needs an extra word to get the meaning in: “when they *were* worked.”
        This brings up a business concept which is crucial to the actual practice of business, but I never was taught in school: downside risk. Downside risk refers to what happens to an investment when things don’t go right. How much money still comes back? If the answer is “nothing,” that’s an investment with a high downside risk, and a business owner needs a whole bunch of ready cash in case that worst case happens. On the other hand, if the answer is “you just get your starting money back,” that’s a low downside risk, and the owner needs little cash on hand to cover what might happen. Ready cash is not productive – money just sitting there cannot be generating profit. So, investors and entrepreneurs always pay great attention to their downside risks and minimize them whenever they can.

        I’ll give an example, *prices fictional, but in the general ballpark,* of what a business owner would see during this period, to show why downside risk might be far more important than pure profit:

        Say you’ve got a large textile mill, designed for two hundred workers. Assume there are:
        -300 working days in the year, each worker gets paid 1 shilling (1/20 of a pound) daily. This means they get each paid £15 yearly, so the whole year’s wages come to £3,000.
        -The mill apparatus, prime mover included, and building costs £30,000, and will be worn out in 10 years. (Some parts will last longer, others shorter, but we’ll average it all out.) We’ll also say the installation cost of waterwheels and steam is the same, just to simplify things. The only difference will be that the water comes free, and the coal doesn’t.
        -Each worker runs machinery that consumes 1 horsepower.
        -For coal, We’ll use 10 s (£½) to the ton. (This is much over the cost at pithead, which is about half that for the period in question, but coal is bulky.)
        -The steam engine, which delivers 200 horsepower, burns 283 kg each hour. (Weird number, but it’s close to the real efficiency of a single-expansion engine of the time, and makes the math work well.) Working days, for simplicity’s sake, will be 16 hours, which isn’t very unusual for the period. This means that the engine is burning five tons of coal a day, or £2/10-, the same as 50 extra workers’ wages, or £750 yearly.
        -Your material inputs less fuel – all the cotton you’ll spin, for instance, costs £30 daily, or £9,000 each year.
        -The finished goods shipped daily net £60 when sold to the wholesaler. This gives a return of £18,000 each year.

        Ok, let’s compare two identical factories like that, but one went with the coal, and one went with the water:
        -They both have costs of £3,000 in wages, £3,000 in physical plant (averaged over ten years), £9,000 in materials, which comes to £15,000.
        -The coal-powered one pays an extra £750 each year for the coal, for a total of £15,750.
        -They earn £18,000, so the profits are:
        -For the waterwheel, £3,000, for a yearly return on investment of 20%.
        -For the steam engine, £2,250, for a yearly return on investment of 15%.

        So why would anyone go for the steam engine? The waterwheel is way better! A 5% difference in rate of profit is utterly enormous, even for an era when rates of profit were usually quite a bit higher than they are now.
        Let’s look at what happens when the shops are *not* working, though:

        If there is a strike at the water mill out in the countryside, then it might take months to either find new workers or talk down the existing workers and get them back to work. Let’s say it takes three months to do this. What happens during and after that time? The costs for the physical plant aren’t going away. There’s probably a loan taken out to pay for it, and maintenance on the building can’t stop. So, the plant is losing £10 daily, for a total of £750 in the red at the end. Then there’s the profits that weren’t being made, since the plant was stopped for a quarter of the year. That’s £750 again. Already, the yearly profit is down to 10%. Now, consider the purchasers – contracts are drawn up sometimes years ahead. Some of your customers aren’t going to come back now that they consider you an unreliable supplier, and some of them may take years to return. (Not to mention your suppliers, who also might not come back.) You’re probably in the red for a year, and profits won’t be back to the normal 20% for a few more years.

        Now consider the coal-fired steam mill in the city. There’s folks desperate for work on every corner, and many of them have some textile experience (remember, this is England we’re talking about). The same strike happens, and a new workforce can be ready in, say, about two weeks, one twenty-fourth of the year. The plant could lose money way faster, by still buying all the inputs, coal and fiber, like normal, paying £42/10/- daily, or £531/5/- total, and still come out ahead of the losses suffered by the water mill. They’ve avoided angering their suppliers by remaining steady customers. Also, they may be able to dodge the loss of customers, as their finished products will only be two weeks late, and they may be able to catch up with a bit of speed-up. The loss of profits will be only £93/15/-, together with the extra supplies they bought, that’s £625 in total losses. They’ve kept their reputation as both a reliable supplier and customer, and come out way ahead of the water-driven mill, even in total profits for the year.

        So when it comes to actually doing business, downside risk often carries the day.

        I hope this example is helpful for seeing how the cheapest option isn’t always the best from the capitalist’s standpoint! It’s considerations like this that gave steam the edge it needed to compete with the cheaper water-power, and take over.

        1. Maybe. My understanding is the Lowell mills drew girls from all over New England, and had high turnover as the girls returned home to marry. So replacing strikers wasn’t necessarily a problem. The boys in Bret’s picture don’t look irreplaceable either. I would need to see some fairly granular labor histories of Britain and New England to be convinced that reduction in the power of organized labor was a big motivator for steam power.

          A slightly less woke but more plausible interpretation might be that it was just generally easier to recruit workers in cities than in isolated rural areas next to powerful rivers. Obviously labor force availability is always a consideration in siting factories.

    2. In China wouldn’t the coal mine owners (state-owned?) have an incentive to develop steam engines to pump water out of the mines like in Britain?

    3. > The backwater English kingdom, pressed on all sides by rivals flush with cash from colonial looting, had only a weak state with which to preserve its social order

      England was no backwater at the dawn of the Industrial Revolution. It was a major colonial power, with a strong enough state to rule over distant and enormous India. Devereaux mentions British control of India as a contributing factor to industrialization.

      1. Sorry, was more than a bit unclear about time period for that comment – that was referring to the situation about two hundred years prior, in the era that led into my example year of 1650, the era where England developed a working class (1450 – 1650 or so). At that time, Spain is the dominant imperial power, they nearly invade England with the Armada.

        You’re quite right that, by the time of the Industrial Revolution, England is a major colonial power. I would actually go further, to say that it was the dominant colonial power at that time. A *lot* changed in the preceding era.

    4. That makes a certain amount of sense, but one question this leaves me with is why, if the government isn’t actively maintaining the bulk of the population on the land, lots of the peasants migrate to the cities even though they just end up poor and desperate there. This is especially puzzling if there are these water-powered mills scattered all over where they could be working. Someone elsewhere in the thread said that pre-industrial improvements in agricultural efficiency had disrupted the rural population and driven them to cities, so maybe that’s yet another prerequisite for an industrial revolution?

      1. Apparently pre-industrial farming could in many circumstances be an extremely poor existence. There’s an apocryphal saying that it is better to be a beggar on the streets of Bombay than to be a farmer.

        1. Though some of that desperate poverty was probably elites coming and taking half your food… I suppose an urban beggar might benefit from the urban elites being the ones taking the food from a large area, and thus having a more stable food supply than any one peasant.

          1. Farmers are a lot easier to tax to be sure. Plus a beggar might be starving poor but at least not have to perform heavy manual labor in the bargain.

    5. To describe the post-Cromwell English (and later British) state as weak is nonsense. England had been the tightest-governed state in Europe since 1020 or so, and became much tighter in the Commonwealth and later. It was able to collect taxes, direct expenditure and control its population in a manner that only Prussia could rival (and then on a much smaller scale, in a poorer state), and to a degree a Chinese government could only envy. And from 1690 to 1815, engaged in what its elite saw as an existential struggle with larger France, it was very dirigiste – spending large sums encouraging technological development and investing in development.

      1. The post-Cromwell state was strong, the pre-Cromwell state was weak. It’s exactly Cromwell (and the generations after) that change this, so that by the 1700’s britain is one of the, if not *the* strongest state in Europe.

        But the preceeding era? Saw england being regularly outcompeted across pretty much all areas of state competency, by the swedes, by the dutch, and by the french. it’s only *after* the Civil War that changes, and the english start catching up, and then rapidly overtaking, their contemporaries.

        1. England was relatively strong 1050-1450, a bit messier under the Tudors, strong enough to field a major naval effort and fight off Spain under Elizabeth, had a down-turn under James and Charles, recovered under Cromwell, took a rest under Charles II, then took off again.

          1. The Tudor era was a pretty darn messy period, Elizabeth was a great propagandist, so it tends to obscure how relatively limited the english successes were.

            During and after Cromwell the english largely went from strength to strength, (though they could take advantage of being relatively safe from french armies, unlike their main rivals in terms of trade and finance the dutch, and a lot of dutch financiers ended up setting up shop in England anyway after their Stadtholder invaded and seized the throne)

          2. Doubtless Bret would say that real historians distrust broad narratives, but this is basically 100 percent correct.

  31. typo: The summarize

    One guess I had for an alternative take-off path was putting natural philosophers discovering electricity and its uses, and then hydropower being used to generate that. Wouldn’t give you as huge a take-off, though it would be a cleaner and more sustainable one! Though there’s also the use of coal in blast furnaces to make lots more steel, for rails and such.

    Rome may not have been anywhere near the spinning path but it seems at least like they could have been; after all the spinning wheel and flying shuttle loom were ‘organic efficiency’ improvements themselves.

    OTOH, at an abstract level, what steam engines brought in was heat engines: the ability to convert heat differences into work. This is a new thing! And then, in reverse, to turn work into heat differences — active refrigeration! This is also a new thing. My hydroelectricity path doesn’t lead directly to those, just to having more work in the economy.

    (Then of course chemistry and synthetic fertilizers provided a way for work to turn into more food.)

    Separately, there’s the question of horse-drawn combine harvesters and such. Even if there’s no coal or hydropower, does such machinery get you to having 50% or 75% of the work force not involved in farming? That seems like a big deal too — no more food, but fewer people involved in making it, thus free for other things.

    1. Some interesting comments regarding water power. I’m not a subject matter expert, but do remember a bit from school. Even a quick check for watermills on wikipedia gives some interesting background – that sort of application dates back to classical history, and was one of the bright spots for medieval technology. So ‘hydro power’ wasn’t some incredible new idea – people had been using it to the limits of local expertise, materials, and geography for many centuries. I interpreted the post’s quick slap at watermills to mean that water power couldn’t really expand to meet new demands because, well, their use had been expanding to harvest power from any economically sensible site throughout those centuries. But once you had a decent steam engine, you could always order more coal.

    2. The Romans did have a horse-powered harvester! No threshing machine, though, despite that being a decent application for a water- or windmill (except the Romans didn’t have windmills).

      Similarly weird missing steps: the Romans had decent literacy and copied some works in massive numbers. They would have had great use for papermaking and the printing press, yet they had neither (they used papyrus, which is both inferior and had to be imported from Egypt). They were great at sieges, and the trebuchet is a 13th century invention. They were making objects from clear glass, but not lenses; they had handheld mirrors, but no telescopes, nor did they use them for communication (heliographs). They had sailing ships, but no experiments with flying gliders, kites or anything else (the Chinese did have man-lifting kites, but mostly as a gimmick). Btw, they didn’t have multiple-segment masts. They (and everyone else) had natural philosophers, they (and absolutely everyone) had noticed that children look like parents, Roman elites were even occasionally trying their hands at selectively breeding animals, but it took until Darwin&Wallace to discover evolution.

  32. This piece answers questions I had in mind years ago, spurred on by a mix of my high school Latin class studies and decades of playing as Rome in the Civilization series. Great stuff.

    Also, your discussion of the preconditions for the Industrial Revolution calls to mind all the time I spent watching James Burke’s “Connections” so many years ago, so thank you for that (I still recall that series with great fondness).

  33. I was struck by the contraposition of “Romans” and “Britons”. Weren’t they one and the same for almost 400 years? If we take that into consideration, then the arguments about the unique suitability of the British Isles for the commencement of the industrial revolution kind of fall flat. Or was Britain of AD 43-410 so different from Britain in the late 18th century (more forests, different population distribution) that the two are incomparable?

    1. Along with what Bullseye just said, ancient Roman Britain had no massive textile export industry. I’m not saying Roman Britain didn’t export wool at all, but it wasn’t on anything like the same scale as what early modern Britain was doing by the 1700s.

      Furthermore, and this is important, Roman Britain was the periphery of an empire, not the metropolitan center of that empire. Wealth- investment capital- political power- in an empire, all these things are centralized in the metropole. Even if all the material precursors had been in place in Britain, the investment capital to build giant steam-powered industrial plants would not have been, because that capital was all in the Mediterranean basin.

      By contrast, as noted above, early modern Britain had vast supplies of imported goods and could export to large captive markets in its colonies, instead of being the “poor relation” border province at the outer edge of the Roman Empire.

  34. My usual late contributions to the proofreading department:
    tinkering with an importance principle > important
    surface coal seems in abundance > seams
    gotten very good and pumping out > good at
    The summarize > To summarize
    there are a few favors that led > a few factors
    were really on true on Great Britain > only true for
    historians tend to be quite skeptic of ‘grand narratives’ > skeptical

  35. I could possibly imagine an Industrial Revolution on the back of high-pressure chemistry. Namely, the production of Ammonia for fertilizer. The development of manmade fertilizers caused such an enormous boom in food production that I could see it doing a similar thing to the centralization of textiles, only with farmers (fewer farmers producing more product on less land). The high pressure requirements would also drive the production and development of pressure vessel technology, and the exothermic nature of the reactions could potentially lead to the development of steam power.

    1. Yes, but to get that, you need an understanding of chemistry and metallurgy that make it practical to build ammonia production plants… And by the time you have that you probably also have gunpowder and cannons anyway.

    2. The Green Revolution is a different thing than the Industrial Revolution. No new energy sources, just greater use of organic energy.

      Also, I’m not sure something like that would be thermodynamically viable without mineral energy. Artificial fertilizers apparently take a lot to make.

      1. Yeah, you could call the fertilizer part of the GR “converting mineral to organic energy”. A lot of the progress of the past few centuries was discovering ways to convert things formerly separate: heat to work, work to heat (or cold)[1], work to food in non-muscle ways, work to light (electric lights)[1]…

        ([1] It’s striking. Say you’re an pre-modern emperor with 1000s of slaves for your personal use. They still can’t *make heat*, or *make light*, for you. Well, you could cuddle with some slaves under blankets for warmth. They can build insulated buildings and gather (or steal) firewood or lamp oil. But they cannot *directly* use their muscles to boil water or make even so much as a candle’s worth of light.)

        There were non-mineral parts of the GR, like breeding dwarf cereals that have more of the biomass as food. I think the ones in use require lots of fertilizer, but since potatoes can be 80% edible without synthetic fertilizer, I don’t see why breeding shorter and stouter cereals couldn’t give some useful gains. But that’s at most giving you 2x the food per labor and land (raising the ‘harvest index’ from 25 to 50%, say)… not trivial! but not like a modern American using 100 “energy-slaves”.

        (A human is roughly 100 Watts, Americans use 10,000 Watts energy per capita — though much of that is either heating energy, or waste heat from making the 1000 Watts of electricity we use… that’s still 10 “energy-slaves” of work. Also waste heat from cars, which only turn 20% of gasoline into work.)

        1. “But they cannot *directly* use their muscles to boil water” Well, not unless you want to do things the hard way and generate friction heat. Like when Benjamin Thompson a.k.a. Count Rumford using horse-driven drills boiled water by vigorously boring out cannon. Which astonished everyone who saw water boiling without a fire, and incidentally put paid to the Phlogiston theory.

          1. Again this is the fallacy that Simon_Jester has commented about before. Just because something is pedantically *possible* does not mean *practical* – even here, the objective was the very practical manufacture of cannon, not generation of heat. Human beings are not always rational economic consumers, but we don’t like doing things the hard way.

            If the alternative to tens of humans frantically spinning a drill shaft is “light a fire and put a kettle on top”, the human powered heat source will not see widespread use or development.

  36. I’d suggest the shipping argument on Rome missed a really big point: once Rome controlled the entire Mediterranean, shipping could supplant land-based transport throughout the entire basin, and shipping is massively more efficient and rapid than ground-based transit, even to this day. Cargo ships are apparently more efficient at moving fuel than are pipelines and railways. Meanwhile humans packing food on their backs are about as fuel efficient as SSTs like the Concorde, and animals pulling carts aren’t much more efficient.

    The tyranny of the wagon equation isn’t just about armies, it’s also about commerce, and Rome beat the equation within the Mediterranean basin with its ships. This becomes obvious when you note how much the city of Rome depended on grain from Egypt during imperial times. Were this grain moved overland, Egypt couldn’t have fed Rome, because the animals would have eaten all the grain in the carts before they got to the Bosporus.

    Another support of this idea was the Chinese digging their Grand Canal to facilitate transport within China. Even though the Chinese have coastline and some pretty innovative sailing ships, the Canal made transport so much more efficient that successive dynasties thought it was worth digging and maintaining the beast.

    The fun question then becomes why the two other bodies of water like the Mediterranean–the Caribbean/Gulf of Mexico and Indonesia and the Coral Triangle–didn’t develop Roman-scale maritime empires. That’s a much nastier question, especially when you consider that maize and rice are every bit as productive as wheat.