A thought experiment I occasionally indulge is to ask myself what technologies and industries could have been possible in pre-modern civilizations given their technological limits. That is to say, what technological and industrial innovations were within the capability of past civilizations, and would not have required anything more than incremental refinements in existing technology to be functional, so that the only thing missing was the proof of concept. Often ancient civilizations got their proof of concept by observing what other neighboring civilizations were capable of doing, and then trying to do it themselves. We call this idea diffusion, which is contrasted to independent invention.
When I entertain this thought experiment, I usually am thinking of Roman civilization at its height, because it had so many resources available to it, and it had developed a commercial and transportation infrastructure that made it possible not only to carry off big projects, but also to follow up on projects by maintaining or extending them. It’s often said that ancient Rome in the second century AD peaked at a population close to a million, and that means bringing enough food into the city every day to feed a million people, and to take away to waste and sewage of a million people. (Not to mention the enormous shipments of wood that had to be brought into the city to fire the massive public baths.) That’s a big job even by contemporary standards, and the Romans did it at a much lower level of technology that we have available. However, they had the resources such that, had an idea been presented to them, they could have developed it on a grand scale, as they developed the institutions that they had on a grand scale.
One of the most obvious things that Rome lacked was steel. They worked a lot of iron, but despite firing their iron in charcoal-fired furnaces, they never learned how to systematically enrich their iron with carbon in a distinctive process of carburization. Partly I view this as a consequence of the devalorization of technology in classical antiquity, which is a theme of James Feibelman’s book Technology and Reality. Most metallurgy, and indeed most technology, was never recorded in written form, and was handed down within families. Techniques and trade secrets tended to be local, and were often intentionally kept secret (there was no system of patents in the ancient world), so that a family might profit from one generation to the next. However, that meant that the technologies and techniques were easily lost. When an invading army came through a region and disrupted production, it was entirely possible for everyone with the specialized knowledge of a craft to be killed or scattered, and even if peace follows in the wake of the invading army, without written records the technology often couldn’t be reconstructed. This contributed to the fragility and vulnerability of the entire technological infrastructure of classical antiquity.
Coeval with Roman civilization, steel production was achieved in Central Asia, India, and China, but these industrial concerns were apparently sufficiently distant that no idea diffusion took place. Also, the conditions of production were similar insofar as metallurgical technologies were often closely guarded family secrets, which we could think of as a form of protection against idea diffusion in order to maintain a monopoly on a distinctive industrial process. In Central Asia and India, the technique developed was crucible steel, which involves putting pieces of pig iron (the result of an earlier process) into a ceramic crucible with carbon (usually charcoal or organic material) and heating this at a sufficiently high temperature for a sufficient period of time to achieve carburization. The Romans never seem to have achieved sufficiently high furnace temperatures, whereas in Central Asia and India they mastered the technique of maintaining a high furnace temperature, and it was only the difference of a couple of hundred degrees, but that couple of hundred degrees makes all the difference in steel production.
In the thought experiment of introducing steel production in ancient Rome, the relatively accessible techniques used in Asia (like crucible steel) could have been introduced, but other methods that could have raised the temperature of furnaces might have resulted in the Romans discovering steel on their own. The intuitively accessible idea of pre-heating the air that goes into a furnace, rather than just letting cold air rush in, wasn’t invented until the nineteenth century, but since, once you hear about it, it’s intuitively simple, and this could have been introduced much earlier in history if anyone had thought of it. Of course, today this is done in a sophisticated way, using heat exchangers to recover waste heat from the exhaust of the furnace to pre-heat air coming into a furnace, but an ancient analogy could have been accomplished without much difficulty at the technology level represented by ancient Rome.
I started thinking about the possibility of steel production in the ancient world indirectly through another approach to ancient technology, and this is interestingly relevant to proof of concept for a new technology. Whenever discussions of technology in the ancient world come up, someone will inevitably mention Hero’s steam turbine, which was viewed as a mere curiosity. The technology was known, but it wasn’t exploited, much less exploited at an industrial scale. This seems to be a clear case of proof of concept of technology being present, yet the technology isn’t adopted. Here the discussion usually turns to the use of slave labor in antiquity, which presumably reduces the incentive to build labor-saving machinery, and this is turn is related to arguments adduced for the high level equilibrium trap, which has been advanced to explain why there was no indigenous industrial revolution in China (China after all, as we have seen, was already producing steel in ancient times).
Thinking this through a little further, the industrial scale exploitation of a steam turbine would require that the turbine be engineered to withstand forces that would be placed upon it in its use as an engine to power industry or vehicles or whatever. A lightly-constructed steam turbine might amuse as a novelty, but it couldn’t do much useful work. To build a steam turbine would require significant engineering skills, and we know that the Romans had pretty sophisticated engineering skills, but there are also a lot of other technologies that would be relevant to the production of a steam turbine capable of doing useful work. For example, the construction of a useful steam turbine would be greatly facilitated by ball bearings, but ball bearings weren’t invented until the eighteenth century. Could they have been engineered in Roman times? Probably so, especially a cage bearing, which isn’t usually as small as individual ball bearings. Another alternative occurred to me. If a steam turbine were built with the kind of bearings used as the main bearings in engines, which are something like thin halves of a cylinder, this might work as well or better than ball bearings. However, to make this work you need to pump oil through the bearings, and that’s not something you would want to do around the open flame of a steam turbine. There’s a way around that too. Produce the steam at some distance and pipe it to the turbine. The Byzantines seemed to have mastered the pressurized delivery of “Greek Fire,” and one reconstruction of the delivery mechanism that I’ve seen could have been easily adapted to pump oil under pressure into the bearings of a steam turbine, making it capable of doing real work, and maybe kicking off an industrial revolution in the ancient world.
This is the chain of thought that brought me to look into why Roman iron production never rose to the level of steel production. Bearings like this would have to be made of a pretty good quality of steel, and in fact turbine assemblies generally employ very high quality steel, as the higher the pressures in a turbine, the more efficient it can be. A pretty sophisticated technology can be enabled by gold, copper, bronze, and iron, all available to the civilizations of classical antiquity, but to crest beyond that threshold it’s probably necessary to have steel. And even steel in and of itself isn’t enough. As we’ve seen, steel was produced in the ancient world in Central Asia, India, and China, but this was effectively a boutique industry. The production reached impressive levels in a few places, but not the level that would be needed to make steel sufficiently common to come into wide use. An illustration of this is the speculation that the high quality steel used in Ulfberht swords (discussed in newsletter 193) came from Central Asia or India. No doubt this was a highly valued commodity that fetched a high price in trade, and we know that Viking trade networks extended into Central Asia. But there are fewer than 200 known Ulfberht swords. Ulfberht swords were a rare commodity, a prestige weapon, and they were produced with another rare commodity as an input: crucible steel.
The Ulfberht sword represents a particular historical milieu: a place and time where what limited resources were available came together in the manufacture of a unique commodity. Probably the steel came from Central Asia, and was worked by smiths somewhere between Central Asia and Scandinavia, as almost all known Ulfberht swords have been found around the Baltic. But this historical moment passed without cresting over directly into a higher stage of technological development. When the Mongols swept through Central Asia, crucible steel production was nearly wiped out—an illustration of the fragility of boutique production of a technology. The industry never recovered and declined into insignificance. A little after this, steel produced in blast furnaces began to appear in Europe, first in Italy, Germany, and Sweden, not as the result of the idea diffusion of crucible steel, but rather the slow, steady, and incremental improvements in the metallurgical technologies that raised the temperature of furnaces and introduced novel techniques for carburation.
In my ancient technology thought experiments the most compelling example I have yet thought of is air balloon technology. All it requires is a reasonably gas-impermeable fabric over a frame, and a source of heat. The Romans could have built a hot air balloon without much difficulty, if only the proof of concept had been available. Air balloons could have been a novelty like Hero’s steam turbine, and would have remained as such unless further developments were made. One way to turn air balloon technology into an air transportation infrastructure would be through thermal airships (a dirigible using hot air instead of hydrogen or helium gases). Without an engine airships couldn’t get far, but slaves could crank propellers, and in this way something like real air transportation would be possible at the technological level of achievement of the ancient world.
Another relatively simple technology is black and white photography, which requires only the most basic chemical engineering and a wooden box with a lens. We can imagine a counter-factual ancient airship carrying a counter-factual ancient camera, extending the technology of aerial photography to the ancient world. This, I think, would have been sufficiently revolutionary that, had these technologies been introduced in ancient civilizations capable to building them and using them, these developments would have radically changed history.
fragile production processes produce magical goods and not commodities