«by Robert C. Allen Professor of Economic History Department of Economics and Nuffield College Oxford University Email: ...»
Do differences in human capital explain why the industrial revolution occurred in Britain rather than France? Literacy in France as a whole was lower than in Britain, but France was a bigger country with a larger population and considerable diversity. Literacy in northern France was about as common as in Britain, and so human capital differences may not have been important. Indeed, it is not clear that there was much difference in inventiveness between eighteenth century Britain and France. There are certainly many examples of the French inventing. Why do we think the British had a more pragmatic engineering culture than the French? Because it was Brits who first smelted iron with coke, invented the steam engine, and discovered how to spin with machines. In the rest of this essay, I will show that these differences in behaviour were due to differences between the countries in the profitability of doing R&D. If that argument is accepted, then cultural explanations become superfluous.
Some famous inventions The only way to adjudicate between the cultural and economic explanations of technical change is to test them against the history of invention. I will examine three famous
inventions–coke smelting, mechanical spinning, and the steam engine. I explore four themes:
• What was the origin of the idea embodied in the invention? Was it an inspired act of genius or a scientific discovery? With the exception of the steam engine, which was based on science, the fundamental insight was copied from other activities. Boyle proposed the publication of craft knowledge to promote invention-by-copying, and Mokyr has made it part of his Industrial Enlightenment. Indeed, copying was the general pattern and shows that the Industrial Revolution was based on little ideas–not big ideas as often assumed.
• What technical problems had to be solved in order to put the idea into practice? How much R&D was involved in making the idea work? R&D was the crux of invention in the eighteenth century, and all of the famous inventions including Newcomen’s steam engine, required substantial development programs to perfect them. These R&D projects exhibited the modern trilogy of development costs, external finance, and patenting. The expense of R&D turned invention from a scientific into an economic activity.
• How were these inventions related to Britain’s unique wages and prices? Were the inventions biased in the sense that they cut costs more at British prices than at foreign prices? All of these inventions make sense in terms of the high wages and cheap coal of the British economy. Despite being known, they were not adopted in other countries where wages were low and energy expensive.
• Why were they invented in Britain rather than elsewhere in Europe? The bias of the technical change implies that R&D was a profitable investment in Britain but would not have been in other countries.
The invention of coke smelting
Coke smelting is one of the famous inventions of the industrial revolution and had an enormous long run impact, for it was essential for the production of cheap iron, which, in turn, was required for the railroad, metal steam ships, and the general mechanization of industry. The invention of coke smelting clearly illustrates the themes of this essay: It was a little idea, not a big idea. Initially, coke iron was more expensive than charcoal iron, and the first problem was to develop a market for the new product. This was accomplished through an R&D program to make thin-walled castings. The second problem was to cut production costs, so coke iron was competitive with charcoal iron for refining into wrought iron. This problem was solved inadvertently as problems of irregular water supply were addressed.
Coke smelting was a biased technical improvement, which was not profitable to use in most of Europe, and would not have been profitable to invent outside of Britain. That is why the discovery was made in Britain rather than France.
How much creativity did coke smelting require? What engineering problems did it pose?
Coke smelting did not depend on any scientific discovery nor did it require an act of genius. In fact, it required almost no thought at all. Coal was a much cheaper source of energy than wood, and attempts were made to substitute the cheaper fuel in most applications during the seventeenth century. If coal was being burnt to heat the house, why not chuck it into the blast furnace instead of expensive charcoal? And, indeed, there are many examples of people doing just that in the seventeenth century. Dud Dudley was an early pioneer who claimed in his book Metallum Martis (1665) to have successfully smelted iron with coke, and he had the iron goods around his house to prove it. Others followed, and there is no reason to believe that they failed. The problem was that the process was not economic. Most iron in the seventeenth century was refined into wrought iron, and pig iron smelted with coal contained too much sulfur for this to be successful. This was a typical problem in substituting coal for wood: the coal introduced impurities, so new technology had to be invented to eliminate them. Wrought iron was not successfully made from mineral fuel pig iron until the middle of the eighteenth century.
Abraham Darby I is usually credited with the invention of coke smelting, but, as noted, he did not conceive the idea. Darby probably learned about coke smelting from Shadrach Fox, who had a contract to supply the Board of Ordnance with cast iron shot in the 1690s. This iron was probably smelted with coke, and the Fox’s furnace was the one at Coalbrookdale that Darby later leased. The furnace blew up in 1701, and Fox smelted some more iron with coal or coke at the Wombridge Furnace. Darby leased the Coalbrookdale from Fox in 1708, rebuilt it, and set off on his career smelting coke iron (King 2003, p. 52).
The link from Fox to Darby solves several puzzles–why Darby never patented coke smelting (although he patented a casting process) and how he had the confidence to use coke from the very inception of his business. He seems to have known the process would work technically, for he did no experimenting with coke nor does he seem to have had a back-up plan to use charcoal if coke smelting failed. Also, Shaddrock Fox’s experience showed that coke iron was suitable for castings, which was the application Darby had in mind.
Darby’s R&D project
Indeed, Darby’s contribution to ‘inventing’ coke smelting was in finding a commercially viable application for the material. In about 1702, Darby and other Quakers established the Baptist Mills Brass Works near Bristol. Most brass was then fabricated by drawing it into wire or by hammering sheets into pots, kettles, and such like. Casting was traditionally limited to church bells and canon. However, by the late seventeenth century, the Dutch were casting many other products using sand moulds and reusable patterns. In 1703, Darby set up his own foundry and tried to cast iron pots with sand moulds, but he was unsuccessful. In 1704, he went to the Netherlands to study sand casting. He brought back some Dutch workers and got them to try casting iron, but they were also unsuccessful.
However, an English apprentice, John Thomas, believed he could do it, and Darby paid him until he was successful in 1707. This was Darby’s principal R&D project, and it resulted in a patent for casting iron with sand molds. Darby’s partners in Baptist Mills did not want to pay for this research, but he found a new financial backer in Thomas Foudney.
When Darby leased the Coalbrookdale furnace from Shadrach Fox, his plan was to smelt pig iron and cast iron pots with sand moulds. Not only were the castings successful, but the silicon in the coke iron rendered it more fluid than charcoal iron, so it proved possible to make pots with thinner walls that sold at a higher price. This was essential for the success of coke smelting since the iron itself was expensive. This was the process that Darby patented (Mott 1957-9, p. 78, Hyde1977, pp. 40-1).
The smelting process involved two further examples of technological borrowing. The first was in the manufacture of coke. Darby had learned how to make coke when he was apprenticed to a maltster, for coke had been invented for that use (Matt 1957, Raistrick 1989, pp. 23, 25). The second was the use of the reverberatory furnace to remelt the pig iron for casting. Reverberatory furnaces had been used since the middle ages to melt the brass for bell founding, and Dud Dudley may have used such a furnace to cast iron. In the 1670s and 1680s, the reverberatory furnace was used to smelt lead and copper by two chartered companies associated with Sir Clement Clark, who may have experimented with melting iron.
Darby was the first to make a commercial success of the reverberatory furnace in the iron foundry (Mott 1957-9, p. 76, King 2003, p. 51). Evidently, Darby’s plan to cast pots with coke pig iron did not come out of nowhere. It was the combination of several recent developments in the iron and copper industries.
For the first half of the eighteenth century, coke smelting was limited to only a few furnaces making foundry pig, for the metal was too expensive and impure to refine into wrought iron, the main product of the industry. This problem was overcome through ‘learning by doing’ or, more exactly, through inadvertent discovery made in the course of solving other problems. The bellows of blast furnaces were run with water power, and a dry summer meant that the water level dropped in the reservoir supplying the wheel resulting in a fall off in blast and reduced iron production. This problem was resolved at Coalbrookdale by installing a Newcomen steam engine to recycle water by pumping it from the outflow of the water wheel back into the supply reservoir. Coalbrookdale was one of the first firms to use this process (Raistrick 1989, pp. 107-115). As was the case with Warmley, coal was mined at Coalbrookdale, and the cheap fuel made the Newcomen engine feasible.
The improved water supply resulted in stronger, more regular blast to the furnace, and that had the unintended consequence of cutting fuel consumption. Lower fuel consumption, which was an energy-saving technological improvement, cut costs enough to make coke iron competitive with charcoal. Coke iron production took-off after mid century.
Coke smelting was a biased technical improvement that reduced costs in Britain more than on the continent.
By replacing charcoal with coke, Darby’s smelting process cut costs in localities where coal was cheap. Since most coal in Europe in the eighteenth century was mined in Britain, coke technology (once perfected) conferred a great advantage on Britain. Conversely, coke smelting was pointless where coal was as dear as it was in most Europe before the midnineteenth century. As late as the 1840s, 80% of French and Prussian iron was charcoal.
Belgium is the exception that proves the rule, for it shifted early to coke, and it was also the only part of the continent with large scale coal mining in the eighteenth century and a price structure like Britain’s (Landes 1969, p. 217). While Landes has argued that Britain’s lead is evidence of superior entrepreneurship, Fremdling (2000) has shown than coke iron did not pay on the continent before the 1850s. Production costs explain the diffusion of the technology–not attitudes to innovation.
Why not France?
It took almost a century from the perfection of coke smelting at Coalbrookdale until its use was widespread on the continent. During that period, the technology was well known and freely available but not adopted. Since it conferred no benefit to French or German producers, there would have been no point in developing it in those countries. It was not the impracticality of the engineering culture that explains the lack of attention to coke smelting.
Inventing the process would not have paid.
The invention of cotton spinning machinery
How much creativity did mechanical spinning require? What engineering problems did it pose?
The spinning jenny and water frame were not based on scientific discoveries. Were they instead ‘macro inventions’ that required enormous leaps of the technological imagination? To know, we must see if the spinning machines really did spring ab nihilo or whether they had genealogies that indicate less dramatic departures from previous practice. I begin with hand spinning to highlight the technical problems that Hargreaves and Arkwright faced.
Figure 7 shows a spinning wheel in operation. The raw cotton was first carded to produce a roving, which was a loose Figure 7 length of cotton fibres. The two key operations in spinning were drawing out the roving so it became thinner and then twisting it to impart strength.
In the late medieval period, this was done with a ‘spinning wheel’. It consisted of three parts–the wheel itself, the spindle, and the string that acted as a belt to connect the wheel to the spindle. Sometimes a treadle was connected to the wheel so that the spinster could turn it with her foot;
otherwise she used her right hand.
She held the roving in her left hand, and its other end was attached to the horizontal spindle. The wheel was spun, and the spindle rotated. The spinster pulled back the roving so that it thinned out and then moved her hand to the left. This allowed the thread to slip off the end of the spindle each time it rotated. Each time that happened, the thread was twisted once. When enough twist was imparted, the spinster moved her left hand to the right, so it was once again between her and the spindle. In this position, the thread was wound onto the spindle. The process was repeated as the next few inches of roving were pulled away from the spindle to be thin out in turn.
It is hard to see anything that came ab nihilo in Hargreaves’ spinning jenny. It was little more than a spinning wheel on its side with several spindles connected by belts to a common wheel. Indeed, the story is that Hargreaves conceived the jenny when he saw a spinning wheel fall over and continue spinning while it was on the ground. A sliding frame replaced the spinster’s left hand and drew the rovings away from the spindles. The difficulty, as with most eighteenth century technology, lay in working out of the details of the linkages Figure 8 and rods that drew out the cotton roving.