concerning Malthusian traps and escapes; spillovers and residuals; the uneasy relationship between population growth and innovation; and the limitations of Chinese emperors, Dutch bankers, and French revolutionaries
IT TOOK ABOUT SIX hundred years for the publishing industry to get from Johann Gutenberg to the book you are now reading.* The most remarkable eight-week stretch in all of those six centuries fell between January and March 1776—a year overloaded with significant dates. On January 10, Thomas Paine published the pamphlet Common Sense (“Society is produced by our wants, government by our wickedness”). February 17 saw the first volume of Edward Gibbon’s History of the Decline and Fall of the Roman Empire roll off the presses (“In the second century of the Christian Era, the empire of Rome comprehended the fairest part of the earth …”). And on March 9, a former University of Glasgow colleague of Joseph Black published An Inquiry into the Nature and Causes of the Wealth of Nations.
Standing in London’s Science Museum, in front of Rocket and surrounded by models of Thomas Newcomen’s beam engine, Joseph Bramah’s challenge lock, and James Watt’s separate condenser, it takes very little imagination to see connections with iron foundries, coal mines, and even cotton fields. The road back to Adam Smith requires more thought, but is just as important, and as enlightening.
Smith’s book, like Darwin’s Origin of Species, was revolutionary in its impact, immediately and permanently, though both are far more frequently cited than read. In particular, Wealth of Nations demonstrates that Britain’s eighteenth-century transformation—the schoolboy’s “wave of gadgets”—was a revolution not merely in technology but in commerce. The founding text of economic science is staggering in its range, with disquisitions on the origin of money, the nature of commodities, interest rates, profitability, the mechanics of trade, bank interest, taxation, public debt, agriculture, and manufacturing. It devotes thousands of words to histories of Europe’s towns from the end of the Roman Empire to the present, and of colonial policy from the time of ancient Greece. It is telling, therefore, that Smith decided to open his magnum opus with a section entirely devoted to the “causes of improvement in the productive powers of labour [and] in the skill, dexterity, and judgment with which labour is applied in any nation.” This was a remarkable bit of insight, the application of Locke’s labor theory of value to national policy.
Smith argued that two conditions were necessary for labor to produce the maximum amount of wealth: perfect competition among sellers—everyone pursuing his or her selfish interest, the famous “invisible hand”—and the complete freedom of buyers to substitute one commodity for another. Under such ideal circumstances (Smith was not the first economist, but he was probably the first to “assume a can opener,” i.e. perfect conditions, in a model), specialization, or division of labor, was inevitable. Ten men could each bake their own bread, weave their own cloth, and build their own houses, but if one became a baker, another a weaver, and a third a builder, the result would be more food, clothing, bricks … and trade.
Smith’s theorems did a spectacular job of explaining the self-regulating character of a free market, in which prices and profits are forced by competition to the lowest possible level.* They inspired David Ricardo’s exposition, in 1817, of the principle of diminishing returns: his argument that the growth of the first decades of industrialization was certain to level off, as each successive improvement produced smaller results. Helped along by the inflation in food prices caused by the Napoleonic Wars, they even set the stage for Thomas Malthus’s Essay on the Principle of Population, with its famous argument that population always grows geometrically, food production arithmetically.
What they didn’t do was explain how wealth, profit, and competition can all grow over time. In short, it didn’t explain the two centuries of growth that were beginning just as Wealth of Nations was being published. It is in no way a criticism of the book to state that it covered everything except the reason the author’s own nation was about to get wealthier than any other nation in the history of mankind. The failure is pretty much explained by what is not in the book. Despite living in the middle of the biggest explosion of inventive activity ever recorded, and even though his illustration of the advantages of specialization was a factory for making pins, Smith’s book hardly mentions the role of the new machines then transforming his world. Next to nothing about waterpower, to say nothing of steam; nothing about the forging of iron,1 and his few paragraphs about the textile revolution are mostly an argument for restricting the export of spinning machines. His pin factory, it turns out, was only a metaphor; he never set foot inside one.
Nor did he show any understanding of Darby’s furnace, or Arkwright’s water frame, or Watt’s double-acting engine—none of the escape hatches out of humanity’s millennium-long Malthusian trap. The efficiencies of specialization are real, and the self-regulating “invisible hand” powerful, but it was the machines, and nothing else, that allowed Britain, and then the world, to finally produce food (or the wealth with which to buy food) faster than it produced mouths to consume it.
A lot more is known about how population increases than how wealth grows. Indeed, the Industrial Revolution was decades old before anyone realized that wealth was growing at all. The first edition of Malthus’s Essay on Population was published in 1798 and convinced nearly everyone that the hoofbeats of the horsemen of the Apocalypse could already be heard throughout England. In 1817, the English economist David Ricardo predicted2 that land rents would increase while wages would approach subsistence level, at precisely the moment when British farmland rents per acre started to plummet and the wages of laborers to explode. Partly this was evidence of the limits of accounting with very little data; Britain’s first census, inspired by Malthus, wasn’t conducted until 1800. But even more it was the lack of a model that carved up overall growth into its constituent parts.
In 1890, another economist, the mathematically trained Cambridge scholar Alfred Marshall, suggested that the century of growth in both income and wealth that began just as Ricardo predicted its opposite was largely due to ideas whose benefits spilled over into the economy soon after they had enriched their creator. An idea—a separate condenser, for example, or a spinning jenny—might be costly for one inventor to develop, but it wasn’t long (not even the fourteen years of a patent) before it became de facto public property, and inspired others to improve upon it.
Marshall’s “spillovers” were intriguing but remained anecdotal until the 1950s, when the Nobel Prize–winning economist Robert Solow incorporated something very like it into an equation known as the fundamental equation of growth. Working from a contemporary economic model that had shown that capital, labor, and land could be substituted for one another—as one component grew more expensive, producers could substitute one for the other—Solow was able to calculate the rate at which average workers increase their output. He found three components of output increase, each one reflecting the key inputs to the national wealth calculus. The first two, land per worker and capital per worker, are, if not easy, at least possible to measure. Except during times of dramatic depopulation, such as the Black Death of the fourteenth century, or extremely large additions to the stock of arable land, as with Europe’s discovery of the New World, growth in land per worker has been negligible for centuries, so small that its effect on growth can be eliminated in the simplest calculations. The second component, growth in capital3 per worker—that is, all the buildings, machinery, tools, and so on—explains only about 24 percent of total growth. However, since the growth in the amount of land and capital per worker together doesn’t equal the overall growth rate, a fudge factor must be used, called the residual: what’s left over.
This also means that the residual, despite the ass-backward way it is calculated, amounts to at least three-quarters of the total increase in economic growth since 1800. That’s a big chunk of activity defined by subtracting everything else, a little like a ten-drawer file cabinet with seven drawers marked “Miscellaneous.” Solow first assumed4 that the residual represented increasing efficiency over time, and he incorporated an arbitrary constant to represent the rate of the growth in useful knowledge.
Useful knowledge, in this formulation, is not all knowledge. The growth in capital includes not just cash, buildings, and machines, but also patents—for the duration of the grant. That’s how a corporation reports it on a balance sheet, and that’s therefore how it is accounted for in estimating national growth as well.
But all the patented knowledge that was originally counted as growth in capital becomes part of the residual once the patent expires, and what it loses in the value it had to its original inventor is gained by the inventor’s nation. Just as public domain books, such as the Bible and the works of William Shakespeare, are both more numerous and more valuable than the universe of copyrighted ones, the universe of useful knowledge is a lot bigger than the universe of patented ideas.
How much bigger? Solow attempted to put a number on the rate of growth in formerly (and also never) patented knowledge—which included everything from calculus to the laws of motion—and assumed, for the sake of simplification, that the rate of increase was not only regular, but independent of changes in custom, law, or historical contingency; that is, knowledge, like Topsy, just “grow’d.” Such simplifications are essential for theory building, but this particular one just pushed the big question back another step: Since knowledge, whether patented or not, is rarely lost, and the sum available has been increasing at least since the invention of written language more than five thousand years ago, what caused it, for the first time in history, to increase faster than the rate of population growth?
Population and prosperity are correlated, albeit imperfectly. Adam Smith was the first to recognize the hugely important but completely obvious correlation (this is a pretty good definition of genius) when he pointed out that the value of specialization utterly depends on the size of the community in which one lives. A family living alone grows its own wheat and bakes its own bread; it takes a village to support a baker, and a town to support a flour mill. Some critical mass of people was needed to provide enough customers to make it worthwhile to invest in ovens, or looms, or forges, and until population levels reached that critical level, overall growth was severely limited.
That level, however, was reached long before it had any impact on per capita growth in productivity. From 1700 to 1820, China grew in population from 138 million to 328 million, which increased its production of goods and services from $83 billion to $229 billion—but both population and production increased at almost exactly the same rate, about 0.75 percent annually. Population growth alone is clearly not sufficient to explain, for example, how the population of Britain, during the same period, could increase at the same rate as China’s, but its gross domestic product nearly one-third faster. Something more than specialization through population growth was at work.
This was the conclusion of E. J. Hobsbawm, who argued that the fuel for the Industrial Revolution was not coal but demography: population—or, more precisely, the growing size of markets both domestic and foreign. While recognizing, generally, the importance of a cadre of mechanics to build and repair steam engines, forges, lathes, and spinning machines, he minimized the importance of the inventions on which they practiced their trade. The presence of a thousand brilliant inventors was far less important, in Hobsbawm’s words, than “the mass of persons with intermediate skills,5 technical and administrative competence … without which any modern economy risks grinding into inefficiency.”
Another currently fashionable explanation for the Industrial Revolution is also a demographic one, though subtle: not the nation’s overall birth rate, but the birth rate in a particular subset. This is essentially the theory proposed in Gregory Clark’s examination of the relationship between differential reproduction in different classes.
Clark’s discovery, arrived at by combing through centuries’ worth of parish records, was that the wealthiest families in England were far likelier to have more sons than similarly wealthy families in China, and therefore had less real property to pass along to the average member of the next generation. Throughout the period 1540–1640, preindustrial Britain exhibited a fair bit6 of both upward and downward mobility, largely due to primogeniture, which obliged a prosperous landowner to leave his land to only one of his sons—and the typical landowner had up to eight. Since all but one of those sons would have to find his own niche in the economy, “craftsman’s sons became laborers,7 merchant’s sons petty laborers, large landowner’s sons smallholders” carrying with them the habits of hard work, deferred gratification, literacy, and a disposition to settle disputes peacefully, all of which showed a decided increase during the eighteenth century. As upper-class habits trickled throughout society, so did economic growth.
This theory explains fairly well why the son of a country squire might find himself learning the craft of a carpenter, and it may explain some critical aspects of historical growth in national wealth, particularly in Britain. A literate population bound by the rule of law and exhibiting middle-class behaviors such as deferral of gratification and low levels of corruption is as valuable to a nation as it is to a business firm. A recent World Bank analysis8 found that a significant amount of wealth worldwide is derived from such behaviors, and that human capital and the rule of law might account for up to 30 percent of the residual.
But the middle-class work habits that proved so vital for a new generation of factory laborers were valuable only because there were factories for them to work in. And Boulton and Arkwright weren’t motivated to build the factories because good workers suddenly became available, but because they had machines that needed housing. Similarly, whatever percentage of Solow’s residual is attributed to a reliable labor force operating in a relatively uncorrupt economy ruled by law, something still needs to explain why growth in prosperity, three-quarters of it derived from the creation of useful knowledge, stagnated for five millennia and then exploded in Britain in the eighteenth century.
This is where Solow’s simplifying assumption—the idea that the growth in useful knowledge occurs at a constant rate, directly proportional to population—however analytically valuable, runs out of gas. It can’t explain the steam engines and reciprocating chisels of the Portsmouth Block Mills, or the boring machines of Bersham Furnace, or the water frames in the Derwent Valley, because it implies that the decisions made to investigate the properties of steam, or iron, or silk were just as probable in eighteenth-century India as in eighteenth-century Europe; more likely, really, since India was home, in 1700,9 to more than thirty million more potential inventors than all of Europe, including Russia. The only thing that can be said in favor of making the growth of knowledge what economists call an exogenous variable (i.e. one without deliberate purpose, a kind of magical growth in knowledge from which everyone could benefit) was that its constant rate of increase made the equations simpler. By the 1980s, however, a number of economists examining the eighteenth century’s wealth explosion thought that Solow’s equations might be too simple.
The best known of them, Paul Romer of Stanford, made his reputation by demonstrating that useful knowledge—the largest component in Solow’s residual, and therefore the most important component of any increase in national wealth—doesn’t accumulate by itself, independent of the larger economy, but rather is almost entirely dependent on decisions made by individuals seeking some sort of economic advantage. Romer showed that such growth depended on real people acting, in their perceived self-interest, to create knowledge. His model was both stable and reflective of the real experience of economic growth since 1800. It also resolved, in a mathematically rigorous manner, the conundrum in Solow’s fundamental growth equation,10 which was precisely how the initial stage of knowledge creation benefits the investor/inventor and therefore is counted in the “capital growth” segment, while the larger sum spills over into the economy at large and forms the residual.
More important, Romer demonstrated that the growth in useful knowledge occurred at anything but a constant rate; that it was, instead, highly sensitive to the economy in which it occurred. Romer recognized that the creation of the idea behind a new invention, or product, was just another fixed cost, like constructing a building, or buying a machine; it was just as expensive in skull sweat for James Watt to design the first Watt linkage as to manufacture a thousand. Because knowledge is the sort of property that can be sold to multiple consumers without lowering the value to any of them—Romer termed it nonrivalrous, as distinct from tangible, or rivalrous, property, which can be sold only once—the payoff for anyone producing it can be very large indeed.
At least, in a large enough population.
IN 1993, A DEVELOPMENT economist named Michael Kremer published a paper in the Quarterly Journal of Economics that examined population growth over time and tracked it against the expansion of what Romer called nonrivalrous, useful knowledge, and concluded that the creation of knowledge is directly proportional to the size of the population. Kremer’s model made two assumptions:11 first, that inventive talent and motivation are randomly distributed throughout any population; and, second, that the larger the population, the larger the total output of inventions. The idea—that since each person has the same likelihood as any other of inventing the wheel, or the steam engine, or the iPod, more people means more inventors—may seem counterintuitive; it took only one genius, after all, to put stirrups on a saddle. But it works. A nation needs only twelve players to field a team to compete for the Olympic gold medal in basketball, and even tiny nations would find such a number easy enough to assemble. But only a few nations are actually competitive, because the twelve players represent the top of a pyramid of competition in which the larger the base, the greater the height.*
Thus, from the first century, when the world’s population was around 128 million and its GDP about $105 billion,* to 1500, when population was 438 million and GDP $238 billion, the various growth equations demonstrate that all of the components of economic growth—land per worker, capital per worker, and therefore the residual—increased at a constant rate. More workers, more income, and, by inference, more invention. Moreover, the country-by-country numbers are relatively close in per capita terms; in 1500, each of the 48 million residents of Western Europe was producing, on average, about $772 annually, while China’s 103 million were good for about $600 a head.
The year 1500 is significant because it’s the one that marks what the economic historian Kenneth Pomeranz calls “the Great Divergence.” Up until 1500, the difference between GDP per capita in the world’s poorest countries and the richest was in general less than 50 percent; by 1820, it was more than 300 percent. After 1500, the lockstep increases in population and knowledge ended. From 1500 to 1820, China’s population nearly quadrupled, and so did its GDP, with no increase in per capita GDP. During the same period, the population of France doubled, but its per capita GDP increased 56 percent; that of Great Britain quintupled, and its per capita GDP increased nearly two and a half times.
Kremer’s theory was intended to explain productivity growth over a very long time frame; his article was entitled “Population Growth and Technological Change: One Million B.C. to 1990.” However, for growth and change after 1500, and especially after 1700, it presents a pretty serious problem for the idea that inventiveness is directly proportional to population. Over the last five centuries, the theory only works by assuming that one can usefully calculate an average for the entire world; but, especially since 1820, the worldwide growth rate has been severely distorted by the huge acceleration in productivity in Europe and North America, which is a little like calculating the average wealth of the patrons in a restaurant before and after Bill Gates enters it.
It’s not as if Kremer was unaware12 of the problem; he acknowledged that for centuries, poor but populous countries like China and India had experienced decidedly low research productivity, which at least suggested that inventiveness is a function of income rather than population, but not enough to change his basic thesis. That thesis, however, is considerably more persuasive when the time span is more than a million years than when one is examining a single century or a single country. But if a large population alone doesn’t improve the chances of the steam engine’s being invented in any particular nation, what does? Why did none of China’s 138 million people invent a working steam engine, while Thomas Newcomen, one of fewer than nine million Britons, did?
China’s size, its extraordinary technological head start, and its relative isolation have made it, for centuries, an irresistible laboratory for theorizing about industrialization and invention. Most of the resulting theories end up emphasizing either China’s geography, science and technology, demography, or political culture.
Geography first. Geographical determinism is the notion that mineral resources, topography, and climate are the key drivers of human history. It remains eternally popular as an explanation for everything from political conflict to pandemics; but China’s missing Industrial Revolution looks like a poor example of it. Pomeranz attributes a good deal of his “Great Divergence” to the relative ease with which British coal was extracted and therefore iron produced. But China had, and has, huge coal deposits,13 just as close to the surface, and China’s forges were using coal to produce as much iron in the year 1080 as all of Europe did seven centuries later. They even did so for the same reason: the Yangtze delta had been deforested by the same demands as were the English Midlands, namely, construction, heating, and smelting. And even though the barbarian invasions14 of the twelfth and thirteenth centuries pushed China’s center of gravity south to less coal-rich areas (China’s nine southern provinces have less than 2 percent of contemporary China’s coal reserves), iron production rebounded almost immediately and by 1700 was certainly larger in China than Europe. Neither Europe nor Britain truly had, to an objective eye, any advantages in mineral wealth, climate, or navigable waterways.
The state of Chinese science relative to the West seems a more promising explanation than its geography. It seems plausible, for example, to argue that if Chinese scientists failed to uncover the foundation principles unearthed by Torricelli and Boyle, Chinese inventors would have found it far more difficult to replicate the work of Newcomen and Watt. This is a small part of what the eccentric Cambridge don Joseph Needham spent his life investigating and chronicling in the two dozen volumes of his masterwork, Science and Civilization in China. Needham’s conclusion,15unfortunately, doesn’t do much to confirm this particular diagnosis: he argues that the basic understanding of both vacuum and adiabatic pressure—the phenomenon that causes a gas to cool when it expands and heat when it is compressed—was present in China from the thirteenth century onward. Even if this overstates the case, a reasonable consensus exists for the belief that by the sixteenth century China had enough awareness of atmospheric pressure to produce not only a Newcomen-style reciprocating steam engine, but one that produced rotary motion.
Chinese engineers had already shown remarkable cleverness in transforming rotary into reciprocating motion. By 400 CE they had developed a system of water “levers,”16 which used a waterwheel to fill a chute with water, tipping it first in one direction, then in another. Even more significant, and well known, was China’s twist on the blast furnace, which, as the careful reader will recall, was being used to make cast iron in China by 200 CE, at least a thousand years before one appeared in Europe. The Chinese version of this device, called a box bellows, used a piston to pump air in and out of a cylindrical box. Early box bellows were hand-operated, but, in legend at least, sometime around 30 CE a Chinese inventor named Tu Shih hooked the piston up to a waterwheel-driven crank, thus facilitating the transfer between reciprocating and rotary motion seventeen centuries before Watt and Pickard. But it was an English historian named Ian Inkster, writing in 1842, who realized the potential for a box bellows piston that produced force on both strokes, writing, “let it [the box bellows] be furnished17 with a crank and flywheel to regulate the movements of its piston, and with apparatus to open and close its valves, then admit steam through its nozzle, and it becomes the double-acting engine.” Pomeranz agreed: “The Chinese had already recognized18 the existence of atmospheric pressure, and had long since mastered (as part of their “box bellows”) a double-acting piston/cylinder system much like Watt’s, as well as a system for transforming rotary motion to linear motion that was as good as any known anywhere before the twentieth century. All that remained was to use the piston to turn the wheel rather than vice versa.”
That was all that remained. It still remained when the first steam engines in China were being imported from Britain. The box bellows was undeniably an inventive leap, but not in any useful direction. The bellows, it turns out, is not a mirror image of the piston, but its opposite. To the degree that it works as a bellows, it cannot work to drive a piston. The Chinese could have a bellows,19 or a vacuum, but not both, at least not without a strong theory of the behavior of gas and pressure like the one articulated by Boyle in the seventeenth century.
And they didn’t have one. Chinese science had taken a very different path than its analogue in the West; most especially it lacked the strong historical foundation—no Descartes, no Galileo, no Bacon—needed to develop the experimental science of the Enlightenment. Even worse: When a scientist like the polymath Fang Yizhy attempted to import Western methods into China in 1644, producing a mammoth collection of works on mathematics, engineering, and natural philosophy entitled The Small Encyclopedia of the Principles of Things, it lacked any readership outside the aristocracy; China’s master artisans were so severely handicapped by illiteracy20 that eighteenth-century Chinese literacy rates never got much higher than 40 percent, at a time when 60 percent of British men were able to read.
If the Chinese handicap was neither geographic nor scientific, was it demographic? Even with Chinese literacy rates barely two-thirds that of Britain’s, by 1700 there were still more literate Chinese than the entire populations, literate and otherwise, of France, Germany, Italy, and Britain combined. How is it possible that not one of them was capable of inventing a spinning jenny? The only difference between Hargreaves’s invention and the machine used by Chinese cotton spinners, for example, was the draw bar, a device whose “fingers” could pull a large amount of softly wound cotton. The draw bar was not a complicated device,21 and yet even though at least five times as many Chinese as Britons were spinning fibers into yarn, there is no evidence that any of them invented one.
Instead, it appears that China’s huge population, and its powerful central government, both of which would seem to be advantages for industrialization, were liabilities when combined. The population’s demands for food were a tiger by the tail for the imperial court for centuries; so many mouths to feed meant that the investment capital available demanded single-minded focus on agricultural innovation. Just as contemporary scientists and inventors “follow the money,” writing proposals in the subjects that granting agencies currently favor, so Chinese innovation pursued those avenues for which support was available. The difference is that contemporary scientists can select from many patrons; in China there was only one.
As with Tudor England, government monopoly of patronage meant control. Virtually all copies of the seventeenth-century Chinese encyclopedia, the T’ien Kung K’ai-wu or Exploitation of the Works of Nature, which included illustrations of everything from hydraulics to metallurgy, were destroyed because, according to Joseph Needham, much of the material touched on industries that had been granted monopoly status by the Qing emperors: “The absence of political competition22 did not mean that technological progress could not take place, but it did mean that one decision-maker [i.e. the Emperor] could deal it a mortal blow.” It is therefore no surprise that a high percentage of both the inventions and inventors we associate with China from the time of the Han Dynasty to the Qings were government sponsored and employed.
Another liability of a strong central government is that it is, well, strong. Europe’s fragmented system of sovereign states23 made it possible for innovative minds such as Paracelsus, Leibniz, Rousseau, and Voltaire to “shop” for more congenial places whenever they skated too close to heretical or otherwise challenging notions; in China, one had to travel a thousand miles to a place where the empire’s writ ran not. And since China, perversely, was able to keep from plummeting down a Malthusian hole by using its enormous geographic extent to expand land under cultivation in the southern and western territories, technological stagnation, until contact with the West, seemed to have few if any costs.
In the end, however, neither territory, nor politics, nor even science is as powerful as culture in explaining China’s inability to produce its own steam engines, puddling furnaces, or spinning jennys. Bertrand Russell translated the Chinese term24 wu wei (usually “doing without effort”) as “production without possession”—a simplification, no doubt, but one with some powerful resonance to the notion that China’s great burden was a lack of the itch to own one’s own work.
ON THE OTHER HAND, that itch was powerfully felt in the seven United Provinces of the Netherlands from their de facto founding in 1581, when the onetime Spanish colonies abjured the rule of Philip II. By the beginning of the seventeenth century, the Netherlands was home to Europe’s most cosmopolitan culture; the world’s first stock market; a large and powerful merchant class, including Europe’s largest bankers; access to millions of consumers through its enormous merchant fleet and colonies; the rule of law; and tolerance for every religious confession in the world, including the best established and assimilated Jewish community in the world. It had developed, out of deference to the North Sea, a huge network of windmills, canals, and waterwheels. It even had, from 1683 to 1688, the prophet of the concept of intellectual property, John Locke, who wrote most of the Treatise on Government while living in Utrecht, Linden, and Amsterdam.
And it had a deep respect for inventions and inventors. Corneliszoon’s crank-operated sawmill was far from an unusual case. From 1600 to 1650, the Dutch government25 issued between five and ten technical patents annually. However, the pace thereafter dropped off precipitously—and the reasons are telling. Petra Moser, now a professor26 at MIT’s Sloan School of Management, spent four years examining more than 15,000 different inventions exhibited at nineteenth-century world’s fairs, and their equivalents, and discovered a fact that seems at first glance to discredit the idea that patent protection was essential for innovation: Nations without patent laws were in many cases just as inventive as those with them. Or even more inventive; some of the nations best represented at those industrial fairs actively discouraged the patenting of inventions.
The reason seems to be that whether or not they enforced a patent law, smaller nations or domains, such as the Netherlands and Switzerland, were vulnerable to the theft of their innovations by competitors in larger nations. The bargain of patent protection runs two ways: The state, in return for making an idea public, offers legal recourse to its creator should someone within the state steal the idea. Since making one’s invention public in a nation with patent protection offered protection against theft only up to its own borders, only a large nation offered a large enough market to make the deal a good one, and (in Moser’s words) the small nations “would have been silly27 to patent [their] innovations.”
This logic inhibited investment in entire categories of innovation. Those nations that relied on secrecy rather than patent tended to specialize in the sort of inventions that cannot be easily reverse-engineered, such as chemicals or dyes. One consequence is that almost all mechanical inventions—particularly, all steam engine innovations—were produced in countries that enforced some sort of patent law, since one can scarcely sell a compound engine and simultaneously keep its workings secret. Another is that any benefit from the cross-fertilization of ideas resulting from the public requirement of patent law, including publication of specifications, was lost.
The result is that while the Netherlands led the world in per capita (really, per man-hour) GDP every year from 1700 to 1820, its average compound growth rate over that period was actually negative, falling more than 13 percent. From there on, Britain took over the lead,28 with a growth rate per man-hour of 0.5 percent—a rate that exploded to growth of 1.4 percent annually from 1820 to 1890. The remarkable growth of the Netherlands during the 1600s essentially stopped a century later, and the only persuasive reason is size, or rather scale. A small country can shelter the world’s largest banks, shipbuilders, and even textile manufacturers, but since it can protect inventors only from their own countrymen, growth that depends on the creation of new knowledge is fundamentally unsustainable, like a nuclear chain reaction with insufficient critical mass. Just as fission requires that a sufficient number of uranium nuclei be, in some sense, accessible to the others, a chain reaction of innovation sustains itself only if innovations are accessible to one another. A few thousand Europeans, no matter how inventive their work in chemicals, or metallurgy, could not create an Industrial Revolution unless they could inspire (or borrow, or even steal) from one another; a few thousand Britons, precisely because they concentrated their efforts on “public” inventions, most especially the steam engine, emphatically could. The consequence is that smaller nations, by avoiding large-scale mechanical invention out of fear that their own territory was too small to make them profitable, deferred industrial leadership to those large enough to take the risk.
So if the Netherlands was too small, and China too big,* what nation was, in the immortal words of Goldilocks, “just right”? The most intriguing candidate29 would seem to be France, which had just about every advantage that Britain had, and a lot more Frenchmen to exploit them. France’s economy grew at almost precisely the same rate as Britain’s between 1700 and 1780, and from a far larger base; only about a tenth of one percent per year less growth in industrial (and, for that matter, agricultural) output in France during those eight decades. In 1789, the year of the Revolution,30 France’s foreign trade was 25 percent greater than Britain’s, and its population was nearly three times bigger.
By the same year, however,31 Britain had a significant lead in any number of significant indices: a third more GDP per capita, far higher rates of urbanization (nineteen of every hundred Britons lived in cities, and only eight Frenchmen), nearly eight times as much money in the hands of banks, and a tax rate less than a third of France’s. Thus, in part because of lower interest rates32 from the middle of the eighteenth century forward, the availability of capital (as opposed to its absolute amount) was significantly greater in Britain.
But it was not only, or even mostly, an advantage in business sophistication that gave Britain a head start on industrialization. Nor was it scientific sophistication, a yardstick on which France was way ahead: Watt was simultaneously a brilliant engineer33 and a gifted scientist, but he still needed to study French engineering texts because there were so few English ones available. His experience remained true through most of the nineteenth century, when French, and later German, scholars were giving a scientific foundation to the laws of thermodynamics and kinematics.
And they were doing so in an environment in which standardization was very highly valued, and strictly enforced. Possibly because it achieved status as a coherent nation-state centuries before England, to say nothing of Britain, France has a far longer history of activism in setting national standards; the Académie Française, as a case in point, has been protecting the purity of the French language since 1634. The project of synthesizing the knowledge held tightly in the hands of French artisans, mechanics, and craftsmen began only forty years later, when the Académie des Sciences—France’s equivalent of the Royal Society—began a national “Description des arts et métiers” intended to establish standard versions of hundreds of apparatuses. Among other things, the project provided Diderot34 and his Encyclopédie with more than 150 drawings and engravings of water pumps, looms, and forges. In the following century, the French government explicitly took on the responsibility of educating and training engineers with the founding of several schools focused on applied science, including the École des Ponts et chaussées in 1774; the École polytechnique was founded by a graduate of Ponts twenty years later, in part to impose technical standards on industry with the same rigor that the Académie governed the French language. The Encyclopédie itself promised “to offer craftsmen the chance to learn35 from philosophers, and thereby hopefully to advance further toward perfection.”
In Great Britain, on the other hand, inventions were much more of a haphazard process, performed by onetime wheelwrights and carpenters competing, rather than collaborating, with one another. Their success did not go unnoticed in France, nor unremarked. In 1824, an École polytechnique graduate named Sadi Carnot wrote Réflexions sur la puissance motrice du feu et sur les machines propres à développer cette puissance—the first theoretical explanation of the thermodynamics of steam power—essentially out of dismay that the great achievements of British engineering had been produced by men, like Watt, with no formal schooling. The snobbery served French science well; less well French innovation. If one secret to sustaining an inventive culture was making inventors into national heroes, it was a secret that didn’t translate well into French. Between 1740 and 1780,36 the French inclination to reward inventors not by enforcing a natural right but by the grant of pensions and prizes resulted in the award of nearly 7 million livres—approximately $600 million today*—to inventors of largely forgotten devices, but Claude-François Jouffroy d’Abbans (inventor of one of the first working steamboats), Barthélemy Thimonnier (creator of the first sewing machine), and Aimé Argand (a partner of Boulton and friend of Watt whose oil lamp became the world’s standard) all died penniless. Other than Joseph-Marie Jacquard, the creator of the eponymous loom, and perhaps the Montgolfiers, the French did not lionize their inventors.37
This didn’t mean they didn’t understand the strategic importance of technology. Carnot himself wrote, “to deprive England of her steam engines,38 you would deprive her of both coal and iron; you would cut off the sources of all her wealth, totally destroy her means of prosperity, and reduce this nation of huge power to insignificance. The destruction of her navy, which she regards as the main source of her strength, would probably be less disastrous.” Competition with Britain alone might have allowed French industrialization to survive the haughtiness that made the nation elevate pure science over its commercial applications, if not for an unfortunate bit of timing. The same year that Joseph Bramah was hiring Henry Maudslay to help build his locks at 124 Piccadilly, several thousand citizens of Paris marched down the Rue Saint-Antoine to the nearly empty prison known as the Bastille. The same year that the British government was certifying a grant of incorporation for Boulton & Watt to design and sell steam engines, the French government was beheading Antoine Lavoisier, the chemist whose research on heat was central to the theory behind those engines.
It was not immediately apparent that the French Revolution would be hostile to invention or inventors. The first law protecting intellectual property in France was passed in 1791, in ringing language that declared,
every novel idea39 whose realization or development can become useful to society belongs primarily to him who conceived it, and that it would be a violation of the rights of man in their very essence if an industrial invention were not regarded as the property of its creator.
Unfortunately for the cause of innovation, the law was abrogated only two years later, as a side effect of the extreme violence of the Terror. And while it was reinstated in 1794, nobody seems to have told the French patent office. They were already playing catch-up in 1792, when Britain granted 85 patents, and France, with a population twice as large, issued 29. In 1793, that number fell to 4. From 1793 to 1800, in fact,40 Britain issued 533 patents to France’s 65. In addition to all their other world-historical effects, the French revolutionists, and the Corsican emperor whose wars were the Revolution’s last chapter, constitute the most important reasons that Britain and America established a thirty-year lead on all other European nations in the development of steam power. As Jeff Horn, whose study of French industrialization is very close to the last word on the subject, put it, “When the revolutionary and Napoleonic wars ended41 in 1815, the British were approximately a generation ahead in industrial technology and in the elaboration of the mechanized factory.”
Looked at through the history of ideas, the French attitude toward invention, and even its revolutionary spirit, share a common origin. To the same degree that Britain’s beliefs about property are traceable to John Locke’s Second Treatise on Government, France’s can be found in the Discourses of the onetime engraver, writer, musician, and philosopher Jean-Jacques Rousseau.* In his First Discourse, for example, Rousseau shared his discomfort with technical progress, which he associated with decadence and moral decline; his Second Discourse argued that the invention of any technology, by demonstrating that some are more gifted than others, promotes inequality and eventually tyranny: “Astronomy was born from superstition … physics from vain curiosity” (First Discourse, volume I). In his “treatise” on education, Émile, Rousseau attacked what has come to be known as amour propre: the invidious striving after excellence in the eyes of one’s fellows. Rousseau’s fetish for compassion and equality have made him a powerful influence on generations of Marxists, but his earliest, and most consequential, impact was on the revolutionaries of eighteenth-century France. When in 1793 the Jacobins closed the Académie des Sciences on the logic that “the Republic does not need savants,”42 they were channeling Rousseau.
And they were hampering their Republic in the race to technological mastery. The finish line for the first stage of that race had been the use of condensed steam to convert atmospheric pressure into the reciprocating motion of Newcomen’s pumps. For the second stage, it was converting the expansive power of steam into rotary motion able to drive dozens, and then hundreds, of spinning and weaving machines.
The third stage was converting steam power into motion. Locomotion.
* This is a particularly strong argument against a belief in progress.
* David Warsh points out, in his own Knowledge and the Wealth of Nations (which has inspired much of this chapter), that Smith’s book was the first of only four dominant textbooks the field has ever known; the only one until Ricardo’s Principles of Political Economy and Taxation, which was followed by Alfred Marshall’s Principles of Economicsand eventually by Paul Samuelson’s Economics. The pattern of successively shorter titles seems finally to be at an end.
* In fact, the same process works in reverse, though at the expense of the pyramid metaphor. Not only does a larger population result in more inventions, the wealth created by those inventions permits even more population growth, and yet another Malthusian trap.
† In units called Geary-Khamis dollars, each of which is equivalent in purchasing power to a U.S. dollar in 1990, and which is the benchmark “currency” of choice for really long time ranges.
* Perversely enough, nineteenth-century Russia was both too big and too small. In 1766, the brilliant Ivan Polzunov, inventor of the world’s first two-cylinder engine, died of tuberculosis, and when his engine, which operated the bellows in Russia’s largest forge, needed repair three months later, no one could be found who understood how to reassemble it.
* It is even harder to calculate the current value of eighteenth-century French currency than British. From 1726 on, an ounce of gold was set equal to 92.5 livres; during the same period, an ounce of gold cost £4.10. Seven million livres, therefore, equaled about £310,000. Using the index of average earnings, this is more than £400 million today.
* The Locke versus Rousseau debate remains one of academe’s almost preternaturally popular, used to explain everything from the differences between modern and medieval perspectives to the reason for Britain’s reluctance to adopt the euro.