The invention of the military-industrial complex

By that point, Robert Whitehead had made several significant improvements to his design. In 1875, he replaced his original two-cylinder engine with a three-cylinder version designed by the British engineering firm of Peter Brotherhood. The original single screw gave way to contra-rotating propellers, and Whitehead introduced a steering engine to amplify the effect of the depth mechanism on the horizontal rudders. In 1889, Whitehead began to build 18-inch (diameter) torpedoes in addition to his standard 14-inch model. By the mid-1890s, his torpedoes could make almost 30 knots for roughly 800 yards. The application of an invention known as the Obry gyroscope (named after the inventor, Ludwig Obry) to torpedoes in 1896 supplied a horizontal guidance system and began their transformation into accurate, high-speed, long-range weapons. Several years before the outbreak of World War I, torpedoes could travel at a speed of 45 knots (51 miles per hour) or run 10,000 yards (5.6 miles). To put those numbers in perspective, Glenn Curtiss, the great American engineer, won the premier airplane racing event of 1909 by flying 47 miles per hour for 12.4 miles—and, of course, he did not have to contend with water resistance. Over a fifty-year period, the speed of torpedoes had increased by roughly 800 percent, and their range by 5,000 percent. They were at the cutting edge of technology.

While torpedo technology changed, so too did the platforms for launching them. Indeed, the half-century before World War I may have witnessed more technological change for navies than any period before or since. The basic outlines are well known. Through the Napoleonic Wars, naval vessels were powered by wind, were made of wood, and fired muzzle-loading smooth-bore cannons maneuvered on carriages. In the mid-nineteenth century, they began a rapid transformation. Propulsion changed from sails powered by wind to engines (first reciprocating, then turbine) powered by fossil fuels (first coal, then oil). Wooden hulls were clad with iron and then replaced entirely by steel, increasing their ability to withstand artillery hits. Muzzle-loading smooth-bore cannons on carriages gave way to rifled breech-loading cannon on mechanized mounts (first hydraulic, then electric), which could shoot farther, more accurately, and more quickly. The growing ability of warships both to endure and to deliver artillery hits involved a celebrated race between armor and armament. Perhaps less well known, yet just as important, were changes in communications and targeting technologies. Navies experimented extensively with telegraph cables and radio for controlling movement at the strategic, operational, and tactical levels. The greater range and accuracy of modern guns were of little use if they could not be aimed and controlled, so navies also developed better targeting (also known as fire-control) systems, which were among the world’s first analog computers.

In addition to capital ships, smaller vessels also carried torpedoes. Torpedo boats, which many navies began to build in the 1870s, were the first vessels designed to use torpedoes as their primary weapons system. A short-lived type of vessel known as the torpedo catcher (or the torpedo gunboat) was developed in the 1880s to defend fleets against torpedo boats, but it soon became clear that the catchers lacked the speed to catch their prey. The most durable type of vessel to emerge in direct response to torpedo development was the torpedo-boat destroyer, better known as simply the destroyer, which began to appear in the early 1890s. Originally intended to take on the defensive mission of the torpedo-boat catchers, destroyers soon showed offensive promise as torpedo boats themselves. Indeed, their greater size, durability, and sea-keeping ability made them better platforms for launching torpedoes than the torpedo boats had been. When firing torpedoes, destroyers used above-water, not submerged, tubes.

Perhaps surprisingly, submarines played little role in driving torpedo development before World War I. France led the way on submarines, introducing the first recognizably modern version in the early 1890s. It was followed by the United States and Great Britain around 1900. (Despite its later association with submarine warfare, Germany actually lagged in submarine development and likely had to rely on pirated French designs.) Prewar submarines had limited utility as torpedo platforms. They were not true submarines but submersibles, spending most of their time on the surface of the water and submerging only to attack a target. Most submarines lacked sufficient surface speed to accompany battle fleets (i.e., they were not fleet-keeping submarines), which moved above 20 knots by 1914, and instead were confined to coastal patrol. They expected to fire their torpedoes at point-blank range of hundreds rather than thousands of yards. Thus it was surface vessels, especially capital ships, and not submarines that drove the development of faster, more accurate, and longer-range torpedoes.

Like many other armaments, torpedoes were built and sold in a global marketplace, featuring (like so many of today’s markets) multinational corporations and transnational flows of capital, ideas, and technology. There were four international producers, who were distinct from those who built for just one country. The first, and most important, was the Whitehead factory in Fiume, which signed its first contract (with Austria-Hungary) in 1868 and its first international contract (with Britain) in 1871. It eventually sold torpedoes to twenty-three countries before World War I. The second was the Berliner Maschinenbau Aktiengesellschaft (BMAG). It was sometimes referred to as the Schwartzkopff Company after its founder, who most likely stole plans from Whitehead in 1873 and began producing a near-duplicate of his torpedo shortly thereafter. BMAG sold to Japan, China, Spain, Sweden, and Germany—until the mid-1880s, when the German Navy ceased to buy from BMAG and instead built all its torpedoes in state-owned factories. The third international producer was the Whitehead factory in Weymouth, England, which was originally established in 1890 to build solely for the British Navy but eventually sold on the open market. Finally, France’s Schneider Company (better known for its guns) began to sell torpedoes internationally, but very little about its torpedo business is known.

The international arms market had several distinctive characteristics. First, a number of armaments firms (like Whitehead) were multinational, with branches in more than one country. Some firms were subsidiaries of larger foreign conglomerates. In 1906, for instance, the great British armaments firms of Vickers and Armstrong-Whitworth purchased the Whitehead Company, including both its Fiume and Weymouth branches. Second, the line between public and private, and thus between state and nonstate actors, was blurry. Governments often operated quasi-private armaments factories to preserve security or stimulate competition, while private firms often received substantial investments from governments, making them quasi-public. Third, the armaments business usually required large upfront capital investments, and thus the number of producers within a given country was limited. Sometimes a single firm had a monopoly on a particular product (as with Krupp in German naval gun production), or a small number of firms had an oligopoly (as with Germaniawerft and Schichau in German torpedo-boat production). Finally, given the specialized nature of the goods being produced and the occasional ban on exporting, there was often just one consumer—the government—creating a so-called monopsony.

Under these conditions, producers faced several challenges. Not only did entering the armaments business require large capital investments, but so too did the constant plant upkeep to remain in the business. Demand was unreliable without a diversified consumer base. If a monopsonist government decided to stop purchasing, for whatever reason, demand collapsed. Government demand itself depended on unstable factors, like financial wherewithal and favorable tactical, strategic, and diplomatic circumstances. Monopsony empowered the consumer to set prices and specifications while depriving producers of leverage to protest. Producers often responded to their vulnerability by combining into rings or cartels.

Monopsonies notwithstanding, consumers faced difficulties as well. If the producers did not find many consumers, neither did the consumers find many producers. To stimulate competition—and thus, in theory, to obtain better products at lower prices—consumers had three basic options. One was to entice more private firms into the business. This task was not easy, despite the potentially lucrative rewards: for the reasons explained above, any intelligent firm would think twice about entering the armaments business. Overcoming firms’ reservations usually required both a cash subsidy (whether in the form of direct injections or payment of artificially high prices) to help firms acquire the necessary start-up capital, and the promise of contracts to assure firms that they would receive returns on their investments. If governments were unable or unwilling to make large financial outlays or to promise orders to private firms, they could adopt the second option for stimulating competition, which was to establish a government factory. The globe was dotted with such plants: the US Navy’s torpedo factory in Newport, Rhode Island; the Royal Navy’s torpedo factory in Greenock; the Japanese arsenal in Kure; the French gun plant at Ruelle; the Russian iron works at Putilov; and the Austrian shipyard at Pola. Of course, these plants also required large financial outlays.

The third option for stimulating competition was perhaps the one most fraught with potential pitfalls: to allow private firms to sell internationally. By doing so, governments effectively gave up their monopsony. The market was flooded not only with additional consumers but also with additional producers because the armaments firms now had to compete with producers in other countries for international customers. Governments could then reap the benefits of international competition in their own countries. Even in the absence of any imperative to stimulate competition, governments might allow firms to sell abroad in order to keep the firms in business at lower cost to themselves. In effect, allowing companies to court foreign buyers stabilized demand, meaning that their home governments did not have to inflate demand artificially through subsidies or unnecessary orders.

Despite such advantages, a significant drawback of this approach is easy to see: allowing armaments firms to sell abroad eroded secrecy. It was possible to minimize that risk by erecting various safeguards—for instance, by physically quarantining especially sensitive parts from the production of less sensitive ones, or by providing for damages if secrecy was breached— but it could not be eliminated. Thus, as we shall see in the following chapters, the global arms market offered benefits, but with costs.

Inventing the Military-Industrial Complex

Beginning with the introduction of the gyroscope in the mid-1890s, the growing accuracy, speed, and range of torpedoes posed grave challenges to conventional naval tactics. Traditional naval tactics called for capital ships sailing in close order and following visual signals from their leader to defeat their counterparts with heavy guns fired at point-blank range. Ships proceeding in close order and engaging at short ranges were extremely vulnerable, however, to torpedo fire. To deal with the torpedo threat, navies experimented with new formations, such as moving ships further apart in the line of battle or even breaking the line of battle into independent divisions, but the new formations created serious command-and-control problems. Navies also experimented with longer battle ranges to stay out of torpedo range, but the greater distances made it more difficult to achieve accurate gunfire. To cope with this challenge, navies sought to improve both their guns and their gunnery fire control. The result was a race for range between guns and torpedoes that raised the possibility that the entire system of tactics built around capital ships armed primarily with big guns would give way to one built around smaller vessels primarily armed with torpedoes.

The implications of torpedo development were equally profound at the strategic level. Traditional naval strategy, as elaborated in previous centuries by the Royal Navy, called for close blockade of enemies’ coasts to stifle their trade combined with decisive battle to destroy their fleets and achieve full command of the sea. Torpedoes threatened both aspects of this system. Expensive capital ships were so vulnerable to torpedo attack by cheaper vessels in battle that fleet actions could seem too risky. Ships engaged in close blockade were overly vulnerable to torpedo attack by surface torpedo vessels under cover of darkness or by submarines at any time. One option was to move the blockade farther from the enemy’s coast, but distant blockade (sometimes called loose blockade) was more difficult to enforce and was considered questionable under international law. By threatening to deprive navies of battle and blockade, torpedo development forced nations to look for fundamentally new ways of defining and applying naval power.

Thus, torpedoes played an important role in the intense naval competition preceding World War I. Navies everywhere poured enormous resources into increasing and conserving their relative power. In a classic example of a challenge-and-response dynamic, no sooner did one navy get a piece of technology than another navy invented a new piece of technology that rendered the former technology obsolete—and with it the massive peacetime investment needed to produce the technology on an adequate scale.

The depreciation of peacetime investment was particularly problematic for navies. Until recently, naval warfare was far more technologically sophisticated than land warfare and required correspondingly greater peacetime investment. “You can go round the corner and get more guns, more rifles, more horses, more men who can ride and shoot,” as Admiral Sir John Fisher once said, “but you can’t go round the corner and get more Destroyers and more Cruizers [sic] and more Battleships.” Lord Kitchener, Britain’s War Secretary for the first two years of World War I, confirmed Fisher’s claim: Equipping the British army, he claimed, “was not much more difficult than buying a straw hat at Harrods.” With so many resources sunk into naval power, representing such a high opportunity cost, the stakes were higher in the event of failure.

Industrialization exacerbated this dynamic, and torpedoes epitomized the process. Although a steamship is the more familiar symbol of industrialization at sea, a torpedo is at least as good a symbol: like steamships, torpedoes were metal and ran on engines, but torpedoes could be produced in much larger numbers because they were relatively small and inexpensive compared to ships. Even as the miniaturization of torpedoes enabled them to be produced in bulk, however, it posed serious design and production challenges. Consider these figures: in an 1882 contract for Whitehead torpedoes, the Austrian Navy required that the margin of error on an overall length of 4.415 meters not exceed 5 millimeters (0.005 meters), and that the margin of error on an overall diameter of 35.6 centimeters (0.356 meters) not exceed 2 micrometers (0.0002 meters). On that order, precision meant margins of error within four decimal places and 0.001–0.0006 percent of overall sizes. Miniaturization on that scale was not easy, and it was all the more difficult in view of the number of parts that had to be crammed into a torpedo. Consider some additional figures: whereas the standard small arm used by the US Army before World War I (the 1903 Springfield rifle) contained ninety parts, the standard torpedo used by the US Navy at roughly the same time contained about 500 parts—in the guidance systems alone.

Given the many small, precisely machined, and tightly fitted pieces of metal that composed torpedoes, sending a prototype into production without putting it through a rigorous research and development (R&D) process could easily create manufacturing, quality control, assembly, and operational nightmares. The small size and relatively cheap per-unit cost of torpedoes did not spare them from the need for an expensive R&D process. In fact, miniaturization and large-scale production made it all the more necessary.

In these respects, torpedoes likely represented a cluster of devices sometimes called control technologies, and they have attracted relatively little interest from scholars. Although historians of that problematic late-nineteenth-century phenomenon known as the Second Industrial Revolution have moved well beyond the classic focus on railroads, electricity, and chemistry, naval historians still tend to study big things, often created by big corporations: armor, guns, and propulsion. If taken too far, this focus crowds out equally important narratives about smaller technologies, built by smaller businesses, that made the big stuff smart—control technologies in communications, data collection, and information processing, which together formed the nervous system for the heavy exoskeleton of the industrial beast. In navies, control technologies included targeting and guidance systems (both of which relied on cutting-edge gyrostabilization) and radio, which had different manufacturing requirements and were built by different types of firms compared to armor, guns, and propulsion. Perhaps most important, these control technologies, like torpedoes, required miniaturization on a scale that many other industrial technologies did not. Although the exact mixture of engineering challenges posed by torpedoes was unique, more generally those challenges typified an important class of industrial technology that has been under-studied by historians.

Solving the challenges presented by industrial technology like torpedoes required a distinctive type of innovation, in which numerous activities occurred together rather than discretely or sequentially. Take basic science and applied science. Although the basic scientific principles at work behind industrial technology may not have been qualitatively more difficult than those behind preindustrial technology, they grew in quantity as the technology grew in sophistication. For instance, the science behind air flow in torpedo propulsion, which rested on the ideal gas law, was in some sense very simple, but applying it depended in part on the metal used for pipes and valves, which had their own chemical science of metallurgy. Discovering a particular scientific principle was easier than combining it with other relevant principles and applying the result in order to create effective technology. Given the difficulty of the latter, basic science sometimes lagged behind applied science (or science sometimes lagged behind technology), reversing an idealized path of scientific-technological progress. To return to the propulsion example, even if the ideal gas law and metallurgical chemistry were not perfectly understood, it could still be possible to build a propulsion system that worked well enough (bearing in mind that the phrase well enough itself constituted a dependent variable), and perhaps later to deduce the underlying science from the technology. Thus, it was possible to have technology-led science as well as science-led technology.

Similarly, invention, development, and production could occur at the same time, conducted by the same people in the same spaces. Contemporary actors struggled to define these activities, the boundaries of which could have legal and financial implications. Did invention consist of coming up with a good idea, or did it consist of embodying that idea in a workable design? Did development end when a torpedo entered production, or did it continue when the design was tweaked during the torpedo’s acceptance tests? Or was tweaking the design invention rather than development? Attempting to distinguish these activities from each other risks not only over-simplifying a complex historical reality but also obscuring the self-interest behind certain distinctions. When innovators seeking patents came up with a good idea but lacked the resources to turn it into a working prototype, it was in their interest to define their contribution as invention and to define others’ contributions as “mere” development. When innovators seeking monetary compensation turned a good idea into a working prototype, it was in their interest to define invention in terms of labor and risk rather than in terms of coming up with a good idea. These issues may reasonably be characterized as being among the ontological and epistemological implications of industrialization.

As if these supply-side problems were not formidable enough, the demand side presented its own challenges. (Of course, those on the demand side— navies—were also on the supply side, engaged in invention, development, and production themselves.) Although many of those demanding torpedoes understood that the weapon had the potential to revolutionize tactics and strategy, determining exactly how that potential would translate into reality was extremely difficult. Even the best guesses had to contend against institutional factionalization in both the American and British navies, and agreements about the desired performance characteristics of torpedoes were temporary. Thus, the specifications that producers had to meet were not only strict but changing. Volatility characterized both the consumption and production environment.

In their ideal world, navies had unlimited resources and could invest heavily in all aspects of innovation to mitigate this volatility. In the real world, navies’ resources were limited, and they had to make choices, all of which came with trade-offs. For instance, slowing production in favor of continued R&D risked having too few weapons in service when a crisis hit, while short-changing R&D in favor of production practically guaranteed more hiccups during the production process and problems with the weapons once they entered service. In the key sector of naval-industrial R&D infrastructure, Britain was far stronger than the United States, despite the traditional depiction of a declining Britain and a rising United States during this period. As a result, Britain was better able to perfect existing technology and test new technology thoroughly, while the United States had to take technological gambles. Precisely this pattern occurred with torpedo technology.

The effort to create an R&D infrastructure capable of developing successful torpedoes profoundly changed the relationship between state and society in the United States and Britain. The historian William McNeill associated this change with the emergence of command technology: technology commanded by the public sector from the private sector that was so sophisticated and expensive that neither possessed the resources to develop it alone. As a result, they had to collaborate, meaning that, while such technology was commanded in the sense that government fiat replaced the market, it was not commanded insofar as governments required the cooperation of the private sector. Indeed, far from the smooth hierarchy perhaps implied by the metaphor of command, this cooperation could be extremely messy, for reasons alluded to above: both parties had leverage, and it was impossible to distinguish neatly among the various activities (science, invention, development, and production) involved in their collaboration.

McNeill’s thesis had three major implications. First, command technology put a premium on the development of a kind of technology—which I will call servant technology—that could generate information needed to improve command technology. Second, the information generated by servant technology was itself a commodity because it had the power to affect market relationships by offering insight into the value of command technology. This commodified information was a distinctive kind of property. Third, the collaboration between the public and private sectors required to develop command technology raised fundamental and complex questions about the nature of property in relation to invention. When more than one party helped to invent a piece of technology, how could ownership of the intellectual property rights be established?

Answering this question generated serious friction between the public and private sectors. Conventional contract language, patent procedures, cost accounting methods, and pricing assumptions provided little guidance, because they were based not on the new collaborative procurement paradigm but on an older one, in which the public sector bought finished goods from the private sector as ordinary commercial products. In a series of legal battles over which side owned the intellectual property rights to technology that both had helped to invent, the governments won. To do so, they exploited two aggressive new legal strategies: applying eminent domain to intellectual property; and using anti-espionage legislation to control exports, that is, to regulate private commercial and proprietary rights—notwithstanding the fact that the legislation had been written for very different purposes. In every case, contractors protested that cutting them off from the global market would damage their property rights, but governments insisted that permitting private actors to share technological information freely would aid the governments’ enemies. Courts tended to lose sight of private property rights when national security seemed to be at stake.