10
Labs in the Woods: The Quantification of Technology During the Late Enlightenment

By Svante Lindqvist

The concept of quantification can be considered an intrinsic component of technology. Two basic characteristics of technology in modern Western civilization are technical efficiency and economy. Practitioners of technology had labored to increase technical efficiency and economy long before the concept of quantification began to be applied to technology. The latter occurred during the 18th century when technology began to borrow its methods from science, that is, when attempts were made to study technical reality by systematic experimentation and quantification in fixed units using precision instruments. These attempts first met with success after 1770.

The phrase "systematic experimentation and quantification in fixed units using precision instruments" describes a laboratory—a table in a quiet, secluded room with sufficient illumination and heat. Such laboratories became more common in universities, academies of science, and affluent homes in the 18th century. But it was quite a different matter to transfer these ideal conditions from the natural philosopher's laboratory to the technical world outside. Where the

laboratory was neat, technology was messy. The central component of the experimental method, controlled experiment, was difficult to achieve in technology, with its comparatively large scale. To apply quantitative methods to technology required a degree of control over the material and social world beyond the means of individuals. The late 18th century saw the emergence of institutions with authority and competence to exercise such control.

Three case studies will illustrate general characteristics of this transformation of technology. Our image of 18th-century technology has by and large been shaped by those technologies seen subsequently as spectacular and/or efficient. But the ingenious and complex machines depicted in books of the 18th century—like those depicted by Ramelli, Zonca, Besson, and others in the Renaissance—were often no more than bold, futuristic speculations directed at a restricted audience, irrelevant to the daily reality in which most people lived and worked. We may identify numerous forerunners to 19th-century industrialization, but technological reality in the 18th century lay a long way from the elegant engravings. The average 18th-century man lived by the sweat of his brow, aided by simple wooden implements at a low level of mechanization. Only occasionally was his work made easier by machines powered by his fellow-men or by oxen, horses, water, wind, or steam. As a period in technological history, the 18th century was characterized by the use of wood, water, and work, and this is reflected in the choice of examples for this study.

Water Power

Industries dependent on mechanical energy in the 18th century included mines, blast furnaces, tilt hammers, other metal works, saw and paper mills, gunpowder factories, brickworks, oil mills, and glassworks. Such industry required high power (work per unit of time) and continuous operation ; the extent to which these requirements could be met by traditional sources of power varied. Muscle power (animal and human) provided continuous operation at limited output. Wind produced high power, but proved inappropriate for industrial

production since continuous operation could never be guaranteed. Only water power fulfilled both requirements. But existing hydraulic resources were limited. In order to utilize fully the available water power, the efficiency of waterwheels had to be increased. For each type of water mill a particular speed gives maximum power. Finding the rules for determining this speed was a problem to which solutions were sought throughout the 18th century. The design of waterwheels was the subject of vigorous debate in learned journals and technical literature in the 18th century, although theoretical considerations had "no direct effect. . . on the construction and installation of wheels," which were governed by a tradition little changed until the end of the century.^[1]

Two approaches characterized 18th-century attempts to establish general rules for the most efficient design of waterwheels.^[2] In the deductive approach mathematical analysis permitted derivation of general rules from fundamental laws of motion. The inductive approach used systematic experiments with parameter variation and optimization to achieve the same end.^[3] Great expectations were matched by great investments of work; the rate of return, at least until the early 19th century, was disappointing. The deductive method resulted in formulas too complicated for ready application. The inductive method yielded unmanageable amounts of unreliable quantitative data. A waterwheel turning peacefully in a stream proved far more complicated than the heavenly clockwork.

Around 1700 French scientists in the Academy of Sciences in Paris made the first analysis of waterwheels in dynamical terms. In his comprehensive history of the waterwheel, Terry Reynolds distinguishes five approximate and overlapping periods following the work in Paris.^[4] The first was the establishment of theoretical analysis by

academicians like Parent and his followers between 1700 and 1750. The second was a period of experimental work from 1750 to 1770 by engineers such as De Parcieux, John Smeaton, and Charles Bossut. A third, simultaneous development was the attempt during the same period by Johann Euler and Charles Borda to reconcile the discrepancy between theoretical predictions and experimental findings. This next forty years, 1770–1810, Reynolds labels "the era of theoretical confusion."^[5] This confusion abated during the fifth period, 1810–50, when Borda's analysis of 1767 achieved general acceptance.

Facilitating the transition from general theoretical confusion to the reconciliation between theory and experiment were institutions capable of large-scale experiments beyond the means of individuals. The importance of such institutions in water power technology during the late Enlightenment and early 19th century is underscored by comparing a successful Swedish attempt to quantify water power with an earlier, unsuccessful one. From 1701 to 1705 the well-known Swedish inventor Christopher Polhem performed some 25,000 experiments with a hydrodynamic apparatus of his own design. But this heroic experimental effort produced only meager results. A century later, between 1811 and 1815, the Swedish Ironmasters' Association (Jernkontoret ), a private organization of independent ironworks established in 1747, undertook a major investigation of hydrodynamics.^[6] Originally intended as a credit agency for the ironworks, the Association assumed an important role in technological development toward the end of the 18th century. Its officers exercised quality control over the various stages in the process of iron manufacturing, and the Association financed a number of large development projects that were

too expensive for any individual ironworks. A large experimental apparatus had been built at the Great Copper Mine in Falun in 1804, which was originally intended for investigating the efficiency of winding gear. However, the metallurgist Eric Thomas Svedenstierna and others supported the idea that the Swedish Ironmasters' Association should finance a lengthy series of experiments in order to establish a general theory of water power.^[7] The resulting hydrodynamic investigation of 1811–5 stands in glaring contrast to Polhem's experiments.

Theoretical, instrumental, and, most importantly, institutional factors led to this successful quantification of water power technology during the late Enlightenment. A more profound theoretical framework was available by the end of the 18th century; a heightened appreciation of the applications of mathematics was reinforced by a broader and more critical knowledge of the international literature in the field and an awareness of fundamental principles of experimentation. The institutionalization of science contributed to this conceptual change: technological innovation became verbalized and was documented in journals and monographs produced and distributed under the auspices of new scientific institutions. A higher degree of precision in scientific instruments resulted from a more vigorous market for scientific apparatus and information. In the social organization of technology, responsibility shifted from individuals to institutions. This last was a sine qua non for applying the concept of quantification to the expanded spatial and temporal dimensions of technology. The importance of these three factors in the Swedish case is described in the following subsections.

Theoretical

Although the experimental apparatus used by Lagerhjelm for the investigations of 1811–5 resembled that of Polhem a century earlier, there the resemblance ends. Polhem's experimental apparatus seems to show the influence of the French physicist Edmé Mariotte, who

had carried out experimental studies of water and wind mills.^[8] Certainly they adhered to the same empirical tradition. Like Mariotte, Polhem was more concerned with articulating and applying generalizations based on experiments than reducing them to fundamental principles; both relied on common sense to guide their reasoning.^[9] Although Polhem's work contains an early example of parameter variation and optimization,^[10] he was never able to convert his many experimental results into general rules for the design of waterwheels.^[11] Nor did he fully appreciate the merit of mathematical analysis of hydrodynamic phenomena. This is evident in his faint praise for the work of his younger colleague Pehr Elvius, who in 1742 published A mathematical treatise on the effect of water mills .^[12] Polhem commented in the Proceedings of the Royal Swedish Academy of Sciences: "Although [Elvius'] book is really written for the learned, who are already familiar with the modern mathematics , which by its discoverer the learned Leibniz is called calculus differentialis and by Newton, fluxio curvarum , so does yet Mr. Elvius show his profound knowledge of such puzzling matters, that he gives hope of becoming a good Mechanicus with time, as well in Practice as now to begin with in Theory."^[13] What Polhem had considered "modern mathematics" was a standard tool in the hands of the mine official Pehr Lagerhjelm, whose report on the hydrodynamic experiments of 1811–5 financed by the Swedish Ironmasters' Association was a highly mathematical treatise.^[14] Lagerhjelm's report also

included a thorough, critical review of relevant international literature. The first fifty pages of the second volume commented on the works of Smeaton, Euler, Borda, Bossut, Banks, Langsdorf, and others.^[15]

Lagerhjelm's treatise also evinces a higher level of conceptual awareness. In his preface to the second volume, Lagerhjelm offered an epistemological program to relate theory and experiment for hydrodynamics. His ideas bear a certain resemblance to Kant's theory of knowledge, and the terminology—"phenomenon," "form," and "content"—is similar. For Lagerhjelm, inductive reasoning cannot produce conclusions of universal validity, because "abstractions from a given experience. . .are only valid under the circumstances and within the boundaries essentially associated with the class of phenomena one experienced."^[16] The implication was clear: the inductive method followed by Polhem and others, with their thousands of experiments throughout the 18th century, was epistemologically pointless. So, too, was the deductive method, the "speculative root" of knowledge in which Elvius and others had placed their confidence, inadequate in and of itself. The path to truth required a synthesis between "form" and "content," specifically, theory and experiment.

Instrumental

Inaccurate measurements compromised Polhem's data. Because the pendulum in the clock he used was not of the proper length, he arrived at incorrect values for the speed of the waterwheel and hence incorrect values for the output. Polhem tried in 1710 to reduce all these figures to their proper value by means of a correction coefficient, but found the work "so difficult and tedious, that no amount of patience would have sufficed." More seriously, the protractor he used to measure the inclination of the water trough gave different readings as the waterwheel was placed at different levels.

Polhem confided to his assistant: "In fact between ourselves, this work is as useful as a fifth wheel on a carriage."^[17]

Lagerhjelm proclaimed explicitly his awareness that calibrated precision instruments were essential if the experiments were to be reproducible and the results of general value. He made linear measurements using "a precisely graduated decimal scale two Swedish feet in length" produced by Johan Gustaf Hasselström, purveyor of mathematical instruments to the Royal Swedish Academy of Sciences. Lagerhjelm also employed "a set of weights calibrated against the Swedish original standard, which is kept in the Archives of the Royal Treasury," and a balance constructed by Gabriel Collin, manufacturer of optical instruments for the Academy, and watched a clock borrowed from the astronomical observatory of the University of Uppsala.^[18]

In quality of instruments, Lagerhjelm enjoyed a significant advantage over Polhem. Polhem had used the most accurate instruments he could acquire or construct. Over the course of the 18th century, however, a real market for scientific instruments had developed in Sweden. The market was largely the creation of the Royal Swedish Academy of Sciences, established in 1739. Rivals for this market competed in precision. Instrument-makers like Daniel Ekström, Hasselström, and Collin won the right to call themselves "Purveyors to the Royal Swedish Academy of Sciences";^[19] this distinction implied to prospective customers that every instrument produced in their workshops promised the highest possible degree of precision. Market pressures did thus increase the degree of precision, and the market itself was a result of the establishment of the Academy. Before then, no Swedish instrument-maker could acquire such status; hence there had been little or no competition in degree of precision.

In this way, the institutionalization of Swedish science contributed to the increased degree of precision in scientific instruments during the late 18th century, a development that contributed to the quantification of technology.

Institutional

The establishment of the Royal Swedish Academy of Sciences contributed, as mentioned above, to a theoretical and instrumental development. But there was also an institutionalization of technology within the largest and most important of Swedish industries, the iron industry: the establishment of the Swedish Ironmasters' Association, which reflected increased interest in general technological problems during the late 18th century.^[20] The general importance of this change in the social organization of technology can be described as a shift in responsibility from individuals to institutions . Technological projects were now being undertaken by and with the competence and authority of the institution.

The competence of an institution, with its hierarchical structure based on academic merit, is apparent in the case of Pehr Lagerhjelm, the leader of the project of 1811–5. He had studied at the University of Uppsala, where he passed the mining examination (Bergsexamen ) in 1807. This university degree, established in 1750, had become a prerequisite for officials in the service of the Board of Mines, the governmental department that exercised ultimate authority over the mining industry. The degree required examinations in physics, mechanics, geometry, chemistry, and law. After graduation from the University of Uppsala, Lagerhjelm was duly appointed to the Board of Mines in Stockholm. He became a pupil of the chemist Jöns Jacob Berzelius, and assisted him in calculating the percentage composition of nearly 2,000 chemical compounds. This work was published as a supplement to the third volume of Berzelius' textbook, Lärbok i kemien .^[21] In 1808, Lagerhjelm was appointed under-secretary of the

Swedish Ironmasters' Association. He thus reached his position in 1812 as leader of the hydrodynamic experiments by making a career within the formally organized educational system of the Swedish mining bureaucracy.^[22] The merit of the system is proven by his many other contributions to Swedish technology during the early years of the 19th century.

By contrast, Christopher Polhem had gained his position as a favorite of Karl XII in the time of the Absolute Monarchy.^[23] Polhem had been appointed director of the Laboratorium mechanicum , the section of the Board of Mines with designated responsibility for research and education in mechanical technology. It was here that Polhem's hydrodynamic experiments were undertaken from 1701 to 1705. Although a mechanical genius by any standard, Polhem lacked the broad formal university education Lagerhjelm enjoyed. His investigations were therefore carried out within the context of a more narrow theoretical perspective and were influenced by important personal idiosyncrasies—including his disdain for mathematical analysis.

Comparing these two examples also illustrates the importance of institutional authority . The project in 1811–5 involved several persons, all highly qualified. Although they did not cooperate throughout the many years of the project without personal friction, the authority of the Swedish Ironmasters' Association led to the completion of the project and to the publication of the results in two volumes in 1818 and 1822.^[24]

On the other hand, Christopher Polhem submitted to the Board in 1705 what he called an interim report; in fact, it was the only report he ever prepared concerning his hydrodynamic experiments. When he resumed work with the experimental apparatus in 1710, he made the distressing discovery that two crucial quantities had been measured inaccurately throughout the whole series of measurements.

This rendered the results incommensurable and the whole series of experiments nonreproducible. But Polhem, whose individual reputation exceeded his relatively subordinate position as an official in the Board of Mines, had no difficulty in forbidding his assistant to mention to anyone that the data in the report were useless. An institution stronger in terms of hierarchical authority would have insisted that he submit a final report on the experiments for which he had already received his fees, and, on discovering that the data were useless, demanded that he repeat the experiments.

Charcoal-Burning

We tend to associate the 18th century with the use of coal and iron, and to look back on the 16th and 17th centuries as, in John U. Nef's phrase, "an age of timber."^[25] But the growing importance of coal technology, especially in England, should not obscure the dependence on forests of virtually all aspects of material culture during the 18th century.^[26] Not only did mining consume large amounts of wood; so, too, did potash plants, tanneries, glassworks, saltpeter works, train-oil works, lime production, and other industries rely on the forests for fuel and raw materials. Domestic demands included fuel for heating houses and drying grain and malt, and timber for houses, fences, ships, carts, barrels, and agricultural implements. This vital natural resource, however, was perceived to be running out in 18th-century Europe. The fear of imminent shortages spurred both legislative actions and interest in technical improvements aimed at reducing the number of trees felled.

In the 18th century, Sweden—a country devoid of fossil fuel resources for all practical purposes—was gripped by general anxiety about a dearth of timber.^[27] It was believed that the forests were laid

waste by excessive felling: "many large areas of the realm are in danger of soon becoming desolate because of the shortage of timber and. . . the mines and towns in many parts of the country are likewise threatened with ruin that cannot long be postponed if an early remedy is not found." In the worst-case scenario, "the fatherland will in the course of time be reduced to a miserable condition."^[28] Whether or not the fear of timber shortage was well founded does not concern us here. What matters is that the belief in imminent shortage was widespread and influential.

The production of bar iron accounted for about 70 percent of Sweden's exports in the 18th century. About 50,000 tons of bar iron were produced every year, and every stage of production required much timber. Blast furnaces and forges consumed charcoal equivalent to three million cubic meters of timber a year. Charcoal-burning amounted to half of the total industrial consumption of timber.^[29] There were two alternative methods of charcoal-burning: stacking the wood either horizontally in piles called liggmilor or vertically in resmilor . Despite the fear of a forest shortage and the great consumption of wood for charcoal-burning, evaluations of the two methods of extracting charcoal from wood were not easily undertaken; the question was not resolved until the early 19th century.

In 1811 the Swedish Ironmasters' Association financed a series of experiments to assess the relative merits of resmilor and liggmilor and to determine the pile design yielding the maximum amount of charcoal. The report was published as a monograph three years later by the mining official Carl David af Uhr.^[30] It was an impressive volume, comparable in scope and depth to the report on the hydrodynamic experiments undertaken at the same time, also with the support of

the Swedish Ironmasters' Association. Uhr's report described the series of experiments in charcoal burning carried out at Furudahls Ironworks in the province of Dalecarlia during the years 1811–3. No less than forty full-scale piles of different types were tested, with twenty parameters recorded for each pile (fig. 10.1). The systematic study was meticulously planned: for example, a specially designed tool was used to measure the diameter of the billets at both ends, in order to compensate for their taper ("frustra of a cone, as they truly are") when calculating the volume of a pile.^[31] The volume of the piles was measured in cubic ells to one or two decimal places. The author discusses the effect of various errors in measurement and ways to compensate for these errors. Output, measured in cubic ells of charcoal, was correlated to the total labor, measured in man-days and horse-days, needed from the day the trees were felled to the day the charcoal arrived at the ironworks. Calorimetric experiments helped determine the quality of the charcoal, prompting Uhr to discuss Lavoisier's opinion of the role of oxygen in combustion. He handled this question with the same facility he showed in computing the number of horse-days needed to build a charcoal pile. Results of this study were summarized in a handbook for charcoal-burners that appeared in three editions—a measure of its success.^[32]

One looks in vain for quantitative methods in the literature on charcoal-burning at midcentury. In assessing the relative merits of resmilor and liggmilor the quantity of wood was not specified; nor was it clear which units were used in measuring the charcoal.^[33] Consider Magnus Wallner, who published at his own expense A brief account of charcoal-burning in Sweden , a Swedish translation of his dissertation under Celsius at the University of Uppsala.^[34] The work

contained a brief description of methods and tools, statements by charcoal-burners, a few quotations from foreign literature on the subject, and some of Wallner's own ideas. Entirely lacking was any attempt to give quantities in defined units and to carry out systematic experiments under controlled conditions.

The failure in applying quantitative methods to charcoal-burning may be attributed to the lack of institutions able to recreate a laboratory environment in the forest. No retort on a laboratory bench could reveal the best type of pile for charcoal-burning. "Systematic experimentation" required building many piles of different types and supervising them day and night for several weeks. It was necessary to take into account the species, age, and moisture of the wood; the length and thickness of the billets; the stacking pattern; the total amount of wood; the outer dimensions of the pile; and the ignition method. After the piles were pulled down, the charcoal had to be shoveled into barrels of known size—"quantification in fixed units." All this was far different from laboratory work. It differed first in spatial terms—not only the size of the piles but also the area of the forest required for the tests was large. The temporal requirements were also of a different order. Because building, watching, and pulling down a pile took more than a month, a series of systematic experiments might stretch over several years. Such was the case with the investigations carried out by the Swedish Ironmasters' Association in 1811–3.

The social organization of the work process also argued against a comprehensive study undertaken by an individual and resulting in useful data. Charcoal-burning was a huge, decentralized system of production: peasants and crofters labored under the tenant's obligation to deliver charcoal to the ironworks. Production was the responsibility of individuals, tens of thousands of peasants and crofters, each working independently, deep in the forests.^[35] Work in the forest was linked with the changing of the seasons and the tilling of the soil. In autumn, after the harvest, wood was burned to charcoal in the forest

where it had been felled; in winter, when the snow made the pathless forest passable, sledges carried charcoal to the ironworks. The methods of charcoal burning were the product of local conditions and tradition; the variety of circumstances was reflected in the names of different sorts of charcoal piles. To hammer this decentralized variety of methods and measures into a standard form amenable to quantitative comparisons seemed impossible. Hence Wallner contented himself in 1746 with publishing the views he had elicited from charcoal-burners he had met and a few quotations from foreign literature, together with his own individual, casual observations.

That charcoal-burning—like most technologies at the time—was a nonverbal technology complicated the problem of quantification.^[36] Witness the failure of the encyclopedists in their attempt to gather and record technical know-how of existing practices like charcoal-burning. John R. Harris has argued that "the difficulties of imparting craft skills by literary and graphic means" limited the "technological gain" possible from the 18th-century encyclopedias.^[37] The difficulties of improving a nonverbal technology were akin to the problems of describing it. A process could not easily be formulated as an intellectual problem and reduced to quantitative terms outside the realm of practical experience.

Science first met technology at the edge of the woods, when the sledges loaded with charcoal approached the ironworks to deliver their products. Only here could the individual ironmaster exercise some control over production by making sure that the peasants and crofters delivered the agreed amount of charcoal. Here the ironmaster might enlist the aid of geometry. Two articles in the Proceedings of the Royal Swedish Academy of Sciences addressed his concerns.

They were devoted to the mathematical problem of calculating the volume of the rhombically shaped sledges that carried the charcoal (fig. 10.2).^[38] Together, however, the ironmasters could marshal resources enough to launch a concerted, large-scale attack on the technical problem of charcoal production.

Utilization of Manpower

It is only too easy for historians to overlook the importance of muscle power during the 18th century. Horse whims and tread-wheels, despite their widespread use and importance in their day, have been overshadowed by more efficient and spectacular technologies like waterwheels, windmills, and steam engines. But an innovation does not immediately cause the abandonment of earlier, inferior technologies as obsolete. On the contrary, the total use of old technologies declines slowly and asymptotically. Treadwheels, for example, were used in Swedish mines as late as the 1880s, and one is even reported to have been in use as late as 1896.^[39] Muscle power was the dominant power technology during the 18th century: the total work produced by men, horses, and oxen in fields, roads, forests, mines, mills, and harbors probably exceeded the combined power of all steam engines, waterwheels, and windmills.^[40]

Industry needed sources of mechanical power capable of high output and continuous operation. The muscle power of men and animals

could provide continuous operation, but output was relatively low. A man working a ten-hour day produced approximately 0.1 horsepower, a horse in a good harness roughly six times as much. Horse whims used at the mines, driven by two or four horses, could thus develop approximately 1.2 or 2.4 horsepower.^[41] Continuous operation in three shifts required as many as a dozen horses. Perhaps eighty or even forty men could have produced an equivalent amount of work, but animals offered certain advantages when continuous power of this magnitude was required. In other cases, however, the power supplied by a few persons was not only sufficient but preferable—turning a crank or a windlass, pulling at ropes, carrying burdens, tramping in treadwheels. Compared to horses and oxen, men were relatively small and movable; their power output could be regulated with a word or a glance. The factor of control was significant when it came to loading or unloading ships, turning lathes, grinding and polishing, operating textile machines, or building.

Early attempts to determine how much physical labor a man could be expected to do in a day were made around 1700 at the Académie royale des sciences.^[42] Inquiries continued through the 18th century. Bernard Forest de Bélidor, Charles Augustin Coulomb, John T. Desaguliers, Johann Euler, and Philippe de La Hire were among those who addressed related questions of a fair day's work and the comparative strength of men and horses.^[43] No consensus was reached, but a large body of data was generated.^[44]

In these 18th-century studies, Eugene S. Ferguson has written, "the most casual and fragmentary data were being worked up, with the help of algebraic operations, into definite and precise conclusions."^[45] A case in point, described in more detail by Ferguson in an earlier paper, is a study published by Coulomb in 1798. On the basis of two single observations of physical labor, Coulomb wrote an equation for the useful work done while carrying one load of firewood upstairs. This equation was then differentiated and set equal to zero. Coulomb claimed to have obtained thereby the optimum load that would lead to the maximum day's work.^[46]

In an article published in 1744, the Swedish mathematician Pehr Elvius compared the efficiency of four treadmills then operating in Stockholm: one at the new Royal Palace under construction, one at a glass factory, and two at the construction of Stockholm's lock. For each treadmill, Elvius measured the rate at which a weight was raised and calculated that "the output of every single fellow is so great that 4 2/5 lispounds was hoisted at a rate of one foot every second."^[47] After discussing the differences in design and output for the four treadmills, he attempted to draw general conclusions concerning the ideal design for various applications.

Though Elvius recognized that quantitative methods could be used to improve an existing technology, his method suffered from two interdependent weaknesses, one mathematical and the other social. He used a single observation for each type of treadwheel ("a time of 4 minutes exactly or 240 seconds"),^[48] not an average value. Furthermore, his study was based on the types of treadwheels that could be seen in action in the course of a leisurely stroll through the

capital. But the large majority of treadwheels were at work in the mines of the countryside.^[49]

The opportunity for a systematic study of manpower did not arrive in Sweden until the early 1770s.^[50] The place was the naval dockyard of Karlskrona in the southeast, where manual labor was used to operate the pumps. To dry-dock a man-of-war, ninety sailors worked in three shifts of ten to thirteen hours depending on the displacement of the ship. The engineer Johan Eric Norberg studied the efficiency of manual labor during dry-docking on ten different occasions during 1772 and 1773, and published his report in the Proceedings of the Royal Swedish Academy of Sciences. His primary aim was to increase the efficiency of the pumping operation, but he also had in mind to determine in general the output that could be expected from human muscle power. This, Norberg wrote, had not been attempted before, save in a very small number of tests of limited scope. The results had fallen short in both reliability and extent.^[51] Norberg thus demonstrated his awareness that many tests under varying conditions were required to obtain results of general validity. Norberg took hundreds of measurements of the work of the ninety men. Figure 10.3, which shows one of his tables, illustrates the systematic nature of his approach.

Norberg's study is the earliest example in Swedish technology of a systematic, full-scale investigation intended to yield a result of general

applicability. It is scarcely surprising that Norberg chose the navy as the locus for his study. The military was the only social institution sufficient in size and authority to assemble and command the large work force needed for so ambitious a study of the efficiency of manual labor.^[52]

In his study in the 1740s Elvius had discussed the performance of only a few individuals. Even the physical characteristics of the individual—in particular, the length of his stride—was an important parameter in Elvius' calculation (dust jacket illustration). But in Norberg's study of the dockyard, individuals are reduced to figures in a table, and the combined product of ninety men's work expressed as a single quantity. Norberg was certainly aware of influential external factors: "harmony, the ration of spirits, the gentle persuasion of the officers, fair or foul weather, etc., have the most important influence on the work, and make a great difference as to output, which should otherwise be nearly the same for equal weights of water." These more human aspects of work at the pumps were not reproduced in Norberg's table, however, which expressed output as a product of measurable physical quantities. Norberg distinguished "the outputs of three kinds of people in various corps." In the sixth column of the table the letter V denotes volunteers (Volontairer ), M , marines (Marinierer ) and B , tenement seamen (Rote-Båtsmän ). The attempt to obtain average rather than individual values necessitated suppression of individual characteristics in favor of general traits.^[53]

The Need for Control

In their aspiration to apply quantitative methods to technology in the hope of improving its efficiency and economy, the engineers of the 18th century found themselves hampered by their own boun-

daries as individuals. They had the ambition to reproduce the conditions of the laboratory in the field: to carry out systematic experiments under controlled and reproducible conditions. But this demanded a much larger spatial and temporal dimension than the laboratory—with its desktop apparatus and time scale of days or weeks. It was beyond the means of individuals to embrace the technical reality in time and space. In terms of space , their ambition required wooden structures weighing perhaps up to a ton (whether they were treadmills, charcoal piles, or waterwheels). It made even large spatial demands on the immediate surroundings, since it required such assets as large areas of forest, hundreds of men, or a substantial water supply. In terms of time , it required that several people work together on the experiments for several years. In short, considerable resources were needed to build these unwieldy pieces of experimental apparatus and to operate them for a period of years, and these resources were beyond the means of any individual.

This argument can be expressed in another way. To say that the spatial and temporal dimensions of technology demanded considerable resources if the conditions of the laboratory were to be reproduced in the field is the same as to say that one must be able to control the necessary physical and social realm in terms of time and space . This chapter began with the claim that efficiency and economy are the two quantities that characterize modern technology, and it will end by asserting that its qualitative characteristic is control .^[54] Technology is an activity that aims at changes of the material world, and this always involves control.

It is therefore not by chance that the first successful attempt to apply quantitative methods to technology discoverable in Swedish history of technology was achieved at the naval dockyard in Karlskrona in the early 1770s. The case of the ninety soldiers, working in shifts at the pumps to dry-dock a frigate while Johan Eric Norberg stood over them with his watch and notebook, symbolizes a turning-point. It marked the beginning of the successful application of quantitative

methods in technology during the late Enlightenment. At that time, only the military possessed the necessary control over the physical and social realm to reproduce laboratory conditions in the field. It would be some time before civilian institutions grew strong enough to be able to exercise the same degree of control, for example, in the federally supported investigation of the Franklin Institute in the 1830s into the causes of steamboat boiler explosions.^[55] These institutions were then strong enough in both authority and competence .

Although this study has been limited to a specific aspect of the question of the relationship between science and technology during the 18th century—the application of quantitative methods to technology—it may have some relevance for the more general question.^[56] Roger Hahn has argued that one way to grasp the relationship between science and technology during the 18th century would be to examine the institutional development of technology, and in particular the many societies of arts that flourished in Europe on the model of the scientific academies. Science made an impact, Hahn wrote, "not only by lending its ideology, personnel and theories to technology, but also by offering its social organization as a model to be copied."^[57] By copying the social organization of the scientific enterprise, these institutions "also accepted the presuppositions of science itself: rationality, objectivity and publicity."^[58] Studies of the history of

individual societies, such as Robert E. Schofield's The Lunar Society of Birmingham ,^[59] have demonstrated and elucidated this institutional development, but the question of their effectiveness in achieving technological change continues to puzzle historians. A. Rupert Hall has, for example, remarked that it requires "a degree of faith" to find a causal relationship between the popularity of science on the one hand and innovation in technology on the other.^[60] But if the benefit of a union between science and technology was conceived only on a cultural level during the 18th century, why was it not put into effect on a political and economic level?^[61]

This study suggests that we should continue our search for institutional developments, but look for projects that were undertaken by institutions rather than individual efforts. The characteristics to look for in these technological institutions are not only the rationality, objectivity, and publicity of their scientific counterparts but also the power to control the necessary physical and social realm in order to reproduce laboratory conditions in the field.

More specifically, we should look for institutions that exercised authority according to rank in hierarchies based on scientific and technical competence . The quest will lead us to that historical borderland that lies between the inquisitive academies of the 18th century and the efficient industries of the 19th century, to the period between 1790 and 1825, which has been recognized by Cardwell as one of the definite periods of decisive change in the course of technological history.^[62] Not only were many of the major technologies of

the 19th and 20th centuries founded then, but "at the same time social changes took place in the organisation of technology and science setting them on the courses that led to modern technological society."^[63]

One of the major social changes was the emergence of institutions that could exercise the necessary control for the successful application of quantitative methods to technology; first within the established structure of the military (military academies, arsenals and dockyards) and later in large industries of national-military importance (mining and chemical industry), major civil-engineering projects (canals), and new civilian institutions. These institutions are probably identical with those that Peter Mathias has called "focal points for developing new skills and educational programs" and that were sponsored by the demands of the state for deploying scientific technology for military or official purposes.^[64]