Ken Alder

This paper documents the connection between the technological and political transformations of late 18th-century France. Its subject is the efforts of state military engineers to produce functionally identical artifacts (interchangeable parts manufacturing). These efforts faced resistance from artisans and merchants attached to the corporate-absolutist ancien régime, for whom artifacts were idiosyncratic, and ‘thick’ with multiple meanings. I argue that to oblige artisans to produce standardized artifacts, the military engineers defined these artifacts with instruments such as technical drawing and the tools of manufacturing tolerance, which the engineers then refined in increasingly rule-bound ways to forestall further subversion by artisans. Hence, I offer a historical account of how the ‘objectivity’ of these artifacts was the outcome of social conflict and negotiation over the terms of an exchange. In particular, I explain why engineers eventually turned to projective drawings (including the descriptive geometry) over alternative ways of representing artifacts (such as free-hand, academic, and perspectival drawings). And I document the origins of manufacturing tolerance, in which the dimensions of an artifact were circumscribed with gauges and machine-tools to preclude possible sources of disagreement. The paper closes with its own ‘thick’ narrative of how standards of production emerged out of social conflict in a particular community on the eve of the French Revolution - a process which reflected the emerging political ‘toleration’ of the French state for its citizen-producers. The SCOT programme can be used to provide a political account of how the operation of seemingly ‘objective’ artifacts can be coordinated across vast physical, temporal and cultural boundaries.

Introduction

We live today in a world of mechanical clones: identical artifacts composed of identical parts. When a piece breaks in our bicycle, our automobile, or our computer, we don’t throw the whole machine out; we replace the broken piece with a piece which is functionally identical. What makes possible this world of identical artifacts? A world in which 10,000 bicycle gears cut in Japan can be shipped halfway around the world to Mexico and fastened successfully to 10,000 hubs? How did such a world of uniformity come into being? And what does its emergence suggest about the way we should conceptualize technological change?

The usual response to these questions about the origin of interchangeable parts is to point to the advent of Fordist mass production in the

499 Index

early 20th century, a period associated with the consolidation of corporate capitalism and the Second Industrial Revolution. Fordism is a form of production predicated on a logic of achieving low unit costs by eliminating the need for skilled labour in the shaping and fitting of pieces. But historians have shown that it was not industrial capitalists, but state military engineers, who first conceived of the ideal of uniform production - and who partially realized it - one hundred years before Henry Ford, back in the late 18th century. This was a period associated with the First Industrial Revolution, and also with the political revolutions in France and America. [1] I will suggest that this earlier timing is no accident. I examine the origins of the ideal and practice of ‘making things the same’, to demonstrate the intimate relationship between the political and material revolutions of the late 18th century.

Understanding how artifacts were made identical, however, will mean paying attention not only to new 18th-century ways of making things, but also to new 18th-century ways of representing them. In particular, the making of identical parts required new forms of technological representation capable of coordinating the efforts of diverse people with divergent interests. Long before the advent of the computer, material artifacts were being produced in conjunction with techniques and representations (‘information technologies’) that were themselves subject to a process of standardization. As we will see, these forms of technological representation - mechanical drawing and manufacturing tolerance - had the property of rendering artifacts with a new degree of ‘objectivity’; but that is not to say that these representations were politically neutral. On the contrary, the form taken by the new representations was part of the new enlightened political order inaugurated in the 18th century. In our own day, computer-aided manufacturing is radically altering the representations and practices which govern late 20th-century production. The designs of engineers are now being realized with hitherto unsurpassed exactitude. Yet as Shoshana Zuboff and others have noted, the process by which these idealized designs are realized is transforming power relations in the workplace, breaking down traditional hierarchies in some places, reinforcing them in others. [2]

For similar reasons, the story of the origins of ‘making things the same’ poses a challenge and an opportunity for the programme in the social construction of technology (SCOT). SCOT has been the ascendant approach to the history of technology for the last 15 years - and for good reason. SCOT has taught students of technology several essential lessons: to pay close attention to the internal workings of artifacts; to value empirical historical analysis; to study the divergent meanings that different groups ascribe to the ‘same’ technology (‘flexible interpretation’); and finally to ascribe the triumph or failure of any particular technology to the clout of its sponsors, rather than the inherent properties of the technology itself (the principle of ‘symmetry’). [3] If anything, these lessons have been insufficiently recognized outside the discipline of technology studies. Many cultural critics still try to address the ‘social life of things’ solely in terms of

500

production and markets, without taking into account the role of technological design and designers. [4]

Yet the SCOT programme, as widely practised, has several limitations worth addressing. One complaint is that SCOT has generally ignored the problem of production. [5] Another concern is that those versions of SCOT which can be reduced to ‘interest theory’ have sometimes collapsed into a form of local social determinism, and have thereby failed to grapple effectively with some important issues in the relationship between technology and society. In particular, these localized studies do not account fully for the ways in which artifacts seem to possess a kind of innate potency, on the one hand, and how they carry social and political values across temporal, geographic and cultural boundaries, on the other. This is not a trivial concern. Technologies travel across boundaries, sometimes with devastating results. And over the course of the past two centuries, bureaucracies have emerged capable of coordinating the operation of these technologies in diverse environments. Understanding the process by which artifacts come to transcend the local conditions in which they are conceived and produced should be one of the central tasks facing any satisfactory approach to technology. In particular, to ignore the potency of ‘travelling technologies’ in the case of modern weaponry would be morally unconscionable. [6] Historians need a genuinely historicist way to conceptualize the process by which artifacts are shaped by local interests, and yet are also made capable of being coordinated across vast distances. Doing so will not prove that these sorts of artifacts cease to bear political values; on the contrary, it will show that they bear the political value called ‘objectivity’ which is characteristic of modern technological systems.

In this paper I seek to develop such a methodology and frame it within a general historical problem. The historical question I will address is the perplexing relationship between the two profound political and economic revolutions which transformed much of Western Europe in the late 18th century. The political transformation led from absolutism to popular sovereignty, and achieved its moment of highest visibility during the French Revolution. The economic transformation led from the guild system of production to entrepreneurial capitalism, and has generally been studied under the rubric of the Industrial Revolution. Of course, neither transformation was fully accomplished within the compass of the late 18th or early 19th century, nor was the pattern of change the same in all European countries, nor even in all regions of those countries. Indeed, 20 years of historical scholarship have emphasized the unevenness and diversity of both of these political and economic revolutions. Still, their conjunction in the later 18th century has been widely understood as marking the boundary between the early modern and the modern period, even if the nature of this conjunction has long been a matter of controversy, especially for those historians who concentrate on France. [7]

In that country, the political transformation led from an ancien régime polity (in which an absolutist sovereign legitimated all roles and recognized no realm of private action) to the emergent system of modern politics,

501 Index

proclaimed in the early days of the French Revolution, in which sovereignty flowed from the people, and which assumed a clear separation between public and private spheres. In the corporatist régime of pre-Revolutionary France, the king accorded distinct legal status to different sorts of subjects (nobles, commoners, city-dwellers, peasants and so on) on the juridical assumption that these groups had agreed to alienate their natural liberties to the sovereign in return for a set of privileges and obligations that were particular to them. On these grounds, the king denied political status to members of religious minorities, and justified the different kinds of justice rendered to different sorts of persons. In practice, this legal particularism had been eroded by the monarchy’s bureaucratic interest in centralizing authority over the military, taxation and justice. But local interests still prevailed in many instances, and the king still governed by personal authority. [8]

Against this ancien régime of dynastic interest and private law, we may set the modern polity based on national citizenship and public law. Crucial to the vision of the ‘enlightened’ nation-state which energized French reformers in this period was the ideal of toleration. This ideal was supposed to govern the relationship among citizens, and between citizens and the state, by carving out a realm of private conscience and public speech, and by punishing (in theory) only those actions which brought harm to others or to the public good. The demand for toleration - particularly for religious toleration - was one of the principal battle cries - perhaps the principal battle cry - of the Enlightenment. One need only think here of the assertions of John Locke and Pierre Bayle at the end of the 17th century, or the declarations of Immanuel Kant or Voltaire in the middle of the 18th century. To be sure, the seeds of political toleration, sown in the ancien régime, were only fitfully realized in the course of the 19th and 20th centuries. But in theory, at least, the boundary between the private and public spheres was henceforth to be defined by a forever-elaborated set of public laws. It is important to emphasize, however, that these Enlightened reformers did not believe that the ideal of toleration meant that the state should absent itself from public life, nor that the populace should directly mete out justice. On the contrary, what Voltaire feared was both the tyranny of the despotic state (which operated according to a system of private and secret justice) and intolerance of the mob (which acted without reason). In this enlightened vision of toleration, the state was expected to play a crucial role as the guarantor and regulator of the public order. [9]

In this paper, I argue that this hope for political transformation was crucial to the concurrent transformation in the representation and making of identical goods.

Index

From ‘Thick Things’ to ‘Objective Objects’

The methodology I will use to develop this argument will consider artifacts as the outcome of a history of exchanges in which parties with distinct interests negotiate their differences. The technology which results from this

502

process, I will argue, is both the bearer of political values and can in some sense be called ‘objective’. In recent years, a group of scholars have made various attempts to define more carefully what they mean by the ‘objectivity’ of techno-scientific results. They have distinguished carefully between the claim that objectivity means the ‘truth’ about nature or some matter of public concern, and the more limited claim that objectivity denotes something akin to ‘impersonality’ or ‘disinterestedness’. In what follows, I take my cue from this literature, applying to artifacts the same sort of analysis with which Theodore Porter has tackled the problem of quantification. [10]

Porter argues that the reduction of a natural phenomenon or some facet of public life to a numerical result does not simply reflect the underlying truth about the subject (though it may do that in part), but also represents the outcome of a process of conflict between mistrustful parties. Experts who resort to numbers generally do so because they find the stability of numbers a valuable tool for managing complex and far-flung operations. But it is only under pressure from powerful outside forces that they agree to make their numbers public. After all, experts understand that the full and public articulation of their rules of calculation restricts their ability to make flexible judgements in the face of changing circumstances. This public articulation, moreover, reduces their private discretion about these matters, and hence, their personal power. What Porter and others call ‘mechanical objectivity’ is the kind of description of nature (or society) which experts provide when they wish to present their conclusions as having been derived with a minimum of human intervention. At the limit, these results are conveyed as if by machine, and mask a different sort of power which operates under the guise of impersonality. T his form of objectivity is part and parcel of the contractual relations endemic to modern, mistrustful polities.

Over the past 200 years, many of the artifacts of commerce and industry have come to acquire a similar degree of impersonality. This was not a trivial achievement. The material world is lumpy, recalcitrant and inconsistent. Connections come apart; parts wear out; things break. Those people who work with material objects - let us call them ‘technologists’ - find it challenging enough to manipulate physical matter so as to build a single artifact which works in the prescribed manner in the workshop, let alone consistently repeat this set of manipulations several thousand times over and still ensure that these artifacts function effectively in a diverse set of environments. In short, things are ‘thick’.

By the phrase ‘thick things’, I mean to invoke two aspects of material artifacts. First, the difficulty of consistently shaping the material world into a working artifact, or what one early modern technologist called the ‘resistance and obstinacy of matter’. [11] And second, the related challenge of assimilating ordinary artifacts to any idealized representation in such a way that their qualities can be captured in their entirety. Here I borrow the term ‘thick’ from Clifford Geertz, who urged anthropologists to provide ‘thick descriptions’ if they wished to capture the diverse layers of meaning with which different human agents imbued their actions and those of their

503 Index

fellows. Geertz contrasted the capacity of thick ethnography to represent multiple (and often divergent) human points of view with the reductive ‘thin’ descriptions in which scientistic anthropologists collapsed actions into a simplified matrix of behaviour or function. [12]

For my purposes, the thickness of both artifacts and their representations can be contrasted with the ‘thinning’ process by which scientific objects are often made amenable to analysis. Here, Gaston Bachelard provides a valuable hint. He notes that the synthesizing power of explanation in the physical sciences depends on a vast array of precision scientific instruments which investigators wield in order to create objects that are mathematically tractable, and can therefore constitute legitimate objects of inquiry. In the extreme case of 20th-century physics, these objects (electrons, for instance) become more than similar: they become ontologically identical; and this in some sense accounts for the fact that their properties can be described with unsurpassed precision and economy. [13]

The ordinary material artifacts of everyday commerce are not, of course, readily amenable to this exacting form of representation, nor this extreme degree of regimentation. But, as we will see, some technologists have been driven to assimilate artifacts to this sort of analysis, and - not coincidentally - to embed them in technological systems. Making things the same, and ensuring their success in diverse environments, requires the coordination of many diverse people - whether by cooperation or by coercion. And common forms of representing artifacts proved essential to this endeavour. [14] The manner in which these representations were achieved, however, did not involve a one-sided imposition of standards by some technologists upon others, but emerged as part of a wider process of social struggle and negotiation. Indeed, I will argue in this paper that it is the pressure of social conflict which has, over time, obliged technologists to define explicit rules for their representation of artifacts. In particular, to guarantee that these artifacts could be defined with ‘mechanical objectivity’, these technologists have been obliged to embed these rules in general ‘instruments’ capable of defining, comparing and judging all manner of artifacts. Two instruments - mechanical drawing and the tools of manufacturing tolerance - were developed by engineer-technologists during the Enlightenment, and were further refined by them in response to outside pressures. In the hands of these engineers, mechanical drawing went from being a pictorial representation of the artifact, to a rigorous (‘thin’) definition of its physical form. The tools of manufacturing tolerance included gauges, jigs, fixtures and even automatic machinery, all deployed by engineers to define and shape artifacts in new and more precise ways. The invention and construction of these tools was, of course, the work of individual technologists - but the way that these tools were actually configured in the workplace was inevitably a matter of wider social negotiation. When coupled with the new scales of measurement introduced in this period (such as the metric system), these instruments have been essential in enabling technologies to travel across physical and cultural boundaries. In this sense, they are akin to those semiotic devices that Bruno Latour has

504

called ‘immutable mobiles’. [15] As we will see, however, such mobiles are themselves the outcome of a social struggle over how to conceive of and enforce standards of production.

Conceptualizing technology in this way has several advantages. First, rather than view technology (including the means of its production) as simply an external resource which generates social conflict, it understands technology (including the means of its production) as the outcome of ongoing social conflict and negotiation, as well as a source of further conflict. Second, this approach thereby folds the making of technology (including the means of its production) back into the historical process without prejudging the relative strength of the parties to these conflicts and negotiations. Third, it thereby allows human agents and contingent factors to set the pace and direction of technological change - even as it points to a shift in the terrain upon which such conflicts and negotiations took place in the 18th century. And fourth, it draws our attention toward the factors which made possible the rise of modern technological systems out of the demise of the corporate order of the ancien régime, and the crucial importance of information technologies in that transition.

In the remainder of this paper, I will proceed as follows. First, I describe the structure of the corporate order - and the agenda of its opponents among the philosophes and state engineers. Second, I lay out the logic behind the two instruments - mechanical drawing and manufacturing tolerance - which these engineers developed in order to tame artifacts and their makers. Third, I provide my own thick description: a detailed case example of how identical artifacts and the instruments which made them possible emerged as the negotiated response to social conflict among parties with diverse understandings of artifacts - and can thus be understood as the outcome of a historical (rather than a logical) process. And fourth, I conclude with some general remarks on the relationship between the modern French state and capitalism, and the political and technological revolutions of the late 18th century more generally.

Index

Replacing the Corporate Order

Social and economic historians have long wondered how and why production in Western Europe shifted from the artisanal workshop to the entrepreneurial factory. The approach of economic historians, such as David Landes or Joel Mokyr, is to couple the rise of factory organization with technological creativity motivated by the heady lure of profits. [16] In complementary fashion, business historians such as Alfred Chandler have emphasized the essential role of the entrepreneur-manager as the organizer of production. [17] And advocates of the ‘proto-industrialization’ thesis have suggested how capitalists first gathered outworkers from rural areas under a single roof in a transitional Age of Manufactures. [18] Each of these schemes (there are others, of course) has illuminated different aspects of this great transition. Yet all spin some kind of teleological narrative. As recent commentators have noted - William Reddy, Tessie Liu and Maxine Berg

505 Index

among them - each assumes the success of the phenomenon it seeks to explain: the rise of machine production, the emergence of the entrepreneurial role, or the triumph of capitalists over domestic producers. [19] Up to a point, this form of teleology is salutary because it focuses the historical attention. However, as Charles Sabel and Jonathan Zeitlin point out, teleological histories of industrialization have obscured important aspects of that process, such as the continued vitality of small-scale flexible production well into the supposed heyday of mass production. [20] A genuinely historical point of departure, then, is to ask how 18th-century elites tried to manage the transition away from artisanal production, and how ‘rational production’ emerged from the resistance these schemes encountered.

The artisanal guilds which controlled craft production in the ancien régime participated in the corporate order whose legitimacy rested on the theory of absolutism. That is, the members of each of the various mercantile and productive associations had collectively surrendered (alienated) their natural liberties to the sovereign in return for the privilege of organizing their own affairs and exercising a legal monopoly over a particular portion of trade. As William Sewell has noted, these collectivities validated this monopoly around a notion of ‘art’, a set of tacit and unspecifiable skills which could only be acquired through a long apprenticeship in the trade, and which governed the norms of their social life. [21] And as Michael Sonenscher has pointed out, these artisans considered themselves to have a natural property right in their own labour power - and this included not only those master artisans who sold goods in the marketplace, but even those artisans and journeymen who worked in large workshops and under an extensive division of labour. [22] For these artisans, the price of their alienation of this labour right was the wage, whether it was paid for a day’s work or for the making of a particular article (the prix de façon). This legal fiction of the ownership-wage is what distinguished the artisan from the slave and dependent servant, and it had real implications for the ability of workers to make claims about the proper division of labour in the workshop, the amount of time they spent on set-up work, and their customary rights to the by-products of their labour. Craftwork, then, was not simply a mode of hand-made production (artisans can use machines too), but a social, cultural, and legal system which validated collective privileges and individual property in skill. [23]

This was part of a larger pattern of legal entitlements which governed not only the production of artifacts in the ancien régime, but also their sale, purchase and use. Not only did guilds superintend the distribution and retailing of most consumer goods, but their consumption, too, might be limited to particular classes of persons, either by formal sumptuary laws or by customary codes. Even the measurement of goods was particular, in that individual guilds used their own units of measures, and these generally differed from one local jurisdiction to the next. Under the theory of absolutism, therefore, to forge a musket barrel, to concoct a new sort of soup, to sell a bolt of linen or even to wear a certain sort of hat, was in

506

some sense a legal privilege. In such a scheme, every artifact was not simply individually ‘custom-made’, but was understood to be idiosyncratic, personal, and particular. [24]

However, a growing number of 18th-century elites - many of them associated with the Physiocratic movement of the French Enlightenment - were convinced that the corporate system of production was deficient. As a practical matter, the monopolies of the various guilds had been eroded by the expansion of rural manufactures not covered by the statutes. But only during the Enlightenment did the corporate system come under explicit political attack. In the last decades of the ancien régime, the Physiocrats and their allies began to argue that the guilds, by zealously guarding technical knowledge in private hands, had restricted innovation, artificially raised prices and involved the state in endless litigation. When one of their allies, Turgot, became chief minister in 1775, he banned the corporations. Although the guilds were revived shortly thereafter, the Revolution abolished them permanently in 1791. It is worth noting, however, that although Turgot was an advocate of ‘laissez-faire’, he expected that the state would continue to play an active role in guaranteeing standards of production and in regulating trade. In other words, these French reformers did not advocate the market principle of unregulated private exchange, but the ideal of the market-place where transactions between parties could be guaranteed by the state. [25]

The question for these elites was: what was to replace the guilds? For all their hostility to the corporate system, these savants recognized that the corporations formed a coherent world which organized the social life of artisanal producers, as well as daily practices in the workplace. In the absence of the corporations, who would decide how to set up work schedules, and how? What would the rates of compensation be? The answers to such questions had important implications for the distribution of wealth and knowledge in society. Yet these theoreticians of the workplace did not necessarily anticipate the outcome that leaps to our lips today: ‘the machine’, ‘the entrepreneur’, ‘the market’. What they called for was the creation of a new kind of public technical knowledge.

This programme for a public technological knowledge was most fully developed in Diderot’s famous article, ‘Art’. There, the cutler’s son made a plea for the mutual aid that the savant and craftsworker should offer one another. Theoretical training was counterproductive unless combined with a practical knowledge of basic physical properties. In the same breath, however, Diderot showed his appreciation of the organizing power of theoretical science by calling for a ‘Logician’ to invent a ‘grammar of the arts’. He deplored the secrecy and venality of the various guilds, which he felt stifled technical innovation. One sign of this secrecy was the chaotic terminology of the trades. The first task of Diderot’s Logician, therefore, would be to devise a quantitative scale to express the various measures of tools (their size, force of action, et cetera) and to initiate a morphological analysis of their shape by means of technical drawing - or what he called ‘the geometry of the workshop’. Where once the tacit and personal ‘art’ of

the guilds had organized production (thereby stifling the free exchange of both goods and knowledge), henceforth an open and public ‘science’ - conducted by means of rigorous analysis - would generate innovative technical knowledge. The Encyclopédie was itself to be the first instalment of this programme. [26]

Diderot’s praise for the ideal of open science, and his denunciation of proprietary rights to technological knowledge, was part of the philosophe’s larger critique of the ancien régime’s world of private justice, personal offices, and privileged status. [27] What was new in the 18th century was the concurrent effort of the French state deliberately to close this gap between science and technology. The French engineers were trained to just this end.

Index

Enlightenment Engineering and the World of Production

The military engineers of the 18th century mediated between the French state and the world of commerce. Trained by the state in the first formal techno-scientific schools in Europe, they were enjoined to partake of neither the routine and secret practices of the artisanal corporations, nor the abstract and purposeless speculations of the savant. Instead, these engineers were to combine theory and practice in a programme of institutionalized innovation. Their school curriculum focused on mechanical drawing, rational mechanics and the practical details of their trade. This cognitive programme was meant to carry particular social lessons: engineers were not to be venal and collusive like the artisans, nor aloof and asocial like the savant. Instead, they were to vie in meritocratic competition (an identity consonant with their dignity as notables), even as they acquired an ethos of hierarchy and subordination. They were to be both technically competent and loyal servants of the state. In short, they were to be professionals. [28]

At the beginning of the 18th century, the military engineers of the artillery service became the sole intermediary through which the army acquired all its weaponry: cannon, artillery carriages, munitions and small arms (muskets, pistols and sabres). No longer would colonels supply their own troops with weapons. This was part of the absolutist state’s effort to make the army answerable to a central command. Yet France, like other states of early modern Europe, did not thereby assume ownership of the means of military production. The military market may have been large and undifferentiated, but it was erratic. Consequently, the state allowed merchants and artisanal producers to absorb the risks associated with these investments, while cloaking these producers in legal privileges and assuring them lucrative (if intermittent) profits. And to make sure that these provincial producers and traders delivered the agreed-upon goods at the agreed-upon price and with some assurance of quality, the state sent artillery-engineers known as ‘inspectors’ into the provincial armouries. [29]

These artillery-inspectors were enjoined to see that the army’s guns were made more precise and uniform, to make their operation more

508

reliable, accurate and deadly. Precision and uniformity are here to be understood as mirror-image twins. Precision, as measured against a background uniformity, ensured that a single weapon behaved the same over time. And uniformity, as measured with precision, ensured that numerous weapons behaved similarly to one another. From the point of view of the army, both attributes promised to make the infantry drill more effective. From the point of view of artillery service, both attributes also allowed them to police their monopoly over this prestigious piece of the ancien régime’s military-industrial complex. In particular, by setting rigorous standards for production, the engineers ensured that interloping colonels and merchants would be unable to strike private deals for weapons, and that all weapons would have to be procured through them.

But how were these rigorous standards to be enforced? In the first half of the 18th century, the artillery-engineers had supervised the armouries through the same mechanisms of privilege which the monarch used to regulate the trade corporations. Only certain designated artisan-armourers could produce guns for the king and, as a mark of their privilege, they received tax breaks and other local legal advantages (exemption from militia service, the obligation to house soldiers and submit to the corvée, and so on). In return, these artisans were obliged to sell their wares at the stipulated price exclusively to certain merchants (known as ‘Entrepreneurs’), who were legally designated as the sole buyers of arms for the king, and who also enjoyed an array of fiscal privileges. In theory, these provisions were backed up by the threat of martial punishment, and the armourers were nominally subject to military law. But armourers and merchants were not always eager to comply with the quality and cost requirements set by the artillerists, and they disagreed among themselves about how to divide the tasks and profits of gun-making. Forced to sell at fixed prices, they cut corners on quality or attempted to leave the king’s service. Already in the 18th century, some 20 different subspecialists contributed to the making of a gun, and each of these artisans considered himself to possess a right in the product of his labour. Moreover, all these artisans and merchants had a real opportunity to make good on this claim by shifting their skills and capital to the private market for guns which existed right alongside the royal armoury.

So, in the middle of the 18th century - under the reform-minded leadership of First Inspector-General Jean-Baptiste de Gribeauval - the artillery inspectors adopted a new managerial role vis-à-vis the armourers. They sidelined the Entrepreneur’s role as the coordinator of production, and began to set the price for individual gun parts themselves, rather than just for the final finished product. But this meant that the engineers had to define detailed standards for each individual gun part, rather than simply asking for assembled, functional guns. But how were the engineers to enforce these new standards, to superintend the fractious provincial manufactures? One of their solutions was to adapt new kinds of technical drawings.

509 Index

Drawing Things the Same

In recent years, a number of scholars have turned their attention to the representation of techno-scientific objects. Many of these studies have sought to uncover the ways in which representations have underpinned the ‘objectivity’ of scientific results. Lorraine Daston and Peter Galison have studied illustrations in scientific atlases, noting that they signal an effort to suppress individual and group idiosyncrasies, and thereby (supposedly) obviate any need for interpretive judgement. Their approach highlights the moral act of abnegation and self-discipline which these practitioners sought to associate with scientific investigation. [30] Michael Lynch has noted how scientists use certain kinds of representations to perform a ‘disciplining of the object’: a process by which the graphical properties of the object are made to embody the ‘natural object’, making the object scientifically knowable and manipulable, much like the docile bodies of Foucault’s prison institutions. [31] This approach implicitly reminds us of the difficulty of ever fully capturing in two dimensions the variety and intractability of ‘thick’ things. More generally still, Bruno Latour has referred to these ‘rationalist’ forms of representation as ‘immutable mobiles’. Latour argues that images in this guise can be transported across physical and cultural distances without undue distortion, and collected at a remote site of power. There, at these ‘centres of calculation’, these images can be analyzed and synoptically compared with other images, so that discrepancies may be noted and corrective actions taken. To the extent that a cathedral plan coordinates stone-cutters and a military map deploys soldiers, an engineering drawing commands workers. Of course, pictures do not in themselves coordinate, deploy or command. These drawings make possible the exercise of power by enabling their possessors to master phenomena on a scale inaccessible to others. [32]

Each of these scholars identifies crucial aspects of scientific representations. However, each slights several important features of the new forms which these representations took in the 18th century, at least as they were deployed in the workplace and in the management of practical affairs. These authors do not pay sufficient attention to the alternative ways of representing objects that were available to contemporaries. Eighteenth-century engineers, for instance, came to prefer projective representations, whereas natural philosophers used perspectival views, and artisans were taught freehand drawing. This omission is serious because these authors do not show how these different forms of representation emerged within the context of different social milieus, and hence implied very different sorts of social relations between image-makers and object-makers. The differences in these sorts of social relations, I would argue, are what made the choice of any particular form of representation so contentious. And this omission means that these authors also cannot give a historical account of why particular types of these drawings emerged in this period as the dominant way to represent artifacts, at least for the management of technical affairs. Finally, all these authors fail to acknowledge the severe limitations on any

510

attempt to master the physical world solely by means of visual representations. Our analyses of representations - at least as they related to activities (like engineering or architecture) which are engaged in manipulating the material world - cannot remain stuck in the two-dimensional world of images, but must follow the efforts of engineers to translate their images into physical objects, typically through their use of mediating physical instruments.

Many 18th-century theoreticians of the workplace agreed that one of the principal tools for organizing the workshop was technical drawing. As I noted earlier, Diderot’s plea for a public ‘science’ of technology culminated in the call for the development of technical drawing - a ‘geometry of the workshop’. [33] Since the Encyclopédie was itself to be a public repository of technical knowledge, Diderot devoted considerable effort to the plates which pictured technology. He recruited many contributors and illustrators to do this work, and thereby convey his message about the value of public discussion in achieving technological progress. Most scholars have recently read these plates as revealing Diderot’s hostility to the guilds. They point out that the artisans in them are generally portrayed as anonymous labourers, cut off from the boisterous life of the workshop, silently bent at their tasks. The argument here is that reducing the artisans’ skill to a set of routine procedures is a sort of intellectual proletarianization. [34] But as John Pannabecker has recently noted, in a project as vast as the Encyclopédie, many of Diderot’s contributors were artisans themselves, and some found scope to offer very different representations of technical work that gave partial voice to the tacit skills that were at the heart of their craft. And as for the artifacts themselves, they are depicted in a variety of ways, in perspectival views and projective views, as cut-aways and in disassembly, in schematic views and in operation. This reflects the tradition of the Renaissance collections known as ‘Theatres of Machines’ - which the Encyclopédie consciously emulated - as well as Diderot’s attempt to reach a larger lay audience [35]

But when we turn from the collections of pictures found in scientific atlases and the Encyclopédie to the sort of technical drawings which were actually taught in technical schools and used in workshops, this diversity of representational forms falls into a clearer pattern. Eighteenth-century France saw the beginning of a vogue for technical education centred on a drawing curriculum. Across the Revolutionary divide and across the divide of social status, drawing education served as the core curriculum in French technical education. We can identify at least three sites where technical drawing was taught, each with its own preferred form of representation: (1) the thousands of workshops where experienced artisans individually taught free-hand drawing techniques to their journeymen; (2) the scores of state-sponsored part-time scholarship schools in which academic drawing masters taught basic geometry and classical drawing to apprentice artisans; and (3) the handful of advanced state engineering schools run by the artillery service, the Corps du Genie, and the Corps des Ponts et

511 Index

Chaussées, in which mathematics professors taught mechanical drawing, including the descriptive geometry, to engineering students. [36]

As ‘instruments’ to assist in the organization of the workshop, the different forms of technical drawing taught in these various sites implied (but did not require) very different degrees of discretion for conceivers of artifacts and makers of artifacts, and hence a very different set of social relations between these groups . But technical drawing is more than a barometer of such changes. The very vehemence of the debates over the most appropriate way to represent technical objects suggests that these forms of technical drawing were also considered to be a tool for creating a new productive order.

A sketch or ‘free-hand’ drawing emphasizes the open-endedness of the design of an artifact - and of the ambiguous roles of its conceiver and maker. The rules of drawing here are ill-defined, even idiosyncratic. This is a quasi-private language, used as an extension of the creative process, or as a kind of private notation to oneself or one’s immediate colleagues. [37] Such a drawing implies a high degree of trust between the designer and executor of the object. At the limit, they may be one and the same person. For instance, artisans in the furniture trades used free-hand sketches as a bridge between their tacit knowledge and their manual skills; their drawings did not exhaust or replace their skills. That is because even when they copied patterns from others, or used geometric forms, they still exercised discretion about how to implement their designs. [38] This was the form of drawing Jean-Jacques Rousseau recommended for his imaginary artisan-pupil, Emile. Rousseau instructed Emile to sketch directly from nature, so he might learn to see for himself and learn skills which would allow him to be intellectually and financially independent. [39] This sort of drawing, then, implied the creative and economic autonomy of the artisan as artiste.

This differed from the form of drawing taught in the more than 20 part-time drawing schools for artisans established by the French state in the middle of the 18th century. The largest of these, the Ecole Royale Gratuite de Dessin in Paris, exemplifies the contradictory attitudes of elite pedagogues as they set out to teach drawing skills to artisans - and to reform craft practice. This Parisian scholarship school, founded in 1766 by Jean-Jacques Bachelier, trained some 4000 student-apprentices in the two decades before the Revolution. The course began with instruction in elementary geometry. Thereafter, students enrolled in one of three curricula - architecture, figures and animals, or flowers and ornaments - each of which involved tracing some 2300 sequential academic drawings in the neoclassical style. None pictured mechanical devices. [40] Bachelier believed that geometry served as a ‘mould for the operations of the mind’, and would make artisanal work more ‘precise’ by teaching students the ‘exact knowledge of the dimensions of objects considered under various aspects’. His real enemy here was the artisan’s ‘ignorant and prejudiced’ imagination; only geometry could ‘prevent the imagination from flying off, and contain it within the bounds of reason’. The neo-classical style, too, would wean artisans from the wild and ungainly designs of their primitive

512

imagination. Self-discipline in taste correlated with self-discipline in the workshop. Bachelier believed that his school gave the habit of work to young men who otherwise tended to be lazy and disorderly. And he asserted that this discipline had practical results: ‘From certainty in work comes promptitude in execution; [and] rapid execution will unleash the industry of the nation by lowering prices’. [41] At the same time, however, Bachelier’s course played to the artisan’s aspirations for autonomy and pride in his craft. The school was to secure for ‘each artisan the ability to execute by himself and without outside help those different works which his particular genius for his art enables him to imagine’. It is no accident that the school’s funding came from aristocratic patrons and the leading guilds of Paris. The productive world Bachelier envisaged remained that of the independent handicraft worker governed by the norms of corporatist culture. [42]

Index

Views from Nowhere

The third, and largely triumphant, form of technical drawing - mechanical drawing - was developed in the engineering schools of Enlightenment France and is still taught today in technical schools throughout the world. Mechanical drawing, it is worth emphasizing, itself comes in two basic forms, each associated with different professional milieu. First, there is perspectival drawing developed by rationalist artists in the Renaissance to convey ‘realistic’ views of figures, landscapes and machinery. [43] Second, there is projective drawing, long used by architects (in profile, plan and elevation) to guide the construction of buildings, and increasingly given mathematical form by technologists interested in designing and constructing a variety of artifacts. Both these forms of representation are rule-based, and both claim to offer a one-to-one correspondence with the material world. [44] And both were taught to engineering students. But the differences between them are important too.

In a sense, projective representations function within engineering culture much the way perspective functions within lay and scientific culture: as a picture of ‘the way the world really is’. But this analogy can be misleading. Engineers and architects use projective views because they avoid the distortions of shape that Renaissance artists intentionally introduced into their pictures to give the illusion of depth. As Descartes pointed out, perspective is a deception set aright by the judgement of the mind’s ‘inner eye’. [45] Perspective drawings are ‘views from somewhere’ and, hence, still within the realm of the personal (albeit a readily translatable ‘personal’). Projective drawings, by contrast, look nothing like the ‘real world’, yet they introduce no distortions of shape. Such drawings are objective in Lorraine Daston’s sense of being aperspectival; they are the negation of subjectivity. First adapted for the fine arts by the Renaissance-mathematician, Albrecht Dürer, they became increasingly appealing to technicians in the 17th century. As Abraham Bosse noted in the latter part of that century, projective views are the equivalent of perspective views

513 Index

seen from infinitely far away - except that they are close up. They are truly ‘views from nowhere’. [46]

Projective drawings achieve this effect, in part, by reducing the representation of objects (and their decoding) to a set of formal rules. The goal is to limit the discretion of both the person drawing the plan and the person interpreting it. In this sense, we may say that a projective drawing is an objective picture of an artifact, even though it ‘looks’ nothing like the artifact. A projective drawing binds those who use it to a common vision of the object by overcoming at least three layers of potential misinterpretation. First, a projective drawing bridges the epistemological mistrust that exists between the inner eye and the external world. For those trained in its rules, it allows for a full reconstruction of the pictured object on exactly the same scale as the original. Second, a projective drawing creates a common intra-group conception of an artifact across space and time. This feature made projective drawings particularly useful for those bureaucratic organizations which had to coordinate far-flung activities. And third, a projective drawing helps bridge the chasm of mistrust that lies between groups by providing a common referent. This feature made these drawings useful at sites, such as the workplace, where diverse individuals had divergent interests.

All these features made projective drawing a particularly appealing form of representation for the French state engineers of the Enlightenment. In the first half of the 18th century, the drawing professor at the Mézières fortification school, Amédée-Francois Frézier, admonished his students to reject perspectival drawings as inadequate if they wished to speak to subordinates with a minimum of ambiguity; for these purposes, only projective views would do. [47] Analogous techniques of projective drawing were being taught at the artillery schools in the same period. In the 1740s, the commander of the artillery school at Metz could claim that the importance of drafting for engineering students was so widely recognized as to need no defending. According to Jean-Pierre Du Teil, who directed the Auxonne school when Lieutenant Bonaparte was in residence, mechanical drawing was indispensable to all artillery officers. Under the guidance of a drawing master, students began with drawings of the natural terrain or strongholds from various ‘geometric’ perspectives. They then moved on to exercises in rendering fortifications, artillery batteries, and civil architecture. And from there they made technical drawings - in elevation and profile - of actual cannons and carriages kept in a special salle des modèles. This drawing curriculum showed students how the design of these cannon and carriages conformed to geometric constructions. The leaders of the artillery touted these lessons as providing students - these sons of petty noblemen and bourgeois notables - with a common body of knowledge, a ready means of reconstructing designs while far from the arsenals, and a set of tools with which to direct craftsworkers and manage the complex tasks involved in producing these artifacts (see Figure 1 – HHC: figures not included). [48]

These attributes of projective drawings were intensified by the descriptive geometry, a mathematicized method of mechanical drawing formalized

514

515 Index

HHC: Figure 1 not reproduced

in the 1760s by Gaspard Monge at the Mézières École du Genie, and taught to successive generations of French military engineers. Monge called the descriptive geometry ‘a [universal] language necessary to all those who work in the mechanical arts’ because it allowed one ‘to represent with exactitude, on drawings which have two dimensions, those objects which have three, and which can be rigorously defined’. Certain artisans, such as masons, had long possessed secret stereographic methods for calculating the various block faces needed to build, say, a Gothic vault. These techniques had been generalized by Desargues in the 17th century. Monge’s descriptive geometry further extended this generality by referring all representations to universal axes, and by tying these views to mathematical analysis. In particular, it showed how regular three-dimensional objects could be mathematically generated by the movement of two-dimensional lines. As a result, the descriptive geometry was also a powerful ‘constructive’ technique, and could be used to search for new shapes and configurations. For instance, it helped engineers solve problems in stonecutting, optimal fortress construction and even machine design. [49]

To be sure, Monge always acknowledged that the descriptive geometry could not be easily applied to the thick things commonly used in commercial and military life. He believed that his limitation, however, only increased the moral value of the descriptive geometry as a tool for training students. As he said:

[I]f, from a young age designers had been trained in the study of the lines of curvature of different surfaces which are susceptible to exact definition, they would be more aware of the form of those lines and their position, even for objects less [readily] defined; they would [then] grasp them [mentally] with greater precision and their work would be more expressive. [50]

This suggests the central paradox of mechanical drawings: these forms of representation seek to preclude the illustrator’s judgement about how to represent an object, but at the same time, one of the central motives for training engineers in this technique is to form their judgement about what are proper objects and how to manipulate them. [51]

Indeed, the very rigour of this training suggests that the descriptive geometry is not a ‘natural’ representation, but a cultural convention which arose historically and reflects its creators’ view of their place in the broader social order. The authority of mechanical representations derives from the self-discipline necessary to make one. Before engineers could use pictures of this sort to command workers, the drawings themselves had to be highly ordered entities. Engineering students spent years learning the self-restraint that enabled them to picture only certain carefully defined characteristics of thick objects. In this way, mechanical drafting defined the social role of engineers in late ancien régime France as the designers of artifacts, placing them as intermediaries between state patrons and artisans: vis-à-vis patrons, projective drawings created a legally enforceable standard which made them accountable to their superiors; vis-à-vis workers, projective

516

drawing distinguished between the conception of an artifact and its execution, suggesting how one might redistribute tasks within the workshop, while still preserving a common language for both elite technologists and artisans. These twin aspects of technical drawing - as an analytic method and as a social marker - appealed enormously to the Encyclopédistes and contemporary engineers.

Index

Why Do Engineers Cast Shadows?

Of course, for these representations to organize the workplace, they had to be readable by all those involved in production, including those ranked near the bottom of the workshop hierarchy. This explains, for instance, why engineers cast shadows. Strictly speaking, shadows provide no information not already given in the projective views; on rational grounds they are unnecessary. Nevertheless, engineering officers in the ancien régime were taught to calculate shadows, since the mastery of this technique was deemed ‘necessary to discipline and perfect drawing’. But shadows offered more than an interesting exercise in geometric construction: they also ‘rendered representations more distinct’. As engineers recognized, it was often easier to draw an artifact in projective views than to reconstruct it mentally from the multiple drawings. By adding shadows and tints, engineer-writers absorbed some of the difficulty of representation so that patron-readers and worker-readers might more easily interpret their drawings, thereby preserving the correspondence between the hierarchy of expert knowledge and the social hierarchy (see Figures 1, 2, 3, 5 and 6). [HHC – not included] [52]

The use of these new forms of technical drawing also required an expanded programme of pedagogy for artisans and shop floormen. Thus Antoine-Laurent Lavoisier made technical drawing a centrepiece of his Revolutionary proposals for popular education. He professed deep concern for the growing split between elites ‘who studied languages and the objects of science and literature’, and those ‘destined for the mechanical arts’. To bridge this divide (and still preserve the social hierarchy), Lavoisier emphasized early training in ‘graphical geometry’ for all youngsters in primary schools.

Just as there exists knowledge that must be common to all men no matter what profession they are destined for, so must there exist knowledge common to all who work in the mechanical arts. Drawing, it seems to us, must be ranked among this type; drawing is a language of the senses that speaks to the eyes, which gives existence to ideas, and from this point of view, expresses more than words; it is a means of communication between he who conceives or orders [an artifact], and he who executes [it]; finally, considered as a language, it is an instrument proper to perfect ideas; drawing is therefore the first study of those who are destined for the mechanical arts. [53]

Implementing this pedagogical programme became controversial in the Revolutionary period, when some of the conflicts over the early École

517 Index

Polytechnique became refracted through the question of how much and what kind of technical drawing should be taught to whom. As a founder of the first, egalitarian, and truly ‘polytechnic’ Ecole Polytechnique, Monge taught the descriptive geometry to his diverse body of engineering students to give them a feel for material objects, practice for their manual skills, and a sense of learning by doing. He and his disciples also tried to see that the technique was taught in the new École Centrales that were to give provincial students access to practical education. [54]

After 1795-96, however, and with gathering force after 1800, technical drawing came to be one of the pedagogical subjects that defined the stratified cognitive order, ranking the state’s various educational institutions and the students who graduated from them. While the École Polytechnique was increasingly reserved for wealthy, elite students, and its curriculum refocused on abstract analytical mathematics (including more abstract uses of the descriptive geometry), a range of ‘lesser’, more practically oriented schools developed in which pupils were taught the forms of technical drawing appropriate to their station. These vocational schools proliferated in the 19th century - including the École des Arts et Métiers and the Conservatoire des Arts et Métiers - and they came to play a crucial role in the dissemination of drawing techniques to the foremen and mechanics who organized production on the workshop floor. [55]

Index

The Limits of Representation: Picturing Guns

Let me emphasize that there is no necessary connection between a particular way of representing an artifact and a corresponding socio-technical order. As Shoshana Zuboff has shown for computerized representations of work, the switch from manually guided machinery to numerical controls did not impose a particular form of power relations upon the workplace. In some work sites, the computer representations permitted a blurring of old distinctions b