DECENTRING SOCIOLOGY SYNTHETIC DYES AND SOCIAL THEORY * Andrew Pickering

Transcription

1 34,200 words DECENTRING SOCIOLOGY SYNTHETIC DYES AND SOCIAL THEORY * Andrew Pickering University of Illinois Department of Sociology 326 Lincoln Hall 702 S. Wright St. Urbana, IL pickerin@uiuc.edu July 2001 * This essay has evolved over a considerable period of time. Several presentations at workshops and colloquia have helped to clarify my thinking. In particular, I thank Geof Bowker, Manuel De Landa, Fernando Elichirigoity, Niels Viggo Hansen, Bruce Lambert, Randi Markussen, Finn Oleson, Amit Prasad, Rick Powers and Boyd Rayward for valuable conversations, and the historians of organic chemistry and the dye industry Ernst Homburg, Ursula Klein, Evan Melhado, Carsten Reinhardt, Alan Rocke, Tony Travis and

2 Henk van den Belt for very useful comments and much additional information. I gratefully acknowledge the award of a Guggenheim Fellowship which made it possible for me to begin work on this project, and I thank the Departments of Philosophy and Information and Media Science of the University of Aarhus and Ursula Klein s group at the Max Planck Institute for the History of Science, Berlin for their hospitality while I was working on drafts of the essay.

3 DECENTRING SOCIOLOGY SYNTHETIC DYES AND SOCIAL THEORY Abstract This essay addresses the difficulties that sociology as a discipline continues to experience in grasping the relations between technology, science and the social. I argue that these difficulties stem from a resolute centring of sociology on the social, which follows a generically Durkheimian blueprint. I elaborate a response to these difficulties which derives from recent lines of work in science and technology studies, and which entails a decentring of the social relative to the material and the conceptual, in terms of both objects of analysis and explanatory formats. In order to display the direct sociological relevance of this approach, I move to the macro level, developing the conceptual apparatus of a decentred social theory via the discussion of a historical example of major sociological significance the systematic gearing together of scientific research, technological innovation and industrial production that originated in the synthetic dye industry in the second half of the 19th century. Besides the interest of the topic itself, the aim is thus to exemplify in some detail what a decentred sociology might look like, both empirically and conceptually.

4 p. 1 We live in an increasingly technoscientific world. All sorts of technologies and sciences permeate and help to structure our daily lives; national and global economies depend on them. Information technologies, to think of a current example, have become almost omnipresent in just the last few years. The streets of Europe and North America are full of people talking not to each other in old-fashioned face-to-face interaction, but to cellphones. When we get chance to sit down on buses and trains, in bars and pubs we tend to ignore our fellow travellers and drinkers in favour of the buttons that send and receive text-messages. The very stuff of social life has been technologically worked over and transformed in all sorts of ways infotech is just a recent and striking instance and sociology as a discipline needs to get to grips with that if it is to maintain its relevance in the third millenium. But it proves easier to say this than to do it. Science and technology are hard objects for sociology to digest. In the next section I will explain some of the difficulties one can perceive in some of the usual sociological approaches, which have to do with their resolute centring on the social. These difficulties have been the object of concerted attention in science and technology studies for some time now, and the solution I want to elaborate here extends an approach developed in the work of Donna Haraway, Karin Knorr Cetina, Bruno Latour, myself and others, which entails the adoption of a more symmetric, decentred, perspective that seeks to grasp at once the social, the extrasocial and their interrelations. 1 This decentring move involves something of a reconception of what sociology is as a discipline, and there has been no shortage of theoretical critiques nor of responses to those critiques. 2 But, adapting Steven Shapin s memorable line (1982, 158), one can either debate the possibility of a decentred sociology or one can do it. Like Shapin, I want to follow the latter course. I want to try to inject the decentred perspective into the mainstream of the discipline, and, my particular strategy here is to move to the macro a terrain that should be of great interest to sociologists. I want to show that a decentred analysis is appropriate to a major social transformation at the heart of our technoscientific social world. My aim is to exemplify empirically and theoretically what a decentred macrosociology can look like, in the hope of encouraging others to follow. 3 The example I choose to look at is not cell-phones and 21st century consumption. I want to go back to the 19th-century, and to industry and production, because there, in a sense, lie the origins of modern technoscience. The second half of the 19th century was the time of the so-called marriage of science and industry (also known as the second industrial revolution ) the first large-scale, organised and systematic gearing together of scientific research and industrial production and I want to concentrate on the major and highly consequential social transformations that went into this marriage. Two science/industry couples underwent this transformation at roughly the same time the physics of electromagnetism and the electrical industry, and organic chemistry and the synthetic dye industry. I am going to focus on the latter. Historical developments there are especially fascinating, involving, as they did: (1) major and highly consequential social transformations the birth of a new industry, the systematic coupling of academic scientific research and industry, the transplantation of scientific research into industry, changes in the law; (2) major technological transformations the development of new material pathways in the mass production of dyestuffs; and (3) major transformations in science the inception of modern organic chemistry. Together these transformations created a blueprint or model for the formation of the academic-scientific-industrial

5 p. 2 complex that now dominates the global economy. In all sorts of ways, for example, the history of the dye industry foreshadowed contemporary developments in information technology and biotechnology. This episode in 19th-century history is, then, worth studying. If we can grasp the coupling of science, technology and society back then, we are well on the way to grasping similar couplings in 21st-century technoscience. The structure of this essay is as follows. In section I, I seek to clarify the contrast between centred and decentred approaches in the social sciences. Section II provides an overview of the historical developments that will concern us. Sections III to V are the empirical heart of the essay, covering in turn relations between the technological and the social, transformations in scientific research, and the birth of the industrial research system. Section VI returns to the general discussion. Drawing upon the historical narrative, I summarise the form of my decentred analysis, and I also discuss the strengths of more centred accounts as well as their weaknesses, as a way of both understanding their claims on our intuitions of how science, technology and society relate to one another and clarifying my own position. The essay concludes with remarks on the theme of time and change, which I first introduce in the following section. I. CENTRING AND DECENTRING I need to begin with some remarks on social theory; not on the details of specific positions, but on the general forms in which relations between science, technology and the social are usually conceived. I want to explain what I mean by saying that these are centred on the social. The place to begin as so often in the social sciences is with Emil Durkheim. Durkheim s strategy in founding sociology as an autonomous science was first to define its pure object that domain of social facts and phenomena which emerges from the collective features of human life and which cannot not be reduced to the subject matter of other sciences, such as psychology or economics (or physics or engineering). Having thus specified sociology s disciplinary object of analysis, Durkheim s next move was to partition it. His idea was that one could divide the social into two parts, effects and causes, the latter often thought of as having to do with social density. Durkheim s principle was that in the social domain, as in the domains of the other sciences, like causes like social effects have social causes. The job of the disciplinary sociologist was thus defined to be that of isolating these effects and causes and displaying how they were chained together. Durkheim s sociology was, then, centred on the social in a double sense: (1) its unit or object of analysis was purely social, excluding the proper objects of the other sciences (which objects were, nevertheless, acknowledged to exist), and (2) the causes of social phenomena were to be sought within the sphere of the social itself. Much contemporary writing in sociology respects this basic Durkheimian blueprint, however much it departs from Durkheim in other ways, and it might seem almost redundant to remark upon such social centring here. But things become more interesting when we ask about how science and technology might fit into the picture. Probably the standard sociological view of science and technology remains the Durkheimian one: that science and technology are, presumably by virtue of their rationality and materiality, other than the social, beyond the purview of the sociological

6 p. 3 gaze. Science and technology thus appear as, at most, a boundary condition on the social, as something given from without. Some extension of the unit of analysis is permitted. The sociologist can notice that our factories and coffee bars are full of computers. But computers themselves are not a sociological topic. The job of the social scientist is to find within the social the social traces or echoes, as it were, of the technology the social meanings attached to computers, say, or social attitudes towards them and to enquire into other social variables that might cause or at least be correlated with those to ask how race, class and gender, say, affect social attitudes towards computers. The analysis, then, remains empirically and theoretically centred on the social. One variant of this position is worth noting, a position often called technological determinism (in section VI, I will also discuss a related position one could call scientific determinism, but for simplicity I will leave that on one side for now). Here science and technology are regarded as something more than a simple boundary condition on the social. Instead and one can certainly find this thought in Marx developments in science and technology are regarded as autonomous motors of social change. The handmill gives you society with the feudal lord; the steam engine, industrial capitalism. In this story, a double centring operates. On the one hand, we have a history of science and technology evolving autonomously from the social, centred, say, on the rationality of science and the powers of technology. This is not the sociologist s to tell; hence labels like technological determinism, which point to the extrasocially centred story. On the other hand and this is where the phrase technological determinism can be misleading added to this one usually finds a story of social accommodation to technology, where sociologists can continue to operate their usual social centring, but now with a moving boundary condition. 4 One attractive feature of technological determinism, I can note now, is that it can indeed recognise science and technology as entities that the social can have consequential and transformative encounters with. On this view the social does not have to bear all the weight of its own history. Finally, there is a more radical position. In The Elementary Forms of Religious Life Durkheim indicated how the apparently extrasocial might be absorbed into sociological analysis. One might think that the topic of Durkheim s book, the spiritual, is demonstrably extrasocial, involving nonsocial supernatural forces and hence quite escaping Durkheim s definition of the social. But in the Elementary Forms Durkheim showed how one might extend the object of analysis to include the substance of religious beliefs and rituals and still be a sociologist. Crucially, Durkheim insisted that to do sociology one had to continue to centre one s analysis on the social, classically defined. Standard variables like social density, residing in the heartland of sociology, were still to be called upon to explain spiritual phenomena and not the other way around. To be a sociologist, rather than a theologian, was to refuse to consider that the spirits summoned Australians to those interesting get-togethers. 5 And, of course, Durkheim s move was one of the threads that flowed into the sociology of knowledge tradition, wherein the contents of many different bodies of knowledge are explained by reference to the Durkheimian social, and eventually into the sociology of science and technology. In the well known sociology of scientific knowledge (SSK) approach, for example, the unit of analysis is indeed broadened to include the contents of scientific knowledge, but the prescription remains to look for explanations of specific scientific beliefs in terms of purely social variables like social structure and social interests. Such approaches to the sociology of

7 p. 4 science and technology, then, broaden the unit of analysis to include the substance of science and technology, but remain determinedly centred on the social in their explanatory structures. So much in review of some stock positions in the sociology of science and technology and their common centring in the realm of the social. I have introduced them by way of Durkheim, but I do not suggest that we are all Durkhemians. What I do suggest is that in their broadest outlines these positions share a Durkheimian blueprint in setting up their objects of analysis and doing a distinctly sociological job on them. Now I want to indicate what is wrong with them, and the easiest way to do that is to set them against one another. On the one hand, empirical work under the banner of SSK and its counterpart in the sociology of technology sometimes referred to as SCOT, for the social construction of technology has demonstrated that the substance of science and technology is not some pure other to the social. 6 One can indeed centre the analysis of the substance of scientific and technological developments on purely and classically social variables. On the other hand, it is hard to deny that the position I called technological determinism has something going for it. If technology is not an absolute other to the social, certainly it is not reducible to it. The specific powers of steam engines, say, to drive locomotives and ships or other machines in the factory, are not themselves explicable in terms of social structures and interests. Those powers can only be excluded from sociological analysis by a kind of methodological bracketting. 7 And this bracketting has unfortunate but unavoidable effect of excluding a fascinating topic for sociological enquiry that technological determinism at least addresses, however defectively, namely, the question of how the social is itself transformed in scientific and technological developments. 8 This is where recent work in the sociology of science and technology comes in. As mentioned above, at least one branch of work in science and tehnnology studies has aimed at finding a form of accounting and theorising that can recognise the reciprocal structuring of science, technology and the social, that while the social does indeed structure scientific and technological developments, those developments also structure the social. And the hallmark of this work is a decentring of the object of analysis and of the way relations within that object are conceived. Following the lead of the Elementary Forms, the object is expanded to include the substance of science and technology; but, departing from Durkheim, SSK and SCOT, the attempt to centre the analysis on the classically social (or the classically material, the second centre of technological determinism) is abandoned. These new approaches thus move out of the socially centred Durkheimian space that has characterised sociology as a discipline for the past century or so which is probably why the discipline has found them hard to come to terms with (much harder, say, than SSK). The aim of what follows is therefore to try to make them more accessible to show how a decentred sociology might be done, and why it should be done, via the working through of an example of evident sociological interest. Part of the job is to develop some basic theoretical apparatus, and I need to say a few words on that before moving on to history. (I will only return to a sustained consideration of centred sociologies of science and technology in the last section of the essay.)

8 p. 5 First, in the narrative to follow, one is confronted with a spectacle of cultural multiplicity. Some of the elements one can straightforwardly name as social: the varying social roles played by scientists in science and industry, for example, and laws concerning property rights. Others can be straightforwardly considered material (chemicals, say, and their reactions); others conceptual (specific scientific theories). Others, like the architecture of the industrial research laboratory, cut across such distinctions. I follow the basic Durkheimian schema in separating out the categories of the social and the extrasocial, but I do not want to go any further in theorising their distinguishing marks, because the object is, after all, to thematise their interrelation in the constitution of the huge, heterogeneous, social-material-conceptual assemblage of modern technoscientific production. 9 I do seek, however, to delineate the extra-social elements of the story as clearly as I can, focussing on the powers and performativity of the material world chemicals, instruments and technologies since I take these to be integral to any understanding of how the social can be transformed in scientific and technological practice. 10 The key notions I deploy in analysing the relations between different cultural elements I call here translation and tuning. Translation refers to the movement of elements from one setting to another in the space of multiplicity; tuning refers to the open-ended adjustments that typically prove necessary if translations are to be successful (and, I emphasise, such success is never guaranteed in advance). The notions of translation and tuning also relate to a second major theoretical point that comes to the fore below. This concerns the problematic of time and change. Briefly, I think it is fair to say that sociology, and, indeed, the social sciences in general, are not good on the topic of time and change. The problematics of the discipline often exclude time completely, in favour of an emphasis on synchronic correlations. When change is on the agenda, the focus is usually, in fact, on the unchanging parameters of change: causes, contexts, constraints, etc. In contrast, in what follows I seek to develop an analysis of change in itself. The notions of translation and tuning refer to an openended becoming of culture, the actual historical trajectories of which emerge in the realtime of practice, rather than being explicable by reference to any pre-existing parameters. I should stress from the start that these two theoretical concerns with decentring and with time and change intertwine in my historical narrative. In my opinion, the way to get to grips with the coupling of the social and the extrasocial is to follow the temporal evolution of both terms. 11 II: OVERVIEW Above ground, the steel rolls out fiery, bright. But to make steel, the coal tars, darker and heavier, must be taken from the original coal. Earth s excrement, purged out for the ennoblement of shining steel. Passed over. We thought of this as an industrial process. It was more. We passed over the coal-tars. A thousand different molecules waiting in the preterite dung. This is the sign of revealing. Of

9 p. 6 unfolding. This is one meaning of mauve, the first new color on Earth, leaping to Earth s light from its grave miles and aeons below... You think you d rather hear about what you call life : the growing, organic Kartel. But it s only another illusion... If you want to know the truth I know I presume you must look into the technology of these matters. Even into the hearts of certain molecules... The shade of Walter Rathenau, speaking to Generaldirektor Smaragd of IG Farben, at a séance, in Thomas Pynchon, Gravity s Rainbow (pp ). 12 Today, we take it for granted that academic scientific research and development are vital to industry and the health of the economy. Our hopes for the future are pinned upon sophisticated developments in information technology, biotechnology, and the like, and we look to university laboratories to keep those developments coming. Research universities seek quite self-consciously to gear themselves into the productive activities of industry, via patent offices, industrial partnership programmes, and the like. 13 The modern university, one can say, is an integral part of a scientific-industrial complex or assemblage the source of new scientific ideas, new devices and new technologies for industry, and, of course, the training ground for the engineers and scientists that populate research laboratories located within industry itself. Reciprocally, research universities are more or less dependent upon financial and material support from industry, or from state agencies themselves concerned with industrial and military initiatives. It was not always thus. In the early 19th century, such systematic and institutionalised connections between the universities, scientific research and industry did not exist; it was only in the second half of the 19th century, and then only in a few industries, that the modern scientific-industrial complex began to come into existence. The empirical aim of this essay is to explore the formation of this science-industry complex. Sociologically, one might want to imagine that this was primarily the product of a convergence of interests that academic scientists and industrialists simply saw that collaboration with the other would serve their own particular purposes and acted accordingly and that all that was entailed was some sort of convenient juxtaposition of two preformed institutions, science and industry. But I want to show that the process was more complicated and interesting, more difficult and uncertain, than that. My argument below is that the gearing together of science and industry entailed a transformation of social interests and social structure themselves in a process which integrally involved material transformations of both science and industry and conceptual transformations in science. Industry, science and the universities were no longer the same social entities after this process had run its course than they were before; and the specific forms they took on have to be understood in relation to the material and conceptual transformations that helped to constitute them. This is the socially decentred picture I want to sketch out and analyse, in which the social, the material and the conceptual can each be seen to have served to structure the others dynamics, without any of them, as it were, calling all the shots. As I said, one of the first sites of this lock-in of science an industry was the conjunction of the synthetic dye industry and scientific research in organic chemistry, and a

10 p. 7 discussion of fig. 1, a schematic diagram of the 19th-century material flows and social relations of dye production, will help to introduce the historical process to be examined in more detail below. FIGURE I: MATERIAL FLOWS & SOCIAL RELATIONS OF DYE PRODUCTION At the heart of the industrial revolution was the industrialisation of textile production. Ancillary to that was an industrialisation of textile dyeing and printing. Until the middle of the 19th century, dyes were extracted from traditional vegetable and mineral sources. The madder plant was a key source of red dyes, for example, and the indigo plant of blues. 14 The plants were grown commercially, harvested and shipped to dye producers, where the colouring matter was separated from its vegetable matrix, and the resulting dyes were then shipped to printing and dyeing works. This is the pattern depicted in fig 1(a). Beginning, however, with the discovery of a new dye, mauve, in 1856, an evergrowing range of synthetic dyes were introduced, at first alongside the natural products but eventually displacing them from the market. 15 Fig 1(b) is a schematic map of the world s first scientific-industrial complex, and shows the very different material flows and social relations of dye production in the late 19th century. Several features need to be noted. First, at the raw material end, synthetic dyes were not derived from living matter. Instead, the starting point was coal. Second, by the late 19th century, the industrial production of dyes had become an intrinsically chemical affair, synthesising brightly coloured dyestuffs from non-coloured materials by distinctly chemical means. Third, the synthetic dye industry, unlike its earlier counterpart, was coupled to the universities and scientific research in at least two ways. On the one hand, academic chemists were directly involved in the dye industry as consultants helping to devise, for example, new or improved dye syntheses. On the other hand, a branch of scientific research had split off from the universities industrial scientific research, which was supported in-house by the dye companies. The universities also functioned, then, as a training ground for industrial researchers. A new career structure had appeared, joining the universities and industry. And, along several axes, the law especially patent law was integral to this wedding of academic science and industry. Figures 1(a) and (b) have the quality of before and after snapshots. My aim in what follows is to trace out and analyse the historical transformations that lay between them. These transformations were indeed social ones. They were transformations in the social relations of production. They involved breaking the social links between madder farmers and indigo planters, on the one hand, and dye producers, on the other, and establishing new links between the dye industry and tar-distillers and the like. More importantly in retrospect, they involved the construction of new links between the dye industry, academic scientific research and the law. 16 But what we will see is that these social transformations cannot be understood in purely social terms. They entailed complex translations and tunings that linked the social, the material and the conceptual. The dye industry was tuned materially to its new source of raw materials via the development of novel chemical technologies; academic science was tuned to industry both materially and

11 p. 8 conceptually; industrial scientific research split off from academic science and was retuned socially, materially, conceptually and even architecturally to its new home; the law was transformed in its very content. These heterogeneous transformations, tunings and translations and their interrelations are at the heart of my historical narrative. In Part III I analyse the formative years of the synthetic dye industry, the period from 1856 to roughly 1870, emphasising the coupling of social relations and material technology. Academic scientific research became strongly coupled to the dye industry from the late 1860s onwards, and that coupling is the focus of Part IV. There I introduce some more specialised analytical apparatus, specifically related to the historical problem in hand. I tackle the problem of differentiating science from extrascientific social formations by characterising it as a distinctive form of life. I then analyse organic chemistry s extrascientific alignments in a discussion of the mundane world as a surface of emergence and return (in many senses) for scientific research, and I discuss the reciprocal tuning of science and industry in terms of a double translation of money and materials, analyses and syntheses between them. Part V focusses on the establishment of the then-novel system of industrial research, linking science directly into industry, which I conceptualise as an enfolding of the former by the latter. 17 III. TECHNOLOGY AND SOCIETY Mauve The synthetic dye story starts with coal. Burning coal drove the steam engines that powered the Industrial Revolution. But other uses for coal were found, too. When coal is distilled, inflammable gas is driven off and tar remains. Such distillation was first undertaken commercially in Britain in the late 18th century for the sake of the tar, used in waterproofing ships. From around 1812 onwards the gas found a use in gas-lighting, to such an extent that the tar produced in gas manufacture exceeded demand, becoming a largely useless byproduct. 18 One response to this was further distillation of the tar into various fractions (light oils, creosote, pitch) for which various markets could be found (Beer 1959, 9). 19 Scientific analysis of light coal oil began in 1843, with the work of August Wilhelm Hofmann in Justus Liebig s laboratory at the University of Giessen. 20 In the space of a couple of years, Hofmann demonstrated the presence of aniline and benzene in coal oil, outlined the manner in which these two substances could be obtained in pure form, established a workable method by which benzene could be nitrated and then reduced to aniline, settled the formula for aniline and some of its homologues, and, finally, determined aniline s basic properties (Beer 1959, 13). Hofmann s work on coal oil was the beginning of his distinguished career in organic chemistry, and in 1845 he moved to England as the director of the new Royal College of Chemistry in London (Beer 1959, 18). There, in 1856, one of his students William Henry Perkin, then aged eighteen sought to succeed in a project previously and unsuccessfully attempted by his master, the synthesis of the drug quinine. Perkin initially began with the coal-tar fraction toluene, but then switched to the derivative just mentioned, aniline. Perkin in his turn failed to produce any quinine, but in the course of his attempts he stumbled upon the following recipe for the production of an intensely purple substance which came to be known as aniline purple or mauve: 21

12 p. 9 I take a cold solution of sulphate of aniline, or a cold solution of sulphate of toluidene, or [etc]... and as much of a cold solution of a soluble bichromate as contains base enough to convert the sulphuric acid in any of the above-mentioned solutions into a neutral sulphate. I then mix the solutions and allow them to stand for 10 to 12 hours, when the mixture will consist of a black powder and a solution of neutral sulphate. I then throw this mixture upon a fine filter, and wash it with water till free from the neutral sulphate. I then dry the substance thus obtained at a temperature of 100 o C. or 212 o F., and digest it repeatedly with coal-tar naphtha, until it is free from a brown substance which is extracted by the naphtha... I then free the residue from the naphtha by evaporation, and digest it with methylated spirit... which dissolves out the new colouring matter. I then separate the methylated spirit from the colouring matter by distillation at a temperature of 100 o C. or 212 o F. Perkin immediately thought of this as a recipe for dye production, took out a patent on it and, together with his father and brother, set up the factory shown in figure 2 for the industrial production of mauve at Greenford Green near Harrow. 22 Commercial mauve production began there in late 1858 and, in retrospect, the establishment of the Greenford Green factory can be seen to have marked the beginning of the synthetic dye industry (Travis 1993, ch. 2). I therefore begin my analytical commentary here. 23 FIGURE 2: PERKIN S MAUVE FACTORY Perkin s dye factory represented an evolutionary leap of the social, a transformation of social ontology and topology. A new kind of entity appeared on the social landscape, a site of synthetic dye production, connected not to traditional producers of animal and vegetable raw materials, but to tar distillers. 24 The social world was thus changed and rearranged in Perkin s industrialisation of his achievement. And it is important to emphasise that this development was indeed a social one. Perkin and his relatives could, for example, have gone bankrupt in this venture, a distinctly social outcome; instead, Perkin went on to make his fortune. The coming of the synthetic dye industry was, therefore, an event in the history of Victorian capitalism. At the same time, it was clearly not purely a social event. It hinged upon new flows and transformations of matter: coal to tar to mauve to coloured cloth. The question then is, how should we think about the relation between the social and the nonsocial elements in these developments? To address that, we can start with the recipe for mauve. Perkin s discovery of the mauve recipe was the discovery of a new transit of matter, a new way in which matter behaved or performed these substances, under these conditions, yielded this product. And it seems reasonable to say that there was nothing intrinsically social about the recipe. Certainly it could not be reduced to anything social no combination of naked human minds and bodies could emulate its performance. The discovery of mauve was an event in the history of matter, and it was a discovery of an emergent material phenomenon emergent in the sense that it could not have been known in advance that it would appear; it just so happened that this transit

13 p. 10 of matter arrived at mauve. However, we can note that the recipe in itself was insufficient to sustain an industry, and this is where things become sociologically interesting. Having discovered his recipe, Perkin s next step was to effect a translation of it. I mentioned translation as a key analytical concept earlier, and here I can note that I use it in a simple and straightforward manner which hangs together with my earlier remarks on multiplicity and that refers to the business of moving elements (material, social, conceptual or whatever) from one cultural (and often geographic) location to another. Thus here Perkin attempted the task of translating his dye recipe from one institution the laboratory, where it had originated to another institution industry, where it was hitherto unknown. 25 Translation can, however, create its own problems, and I can now approach a discussion of my second key term, tuning. Perkin s laboratory methods were sufficient only for the production of small quantities of mauve. Industrialisation thus entailed a scaling-up of the recipe and the development of production technology adequate to the manufacture of economically significant quantities of mauve. And here we can note, first, that in moving towards industrial production, Perkin s laboratory version of his recipe clearly functioned as a model for creative extension. Without the laboratory example before him as a model for what might be accomplished, Perkin s industrial dreams would have been inconceivable. But, second, we can note that models do not prescribe their own future trajectories of extension. An indefinite number of modelling vectors can be constructed in relation to any given model; modelling is an open-ended process. 26 In this case, for example, given that Perkin was initially looking for quinine, it would have been no surprise if he had regarded the appearance of coloured sludge in his experiments as a sign of failure and abandoned them. Likewise, it would have been no surprise if, reasoning by analogy to quinine, he had sought to explore mauve s pharmaceutical properties rather than its utility in dyeing. Of course, given the pre-existence and importance of the printing and dyeing industry, it is no special surprise either that Perkin s thought moved in that direction, and one can see that his goal of industrialisation was, in that sense, socially structured. The social picked out a set of modelling vectors along which Perkin sought to extend his material model, lending a degree of closure to his practice. In effect, then, the development of the industrial technology of dye production represented a tuning into the social of the original mauve recipe (as in tuning a radio set or a car engine). My notion of tuning is thus intended to conjure up a notion of the open-endedness of cultural extension, coupled with the idea that closure in such extension is a function of the achievement of alignments with other cultural elements. 27 The third point to note is that this tuning could have failed. The material world need not have gone along with Perkin s ambitions. Working with large quantities of the relevant chemicals was physically hazardous, for example. Travis (1993, 38) remarks that Hofmann considered Perkin s industrial ambitions foolhardy and he speculates that Hofmann s caution may have been linked to the memory of a fatal accident that befell another brilliant student and aspiring chemical entrepreneur, Charles Blanchford Mansfield. Early in the previous year, a still for separating hydrocarbons in coal tar naphtha caught fire, and Mansfield met a slow and agonising death from the severe wounds that were inflicted. 28 Specifically in the production of mauve, a great deal of heat (as well as obnoxious fumes) was produced in the production of nitrobenzene from

14 p. 11 benzene and in the subsequent conversion of nitrobenzene to aniline, leading to considerable danger of explosions. As Perkin later recalled, discussing his early attempts to produce nitrobenzene, several explosions occurred, but fortunately without causing any injury to the workmen attending the apparatus (quoted by Travis 1993, 52). 29 All of this should, of course, help to make clear that the material powers of technology are not reducible to anything social, that they are not subject to human will, for example that here the social, the birth of a new industry, was at stake in relation to something nonsocial. Eventually, however, in this instance it turned out that these hazards could be dealt with, and Perkin survived the translation of his recipe into industry. Initially he used large glass flasks for the production of nitrobenzene and aniline, which were kept cool (and thus discouraged from exploding) probably by immersion in tubs of water or by being hosed. In 1858 or 1859 Perkin switched to the use of larger and stronger cast iron reactors, shown in fig. 3, for the same purposes (Travis 1993, 47, 53). 30 And here I need to stress three points. First, a certain kind of material tuning was at work in Perkin s development of the technology of dye production. We should think of him open-endedly exploring a space of material arrangements and their performances. Second, the successful performance of the technology he settled on was again an emergent material phenomenon, irreducible to anything social. Perkin could no more have known in advance that he would succeed than Mansfield could have known that he, Mansfield, would fail; it just so happened that Perkin achieved a certain technological grip on the material world, leading to the reliable and repeatable production of mauve in quantity. But third, despite the irreducibility to the social of both Perkin s recipe and production technology, still, as already discussed, the trajectory linking the two was socially structured. Perkin s traverse of the space of material possibilities leading from the recipe to industrial technology only makes sense considered in relation to the pre-existing textile, dyeing and printing industries. 31 FIG. 3: PERKIN S IRON REACTOR FOR ANILINE PRODUCTION Where does this commentary leave us? It leaves us, I hope, with a clear image of both the irreducibility of material performances to the social and, nevertheless, of a coupling of the technological to the social. The performativity of Perkin s original recipe and his industrial technology existed in the material world, but the trajectory leading from the former to the latter was structured by considerations of production and capital the technology was tuned into the social world. But at the same time, the shape of the social world was itself at stake in these developments. As I mentioned earlier, the social became something new ontologically and topologically in the founding of the synthetic dye industry. And the last point to emphasise here is that the becomings of the technological and of the social hung together and interactively stabilised one another in this episode, becoming parts of a unitary but heterogeneous assemblage. 32 The evolutionary leap of the social was sustained by an evolutionary leap in the space of material performance the discovery of the dye recipe and a further translation and evolutionary tuning of the recipe into the space of the social, a successful tuning, as it turned out. One cannot, then, centre one s understanding of this social transformation within the Durkheimian social. The transformation of the social has to be understood instead in relation to the evolution

15 p. 12 of the material technology at the heart of the dye industry. But neither can one understand the latter in classic technocentric fashion without reference to the former. The image is of the material and the social co-evolving, structuring each other s dynamics, and underwriting one another in the Greenford Green dye factory. This is the basic decentred image that I want to elaborate in what follows, though the story gets richer. MAGENTA I described Perkin s mauve recipe as a model for open-ended variation, and, as just outlined, one such line of variation led to Perkin s production technology. Another line of variation that followed Perkin s success was the search for alternative recipes, other synthetic pathways to mauve. The third and most original line of development was the search for recipes like Perkin s but leading from aniline to other new dyes. Tinker with aniline was the strategy (van den Belt and Rip, 1987) a kind of groping around in the space of material transits and aniline was accordingly subjected to the action of just about every industrial reagent ready to hand. 33 The first productive outcome of this tinkering was a recipe for the production of a red dye variously known as aniline red, magenta or fuchsine discovered by Emmanuel Verguin in Lyon in late 1858 or early The magenta recipe was even simpler than that for mauve: it consisted only in the heating of aniline with anhydrous tin bichloride... in enamelled iron pots without any subsequent purification of the product. In the patent specification the product is dissolved in water and the dye precipitated by adding salt (Hornix, 1992, 70). 34 Verguin sold the details of his process to the Lyon company of Renard Frères for 300,000 Fr. payable over fifteen years; in April 1859, Renard Frères patented a process based on Verguin s, and manufacture began in May of that year (Travis 1993, 69). Magenta proved very popular with consumers, and in various quarters further modelling then ensued, aimed at finding alternative synthetic pathways. 35 Another familiar reagent arsenic acid proved to do the trick, and in 1860 the aniline producers Simpson, Maule & Nicholson acquired the British rights to the arsenic acid method and went into magenta production themselves. In July 1860 a new blue dye was discovered, apparently by chance, when excess aniline was included in the preparation of magenta by the arsenic acid method. Renard Frères acquired the rights to this aniline blue, and Simpson, Maule & Nicholson produced it under licence in England beginning in Other new dyes quickly emerged from further tinkering, including greens and violets, so that by 1866 a whole family of differently coloured aniline dyes was in production (Travis 1993, 100). Here again, then, we see the social structuring of explorations of the space of material performance. From the indefinite number of ways in which first Perkin s and then Verguin s recipe could have been extended, commercial considerations picked out certain avenues of investigation. And, conversely, inasmuch as those investigations succeeded in pointing towards commercially significant dyestuffs they reinforced the nascent social ontology of the synthetic dye industry. Social and material developments continued to hang together and structure one another in a decentred fashion as they had in the mauve story.

16 p. 13 Beyond that, though, in the post-magenta period another shift in social topology began to take place, linking the new industry now to academic research in organic chemistry. It would be a mistake to overestimate the role of chemists in this period up until 1868, say. Simple trial-and-error tinkering was probably more important. But still, the first systematic links between the industry and university science were established then, and I want to discuss a couple of axes along which they developed. 36 At issue here are what one can generically describe as consultancy relations. One coupling of science and industry had to do with patents. 37 In the wake of magenta s success many companies set about producing it, leading to a sequence of legal actions in France and Britain centring respectively on Renard Frères and Simpson, Maule & Nicholson, who sought to exercise monopolies in production. The details of the patent cases are complicated and unedifying (van den Belt 1989, Travis 1993, ch 5), but the following remarks are sufficient for our concerns. In the patent cases the question was one of sameness and difference: were the dyes produced by different companies the same or different, and were the processes of manufacture the same or different? Answers to these questions were not obvious. One cannot tell just by looking whether one red dye is the same as another, especially when the dye companies sold (as we would now say) impure mixtures of chemicals that had rather different effects in dyeing. As far as production processes were concerned, one might think that matters were more straightforward: did each company use the same reagents and reaction conditions or not? But still questions could be raised about whether differences were important or not was the substitution of one reagent for another a trivial change, or did it represent a new synthetic pathway? In the attempt to decide such questions, the dye companies appealed to the expertise of academic organic chemists, whose life revolved around questions of analysis and synthesis (as discussed in section IV below). Chemists were produced in court to testify as to the composition of the dyes in question and to the relation between synthetic pathways. We should note that the chemists rarely agreed in their opinions. 38 In fact, the correlation between opinions and paymasters was remarkable. 39 Nevertheless, scientific evidence was the primary evidence in these legal actions, and in this way a solid link was established between the synthetic dye industry and organic chemistry. Some commentary is appropriate here. Clearly, this link constituted a further transformation of social topology, constituting a systematic bond between two social institutions that had initially had only opportunistic connections, via Perkin, for example, and his move from the laboratory to the factory. After the formation of the synthetic dye industry itself, it was the next step towards the construction of the scientific-industrial assemblage that is the focus of this essay. At the heart of this assemblage lay what one can call a double translation. Materials, such as dye samples and money (as fees to expert witnesses) flowed from industry to the laboratory, and information, analyses of new dyestuffs and opinions on syntheses, flowed back. And we need to recognise that this double translation served to reconfigure both of the entities that were party to it. Thus, the contours of the emergent dye industry were themselves at stake and liable to transformation in this intersection with science. The most basic issues of ownership and capital were decided with reference to scientific

17 p. 14 opinion (Renard Frères held onto their monopoly on magenta in France; Simpson, Maule & Nicholson lost theirs in Britain in 1865). At the other end, within science, a new social role was established, that of the academic consultant to the synthetic dye industry. Many organic chemists undertook such consultancy work, in patent cases and beyond; Hofmann was pre-eminent amongst them as the world s leading expert on the chemistry of aniline. 40 In the same process, of course, organic chemistry was encouraged to develop in a particular direction, with chemists developing often intense interests in all matters concerned with synthetic dyes a point I will return to in the next section. Science and industry not only came together, then; they were reciprocally transformed in the patent suits over new dyes. They remained separate entities they could not be reduced to or substituted for one another but via the double translations that linked them they were, to a degree, tuned to one another. The becoming of each helped structure the becoming of the other. Now for an example of the second axis along which science and industry began to intersect in the aniline-era. The story is again of a double translation between the factory and the laboratory and, as often for this period, it concerns Hofmann. Following the discovery of magenta, Hofmann was quick to obtain pure crystals from Simpson, Maule & Nicholson, which he analysed (as he had aniline, years before), assigning magenta a chemical formula and giving it a scientific name, rosaniline. 41 This work was central to Hofmann s contribution to the patent suits. But he actually went further than that. On the basis of his analysis, Hofmann reasoned that it might be possible to synthesise more new dyestuffs and this proved to be correct. Hofmann succeeded in creating new violet dyes, patented in 1863, which Simpson, Maule & Nicholson then manufactured and sold under the name of Hofmann s violets (Beer 1959, 27-29). This episode is important because it illustrates a quite different interrelation of science and industry from that displayed in the patent suits. In the patent suits, the chemists had a largely parasitic role, analysing the products of industry; the synthesis of Hofmann s violets, in contrast, exemplified a growing generative and dynamic role for scientific research in academic settings. 42 Once more, though, we can note a reciprocal tuning at work. Organic chemists, typified by Hofmann, increasingly oriented their research towards future industrial possibilities, while industry was itself transformed as the producer of specific new dyestuffs and, more generally, in developing an increasing interest in possible pay-offs from chemical research. Again we can see here how specific social transformations in industry and its relations with science were coupled in a decentred fashion to material ones in science, to particular analyses and syntheses. The social convergence of the two institutions, industry and science, depended upon the upshots of chemists explorations of the powers and properties of matter. It was not a purely social process, and the social was itself reconfigured in it. ALIZARIN AND AFTER The 1860s were the years of the aniline dyes and of the dominance of the synthetic dye industry by England and France, the leading industrialised nations. In the background, however, the forerunners of some companies that are very well known today were emerging in Switzerland CIBA and Geigy and Germany Bayer, the Badische Anilin- und Soda-Fabrik (BASF), Hoechst and AGFA (later to join forces as the massive IG Farben conglomerate). Initially these companies took their lead from the English and