Enacting Environments: An Ethnography of the Digitalisation and Naturalisation of Emissions Inaugural-Dissertation zur Erlangung des Doktorgrades der PhilosophischSozialwissenschaftlichen Fakultät der Universität Augsburg vorgelegt von Ingmar Lippert aus Mainz 2012 Supervisors Christoph Lau (Augsburg University) Lucy Suchman (Lancaster University) Date of Defence 12th December 2012 Enacting Environments An Ethnography of the Digitalisation and Naturalisation of Emissions Ingmar Lippert PhD Dissertation in Sociology Augsburg University Augsburg Cite in a version of the following form: Ingmar Lippert (2013). Enacting Environments: An Ethnography of the Digitalisation and Naturalisation of Emissions. PhD Dissertation in Sociology, University of Augsburg, Augsburg. c ⃝ Ingmar Lippert 2013 Typeset with LATEX 2ε on Tuesday 8th January, 2013 Licence Copying is permitted only for private or the own scientific purposes of the person who performs copying (according to §53 of the German Copyright Act). This publication may not be edited or changed otherwise. Commercial, business lobbyism, military or policing use is not permitted. For any other type of use request permission from the author. The author has no responsibility for the persistence or accuracy of URLs for external Internet Web sites referred to in this publication and does not guarantee that any content on such Web sites is, or will remain, accurate or appropriate. Some material of this book has been published elsewhere: an earlier version of Section 2.1 can be found in Ephemera 12(1/2), 2012; segments across this book have been published in Computational Culture 1(1), 2011. Abstract Enacting Environments is an ethnography of the midst of the encounter between corporations, sustainable development and climate change. At this intersection ‘environmental management’ and ‘carbon accounting’ are put into practice. Purportedly, these practices green capitalism. Drawing on fieldwork of day-to-day practices of corporate environmental accountants and managers, Ingmar Lippert reconstructs their work as achieving to produce a reality of environment that is simultaneously stable and flexible enough for a particular corporate project: to stage the company, and in consequence capitalism, as in control over its relations to an antecedent environment. Not confined to mere texts or meetings between shiny stakeholders co-governing the corporation – among them some of the world’s biggest auditing firms, an environmental non-governmental organisation (NGO) and standards – control is found to be distributed across as well as limited to a myriad of practical work situations, involving spreadsheets and slide shows. Carbon accounting takes place in the midst of docile as well as dissident humans and nonhumans. As a result of this analysis, Enacting Environments establishes how carbon emission facts are produced and co-configure climate change realities. Ingmar Lippert argues: within capitalism, environment does not exist in the singular but in the plural; and these environments are not existing out there to be read off some anterior Nature but they are brought into social, economic and political existence in the practices of accounting for them. Providing a portfolio of methods to study techno-managerial engagement with carbon, Ingmar Lippert shows how much is overlooked in received theories of corporate environmental accounting, theories of the performativity of environmental economics and, ultimately, the epistemic and ontic effects of fact-making in the heart of neoliberal capitalism. Keywords: Science and Technology Studies (STS), environmental sociology, organisation studies, actor-network theory (ANT), apparatus, multiplicity, carbon accounting, corporate environmental management. Ingmar Lippert is Residential Fellow in Tembusu College at National University of Singapore. Before his lectureship at Tembusu, he was a Research Fellow at the Institute for Advanced Studies of Science, Technology and Society, Graz and a PhD student at Augsburg University. His research focussed on the situated social and material practices in the doing of ecological modernisation as a technomanagerial attempt to green capitalism. Outline of the Scholarly Argument Capitalism manages to enact environments in the midst of its centres by means of keeping Other environments out. The fundamental practice which allows for this contradictory and generative move is that capitalist agents enact environments. Capitalism does not require a clear, neat, distinct, singular environment. Multiple, fluid, dynamic environments allow far better the tactical and strategical project of staging capitalism as having its destructive environmental impacts in control. That control is a decisive fiction sustaining the unsustainable. These theses are the result of an ethnography, reported in this book, that scrutinised corporate carbon accounting practices as a site at which we are able to simultaneously explore two significant issues for the management of environments: on the one hand studying practices of corporate environmental accounting allows us to engage with agents’ practical work reality by which capitalism seeks to render itself ‘green’ and ‘sustainable’; on the other hand the focus on precisely how accountants achieve taking carbon into account is able to sharpen our understanding of how quantifying practices perform in a non-substantial area of business, such as engaging with climate change. In the received view, corporate carbon accounting is about providing the facts and figures about the emissions which a company produces. Accounting for these emissions is supposedly a condition to take carbon into account – economists would call this process internalisation. The discourse which assumes that ‘if only capitalist society is able to internalise its environmental problems’ (like carbon emissions which are identified as the culprit of global warming and, in consequence, climate change) ‘then capitalist society will be able to solve environmental crises’ – this discourse is called ecological modernisation. Within environmental sociology arguments over whether that discourse is actually materially reflected abound. Ecological modernisation theory proposes that capitalist organisations do get green(er). In the midst of debate, little attention, if at all, has been paid to those agents who are, supposedly, implementing the programmes of ecological modernisation, such as environmental management systems (EMS) and carbon accounting. This study contributes to understanding how capitalism organises its relation to environments by means of scrutinising the work practices of these agents. To conduct that study, I have carefully avoided to make assumptions about whether a particular organisation would be conducting greenwash. Much rather, the intentionally open question was: what do agents of ecological modernisation do and how do they achieve it? With this orientation, this study turned to sociological theory and methodology which does not presume any overarching structure as determining agents. Instead, by means of methodological triangulation between ethnomethodology, actornetwork theory (ANT) and Pierre Bourdieu’s concepts of field and habitus, this study reconstructs by which specific practices and discursive action agents manage to make greening more central to capitalism. The decisive finding is that while agents do manage to bring environmental data into the heart of the corporation – the centre of capitalism – what that data is about (the things this data is related and presumably representing, the material hinterland of that data) is simultaneously distanced from the corporate core. This study, thus, shows how capitalism manages to enact a epicentral movement of ‘environment’ and, in parallel, to ensure that environmental issues and concerns do not challenge or interfere in that centre. It manages by means of keeping the largest degrees of environments out. The overarching thesis of this study is, thus, that environments, such as carbon, are not existing – for all practical purposes of corporate agents – out-there but, rather, they are carefully crafted and enacted into corporate, social and, eventually, economic reality. Environments are enacted. The plural matters. Within the organisational practices of capitalism, agents may imagine to refer to ‘the’ environment. Their everyday practices of taking environments into account, however, relate to specific materials, such as spreadsheets, pieces of papers, flipcharts. Environments exist through these multiple materials, in multiple versions; ontologically, thus they do not exist in the singular but they are staged as such. If the carbon emission fact of a company is established, that fact may well be out-dated a few micro-seconds or years later; it may differ several kilometres off or in a neighbouring storage unit in a computer. Any global fact is enacted in particular located situations. Emissions facts are not stable but fluid, flowing in and between myriads of situations. They are hold together by means of humans’ material-discursive performances. And they shift with agents’ practices just like with the dynamics in-built into materials, like a database. These processes produce artefacts, versions of environments. And these versions matter. What a society is dealing with when encountering a corporate emission fact is not Nature but a version of an environment. Next year, the same fact (as in, signifying the same imagined out-there) may have changed. Vis-à-vis Science and Technology Studies (STS), my analysis of the effects of enacting environments is indicative of a potentially general characteristic in digital quantification practices – whether in offices or in laboratories: data flows are not that clean and under control. While workers may achieve staging being in control, in practice parallel versions of realities may proliferate – for the better or worse. What we find is that the reality of corporate carbon emissions is enacted as mutable, mobile and multiple. In the practical work of corporate fact finders, it is not necessary, albeit it is deemed required, that facts are singularised and immutablised. In consequence, social and economic reality is confronted with diverse carbon emission accounts, all implying universal truths. Societies and politics which resist engaging with parallel realities and insist on singular ones may not be well equipped to manage those crises that may be co-constituted by these parallel realities. Ethnographic work underlying this argument involved participant observation over a period of twenty months as well as document analysis. The study took place at a Fortune 50 financial services provider positioned in a legitimising network involving one of the largest international environmental NGO’s and one of the four largest auditing firms. The findings of this study are, thus, considered to point to practices indicative of widely organisationally accepted and shared realities within hegemonic modern capitalist culture. Note: This text is NOT to be read by investors; especially not by those in the ethical investment business. Disclaimer : Names and numbers within this text have been altered to guarantee anonymity of subjects encountered during the underlying field research. Numbers of the text have been designed to mirror realistically magnitudes and the claimed accuracies present in the actual field. However, in this book, seemingly accurate numbers serve merely illustrative purposes. Language: This book has been written in British English. Quotes have been rendered accordingly. The appeal of numbers is especially compelling to bureaucratic officials who lack the mandate of a popular election, or divine right. Arbitrariness and bias are the most usual grounds upon which such officials are criticised. A decision made by the numbers (or by explicit rules of some other sort) has at least the appearance of being fair and impersonal. [. . . ] Quantification is a way of making decisions without seeming to decide. Objectivity lends authority to officials who have very little of their own. (Porter 1995, 8) No doubt some of the things in the world can indeed be made clear and definite. Income distributions, global CO2 emissions, the boundaries of nation states, and terms of trade, these are the kinds of provisionally stable realities that social and natural science deal with more or less effectively. [. . . ] So what are the textures they are missing out on? (Law 2004a, 2) The necessary point of departure for renewed socio-environmental theoretical elaborations [. . . ] is the hyper-modernist, post-calculative, disembodied, entrepreneurial agent that represents the theoretical engine of neoliberalism and the primary responsible actant for the conflation of manufacture or assemblage and proprietorship in its politics of nature. (Pellizzoni 2011, 802) Contents 1 Introduction 1.1 Greening Carbon . . . . . . . . 1.2 Accessing Carbon . . . . . . . . 1.3 Doing Data . . . . . . . . . . . 1.4 Researching Corporate Carbon 1.5 This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interlude I: Contested Carbon 1 4 21 31 41 57 63 2 Data Construction in a Basement 69 2.1 Classifying Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . 70 2.2 Constructing Consumption via Calculation . . . . . . . . . . . . . . . . 88 2.3 Commanding Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 3 Translating Data Into Sustainability 3.1 Breaking Down Sustainable Development 3.2 Controlling Data Collection . . . . . . . 3.3 Boundary Emissions . . . . . . . . . . . 3.4 Extension of Carbon Coverage . . . . . . 3.5 Visionary Emissions for Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 123 142 168 179 204 Interlude II: Normal Anomalies 227 4 Situated Control Over Carbon 4.1 Performing Qualities of Data . . . 4.2 Failing the Market . . . . . . . . 4.3 Aligning Legitimacy with Carbon 4.4 Transforming the EMS . . . . . . 4.5 Meeting Data . . . . . . . . . . . 231 238 256 279 320 344 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Achieving Order 363 5.1 Adjusting Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364 5.2 Blurring Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378 5.3 Betting on Bodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395 xi Contents 5.4 xii Accommodating Mess . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405 Interlude III: Sustaining Labour Relations 421 6 Carbonscape 6.1 Boundaries 6.2 Materiality . 6.3 Multiplicity 6.4 Temporality 6.5 Scale . . . . 437 443 457 474 488 499 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interlude IV: Companion Tales 517 7 Conclusion 531 A Fieldwork and Methods 555 B Analytic Materials B.1 A Comment on Relating Bourdieu and Science and Technology Studies B.2 A Bourdieusian Analysis of Data Reporting . . . . . . . . . . . . . . . B.3 Analysing the Relations between the Corporation and External Organisations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.4 Analysing the Structure of the Workplace . . . . . . . . . . . . . . . . B.5 Auditor’s Critique of GFQ’s Carbon Accounting . . . . . . . . . . . . . B.6 Outsourcing Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.7 Materials on 3C Group’s Carbon Offsetting Project . . . . . . . . . . . B.8 Additional Field Materials . . . . . . . . . . . . . . . . . . . . . . . . . 583 583 587 Lists List List List List List List List List 612 615 619 621 623 625 627 629 631 of of of of of of of of Field Note Extracts . . . . . . . . . . . . Artefacts . . . . . . . . . . . . . . . . . . Figures . . . . . . . . . . . . . . . . . . . Tables . . . . . . . . . . . . . . . . . . . Members’ Names Anonymised . . . . . . Organisational Designations Anonymised Organisations . . . . . . . . . . . . . . . Standards and Rankings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601 605 608 609 610 611 Glossaries 633 Bibliography 639 Index 673 Acknowledgements This book would have not been possible without many people. First, I like to express credit to members of the company I studied (its pseudonym is Global Finance Quality (GFQ)): I am more than grateful for their time and their trust – access to their everyday work territory was invaluable; they shared their thoughts, happiness, joy and frustration. I am protecting the identities of these informants and, therefore, render their names and the names of most organisations subject to this study anonymous. Greatest thanks to Christoph Lau and Lucy Suchman who were providing me with enormously helpful scholarly and practical guidance, who were willing to supervise my engagements – with GFQ, sociology and STS. Christoph Lau made me ask how this study challenges social theory. Lucy Suchman carefully engaged with the nuances in carbon work situations that I reported and supported me to delve into the differences and resonances between my study and approaches in the conceptual reality of STS. How to express the gratitude to all of you? Let me approach this by way of telling the story how the fortunate path to this book has woven through intellectual and practical land- and taskscapes. All people I met on this path enabled the journey. I am deeply grateful to you for accompanying me on that path, finding it, mapping it, showing me directions and supporting me in some many ways – too many to be named! At Brandenburgische Technische Universität at Cottbus I had been familiarised with the hegemonic mode of approaching the management of ‘nature’ and ‘humans’. Gerhard Wiegleb, Wolfgang Schluchter and Klaus Kornwachs supported the development of my critique on the hegemonic ways of studying environmental management. During my first semester at Boğaziçi Üniversitesi, Belgin Tekçe led my path towards sociology and it was there that I first engaged with the theme of environmental managers’ agency that I scrutinise in this book. In parallel, the discussions with Susann Eisert, Lea Herberg, Gudrun Pollack, Anke Schaffartzik and Stephan Wolf opened the wider frame of questioning the limits of environmental management. Thanks to head librarian, Magdalene Frewer-Sauvigny, the library soon shelved critical management texts. In the intellectual space spanning Lancaster University’s Institute for Environment, Philosophy and Public Policy (IEPPP), the Centre for the Study of Environmental Change (CSEC) and the Centre for Science Studies (CSS) I developed the methodological and theoretical frame underlying this book. Noel Cass decisively classified one of the key objects of my study when he suggested to think of environmental managers as ‘agents of ecological modernisation’. In conversations with Brian Wynne, Bron Szerszynski and xiii Claire Waterton I learned to scrutinise conceptual and theoretical spaces and deepen my interpretations. Helen Jackson (whom I explored Lancastrian environmental STS with, alongside Uli Beisel, Aurora Trujillo Peraire and Michaela Spencer) helped me to align my inquiries towards societally and politically productive struggles. All the staff and students I met at Lancaster have shaped this journey. At Graz’s Institute for Advanced Studies on Science, Technology and Society I met Anup Sam Ninan, Torsten Wöllmann, Corinna Bath and Kolya Abramsky with who I had generative discussions of how to operationalise critique in STS. At the intersection of the spaces of Lancaster, Graz and Cottbus soon formed the Environment, Management and Society Research Group – in which I intensified my engagement with Anup Sam Ninan and Niklas Hartmann. Their thoughts have been reassuring and critical throughout this study. This study travelled. Parts of it have been presented at Graz’s Institute for Advanced Studies on Science, Technology and Society as well as the city’s social centre Spektral, at Nottingham at the UK Postgraduate Conference in STS, at the Amsterdam Social Technology Workshop, at the XVII World Congress of Sociology at Gothenburg, at Trento’s conference of the European Association for the Study of Science and Technology, at the Workshop ‘A Billion Gadget Minds: Thinking Widgets, Data and Workflow’ at Swedenborg Society, London, at the Workshop ‘Carbon markets and their future: A Social Science Perspective’ at Hamburg, at London’s Corporate Watch, at Lancaster’s Centre for Science Studies/Centre for the Study of Environmental Change Mixture Seminar, at the Berlin Innovation in Governance Research Group’s Lunch Seminar, at the Cardiff Interpretative Policy Analysis Conference, at Naples’ Critical Management Studies Conference, at a Seminar of Berlin’s Humboldt University’s Institute for European Ethnology as well as the Berlin Frozen Flow Workshop and at Bielefeld University’s Centre for Interdisciplinary Studies. So many helpful and encouraging voices responding to elements of this study! Among them the supportive discussions – some longer, some on the spot – with Steffen Böhm, Colin Brown, Günter Getzinger, Les Levidow, Adrian Mackenzie, Andrew Sayer, Arno Simons, Steven Yearly and all the respondents to my work at these gatherings. Others providing valuable comments on this study, at different phases, have been Judith Ancke, Rebecca Ellis, Joseph Huber, Elham Kashef, Reiner Keller, Alfredo Ramı́rez-Ogando, Peter Seyferth, Lydia Stiebitz, participants of the sociology colloquium run by Christoph Lau, activists of anarcho-sydicalist organisations and anonymous reviewers of an article published in Ephemera (2012). This study would not have been possible without incredible amounts of support in infrastructure and reproductive work. In that respect I am thankful to my family, friends, the anarchist community, the couchsurfing network. The study was not financed by any organisation; still my work was supported with scholarships by the German National Academic Foundation as well as Hans-Böckler-Foundation and the ManfredHeindler grant administrated by IAS-STS. Of course, I take responsibility for this text and hope errors and problems will be signalled to me. xiv CHAPTER Introduction 7th of July 2009. We are in the heart of a Western metropolis, in one of the G8 states, on the third floor of an office block, hidden behind urban flats. The block which I am in houses the Head Quarters (HQ) of a multinational: Global Finance Quality (GFQ). I am in an office. The office is equipped with a flip-chart, a grey filing cabinet – on it, note the green leaved plant, neighboured by an event flyer; the flyer carries the logo of The Club of Rome. Turn around. More furniture: a small desk with two chairs and a larger desk. I am visiting a woman in an unobtrusive pantsuit. Her name was printed on that flyer. She must be important. Important – in Latin – means ‘being of consequence’. Is she? Well, she occupies the boss’s chair. We sit at the larger desk and look at a screen. Right-handed, she moves a mouse, connected to a computer. Synchronically a pointer takes its way on the screen. She presses the left button of the mouse. Click. A consequential digital change. It operates the ‘ok’ button at a website of a global carbon reporting organisation. Yeah!!! Triumphantly, she raises her arms. In the moment of pressing the mouse button to click onto ‘ok’ she released tens of megatons of carbon emissions into the global discourse of carbon, into an utopian atmosphere of green capitalism and techno-economic fixes to climate change. It is people like her who bring carbon emissions into being. This book substantiates this thesis by way of visiting the specific work practices of her and her colleagues – those who collectively produce accounts of GFQ’s impact on ‘the’ environment. Before detailing the book’s position in relation to sociologies of the environment and unsustainability, to the debate of the performativity of economics and Science and Technology Studies (STS), I provide a quick sketch of the political concerns. So-called carbon emissions were something many were concerned about while I was writing this book. Substantially, however, this book is not about carbon. It is about practices which bring corporate impacts on ‘the environment’ into social, economic and political reality. These impacts are now framed in terms of carbon; tomorrow they may be in terms of water, biodiversity, ecosystem services. At a glance, this is the political discourse against which this book emerged: major play1 1 1. Introduction ers – such as governments, the United Nations (UN), multinationals, non-governmental organisations (NGOs) – figured out that ‘we’ need sustainable development. Companies are engaging with environmental crises, such as climate change. It is time to act. Whether this imperative is cast as hegemonic ideology (Gramsci 1971), symbolic violence (Bourdieu and Wacquant 2004), discourse (Foucault 1981), matter of concern (Latour 2004b) or of hope for business innovation (Giddens 2008) it is clear that humans and non-humans are framed as dependent on companies – to go green. This is the prescription of the economist-managerial take: following governments, societies are to alleviate the problems by means of ecological modernisation. According to this approach, modern institutions have to use science – rational knowledge practices – to understand reality and develop technologies to repair institutions. Industrial and service sectors are main protagonists of greening: companies use so-called environmental management systems (EMS) to optimise their environmental conduct; climate change has become an issue of greenhouse gases (GHG) – carbon – which, unfortunately, was not allocated efficiently enough.1 A market failure resulted; this failure can be repaired by economically internalising carbon emissions. 2011’s markets internalised carbon, resulting in transactions worth 176 billion US$ (World Bank 2012, 10). The capitalist economy can be sustained if market participants are equipped with the required information to make decisions. Correspondingly, nation-states produce carbon inventories and companies provide environmental reports, detailing their emissions. The solution is carbon accounting. If only modern institutions internalise their emissions into their operations, economic consequences of climate change would be optimally allocated: intelligent management for a green economy (UNEP 2011, cf. Brand 2012). Or so the story goes. These discourses also detail how exactly organisations are to account for carbon. They specify standards, measurement technologies, softwares, databases and quality assurance. If it comes down to it carbon accounting is carried out by particular people. Agents who are to put the narrative of ecological modernisation into practice. In heavily regulated and standardised workplaces, they have to figure out their organisations’ emissions. Management technologies abound. Using specific conversion factors, calculatory techniques of extrapolation encoded into database functions, they produce a result. The solution to environmental crises is that such agents accomplish their job. Salvation depends on their practices of managerial-technological knowledge production. This book offers a study of the practices around this workplace. It provides an account of carbon accounting – an account sensitive to the lived realities of workers, of the problems they coped with and the materials they were organising around. Carbon. Data. Corporate emissions, for all practical purposes – of carbon accountants – were data. I argue, first, that this quality of emissions has consequences for how we should conceptualise both, ontology and epistemology of carbon. Interwoven with this line 1 Economist discourse can be understood as dealing with the distribution of resources and of damages within society. In their normative take, distribution (somebody distributes something to a location) and allocation (some location to which things are al-located) is to take place as efficiently as possible without causing damages to others. Labour unions share this discourse (The International Labour Foundation for Sustainable Development 2008; Allespach and Bartmann 2011). 2 of argument, we encounter how ecological modernisation is put into practice: I argue, second, that while ecological modernisation manages relations between organisations and some environments, mostly ecological modernisation increases the distance between organisations and other environments, resulting in a failure to engender environmental accountability. At the intersection of these lines, I argue, third, that digital knowledge practices unsettle the foundations of environmental economics and politics. My account of these practices adds to our understanding of a fundamental concern in our society: quantification-as-digital-practice. Enacting environments means that what an environment is is not given but more or less carefully assembled by particular humans, their practices and their devices that mediate their relations to what they imagine as ‘out there’. This book investigates how environments can be enacted by means of ethnographic analysis of such entities and their relations. GFQ provides us with a case of neoliberal enactment of particular environments. I argue, however, that what we can learn from this case points us beyond neoliberal enactments: we gain insight into infrastructural practices that any mode of modern capitalisms would need to utilise to take nature into account. In the remainder of this introduction I provide a systematic account of the case study and I show how it relates to debates in sociology and STS. This takes five steps. (1) I begin by tracing the social scientific takes on the greening of capitalism to the subject carbon accounting. This section should provide a feel for how carbon accounting occupies a firm position in hegemonic unsustainable politics. Carbon accounting is about producing data. That data matters for managing carbon. Readers who are familiar with social constructivist takes in environmental sociology and studies of accounting may consider skipping this section. (2) An understanding of how capitalism relates to carbon accounting should allow to make sense of the position of the case study. In the second section I present the field in which I observed carbon accountants and my position in it. I will also draw out a frame of how the company I studied, GFQ, related to carbon emissions. (3) The third section introduces how data and knowledge are produced. This entails engaging with the plural of knowledge – for we all know things differently – and with how humans shape and produce reality through their practices. Readers who are familiar with feminist technoscience will not miss much when skimming over this section. (4) Following this, I circumscribe the apparatus which I mobilised for this analysis. This entails engaging with the rationale of this study, methodological assumptions and the analytical process underlying this book. (5) This Introduction ends by summarising what this book is concerned with, how it seeks to contribute to sociological discourse and I lay out the organisation of the argument. 3 1. Introduction 1.1 Greening Carbon Sustainable development, climate change and carbon economics The notion of sustainable development serves well as the starting-point of this book’s narrative. Since the end of the second millennium, this notion has become a common reference point which agents have to mention when addressing the future of Humanity and Earth itself. The notion gained global momentum as a label for policies that promised to sustain economic growth while ensuring social justice and ecological equilibria. In these discourses, commonly it is held that sustainability may be achieved when three systems are reconciled, the economic, the social, the ecological. Graphically, three circles, more or less overlapping, or a pyramid are employed to map sustainability. It is in the centre, in an area where the three so-called aspects overlap. Critique of this model, even on the very notion, is everything but new (Mebratu 1998; Dingler 2003). Yet, even if this label is privately seen as flawed, agents publicly reproduce it; it has became a decisive anchor in global discourses, the reference for powerful agents to exhibit doing good. Hence, stakeholders, globally, in multinationals as in academia, in governments as in NGOs, all talk about the need to identify and employ workable transitions towards sustainability. Any such approach, in this discourse, requires some form of transition management: a goal-directive, learning-based more or less collective process. These processes, call them governance or management, need to be informed. With data. Realist data. These discourses do not provide much scope for constructivist agendas suggesting reflexive reconstruction of the priorities of economy, science and technology; there is not much scope for actually reorganising societal relations around ‘what kind of human we aspire to be, and in what kind of human world’ (Wynne 2002, 460). Much rather, realist data is used to determine what is best for the ‘public good’ – using economists’ instruments, above all, markets. Empirically, this book is about what was going on in the Sustainable Development Unit of the multinational GFQ. Whatever you think of sustainable development, to introduce this book’s theme, it suffices to accept that this firm related to ‘sustainable development’. In Chapter 3 I detail this relation and show that, indeed, the company did not only link to the label but also took part in reproducing the discourse, internally and publicly. Only at that point of my argument, it is required to engage in more depth with the sustainability debates, its history or major actors. Interestingly, the discourse of sustainable development was (maybe only temporarily) overwritten by a new ‘paramount environmental problem’, climate change (Szerszynski and Urry 2010, 2).2 Climate change was coded in a language of carbon equivalents, translated into emissions and emission reductions. This framing was the effect a contingent history of climate science, policy and economics. For social scientists, it is of interest how climate change is played out 2 Despite risks like nuclear power, exemplified in Fukushima in March 2011 while I was writing up the first draft of this argument, or, say, nanotechnology. 4 1.1. Greening Carbon because descriptions of climate change entail normative prescriptions about society (ibid., 4). Above all, climate change is often seen as necessitating a social ordering which is able to ‘master’ climate by means of earth system governance including international treaties like the Kyoto Protocol – environmental management on all scales ranging to consumption pattern changes (Hulme 2008). All these instruments to conquer the climate assumably allow humans to model and mould it. However, the science making climate change readable for so-called decision-makers – that is, bureaucrats, governmental representatives, business and NGO functionaries – is organised as regulatory science (Jasanoff 1994). This means, it is oriented to its presumed application (Szerszynski 2010, 19), well exhibited in discussions of e. g. general circulation models shown to involve – for good reasons – political concerns in modelling itself (Shackley and Wynne 1995). Therefore, we cannot make sense of climate change science in simplistic notions of science and objectivity. What science, objectivity and, thus, knowledge means has turned into an open question. In Latour’s (2004a) account, science and nature are political; yet ‘moderns’ pretend that politics and science/nature are distinct realms. Contra Latour, Asdal (2008) claims: the doing of nature-wholes explicitly opens spaces for politics. Eventually, Latour (2011b) points to climate change negotiations, e. g. at Cancun, as evidence of moderns’ successful operational engagement with the concern that nature is political, i. e. not distinct. This book, accordingly, sets out from a position, which sees scientific arguments involving so-called nature as political. Here, then, is a concern with how ‘objective knowledge’ of nature is political. Yet, rather than theorising this, I ask empirically what kind of politics is black-boxed into ‘nature’, into ‘carbon’. The climate change narrative usually involves a technological reading. In Szerszynski’s (2010, 10) account, it is the story’s constitution that ‘timeless truths of science will legitimate technological interventions’; science allows for reading climate change and suggests which so-called solutions are apt to write the climate. As it were, governments, business and NGOs agreed on specific marks to be written into earth’s climate, such as a limiting average global warming to +2◦ C. In this rationality it was seen as straightforward to identify the causal agent of warming. It was determined; GHG are the culprits, abbreviated as carbon. Below I return to a closer introduction of this entity. Right now, note that CO2 has become the enemy (Swyngedouw 2010a, 223). The primary mechanism to delimit this enemy was the market; that is, capitalist economics organised to assumably allow the optimal allocation and reduction of carbon. This presupposed the privatisation and commodification of carbon emissions. Ingeniously, governments inscribed this approach in international treaties – such as the Kyoto Protocol. This protocol proposed emission trading between entities as a rational mechanism to determine the best allocation of emissions and their reductions. For economists the idea is straight-forward – at least on paper. Here is an exercise. Take a piece of paper, construe three entities (it does not matter whether you imagine firms of nation-states; they are all the same – on paper). Entity A produces soy beans, 5 1. Introduction entity B manufactures military drones and entity C sells bank accounts. Three entities specialised on certain production. Now we copy this paper. As a result, you have six entities (label them A to F); two are competing within each sector. Here comes the sovereign (be it a friendly green benevolent dictator or an elected democratic head of state); governments set a limit: these six entities are to reduce their emissions by fifty hypertonnes. Now, who is to rearrange their production processes such that they emit less? The economist suggests that each entity calculates how much costs would result from measures to reduce emissions. Imagine a table: column (entities) and in the first row you note the particular costs for the means to prevent a first hypertonne of carbon to be emitted, for each entity; second row you inscribe the costs for the second hypertonne of emissions prevented – and so on. ‘Obviously’, the economist would predict, in different sectors and in different entities, different measures to reduce emissions would be possible, resulting in different costs. Now, for the economy involving all these entities, the optimal choice of measures would be, economics argues, to carry on with those measures, and only with those measures, which are cheapest. Emission trading means this: an entity can buy reduced (saved) emissions from another entity. Say, entity B buys five hypertonnes of emissions reduced by entity A. On the balance sheet, then, entity B can subtract these five hypertonnes while entity A received, instead, financial compensation. A simple deal. Of course, this would only happen, if entity A can produce these – let’s call them – negative emissions cheaper than entity B. Similar deals could happen with other entities. Eventually, entity B can declare having zero emissions, being carbon neutral. Figure 1.1 illustrates this calculation. Entity B has, say, 50 hypertonnes (h. t.) emissions. If it buys negative emissions from all the other entities, entity B’s emission sum is zero. Table 1.1: An illustration of a balance sheet with subtracted emission reductions Balance Sheet: Emissions 2006 from Entity B Source Emissions (in hypertonnes, h. t.) Entity B: total emissions 50 (prior to offsetting) Emission reductions bought from: Entity Entity Entity Entity Entity Entity B: sum of emissions A C D E F 5 10 5 10 20 0 (after offsetting) Such trade can take place between existing polluting entities. And it can take place between existing polluting entities in the global North and not-yet-polluting entities in the global South. The latter is usually referred to as Clean Development Mechanism (CDM). Project developers in the global South can develop entities which have the only 6 1.1. Greening Carbon purpose of producing negative emissions which, thence, can be sold. These negative emissions are called Certified Emission Reduction (CER) if they are produced under the allegedly heavily government supervised scheme CDM. In parallel, a voluntary market is developed, the Voluntary Carbon Market (VCM). On that market negative emissions could be traded voluntarily, according to a variety of idiosyncratic standards; these negative emissions are called Voluntary Emission Reduction (VER). These markets are also called carbon markets. And they do not work well (Gilbertson and Reyes 2009) – yet, as some hope (Lederer 2012); whereas others question the foundations of these mechanisms (Lohmann 2005). Not only it is argued that the occupation with market construction drifts attention away from dealing with climate change (Lohmann 2009a, 27) but making climate change dependent on market mechanisms also involves the risk of what Friends of the Earth author Chan (2009) identified as the risk of a carbon finance bubble, subprime carbon. According to World Bank (2012), carbon markets are rapidly evolving; a high-volume market with corresponding systemic opportunities, framed by governments as problems, like ‘framework loop-holes and criminal activities’, fraud and cyber-thefts (World Bank 2011, 10). Climate change, thus, is related to not only through the rules set by governments but very much through what is going on in and around market participants (Bernstein, Betsill, Hoffmann, and Paterson 2010). Consequently, this book is based on the understanding that to make sense of how capitalism performs in governing climate change we need to actually study how carbon counts and numbers are brought about. It is these inner worlds of governance which matter, not just big talk. Then, this study should not only be of interest to those who invest their trust in these mechanisms to work; it may also be relevant for those who do not trust governmental regulations and market participants but locate the source for the required reactions to climate change in ‘autonomous movements self-organising from below’, workers who may organise the greening of production processes themselves (Abramsky 2009, 6). Studies of carbon governance and the CDM trace these instruments back to the discourse of ecological modernisation (Ninan 2011; Bailey, Gouldson, and Newell 2011). While it suffices to revisit the discursive link between climate change and GFQ in Chapter 3, as the next step we need to narrow down what the discourse of ecological modernisation entails and how it frames carbon emissions. For now, the take-home point is this: accountants establish the link between an environment (in our case carbon) and the market. Ecological modernisation and corporate greening Originally, the notion of ecological modernisation has been developed in Western European Countries as prescription of how capitalism might be reconcilable with an ecological or environmental rationality in the 1980s. Huber (1988), as one of the first authors to develop this notion, argues in favour of a new social-ecological order which would be characterised, inter alia, by ecologically modernising production and consumption. This should be achieved by employing ‘new and more intelligent technologies’ such as Information Technology (IT) 7 1. Introduction in industry, office and home, improved energy and environmental technologies, solar technologies, genetic technologies, biotechnologies and their applications in agriculture (ibid., 174). Of that, the ecological modernisation of industry was the most important; he argues for integrating environmental protection into production processes rather than cleaning up afterwards. When, a couple of years later, Hajer (1995) published a discourse analysis of ecological modernisation’s move into the policy world, he identifies several of Huber’s (1988) postulations as having found their way into dominant discourse. In Hajer’s (1995, 26-29) account, six conceptual shifts to ecological modernisation can be recognised: From reaction to prevention Bureaucratic-judicial administration is criticised for being focused on ‘react-and-cure’. The innovative solution against the former approach was ‘anticipate-and-prevent’. In his reading, this lead to two consequences: integrated environmental protection became the new norm and, together with de-regulation policy, it lead to a discourse favouring economic incentives and market instruments for environmental protection. A new science While he identifies earlier models of science with the identification of environmental problems, in ecological modernisation discourse science moves towards the centre of policy-making: scientists got the task to predict how much human load the planet could carry (the 2◦ C aim is a perfect instance). Micro-economics Earlier models of economic behaviour of individual entities stressed that measures to protect against environmental damage implied ‘costs’ for this very entity. A new version of micro-economics argued that ‘pollution prevention pays’, i. e. that investment into measures would prevent damages in the future which the entity might have to compensate. Macro-economics He also observes a shift in public households’ and nation-states’ economics. He sets a conception of nature as a free resource and sink for pollution against a new understanding of nature as a public good. If nature is a public good which, in principle, ought to be or is available for ‘the public’ then the state has to ensure that economic actors do not externalise costs, e. g. using nature as a sink without compensating ‘the public’. The positive account goes like this: actors have to internalise the costs into the actor’s economic optimisation. Legislation changes Legal norms are also adjusted. In ecological modernisation discourses it is not anymore the damaged party who has to carry the burden of proof but the damaging party. Participation Whereas for several years environmental groups staged practical opposition and engaged in direct action against environmentally destructive projects, with ecological modernisation a selection process takes place. So-called civil society (the civil components of it, i. e. those parts of the social movements that regrouped as NGOs) are invited into policy-making as stakeholders.3 3 8 For the exposition of this last point I also draw on Bergstedt (2005). 1.1. Greening Carbon Ecological modernisation is still the dominant discourse in environmental policy, among businesses and NGOs. The decisive question, which I am not alone in addressing, can be posed in this way: is capitalism able to innovate itself? And, if this discourse informs actual innovation processes: what are the consequences of such innovations? Seeking to contribute to such questions, this book shows how ecological modernisation takes place in work practice. Crystallising in relation to the notion of ecological modernisation, an own school of thought has emerged: Ecological Modernisation Theory (EMT). This theory interprets ecological modernisation as primarily promising to reconcile capitalism with ecology. To test whether this promise holds, scholars study how the promises are reflected in policies, in corporations’ accounts or in economic data. Proponents of EMT suggest that, indeed, ecological modernisation is taking place and that there is reason to expect that nation-states and corporations can become green in this way: deploying more, rather than less, market-driven science, technology and expertise (Buttel 2000). To make sense of EMT, four issues seem relevant. (1) EMT positions itself firmly against other environmental sociologies which claim that industrial or capitalist societies create detrimental environmental problems; thus this theory takes an optimistic vantage point proposing that sustainable development is possible and that industrial or capitalist societies are approaching this utopia (Mol 2001).4 (2) There is not yet enough of ecological modernisation. Scholars argue that more eco-technical and eco-efficient innovations are needed (Huber 2008); these are understood as synonymous with ecological modernisation (Jänicke 2008, 558). Of course, these innovations have to spread. Huber wants to see a global diffusion of ecological modernisation. Jänicke (2008, 563) accepts that, unfortunately, the state has to intervene by means of ecological politics and regulation which is to render restructuring of industry as well as economy both ‘socially and economically acceptable’. (3) EMT scholars point to two findings seemingly affirming their theory that ecological modernisation is taking place. Macro-economically, they identify a decoupling of economic growth from resource utilisation in ‘in most of the ecologically advanced nations’ (Mol 2010, 23). Culturally and institutionally, they (re)cognise an economic and an ecological sphere becoming ‘autonomous’ from each other, allowing business to consider both spheres – semi-permanently and in parallel (Mol 2001). As Mol puts is: The basic idea of ecological modernisation is that, at the end of the second millennium, modern societies witness a centripetal movement of ecological interests, ideas, and considerations in their institutional design. This development 4 Mol (2001) and other proponents especially struggled with those approaches which have leanings to (neo)-marxisms. These suggest especially two points: that capitalism could not be reconciled simultaneously with environmental and social justice; and they stressed that ecological modernisation cannot optimise capitalism at a scale large enough in order to cancel out the catastrophic effects of capitalist dynamics (Christoff 1996; Pepper 1998; Pellow, Schnaiberg, and Weinberg 2000; Blühdorn and Welsh 2007). Beyond statist approaches, except of Bookchin’s (1993) work, not much academic anarchist thought has informed environmental sociologies (although anarchist direct action like other environmental social movement activities have repeatedly been studied). For a more detailed account of these debates, see e. g. Lippert (2010a) (as a critic of EMT) or Mol (2010) (as a proponent). 9 1. Introduction crystallises in a constant ecological restructuring of modernity. Ecological restructuring refers to the ecology-inspired and environment-induced processes of transformation and reform in the central institutions of modern society. (2010, 23) They treat as evidence for such changes e. g. the widespread emergence of environmental management systems in companies (ibid., 24). Such systems play a key role in this book – and I will get to them in a moment. Before, note a last general point on EMT, valid also for these corporate systems. (4) Ecological modernisation is very much about knowledge. Jänicke (2008, 558) defines modernisation in economic terms as ‘the systematic, knowledge-based improvement of production processes and products’. This entails also a risk: if ecological modernisation is about knowing and changing accordingly, it is also possible that actors decide not to learn, i. e. to ignore (ibid., 562). To ecologically modernise an institution, following various EMT accounts, ‘the environment is reduced to a series of concerns about resource inputs, waste and pollutant emissions’ (Christoff 1996, 485). Knowing environments in these ways is a prerequisite to optimise processes. Studies of how exactly environments are known by those people who are to put ecological modernisation into practice, however, are missing. We may use EMT to show how social sciences allowed themselves to talk about evident changes in how seemingly modern institutions related to environments: they used data representing environments and corresponding data-generative mechanisms (like environmental management systems) to speak for these environments – which constitutes a realist take. With this take, social sciences also limit themselves. This book seeks to employ an alternative epistemological take on environmental data and representation devices to engage with ecological modernisation’s knowledges (and carbon is an example). Vis-à-vis EMT, this book questions their findings of decoupling and the centripetal movement of ecology into institutional design and suggests as an alternative a material-semiotic take (introduced below). The next step is to look more closely at corporate environmental protection, business’s ecological modernisation take which is mostly called (corporate) environmental management.5 Even critics like Levy (1997, 131) recognise that environmental management ‘is much more than ideological greenwashing that masks production and pollution as usual’. He finds that corporations can and do employ approaches like total quality management which allow them to ‘profit from cutting resource use or disposal costs’ (ibid., 132). Beyond profit, Hoffman and Bazerman (2005) argue, the for-profit sector has to act because companies have the knowledge about their internal processes and the environmental impacts they produce – and they have the power to change them. In this logics, where governments do not have the capacity to steer firms, firms have to steer themselves. 5 The notion environmental management is widely used, beyond corporations. While this book emerged in reaction to my dissatisfaction with the unreflective and unreflexive stances taught in an environmental management degree programme (resulting in a research call to study environmental management in practices (Lippert 2010b)), here I focus on corporate environmental management. 10 1.1. Greening Carbon Environmental management systems constitute the reaction to governments asking companies to ecologically modernise themselves (or were it companies which originally pressured governments to stop interfering in firms internal processes?). An EMS is supposed to satisfy both: interests in profit and in documenting vis-à-vis regulators that the company is ‘indeed’ ecologically modernising itself. It, thus, combines issues of capitalism, ecology and accountability which makes it an archetypical instrument of ecological modernisation (Buttel 2000, 58). How an EMS works in practice is a different question, of course. In theory, an EMS is a standardised procedural package which an organisation may, voluntarily, adopt. It is advertised to managers as containing ‘tried and tested principles’ of corporate greening, reports Fineman (2001, 21); he finds it presented as ‘forward looking, a “challenge”, “strategic”’, rendering greening refulgent. If we open the box of promises we find more specific ideas of how it supposedly works. In Chapter 3 I show in detail how GFQ related to the idea of an EMS. For now, it suffices to point to a fundamental design characteristic shared by various prescriptions of how an EMS is to be implemented. This design characteristics is a circular structure representing an imagined causal and temporal loop.6 In an EMS’s loop, you would commonly find this order of elements: setting an environmental policy, planning how to implement it (devise measures), implementing measures, reviewing the outcome (US Environmental Protection Agency 2012). This order is standardised, for example, in a norm set by the International Standardisation Organisation (ISO) (14.000 series).7 Relevant is that this cycle of steps is repeatedly gone through. That is, once the organisation has reviewed the outcome, it would revisit and potentially update its policy, set new commitments and the cycle starts anew. This promises ‘continuous improvement’ (ibid.). Now, those environmentalists who would try to involve themselves in a corporation’s environmental conduct would usually either attend to the corporation’s environmental policy (in Chapter 4 I present an analysis of the reality of such attention) or they would turn to the actual measures taken by the company. For this book employs the climate change and carbon reality, I illustrate measures in terms of saving carbon emissions. The specific techniques to reduce carbon emissions would – in principle – be subject to the economic technique called cost benefit analysis, comparing various options to achieve the target: from (1) in-house changes to heating requirements (e. g. introducing better insulation) or (2) the substitution of carbon emission intensive coal-fired power plant generated energy by less carbon emission intensive and cheap nuclear power based electricity or even carbon neutral (more expensive) hydropower based electricity to (3) buying CERs. Thus, the company would select among various available measures to go greener – the most cost-efficient. 6 Imagining work processes as designable loops is a wide-spread approach in business process modelling. For an illustration, see, e. g. Denning and Medina-Mora (1995); imagining humans as programmable in such ways has been criticised in-depth by Suchman (2007). 7 In parallel to this norm, many other standards of how an EMS is supposed to work exist. This includes, inter alia, the European Union (EU)’s Environmental Management and Audit Scheme (EMAS) as well as several national standards (see e. g. Kollman and Prakash 2002). 11 1. Introduction So far for the shiny reality of EMS promises. We are approaching, though in small steps, the core concerns of this book. Reviewing the outcome of any measure which actually interferes with the business process requires actual measurements of the effects of the measure on the corporation’s environmental performance. For the two latter kinds of measures (2 and 3), such measurements would not be needed for they could be represented easily on paper as indicated in Figure 1.1 (on page 6) (one may simply subtract saved emissions from the sum of emissions). Yet, wait a moment. The row stating ‘Total emissions: 50 hypertonnes’ has been the base for the calculation; but where did it come from? This book revolves around a fundamental presupposition of ecological modernisation, of any EMS: to manage environmental goods and bads rationally, to ecologically modernise operations, one requires knowledge of the state of environmental impacts. This is a classic recognition which is of concern to environmental managers. An EMS requires what they would call environmental data, an ecological information system, to inform management (Richardson 1994; Birke and Schwarz 1994). In a management school approach to environmental management, such as implied by Schaltegger and Synnestvedt (2002), environmental data constitutes environmental performance data; and environmental performance may well be related (some hope, others fear; some agree, others disagree) to economic performance. To ‘gather’ such ‘data’ an organisation would employ some form of environmental accounting. More on this later. We now turn to problems corporate greening scholars have noted. In a discussion with the field of industrial ecology (a compatible ecological modernist approach), Hoffman (2003) notes that in corporate environmental and resource management only natural and material flows are recognised. He criticises that values and social factors, more generally, are omitted. Even though he mentions that issues like ‘waste’ are both, social and technical, fundamental to his conception of corporate environmental reality is its existence in two separate systems, the social and the technical, normally clearly distinct. In this understanding, entities like any material or energy are non-social, non-cultural. Waste is the exception to the rule. The social, cultural, in his account, consists of values, beliefs, etc. They are aspects. Without an appreciation of these aspects, industrial ecology perpetuates an engineering mind-set that relies on input-output data without appreciating the social and political aspects of both how those data are developed and, more important, how they are interpreted and acted upon. (ibid., 82) His take exemplifies the ecological modernist realism and ontology. Reality consists of the natural and the social. They are clearly distinct. Well, they are not. The social gets constantly back into the materials under management. This is what EMS scholars discuss when attending to so-called cultural and national variation in how environmental management systems are implemented and how they work. Baumast (2003), e. g., finds that environmental management is subject to cultural influences or Branzei, Vertinsky, Takahashi, and Zhang (2001) identify culture as a contextual 12 1.1. Greening Carbon factor, which shapes values, serves as a ‘context’ guiding action and endorses ideologies. Though some of this literature recognises that environmental culture transcend national boundaries, it is always located external – off materials under management and, ideally, off the manager. Culture as an impact – it interferes. Consequently, most sociologists will not be surprised, empirical studies find that neither environmental management nor restructuring business processes mirror the celebrated circles – and that they cannot. A study by Birke and Schwarz (1997), for example, reports that while an EMS can be easily sketched on paper, its implementation is a different matter altogether: environmental management is not determined by economic, technical or organisational prescriptions; much rather, environmental management is always confronted with multiple decision and development alternatives which are contested and contingent, ‘conflict-laden, containing politics and, therefore, open-ended’ (ibid., 195). Similar, Howard-Grenville (2006, 48) reports that corporate reactions to environmental issues depend on internal cultural interpretations, finding ‘that multiple interpretations of a given issue can coexist in a given organisation’. As significant as these reports on multiple perspectives are, we need to go further. In relation to these stances, this book argues that it is a limiting approach to understand environmental management only in terms of interpreting reality. Yet, before we move on to extend this discussion of and beyond interpretation, we have to engage with those who position themselves against interpretation. In the received view, it is rare that divergent interpretations are accepted – let alone made recognisable. It is hoped, as the story goes, at least if we have all the facts on the table, then we can add our cultural beliefs, value-systems or ideologies; i. e. subsequent to fact-finding we can contest the meanings of data and the meanings of different measures. That is, many do not like this unfortunate narrative of a reality which is interpreted – even if they recognise that, if it gets down to discussing what to do about certain information, they have to negotiate. They would stress again and again that an EMS is a standardised instrument, a meta-instrument; that is, they would stress that as part of an EMS one has to ensure that all the legal requirements, standards and other prescriptions are properly reacted and adapted to. After all, it is one of the ‘basic elements of an EMS’ that it ensures ‘compl[iance] with legal requirements’ (US Environmental Protection Agency 2012). And these standards put companies under pressure. Hoffman and Bazerman (2005) argue exactly this and suggest that standards allow governmental agencies to exercise oversight over companies. Governmental standards ‘dictate’, in their view, which corporate materials and wastes a firm has to control, how and how much. This take does not fit well with the interpretative findings noted above. Either corporate environmental management is determined by standards or its managers have same leeway in interpreting and acting on the world. Now, neither business nor governments would want firms to appear as interpreting standards. Ecological modernisation provides a solution to this puzzle: allowing organisations to govern themselves. This is exactly what an EMS is about. With 13 1. Introduction an EMS, an organisation can set itself certain policies being ‘[i]n principle [. . . ] free to choose its own procedure in order to achieve the given ends: a way of governing through self-government’ (Asdal 2011, 4).8 In a regime that allows for environmental management systems adopting an EMS is a voluntary act; although an organisation may experience economic effects if they do not adopt (such as more frequent governmental inspections, higher insurance premiums). The voluntary characteristic of environmental standards can be understood as laid out in ecological modernisation’s emphasis on corporations to take voluntary preventive action. In this discourse it is often voiced that an organisation can gain favourite public attention by aligning itself voluntarily to a standard (assumably resulting in higher standard practices than without claiming to adopt the standard) (Lippert 2010c). Power (1996, 301-302) proposes to read such standards as splitting form and substance. In an EMS the substance (i. e. specific environmental conduct, impacts or material and waste flows) are not anymore directly governed; rather, through the EMS’s form the organisation’s approach to govern the substance is circumscribed. Drawing on Shayler, Welford, and Shayler (1994), he points to the ‘unintended consequence of this division of labour [which] is that environmental performance has come to be closely identified with having an (auditable) system’ (Power 1996, 302). In this rationality it follows that environmental performance does not need to be inspected on the ground but it suffices to check whether an EMS’s policy is compliant with external regulations and whether some evidence can be found that the EMS is properly put into practice. This, of course, involves a further dimension of politics and interpretation: who would be allowed to audit environmental management systems, scientists or accountants; which and whose expertise would be recognised as apt to lend legitimacy to a firm’s EMS (Power 1997)? Of course lots of critique on corporate greening exists. Rikhardson and Welford (1997, 62), for example, call out: ‘[s]ooner or later the eco-modernist approach will be exposed as the “greenwash” what it is.’ A new institutionalist take sees environmental management systems as a rational myth (Meyer and Rowan 1991), indicating that we can only expect myths and ceremonies from companies, mere green talk, rather than seeking real change (Boiral 2007). In this account, corporate greening statements are only loosely coupled with real activities. Even while, e. g., certified ISO 14.001 environmental management systems may ‘project an image of rigour, objectivity, precision, and control borrowed from the exact sciences’ (ibid., 130) it is clear that the work reality in a corporation may develop on a different trajectory than those of an idealised science. In Boiral’s study, labels signifying that an EMS was run according to international standards like ISO 14.001 were seen, ‘both within and outside organisations, as a formal acknowledgment of good practices’ (ibid., 137). He found that business process were already imagined as good (enough); so they were not changed very much under the assumably strict and ‘consistent control’ (US Environmental Protection Agency 2012) of 8 Social scientists may see the resonance of this statement with Foucault’s (1991) notion of governmentality; a neoliberal version of governing subjects’ and organsiations’ conduct (Gordon 1991; Lemke 2001). This theme, however, is one which this book does not cover. 14 1.1. Greening Carbon an EMS. My study was aware of such critiques; however, rather than engaging in such a, seemingly, evaluative discussion, I studied the nitty-gritty of day-to-day knowledge practices among EMS agents. This interest in the day-to-day knowledge practices, in how these agents managed to sustain an EMS, was supported by the methodological considerations reported in some of the studies mentioned above. Burschel (1997, 309-310) concludes that structural takes to study workers were not sufficient to understand them. They call for methodologies which are apt to study agents’ social processes.9 Hoffman (2001) recognises the limits of survey-based research.10 A large scale study, conducted by Gunningham, Kagan, and Thornton (2003), uses a mixed-method approach, involving qualitative interviews. Boiral’s (2002) ethnography raises the question of how the requirements on knowledge for managing environments should be conceptualised. He points to tacit knowledge which would not easily fit organisations’ preference for clearly definable information. Howard-Grenville (2006), based on an ethnography, identifies different subcultures within an organisation which configures the social reality in which environmental issues are interpreted. And Hargreaves (2008) shifts the methodological apparatus towards practice theory, which I discuss below, employing ethnographic sensibilities to recognise not only symbolic meanings but also procedures, skills, materials and technologies relevant in co-constituting agents’ environmental practices. The approach labelled practice theory has also now been recognised in the centre of EMT. Spaargaren (2011) proposes to employ practice theory to study consumption practices. According to him this allows to conceptualise and investigate how ecologically modernised practices are. He identifies them as ecologically modernised if they incorporate objects, meanings – e. g. the ‘ways of doing and saying’ – which are important for monitoring, assessing, valuing, and improving the practice with respect to its environmental or climate performance (ibid., 816). In that respect he clearly opens studies of ecological modernisation to practices and the objects which are part of them. However, the same concern which I noted above still holds: environmental performance and data is conceptionalised in a realist way – as if that data was waiting for the ecological modernist to gather it. This review of ecological modernisation and corporate greening suggests that the concepts employed by these discourses (such as EMS, control, standards, implementation and, above all, knowledge, data and information) should be the object of analysis, rather than analytical categories, taken-for-granted and employed. In other words, this book claims that ecological modernisation should not be used analytically to understand corporate response to environmental and sustainable development crises or to climate 9 While Asplen (2008) misses to observe managers, I read her, too, as calling for studies of practices. While Hoffman (2001) imagines ethnographic in-depth studies which would help to fill his variables with meanings, my study accepted the possibility that, indeed, the concepts through which corporate environmental management was thought, i. e. his variables, might have to be reconsidered. 10 15 1. Introduction change. Rather, ecological modernisation is a discourse which needs to be studied in ethnographic detail. This book does so. Carbon In the following I link carbon economics to corporate carbon accounting and to a first conceptualisation of what carbon might be. In environmental and ecological economics a fundamental idea is that market pressure selects in those firms which can perform better, economically and ecologically (especially if both dimensions causally relate). In other words, those companies which are neither green nor profitable would cease existing; and economists see this as better for society than sustaining an economically unsustainable corporation. Now, unfortunately, market pressure does not always act fast enough. Polluting companies are still on the market, even the relatively more ecologically destructive ones in a given sector (this is not to claim that some firms have not been selected out). In economics, the failure of markets to allocate environmental goods and services (as economics would see it) is lamented. Unfortunately, the market did not perform well enough. Fortunately, repair is possible: committed consumers and investors can switch their consumption and investment choices to favour those products and producers which perform greener, or in any other way more ethically to their like. The corresponding discourse refers to these choices as Socially Responsible Investment (SRI). This book addresses SRI and its prospects to alleviate market failure especially in Chapter 4. Another take to repair the market is the governmental construction of new markets. Carbon markets are an example for this (Callon 2009). And, as alluded to above, carbon markets do not work perfectly. Relating to carbon markets in this book is not so much necessarily meaningful in itself. However, engaging with how scholarly debate attends to the commodification and trading of environmental entities is useful to learn from – to inform future debates on the next version of environmental good or service. In 2012, Mol publishes a new take on carbon markets, aiming to go beyond EMT. His account is illustrative for a more widespread lack of attention in seemingly critical takes on carbon economics. He engages with all kinds of what I see as negative carbon entities, ‘carbon credits and reductions’, and finds them to be ‘increasingly abstracted and detached from on-the-ground time-place settings’ (ibid., 17). And, of course, he is very right about this. He relates this abstracted and detached character to the deplored reality that the abstract carbon markets increasingly become subject to and partly dominated by instruments, practices and products of creative investors, banks, traders, brokers, and speculators who see these GHG emission rights and offsets just as financial products, as a means of profit making. (ibid., 18) Similar to World Bank (2011), he finds that threats to the ‘proper’ functioning of markets exist. For example, he recognises that carbon markets may be subject to accounting scandals (Mol 2012, 20). However, even against this background he manages to sustain the optimism cultivated and exercised in EMT. He hopes that introducing 16 1.1. Greening Carbon distinguishing indicators in the market, as (if) classifying some carbon products as ‘gold standard’ – i. e. high quality, particularly green carbon products – will help to manage carbon economy; for him the aim is ‘to ensure that climate change mitigation rationalities remain the dominant logic in these new institutions’ (ibid., 23). And he is not alone with such a hope. MacKenzie (2007) also sees a market (segment) emerging in which market participants are willing to pay higher prices for particularly green project credits. They reference their hopes to institutions like the ‘Gold Standard’, an organisation providing a voluntary standard for projects generating negative carbon emissions, i. e. voluntary carbon offsetting projects. GFQ, was only to buy carbon credits which were certified through the Gold Standard to ensure participating only in this very green carbon market segment. This book relates to these promises in Chapter 6. In economic terms, these negative emissions, whatever their quality is, constitute supply. CERs and VERs are offered to customers, say a company, to offset emissions. Social sciences, hitherto, have underexplored these emissions. Economically, they are a significant factor in determining demand. A company will only buy so much negative carbon as they can actually add to their balance sheet to become carbon neutral. Whereas EMT scholars, like Mol (2012), and poststructuralist critics of carbon markets, e. g. Descheneau (2012), miss to ask where emissions come from, businesses recognise that they have to take a stance on this issue, that is, on the ground of their emission offsetting activities. For example, The Climate Group (2008, 2), a finance industry sector group states as one of the principles to which signatories have to commit: We have measured a significant proportion of our operational GHG emissions using an internationally recognised or equivalent domestic standard and we disclose this information. (Principle 1.3) Carbon accounting is consequential. Measuring emissions determines how much emissions an organisation ‘actually’ had. However, finding The Climate Group relating to this issue, indicates that these measurements are not self-evident. As this quote shows, obviously, different standards to measure how much emissions their business operations caused exist. MacKenzie (2007) relates to this problem: Measurement and independent verification, the foundations of any emission market, are getting better. There were a lot of difficulties in the first year of the [EU Emission Trading System (EU ETS)], simply as a result of companies’ unfamiliarity with what they had to do, but I’m told that the 2006 measurements, currently being collected and aggregated, are better in that respect. He also finds ‘there is some room for “gaming”’, for example by deliberately underestimating emissions or using specific standards. He is optimistic though. ‘Overall, though, such problems appear tractable.’ However, he also recognises that how accounting works in practice is an empirical question. Correspondingly we find him calling for ethnographic studies of accounting (MacKenzie 2009b, 130) and Lohmann (2009b, 529) 17 1. Introduction wants us to engage with the limits built into carbon accounting. Indeed, we need radical research on environmental accounting, say Gray and Laughlin (2012). Here we go. To start with, let me offer an indicative definition – one of the few ones in this book – on accounting. A shiny book for corporate environmental managers by Ditz, Ranganathan, and Banks (1995, 4) defines accounting as the systematic collection, organisation, and communication of information on an organisation’s activities. This gives us some orientation in what we are looking for: agents who do something; agents who (1) collect, (2) organise and (3) communicate – in our case – (4) carbon information. This book attends to the practical reality of all these four points. Above, in Figure 1.1 (on page 6), I sketched a segment of a means of accountants to communicate, a lower part of an environmental balance sheetZ, i. e. the sum of emissions produced in an organisation and the carbon offsets.11 In this book we engage with the upper part of such a balance sheet and, thus, answer these question: where do emissions come from; how do they get from there to here; and what happens on this way? MacKenzie (2009a, 447) declares that accounting is key to make any economic item, like carbon emissions, visible. If accounting is so fundamental we can expect environmental management scholars to comment on how accounting takes place. Indeed we can find their accounts on this. Hoffman (2010), for example, engages within the same article with the problem that metrics to measure performances are problematic (ibid., 298) and postulates on the succeeding page that a company should quantify or measure (he uses these terms synonymously) GHG emissions in order to understand what climate change means for the corporation. Engaging with this – often cited – scholar of environmental management reveals a differentiated understanding of environmental accounting. In 2005 he affirms the point that a company has to establish its GHG balance. But, he points out, it can influence how it measures its emissions; ‘multiple methods’ are possible. For example, emissions can be measured directly (laboratory based) or estimated (using fuel or material based calculations); but the latter is complicated by a variety of available methodologies (2007, 10). We may read him as indicative of a recognition among environmental management scholars that accounting for environmental management is not a simplistic issue.12 Burritt, Schaltegger, and Zvezdov (2011) published results of a study which attracted my attention for two reasons: I know one of the authors and am familiar with their research methods and the study sought to scrutinise how carbon accounting works in corporations’ practice. Going beyond old-school survey research, they employed semistructured interviews carried out face-to-face or by telephone with environmental 11 I use glossaries to briefly describe the meanings of acronyms as well as members’ terms. Key concepts are introduced with this sign: Z. See the appendix for brief explanations of such terms. 12 Correspondingly, a wide variety of literature on the topic exists. Schaltegger, Gibassier, and Zvezdov (2011) and Stechemesser and Guenther (2012) provide a bibliometric literature review on environmental management accounting as seen from within their disciplinary affiliation in the management realm. 18 1.1. Greening Carbon accountants and users of carbon information in several of the largest German companies.13 This approach is of interest to this book because my study was based on a similar high-profile multinational as they have studied and we were interested in similar agents: those corporate environmental managers and accountants who were processing carbon data. Five notions reoccurring in their study are noteworthy. Control First, they imagine that carbon accounting is well under control. According to them, designs of accounting systems are determined by ‘managers seeking information for their decisions’ (ibid., 83; referring to Burritt, Hahn, and Schaltegger 2002). This implies the design of the system is actually expected to be helpful to guide managers’ decisions. In addition we may interpret them as implying that the accounting system could be shaped and controlled by all those legitimate users who want actually information. Physical information as available Second, they claim that ‘physical information is more often than not available and kept for bookkeeping purposes and is thus available even if no environmental or carbon management department is officially responsible for the management of such information’ (ibid., 86, added emphasises). Thus, their report claims that information about material operations of companies is normally existing, waiting for users to be taken into account. Physical information as necessary Third, they argue that precisely such information inevitably has to be collected ‘so that the economic linkages and effects are made explicit and thus manageable’ (ibid., 87). This normative statement implies that management is not possible without knowing what it seeks to manage, i. e. without knowing carbon emissions. Data collection Fourth, throughout the paper they employ a vocabulary suggesting that data or information (two concepts used synonymously by them) can be collected or gathered. I read this vocabulary as indicative of a simple realist ontology, consistent with EMT. Data generation Finally, however, at two points of their article, they write that data is generated (ibid., 84) and produced (ibid., 91), respectively. The language of generating and producing data points to a constructivist understanding of data; data is not out there waiting for users but it requires some form of agent who does something which results in data. While the first four points seem theoretically consistent, the last stands apart. A question poses, seemingly, itself: is data ‘out there’, waiting for its users, or has it somehow to be achieved? Where do carbon emissions come from? Till now I used the notions CO2 , carbon, carbon emissions and greenhouse gases synonymously. And I will continue to do so. For a reason. In the fieldwork I undertook on carbon accounting, these notions were also used simultaneously. Who am I to imprint their world with another, a seemingly better or more legitimate way of 13 I admit, I feel challenged to engage more with their study; but this has to be postponed. 19 1. Introduction drawing distinctions between these notions? If carbon accountants use these concepts interchangeably, we should take their day-to-day language use seriously and study what kind of entity these notions refer to. If explicitly discussed, however, they would point to the formally correct statement. In their discourse, a ‘correct’ expression exists; when asked about these notions, carbon accountants would say they signify this expression when using any of the former notions: it is equivalent carbon dioxide emissions, abbreviated as CO2 e. The notion of equivalence points to several gases or groups of gases which are legally recognised in the Kyoto Protocol as contributing to globally warming the greenhouse (the earth’s sphere); they are collectively addressed as greenhouse gases. How equivalent these gases are is defined by referring to the different gases’ global warming potentials (GWP). These potentials are subject to change, depending on shifting scientific consensuses and on so-called politicians or other agents to refer to scientists’ updated versions of GWP tables (Lippert 2012b). Thus, through the construction of GWP, different gases are made commensurable, which MacKenzie (2009a) describes as making things the same. To state this clearly, this book’s concept of carbon is not fixed. Rather than imprinting carbon accountants’ work reality with a somehow politically-scientifically defined prescription of how carbon ought to be voiced, this book is about how carbon is being voiced by carbon accountants. Thus, I am not so much here interested in a chemical, physical or climatological discussion of carbon – or attending to the carbon molecule’s exciting extraterrestrial occurrences, engaging with diamonds, etc. This book is about the social form of carbon, the material way of how it is ‘collected’. The carbon accountants I studied did not use laboratories to discover their corporation’s carbon emissions.14 Rather, these workers used spreadsheets, calculators, databases, team-meetings to establish carbon emissions. For all practical purposes, for a company’s environmental accountants carbon emissions are data, rather than some kind of molecule. MacKenzie (2007) establishes a first glance at this reality. In sulphur trading in the US, each smokestack is fitted with automatic measurement devices. European carbon dioxide emissions are measured less directly, using the method known as ‘mass balance’, in which gas-meter readings or invoiced quantities of coal or oil, for example, are multiplied by appropriate emission and oxidation factors. I concur. But how do mathematical operations on ‘gas-meter readings or invoiced quantities’ turn into carbon emissions? Where is carbon? Rather than searching for molecules because natural scientists tell us that greenhouse gases are first of all natural entities, this book is based on searching for the ways through which carbon was made present and the materialisations of carbon in an actual company. Following Wynne (2010), scientists do not hold the singular authority to represent carbon facts but I 14 And if they had utilised laboratories, surely, this would have turned into a wonderful story of carbon laboratory life (Latour and Woolgar 1986). 20 1.2. Accessing Carbon accept any spokesperson or entity in the field which claims to represent carbon facts and engage with these representations. This book, thus, sets out from the conviction that at the intersections between sustainable development, climate change debates, EMT, corporate environmental management studies and carbon accounting we find an exciting practice of knowledge-making and materialisation: agents doing carbon. STS scholar Blok (2010) provides us with an indication of what such a study might entail in terms of carbon: he finds that carbon does not fit a simple and singular topology. Earth’s ‘global’ emissions are not out there in the atmosphere but on computer screens, at conferences, in articles. Accounting scholar Gray (2010) underlines that restricting environmental accountants to provide a singular account of how sustainable their company is will not represent well what is going on within the company. Consequentially, he wants us to accept multiple accounts of how sustainable a company may, or may not, be. A clarification: studying how carbon accountants do carbon promises empirical contributions relevant to understand how agents of ecological modernisation actually go about their work and to understand the ontology of their products, emissions. If indeed making capitalist carbon green is ‘experimental’, as MacKenzie (2009a, 453) would have it, these ‘experiments need “witnesses” [Shapin and Schaffer 1989], and those witnesses must be multiple: lay as well as professional, from many countries, and if they are academics from many disciplines’. 1.2 Accessing Carbon Between late 2008 and mid 2010 I conducted ethnographic fieldwork in the multinational company GFQ. This name is a pseudonym.15 GFQ, at that time, has been among the world’s top 50 companies; it has been positioned there for several years, sometimes among the top 25 – listed in the magazine Fortune. Hence the expression, Fortune 50. GFQ is a global financial services provider – banking, insurance and other capital techniques. Contractually, the company was engaged with tens of thousands of customers and a workforce of over 10,000. The multinational could be found in over 50 countries; this book links to accounts from, inter alia, Australia, Columbia, Germany, India, Malaysia, Russia, and the United States of America (US). This corporate group performed well – economically. During one of my fieldwork years their accumulated profits summed up to over 7,000 millions United States Dollars (USD) and their assets exceeded 1,000,000 millions USD. These numbers have been rendered anonymous sufficiently.16 15 Following Suchman and Bishop’s (2000) approach, this pseudonym serves less to obfuscate the corporation than to mark the possibility that the study has not produced results that are idiosyncratic to GFQ but, rather, are emblematic of a more general direction in relations between capitalist accounting practices and doing nature. 16 Find details on the material and data rendered anonymous in Appendix A.2 (on page 555). 21 1. Introduction Climate change and the finance industry In discussions of climate change, often biggest polluters are named (and shamed). Some discuss them in terms of nation-states (then they name the EU or US, e. g. Bachram 2004) others in terms of industries (then, one would point to oil and coal, e. g. Perrow 2010). The financial services sector is seldom named. If it is named, it is often addressed as an enabling industry; it is seen as enabling through its power to steer investments – investments into a low carbon trajectory, into sustainable development – or as enabling through its financial compensations for damages caused by climate change (e. g. Mills 2005; Kahlenborn, Dierks, Wendler, and Keitel 2010). However, the finance sector in general is also seen as especially at risk through climate change because its profits have to be reinvested in specific objects – investment, rather than keeping money for oneself, is the classical marker of capitalism (Marx 1887, Ch. 4); and these globally spread objects which attracted banks’ and insurers’ investments are at risk of climate change. Thus, financial investments at a global scale, draw climate change damages to the financial industry. Insurance is additionally burdened by climate change and other dynamics with damaging effects because they have to compensate the insured. Consequently, the financial services sector is financially interested in damage-adverse behaviour by insured parties or investment objects. Insurance logic dictates that customers should be pressured to invest into damage-adverse actions; and finance industry can threaten potential investment objects with switching their investment to less risked objects. For these pressures and threats to work, the financial services sector, especially insurance, is, arguably, interested in indicating to insured and investment objects that they themselves take environmental catastrophes and climate change serious. We could also consider less direct approaches – such as finding that financial services providers publicly signal pressure towards governments that they want strict climate change mitigation policies. Whatever the precise configuration of reasons for finance industry to fight climate change is, we find, as indicated with the above reference to The Climate Group (2008): initiatives in this sector express their concern for climate change and announce action.17 The finance sector, thus, has to go green. To go green, as the discussion on ecological modernisation showed, in managers’ rationality they need to know their business’s pollution. Now, engaging with complete life cycle assessments (LCA) of financial products is not an easy task.18 Facing this task, another polluting source is easier trackable: the emissions caused by burning fossil fuels and all the indirect emissions linked to the products and services they buy to be able to operate their offices and server farms. Thus the finance sector can also reduce its carbon footprint in an easy way. 17 For a well-accessible summary of the risks of climate change for the finance industry and some of their reactions, see e. g. Richardson (2009, 612-619). 18 An LCA is a technique claimed to establish the environmental and resource implications ‘attributable’ to a product over its total life (Rebitzer, Ekvall, Frischknecht, Hunkeler, Norris, Rydberg, Schmidt, Suh, Weidema, and Pennington 2004). Usually imagined as a scientific, and therefore, objective approach, Molloy (2000) shows how, of course, the grand aim of objectivity cannot be realised and reconstructs ‘LCA as politics by other means’ (ibid., 2). This book does only at the margins address financial products and, thus, product ecology, explicitly (but, see Chapter 3 and Chapter 4). 22 1.2. Accessing Carbon GFQ’s take on carbon GFQ was approaching greening precisely in this way: running an EMS to establish their carbon footprint and devise actions to reduce it. And when asking GFQ where its EMS was, ‘the company’ would direct one to a specific office in the HQ, to the office which we encountered already at the opening of the Introduction. (Of course, if you were a mere citizen asking about GFQ’s concern for ‘the environment’ you would, more likely, have ended up in the Public Relations (PR) Unit.) My narrative, sets out from that location to which I was immediately led when visiting GFQ the first time. This office was neighboured by further offices. People in these offices were working on a huge variety of issues. Luckily, they had at hand resources, symbolic (a common designation on GFQ’s chart of the division of labour), bodily (a common boss) and spatial (a meeting room), to imagine themselves as working together. Broadly speaking, their activities were concerned with what is usually called sustainable development, Corporate Social Responsibility (CSR) or corporate citizenship. We are in the unit of GFQ which was responsible to develop and operationalise new ways and products of how GFQ could ‘be’ a responsible actor in society while being a profitable actor in economy. Coordinating GFQ’s carbon and environmental performance was part of this field. These people were to help GFQ to turn towards the grand aim of sustainable development. In the corporation’s view, this included that GFQ had to engage with their carbon emissions (Lippert 2012a). For my fieldwork I had to be equipped with the corporation’s reasoning why they needed to worry about carbon at all. For this book, this reasoning provides a clear entry point to GFQ’s carbon emissions. It has been materialised in several ways, on paper and digitally. I share this equipment – the materialised reasoning – with you. Right on my first day in the field (that is after access had been negotiated) I have been given a copy of a presentation which people of this unit had prepared for a meeting of their board of directors. The presentation was entitled ‘CO2 Reduction Concept for [GFQ]’. Some days later, I came across another presentation, called ‘EMS for GFQ’ (EMS4GFQ). Both included the same slide, ‘Why does [GFQ] need a CO2 footprint reduction program?’ Well, they have not been exact copies. In the version for the board of directors the subheading is simply ‘Rationale’. The other version is reproduced here (and rendered anonymous) as Artefact 1.1 on the next page. I use this slide to make eight introductory points about GFQ’s relation to carbon and, simultaneously, formulate how this book attempts to engage with these specific elements of their relation. Together, these eight points provide us with an outline of the shared formal reality of GFQ’s carbon policy. That is, this reality had been shared between the authors of this slide (people in the Sustainable Development Unit), the board of directors and the audience of EMS4GFQ, – the environmental managers working in GFQ’s subsidiaries all over the world. To make this interpretation traceable, I follow the textual inscriptions on the slide – mostly at least. First, the heading’s symbol ‘CO2 ’ signifies carbon as greenhouse gases in relation to 23 1. Introduction [GFQ] [GF Q] [25] [2015] [2006] xxx [GF Q] X $ [2015] Artefact 1.1: In a Nutshell: Making the Case for Carbon Foot-printing climate change as a ‘challenge’ (as they would have it) of sustainable development. In GFQ’s reality, CO2 is a shorthand for this complex. Carbon and carbon dioxide need to be addressed by much more than what natural sciences can offer. This book argues that carbon is to be understood as a social and organisational achievement; it could, therefore, be otherwise; carbon is politics. Second, the notion of carbon footprint indicates the discursive continuity between ecological modernisation and GFQ’s take on carbon. Chapter 3 serves to establish this continuity in detail. Footprint refers to the idea that any activity requires resources and produces wastes. They are the trace of that activity on something. That something is, usually, imagined as Earth, Nature.19 Third, the subheading (‘rationale’/‘in a nutshell’) announces the subsequent presentation as informing the audience efficiently (that is most effective given the limited amount of space on that slide) about what constitutes the need for GFQ to have a carbon emission ‘reduction program’. 19 The concept of footprint is closely related to the concept carrying capacity. To morally or politically judge ones footprint against another’s footprint, a common metric has to be available. Earth’s carrying capacity is imagined as apt to be employed for such a commensuration. A much cited entry point for this approach is an article by Rees (1992). 24 1.2. Accessing Carbon Fourth, we are provided with a single sentence which establishes the answer to the question (of the heading); we learn how GFQ links to carbon. This link is specified in several dimensions. The link exists because of a past activity of the company. It issued a statement. To whom we are not told on this slide. The slide can be read as implying that having made a statement requires GFQ to act accordingly. In fact, one of the audiences of that prior statement may very well have been the NGO that I call Global Greening of Capitalism Association (GGCA). GGCA and GFQ enjoyed a partnership agreement (see Chapter 4). I argue that the NGO was used as a legitimising agent in several ways. With respect to this slide, GGCA can be assumed to have made very explicit vis-à-vis GFQ that they expected the corporation to stick to their statement. The imagined presence of audiences legitimises the postulation that the corporation should adhere to its prior statements. So far for the reality of an announcement: it matters. How did it matter? This statement was of a reflexive nature: it referred to itself as voluntary. The slide stresses with this notion that its link to carbon emissions is contingent on GFQ wanting to exercise this link. A technical reading would suggest that the heading’s ‘need [. . . ] reduction’ contradicts this statement’s ‘voluntary reduction’. But this book is not about evaluating the technical coherency of GFQ’s documents or GFQ’s worker’s statements. Much rather, I propose to use the friction between both concepts as a generative force. Friction is a term used by Tsing (2005) to point to characteristics within capitalism in which the presence of conflicts allow capital to work – while at the same time hope for alternative realities remain. For friction implies movement and how things move cannot be exactly controlled. I argue that this friction between ‘must’ versus ‘may reduce’ is constitutive for the emissions which GFQ, eventually, emitted. It is also made clear that the responsibility which has been announced was bounded. GFQ only committed to reduce ‘its own’, rather than anybody else’s, emissions. This statement signals a limit in ownership. The slide’s authors construe themselves as reasonably concerned. For many it might seem self-evident that GFQ would only take care of its own emissions. However, critics might postulate that GFQ – one of the largest corporations on earth – takes a bit more of responsibility on board: e. g., partial responsibility for clients’ emissions that GFQ’s financial services enabled. To rule this out, the limitation needed to be enunciated. Only GFQ’s ‘own’ emissions are of relevancy for us. Chapter 3 shows precisely how these limits have been drawn. I show that the ‘own’ of GFQ’s emissions were not self-evident. We are then presented with a set of three numbers. And these concern us throughout the book. For they were significant calculatory reference points for carbon accountants. They also provide limits to the carbon reduction statement. Now the reduction aim is limited through two markings on a one-dimensional line of time and the aim is as well limited in how much they wanted to reduce. I changed the values of these identifiers in this book to ensure anonymity. Yet, in orders of magnitude these are correct numbers for the period of my field work. Nevertheless, these numbers have also been unstable over 25 1. Introduction time. Most noteworthy is that for half a year or so, GFQ had publicly committed to be carbon neutral. This marks an absence which is a significant part of this presentation. Observing absences and presences, following Mol and Law (1994), provides us with an analytical sensitivity which we may draw on to develop a fuller understanding of the phenomenon. Having provided me with this information, GFQ let me know that it did not, anymore, claim carbon neutrality as its aim. In this book we study how flexible these hard-core limits have been in practical work. Further, the statement specifies the extent to which this statement is relevant for the company. It was not merely the HQ that was to reduce emissions but the complete corporate group. This means that emission reduction was to be achieved across the complete GFQ imperium, across all its subsidiaries. These subsidiaries were distributed all over the globe and were themselves quite complicated business constructs (some of which contained and owned even further companies; this means that GFQ owned not only companies but also – folded into GFQ – further corporate groups). This book identifies a subsidiary as GFQ Corporate (Group) Entity (GCE). And GCEs could be bought and sold. For this complicated and shifting organisational nexus, emissions across the totality of GFQ were difficult to establish. I make explicit the (im)possible qualities of such facts and of practices of knowing such a conglomerate’s emissions. Finally, the statement is specified in four further directions, politically, strategiceconomically, legally and micro-economically. Political demonstration Fifth, GFQ established its carbon reduction targets in relation to audiences vis-à-vis whom they manifest that they are committed to mitigate climate change. The wording is relevant: they did not imagine that climate change could be prevented; rather, they accepted climate change as happening. And the damaging consequences of climate change had to reduced. However, their relation to mitigation was mediated via two modifiers. GFQ had committed to mitigation, rather than started mitigation. They had voiced a pledge that they want to mitigate, but whether they were contributing to mitigation was not an issue. Furthermore, carbon emission reduction was not directly related to this commitment but to perform this commitment, i. e. to show their dedication to other parties. Carbon reduction, thus, was about showing audiences something rather than taking action to mitigate climate change. Strategic economic positioning Sixth, GFQ wanted to position itself favourably in SRI ratings. These inform investors about how, inter alia, green a company performs. Dow Jones Sustainability Index (DJSI) and FTSE4Good are the labels of such ratings; ‘etc’ refers to further ratings carried out by e. g. OEKOM or the Carbon Disclosure Project (CDP). To be positioned at the top of such ratings, the authors of this slide implied, GFQ needed to reduce its emissions or, at least, have a footprint reduction programme. They specified that GFQ’s environmental performance was influencing the positioning in ratings, i. e. that being green was not negligible. To bring the promise of such a high rating closer to the audience of the 26 1.2. Accessing Carbon slide, authors added a visual element, the DJSI logo,20 which indicated that GFQ had been listed in that rating in the early 2000s (exact years deleted by me). This bullet point of the slide indicates a measurable target for GFQ. The corporation would be listed at a specific position and vis-à-vis this scale the company could measure and indicate its performance, including the performance of any carbon management by GFQ’s carbon accountants. As a shareholder company, GFQ was dependent on investments by other parties. Greening economics presupposes that ethical investors actually switch their investments to green companies. In economics, the function of such ratings is that they improve transparency within markets. Only if investors know how green a company is, they can they channel capital to green targets. This book follows the information flow into the world of ratings and challenges the belief that ratings are able to produce this transparency; I argue that markets are systematically failed through these ratings (Chapter 4). Legal compliance Seventh, a corporation like GFQ lives off the trust of publics that its operations are coherent, correct and to the norm. GFQ had inscribed into their Code of Conduct that the company was to engage in straightforward environmental protection. Governing itself through implementing its own rules is key to sustain public trust. Otherwise publics and governments might (have to) suspect that the company acts unaccountable, irresponsible, out of control; if a company does not even manage to follow its own rules, how could publics trust that the company follows publicly set norms? GFQ translated ‘protection of natural resources’ into reducing carbon emissions. The central empirical concern of this book is the monitoring of emissions. For anybody, GFQ or publics, to evaluate whether GFQ was complying with its rules knowledge of GFQ’s emissions is required. I attempt to deconstruct that such a quality of control is possible. Tangible micro-economics Eighth, the bottom (line) of the slide is very straight forward: even if all the above mentioned concerns do not work out, the company would still save costs. There is more to this point, however, than making microeconomics explicit. Making these savings this obvious was safeguarding the carbon footprint reduction programme. Whereas the other elements of this slide pointed to legitimate translations of environmental concerns into GFQ’s rationality, the promise to save a couple of hundred thousands USD would also be appealing to actors not interested in ‘the environment’, let alone ‘sustainable development’. This book positions the work reality of carbon accountants in relation to such actors. The following sentence contains a puzzle: I show how modernising carbon accounting led to move GFQ’s carbon reality closer to profits and more distanced from ‘the environment’ and ‘sustainable development’ (see Chapter 4). The puzzle is how one can move carbon issues away from the environment. In this book I show how GFQ solved it. 20 DJSI should be understood as a sign of enormous symbolic capital, i. e. a sign which can be transformed into economic capital if it is known by actors (Bourdieu 1989). 27 1. Introduction To summarise, I argue we can use GFQ’s take on carbon to learn how carbon has become a promissory act, claiming to appease nature’s representatives, investors, publics and management while, yet, increasing the distance between capitalism and environments. Fortune carbon fails. But, surely there must be hope! I attempt to ground any hope in an analysis of actual practices. Any utopian project (whether they are called carbon accounting, corporate environmental management, ecological modernisation, greening capitalism or sustainable development) that fails to be grounded in understanding how current practices work buys into the risk of sustaining unsustainable routines. Therefore, any reasonable utopian project needs to be based in a critique (that does not signify shaming!) of the work practices governing the respective realm. I developed such a critical appreciation of carbon accountants by way of studying how they conducted their work, how they achieved determining GFQ’s carbon reality. Actors, access, materials and first-order data generation Pure luck, or privileged access, I cannot know exactly, provided access to GFQ. After two years of negotiating fieldwork access with various multinationals with limited success I let a contact at the German National Academic Foundation know I needed his help. He asked for a short description of the project. I replied with a couple of lines (I wanted to study the working conditions of environmental managers; using participant observation; I announced wanting to identify chances and problematic factors in their work); three days later GFQ’s head of sustainable development got in touch with me. Her name was Victoria Miller. Three weeks further on I went to meet her at the HQ. (On some other day I figured out, my contact had sent a message to the very top of GFQ.) There is only so much I can reveal about accessing the field. Well, probably you can get a high speed train to – imagine – one of the Western European metropolises, then catch the tube, change, another tube, exit, turn right, left, left and then: if you search for a huge building with big labels of the corporation, you will not find it. Rather, look out for a simple, decent, blackish glass front and lots of office workers – the males normally with suit and tie – wandering in and out, waving ID-cards with electronic chips at a post next to a door, which opens and marks the transformation of the human into a worker. Behind the doors I encountered a large reception desk staffed with security and a waiting area. It was the entrance area into a labyrinth of offices located in the midst of a block of flats. GFQ’s offices, located in the block’s courtyard, were concealed except for a small number of doors or gates through which one could enter from the streets. (Of course, also some representative buildings existed, making GFQ present in the city; but most of GFQ’s HQ was built not to be seen.) Inside GFQ’s private office block, floors with fitted carpet led to standardised offices, meeting rooms, corners with vending machines providing coffee (to be paid for); workers wearing suits and pantsuits, gait often more strut than relaxed, heads up in the air, confident, no marks. I tried my best to adapt. Living during fieldwork periods with a 28 1.2. Accessing Carbon comrade, at daylight I turned into a, more often shaved than not, white-collar worker.21 This physical transformation was part of learning the ways of doing GFQ. GFQ was, precisely during these weeks when we first got in touch, looking for somebody to support them in managing their environmental database. Having been trained as an environmental manager and showing records of working in the IT business, plus being recommended to them by an officer at the top of GFQ, I fit their profile. Dissent from ecological modernisation was unthinkable. I got a contract for the database job. Access granted. When first entering, my position was that of a low-paid worker. My job was to manage the database. It’s designation was Environmental and Social Data Reporting (ESDR) – it is now a key actor in this book. This database was used by sustainable development staff to store and administrate all kinds of data, including workforce achievements, community engagement, green products and masses of consumption data, say electricity, from all over the globe. ESDR was GFQ’s sustainability memory. My task was to analyse the current version of the database, co-ordinate updates and manage their implementation. As part of this job I was routinely in touch with users, their data, the database’s programmers and GFQ’s IT Unit. This job was providing legitimacy to my presence in the midst of GFQ’s carbon data. Within the CSR Unit everybody knew that I was a researcher. Early on, Victoria had introduced me to the team. She explained my technical role and then said that I also was doing doctoral research on cultural aspects of environmental management. My double role was accepted. When meeting new staff, Victoria repeated this narrative and, soon, I reproduced it. Within GFQ’s hierarchy it was sufficient that Victoria (and her superiors) accepted my research activity. In addition to her consent I negotiated one-to-one research agreements with the central agents managing carbon data. Thus I was allowed to observe them during their work, take notes, gather and analyse artefacts under the condition that their identities are not revealed. Most workers were incredibly busy with their work and they left me alone with my work. Occasionally some showed interest in my research. A few workers, however, took greater interest. Among them were a sociology as well as a political science student (both also employed as cheap labour) who became important informants for my field work. One of the key figures in this book was a close colleague, Dieter Klar. A historian by education, he had worked before with some of the theoretical resources I employed (actor-network theory (ANT), which I introduce below); at that time he was a distant-student of sustainability management and, often, we were discussing in depth the issues and practices revolving around GFQ’s environmental management reality. Researching GFQ’s carbon accountants led me to a number of locations. While my main field site was GFQ’s HQ, I also conducted studies at subsidiaries that conducted their business in Western Europe and in Western Asia. In addition I had an informal meeting with a contact of a rating agency. Fieldwork extended to 96 days across twenty 21 Artefact A.1.1 (on page 556) provides an illustration of the visual appearance. 29 1. Introduction months. After this period I still had access to the field; this resulted in further data. These are the sources and processes of first-order data generation: during the field work period I generally had free access to the offices of GFQ’s CSR Unit, their meeting room and, especially, their carbon data (stored mostly in ESDR) and file server. Most generative has been participant observation in meetings of GFQ’s carbon accountants and strategists at which I was receiving access to drafts of GFQ’s shifting carbon policies and facts. During meetings I took detailed notes, visible to all participants. I further collected artefacts and carried out several informal interviews.22 As well as participant observation in meetings, working with members involved listening to colleagues when they carefully explained tasks such that I would contribute most effectively. Sometimes I joined members when they worked on their tasks and allowed me to observe them; I supported members when they met obstacles and joined colleagues for lunch, having a coffee or going for a walk outside the office in a nearby public park. Working alone on and with the shared materials used by my colleagues helped me to deepen my understanding of carbon accounting. I generated digital field notes, following Emerson, Fretz, and Shaw (1995); sometimes this involved taking jottings on paper beforehand. Writing up field notes as the central body of data took place sometimes simultaneously to the unfolding of action (e. g. when others interacted and I was working on my research computer anyway), mostly after events took place (e. g. after a meeting took place or after returning from a fruitful discussion over lunch) or at the end of the day (with the consequence that sometimes I found Victoria to work even later than ten pm); and occasionally I was only able to formulate specific notes only many hours later (in these cases I prioritised writing up those field notes which seemed most relevant). This project’s body of field notes, at the time of writing this sentence, includes 105 notes with an average of 1457 words (amounting to about 300 pages). It occasionally happened that I have been asked to not reveal information about specific occurrences (e. g. concerning classified data and strategies or action non-compliant with GFQ’s internal rules or decisions and opinions on specific individuals). While these occurrences have been included in my analysis this book does not detail them. Except for these cases in which I have been explicitly asked to not reveal data, this book is committed to faithfully representing the realities in which I participated. This ethnographic programme yielded a large body of data. My analysis has selected a tiny fraction of that data for in-depth analysis and presentation in this book (I discuss nth -order data generation in relation to the wider methodology and analysis below). Outlining GFQ’s approach to establish its carbon emissions In the corporate rationality of an ordered reality, an applicable account of the structure by which they established their carbon emissions would involve a clear organisational hierarchy, imply clearly defined individuals and an unambiguous process description. 22 I was also able to participate in a general introductory training course on what working for GFQ means; my general observations on GFQ have also been informed through their intranet. 30 1.3. Doing Data Organisation GFQ used the CSR Unit as a broad container, ranging from charity, development of micro-finance policies to sustainable development. In this unit, a Sustainable Development-Team existed. Victoria was the Sustainable Development head (and my boss). Sustainable development involved environmental concerns and, therefore, she was politically responsible for the EMS within the HQ. Operationally, a coordinator of the EMS existed. His name was Frederik Steine. He had an assistant, Elise Richards. Frederik and Elise were the core staff responsible for carbon accounting. I was responsible to support the database they employed (ESDR). And Dieter was employed as a liaison between subsidiaries’ environmental managers and the EMS. Together, we formed the EMS-Team. Process Year-to-year the EMS-Team would use the same process of carbon accounting. The subsequent three points should suffice to give you a taste. 1. Towards the end of a year the EMS-Team would inform subsidiaries’ environmental managers that they were to collect data on five groups of indicators (until a given deadline, typically end of January of the next year). These groups were called Key Performance Indicators (KPI) and the latter referred to water, paper, energy consumption, distances travelled and wastes generated. GCEs’ agents were to collect facts from the office buildings of their respective subsidiary for these five groups; and, finally, they were to enter that data in the central database ESDR. 2. Following data collection, Elise would scan the submitted data and check for errors. After errors have been repaired, Elise and Frederik would produce environmental balance sheets, presenting a one-page summary of a GCE’s consumption and emissions. 3. Using the data covering these subsidiaries, Frederik was then extrapolating further emissions for the remainder of the total GFQ group. This resulted in GFQ’s statement on its emissions. Victoria would communicate it within the corporation and to rating agencies. According to this story, data flows from consumption practices all over the world to ESDR and then to its audiences. This book follows this data throughout this flow, how it circumvented obstacles, how it sunk into storage devices, how it was cared for, cultivated, how it was torn apart, manipulated, repaired, restored, how it drifted into creeks, how it circulated corporate carbon canals, how it was ebbing onto distant shores. Digital data must be a peculiar fluid. Yet, what is its consistency, which forces soak into it? What is carbon data? 1.3 Doing Data Studying carbon – the concern of environmental accountants and matter of global politics – requires to delve and dive into data. Orientation in that business provides us 31 1. Introduction the field of STS. The studies undertaken in and around this interdisciplinary field make available insights into knowledge-making practices in science, technology, accounting and many other fields. The common understanding developed in this field is that knowledge-making in practices looks very much different compared to textbook-versions prescribing how facts ought to be produced. In the following I point to the key conceptual resources which I employ to follow carbon-in-the-making. To open this process of reconceptualising carbon an example from a ‘different’ realm seems apt; it shall provide us a sense of the range to which reconceptualising is pursued in this book. The example is sex. Following this brief outline, I turn to analytical signposts concerning knowledge, practice, devices, reality and data. Step-by-step I introduce, illustrate and develop the way I use these notions throughout the book – presupposing least familiarity of the reader with them. Thus, for the following it should suffice if I limit myself to only sketch the relations of what it means to engage with how actors know, their practices, the devices they use, how they actually relate to reality and what the status of data will be. Sex and gender as an eye-opener While certainly this is not the place to provide a complete history (and I would not be able to), my humble outline of what sex is (not) might set out from a widespread knowledge in Western societies: humans are either female or male. Immediately, some will intervene and point to the understanding that while this dichotomy is relevant for biology, human culture is more diverse. Human individuals can have much more complex identities and display much variation in how they culturally interpret sex. Luckily, how people live their sex is not determined by their biology. De Beauvoir (1953, 273), famously, states ‘[o]ne is not born, but rather becomes, a woman’. As a result of consequential feminist intervention, the differentiation between between sex and gender became accepted. Sex is then seen as pointing to a biological black-white pattern, a distinct invariant, while gender can be much more colourful variable; gender became unnatural, an issue of culture (Butler 1986, 35). Gender is socially, not naturally, constructed. Now look at the natural, the biological, ‘side’. In human population, in common sense, new borns are either male of female. However, if you recall science classes at high school, you may think of a Gaussian (often considered normal ) distribution. Statistics are fundamental in biology. As most biological phenomena we can interpret the male/female dichotomy through such a normal distribution. Having been parented by two natural scientists, among them a human geneticist, this is a sensitivity which seems self-evident to me – a realist position: in empirical biology (field or laboratory studies) you rarely get 100 percent data; reality is not black/white but you find ‘normal distributions’. For the statement ‘humans are either male or female’ this means that we have to expect that this dichotomous statement holds only for a largest part of a population, but not for the entire population. The curve also has margins; and at the margins Others are represented. In other words, if you take empirical natural science and statistics seriously you can expect that not all humans fit the ideal of being biologically male or 32 1.3. Doing Data female. In fact, Blackless, Charuvastra, Derryck, Fausto-Sterling, Lauzanne, and Lee (2000) report that two percent of human live births do not fit this ‘ideal’. Unfortunately, those who do not fit this ideal are often stigmatised through the very process which represents them as non-fitting. Science, medicine, societies often label them un-normal. Normality turns into an evaluative account. Those who do not fit the expectations have to be adapted, e. g. surgically. How societies treat those entities that do not fit their expectations is of highly moral and political relevance. Some argue that even if naturally, and not only culturally, such diversity exist, this does not have to be accepted. Sex-ratios can be made subject to population control debates, making humans fit desired social orders (e. g. Parkes 1926). However, what constitutes the desired social order? Again, we may be pointed to natural facts – such as that ‘we’ need a monogamy between two and hierarchy among many, that this is the natural order of human interaction. What is the evidence for such claims? We are referred to a close species to us, non-human primates. They would, after all, organise themselves just as ‘we’ are supposed to – employing monogamic and hierarchical structures to order themselves. Historian of science, Haraway (1991b) has brought attention to the contingency of precisely such facts. In her analyses she shows how the knowledge produced by scientists on primate behaviour (on non-human primate social order) is dependent on, inter alia, particular assumptions and material arrangements employed in their studies. She reconstructs facts as not only linked to what scientists assumably study, say primates, but very much inextricably interwoven with the researchers themselves, their position in society and their epistemologies. She identifies claims to singular objectivity as shortsighted for one cannot see everything, see entire reality, but one can only put into practice partial vision. Whether in the field or in laboratory, facts are not a matter of the natural alone but also of the social – and, identifying the boundaries between both remains impossible. Sex and gender are culturally constituted. The distinction looses analytical value. And humans are not inert. They put sex/gender into practice, every day, every hour. We learn to practice sex/gender in the right ways (think of sex education, pornography or simply, again, high school). Butler (1993) argues that humans cannot simply choose some kind of sex/gender but, rather, that their bodies are ‘sexed’. Through constant citation of the norms to which humans have to submit on and into the body, sex/gender is materialised. Then, whatever we do changes, transforms or stabilises a particular phenomenon of sex/gender. Sex/gender is a life-long performance of actors. Sex does not simply exist but is done. To sum up, the natural fact of sex has been de- and reconstructed. The self-evident correspondence of a signifier with a signified has been questioned as much as the adequateness of the imagined signified. This certainly is not a full account and many will produce a better one; but it fired my imagination. Hence this study sets out with a social constructivist sensitivity. This means that entities can be fundamentally questioned in terms of how precisely they are brought into, transformed and sustained in social 33 1. Introduction reality – be it scientific, marital or managerial. One of the core implications is that if I meet notions like fact or knowledge, that elevate certain claims as beyond the social (Hacking 1999, 22-23), I am prompted to question them. I question naturalisations. On knowledge, objectivity, the real and their plurals Knowing something is often associated with taking a step back, and another step back, to know the phenomenon more fully, to encompass more of reality out-there. But, inevitably, by moving away from the object we also come to see less. More and less is not a helpful distinction. From different locations one sees differently. Seeing depends on perspective. For a book contributing to the social studies of environments, an example from seeing environments seems appropriate. Latour (1993a, 219) provides us with a juxtaposition of two ways of finding orientation in space. [Latour’s stereotypically imagined] hunter who covers dozens of square miles and who has learned to recognise hundreds of thousands of signs and marks is called a ‘local’. But a cartographer who has learned to recognise a few hundred signs and indices while leaning over a few square yards of maps and aerial photographs is said to be more universal than the hunter and to have a global vision. Two practices of seeing and knowing a space can result in significantly different, but both adequate, knowledges. One is labelled a local perspective and the other a global view. Often these labels entail an evaluation. Local knowledges are seen as inferior to global knowledges. In STS the recognition of the adequateness to conceptualise knowledge in the plural is is linked to Haraway’s (1991c) feminist work on knowledge and objectivity. She positions herself firmly against both relativist (we cannot know and all knowledge is relative) and totalising (we know a phenomenon in its totality and can establish a singular correct way of seeing the object) stances and suggests as an alternative ‘partial, locatable, critical knowledges sustaining the possibility of webs of connections called solidarity in politics and shared conversations in epistemology’ (ibid., 191). According to her, relativism and singular vision share the denial of ‘location, embodiment and partial perspective’. Following her account we find that ‘[v]ision is always a question of power to see – and perhaps of the violence implicit in our visualising practices’ (ibid., 192). Vision, thus, loses its innocence. Whatever we make visible, we change it. Science, thus, cannot leave the nature out-there intact; Haraway points to entities like microbes and quarks – and surely we can add CO2 e – which are all produced under very specific regimes of making them visible, at particular places and in specific situations. They are momentary traces focused by force fields, or they are information vectors in a barely embodied and highly mutable semiosis ordered by acts of recognition and misrecognition. (ibid., 185) In the laboratory as in the field, the objects ‘seen’ neither can be accessed completely, nor does a singularly correct visualisation technique exist. Knowledge of objects are traces 34 1.3. Doing Data of specific configurations. Configuring the laboratory, equipping the researcher with measurements instruments in the field, yields the artefacts of this specific arrangement. The assumptions, recognitions and misrecognitions, about the object in question coconfigures, to say the least, what will be visible. We cannot, therefore, fully connect to any object. Partial connection, however, is possible. Objectivity, then, requires giving accounts of the partiality of ones knowledge. In my engagement with STS, Haraway’s (1991c) message has been accepted: knowledges are always situated in specific webs of relations among viewers. This has consequences for politics. ‘Struggles over what will count as rational accounts of the world are struggles over how to see.’ (ibid., 194) There is neither any neutral point of view nor do neutral, innocent and non-implicated technologies of vision exist. Alas, this understanding of objectivity – while having ebbed into feminist, anarchist or queer social movements – has not spread into many scholarly debates and, least, into regulatory science. All too often, scientific statements are packaged into a specific genre, a genre of direct representation. Law (2004a, 88) writes: [i]f a statement in endocrinology (or medical sociology) corresponds to a reality out-there, if it simply seems to describe it, then this is because most of the assemblage within which it is located has been rendered invisible, Othered. The authorship, the uncertainties, the enactment of out-thereness, all of these have disappeared. The appearance of direct representation is the effect of a process of artful deletion. Vision, in scientific practices, is embedded in collections of heterogeneous elements. In specific configurations of these, devices can be brought to make present specific visualisations. These are then published, e. g., in journals (Latour 1999a). The appearance of on object being out-there is the desired effect of this kind of scientific practice. However, this appearance is entirely contingent on the precise configuration of entities like measuring devices, samples, assumptions, scientists or cups of coffee. All these relations need to be deleted when aiming to represent the object in the ‘correct’ way. This is what students learn at university; a scholar is not to mention how they got basic results but to reproduce the textbook account of standard techniques. Law (2004a) proposes, therefore, to conceptualise such stories as a specific type of account. Direct representation is a genre in which only a select view relations are authorised to be made visible. And a story in this genre is carefully crafted. Knowledge claims are a matter of practice. And how specific enduring facts are brought into social reality is still underexplored (Doing 2008). At least we have an indication where to start our investigation: at knowledge makers’ practices of describing some things as out-there.23 Out-thereness deserves a further introductory note. My study presupposes a reality. And here is not the space to develop a philosophical treatise on this question. But to make sense of this book, so much needs to be made explicit. Reality becomes. It 23 In the sociology of knowledge further approaches to modes of knowledge exist. Wehling (2006) discuss a range of ways to not know. This study does not investigate knowledges in Wehling’s terms. 35 1. Introduction encompasses every thing, every relation, every past, every future. Hacking (1999, 23) refers to a game of Vish! It refers to circular definitions. I do not want to play it. To define a vantage point of the study, Wynne’s (2002, 472) account of realism fits. It emphasises that reality is • • • • a natural–social–artefactual hybrid; contingent (ultimately in those classifications themselves, too); unfinished, always in the making; and forever incompletely represented. This understanding informs my analysis as well as the status of this book. Thus, I assume that reality is neither entirely antecedent to human action, nor solely causally dependent on humans. While some form of reality may be independent of humans (the critical realist version of real), reality also consists of the imaginations and co-constructions of humans (Sayer 2000, cf. Law 2004a, 158). Any account of reality is partial. Statements signified as mirroring the real are partial. Alternative statements exist and are possible. Reality, while imagined as out-there, is always also a representation. As a consequence, not a singular but, multiple representations exist. I call them realities. This book shows that a carbon accountant’s environmental balance sheet is as much a particular visualisation technique like a laboratory, that these techniques are linked to multiple realities and that this causes friction. What my account on knowledge-making provides, then, is not so much a suggestion to shame the problematic and frictional realities but to allow the reader to become familiar with the practices, assumptions and politics built into the knowledges produced. Practices and performativity Producing environmental facts is a matter of practice. But what does that mean? I follow a question that Suchman’s (2000, 316) study on the production of an environmental impact assessment (which is a statement of the impacts of an, e. g., engineering project on an environment) phrases well: ‘[t]he question [. . . ] becomes how, as a practical matter (rather than only as a matter of faith, although I think it must remain that as well) do they do it?’ Studying practices does not only consider the accounts of practices, such as a laboratory protocol, a standard’s methodology of how to calculate certain carbon emissions or in an interview, but it fundamentally engages with all the elements that are part of practices by which members in a particular situation achieve some thing (while, potentially, failing in other respects). Studying practice in this book is about how some thing has been achieved, rather than about evaluating whether. Setting out to study practices relies on an understanding that through actual, material, practices reality is shaped and transformed. This take breaks with both structuralist understandings (in which, say, class acts by itself24 ) as well as the rational actor paradigm (suggesting that actors have particular goals and calculate their actions). This take may be conceptualised as contributing to – what has been called – the practice 24 36 Of course concepts like class can be made to act – with real effects. See Chapter 4. 1.3. Doing Data turn. Schatzki (2001, 2) (aiming to meta-theoretically conceptualise various turns to practice in social theorising) identifies as the core of these turns: ‘practices as embodied, materially mediated arrays of human activity centrally organised around shared practical understanding’. Turning to practice, thus, starts with a distinct ontology. Practice is through what this book understands carbon accounting. Yet, saying we turn to practice does not make clear how we study it. I follow an accepted practice in STS that is to visit actors in particular situations. I refer to those humans who are part of situations as members. Members’ action is situated. Suchman (2007) suggests to study situated actions by turning to how actors bring into their reality the resources they, then, can use to shape the situation which, simultaneously, shapes their actions. Resources may be material (e. g. a desk), conceptual (e. g. a plan) or social (e. g. a hierarchical relationship). These, however, are not managed by humans according to the rational actor paradigm but in contingent, that is neither in arbitrary nor structurally determined, yet patterned ways (Garfinkel 1967; Bourdieu 1990b; Amsterdamska 2008). This holds even for something as – assumably rational – as mathematical calculation, as shown by Lave’s (1988) study of maths in a supermarket. The practical production of knowledge (think of a laboratory) is not a simple, one-time act. Much rather, for knowledge to be produced ‘things need to be done and redone to have a presence in the world’ (Waterton 2010, 650). This means that for knowledge to exist, repeated engagements with elements are necessary that make certain things, say a species, thinkable and manageable. Thus, members’ action of knowledge-making have effects. Statements do something; they inform members (Moser and Law 2006, 66). That quality of practice that makes things present, that changes things, is usually called performance. Like a marriage is performed through particular practices (Austin 1962), the status of the relationship between a company and their carbon emissions is performed. By way of publishing a sum of their emissions, e. g. in a CSR brochure, the company performs that number as a fact; the companies makes present their identity with the fact. Studying the performance of facts means that we are concerned with members’ practices. Of course, some will say, knowledge has to be performed vis-à-vis others; therefore, a study of corporate carbon emissions should attend to the greenwashing activities of a PR unit. I want to take another step. Not only is knowledge performed in relation to audiences of that knowledge, but the production of that knowledge is also a performance. Producing the knowledge requires skilled engagement with props (the mass spectrometer, the invoice). Only when practitioners manage to perform well, i. e. assemble props in a certain configuration, the desired effect (the knowledge product) will come into existence. Mol (2002) calls this generative quality of performance enactment. She uses this concept because she wants to emphasise that, in her study on doing diseases in a hospital, there is no backstage to the performance. The notion of backstage refers back to Goffman (1959). He suggests that people, day-to-day, perform specific characters and masks vis-à-vis each other and that, however, behind the scenes another layer of reality exists: 37 1. Introduction the individual performer’s true identity. In her version of theorising, realities are said to be enacted.25 I take this notion to stress that reality does not consist of some objectively truer backstage area, but just of multiple stages. Nevertheless, in this book, I also address as staging those performances by members which are based on a stage-performance conception. That is to say: members sometimes perform performances; they may stage a reality in a way that does not correspond to their understanding of another reality that they consider more truthful. For such occasions the notion of staging seems helpful. Members’ performances are also constitutive of discourse. By means of, e. g., their speech, their textual artefacts and visualisations they produce, reproduce and proliferate specific patterns of understandings (Keller 2011b, 48-49). Gilbert and Mulkay (1984) stress that organisational internal discourse and their external presentation may differ. They make this case for science, showing that often multiple accounts of an object exist (ibid., 2). Most relevant for this book is that in the laboratory, when members were not struggling, the contingency of facts on particular arrangements and their fictional character were accepted. However, when accounts needed to be defended (e. g. in textbooks for students) scientists performed realist depictions. If these findings from science are indicative of the reality of carbon accounting we can expect that multiple discourses are maintained by members in the field. Devices and the enactment of reality Imagining a study of knowledge-making practices – of particular situations in which facts are performed – inevitably involves noting material elements. Haraway (1991c) points to the role of human eyes for vision (and, she critiques the politics of making visible without marking who sees how). Eyes can be enhanced: using devices like a magnifying glass or an electron microscope. Behind the eyes we imagine a brain. The human brain is so limited (is it really?); we can support it with spreadsheets, we can extend memory to server farms. ANT is an analytical toolkit, says Verran (2010, 101), which I made much use of for the analysis underlying this book. Chapter 2 introduces how I utilised this toolkit. Here, it suffices to signpost how it supports studying the doing of carbon data. Latour and Woolgar (1986) study what they call laboratory life. Their understanding is that science is not so much revealing a reality which existed before; but rather that reality is a consequence of scientists’ practical work, i. e. a micro-social phenomenon. Scientists’ approach to reality is precisely to ‘make it a given’ (ibid., 236). To achieve that, they enlist various resources. Every thing in the laboratory shapes the facts produced. The fact emerges as an effect of particular circumstances, the network of things in the lab. And producing certain truths, rather than others, is simultaneously about struggles with colleagues and about what counts as real. This approach to understand how reality is produced can also usefully be applied on other social reality, Latour (2005) argues. ANT, understood as politically naı̈ve by critical carbon market 25 38 I return to these theoretical discussions in Chapters 2 and 3. 1.3. Doing Data scholars (Descheneau 2012), it is also recommended for examining how macro-social order is constructed at particular micro-social sites (Michael 1994). Most significant for studying carbon accounting is that with ANT we can systematically trace and reconceptualise the ontology and ontological status of carbon. Like feminist work now recognises that sex is not pre-discursively given but is materialised (Butler 1993), this book argues that environment, and specifically, carbon is an effect of social and material practices shaping the apprehension, filling the latter with content. As a resource for this argument I draw very much on Mol’s (2002) work. She proposes that when studying the doing of a disease – her example is atherosclerosis – we find that at different sites at which the disease is enacted it can diverge. In one hospital unit atherosclerosis may be located in the leg; in another unit it is located in the population. ‘Reality is distributed ’ (ibid., 96), it is multiple. Different devices enact different realities of the body. A singular patient’s body is made to exist in multiple ways and the workforce within the hospital is needed to singularise them. Barad’s (2003) account goes even further. She proposes that realities do not and cannot leave the apparatuses in which they are produced. In Chapter 6, in a first step of drawing together my analysis, I engage with how these notions help to understand carbon. That environmental realities may well be complicated has now reached environmental sociology. Unfortunately, in the EMT proponent Mol’s (2010) account of an environmental sociology of networks and flows a significant incoherence exists. Drawing on ANT he notes that ‘[t]he concept of nature as external to society is outdated’ (ibid., 34) and sets as the paradigm for this new environmental sociology a study of flows: (a) by analysing flows of information, capital, goods and persons from an ecological rationality point of view (by looking at environmental information, green products, green investment funds, sustainable management concepts, environmental certifications schemes, flows of environmental activists, and their ideas); and (b) by analysing environmental flows as such, that is: energy, water, waste, biodiversity, natural resources, contaminants, and the like. (ibid., 30) Arthur Mol takes a realist stance by assuming that these flows exist as such (where Annemarie Mol (2002), likely, would question their status). Pellizzoni (2011, 800) makes the same point. ‘The ontological underpinnings of ecological modernisation theory are straightforward: nature and society are seen as interacting but distinct realms.’ This book challenges this distinction by way of investigating how accountants enact a flow of data resulting in a flow of emissions. Data and information In science but also in accounting, the production of facts involves data that is translated into information, interpreted; facts are the effect. That knowledge can be used, again, to inform others – a company, a government, an activist. Struggles over substantive questions, from the construction of a nuclear power plant to climate change or the insurability of potential damages through nano-technology, assumably, are translated into struggles over ‘objective’ information (Beck 1996). These 39 1. Introduction knowledges are often not well accessibly for many affected and, therefore, people have to trust the producers of data (Porter 1995, 214). In economist terms of knowledge production, one might say that information should be allocated optimally. Where no information is demanded, none should be produced. However, what happens to the argument if we zoom into what information is? We find that information is a contingent product, bound up in interaction – not static at all. Heath, Knoblauch, and Luff’s (2000, 315) account of studies of knowledge workplaces underline this: information as a blanket term, to encapsulate a disparate and unbounded array of materials, matters, and the like, becomes untenable, as we turn analytic attention to the ways in which particular organisationally relevant information gains its significance and determinate character in actual courses of action and interaction. Information is inextricably embedded in practice and practical action. The question we have to struggle with, then, is how information is enacted. What are the qualities of information? How can they differ? Studying information we can expect more information. Depending on the position of a speaker, some information is taken as given; that information is called data. In knowledge production, members translate data into specific forms; for instance from a petri-dish into a table; from an invoice into an accounting form. The result, therefore, is called information (Latour 1987). Accounting is all about this (Robson 1992). Engaging with such translations is a recurrent theme in this book. Thévenot (2009) underlines that any standardised methodology presupposes users to invest in specific forms and, by that, to disregard other potential forms of making sense of out-there. Users of manifold information often employ second and nth -order data formation devices: databases. These can range from collections of information incorproated in bodies (e. g. specimens in a museum (Star and Griesemer 1989)) to paper-based (e. g. an offices’ filing system (Suchman 1983)) or digital devices (e. g. a representation of earth’s biodiversity (Bowker 2000a, 2000b)) . These databases, in addition to the data they ‘carry’, incorporate even further histories. Databases are formatting what they store as well – which makes them less a storage than a transformation device. Waterton (2010, 648), therefore, argues ‘that they obscure some fundamental realities: that all archives both reveal and hide, that all archives harbour their own tacit politics, histories, and powers’. A decisive point that occurs when we study how information/data is produced and translated is that we come across incredible amounts of mundane work. The database designers, the cleaners of the server room. Their work enables the forms, the databases, to not break apart (think of the secretaries helping you to fill a form). Often this work is carried out by women. Feminists, thus, draw attention to the significance of this enabling and maintenance work which helps to keep things together. I refer to these qualities of work several times in this book. It is called articulation work (Strauss 1988; Star 1991); and database managers do precisely such work (Bowker 2000b). Important to manage data is the availability of data. Thus, an issue this book is struggling with 40 1.4. Researching Corporate Carbon is how members engage with missing data. Porter (1995, 81) claims that while ‘[i]ndex numbers could never simply be observed; they normally involved extensive data collection and often difficult or at least tedious calculations’. He stresses that calculation, thus quantitative operations, and the existence of data are fundamental to number generation. However, Lampland (2010) argues that to get large scale quantification processes working, making up numbers which meet audiences’ expectations of that data is well working in practice; effectively, numbers may be found to be placeholders (Riles 2010), i. e. they may be used as-if they represent some out-there well enough while some agents recognise that the representation is highly uncertain. Doing data is also always about reading data, making sense of what one sees. Moser and Law (2006, 66) propose to conceptualise information as something that informs a decision; it ‘is what is present and relevant’. While at first glance a useful hypothesis for a study of doing data, we also have to consider that the absence of information is highly relevant. Summarising facts means members have to classify the relevant from the non-relevant; sorting things out (Bowker and Star 2000) implies that certain narratives are silenced, deleted.26 Their absence is highly relevant. Decisions could not proceed with all possibilities still on the table (Suchman 2000). Information, hence, should be seen as relationally constituted through what is made present and absent. It is not surprising then, that financial analysts and managers are interested in alternative narratives, rather than restricting themselves to accounting information, to form judgements (Davison 2008, 794). At the same time, uncertainty about out-thereness may sustain current modes of governing despite calls for the contrary (Pellizzoni 2011). Indeed, whether more data, more information or more knowledge is helpful to solve problems in the assumably represented realm of out-there is highly contested. Hetherington (2012), for example, shows that more information about resources distribution does not solve struggles over resources. Adding more knowledge allows, much rather, for additional layers at which struggles can be plaid out. This, of course, may help some actors (and not others). Holmes (2010, 229) challenges the veil of transparency connoted with data. Showing off lots of data may well be used to stop people from asking uncomfortable questions. To me it seems, therefore, even more important in order to know which questions to ask, to understand what data is. 1.4 Researching Corporate Carbon The credibility of (ac)counts of carbon emissions is a social, political and indeed a material problem. My aim in this book is to articulate a view of emitting as a practical achievement by workers who assembled and ordered heterogeneous epistemic, material and technological elements; my view is inspired by methodological reasonings of debates on ANT, ethnomethodology, grounded theory and by Pierre Bourdieu. With these I 26 Bowker (2005) studies memory practices – different ways of remembering knowledges. My study does not closely relate to Bowker’s account. 41 1. Introduction developed a multiply constituted ground for analysis. Thus, employing these approaches I secure the foundation of my account of carbon emissions. I do not claim to have woven these methodologies into a coherent toolkit, where different methodological devices are arranged into a neat division of labour. Much rather, being inspired by them means that I draw on sensibilities developed within their respective discourses. I use these sensibilities to triangulate workers’ practices within my epistemic practice of knowing the doing of carbon counts. The credibility of my account of GFQ’s carbon counts relies on making present that the foundation of my analysis has been multiply tested and strengthened. Employing disparate methodologies allows to underline the credibility of my account better than using a single streamlined methodology. To make the logics of my study debatable, in the following I sketch my account of carbon as grounded in five parallel considerations: the rationale of the study, the selection of a case, a clarification of my central assumptions, the structure of the method-assemblage I developed to engage with GFQ’s workers and the emissions they enacted and, finally, a technical account of the ethics involved in this project (I discuss the wider ethical-political ramifications as part of the rationale). Rationale of the research project The choice to research day-to-day work practices of agents of ecological modernisation within a multinational has a clear fundament. At the end of my undergraduate degree in environmental management I worked on a programmatic vision of what kind of interventions in academic reality could help to make environmental management the object of study rather than the taken-for-granted aim (Lippert 2010b, Part I). I argued that a significant step would be to study these goal-directed practical activities (i. e. environmental management) as a social process which is conducted under certain conditions in a given context (ibid., 79). Along these lines and specified in the reasons laid out in the first section of this introduction (i. e. that to study sustainable development in its hegemonic form we need to turn to organisations’ environmental managers) I conducted a study to conceptionalise agents of ecological modernisation (Lippert 2010a, Part II). This study concluded that their practices constitute an apt site to study how the hegemonic order of unsustainability is sustained and can be transgressed (ibid., 111). In parallel to narrowing down this project’s object, i. e. agents’ practices, and constitutive of this focus was also an analysis of sustainable development and in particular greening activities in relation to Capitalism, Democracy and the State (ibid., 2010a, Part I; 2010b, Part II). Engaging with these relations and academic knowledge production made clear that not only was objective and neutral knowledge production not possible but also that knowledge production dedicated to emancipatory struggles would be well-advised to connect to those agents of change who are interested in conducting the radical transitions towards some form of reconstructed interpretation of sustainable development so much in need (cf. Lippert 2007). Some of these agents connect with academic discourses. While engaging with respective discourses around feminist techno-science studies (especially Haraway (1991a), 42 1.4. Researching Corporate Carbon Barad (2003), Suchman (2007)), marxism-oriented Critical Management Studies (CMS) as well as discussions of carbon political economy and of critical realism (foremost Levy (1997), Sayer (2000), Böhm and Dabhi (2009)) I turned towards debates in a particular branch of libertarian studies: an engagement which would be sensitive to multiple axis of domination (not only over women, workers and the environment) and which would not look for simplistic ways out of the crises that humans and non-humans, men, women and intersex people, rich and poor, are so differentially subjected to – such as proposing that ‘the’, or, ‘a market’, ‘a government’, ‘a political party/movement’ or ‘the state’ would be realistically able to lead the ways to the needed transitions. A political theory of democracy which seems well aligned with these criteria is anarchism. In response, this book intends to contribute to an anarchist sociology – a sociological study which does not presume that any of the existing dominant ways of societal organising provides the desired solution or needs to be sustained as an end in itself. Now, anarchism, some may object, is a horrible approach – full of terror and chaos. Engaging with these kinds of discussions needs other texts.27 It suffices to say that in the kind of poststructural anarchist theory, which I am interested in, revolutions are theorised as being concerned with the practical enactment of emancipatory relations among people and other entities. The archetypical anarchist is concerned with prefiguring the desired relations in the present (Franks 2006, 193) – ‘revolution’ is not imagined as a final moment but as ever continuing every-day practice (Gordon 2008), change towards the desired configurations of social-material relations. With respect to my field of study – studying technologies of carbon accounting – I should mention: a long history exists on differentiated anarchist engagement with technology – ranging from the Luddite uprisings (Grint and Woolgar 1997) via philosophy (Winner 1978) to gardeners and hackers (Gordon 2008).28 Thus, anarchist sensitivities informed this project. Similar to Fernandez (2009, 95) I am aware that academic knowledge production may be useful for social and corporate control. Therefore, rather than seeking detachment and developing conceptions useful for large-scale and top-down implementation projects (by a multinational, by a nationstate, by one of the globally working NGOs) I aimed to produce a knowledge useful for individual workers to understand their practices and its consequences, a knowledge useful to build a common ground for green-collar workers to connect and organise emancipatory movement. This knowledge is not to serve for seeing like a state (Scott 1998) and I aimed to avoid providing a grid which would deconstruct the field of carbon 27 Introductory readings to anarchism. . . well, you are going to look at wikipedia first anyway. Go for it. If you are done, go to the website of the British Political Studies Association’s Specialist Group for the Study of Anarchism, called the Anarchist Studies Network (www.anarchist-studies-network.org.uk). Obviously there are classical readings (on workers e. g. Rocker (1938), on women e. g. Goldman (1910), on nature e. g. Bookchin (1993)); yet I would rather point to Le Guin’s (1974) The Dispossessed or the film Land and Freedom directed by Ken Loach and written by Jim Allen (1995) to dive into the wide range of historical and utopian considerations. 28 Also, in general, anarchism is no complete stranger to anthropology (Graeber 2004; Morris 2005), sociology (Jaworski 1993; Ehrlich 1977) and STS (Winner 1980; Michael 1994). Much rather, as these scholars show, in these fields anarchist considerations were influential among these fields’ ‘founders’. 43 1. Introduction accounting as a normatively neutral space (Gibson-Graham 2006, 75). My intention is not to prescribe any particular regime of practices. Thus, this book is not followed by a list of check boxes.29 Rather, inspired by STS’s reconstructivist tradition, I hope this book acts as a catalyst of change (Woodhouse, Hess, Breyman, and Martin 2002) by way of helping agents involved in diverse struggles (inside and outside corporations) to learn from another. Following Mol (2002, 164) this book is not seeking to command readers but rather to point and lay out avenues for learning together. Not only sociological theories and methodologies have followed a turn to practice; also poststructuralist political theory and theories for/of practical intervention in every-day affairs have taken paths towards engagement with practices. In May’s (1994, 87) account of a poststructuralist political theory of anarchism practices are claimed as the ‘proper unit of analysis’. He relates practices to ‘goal-directed social regularity’ and, at the same time, recognises ‘that the goals people think they will achieve when engaging in such practices and the consequences they actually do promote are often very different’ (ibid.). I like to rework his understanding of practice sociologically and contribute with this book to an analysis of practices which is not only of purely academic interest but is helpful to identify practices which are apt to inform social struggles seeking practices going beyond the dominant approach to sustain the unsustainable.30 Studying action imagined as rational or nearly scientific, like accounting, is of interest to anarchism for at least two reasons: first, anarchism has a long history of an utopia grounded in the possibility of reason, classical modern rationality, science (see, e. g., Bakunin 1916). Second, as Bakunin (1916) recognises, a society that is governed by scientists (such as imagined by Bacon (1626)) is well likely to reproduce oppression of many through few. Purkis (2004, 47) interprets Beck (1992) as saying that modern societies struggle with the consequences of their modernisation project. And Purkis questions the reaction taken by administrators of reflexive modernisation; i. e. he challenges that those institutions which created the social and ecological problems are asked by societies to solve them while not recognising ‘the part that hierarchical structures and the profit motive have in perpetuating these problems’ (ibid.). In my understanding, transitions towards any kind of utopia need to be grounded in an understanding of the enactment of relations in the here and now. Therefore I turn Purkis’s (2004) critique into an empirical question: what are these institutional actors (who supposedly perpetuate these problems) doing and how do they achieve it? Now, when engaging with hegemonic actors we find that they have huge claims to objectivity. Poststructural sensibilities, taken on board in anarchism, question claims to 29 Drawing on Crozier and Friedberg (1979, 1), I suppose, this is an acceptable move in the body of literature on management. 30 Goal directed practices link to issues of strategy and tactics. This is a main strand in May’s thought. He views strategic political philosophy as focussing on singular goals. Actors may use tactics in every-day which may stand in complicated relationships to formal strategies. Often strategy and tactics are considered as distinct (see also de Certeau 1984). I am not following and reproducing this distinction to the letter because, in practice, doing strategy and tactical movement overlap. 44 1.4. Researching Corporate Carbon singular truths and neutral relations of representation (Kuhn 2009).31 A central notion in poststructural theorising, Deleuze and Guattari’s (1987) rhizome deserves a comment before linking it to our debate. Deleuze and Guattari draw this concept from biology where it refers to horizontally growing rootstock entities of which smallest parts broken off can live and reproduce as clones. If you know plants, I would propose to imagine their concept in relation to fungi and the characteristic of mycelia to exist in all kinds of sizes (with some being considered as the largest living entities on earth) and the characteristic of mycorrhizae to form symbiotic relations with other life forms. Kropotkin (1902) would call it mutual help. Deleuze and Guattari (1987, 6-7) formulate it this way: A system of this kind could be called a rhizome. A rhizome as subterranean stem is absolutely different from roots and radicles. Bulbs and tubers are rhizomes. Plants with roots or radicles may be rhizomorphic in other respects altogether: the question is whether plant life in its specificity is not entirely rhizomatic. Even some animals are, in their pack form. Rats are rhizomes. Burrows are too, in all of their functions of shelter, supply, movement, evasion, and breakout. The rhizome itself assumes very diverse forms, from ramified surface extension in all directions to concretion into bulbs and tubers. When rats swarm over each other. The rhizome includes the best and the worst: potato and couchgrass, or the weed. Animal and plant, couchgrass is crabgrass. And they extend the notion beyond the plant kingdom to animals and even further: to collective assemblages of enunciation, to a whole micropolitics of the social field. A rhizome ceaselessly establishes connections between semiotic chains, organisations of power, and circumstances relative to the arts, sciences, and social struggles. (ibid., 7) Anarchists are inspired by this. Some authors interpret rhizome as an antiauthoritarian and micro-sociological approach, stressing that people are shaped by power relations and implicated in them (Purkis 2004, 50). Kuhn (2009, 23) sees poststructuralist informed anarchism (i. e. postanarchism) to be based, inter alia, in ‘[a]n uncompromising commitment to plurality and all that goes with it – rhizomes, cracks, shifts, fluidity, etc.’. Some, like Franks (2007), struggle to emphasise the long tradition within anarchism to recognise multiple vectors of struggle (I noted some above, sexism, class, religion, environment). The empirical question is how powerful which dynamics are. Now, Law (2009, 145-146) suggests ANT as an empirical translation of poststructuralism. With ANT we are offered a toolset which does not presuppose an ordered society (like a properly functioning State, Marriage or Accounting System); rather ANT is to enable us to study rhizomes, ‘the provisional assembly of productive, heterogeneous, and (this is the crucial point) quite limited forms of ordering located in no larger overall 31 With Feyerabend’s (1975) Against Method: Outline of an anarchistic* theory of knowledge anarchism is equipped with a text which is seen as a significant reference point far beyond anarchist debate. (The star ‘*’ refers to Feyerabend’s (1975) opposition to Puritanical seriousness and his embracement of ‘joyful experiments’ (ibid., 21).) See also Purkis (2004, 41). 45 1. Introduction order’ (ibid., 146). The ontology assumed in poststructuralist thought is this: there is no pre-existing order but multiple, heterogenous, more or less extending relations and orderings. (For the biologist: think of hyphae as a metaphor.) Now, this kind of ontology breaks – with the classical self-depiction of the modern state, the separation of powers (trias politica), natural conceptions of class struggle, the self-evident existence of property rights, the natural dominance of men over women, a binary conception of sex/gender, any fixed conception of human nature. Poststructuralist ontology breaks with the fundaments of the State and, therefore, argues Koch (1993) can be considered systemically anarchist. Consequentially, STS scholar Michael (1994) advertises ANT in Anarchist Studies as a method to study phenomena like the State, to study how ‘political inertia is maintained’ as well as how change could be brought about. Studying these assumably powerful institutions, like a multinational, means studying up (Nader 1972). This approach may precede ANT but, nevertheless, practically has to go together. The urgency expressed in anthropologist Nader’s (1972, 1) paper still holds: ‘the quality of life and our lives themselves may depend upon the extent to which citizens understand those who shape attitudes and actually control institutional structures’. Among her original proposals was to study institutions like air pollution agencies. Now it is carbon. Staged as a global problem, enacted as emission rather than pollution (Asdal 2008, 126). Studying up is still relevant, especially in studying techno-science (Forsythe 1999). And it is clear: the perspective should not be locked in at studying up, we may also study down and sideways (Nader 1972, 8). Case selection criteria and possibilities to generalise For this study (to my best knowledge) was to be among the first ethnographies of agents of ecological modernisation as such (based upon literature research for several years32 ) I intended to provide a detailed study of the practices of a few individuals and study how they achieve whatever they were doing. A case study in one corporation seemed fit. Criterion 1 For EMT postulates (theoretically and politically) the global diffusion of ecological modernisation (e. g. Huber 2008) I decided to search for a multi- or transnational corporation. Criterion 2 As some green niche company might well be able to satisfy a niche market, I wanted to ensure that the case would be representative of leading capitalist practice (criterion 2). The best way to specify this criterion was getting access to a Fortune 100 or 50 company, i. e. one of the globally largest corporations. Criterion 3 Finally, studying agents of ecological modernisation in-depth required access over an extended period and to, at least, large parts of their work practices. I started to search for field sites end of 2005. Initially, after only a couple of months search, I was able to forge an agreement with one of the world largest telecommunication 32 My literature review was restricted to studies in English and German, unfortunately. I am very interested in contact with authors of similar studies in other languages – if you exist! 46 1.4. Researching Corporate Carbon services providers to study them. Unfortunately, when I was ready to start conducting the fieldwork, this agreement had vanished (the gatekeeper had lost the position that would have allowed me to study their environmental managers). A new search started. About 18 months later GFQ answered to my call. The corporation fit the named criteria. The specification of the site ‘GFQ – EMS-Team – carbon accounting – ESDR’ seemed particularly interesting for it allowed even further generalisation drawing from this study. To introduce this study, four factors that allow generalisation are particular noteworthy: the techno-political category, the industrial sector, the state of techno-managerial development and the particular organisational relations GFQ enacted. Techno-political category: Carbon Given a hegemonic discourse of climate change and ecological modernisation, carbon emissions accounting is relevant and present in businesses across many markets. Therefore, this particular techno-political instance of environmental management can be seen as an example of how capitalism enacts environments. Industrial sector: Finance In the discourse of neoliberal climate change politics, free market forces are best suited to identify best practices. The finance sector was not, yet, regulated through the Kyoto Protocol (in terms of, e. g., the EU ETS) and is not identified as a culprit politically, which, I assume, means that this industry sector is less distorted by state ‘interference’ (than, say, the oil sector). Thus, the finance sector should generate good solutions (in terms of the market) to climate change (rather than solutions significantly shaped by the Regulator). With the finance sector we can identify what kind of solution to carbon accounting seems fit for the market. Does the market take carbon seriously?33 State of techno-managerial development: Before closure Closure means that consensus over the proper ways of saying and doing things has been established. The study of GFQ’s carbon accounting system showed that their practices were highly evolving along a specific trajectory of change. This book analyses the trajectory of change within the distributed socio-technical apparatus of carbon accounting. Precisely because I was able to study a process ‘before closure’ I was able to analyse techniques of closing, i. e. the qualitative shifts enacted in the 33 The practices observed revolving around GFQ’s carbon accounting system may be significantly shaped by the fact that the financial services sector has not been addressed by the Kyoto Protocol as a polluter. Rather, the corporation participated in the VCM to ‘offset’ emissions. Under the EU ETS those polluters who are obliged to reduce their emissions are grasped through their installations. This book does not investigate the environmental and carbon accounting systems of installations’ operators. These may be structured very different and, thus, may imply very different translation practices compared to the ones observed in this study. A companion study to this book might investigate what is materially black-boxed into measuring devices (e. g. ones which measure emissions at a smokestack). After all, however, my study speaks predominantly for the accounting processes after the ‘measurement device’. Therefore, I propose, this study is indicative for all those corporations which account for their indirect emissions. If we can trust Hoffman’s (2007, 11) statistics, then my study should be very normal for he found: 77 % of his sample companies measured indirect emissions (and so did GFQ) and 63 % of the companies have developed new IT to track GHG emissions (ibid., 13) (and so did GFQ). For GFQ seems to fit well into the pattern, I feel positioned well to argue that the practices I observed should be relevant to a significant part of companies producing carbon emission knowledges. 47 1. Introduction doing of carbon accounting to render carbon accounting docile and in control. We can assume that qualitatively similar techniques of closing have taken or will take place in other cases of carbon accounting. Organisational relations: Well-linked to other institutions The carbon market is highly concentrated (Dorsey 2011). That is to say, a few actors dominate the field. Part of this ethnography are the links between GFQ and one of the most relevant ranking organisations informing this market (which I also render anonymous) and the links between GFQ and one of the globally largest nature conservation NGOs. In addition, this book covers the ways by which their carbon accounting practices have been governed by particular standards (including the Greenhouse Gas Protocol (GHG Protocol), a Global Reporting Initiative (GRI) standard and the Standard of the Verein für Umweltmanagement und Nachhaltigkeit in Finanzinstituten (VfU)) as well as how GFQ’s implementation of these standards has been audited by on of the largest accounting firms (one of the so-called Big Four), one of the six largest consultancy players in the carbon market (ibid.). And Burritt (2012, 397) believes that accountancy firms are decisive in establishing corporate carbon footprints. In addition to these formal linkages this study covers how agents of GFQ rendered their practices comparable and similar to their competitors. For these points I argue that albeit GFQ is only one multinational, folded into the object which this ethnography studies are capitalist solutions to greening, the finance industry sector, and key actors of the carbon political economy and GFQ’s competitors. This book provides the evidence for this. A comparative analysis of carbon accounting in different organisations would show both similar patterns as well as differences. We will see that many of the statements I make do not only hold for this case. To give you a taste: extrapolating data is necessary in all sorts of empirical work; one often simply cannot gather data for complete populations. For the case of carbon this means that emissions are extrapolated. The analysis will detail the implications; and several implications do not only hold in GFQ but in general for any performance of extrapolating. Assumptions Of course, no study can commence without a number of assumptions through which the researcher can think and use any senses. As Clarke and Star (2008, 116) put it: we cannot set out ‘with the image of some sort of tabula rasa’. In addition to the introduction so far (that laid out and argued for the frame which I assume), four areas deserve further elaboration: my ontology, epistemology, the use of concepts in this study and methodological considerations. As I said, this study sets out with a poststructuralist ontology, which is to assume no particular ontology except of, in my take, that a reality exists which is shaped and exists through natural-social hybrids, which is highly contingent and not stable. In this changing and shifting reality, things do not necessarily add up, they cannot necessarily be known through deterministic patterns and hierarchical forms. Law (2004a) calls 48 1.4. Researching Corporate Carbon it mess. With him I argue that the ontology we are dealing with is outside of the area typically imagined by social science research, which is as he puts it, is ‘that the world is properly to be understood as a set of fairly specific, determinate, and more or less identifiable processes’ (ibid. 5). Showing the evidence for my argument takes the reader through the book itself. Necessarily to accept before reading the book is merely the open-mindedness that the world is not clear and definite. My analysis did not presuppose such clearness. However, to ease your stomach, I do find a number of clear elements and processes which I am able to satisfactorily and rightfully describe and analyse as such. This is to say that the ontology of the underlying study did not rule out deterministic patterns; I did not want to write them into ‘my data’ a priori. A significant assumption I do entertain about reality is that many things are related; and that through their particular relations reality is changed and enacted. Of course, speaking of relations implies substances, or, at the least, containers or apparatuses; which one could iteratively zoom into. So, substances are relational effects; substances are ‘made real in practice’ (Law 2009, 147). Vis-à-vis this logics I should point out that, again, making reality does not start with a tabula rasa; reality is not constituted right now in front of you. But relational effects sediment into history – and that history may be material, like sex, like a Nation. However, if performers stop performing nation-state boundaries, or their sex, much may change; and they may encounter resistances by others who profit from certain performances. So, the world is relationally constituted and the effects can be very real to its performers. Also in terms of epistemology I employ a notion provided by Law (2004a): methodassemblage. He uses this term to suggest that through methods and techniques of vision we enact reality. This is certainly the case with this book. I make present a particular version of GFQ’s carbon emission practices; I enact this out-thereness. Now, if we accept that the realities ‘measured’ are at least co-constituted by the methods through which they are ‘measured’ then the concept of triangulating can be either seen as flawed or of particular help. Triangulating the object-out-there seems to assume such an object to exist out-there; a realist stance which I cannot share. But, obviously, the constructivist realism I exercise assumes things and relations to be out-there at the same time. Yet, I cannot access them easily. To know them, then, better I employ several techniques of knowing them and, through that, try to develop a more balanced understanding of the phenomenon (and this is seen as triangulation (Flick 2007b); more on methods below). My understanding of research quality is that by using more method-assemblages we can develop a more saturated understanding – that is saturated with additional relations of accountability that my claims have to endure. The final understanding I develop is far more removed from the object (that I originally tried to engage with) for it is distanced through all the subsequent chapters.34 Accordingly, to show how distancing was performed through my analysis I need to discuss the material 34 As an example of distancing techniques, consider documents. In the field I produced field notes. Then I operated on and with these notes and produced further documents. The book is, hence, a product of specific ethnographic documentary-discursive practices. See also Riles (2006b, 17). 49 1. Introduction in-depth and iteratively throughout the subsequent chapters. A significant implication of an open-minded ontology is that predefining concepts to signify entities out-there does not fit well to this study. To illustrate with Bewernitz (2005, 65): some ontologies would assume that ‘workers’ exist out-there; but what about managers? Analysing work relations around managers indicates that orthodox marxism would have to conceptionalise them as workers for their are paid and dependent on the owners of capital; but at the same time they may be paid to perform being a capitalist; they may also be provided some capital ownership.35 Following Gilbert and Mulkay (1984, 2) I propose that defining a priori a number of concepts to make sense of members’ realities does not allow accounting for the ‘multiple and divergent versions generated by the actors themselves’. This precautionary relation to ready-made categories also extends to concepts provided by ANT. My analytical approach requires us to establish the applicability of ANT’s concepts before using them (I do that in Chapter 2). Of course, I am not able to operate without concepts. The point of the argument is that the development of concepts, especially central concept like ‘carbon’ (its ontological status), is the result of this book, rather than the starting point. I have pointed to four methodological bodies which influenced the analytical design of the study: grounded theory, ANT, ethnomethodology and Pierre Bourdieu’s work. To emphasise, this study does not aim to reconcile all these bodies of texts. Nevertheless, the analysis has been influenced by distinct methodological assumptions which all add, I argue, to the quality of the study (and not criteria drawn from quantitative social sciences36 ). I briefly allude to the range of these assumptions; subsequently I spell them out in relation to a description of how I actually conducted the study. Grounded theory influences This methodological approach shaped my research design in a distinct way. Grounded theory is about abstraction from and building theory of data collected in the field; it is not about descriptive accuracy but about theory generation which is freed from the particular circumstances, i. e. an ‘abstraction from time, place and people’ (Glaser 2002). I planned to generate data in the field. This data was to be qualitatively coded and first hypotheses were to be generated. Following that, I wanted to conduct theoretical sampling, i. e. focus on data generation in relation to specific categories of practices in the field. Accordingly the field study was to be organised in different phases of research with increasing focus on particular categories of practices. Analysis of data within these categories was to look at diverse cases including seemingly deviant cases. An analysis of a variety of, at the best, conflicting cases promises, according to grounded theory, a good chance to develop a theoretical account of what is going on in the field which transcends any particular situation. Rather than following 35 Consequently, I argue against the understanding, e. g. by Katz (2004), that the researcher may not define concepts in order to impose the researcher’s theories on the researched. It seems not applicable for my study. 36 Criteria from quantitative research are reliability, validity and objectivity. Flick’s (2007b) account of quality in qualitative research shows why they are not applicable for a qualitative ethnography. 50 1.4. Researching Corporate Carbon grounded theory along all this study, my take emphasised also that I needed to ensure descriptive accuracy to ground the theoretical conclusions of the study. Beyond Glaser’s (2002) account, the research design was particularly influenced by Emerson, Fretz, and Shaw (1995) who draw on grounded theory. ANT My study was shaped by ANT’s dictum to follow the actor (so well exemplified by Latour (1999a)). This meant that I expected to follow environmental managers around and see how they did what, follow the entities they engaged with and zoom in even further. The ontology laid out above suggests ‘the need for methodological humility. If the world is messy we cannot know it by insisting that it is clear’ (Law and Singleton 2005, 350). Drawing on earlier work of mine (Lippert 2010a, 76) I expected that using visualisation techniques might be of help to gain instructive perspectives on the relations on the field. Trying to visualise what is going on in the field can never show everything (for the best model of the world is the world itself); but particular graphs have been productively employed in ANT (Latour and Woolgar 1986). ANT approaches were to be used throughout my study. Ethnomethodological considerations Ethnomethodology especially influenced the fundamental orientation of my study and the opening question I posed: what do environmental managers do and how to they achieve it? Garfinkel (1967, 1) posits that members use the same activities to ‘produce and manage settings of organised everyday affairs [and] for making those settings “account-able”’. Correspondingly my study was based upon the general assumption that whatever members did, I was considering their actions competent. And the study was about, in a circular logics, what competent members did and how they made accountable to others that their actions have been accountable. Bourdieu’s concepts This study was also drawing on the concepts capital and field as laid out by Bourdieu and Wacquant (1992). Based on them the study was to reconstruct how members constituted through their practices a structure/field and simultaneously members’ rules and possibilities as well as the structurated mechanisms of how positions within that field could be taken and changed, how the structure could be transformed.37 With Flick (2007b, 41, 73) we can understand this combination of approaches as methodological triangulation (triangulation version 1). It means that a fuller picture of the phenomenon can be developed by studying the phenomenon through the productive facilities of each of these approaches. Methodological triangulation is one of the techniques through which this study’s quality has been secured. Method In the following I make explicit how I combined methods which generated data and how I analysed that data. My study of GFQ’s carbon accountants took place over many months with several periods of persistent observation. This approach 37 Compatible takes might have been as well Giddens’s (1986) or Crozier and Friedberg’s (1979); which I did not choose. And this book does not aim to theoretically compare their takes. 51 1. Introduction contributes to securing the quality of my analysis (Flick 2007b, 19). During these periods I was granted access to follow members and their activities. I focussed my observations on the EMS-Team, their shared activities and, more to the point, their relations to carbon. My field work has been structured into two phases: first I was trying to collect data on virtually everything. Later, I narrowed down what issues and themes to focus on – until I arrived at the focus which this book has taken. Throughout these phases I undertook fieldwork at several locations (a multi-sited ethnography (Marcus 1995) in principle, but the focus was on the HQ-based EMS-Team). During this work I have engaged with members as institutional actors. This means I have not followed them outside their work life (which renders this work not into a multi-integrative ethnography (Weber 2001)). Throughout the interactions I was not aware of me struggling intensively with members over resources (as Forsythe (1999, 9) describes it); rather, because my income was secured independently of the work for GFQ, I did not have to struggle over the economic positions within the corporation.38 Throughout field work I conceptualised the observations, in Hammersley’s (1992) terms, both as ethnographically ‘realist’ as well as ‘methodist’. This means that I was quite relying on my perceptions in the field as making available the relevant circumstances for members. For members accepted me as another member we can assume that my schemes of perceptions and action (which Bourdieu (1989) would call habitus) fit the field well. I was a methodist in so far as I structured both observations and analysis in a specific way, allowing me to orient the research process according to clear temporally constrained ordering prescriptions throughout the study. The following paragraphs serve to sketch these orderings (both the pre- as well as descriptions). (1) As mentioned above, first order data generation was focussed on fieldnotes. Emerson, Fretz, and Shaw (1995) explain: ‘[f]ieldnotes are accounts describing experiences and observations the researcher has made while participating in an intense and involved manner.’ These notes, of course, act as a filter (Flick 2007b, 80) for not everything can be inscribed into these notes. I used a relational note editor, Tinderbox, to author and organise these notes.39 I tried to leave these notes as intact as possible in this book – in that sense this book is full, as Mol (2002) would say, of snapshot-stories. I have sketched the body of notes above but I did not yet clarify what data I generated in addition to these notes. First to mention are 28 drawings of architecture of sites at which members practiced their work and of their interactions (e. g. gestures). Also independently of specific interaction I documented (shifting) office architectures of relevant members. I ‘collected’ and catalogued 281 artefacts of members’ work, such as flip-chart slides, and of their work context (ranging from wastes to PR materials they equipped their offices with). I prioritised data generation for the EMS-Team and then (in lessening importance) other Sustainable Development workers, the CSR unit, the office block and the total corporation. 38 Throughout this project I received a research grant. I also received a small wage by GFQ for the months helping them with ESDR – similar to an intern. 39 Tinderbox can be found at http://www.eastgate.com/Tinderbox/. 52 1.4. Researching Corporate Carbon In addition to this first order of data generation, the analysis process entailed ten further orders. The first two (orders 2 and 3) of them took place well in the midst of the field research period. The others followed subsequently. (1.5) This is serious and a joke: of course during field work I was already starting to analyse the data and tried to gain or keep oversight of what I was learning. There was so much, I could not keep track.40 I soon decided to focus on the quality of fieldnote generation and to leave second order data generation to a dedicated time period. (2) After the first phase of field work I retreated from the field (bodily anyway; I was staying in touch with members via one of GFQ’s ThinkPads). I coded all the fieldnotes and associated artefact descriptors in terms of, inter alia, members’ practices, members’ themes, their terms, their emotions, the entities they operated with, ESDR (and research reflections). 435 codes resulted. Coding has been conducted using the application Text Analysis Mark-up System (TAMS), developed by Weinstein (2006).41 (3) These codes did not allow me to gain oversight fast enough over what categories of practices I had identified. The solution was to organise the codes in terms of productrelated practices.42 Ten potential foci to zoom into resulted.43 I decided to focus my observations in the second research phase on six foci: (a) on carbon and the database ESDR; on the link between carbon and an (b) auditor, with an (c) NGO as well as with (d) a ranking agency; (e) on quantification of qualities and finally (f) on a schooling device they employed to train further environmental managers. These foci should be understood as a first move to theoretical sampling. I aimed to focus subsequent data collection on these entities. (3.5) During that phase I was, again, trying to analyse the relations in the field, to map practices. I mention this for these mapping exercises shaped my emerging understanding of how messy organisational realities of carbon accounting have been.44 (4) Towards the end of the second field work phase I started to outline how I would engage with an in-depth analysis of the data.45 Roughly this is also what I have done – albeit, the path was a bit more bumpy. 40 For example, right within the first week of in-depth study I tried to map the formal and practical work relations of those members I encountered. Figure A.1 (on page 557) shows the first structure I reconstructed. However, this did not represent the reality well enough, so, soon afterwards, I developed another representation, Figure A.2 (on page 558). 41 The highly programmable and open-source qualitative data analysis software application TAMS fit my needs throughout this project better than alternative softwares I tested for this research project. 42 I derived the categories of products directly from a list of all codes. Before that, however, I engaged in a mapping exercise. I organised the codes into a wide variety of code sets. As one could expect, members’ practices around products were not neatly separated. Rather, very economical, they allowed themselves to prepare several products by shared practices. As a result, when I organised data into code sets, these sets shared some codes. Say, the practices around the EMS Handbook (one code set) shared with another code set on the product environmental database (ESDR) the code ‘[ESDR] > handbook’. Therefore these code sets have been linked in a particular way. Figure A.3 (on page 559) indicates what I was up against. 43 Figure A.4 (on page 560) provides an impression of the structure of these foci. They were all linked to each other. 44 Figure A.5 (on page 561) is an example for this period. The example revolves around material now shown in Chapter 2. 45 See Figure A.6 (on page 563) and the corresponding explanation for details of this plan. 53 1. Introduction The in-depth analysis started with open coding, which was trying to capture various readings of the data. I engaged with the data by way of asking questions on members’ practices, terms and understandings, their assumptions as well as on my characterisations of what was at stake, which actants I found and how I described my research practices.46 Coding resulted in 1704, fine-grained, codes; each code provided an answer to a question. I used code definitions to use the codes distinctly. (However some of the codes surely overlapped.) Much data was classified through coding to relate to several questions. In parallel to coding I generated 89 memos, producing first theorisations grounded in the material. Consequently, I created groups of codes around both, members’ as well as analytical STS questions. For that I organised the codes into 67 ‘code sets’ (in TAMS nomenclature), i. e. groups of codes.47 Each code set was then linked to memos. I called specific combinations of a code set and memos a theme. (5) As a next step I selected which themes the subsequent analysis was to focus on.48 For that I arranged the themes into an even further level ob abstraction, called core themes.49 In discussions with peers I developed the decision to focus on ten core themes which I was able top group under the question: how did member’s practices configure carbon emissions? 50 (6) The next step was to analyse the data represented by this question (by retracing the abstraction ‘backwards’51 ). Trying to ‘completely’ map the actor-networks represented by my data (‘under’ the question I posed for this book) failed.52 Instead I scanned through all the data instances represented through the research question (this referred to 3,390 data instances), summarised each instance (with respect to the question which its theme asked) in a file which was to be the first phase of this book. Subsequently, I rearranged the summaries and made introductory notes on the new groupings. Through 46 This approach was inspired by Emerson, Fretz, and Shaw’s (1995, 146) questions. See Figure A.6 (on page 563) and the corresponding explanation for the list of questions I asked specifically. 47 In addition I produced ten lists of diverse classes of dimensions through which I could access the data. A list of the code sets can be found in Appendix A.3 (on page 562). 48 Again, I mapped the relations between code sets and codes in TAMS. The result was highly unreadable (both, in small as in Figure A.7 (on page 565) and in large (several metres wide)). 49 This process entailed transforming the TAMS map of code sets into a map in which I could link the sets to the memos (see Figure A.8 (on page 566), Figure A.9 (on page 567)) and group these themes into sets of themes, i. e. core themes (see Figure A.10 (on page 568), Figure A.12 (on page 570)). These graphs had about 174 nodes and 666 relations (edges). A human-eye-readable and, thus, usable version of this order of data was black-boxing the content of each core theme, showing, thus only the core themes and remaining code sets and memos which ‘did not fit’. The result can be seen in Figure A.11 (on page 569). 50 Find a visualisation of these ten core themes and the themes which they integrated in Figure A.13 (on page 571). 51 See Figure A.14 (on page 572) for an illustration of how the higher level representation are linked to fieldnote fragments. 52 I started mapping by way of using nodes as representatives of both, actants and their relations and drawing all the links. Figure A.15 (on page 573) shows a segment of such a map. While this is readable for this small number of entities involved, I was not able to handle a much larger number of actants and relations. Figure A.16 (on page 574) shows the extent of a resulting map for only a few observations from fieldnotes. One way to relate to this problem was in terms of time and schedule. In Appendix A.3 (on page 575) you can find a (slightly edited) reflection of mine on the decision-making process of how to analyse the data. I made frequent research process notes in text, image and video to make the research process development traceable. 54 1.4. Researching Corporate Carbon a process involving printing, cutting, distributing these summaries over several square metres, I was able to identify eighty data instances in fieldnotes which represented well the heterogeneity of the material (that is, these fragments included highly different types of observations and seemingly non-fitting cases).53 These eighty data fragments built the core of this book: after linearising the networky relations among them, I turned them into fieldnote extracts (which you find throughout the book).54 This allowed to analyse them in-depth and structure the book around their relations.55 While grounded theory demands a constant comparison of all data, this did not seem feasible; in practice I was only able to perform partial comparison, partial connections if I may draw for that on Strathern (2004). Vis-à-vis Flick (2007b, 32) I would maintain that honesty about the reality of partial connections seems to generate a higher quality than a (potentially even less than) superficial comparison of all data. In any way, the comparisons I enacted add up to another version of triangulation (version 2) for different cases have been compared for the same issue; and different issues have been used to triangulate an answer to the book’s empirical question. Throughout the book you find evidence of the comparative and iterative analysis at work: I use a sign, ‘ ↩→56 ’, to point to analysing, unpacking, material differently and, thus, a comparison enacted. (7) While analysing the selected fieldnote extracts in comparison to other fieldnotes or by zooming into the situation, several times I came across further artefacts or for explanatory reasons I had to dig in the situation’s history. Thus, at the order of data generation I was adding additional data fragments from the case repository to the elements through which my account of GFQ’s carbon emission practices was to be constituted. Following Flick (2007b, 42), this constitutes data triangulation (version 3) This newly added data, presentations, letters and the like can be understood as traces of past practices and processes within GFQ (Flick 2007a, 90). To analyse the data I employed partial mappings of the situations observed. These mappings accompany the discussions across the book. (8) The next phase of the book was shaped by bringing the analysis into a conversation with STS, accounting, ecological modernisation and sustainable development and CMS literature. I, thus, exercised a triangulation through theory (version 4). (9) The final book emerged in relation to imagined readers (see Van Maanen 1988, 25-33). It involved editing materials to render the argument more coherent. (10) This final book phase was also influenced by another order of data generation: a validation process. Through communicative interaction with core EMS-Team members I have been developing consensus over interpretations. This book has also been validated 53 An impression of this process can be gained from Figure A.18 (on page 577). Flick (2007b, 45) argues that high quality qualitative research should allow readers to understand the research setting in its own terms. I react to this call by providing as least edited fieldnote extracts as possible. Especially, this involves keeping signs of my temporal identification with the work setting that I researched visible; you will encounter occasional ‘we’s and ‘us’s in the notes. These collective identifiers refer to GFQ or EMS-Team members. In the analysis, I restricted ‘us’ and ‘we’ to denote the collective of author and readers of this study. For a list of extracts, see the appendix. 55 The relations of these extracts are shown in Figure A.19 (on page 578). 56 The following discussion is relating To itself; in Section 1.4 (on page 55). 54 55 1. Introduction through other techniques, such as peer debriefing and passing as a member.57 In addition I exercised a validation situation by way of sharing my interpretations with students of the field of environmental management.58 In conclusion of this account of methods underlying this study I wish to underline two relevant characteristics. As the mass of prior footnotes indicated, this method – and, especially, the intensive use of visualisation techniques – was an experiment, trying to make the mass of data manageable without giving up the claim that my account is based upon a systematic engagement with the data which I have generated. To me, at least, the visualisations have provided added value: they allow to exercise additional perspectives. Most relevant, however, is that the technology of mapping repeatedly brought to my attention missing relations.59 Following a clear methodic outline, the most recent version of my programme can be found in Figure A.22 (on page 581), was making sure that this study’s account is traceable to its ground. This seems to be, after all, one of the most significant demands on social scientific contributions. Ethics Most of the significant points on the ethics involved in this book have already been made. There is a technical point: I conducted the study in line with the International Sociological Association’s Code of Ethics (ISA Executive Committee 2001). Beyond that, I need to recap. I do not claim access to any superior or omnipotent perspective to members’ practices. Thus, rather than performing what Haraway (1991c) calls the god-trick, I intentionally do not constrain my voice to an invisible passive actor or an ‘we’ (even though I also use these voices); often I exercise reflective and reflexive writing; all over ‘I’ remind(s) the reader of my presence, in the book and in the field. However, as this book is likely to be read as about GFQ’s carbon accounting practices I try to not indulge myself in excessive postmodernist reflections. I engaged in very specific agreements on anonymity with informants. Our relations have been governed by a contract (me as an employee of GFQ) and in addition to that by subsidiary agreements (verbal and by email). I received consent by the core subjects engaged with in this book to study them. Some people asked me to reveal the identity of members or of the corporation.60 Besides the specific agreements with members and the Code of Ethics it is precisely the anarchist take, if I may rephrase Williams (2011), 57 See Flick (2007b, 16-18, 33, 35) for accounts of validation. Regarding passing as a member, I can report: my work performance was perceived as native and well enough, resulting in the effect of superiors asking me to work for a longer time in their unit. Thus, we may assume that my position and observations therefrom fit well into the field. 58 For an artefact of this exercise, see Figure A.21 (on page 580). I performed an analysis for the students to test whether they would accept the reasoning. 59 I should note that the visual inscription devices (Savage 2009) which I developed do not claim to be linked to Clarke’s (2003) cartographic work. Against Lynch (1991) I would maintain that the maps employed in this book add something to the text. 60 Indeed, my take has been challenged by pointing to the ‘banality of evil’ or what Platform (2010) calls the deskkiller, in German Schreibtischtäter : office workers have to be held responsible even if only they act-at-a-distance (Latour 1987). However, as problematic as I consider the particular carbon enactment practices covered in this book, I would not see them as close enough to the immediate fascist crimes as shown by Arendt (1963). Towards the end of the book, I explicate normative discussions. 56 1.5. This Book suggesting that not any, say, particular manager is the problem, but the relations within society. This book is not fighting with individuals but seeks to contribute to wider social change. From the same position I have to reply to Flick’s (2007b) concern with ethics and data protection. Yes, of course, most sensitive data has been encrypted; but the risk that so-called democratic police forces seize engaged researchers’ data seems, to me, higher than the risk of burglary. This excursus to formal defence shall end with the most important point: I engaged in repeated feedback and validation interactions with all the key members under study. The constructivist take stressing the contingency and fluidity of emission realities did not alienate practitioners. It primarily alienates those holding on to the belief that practitioners’ numbers can represent reality precisely. 1.5 This Book This book is concerned with carbon emissions. If a company says they emit 1 tonne (t) of CO2 e emissions then they have to construct this statement. They do not say they represent their carbon emissions as equalling 1 t; but, rather, they claim that their emissions really equal 1 t. The phenomenon this book is dealing with is how they construct such a particular reality. They would reply they employ a technique called carbon accounting. A study of what that means also promises being informative for engaging with similar accounting instances: environmental, sustainable development, CSR management accounting, scorecards or, even, national inventories. Therefore, what this book tries to shed light on, is not only how capitalist actors account for carbon but, wider, for environments. In that respect, this book is about a phenomenon classical for reflexive modernisation: a modern project of transparency formally committed to modernise the damaging dynamics of capitalism; studying carbon accounting is a post-structural project of radical reflexivity, questioning the universalising project of transparency. In order for corporate carbon emissions to exist, agents need to be equipped. I ask: how on earth do carbon emissions turn into existence if their emergence cannot be explained with a rational actor paradigm, EMT or with the reality of the emissions out-there? What this book provides is an ethnographic account, built on techniques of observation, writing and sociological analytical methods. Thus, this book is not about thought experiments by practitioners or academics on how carbon facts might or ought to be produced, it is not about presenting results from a survey and it is, not at all, about bashing workers. In this book I consider some new resources for thinking about, and acting within, the space of agents of ecological modernisation, their tools and environments. The first shift which this book proposes is located within feminist techno-science and involves reconceptualising the implicated and interwoven relationship between the social and the material with implications for digital quantification practices. The second shift which this book engages in is from ecological modernisation to a feminist techno-science57 1. Introduction inspired postanarchist sociology and involves rethinking the grounds on which control over environments and imagined Society-Nature relationships can be exercised. For that the book draws on ongoing explorations within relevant areas of STS, environmental sociology, CMS and anarchist theory, each of which is heterogeneous and distributed in themselves; and for none of these I can do full justice here. I hope nonetheless to trace out enough of the lines of debate – that these fields of research and scholarship are committed to – to indicate the generativity of employing these ropes and threads, specifically with respect to rethinking and creatively enacting prefigurative relations between humans and their social and natural environments. My account is grounded in a sustained period of field work and emerged in interactions with members of the field as well as with critical practitioners, activists and peers. The stories of my account travelled; to Amsterdam, Augsburg, Barcelona, Berlin, Cardiff, Göteborg, Graz, İstanbul, Lancaster, London, Singapore, Trento. Doing an ethnography is promising because it allows to go beyond the focus on entrepreneurs or intrapreneurs of the company. The method is involving also their assistants, their machineries, their drafts of techniques, their successes and failures. An ethnography of carbon allows to bring to light the invisible work required to keep carbon emitting. If that stopped, what would happen to climate change discourses? My concern is not to evaluate any particular quality of the numbers enacted in the company. Rather it is to show how the representational convention of carbon reporting (which give the reports distinctive advantages) also are sources of trouble for members, and, potentially of catastrophic consequences for social and ecological justice. This book speaks to three groups, scholars of science and technology, of accounting and to environmentalists of all colours, including corporate, NGO, academic and autonomously organised agents. Broadly speaking, this book is concerned with how globally distributed environmentalists are positioned to relate to the producers of carbon emissions. They relate to each other through intermediaries – accountants, corporate databases, greenwashing PR reports. Most carbon knowledge is knowledge utterly transformed on the way from the site where it ‘originates’ to the environmentalist. In public discourses, nearly no ethnographic data is available about carbon knowledge producers, corporate agents of ecological modernisation. Whether to inform regulation, consumption or direct action, environmentalists cannot, at the moment, be aware of the ground which they employ to make their strategies work, that is about knowledge of how carbon is actually woven. This book is about the lore and practices of agents at the centre of carbon knowledge production. It problematises the division of labour and the assumptions of who has what knowledge and who is entitled to what knowledge and control. Instead I provide a vision of mutual learning and partial translations. I hope this book will also be a point of crystallisation for environmental management practitioners, open-minded to changes at the rootstocks of ecological modernisation. Contributions The site ‘GFQ’s emission enactment’ built a great ground for my argument. As one of the largest companies on Earth, it allows us to empirically ground 58 1.5. This Book our understanding of the relation between Nature and natural sciences with finance and capitalism. This link was shaped by common objects, like spreadsheets or software by SAP. And this book involves an account of how GFQ optimised its practices to account for carbon. Thus, I believe, the case allows to not only enter sociological theoretical territories but also into the territories of greening capitalism. This book attempts to contribute in four areas: in STS, in sociologies of work, accounting and economy, in environmental sociology and environmental management studies as well as in qualitative research methodology. In STS, the sociology and anthropology of science, technology and society, of quantification and knowledge production my contribution circles around digital quantification practices. The phenomenon of digital labour (Burston, Dyer-Witheford, and Hearn 2010) and practice poses, indeed, reasons to elaborate the enactment of materials like carbon. Rather than posing the multiplicity of accounts of carbon as a scandal I take it serious. This leads me to retrace much of Mol’s (2002) work on multiplicity. I go one step further, I believe, in so far as I was never imagining that there was a singularity out there to which a singular account was to be linked. The notion of multiplicity which I develop exceeds Mol’s; for her a singular body matters and it is multiple. I argue that carbon, for corporate purposes, is data. Carbon-as-data in itself is a multiple which is enacted differently and in parallel in multiple practices. (As for molecules: if they are at all, the signified is outside environmental management practices; ‘real’ molecules do not matter.) I underline also that GFQ itself is multiple and, thus, multiple multiples shape the knowledge maps through which (reflexive) modernity must fail to navigate. Anarchist sensitivities add to our engagement with this situation. As an outlook, I explore how decision-making might be organised in a way which would prefigure socially and ecologically just relations among humans and with (not their ) environments. With this work in STS I do not simply challenge any particular numbers but, more profound, the situatedness of these numbers. This constitutes the ground to contribute to sociologies of work, accounting and economy. For quantifications are among the foundations of modern control, the situatedness of quantification results in situated control. In a circular organisation of accountabilities between GFQ, their auditors, standards, ranking agencies and an NGO, I show how ‘being in control’ and ‘rulefollowing action’ is successfully performed by members: how members and collective actants manage to stage numbers as controlling realities. This challenges not only intraand interorganisational potentials to control but also destabilises the ground on which market-based policy instruments are positioned. By introducing the situatedness of numbers and control to environmental sociology and environmental management studies I question not only the grounds of ecological modernisation and EMT but also the realist foundation of Mol’s (2010) environmental sociology of networks and flows. In a manner of hypothesising, consider this case: in some closer or more distant future, an institution like the IPCC would figure out that environmental realities on Earth are inextricably stuck in a trajectory of climate 59 1. Introduction chaos. At the same time the World Business Council for Sustainable Development (WBCSD) would state that the IPCC must be overstating the seriousness because, WBCSD’s member organisations report that business has reduced emissions globally by 500 %. This book shows how such disparate realities are not only possible but likely to occur within the current configuration of actors, regulators and environments. Thus, I establish how the practical discourse of ecological modernisation allows to not engage with problematic issues, like core business or providing useful accounts for society about corporate greening. Ecological modernisation is discursively stabilised as an idea while its effects are not those which are claimed by ecological modernisation. Carbon market governance might provide bold statements of the reality of carbon emission reductions. In realist terms, the carbon they are dealing with is not closely linked with Nature. I show that engagement with that link is not in the interest of capital. Finally, by way of not jumping from this introduction right to the conclusions, but by making visible how my analysis forms the emerging understanding, this book contributes to methodological concerns of qualitative research. I show in detail how members’ practices are situated within discourses and how they reproduce, sustain and shape discourse trajectories. The book, thus, links members’ practices with macrophenomena, like ‘the market’ or ‘sustainable development’. Organisation of the Argument Rather than dedicating a disjoint discussion to theory and method after this introduction, I work iteratively. In each of the following five chapters I mobilise theory that helps me interpret the data or suggests fruitful methods to direct my inquiry. Interspersed are four interludes that establish relevant signposts along the paths of studying GFQ’s emission practices. Chapter 2 (entitled Data Construction in a Basement) investigates data construction practices on the ground. It traces the work by agents of ecological modernisation in a GCE and suggests that practices are to a distinct degree under the control of these agents while they are oriented to prescriptive statements by superiors. Empirically, this chapter engages with three of the fundamental practices on which carbon emission statements rely: the correct classification of data, the correct quantitative engagement with data and the ability to command data practices. Rather than simply showing that, as many might expect, data handling practices were not correct, I argue for their necessarily generative character which socially and materially construct carbon emissions. In that account, correct or false numbers do not matter as much as the point that practices establish how data turns into carbon matter. At a methodological level, this chapter introduces the reader to how an ANT account of knowledge practices can be performed. This involves pointing to some qualifications of how I bring ANT to bear in analysing fieldnotes. Sceptical readers will want to challenge that my account of practices observed in the subsidiary can be generalised to the level of the multinational and, indeed, to ecological modernisation and sustainable development. In response, Chapter 3 (entitled Translating Data Into Sustainability) provides an analysis of how, materially, GCEs, 60 1.5. This Book GFQ’s HQ and the global discourses of climate change, environmental management and sustainable development hang together. For that I trace how these discourses have been translated into the company, reproduced, shifted and given direction to, and, finally, released into the public sphere in a consequential configuration. I make members’ categories, practices and artefacts within the corporation visible as clearly part of hegemonic techno-managerial carbon discourse. I show how, in specific translations, some concerns of sustainable development and ecological modernisation discourses were silenced within GFQ. This explains why these discourses are not apt to conceptualise what happened as part of members practices. The chapter establishes that ecological modernisation reality cannot be understood adequately through its own discourse. Through practices, GFQ managed to signal publics that they are in control over their emissions. However, internally, it was well recognised that they were ‘out of control’. Discourses of sustainable development imply that corporations are able to control their own conduct and that market forces can be used to govern corporate governance. Chapter 4 (entitled Situated Control Over Carbon) attends closely how carbon accountants’ work has sedimented in particular structures of their work places. With the help of Bourdieusian sensitivities, I inquiry into the possibilities of their, thus, structured work to be in control by anything or anyone. The chapter claims that the practical reality of control can be best grasped as situated control. I argue for this claim by way of presenting an analysis at five levels at which carbon emission enactment has been structured. Consequential for that enactment is that situated control allows for multiple diverging practices to be present in parallel. However, members still would maintain they aimed at putting GFQ’s carbon account to order. This points us to focus on particular ordering practices. Chapter 5 (entitled Achieving Order) approaches these practices with ethnomethodology as an issue of practical achievement and doing of accountability. The study of methods and practices by which members attempted and managed to order their social environment and, in effect, also the corporation’s carbon emissions points to four core findings. Members routinely used their agency to adjust data – either by changing quantifications or by changing definitions. For all these practices were struck with obstacles and friction environmental managers tried to repair and optimise all kinds of elements, human and non-human. Whilst automatisation was always a grand hope, members practically resorted to insert human bodies into the carbon accounting machinery as creative problem-solvers; the EMS bet on these bodies to repair the near-breakdowns caused by techno-managerial configurations. After all, yet, disorder remained. Carbon data was shifting and uncertain. Consequently, members established ways to work-around and accommodate the messinesses they were confronted with. Chapter 6 (entitled Carbonscape) reconceptualises the ontology of carbon and proposes to rethink carbon from a punctualised fact to a fluid landscape. The notion of carbonscape, then, is the result of drawing the discussions of this book together, arguing that the enactment of carbon can well be conceptualised by way of turning to five key 61 1. Introduction perspectives: boundaries, materiality, multiplicity, temporality and scale. Rethinking carbon through these lenses serves to grasp how members’ practices constrained and made possible carbon. I argue that modernist oligo-dimensional accounts of carbon can only fail to consider the pattern of phenomena emerging from members’ distributed tactical but always messy engagement with carbon. Politically, this means that any intervention targeted at sustainability needs to address practices in this multidimensional socio-material universe. And it means that any responsible account of carbon emissions ought to engage with those elements and issues which did not fit into the picture. Silenced practices, masses, effects floating and taking place within carbonscape need to be detailed, in response to this reasoning. In conclusion of this book’s argument I draw together what it means to discuss the enactment of environments. For that I focus on the conceptualisation of apparatuses by means of which environments are enacted. This analysis indicates that enacting environments is inherently and constantly political and that such enactment work is only appropriately interpretable and accountable for in situated contestation of that situated enactment. Doing carbon, I generalise, is an instance of enacting environments. And enacting environments is systemically a troubling activity in need for political reconfiguration. For this political implications of the argument I turn to normative scenarios and investigate how useful they may be in allowing democratic engagement with the enactment of environments. I show that none of the hegemonic – neither markets, the state, environmental management systems nor the professionalisation of agents of ecological modernisation – is promising. Alternative ways of relating humans and environments, I propose, may be imagined by looking at resources at the intersection of feminist techno-science and anarchist political theory. 62 Interlude I Contested Carbon This interlude establishes a hint of HQ mentality. Mostly it will puzzle the reader. That is why this interlude is located here. Within only a couple of pages we find that GFQ’s capital practices are linked to carbon accounting. GFQ worked on the premise that climate change is taking place. Actually, the corporation was publicly outspoken about that; they appeared on stages in the policyarena (e. g. around the Kyoto Protocol Follow-Up conference at Copenhagen in 2009), portraying their corporation as a so-called good citizen, caring for the environment and looking into climate change issues. This book is not about PR. Therefore, I have not concentrated on publicly accessible trajectories of GFQ, but rather on their internal practices and how the effects of their practices surfaced to wider publics. I introduce the empirical discussion of this book by running through moments of strategising and contestation among GFQ’s CSR staff and the EMS-Team. I set out by linking climate change to GFQ’s core business – capital. Subsequently I turn to a discussion of a specific technique of carbon emission reduction, that is, offsetting. I conclude this interlude by showing how the prior narratives link to the concern of this book, i. e., carbon accounting. Economising climate change Nearly everybody was present; the room was crowded. Some of the young males have been sent off to get further chairs. The air was sticky, torrid. We are at the outset of a team meeting of the CSR Unit. Victoria and her boss, Herbert McJames, discussed the natural science of climate change: what are the tipping points? Are the collapse of the Gulf Stream and loss of sea ice in the Arctic all related to global warming? What is the role of fresh water in this? Soon afterwards, three guests entered the meeting. Working at the fringes of the CSR Unit, they were involved with GFQ’s insurance section and their climate change policy. They introduced the CSR team to the economic reality circulating around that policy. It was made clear to the meeting, GFQ’s engagement with climate change can be reduced to a fundamental question: 63 Interlude I. Contested Carbon Field Note Extract I.a (Climate change policy: trading off ) ‘What do we save’ if we put policy into practice – in contrast to what do we pay if we have to compensate damages by hail? GFQ is perceived as a natural partner in the politics of climate change mitigation and prevention because we are able to save money. In this meeting, climate – the atmosphere – was construed as an object in a rational process of trading off: what was cheaper for GFQ – mitigating and preventing climate change or paying for damages. It is clearly recognisable already now: for GFQ, the ‘right’ approach to climate change was depending on numbers. Relevant for them were figures representing two alternative sets of economic costs. Losses GFQ’s insurance section would experience as a consequence of climate change versus the costs GFQ would have to bear when they would address climate change as a corporation. This moment can be used for hypothesising that a widely shared assumption about the nature of capitalism holds also for GFQ: within the firm the regime of the ‘sayable’ is driven by cost considerations. Additionally, it was considered likely that publics would perceive GFQ’s new role as trustworthy. Knowing that publics frame GFQ as a capitalist entity, they assumed, a policy which would promise savings would be considered as a genuine interest of the multinational. When the meeting continued, they pointed in more detail to the implications for the insurance branch of GFQ: Field Note Extract I.b (Climate change: investment as solutions) [...] one can ‘technically’ relate to climate change: by excluding specific risks or raising deductibles. [...] In general the following is valid: ‘insurability [...] is questioned’. One option to that is a technical solution. Another option are green investments. Frank Jones: ‘thus, we require a definite framework’, to be provided by politics, in order to know whether it is worth to buy certificates and reduce carbon. Victoria: the ‘charme’ of an investment is that it behaves like a normal investment. These statements elaborate the beginning of the meeting: they considered it rational to not ignore the accumulating effects of climate change, i. e. to not sideline the issue to merely legally ensuring their position against the immediate effects of climate change. Rather, they wanted to perform action against climate change. The preferred means for such action were green investments because the latter could be dealt with just like any other investment – which was the core of GFQ’s business practices. Additionally, to be able to move within the realities of climate change and carbon economies, the corporation needed definite fixed points in the political-financial realm. The latter would involve, e. g., definitions of carbon, a predictable price structure in the VCM and a legally reliable frame of how carbon calculations were to be conducted. In GFQ’s approach to climate change, the essence of this problem could be extracted from natural science and contained in micro-economics. Climate change was translated into an informational problem: what are the figures? Contesting carbon offsetting I now turn to how micro-economics was rendered complicated in the reality of the EMS-Team. When I started my work for GFQ 64 officially, I have been introduced by Victoria to all the significant issues. One of the key coordinates for their work was GFQ’s emission reduction target. Unfortunately, different authoritative voices of GFQ ventured varying targets for the reduction of carbon emissions. Field Note Extract I.c (How to accomplish carbon reductions?) Meeting the goal of carbon reduction is difficult to achieve, Victoria told me. Until recently, the aim was 25 % reduction of GFQ’s emissions. However, now, that amount is to be doubled. Thus, the board of directors wants us to cut emissions by 50 %. Victoria suggests: ‘we won’t manage’. Those emissions, which are not reduced, will have to be offset. The reduction of carbon emissions can be created through two strategies: on the one hand, emissions can be physically reduced (which I explain in a moment). On the other hand, the corporation can buy certificates which would offset the emissions caused by its practices. This means emissions of GFQ can be swapped against negative emissions offered on the VCM. These negative emission could be sourced from a CDM project; GFQ was neither legally bound to offset at all nor to buy CDM CERs if they were buying. Thus, GFQ could source negative emissions from any other project as well. In fact, GFQ had the policy to only buy negative emissions if they have been produced under the so-called Gold Standard. I zoom into this standard at a later moment. Fundamental to offsetting was, of course, the willingness to buy negative emissions at all. This willingness, however, was contested. We switch into a strategy discussion of the EMS-Team. Next to Victoria, two other participants were present: Elise and George Kruger. George was an internal consultant with many years experience on environmental issues within GFQ. The discussion revolved around a service which GFQ offered to its customers. It was called ‘green mobility’. When buying this service product customers were offsetting some of their personal mobility-related carbon emissions. GFQ’s partner, GGCA, demanded that the corporation follows a strategy favouring the ‘actual’ reduction of emissions. In that policy, second to these imagined ‘actual’ reductions was substitution of carbon-rich consumption activities by less carbon intensive consumption. Only as the last option, did the policy accept offsetting. Hence, ‘green mobility’ was supposed to not be used; it was not considered an ‘actual’ reduction. Later, their discussion was rendered more complicated by this characteristic of GFQ’s Environmental Management System (EMS): GCEs were allowed much freedom in their environmental decision-making. Field Note Extract I.d (Nature of offsetting) Elise suggested to reduce emissions through ‘green mobility’. Victoria opposed this immediately. The discussion with GGCA did show that we should not do this ‘at full-scale’. George replied: this ‘is irritating me a little: [...] if certificates do not save anything’ this constitutes selling of indulgences. We do not want this. However, he had a look at where they money goes to. People ensured that the projects work out properly. George was convinced: a ‘solid process’ is in place at these offsetting projects, guaranteeing that these emissions are actually saved. Z discussion at Section 4.3, Section 6.5 65 Interlude I. Contested Carbon The approach of George indicates that offsetting carbon emissions through CDM (or equivalent) projects can be considered as the right and effective approach to get the atmosphere right. However, Victoria pointed out that the NGO would not like GFQ to use much offsetting. How were members dealing with these contrasting approaches? Field Note Extract I.e (A technical problem?) George argued: it ‘is only a moral-ethical problem, isn’t it?’. We should separate this from environmental problems. ‘We should not link [the practical dimension] to norms.’ Victoria replied: we want reduction to take place as much as possible. Only afterwards offsetting should take place. George: ‘I want to move away from the debate about norms. [...] We have a technical problem’: if a GCE plans buying a new heating system – at this point a debate about norms is appropriate: buying green electricity or a better heating system? We should not ‘demonise’ offsetting. The issue at hand is not whether or not we use offsetting, but rather: when do we start with it? Interestingly, while George implied a conflict over the normativity of GFQ’s approach to accomplish their carbon emission reductions, Victoria referred to the external authority, the NGO. She then elaborated her appeal to authority and included a link within the argument to the responsible member of the board of directors, Ivo Hampson. Z discussion at Section 4.3 Field Note Extract I.f (A political problem?) Elise contributed again: ‘green mobility’ is our ‘own product’. Why shouldn’t we use it? Victoria immediately: GGCA ‘is our partner’ [...]. That’s a [four million dollar] partnership. Victoria, George and Elise were having quite a heated discussion. Victoria: ‘GGCA articulated precisely: no.’ Both, Ivo and GGCA communicated: reduction rather than offsetting. Elise and George: Employing our own product indicates our credibility. George: ‘if there is a global problem’, one needs to solve it globally, rather than locally. Of course, one should not merely buy a way out, George said. Finally, by evoking the mighty board member, Victoria successfully established order: even though George and Elise disagreed with the boss, they accepted that reduction should be carried out before buying negative emissions. Victoria, in this discussion, performed a counter-hegemonic stance to emission management. This is of interest because it may provide the analyst with some hope that not the complete environment is turned into a simplistic economic calculus. While this hope will accompany us throughout the book, here is the dominant discourse in plain text: Hoffman (2010, 296) agrees with George as a matter of fact. ‘The coal burned in Ann Arbor, Shanghai or Moscow has an equal impact on the environment we all share.’ Carbon management If carbon emissions were to be reduced, the corporation’s approach to that was to quantify its emissions and reduce the latter. Who was responsible for these quantifications? Environmental managers. Early on in my fieldwork: Z discussion at Interlude IV 66 Field Note Extract I.g (Carbon managers) Dieter emphasised again and again: ‘They are actually not environmental man- agers, but rather CO2 -Managers’. The only job is to save CO2 . He reconsiders and adds: this is about reducing CO2 emissions, save costs and deliver data for the footprint. Thus, these agents were supposed to reduce emissions and by that costs – and in for preparing reduction of costs, they had to deliver data about the environmental effects of their GCEs. The actual reduction of emissions would take place through so-called ‘measures’. For instance, these included buying carbon neutral car fleets or hydropower. However, how much was to be reduced, eventually, depended on how much emissions GCEs had in the first place. Thus, GFQ also needed to establish the amount of their carbon emissions. In a swift movement we covered some ground between climate change and GFQ’s environmental managers. The HQ wanted to produce a carbon footprint. With Latour (1987, 185) we may ask the defenders of GFQ’s carbon footprint knowledge: ‘what are their proofs, who are their witnesses, how is the jury chosen, what sort of evidence is legitimate?’ The answer lies in Dieter’s statement: GCEs’ environmental managers’ job is to provide data. That data is the proof the HQ would always resort back to. The subsequent chapter opens the black box of that data. By that I establish what delivering data to the HQ – into the hands of Dieter, of Elise, Victoria, mine – meant. 67 CHAPTER Data Construction in a Basement Data is constructed. Get down to earth: visit the engineers, working in the basement of office blocks. Find how carbon takes its form and you will ‘see’. This chapter draws out a constructivist understanding of carbon emissions. To start our investigation, I provide a detailed account of the underlying practices at GFQ that were part of the construction process of carbon emissions. That is, I turn to the practical groundwork of constructing the data which was to be translated into emissions. Taking place at a subsidiary (I call subsidiaries GCE), the practices presented in this chapter introduce us to members, their means of production as well as their relations to the making of data. Empirically, this chapter reacts to Burritt, Schaltegger, and Zvezdov’s (2011, 89) concern that carbon accounting research has no access to ‘how climate-related information is input [and] who provides the input’. More precisely, we engage with near-first time encounters of subsidiary staff with the data collection and classification approach as devised by GFQ’s HQ for the Environmental Management System (EMS). Ethnomethodology suggests that it is useful to follow first time users of classification in order to study the disclosed work required to make the system work. Hence, I focus on users’ practices of classifying data and entering it into the EMS’s central data collection database, ESDR. Unbracketing the practicalities of the encounter between the request for data and the GCE’s staff promises to reveal key characteristics of the data emerging from that very encounter (Mol 2002, 163). The specific take I employ to study and discuss the practices at this GCE is using actor-network theory (ANT). Following Bowker and Star’s (1996) thrust, if an organisation is particularly good at delegating and distributing, ANT is well positioned to trace and reconstruct how activities have been coordinated. Thus, in this chapter I argue two points: (1) the configuration of activities delegated and prescribed by GFQ’s HQ to its subsidiaries are decisive to understand the quality of emissions (and that latter were simultaneously rendered ‘dis-coverable’ through the subsidiary’s provision of consumption data); (2) an ANT-inspired analysis of the material is fit for the job of critically engaging with the stuff of which carbon is made of. 69 2 2. Data Construction in a Basement Here is a beginning of the story: my boss, Victoria, had asked a Western Asian country’s GCE to welcome me in order to support my study as well as to ensure that the GCE participates in the data collection process. The latter was key to the EMS: from members’ point of view, without data, the EMS did not have any base to know GFQ’s emissions. Specifically, the aim of my visit to this GCE was to learn about how a GCE environmental manager would gather the piecemeal data about the environment. Reconstructing the practices yielded three modes of engagement with data. Consequently, this chapter is organised into three sections, each of which serve to allow appreciating the hard work undertaken at the subsidiary: this work allowed that data to exist and to be available for being subsequently employed at the HQ for calculating carbon emissions. First, we trace how staff classified specific kinds of consumption related artefacts, such as invoices, into environmental data categories. While reconstructing the classificatory practices I establish that an ANT take on analysing the construction of carbon emissions is apt. Subsequently, I turn to the decisive quality of the way how data was treated – as quantities. Hence, the second section looks at what it meant to do calculations – visiting the literature of the performativity of economics and the emerging sociology of calculations. Finally, I visit a meeting in which decisions were taken. This allows us to clarify what was needed to command the carbon construction process. GFQ’s EMS was delimiting its scope of data gathering to five so-called Key Performance Indicators (KPI). In the course of this chapter I introduce and illustrate several types of these environmental indicators: in the first section I approach three of them – electricity, paper and water consumption; in the second section travel; and, in the final section, waste. For each type I reconstruct a case of data collection. In conclusion I draw out the generalisability of the findings, suggesting how to take a critical distance to the notions of consumption data and facts: data is not anyhow given and found but it is made; behind simple facts performed as number–unit–account lurk wide landscapes of heterogenous qualities and assumptions. 2.1 Classifying Consumption1 Thus, I travelled to the GCE. After arriving in the city and settling in for a couple of days, soon I was to visit GFQ’s local offices. Going by bus, down a large street, many people crossing busily in front and behind, I saw a huge sign of GFQ on top of a building on the right hand along the street. After leaving the bus, crossing a smaller street and the GCE’s forecourt, I entered through its revolving door and went to the information desk – clearly the point to go to. After recognising that we were not able to communicate in our disparate versions of English, one of the porters wrote on a small piece of paper something along the line of ‘I want to speak to Mister: . . . ’. I entered the name of this GCE’s head of the EMS, Simon Jacobs. The porter talked on the 1 Parts of an earlier version of this section have been published under the title Carbon Classified? Unpacking Heterogeneous Relations Inscribed Into Corporate Carbon Emissions (Lippert 2012a). The present analysis establishes an elaborated argument. 70 2.1. Classifying Consumption phone and, soon afterwards, Simon’s assistant came to meet me and lead me to the manager. He had been the addressee of my boss’s request to welcome me. As one of this GCE’s top officers he was occupying a large office, with a glossy wooden desk and several square metres of windows at the top of the building. Early in our meeting Nick Xi joined us and presented a list of numbers to Simon. Later on I learned: Nick was the office site’s head engineer. Simon had asked him to collect the environmental data which the HQ was seeking. Subsequently, Nick showed me around at the site, and, eventually, we went to his office, located in the building’s basement. The practical responsibility for environmental data collection, I soon came to recognise, had shifted with our way down, from the top of the building to the basement: from whom the HQ had responsibilised to an agent responsibilised at this GCE. It was at the basement that I learned about the practices by which this GCE came to know the data requested from the HQ. Interestingly, I came to know, Nick had only been ordered to gather that data a week before I arrived. Within the week, then, he had been running around, calling people, to acquire some data to deliver to the HQ and present me with. This section introduces three cases; each helps us to understand the range of elements required to be in place for the desired data to exist and the effects implicated in the classification of this data. Thus, we reconstruct how the numbers that Nick had later on shown upstairs to Simon have come into being. I start off by introducing a seemingly clear-cut case – a case in which all the elements were assembled in a smooth manner. Next, we turn to engage with a case of contestation: data was not naturally given but was emerging through practices of sorting out tensions. Finally, this section shows how the classification of data involves room for resistance in the midst of accounting practices. Electricity: A Clear-cut Case The windowless room was populated by round six work places. Nick’s beige desk was equipped with two land line telephones, a computer screen, mouse and keyboard. An opened guideline issued by the HQ prescribed the parameters of environmental data reporting. Some of the words were neighboured by scribbled translations into Nick’s native language. The wall behind his red chair was supporting (or supported by?) a picture of the national leader. He offered tea; and I asked where he got those numbers from (that he had earlier presented to his boss). As a response he got out a file folder. Opening the folder, Nick presented its content: Field Note Extract 2.1.a (Measuring electricity consumption) Electricity invoices of 2008 and 2009. He pointed out that the invoices include the amount of kilowatt-hour (kWh)s consumed by highlighting the data. Thus, he had seen the numbers, not only calculated. This presentation of electricity supply invoices indicates several key elements, which we need to keep in mind to fully understand their social and environmental implications. This GCE consumed electricity for which it had to pay. An invoice listed the amount of consumption for which the GCE was charged. By showing the invoices Nick pointed out that he relied on measured data of the consumption (rather than on what he 71 2. Data Construction in a Basement would imagine as a calculation; see below). Within GFQ it mattered whether data was designated as ‘measured’, rather than ‘calculated’. I was interested in this distinction because the HQ based EMS-Team was observing – what they called – ‘data quality’Z. In order to engage with this object I exercised, and jotted down, how to classify them with the HQ’s data quality filter. This filter consisted of a ranking: the EMS-Team considered measured data to be of higher quality than calculated data. The guideline, opened on his desk, was supposed to orient Nick: it included a hierarchical order of data qualities, depicted verbatim as Artefact 2.1.1. The object ‘data quality’ is central to the discussion in this book; therefore, Chapter 4 focusses on its existence. Level of data quality: 3 2 1 0 = = = = Data Data Data Data based on exact measurement based on calculation / detailed estimate based on rough estimate not reported Artefact 2.1.1: Data Quality When I had asked Nick about the source of the numbers, he got out the invoices. For him, the numbers printed onto the latter constituted an appropriate reply to my question. Picture an invoice. He had highlighted those data representing the kilowatthours consumed which had been billed for. Thus, the measurement had neither been carried out nor observed by Nick. Rather, he must have assumed that these numbers were pertinent to the data request by the HQ. Therefore, in my reading, it was acceptable for him to copy these numbers and present them as data to Simon. Using ANT, we are able to unpack the relations involved in this transfer of data from invoice to boss. The extract above provides us with a range of diverse actants having diverse qualities. Using a graph editor, I visualised them in Figure 2.1 on the next page.2 This figure serves both to illustrate the relations implied in this transfer of numbers as well as to document a perspective in which the elements appear as positioned at equal level, corresponding to ANT’s postulation of a ‘symmetrical and tolerant description’ (Callon 1999, 81). The following discussion, in consequence, analyses the figure, and results in a reconstruction of the power relations which does not presuppose a priori that humans, or entities like ‘society’ and ‘nature’ determine the situation. First, we encountered Simon who acted as the environmental manager of the GCE. His task was to report environmental data to the HQ, represented by Victoria. To do this he had asked Nick – an engineer – to collect the respective data. Thus, Nick started to check where he may get the data from. He found data in invoices and provided selected information of these to Simon. To legitimise the numbers presented to Simon, Nick pointed to the ‘facts’ shown on the monthly invoices. Following Latour (1987, 23) 2 This figure – as well as those below – indicates actants through the colour orange, a description of their relation in green. Actants which I have not directly observed are shown in violet, their relations into the field under study is represented in grey boxes. A rounded rectangular depicts human actors and organisations. I use octagons to point to (digital) materials and ellipses to informational actants. 72 2.1. Classifying Consumption Figure 2.1: Using amount of electricity billed for to speak for environmental effects a so-called fact is something which does not invite to question ‘ownership, construction, time and place’. Further, he suggests: ‘A sentence may be made more of a fact or more of an artefact depending on how it is inserted into other sentences. By itself a given sentence is neither a fact nor a fiction; it is made so by others, later on.’ (ibid., 25) In the case we are exploring Nick had enlisted the authority of the invoices to speak for the consumption of electrical energy. By that, he gave more weight to an element of the electricity provider’s statement while deleting contextual information from the facts, which enacted a dichotomy of content and con-text. Thus, he had screened the documents and classified the information into different groups. Nick quoted the class, quantities and the units – but did not refer to the author of the consumption counts nor did he investigate how these counts were constructed. The invoices, thus, were a carrier of a range of information, some of which was translated to the sheet which Nick had used to present the number to his boss. Latour conceptualises the context 73 2. Data Construction in a Basement of a statement as its modality (1987, 22). A modality is a qualifying and, by that, modifying statement associated to the former statement. Another take on Nick’s act of demodalising is to understand it as a form of purification (Latour 1993b). The invoices’ offers of text were processed by Nick: first by purifying its information, resulting in a single number for each invoice, its unit (kWh) and class (electricity consumption generated by hydropower); then by adding the numbers up and copying this to Nick’s sheet which he used to collect the environmental data. The effect of purification was that traces of the hybridity (in this case e. g.: authorship, materiality, historical position and economic relation between provider and customer) of the entity in question were silenced. GFQ was asking data collectors, i. e. Nick in this case, to assign the numbers reported a value for data quality. Thus, later he continued altering the modality of the statement by assigning the number the data quality signifier ‘3: measured’, so signalling the HQ that the number has been read off a measuring device – or at least that it could be traced back to some form of document certifying the information. Throughout this process of altering the modality, Nick had to classify his own data access practice according to the list of data qualities. He had linked the information by the energy provider to the category measured, thus adopting this category to qualify the prior relations between provider and the invoice’s numbers. Following Latour (1987, 109-121), the set of relations performed by Nick can be understood as translation. By using a single number to represent the electricity consumption Nick created a fact which all other actants had to use. No actor attached to GFQ’s EMS who had as a goal knowledge about this GCE’s electricity consumption would look again at the invoices or look out for alternative facts. Nick’s translation to the sheet made his fact indispensable. He became the macro-actor on top of the black-box called electricity consumption (Callon and Latour 1981, 284). This has politically non-innocent implications. Any translation is rendering two statements equivalent and, thus, shifts their meaning.3 Law (2007, 4), therefore, points out that translation also implies a dimension of betrayal. According to this then, analytically, Nick and the electricity consumption number can be understood as allies. He allowed the energy provider’s statements to circulate even further in the world and used his power to translate the invoices into a single number. And the invoices allowed him to translate some of their elements such that he could consider his job done. No doubt then, Haraway’s (1991c, 195) finding that ‘[t]ranslation is always interpretative, critical and partial’ can be confirmed: it was Nick making sense of the information which he transferred; he was only partially having access to the reality represented in the invoice and was only carrying over selected elements of the invoice; thus, he engaged in an act of deleting information which had the critical effect that Simon would be presented with a clear-cut figure without explicit authorship. How did this information gain the possibility to travel? Latour (1987, 68) uses the concept of inscription to rethink the material foundation for the mobility of information. 3 This equivalence is achieved for all practical purposes and is not necessarily assumed as given. The same holds, of course, for the translations undertaken by the author of this text. 74 2.1. Classifying Consumption The energy provider inscribed some selected information into the invoices sent to the GCE. An invoice, thus, can be understood as the ‘visual display’ of the electricity consumption. Latour refers to anything or any institution which provides such displays as inscription devices. Nick inscribed the electricity consumption number onto a sheet of paper, thus performing a further inscription device. Or, seen the other way around, Nick’s environmental data sheet transformed – as Law (2004a, 20) points out – the material, i. e. here the invoice’s information, into a purified form which was more apt to usage for the worker. The effect was simple enough: only traces of the originally rich invoices was left for the next level of translation. Most information inscribed in the invoices was not carried over by Nick’s application of this specific inscription device. Such information got lost, was deleted from further consideration, muted. The associated implicit claim was that the product of the translation represents the reality of the GCE’s electricity consumption. Latour (1987) points to two further competences of a statement which would strengthen it: combinability and stability. The invoices’ numbers were easily combinable by Nick; he had to perform the mathematical operation called addition. At the same time, these numbers were stable. Neither was there a reason to question them, nor did they seem prone to erroneous copying. This discussion showed that the simple administrative practice of adding up numbers provided by invoices in order to represent the total amount of kWh billed for has to be understood as consisting of constitutive and emerging relations giving rise to a specific configuration of power: among these was Nick gaining the power to construct, a fact which was not questioned in itself anymore.4 This analysis, thus, suggests that the agent who identifies, classifies and translates data has much more power over the forming of the informational basis for constructing carbon emissions than the invoice itself. The latter could not force Nick to translate it in any specific way, it seemed. While this first case did not question the class onto which the kWh consumed were to be mapped, next we look into Nick’s practices and the relations he co-performed when he and Simon had to deal with the question of what type of environmental data some numbers were related to. Paper: Classification Contested When Nick had joined Simon in his office, I introduced them to the database which GFQ used to collect environmental data. It was called ESDR. While I felt that I should introduce them to the details of the accounting technology, Simon reacted by stating ‘in ten minutes we will be done’. I had just started to show the structure of an ESDR form to enter paper consumption data, when Simon interrupted me and decided to use this category as an exercise or showcase. Whether or not he actually intended to show that his GCE had no problems with entering data, Simon effectively employed the KPI paper as an example. 4 To be more precise: the fact has not been altered during my field work and I see no reason why it should be; except, maybe, if this text sheds light to its conditions of existence. 75 2. Data Construction in a Basement Field Note Extract 2.1.b (Paper case) Simon was getting up and fetched a 500 page package. He asked: what paper is it? And he looked at the labels but could not find much information. Then Simon suggested looking at the producer’s website. They failed to identify the required information. After a short while, Nick claimed it to be recycled paper. Simon replied: no. This isn’t recycled paper. We will recycle it. [. . . ] Finally, Nick agreed. Paper does not necessarily equal paper. Different production processes and kinds of disposal are associated with different emissions. GFQ, thus, differentiated several types of paper, e. g. chlorine-bleached and recycled paper. Simon and Nick undertook an act of classifying the paper they normally used. For that, Simon briefly went to his secretary, getting a package of paper. Figure 2.2: Establishing the Class of a Package of Paper The very aim of Simon’s move to get the paper package was to relate a local object to the database, central to GFQ’s environmental universe. To translate the paper package existence into ESDR, they had to determine the package’s class. It was this class which 76 2.1. Classifying Consumption was at stake in this situation, illustrated by its central position in Figure 2.2 on the facing page. Simon and Nick tried to ascertain the correct class of the package, choosing one of the classes offered on a list of available ESDR forms. Thus, a normative orientation was existing from the start: the local object was to fit the existing classification scheme. This likens the observation by Waterton (2002, 186) in her study of the classification of natures: she found that field ecologists are trained to fit their observations to existing categories. Troubles with lists of available classes and what they imply are a classic theme. Law and Lynch (1988) point, for example, to the problem of classifying entities for which no classes are available in the classification scheme or to the problem that the specification of classes cannot be applied by the agent who is supposed to apply them. An application of ESDR’s classification presupposes that agents check which class a consumption belongs to. The problem is that classification schemes do not work as scripts which would determine members’ classification practices; much rather – drawing on Waterton (2003, 126) – within the situation members have to improvise in order to determine any class. The way by which the scheme’s prescriptions can be replicated in the encounter with paper package is contingent on the situation. Latour (1987, 89) suggested that in science in action things are ‘defined by their performances [where] each performance presupposes a competence’ (quoted without emphases). In our case the material object had not been assessed by members based on some form of scientific or laboratory analysis, but rather, they questioned the object in two senses; the paper package’s performance was questioned with respect to globally and temporally distributed competences: first, Simon asked the object’s label to reveal its class, and when that failed, he suggested retrieving the required information from the producer’s website. Second, Nick enquired into the near future of the object and found it will be recycled.5 While Nick concluded, thus, that the object were to be classified as recycled paper, Simon opposed Nick’s understanding. Thus, temporarily Simon and Nick were allies to competing claims: they considered the paper package to be the case of different classes. Following Beunza and Stark (2008, 272), we may expect that members would make a case for their respective claims. So, how was this struggle resolved? Rather than using the object’s future as the indicator of its class, Simon implied, we should categorise the object based on its past. None of the two competing claims can be considered intrinsically right. Instead, drawing on MacKenzie (2009b, 26-30), we may conceptualise categorisation through finitism. A finitist take allows focus on the factors constraining the in-principle flexibility of classification-in-the-making. In this case, the constraining factor may be interpreted as the combination of formal authority of Simon in relation to Nick and the recognition of Simon’s better understanding of English language. In both respects Simon was able to exhibit control because he related closer6 than Nick to the sources of formal authority within the company and to the 5 A note for the Northern/Western reader: In this particular culture it was not self-evident what the English term ‘recycled paper’ conventionally refers to. 6 I recognise that ‘closer’ is not qualifying the relation in detail. Chapter 4 engages in depth with the power relations in the field and the relevancy of formal assignments of authority. 77 2. Data Construction in a Basement language in question. Thus, they moved on to inquire backwards in time. As they did not find any hint of the paper having been produced using recycled materials, they finally concluded the act of classification by deciding for the class ‘new fibres elementary chlorine bleached’. In effect, having this particular paper package linked to the latter class of paper emerges as a result of specific (re)cognition and inscription practices. Hence, to classify elements, they need to be rendered as knowable, made the object of enquiry and the result needs to be attached to a category. Yet another effect of their interaction is one we may call enculturation. With Latour (1987, 201) we are able to think of disputes over classification claims as signifying the boundaries of a culture: ‘“culture” is the set of elements that appear to be tied together when, and only when, we try to deny a claim or to shake an association’. In our case, Simon performed a boundary of a culture offering Nick to join. The former tied together the term recycled paper to a temporal view, i. e. determining whether a given paper object is recycled paper through its past. Under the circumstances of capitalist work-relations, Nick probably did not have much of a choice as to whether to accept his superior’s offer. His application of the classification scheme was disciplined in the situation. In this new culture it is possible for a Nick to be green by means of switching from chlorine bleached paper to recycled paper. Porter (1995, 42-45) points to the significance of such effects. An accounting category may transform business practices such that the company performs well in this new dimension. Finally, I employ this extract to emphasise a key assumption in the field: using paper, just like electricity, or engaging in any other consumption activity causes carbon emissions. Figure 2.2 shows how the conversation between Nick and Simon was related to carbon emissions. Here we approach the central concern of this book: the classification of the paper they used was linked to carbon emissions which they were about to construct as being caused by their paper usage. Depending on the class, different emissions would be constructed as the effect of this GCE’s paper consumption. This analysis suggests a number of effects: the particular choice for classifying the paper, Nick’s enculturation, the possibility that members encountered a new possibility to render GFQ green(er) and the impact on the amount of carbon emitted. All these effects can be considered political. Clearly, then, categories have politics (Suchman 1994a). The classification scheme disciplined its users and its users disciplined each other in shaping their relation to the scheme. The ‘correctly classified’ consumption object presupposes the discipline of users and of the objects themselves. Water: Classification Complexities The final case focuses on the mode of existence of carbon emissions. Above, we learned that to file an object it needed to be classified first. This implied giving voice to some modalities while black-boxing alternative ones. What happens if the available system of classification fails? This discussion opens a view onto the hinterland (Law 2004a, 27) of classification implications. The encounter between Nick and the class ‘Water’ allows us to unravel the implicit 78 2.1. Classifying Consumption assumptions and realities further: towards relating the practices at GFQ to external worlds of carbon fact production, in this case, Life Cycle Assessment (LCA). When Nick approached filing water consumption, he diligently gathered data from different sources of water consumption, including a well, tap water, and, a huge amount of drinking water. However, then I had to interrupt his endeavour: Field Note Extract 2.1.c (Water) 15:50: I told Nick that Elise was saying that drinking water is tap water. Nick asked: What shall I do? I replied: I can only inform you. I cannot decide. You have to decide. I told Nick that the HQ does not need the drinking water data that he had collected. But, I suggested, he could collect the data for his GCE. Then Nick was calling the canteen and cafeteria to ask how much drinking water they use. He added this information in the ESDR form for water drinking water. What had happened? GFQ used ESDR to collect environmental data from all over the globe. Members of the EMS-Team administered the database. Nick, as a local agent, put data into the database – by means of a form referred to as ‘task’, reproduced and anonymised as Artefact 2.1.3 on page 81. At the HQ, superior, in terms of the EMS-Team’s hierarchy, members were able to access the data. Elise was one of them. I had stayed in close email contact with her in order to ensure the data which Nick constructed during my visit fitted the EMS’s requirements. At 14:50 I had sent her an email differentiating the consumption of water according to different types of sources (a well, bottled water, tap water). At 15:01 she replied and dealt with the different classes of water. Her email provided, firstly, a quote from my original email to her, and secondly, an answer (a verbatim, anonymised extract): [My text:] They are having a) a well (natural water), b) drinking water (in large cans – in heaps), c) tap water. What is tap water – which account are we supposed to use? [Her reply:] Drinking water in cans is not included into the calculation, merely the water got from taps ([use the account] drinking water). Well water is natural water. This account is correct. Artefact 2.1.2: Email at 15:01: Drinking Water defined Key to understanding this email is first of all recognising the means which Elise and I drew on: accounts. ESDR can be understood as an effect of an accounting and audit society (Power 1999). Within this, ‘environment’ is managed through an accounting system which associated each consumption data set to an account. Such an account was characterised through an ‘explanation’, drawing out, i. e. defining, the class – indicated in the database form, Artefact 2.1.3, as mark ‘A’. Elise, interpreting the definition, thus, specified that water received by means of a system of pipes, rather than by means of transportable cans, had to be filed in the account ‘drinking water’. Water received through cans was to be excluded. Yet, after I informed Nick at 15:50 about Elise’s filing prescription, he decided to collect the data on bottled water anyway. At this time it was significant that in Nick’s culture water to be drunk was normally not obtained through a system of pipes, i. e. tap water, but through bottles and cans. This understanding 79 2. Data Construction in a Basement was not shared between Elise and Nick. Their approaches to how elements were to be associated with the term ‘drinking water’ exhibited a discrepancy. Culture is, thus, key to understandings which are drawn on in the practices of classification. Any definition needs to be translated into the intended practice to produce the targeted effect. However, actual practice is informed and hold together culturally which may contain HQ prescriptions but is not determined by them. While Bowker (2000b, 659660) generalises the effect of convergence – realities being fitted into the classification scheme, disciplined by it – and claimed that, because of this effect, databases can be expected to represent political economy well, Nick’s treatment of bottled water questions challenges Bowker’s interpretation: similar to Waterton (2002, 188), we find that, actually, an effect of convergence is precarious, contingent on how entities are translated into the database. Another key element inscribed in the ESDR data entry form was its relation to a carbon conversion factor (mark ‘B’). The respective factor stored in ESDR for the drinking water account was 0.3747 kilogram (kg) of so-called scope-3-emissions for each consumed cubic metre. This refers to indirect emissions caused in the life cycle of a product. Where did the number come from? GFQ was referring to the voluntary industry specific environmental reporting standard called VfU. This is a standard organised by several large international banks and insurance companies and is freely available to everybody.7 Study of this document reveals that the number displayed is the sum of two factors, 0.283 and 0.0917. The former factor was presented as being derived from the emissions associated with a waste water treatment plant in Switzerland and the latter factor with the production of drinking water. These factors are prime examples for Latour’s (1987) immutable mobiles. The documents refers to two sources, ‘Althaus (2003)’ and ‘Doka Gabor (2003)’. These authors managed, by reducing their local investigations to factors, to make their judgements mobile (Almklov 2008, 876). The factor as the result of their calculations was not entirely isolated. Doka Gabor also made accessible information on the methodology of data construction. However, as he let me know, more details associated with the factors were only accessible commercially, at 1,800 Euro (EUR) plus taxes. The detailed contestation of carbon factors is out of scope of this book.8 Following Mol (2002, 49), I do not question that somehow such records may be related to the bodies they refer to: like a patient record is very much the effect of heterogeneous elements like buildings, knives, administration work as well as a body, carbon emission factors may be assumed to be related in a complicated and contingent web of relations with the realities they aim to represent. If authors of carbon emission factors of, e. g., a water treatment plant, however, were to include the emissions associated with the financing of the plant, its insurance or investments, then, the authors should be recognising that they are relating back to their own work. These are intransitive relations (ibid., 120-121): GFQ’s carbon report included the factor 7 8 80 I discuss this standard in more depth below. See Section 4.3. For that see, for example, MacKenzie (2009a). 2.1. Classifying Consumption • F) • B) • C) • E) • D) • G) • A) Artefact 2.1.3: Task for Water Consumption 81 2. Data Construction in a Basement which should (if they were to take the emissions associated with the treatment plant’s finance world into account) include GFQ’s report. Mol emphasises that when looking at work practices, such intransitive relations can be found. Nick’s act of classifying water consumption as drinking water linked GFQ to emission factors which themselves should have been informed by GFQ’s carbon reporting. Doka Gabor presented himself as an LCA practitioner based in Switzerland and was aware of the contingencies of the facts he mobilised. This, at least, is the gist of his statements on ethics. His comment on epistemology (a verbatim copy from doka.ch/ethics.htm, accessed 2010-11-28) is particular relevant for our analysis: Contrary to popular belief an LCA study is never objective or provable. Subjectivity occurs in all phases of an LCA study. In that struggle I always take the side of the environment. With precaution I strive to make sure potential burdens are not lost from view. Artefact 2.1.4: Fragment of the Ethics Statement on Fairness This author of carbon emissions factors, thus, did not imagine his work as free of values and politics. The modality lent to his factors was one which invited caution. Whatever GFQ was doing in carbon accounting, the emission factors linked their consumption to carbon emissions. And this link was not regarded from the point of view of the author of the factors as a positivist relation to facts-of-nature. The specific configuration of Nick’s classificatory act, furthermore, was resulting in a significant constraint. While the carbon emissions assigned to drinking water production and sewage treatment were included in the factors, the covering of the additional emissions of bottled water (implied within bottles’ material and transport cycles) was uncertain in two ways: the emissions were multiply out of scope for GFQ because, first, Elise asked for excluding water from cans and, thus, emissions associated with this water. Second, this exclusion also led to not accounting for the emissions caused by the synthetic cans. Thus, while GFQ tried to account for its water consumption, their particular classification scheme’s design could not access the emissions associated with packaging of bottled water. Bowker and Star (2000, 304) draw on Haraway’s (1992) notion of monsterZ, referring to those elements which are considered exceptions to natural law, to emphasise how classification schemes necessarily imply that some elements do not fit and, therefore, cannot be clearly accounted for. GFQ’s classification scheme did not allow to account for this exceptionalised water consumption; the latter can, hence, be understood as monstrous. I mapped the water case in Figure 2.3 on the facing page; note the contested area. The highlighted (dashed black-boxes) relations were those performances which were giving the ultimate shape to what water consumptions were filed in the database, ESDR. This shape exists now as carbon matter. Law (2008, 10-14) refers to practices which perform such shapes as mattering, enacting matter. Nick’s practice matters because through it carbon matter was ultimately presented as emitted by GFQ. The reality perceived is shaped by effects of performances of mattering. Struggles over the real are inherently political. Thus, doing classification implies ‘ontic’ political action – politics 82 2.1. Classifying Consumption Figure 2.3: Establishing the Right Choice of Accounts to File Water Consumption 83 2. Data Construction in a Basement of what metaphysics of water come into being, politics of, understood with Verran (2010, 109-113), what is. This practical politics of making things matter did not require explicit debate about the ontology of drinking water – ontological politics (Mol 2002; Law 2004a, 162). This is no harmonious ontic ethics-politics, in which agents recognise ‘that different practices do different metaphysics’, i. e. that things can exist differently, respect this finding and make it explicit (Law 2010, 127). More precisely, Nick inscribed one reality of water into ESDR, rather than others, while he recognised that this implies different meanings of what drinking water is. Like other databases (Verran 2010), ESDR decisively constrained this ontic action. By my act in the field of asking Nick to decide about the inclusion of bottled water I tried to prevent too early a closure of the dispute. Rethinking this moment with MacKenzie’s approach to finitism allows a view of how an implicit decision necessary to carry out the act of classifying has gained the possibility to play itself out more explicitly. Thus, in Artefact 2.1.3 (on page 81) Nick hinted in a ‘Comment’, mark ‘C’, that he added 171 cubic metre of bottled water in the account. Thus, the 64 kg CO2 e emissions assigned to this water (however, maybe not those to the bottles) were, eventually, included in GFQ’s carbon emissions count. In that respect, Nick acted as a dissident (a phrase suggested by Callon 1999). That is, because classification is a practice, an engagement with classes opens up interpretative spaces which can be appropriated (Waterton 2002, 193). To classify, members have to draw or reproduce boundaries. This implies they can select meaningful things to count into the bounded area (Almklov 2008, 880). Nick could dissent from the classification scheme’s intention because resistance to given categories is possible (Suchman 1994a). Nick did not let himself, and through that the product of his translation (the reported amount of water consumption), become enrolled by Elise; a betrayal took place – and I was an accomplice. Elise’s act of purification was not completely successful. While she had the formal authority within the organisational structure of the EMS to interpret the definition provided in the form and to inscribe its application in a directive email to me, the relations involved did not lead to the prescribed effect. And, now, one might ask who or what was betrayed? Several actants are thinkable – among them the imaginary actant of GFQ or Nature. Analytically seen, the network of Nick, his data sources represented in various inscription devices, ESDR, me and our collective performance constituted an Obligatory Passage Point (OPP). Any water consumption fact of this GCE had to pass through the classification and translation filters of this network. From the point of view of the HQ actants this network would normally only be perceived in a punctualised form (Callon 1991). This resembles the punctualisation referred to by Law (1992, 5) who suggests that normally a television is perceived in a punctualised form; the network producing the TV effect remains behind the stage. In Figure 2.3 on the previous page we are more specific than in the prior figures: what we previously referred to as ‘Environmental Data’ we are now detailing as ‘ESDR tasks’, i. e. ESDR task-formsZ in which the data was contained. ESDR was a significant 84 2.1. Classifying Consumption part of the machinery (Lippert 2011a) in construing carbon emissions. The accounts were linked to conversion factors, which were used to calculate the amount of carbon emissions for the amount of water consumed. These conversions factors, as in the case of paper, were often specific for each class of material consumed, i. e. in this case differentiated for three types of water. The statements produced by the Swiss LCA practitioner Doka Gabor had been enrolled by an enterprise which was in charge of producing the nitty-gritty for the VfU standard. This case followed the chain of translations further than the prior cases and, in this course, came across an exciting modality of ethics, which has been completely neglected in the public statements of GFQ: an author of carbon emission factors pointing out that these factors should not be interpreted as an objective and neutral entity. Classifying consumption into classes, thus, does not link consumption to some independent carbon reality but to authors of yet another informational entity. ♦ End of Section ♦ This section discussed the work required to classify the consumption of – as GFQ would have it – environmentally relevant goods and services into GFQ’s central database. Following a subsidiary’s environmental agent, we came across a range of elements and relations required to be in place for the desired transmission of data to the HQ. With Bowker and Star (2000, 310), we could conceptualise the work observed as articulation work : this concept refers to managing overheads, i. e. those entities which do not fit the established categories of an organisation. This is the work of secretaries, parents, airtraffic controllers and nurses. Just like them, the agents entering data into carbon databases do invisible work. In Bowker and Star’s words, ‘articulation work [. . . is . . . ] work done in real time to manage contingencies’. This work matters because it is continually achieved by Nicks and Simons, agents working all over the globe. It is their work which gives rise to new facts about the carbon pollutions of the multinational. These facts still perform and can be expected to continue performing – as part of the corporation’s emission history tables. At GFQ, carbon emissions have been computed based on consumption data. In this section I have established three key points. Partial translation: consumption data was the result of partial and always interpretative translations. While data was found and translated, its authoredness was rendered invisible. Only because of this was it possible to present clear-cut, pure, facts. By cutting, the con of the con-text has been deleted. The human actant, Nick, seemed to have much power in translating consumption representations into environmental data: these powers have been entities like invoices, telephones, emails, pieces of papers, calculators and his own body. In general, we found him to not have made data. Much rather, his work consisted of bringing consumption statements into relation with each other, rendering them considerable as data. I argue that his approach was distinctively structured by the availability of these statements: he con-figured and con-texted them in ways which allowed him to pull out a folder of invoices or open a file, indexing the sources of his data. These sources were claimed to be representing specific pasts. 85 2. Data Construction in a Basement However, pasts are always indeterminate. They can only be reported and interpreted based on current acts of experience, conceptualisation and classification (Bowker and Star 2000, 40). The translation of statements into seemingly mere numbers allows for both, application of quantitative classes (and respective further quantitative operations) as well as the ignorance or reconfiguration of ‘much of what is difficult or obscure’ (Porter 1995, 85). If pure facts emerge from such translation of past consumptions into present environmental data, later political attention to the facts is unlikely, if not impossible (Bowker and Star 2000, 325). Classification as a political activity: we found that classifying objects presupposes that the latter are rendered threefold, as knowable, as known and into an inscribed relation to a category. Through the last act, the attachment of a quantity to a category, a further relation is achieved: the quantity is linked to the carbon conversion factor which, in itself, is attached to the category. If emitting carbon is a politically relevant activity, then linking consumption to carbon renders this linkage political itself. Precisely because of the political – and economic, as I will show – relevancy of this link, the choice which categories to relate to quantities can easily be subject to contestation. Classification is contested. I showed how an organisation which has an idea of what is the ‘correct’ classification of an object presupposes the disciplining of the classification scheme’s users. The latter have to realign their understanding of what is to the scheme’s (or its representatives). In effect, ‘local’ conceptions can be marginalised through disciplining. With Riles (2010, 799), however, we find that statements were not existing so much as immutable mobiles (say, the invoices) that were disciplining the user (as Latour (1987) could be read). Rather, we may propose, it were the abilities of those actants that the data entering agents had to go through, i. e. the OPPs (such as databases and bosses, but see Section 2.3), that were decisively disciplining. A monstrous enterprise: this section established also that culture is a key for members to conceptualise consumed entities and categories. While the corporation’s HQ tried to prescribe how classification was to work, in practice this control was not possible: some prescriptions’ intended effects could not materialise because they were not translated into work practice. This relates to the problem that, as MacKenzie (2009b, 120) put it, even ‘the most detailed rulebook will on its own be insufficient to determine the practice of bookkeeping and accounting’. Even though Nick was accompanied by the HQ’s set of prescriptions as well as by myself (also acting as HQ’s representative in this case – and I did honestly voice HQ’s prescriptions) it was he who was supposed to and managed to enter consumption data into the database. I am glad for this – yet in most organisations, humans cannot be totally steered by their superiors. With Latour (1993b) we can summarise: there is no such thing as the modern bureaucracy in which workers are merely cogs in the machine. They are creative, interpretive actants. Through their relational engineering achievements, they manage to link consumption to carbon emissions. In the course of this, the emission conversion factors are cited and recited, effectively strengthened (Lippert 2012a, 158). At the same time, however, they 86 2.1. Classifying Consumption are positioned even more precariously. If the factors also start to, or already, include the emissions attached to the financial relations in place for productions, i. e. if products’ and services’ conversion factors internalise the emissions ‘caused’ by their necessary financial overhead work, then a circular structure exists: factors are black-boxed into emissions and these emissions are black-boxed into factors. In many respects, then, Nick’s practical work of categorising consumption mattered – and it still matters. Also black-boxed into the carbon matter, now linked to GFQ, are the meanings which members attached to the entities they needed for their work. As we have seen, however, meanings have been contested. Some meanings could not materialise. The water case provided a complicated story: while neither the HQ could ensure that their version of water determined which carbon emissions would be produced by Nick nor he himself had total control over the emissions, the conversion factors emerged as even more relevant. Nick could increase the quantity linked to a category and the respective factor, but he could not reconfigure the factor in itself. Some more emissions resulted than the HQ wanted, and some less than would have been appropriate for Nick’s wish to represent bottled water consumption. Thus, in the midst of multiple meanings, some emissions became invisible. Bowker and Star (2000, 310) call these monsters. The latter emerge when the multiplicity of meanings cannot be accounted for. The bottled water may only just be a monster, maybe an imp; but are there meanings than are not categorisable?9 Nick’s boss made clear he did not want all the meanings made explicit – the boss wanted pure facts. Juggling meanings is what Bowker and Star call categorial work . A question which we have to follow throughout this book, then, is where monsters are constructed, how they come into being or are rendered explicit. One way of attempts to balance out monsters was enabled through the quantitative character of the data. The combinability of quantitative facts provides for possibilities to add numbers quite freely: a data-entering agent can simply increase numbers as translations of imps into emissions. Another way is the one this book takes: re- and alter-modalising emissions. This includes pointing to the non-positivist relation of the author of conversion factors to his products: if a factor’s author tags his products as subjective, it seems monstrous for an enterprise to get rid off such a statement. Two positions need stressing. Agreeing with Bowker and Star (2000), classification is a practice configured in a heterogeneous network that does not produce neutral facts. From this analysis, classifying consumption emerges as a fundamental issue of the politics of what is counted as reality. Also, tracing Waterton’s (2002) argument, at the same time, a classification scheme cannot be expected to produce orderly effects: because it is an issue of practice, its effects cannot be predicted through an engagement with the history of the scheme. While the scheme influences how consumption is linked to carbon matter, human actants’ classification practice can dissent from prescription, such as the subversive calculative operations as seen by Nick. In the following, therefore, we turn to how numbers and calculative work was performed by him. 9 I am grateful to Lucy Suchman for making me aware of the difference between things that are uncategorisable (monsters) vs. those that are translated across normative differences (that I call imp). 87 2. Data Construction in a Basement 2.2 Constructing Consumption via Calculation In the prior section, we encountered Nick classifying quantitative data and, thereafter, translating that data into account forms provided by GFQ’s database ESDR. For these successful translations he had carried out three simple calculations; two of these even did not seem noteworthy. In this section we revisit these instances and wonder about the quality of calculations. It immediately comes to mind that calculations are different from classifications because the former is concerned with quantitative information. More precisely, whereas classification implies that entities are sorted into groups, calculation presupposes a common metric allowing commensuration (Espeland and Stevens 1998, 316). Similarly, carrying out calculations also translates some things into other things (Latour 1987, 239). How does this kind of operation matter? This section shows that maths matters, that the practical reality of calculating constructs emissions.10 For our understanding of society and economy, calculations matter a lot. According to Köhler (2010), modern institutions are fundamentally characterised by people being convinced of the calculability of societal phenomena, organisations and issues. Following Weber, he argues that modern administration uses numbers and their rational calculable transformations to translate social processes into expectable-predictable orders. Accounting is one of the main calculative mechanisms orientated at that goal. Callon (1998b, 23) reads Weber as suggesting accounting to be a ‘key prerequisite of modern capitalism’. And for Guattari (1995, 34-35) ‘formulae, equations and calculations’ may be part of the machines which populate our universe and shape our relations. If indeed number operations, their translations into accounting systems, are of such importance, then we need to investigate how they are practically performed and what is silenced in that process. Callon (1998b), continues, by calling for studies of this ‘humble, disclaimed and misunderstood practice: accounting and the tools it elaborates’. By that, he carried over the proposal of Latour (1987, 246), to follow the calculators and how they tie together information, into making the economic sphere a sociological object.11 In this book, I draw on their approaches to study the inside of an economic actant, GFQ. In the edited volume The Laws of The Markets (Callon), Miller (1998) argues that accounting shapes what it measures. For the accounting practices shown above, as they incorporate various instances of classifications, we can comprehend this: we saw that classifying performed water, paper and electricity consumption. In this section we broaden our perspective by engaging with this quantitative type of translation. Some readers might wonder: calculation should be left where it belongs: in the realm ob abstract cognitive engagement. Calculation is to be studied (and developed) in high school or at university. However, Lave (1988), in her study of supermarket bestbuy decisions, pointed out that mathematics in practice is a very different enterprise 10 For this argument, I am not engaging with a sociology of mathematics (as an academic discipline). The latter has been undertaken by e. g. Heintz (2000). I use the word calculation to refer to a practical and material activity by humans, rather than to a disembodied mathematical logic. 11 This theme resonates with further texts (e. g. Hacking 1990; Porter 1995), here not in focus. 88 2.2. Constructing Consumption via Calculation compared to high school problems.12 She shows that it makes a difference whether you solve a ‘going to the store’ problem in high school (in which the store is actually not relevant but you need to do the maths ‘correctly’) or if you are going to the store (and need to buy things). In the latter case, in actual practice, people have different aims compared to the classroom; they act within a physically structured space and they gather local knowledges about the shopping space as well as about the organisation of prices. If we are to understand how calculation practices affected the construction of carbon emissions, we need to study the ‘actual calculative practices of actors at work’ (Beunza and Stark 2008, 253). This requires to do a sociology of calculation.13 Following Kalthoff (2005, 70), I read calculation as a type of epistemic practice through which new objects are performed. By investigating four calculative instances, this section contributes to this field. First, we revisit the additive calculations which we glossed over in the prior section. This allows us to situate the rationale of these translations. Second, the section turns to calculations in which Nick used cost information to derive consumption. By studying this example we draw out the analytical implications for understanding calculations. Third, the story follows Nick in a combination of calculation and classification practices. This allows us to study the materiality required to render the issue calculable. Correspondingly, fourth, we engage in a conversation about purifying a calculation from unwanted information. In conclusion to this section, I draw out some cautious suggestions for an ethical reconsideration of the received understanding of calculations. The prime case discussed below concerns the determination of ‘kilometres flown by Nick’s GCE’s employees’. To produce a fact about this distance Nick had to construct a complicated network of relations between data and assumptions, rendering specific fragments visible or invisible. Figure 2.6 (on page 105) sketches these relations which I iteratively introduce. Above, we encountered Nick who was engaged with mapping representations of environmental consumption onto accounts provided by the Environmental Management System (EMS). For example, he made electricity invoices speak for the electricity consumed at that particular site of his GCE at which he was working as an engineer. He needed to check the monthly invoices for the relevant information, the amount of kWh billed for, and added them up subsequently. Later, he qualified the year’s consumption sum which he put into the database: he inscribed into ESDR’s form for ‘data quality’Z that the amount of consumption had been ‘measured’ (rather than ‘calculated’). However, from a naı̈ve point of view, we should wonder now why he designated this modality to the number. After all, adding up can easily be understood to be a mathematical operation. ↩→14 For the case of water consumption, he also added up several consumption information (in this case in cubic metre). However, in this 12 See also Verran (2001). A sociology of calculation is usually implicit. However, Kalthoff (2005) and Cochoy (2008) use this notion. I read them as compatible with the work following the paradigm of Callon (1998c). 14 The following discussion is relating to Artefact 2.1.3 (on page 81). 13 89 2. Data Construction in a Basement situation he qualified the numbers as ‘estimated’. If we revisit Artefact 2.1.3, we find right of mark ‘C’ a so-called comment, provided by Nick. The comment can be read as proposing that 800 employees used 154 m3 of drinking water. The amount of employees was relevant because he knew that at his site of his GCE 800 employees worked. I did not ask where the number came from; and at another situation I asked whether he were sure about the number. He was not. However, the 800 employees number was adequately correct for all intentions and practical purposes of the situation. Both of us shared, thus, the judgement that it was right to enter a false number under the specific circumstances: a couple of workers more or less did not make what members would consider a ‘significant’ difference). This finding resembles Lampland’s (2010) account of false numbers in the context of rendering quantification routines acceptable. Note, the situation, in terms of present humans, was immediately shaped by himself. I was merely present as a support worker and researcher. Thus, I did not constitute a dangerous challenge to the number. Nor would I want to challenge it now in ethical or evaluative terms. I draw attention to this fact in order to argue for the situatedness of the calculation we are dealing with. A calculation can only be understood by engaging with the situation in which it was present. The number was not present by accident but for particular calculatory purposes. Rather, Nick drew it onto the scene in order for it to do work and speak within the following continuation of the calculation: he implicitly argued that if 800 employees consume 154 m3 , 1299 employees would consume 250 m3 . The premise was that he can apply the same level of consumption that he calculated for the 800 workers to people somewhere else. This is a further order of qualification – a judgement about offices and workers whose consumption rate he did not have immediate information about. Nick’s practical solution to the lack of information was, thus, the construction of a generalising calculus: with it he could remove the site-specificity of the data he already had. This calculus required further assumptions. The 1,299 employees were taken as the amount of employees of his total GCE. By dividing the local consumption (154 m3 ) by the amount of consumers (800 employees) he derived a drinking water consumption factor of about a fifth cubic metre, surmised and immediately assigned as average consumption of employees. He then multiplied this factor with the amount of all employees, a calculus called extrapolation. In this way, he inscribed the estimation that the employees at his site consumed as much drinking water as those located hundreds of kilometres away. Then, in friction with HQ advise, he added the water consumption amount he had gathered from the cafeteria. Thus, this case is one of calculations and judgements being built upon other calculations and judgements – focussed precisely on the building and entangling with the organisation, rather than disentangling numbers from the organisation and scrutinising them in isolation. This description differs very much from Kalthoff’s (2005). He had carried out field work; in the financial services sector, too. However, he focusses on risk analysts. They made their business by questioning the data they received in order to evaluate the risk 90 2.2. Constructing Consumption via Calculation reality supposedly represented by various forms and tables. The rationale was that better checking would result in a more reliable understanding of the economic reality represented. Thus, they were orientated towards an external actor (e. g. a firm which wanted to borrow money from their bank). In my case, however, the figures produced by Nick were directed at internal actors who did not have any reason to question the data. The figures were not linked to the core business of GFQ. Rather, they were seen as an add-on which Nick invested only a few days in – his boss, Simon, would not have wanted otherwise. The rationale of the figures produced by Nick was to account for his GCE’s environmental consumption reality. He would try to give a good and defendable account. However, he did not have to be interested in uncovering the reality underlying the invoices or the water consumption reality far away from his site. He could safely assume that any of the calculations and judgements he was performing would not alter the objects’ characteristics (Kalthoff 2002, 29). I, thus, propose that by referring to the orientation of a calculative agent we understand better whether the agent would focus on opening black boxes or, rather, closing, stabilising or sealing them. Translating Costs into Consumption We just encountered that Nick had inscribed the generalised consumption of abstract employees into the sum of drinking water consumption. He had assumed them to, on average, consume the same amount of drinking water. What do such generalisations entail? I turn now to another instance of generalisation to spell out the steps needed to perform such a calculation. Early on during my visit to Nick, I had asked him how he had gathered the data. He provided this overall narrative: Field Note Extract 2.2.a (Collecting data) First he sent emails to the secretaries of the managers at [corporate sites at three other cities]. He expected an answer from them soon. Further, he took data from the accounting department, asking them: how much money did they pay for water, gas and electricity. He told me the prices for these commodities. 1 m3 of water equals 2.3 EUR, 1 kWh of electricity equals 0.08 EUR and 1 m3 of natural gas equals 0.033 EUR. I asked whether getting the data was a problem. His reply was straight forward: no. Second, he said, I wrote and calculated. Nick suggested that it would be great if the accounting department would provide a programme/software for this calculation. His completion of the narration was: this is the story. In this extract we learn about how Nick mobilised data and transformed them into consumption data. According to Barry and Slater (2002), we should ask on what calculativeness depends in this case. Prior to relating numbers to each other and transforming them in any way, Nick had to gather data. He requested that specific data from other organisational actors. In this set-up, he would accept the data delivered by these other actors, i. e. the secretaries (of other sites of his GCE) and his contacts at the accounting department of his GCE’s site. By that he used these other actors each as a version of, as I will argue, an Obligatory Passage Point (OPP). While in Latour’s (1987, 91 2. Data Construction in a Basement 245) account, centres of calculations appear as focussed OPPs, collecting, reviewing and creating more abstract forms of data, when following the flow of data at GFQ – from consumers via secretaries to Nicks, Simons to the HQ – each of them can be considered as a point through which data has to pass. As a source point and control centre at the same time, depending on from which direction in the imagined flow you are looking. In the account of Nick, we find his ability to enact a centre of calculation to be dependent on the distributed OPPs through which data was passed on to him; at the same time these source points were dependent on Nick to pass on the data to the HQ. They enacted the punctualised sources of data for him. For him, to request the appropriate data, he needed to know what data counts.15 His practice of requesting data can be understood as initialising a process fundamental to calculations in general. Callon and Muniesa (2002; 2005, 1231) propose that as a first of three steps to any calculations, ‘the relevant entities are sorted out, detached, and displayed within a single space’ (Callon and Law 2005, 719), i. e. classified into the account. In Nick’s case, the sorting was performed at the moment of requesting specific data and retrieving precisely that data. Note that this was not self-evident. It can easily occur, as we shall see below, that actually data is delivered which does not fit to the imagined space. However, in this case, Nick treated the data he got as compatible. This implies a significant performative move. By treating the data in this way, he positioned the data as ready for and to be treated by commensuration. As a second step of the calculation, the entities within the space can be positioned in new relations to each other, be transformed. Lohmann (2009b, 503) points to the laborious process partially necessary at this step to ‘rework the objects’ in order to make them ‘“passive” and tractable to the agencies of calculation’. Nick needed to invest such work. It consisted of making a metric and a spatio-temporal assumption: that the consumption of drinking water, electricity and natural gas were priced, and equally so. With that, I point to the qualities of two kinds of generalisations. Consumption was commensured through the metric of costs, counted in Euro. As Espeland and Stevens (1998, 316, 317) suggest, in this moment all uncertainties have been absorbed; data is rendered as certainly commensurable through a common metric. Thus, following the classificatory step, drawn into a single space, the entities can be compared through the same kind of measurement. In this case, Nick transformed the prices into an ordered set of six types of data: for the consumption a quantifier (the amount consumed), a unit (through which the amount was measured) and a category (the category of service consumed); and again, for the referenced costs a quantifier (the price), a unit (the currency EUR) and a category (costs); all of this information was explicitly inscribed in ESDR task-formsZ. Thus, the commensuration moment in this step of calculating consisted of organising data in specific relations, allowing for mathematical operations on them. Furthermore, his approach also implied stability in both dimensions: consumption 15 In that respect, Nick was performing a similar function of knowledge compared to the one used by the traders described by Beunza and Stark (2008, 253). Their arbitrage traders needed to know what counts as a condition for performing any quantitative calculative operations. 92 2.2. Constructing Consumption via Calculation prices for these entities were judged as the same at the different sites of consumption and at different times of consumption. By this, Nick effectively extended the network which he was shaping. Rather than describing the information in words and telling his superior that he could not get fitting data, Nick translated the data into entities which allowed further calculations (Latour 1987, 251). The calculation concluded: third, a result is extracted. A new entity is produced. A ranking, a sum, a decision. A judgment. A calculation. And this new entity corresponds precisely to – is nothing other than – the relations and manipulations that have been performed along the way (Callon and Law 2005, 719). What does judgment mean in this respect? Haven’t we just described a quite ‘normal’ calculation? No, Nick had to take judgements which he wove into the calculation’s thread. He first separated what data he wanted to commensurate, i. e. make equal, and, later, he added spatio-temporal assumptions about prices. Beunza and Stark (2008, 256) generalise: ‘calculation involves judgement’. Callon and Muniesa (2005, 1231) (re)draw the boundary of calculations such that they include the distinctions needed to be made for calculativeness to be established. All the numbers assigned to categories are performing judgements (Robson 1992). Thus, a statement like ‘1 m3 of water costs 2.3 EUR’ establishes seven relations and two sets; the set consisting of the triangular relations between the quantifier one, the unit and the category water as well as the set of, again, triangular relations between 2.3, the currency and the category costs. And these two triangles are, again, related to each other, illustrated in Figure 2.4. Note, such a statement is everything but universal. It does not translate into arbitrary languages (Guyer et al. 2010). Figure 2.4: Relations in a Simple Quantitative Statement (W ater) Category • (Costs) Category • • • Quantif ier (1) • U nit (m3 ) equals • • • Quantif ier U nit (2.3) (EU R) Nevertheless, Nick had to get on with his work: he employed this structure of data to relate it to quantified representations of his work environment, the count of workers of his GCE and the costs of consumption at his office site. (I reconstructed their respective 93 2. Data Construction in a Basement assumptions and judgements towards the beginning of this section.) With some further mathematical operations, multiplications and divisions, Nick achieved to calculate for several environmental indicators the total consumption amounts of the GCE. The newly made entities – each a complicated network of relations and imagined entities – are, thus, the result of processes which included both, quantitative transformations as well as qualifications. Callon and Law (2005) propose to call this kind of process qualculation. Such qualculations ‘are all about arraying and manipulating entities in a space in order to achieve an outcome, a conclusion’ (ibid., 719). I introduce this concept in order to distance ourselves from the received view on carbon accounting as some simple number juggling. The notion qualculation, thus, should be understood as a sociological designation for what actually happens in the practice called by members ‘calculation’. The outcome of a qualculation is turned into a thing because its constituents ‘hold together’ (Callon and Muniesa 2005, 1233). This new statement is stable and mobile. An open question, however, is this: what is lost by drawing these diverse entities into a single ‘spatio-temporal space’ ? I explore this question by turning to a specific formula which Nick used. Qualculation in Practice It has become recognisable now that, analytically at the least, it is apt to not discriminate between (qualitative) judgement and (quantitative) calculation. When Nick was attending to the travel account, he drew together a number of entities. The EMS prescribed members to account for the distance travelled by various means; travel was a Key Performance Indicator (KPI). One of its sub-accounts was called ‘short-haul flights’. This account was represented to the ESDR user Nick through a task-form. As the starting point of this calculation Nick used, again, a sum of costs. The sum he used was the total of the GCE’s financial cost account for domestic flights. For our purposes, it suffices to introduce the calculation as a formula. The total costs of domestic flights is indicated as total flight costsdom. f lights . Yet, the concern is not about anything measurable as ‘EUR’ but only in a distance, like ‘kilometre (km)’. Field Note Extract 2.2.b (Calculating travel) He calculated the following: total f light costsdom. f lights ∗ average f light dist.dom. f lights average f light costsdom. f light 168,078 EU R ∗ 500 km 230 EU R = total dist.dom. f lights = 365,387 km I suggested writing the comment in the record of the database how he calculated the estimations. He was a bit hesitating but then did so. As a way of scrutinising this calculation, I tease out what has been included in it. Above all, Nick had identified a particular partial statement which he took into account: it was the triangularised statement, consisting of a known category and unit, both prescribed by the task-form, as well an unknown quantifier – in short: the total distance (total dist.) of short-haul flights. In order to make this quantifier known, he engaged in a series of routine translations. With Verran (1999, 150) we can assume his quantification 94 2.2. Constructing Consumption via Calculation routines as resulting in certainty. The result of his calculation was to be certain – this was the very aim of the activity. Otherwise, he could have simply made up a number. She proposes that by studying how certainty of numbers is achieved we can make the quality of their certainty experienceable. To commence the qualculatory path towards a (re)solution for the unknown quantifier, first, somehow, he had excavated a statement on the total costs of domestic flights paid by his GCE (168,078 EUR): whilst Nick himself was tied to the building in which he worked, he was able to mobilise via phone and emails actants at other places, motivating them to submit data to him. For this calculation, he relied on one such datum which he retrieved from others. Second, he mobilised his estimating imagination. By this, he was able to draw two further statements into the calculative space which he was forming. These were statements about the average distance of domestic flights (500 km, denoted as average flight dist.dom. f lights ) as well as their average costs (230 EUR). By saying ‘average’ we are referring to a spatio-temporal set of assumptions including, as discussed above, the imaginary of commensurability. Verran’s (2010, 106) comparative qualculation studies emphasise that any such generalisation logic is particular, premised upon a particular ontics. These statements were, thus, not purely quantitative but interwoven with judgements. Stabilised by a variety of assumptions, his calculative space consisted in total of four statements, one of which was not completely known. These statements needed to be rendered commensurable, i. e., calculable. Table 2.1: Organised Qualculative Statements Knowledge Status Framed triples of data Quantifier Unit Category of Flights Partial Statement ? km short-haul (s. h.) Complete Statements 500 168,078 230 km EUR EUR average s. h. flight distance total s. h. flight costs average s. h. flight costs In Table 2.1 we find these statements organised into a common frame, into triples. This can be understood as a first transformation of the elements within the calculative space. Cochoy (2008) suggests, further, that calculative spaces can be structured. Then, this reconstruction of the statements as a threefold-structured space can also be understood as a manipulation of the space itself. By translating the statements into such triples of differentially treatable but directly related elements, Nick rendered all these flights and costs as commensurable – in two metrics for costs and distances. The assumptions could not exist anymore explicitly in this space. To illustrate, by conceptualising the average flight costs as quantitatively measurable and representable in this triangular set – rather than, say, as ‘too expensive’ or uncertain, ‘230 EUR?’ – it became commensurable (Porter 1995, 86) with the dissimilar thing total flight costs. 95 2. Data Construction in a Basement Through commensuration, the difference of both things is translated into a question of magnitude, Espeland and Stevens (1998, 316) argue. Nick then searched for a way to relate these triangular sets in a way which would allow him to produce a result which he could enter into the form’s field for the desired quantifier ‘?’. While commensurability in itself is often questioned, the choice for the (usually as singular conceptualised) mathematical apparatus as an appropriate calculus is normally not questioned. It has been normalised. Even more, working with numbers has become a social norm. Porter (1995), correspondingly, studied ‘trust in numbers’. He sees quantification as a social technology (ibid., 49). ‘Strict quantification through measurement, counting, and calculation, is among the most credible strategies for rendering nature or society objective.’ (ibid., 74) Historically, he traced back, objectivity became to be seen as subordinating human, personal interests, substituting the latter by public standards. Nick, maybe influenced by being a well-trained engineer, reproduced this normative choice to operate a standard division using this apparatus, to employ its operation rules. The latter imply not only a set of axioms16 – what mathematicians call ‘foundation’ – (which in itself is changing, and is contested, Shulman 1996), but also very particular relations, among humans and other materials (Restivo 1993; Barany and MacKenzie 2011). The solution he found to identify the unknown consisted of rearranging the triples into triangles which would form what mathematicians call an equation. This formula brought together many relations – illustrated, symbolically, in Figure 2.5 on the facing page.17 On the left hand, the outer circle depicts the formula as a total. The right hand circle represents the class (short-haul flights). Nick linked both circles. Inside the left hand circle, three punctualised data sets are found (each small circle). Nick constructed the relations for each data point between quantifier, unit and category (the triangles). In order to imitate the shape of the formula, the figure presents the relation between the data points with a denomination line and a multiplication point. The result of this calculation, which Nick extracted after a series of transformation of the calculative space and the statements he assembled, could be referred to as the total distance of domestic flights. His arrangement, thus, produced a punctualised entity (365,387 km). By this I mean that not only was this entity black-boxing the various assumptions and data constructing methods (including: aligning actors, communication devices, statements, a formula), but also that this entity appeared as a single point to which Nick could later on refer to and calculatively work with.18 However, while the equality sign ‘=’, only introduced in the sixteenth century (Cajori 1909, 175), suggests 16 The standard axiomatic system for set theory, at the moment, is staged as a 20th century enterprise – referenced by Wikipedia authors e. g. to Fraenkel, Bar-Hillel, and Levy (1973), Zermelo (1908), see Kunen (1980). 17 It follows the logics of Figure 2.4. For an explanation see also the introduction corresponding to Figure 2.4 (on page 93). 18 Kalthoff (2005, 71, 74) emphasises the distinction between ‘calculating something’ and ‘calculating with something’. The former refers to the production of numbers and the latter refers to the expectation that one can use the numbers to interpret a reality and built on this interpretation. 96 2.2. Constructing Consumption via Calculation Figure 2.5: Relations in a Simple Qualculation that both sides of the formula are substitutable, I propose, that this formula did not at all result in an absolute equation. The very point of the labour put into this qualculatory work, was to produce an entity of higher value. If both sides had been equal, there would have not been any progress with his work: the information after the translation must differ (Latour 1987, 51, 238-239). His investment in the calculation was worth it because the task-form demanded from Nick to insert a numeric value into the form’s field entitled ‘Value’. Technically, the form did not allow the user to inscribe it with the pile of triangular sets in the field. And, how could he have easily represented the left hand of the so-called equation? Equality between both sides of the relation emerges, thus, as restricted to a narrow mathematical sense. For all practical purposes, both sides did not have the same value for the user. The appropriate choice of a reduction apparatus was key for Nick in order to make present an entity which ESDR and he, himself, could easily calculate with. We find, thus, that the mathematical apparatus performed by Nick should not be understood as operating primarily on numbers. Two-thirds of the elements which it transformed into invisible were qualifiers – units and categories. This very achievement made possible that the traces of the qualitative relations underlying a quantifier could be hidden from users along the numbers’ production cycle. Nick rendered this specific entity temporarily visible. Hence, Nick presented an entity, brought it into presence. However, as the literature of the performativity of economics – which I read as an element of the sociology of calculation – reminds us: any presence also implies an absence. Nick rendered this new entity into the visible, whilst at the same time blinding himself and all other actants who would deal with his constructs from recognising all the other assumptions drawn into his calculation. Beunza and Stark (2008, 279) remark: seeing implies blinding; moving things away, out of the focus, out of the frame – framing being the conceptualisation offered by Callon. Drawing on Goffman (1975), he proposes the notion of frame to mark the boundary of that which is self-evident (Callon 1998a, 249). Things which are not or cannot be accommodated in the frame are out of the frame, are overflowZ. In their 97 2. Data Construction in a Basement later work, Callon and Law (2005) generalise that all qualculations co-exist with ‘absent presences of non-calculations’. Some things which are not part of the qualculation are, nevertheless, present; they continue to be part of the imagined total world (i. e. of the critical realist real). However, these things do not enter the calculative space, nor are they in other ways penetrated by judgements. Rather, they are judged to be left out of further instances of qualculations. This includes all those monstrous flights which did not fit Nick’s spatio-temporal assumptions. The formula used by Nick was key in the formation of this specific new entity, i. e. the result of 365,387 km. If the formula had been different, the outcome would diverge from this quantitative reality. And such calculations could in principle take place so they are part of bills; but that does not make them less contingent and precarious. Nick expressed for the calculations discussed earlier, that he would have liked to automatise such calculations. He voiced how helpful it would be if GFQ provided him with a software to apply such formulas, without him needing to actively decide for the arrangement of numbers and pressing buttons on a calculator. In this situation, my immediate response had been: I started considering whether I could convince my superiors to implement such a device within ESDR. The material form of a calculative device black boxes assumptions and methods (Knorr Cetina and Bruegger 2002; Kalthoff 2005; Beunza and Stark 2008, 254; MacKenzie 2009b, 59). An actor with a device differs from an actor without one. They enact both, framing and overflowing, while making overflows less visible. Nick constituted a new entity with his particular material arrangement of other entities. The travel ‘consumption’ which he brought into presence represented this arrangement, the relations inscribed in it, and not something outside. Knorr Cetina and Bruegger (2002) refer to this as not representing a reality out there, but as constituting an own reality, without any counterpart ‘outside’. They call this form of making present appresentation. Kalthoff (2005, 71), thus, suggests that ‘through technical devices of calculation [. . . ] companies are constituted anew’. We return to how Nick proceeded with the calculation. He navigated to the ESDR application in his web browser, and opened a task-form that was linked to the account short-haul flights. He entered the calculation’s result, as well as other necessary data, declared the data quality of this task-form as ‘calculated’ and ‘completed’ by checking a box; mark ‘D’. These actions were achieved in few moments – and they achieved very much: Nick declared domestic flights as short-haul flights. This was consequential in three ways. (1) He mapped domestic flights of a variety of distances onto the distance 500 km. The indicator ‘short-haul flights’ was specified within ESDR as having a carbon emission conversion factor of 180 gram (g)/km. Nick, thus, enacted a classification of all the domestic flights as short-haul flights, which, by that, would become visible as having caused in total about 66 tonne (t) of CO2 emissions. All the domestic flights’ emissions longer than 500 km (identified as monstrous flights above) would accrue in the absent realm of overflows. Thus, a first effect was that the calculated emissions were lower than if he had assumed flights to span longer distances. (2) In addition, 98 2.2. Constructing Consumption via Calculation this overflow of monstrous flights would, if Nick had ‘implemented’ HQ’s prescriptions, actually have resulted in lower total emissions. This specific overflow needs spelling out. HQ wanted GCEs to classify flights, spanning more than 500 km as long-haul flights. The indicator for long-haul flights was associated with a conversion factor of 110 g/km. Thus, long-haul flights were considered to cause less relative emissions than short-haul flights. If Nick had classified flights which had spanned distances of slightly over 500 km as long-haul flights, total emissions would have been lower. Thus, the second effect was that Nick increased his GCE’s emissions. (3) Finally, however, by assuming that all international flights were long-haul flights from HQ’s perspective, all those international flights which spanned less than 500 km were about to be related to a ‘too low’ emission factor, resulting in lower emissions. While we find that the actual performative effects of Nick’s qualculation are not easy to sum up, it becomes clear that it is precisely through his qualculative engagement with these entities that the accounting model was extended, i. e. that the calculuses of standard mathematics as well as of carbon markets were applied on his data. However, as Miller’s (2002) intervention in the performativity of economics debate argues, actually, calculuses do not represent well the relations on the ground. The calculuses are staged as data-being-governed. But, as we have seen, data is translated to fit the models, leaving monstersZ aside. Nick, and similar agents, were situated to provide the data demanded by the corporation. This does not mean the results of qualculations actually reflect what they appear to represent. Appresenting facts, thus, necessarily implies absences; and those entities which are present are present because of normative and political moves. The structure of the qualculation employed statements; these were translated, transformed, into data points, rendered into particular relations to provide a result for the EMS’s forms. Yet, none of these calculations directly link to real emissions. Finally, however, Nick had not been too happy with his calculation, even though he had produced a clear-cut number which nicely fit into the form provided by ESDR. Shortly after carrying out the calculation, he asked his accounting department for further information. How their reaction changed the situation is the subsequent topic. Exclusive Materialities The final part of this section serves to rethink the ambiguous role of materials which enable and are formed by qualculative practices. This involves attending to how these materials have been configured, i. e., their materiality. As a point of departure, I shift forward a couple of hours. We were both waiting for Nick’s boss, Simon, to call us. Nick and me were about to report to Simon about our data construction outcomes of the day. While we were waiting, Nick allowed me to shadow him as he was checking emails. At some point, I was losing track of time – it was after 5 pm, he received an email detailing what had been actually paid by the accountants through the domestic flight account. The email contained an spreadsheet. Field Note Extract 2.2.c (Problem with calculation) He was scrolling through a Microsoft (MS) Excel spreadsheet and found: domestic flight included items such as: 99 2. Data Construction in a Basement • • • • • • boat [to cross river in the city at which his GCE site was operating], taxi restaurant visit: 12 EUR McDonald’s [other city] visa costs: 265 EUR 1166.5 EUR for former business partner, Jones. Nick called out: the list includes Lunch! Checking emails can be considered a routine activity. In this case, the digital material which Nick immediately hove into the qualculative space constituted a problématique. Looking at the spreadsheet’s contents, he found statements which destabilised the arrangement which he had constructed before. Learning about these items shacked the relations holding together total flight costsdom. f lights . This triple data point set (quantifier, currency and account, i. e. 168,078 EUR paid for domestic flights) was built upon the foundation of a pure account class. However, with this email, Nick learned that some items had been classified into the category – practically speaking: put into the account – which he considered out of place. From his point of view, all the items named above should have not been linked as co-constituents to be represented in the account. Rather, the relations should have been ones signifying exclusion. This moment points, thus, to two issues of materiality: of what kind of material consisted a flight; and, what was specific about the material Nick now came across. The materiality of flights: the items, Nick just learned about, deserve our attention. Reading through them, we may ask: how are they constituting domestic flights or not? Where does a flight start, where does it end? Does it start when you book the flight? Does it start when you imagine yourself driving to an airport, rather than going by train? How are these boundaries drawn? Deciding upon these questions configures how flights are material, which materials flights consist of and, hence, does (Butler 1993) (the ontics of) flights. Thus, relating to the materiality of some thing, addresses how that thing is configured, how it is material. Nick might have shared the understanding with other actors within GFQ that a flight starts when a GCE agent entered a plane physically and the plane moved into the air. Then, the materiality of flights would have been related to the agent using a plane. Nonetheless, as we have seen above, he only related to flights covered by the costs account of his GCE. However, if I were to travel somewhere by plane, I can easily see how this entails all these items (except of the last, 1,165 EUR): I would organise to receive a visa – well, maybe for international flights, normally not for domestic ones; I would take a taxi and maybe also a ferry to get to the airport; eat at a McDonald’s while waiting for the flight; and, after I arrived go for a proper meal. In that way, all these items are well accountable for. Thus, while we can easily see how actors might associate certain costs with a flight’s account, an environmental management agent would be characterised by classifying only the paid for service of being flown in itself as an orderly instance of the category short-haul flights. The implicit claim is that a flight account has clear cut boundaries 100 2.2. Constructing Consumption via Calculation of what should be included in it and what not. For Nick, the materiality of flights was, therefore, an accounting issue. However, while certainly GFQ agents had clear cut understandings of how their financial accounts were to be used as containers for a GCE’s expenditures, the case of environmental accounting is more complex. Here is a brief excursus: a colleague of mine at the HQ explained once that they recently altered the carbon conversion factor reference for flights. The former reference19 provided a factor including the carbon emission caused by operating some of the infrastructure necessary for flights to be possibilised, e. g. airports. HQ’s EMS agents adjusted the flight carbon emission factors to draw the boundary around the flights, excluding the infrastructure operation’s emissions. The result was a decrease of the carbon conversion factor for short-haul flights by a factor of 1.8 and of the long-haul flights’ factor by 1.06 times – when comparing the employed conversion factors20 with the factors GFQ was employing for other categories; the differences are depicted in Table 2.2. Clearly, both standards measure different objects. Conveniently, the qualculations followed on after Nick’s data input, would translate his data into CO2 e, measured as metric weight. That different instruments measure different objects (Mol 2002, 75), and that Nick’s data input would be measured through competing instruments which all produced results (and were rendered commensurable via their metric weight specifications), would be made invisible. Table 2.2: Carbon Conversion Factors for Flights Source Indicator Conversion Factor Travel: Flights kg/km VfU short-haul long-haul 0.326 0.116 WBCSD short-haul long-haul 0.18 0.11 Above, we have already introduced the pair of concepts frame/overflow coined by Callon (1998a). This pair is useful to immerse ourselves in the significance in the boundaries drawn – those enacted by Nick and by GFQ through its database ESDR into which the lower two conversion factors of the table have been inscribed. Nick framed what he wanted to include into his calculation. He considered the items listed above out-of-scope for what he had to report. Hear the exclamation mark: Lunch! – reminding us of the everyday achievement of practical office work (Suchman 1983). It became clear to me, that he felt that the list was not pure. What were Nick’s options? He could be annoyed about his GCE’s accounting department not maintaining pure records. If he was to purify the record, it would be a local purification; it would 19 20 Normally, GFQ used the VfU as a reference for conversion factors. Referenced to the WBCSD’s GHG Protocol. 101 2. Data Construction in a Basement not spread. However, purification was in order (Latour 1993b). Even though I did not observe any purification, let us understand what was supposed to happen after that point. One by one, Nick had looked at all the items listed in the flights account spreadsheet. Most of them had been flights. Cleaning up the record would consist of evaluating each item in terms of whether it would fit to the category domestic flights. Let me comment on some of the items named above. How could GFQ’s EMS account for them? Table 2.3 provides a humble approach to provide you an idea of how GFQ would deal with these items. Without going into details, it suffices to note that some consumption types have not been provided for in the classification scheme; no classification scheme can attend to all characteristics of the objects it seeks to frame (Waterton 2002, 195). Table 2.3: Accounting for Wrong Items? Accountable? Item Within EMS Boat A travel indic- However, it did neither have categories for ator existed. travel by boat or a category ‘other’. A travel indic- However, only the carbon emissions caused ator with a sub by GFQ owned, officially hired or private cars indicator for car were considered to be accounted for. A taxi travel existed. trip was not part of this framing. Taxi Outside EMS Restaurant visit; McDonald’s I learned explicitly: food is not considered a topic for the EMS Visa costs; Transfer of funds Probably out of scope of the EMS. For GFQ’s operations – ‘serving’ global financial markets – all these consumption items, goods and services, and myriads of others were required to be in place. The hegemonic form of markets would not work without the infrastructures connecting people with their mobility devices (Bowker and Star 2000; Jackson, Edwards, Bowker, and Knobel 2007). However, Nick was about to exclude these infrastructure network instances from the calculative space which he had drawn up in order to provide data designated to represent his GCE’s carbon emissions. The items shown on the list are what Callon (1998a) called an overflow. Mainstream economic theory refers to this overflow as ‘externality’ (Callon 1998b, 16-17).21 Callon suggests that not everything can be framed. If social or environmental problems occur because some material entities have not been part of the frame which has been optimised, this is referred to 21 102 See also for competing engagements with externalities in economics Pigou (1920), Coase (1960). 2.2. Constructing Consumption via Calculation as ‘market failure’ (1998a, 247).22 Caring for the environment is not easily compatible with qualculation (Callon and Law 2005, 725). The materiality of the flights accounted for excluded many elements required for actual flights to be possible. The material devices carrying the qualculation: in the instance encountered above, Nick’s situation had been utterly changed at the moment the materiality of his qualculative space was, literally, re-con-figured. The arrival of the spreadsheet transformed the network, it introduced figures to be recognised together with other data points anew. It made the presence of a number of items visible which Nick did not want in the space. Thus, as a consequence of a new material entering the calculative space, the actual materiality of flights was questioned: flights consisted of costs taken into account and through the spreadsheet Nick found that the total costs were have not been purely consisting of ‘proper’ flight costs. Callon and Law (2005, 731) suggested this as a general property of qualculations:23 ‘Quantitative methods, qualitative procedures, professional judgments, or the tinkering of daily practice, all of these are qualculative. And how they are done is a function of the material arrangements, including the bodies, in which they are produced.’ A qualculation is dependent on the materialities which provide the entities to be drawn into the judgement-calculative space. At the same time, drawing some entities into that space may exclude other entities from being recognised in that space. This can have material effects. Thus, the material called ‘result’ may have inclusive and exclusive effects. In the case discussed above, some carbon emissions were not accounted for: a couple of days after my meeting with Nick, a number was in the database for this type of flights. The number was slightly less than 60% of the number which we had calculated together. The comments were gone. We learned to see that the seemingly known part of the total sum of flight costs was ‘non-accurate’. It included expenses which should have never been part of the number associated with total flight costsdom. f lights . Thus, the quantifier for total dist.dom. f lights was too high a result. From that point of view, a reduction of the quantifier reported for the account short-haul flights seems to have been a reasonable choice for Nick. As a reaction to the recognised shift in materiality, Nick had purified the statements entering the qualculation and transformed them in a way, probably along the lines described above, to conclude with an equivalent valuable result – one he could enter into the task-form. Thus, effectively, the spreadsheet made Nick re-materialise the short-haul flights flown accounted for by his GCE. And, I argue, the spreadsheet itself was reconfiguring as well the situation: Nick knew that this spreadsheet must have been a copy which also was existing at other places within the subsidiary. While, more easily, he could have ignored a phone call, specifying the 22 Note, in other respects it might also be good that some entities are not drawn into the space of qualculability (Callon and Law 2005); maybe we do not want all areas of life to be governed by judgements and calculations. 23 Kalthoff (2005, 75) makes an equivalent point. 103 2. Data Construction in a Basement impurity of the numbers in the cost account, the spreadsheet’s multiple existence, in his inbox as well at uncontrollable other locations, urged him to lift (it into) the calculation, and through that take it into account. However, we need also to consider this: what is about flights paid through other accounts – just like visa were paid through the flights account? This constitutes an unsolvable ambiguity, overflows out of Nick’s hands. ♦ End of Section ♦ This section set out to ask how calculative operations matter in the construction of carbon emissions. We found that when we turn to the specific numbers, they emerge not at all as absolutely stable. Much rather, entering and working with false numbers can be very adequate. To be a competent practitioner of carbon accounting, at GFQ, in specific situations, one better makes not a fuss about small differences in numbers. However, at the bottom of GFQ, members had to use exactly this kind of entity, numbers, to construct facts which would fit the HQ’s forms. The corporation used a database which did not allow for complex texts. Correspondingly, I found that the complex realities encountered at the subsidiary were stripped off some of their complexities by reducing them into a specific frame which allowed to calculate with them. This form was characterised through a threefold differentiation between quantifiers, units and categories to describe consumption of goods and services. I, thus, identified a triangular structure which consumption information (Robson 1992, 691) took. At the same time, assumptions were rendered invisible. These triangles could be arranged such that the standard mathematical apparatus could digest them. This apparatus was, as one could expect, transforming the quantifiers, producing new numbers. However, we need to emphasise that these rules of calculation were also acting on the remaining qualifiers which were part of the triangles. The result was that the new consumption fact was neat, stripped off, once again, of its qualifying history. While a formula describing the calculation would normally be called ‘equation’, this designation tricks the reader. I was able to establish the – for all practical purposes – in-equality between both sides of data arrangement. The very point of this calculative work was to transform data, to make it different – much rather than keeping it the same. This process of manipulating data required constant judgements. Calculation carbon emissions is full of such judgements. It would be wrong to say that this did not have consequences. The effect of taking these judgements, of translating data into different data, was that members were able to format data such that it would fit ESDR. Only by taking these normatively and politically consequential moves, a person supposed to enter data into the database could get her job done. Therefore it is apt to conceptualise carbon not as simply calculated, but as qualculated – as implicated in both qualitative and quantitative manipulations of data and their relations. Figure 2.6 on the next page sketches the key relations and practices which the example of Nick, qualculating the distance of short-haul flights by his GCE, entailed. All these relations had to be in place for Nick’s GCE to ‘posses’ its ‘total distance of travel by short-haul flights by all its employees for corporate purposes’. This included 104 2.2. Constructing Consumption via Calculation Figure 2.6: Calculating the Total Distance of Short-Haul Flights Travelled 105 2. Data Construction in a Basement the relations between Nick and the items which he understood to have been wrongly classified into the flight account. One aspect deserves more attention: the deletion of the comment in the ESDR form. I attend to this in the next section. The result of this qualculative network which Nick conducted was a pure number without easily recognisable traces of the qualculative work which it was based upon; no comment would suggest to its readers how the distance reported had been assembled. Following the observation by Kalthoff (2002, 33), we notice that this GCE did not possess some singular objective or independently existing fact like a ‘total distance of travel by short-haul flights by all its employees for corporate purposes’. A story suggesting that this fact can be simply figured out on the basis of ‘hard accounting facts’ to be read off the records is glossing over too many ambiguities. Rather, during the process of data collection at this GCE, many ‘total distances of travel by short-haul flights by all its employees for corporate purposes’ existed. ‘[T]hese calculations always remain incomplete’ (ibid.). While appresenting any such fact, overflows resulted. Realities were silenced when trying to purify data. However, even the merely mathematical application of rules attached further relations to people, texts, assumptions, world-views to consumption data and, thus, carbon. Any informational entity produced while qualculating towards carbon emissions has to be conceptualised as hybrid, intrinsically related to those entities considered meaningless along the way, monsters. For calculating is taking place in the real world, and not in some abstract logical pure space, the materials employed for calculation affect the calculative work. Material devices like calculators or spreadsheets enable their users to organise their data manipulation practices – and they can get in the way. Any quantity entered into an environmental accounting database has undergone qualculative moments. Therefore, it makes as much sense to assign a number which Nick en-calculated himself the data quality sign ‘measured’ as the status ‘calculated’. The scale of data quality used by GFQ did not correspond to the qualculative data processing reality on the ground. Environmental economics proposes to ‘internalise’ those realities which are significant for the social or environmental outcome into calculations (Clark and York 2005, 409). Global climate change is seen, correspondingly, as a market failure (Stern 2008). The implication of this understanding is that the economic and qualculative space which are part of allocating ‘goods’ and ‘bads’ (Beck 1996) need to be more inclusive. This would require the metrological networks to extend and incorporate more entities (Latour 1987). This is exactly the move which can be observed in global environmental politics. The introduction of more and more accounting systems – to account for nature – can be observed (Asdal 2008). However, the discussion in this section clearly suggests: numbers do not mirror reality ‘out there’. Numbers represent the relations and materials which have been arranged to fit together and perform a result. However, all these relations can be called into question; the materials – and this can be observed well at their boundaries – do not produce stable self-evident entities. When we see numbers which make sense then it is 106 2.3. Commanding Carbon because they are relating to some forms of qualifiers. However, whenever we encounter such constructs we should keep in mind that these are translated, alter-materialised, versions relating to invisibilised realities. They are signifiers of loss of accounts of realities and, thus, of lost chances to exercise transparency, democratic decision-making. 2.3 Commanding Carbon Recall, it was Simon who ordered Nick to gather consumption data from all over the GCE. Consequently, we have investigated what gathering data meant as a practical activity: we found two fundamental types of practices underlying the translation of consumption of goods and services into carbon emission data – the classification and calculation of consumption. This section underlines a moment of these practices’ situatedness, the moment their arte-facts returned to so-called superiors. For that I am shifting the focus from the construction of data sets to the performance of relations between bosses and the facts produced. By way of engaging with the presence of Simon (the boss) in Nick’s (the subordinated engineer’s) work to enter the data in ESDR, a re-conceptualisation of what it means to command carbon becomes possible. Agents who were engaged in the commanding of carbon were utterly dependent on heterogeneous elements in the network. To illustrate this argument we follow the finalisation of ‘decisions’ over carbon, from Nick’s office in the basement to the GCE’s top level. Towards the end of my visit to Nick, he went through all the ESDR task-formsZ, checking whether we had to discuss any more issues. Field Note Extract 2.3.a (Excluding paper consumed) Regarding counting the amount of paper consumed, Nick declared: Jones & GFQ was different. We should not include their paper. I suggested to ask Simon about this. This moment connects some very differently existing actants. Nick and me were present and related to a recently departed organisation ‘Jones & GFQ’ and paper labelled with the name of this organisation. I challenged the way Nick was about to materialise the relation between GFQ’s carbon emission machinery and Jones & GFQ’s paper. As a background for the analysis, the designation of the entity Jones & GFQ deserves a note: this entity was the condition for and consequence of GFQ’s move into the West Asian country this chapter is about. In order to have access to the financial market in that country, GFQ had to establish a partnership with a powerful local player, Jones. However, a year before my field visit to Nick, GFQ was able to leave the cooperative relationship and run this GCE on its own. Thus, Nick’s current employer was not anymore staging itself as related to Jones. Nevertheless, in a storage room, which Nick had shown me, large amounts of Jones & GFQ-labelled paper was waiting in boxes to turn into waste. Nick was aiming to complete data collection. This meant, after collecting or calculating the amount of paper consumed and classifying it, for finalising his task it was still necessary for him to complete filling the task-forms. In this case he did not start to 107 2. Data Construction in a Basement Figure 2.7: Challenging the Exclusion of Paper Consumption from Counting question what category of paper he should link the amount of paper to. Rather, he considered that Jones & GFQ was a different organisation than GFQ and, therefore, should not enter the accounts of GFQ at all. I was not agreeing with his perspective and, therefore, intervened. However, as I did not simply want to prescribe what the ‘correct’ relation between the paper with the Jones & GFQ label and GFQ’s emissions was, I suggested to include Simon in the decision. The upper circle of relations visualised in Figure 2.7 indicates the classification trajectory and relations around Nick. I, however, initiated the lower circle of relations, trying to enrol Simon, in order to challenge Nick’s intention. Thus, this classification case was contested. The theme of classification raises two points: first, Nick aligned his reasoning to temporal coordinates. According to him, because GFQ’s relation to Jones had ended, GFQ’s carbon balance was not to be polluted by past consumption of Jones & GFQ’s paper (materialised, as we have argued above, precisely through their costs). Second, my disagreement was premised upon my understanding that, if GFQ did not account for this paper, no organisational entity would do so. Then, the emissions related to that – now unusable? – paper would have been lost from being expressed in corporate emission statements. Therefore, I wanted to do something about the relation between Nick, GFQ’s environmental database and 108 2.3. Commanding Carbon the paper. Retrospectively considered, I propose that I undertook this reconfiguration of their relation because of two reasons. First, it was my part of my job to teach Nick about the correct engagement with this issue. I must admit that I did not perform this job well: in my interpretation of my task, it would have been more appropriate to stage myself in this situation more authoritatively, i. e. command Nick how to materialise carbon. However, and this is a second reason for intervention, I wanted to not close this issue, but, rather, engage Nick’s boss with this decision. As a researcher, I wondered how Simon would relate to the question. Agency to shape carbon, thus, can be redistributed among humans. Late afternoon on that day we went upstairs in the building to meet Simon. This was my final meeting with Simon and I wanted to let him know some of the uncertainties I recognised during my trip to their GCE in order to give feedback both as a representative of the HQ, i. e. as a friendly internal auditor who would not immediately provide a report to the HQ about their practices, and as a researcher. The following extract is taken from the midst of my field notes on this conversation. Nick was mostly silent and, so, Simon and I exchanged some ‘technical’ thoughts about the local situation, i. e. the measurement of Key Performance Indicators (KPI), as well as wider considerations about GFQ’s move into sustainable finance products. Field Note Extract 2.3.b (Approval by Simon) Regarding Jones & GFQ, Simon asserted: this is just a brand change, therefore we need to integrate Jones & GFQ data. Concerning waste they informed me: it does not create costs. I learned, however, getting rid of batteries seems to generate costs. Nick pointed towards some estimations. Simon replied: ‘We will put in what we know’, no insecure estimations! Simon pointed out – in general ‘we need to work on’ the data, we need time. He made clear that he communicated this to Victoria. This series of statements includes four different accentuations of how Simon shaped the relation between consumption and carbon emissions: consumption may be included in counts, it may be excluded and it may be reconfigured or modalised. This is not to say that these accentuations should be read as analytically exclusive from each other. In practice, they may and partially do overlap. Seeking to study how Simon was able to control the doing of carbon, I attend to each of these ways one after the other. Inclusion Simon decided to include the paper produced for Jones & GFQ, reasoning that the label, rather than the corporation had changed. Before he provided this statement, I had asked him about this issue. Nick had not commented. Simon was able to provide an authoritative decision because both, Nick and me, accepted it as such. The kind of hierarchy enacted here is a symbolic one. Even though data passed, in that very moment, to and through Simon, this established in no way an Obligatory Passage 109 2. Data Construction in a Basement Point (OPP). Being a boss depends on being recognised as an authority, being a point of passage depends on being enacted as such. If an actant is enacted as a passage point, that point may then perform as an OPP. As it turns out, the latter, again, may be contingent: in this case, Nick was the person who ultimately entered the data into ESDR; and I was anyway technically able to alter data within the database. Despite these contingencies, we performed Simon as authoritative; and that made Jones & GFQ-labelled paper count. This suggests, control over carbon through formal authority may be quite precarious. Exclusion The HQ wanted this GCE to state how much and what kinds of waste they had ‘produced’ in the prior year. Accordingly, ESDR was providing several categories for waste. However, at this GCE’s location, producing waste and getting rid of it did not incur costs (at least not ones which the GCE had to compensate for). Elements which were part of the business process but were freely available (as public goods or as public ‘dumps’, Lohmann 2005) have not entered explicitly the symbolic representations of this GCE’s accountants. This finding points to an enormous implication about GFQ’s Environmental Management System (EMS). We underline here that carbon accounting can only translate those acts of consumption and pollution into carbon emissions which have resulted in some kinds of costs. An enterprise is not supposed to engage with or optimise any processes which are not cost relevant – otherwise it would not act in the interest of its owners (Corporate Watch 2004). This GCE did not posses an inscription device to record waste data. Nobody would provide them with waste invoices. I assume, the informal economy did not issue appropriate material statements which Nick could have accessed and filed. Power (1996) discusses various options of how things may be made auditable. One option, he finds, is that the environment within which external auditors find facts, including the facts themselves, need to be put in place actively. Otherwise, the auditors would arrive and do not find any of the things they are expecting. Power’s point likens the waste situation in this case. Simon and Nick, otherwise friendly and supportive, simply did not seem to feel they had a substantial clue about the amount of waste their GCE produced. Power (1996, 305) calls this ‘beyond expertise’. Locally, they did not recognise suitable knowledge. How GFQ dealt with this issue I explore in the next chapter. Above, we have also seen that even if consumption was to be included, the quality of that inclusion was quite uncertain. The decision to exclude waste from being accounted for, overall, was out of their hands. While Simon could have commanded Nick to exclude specific consumed items, without outright lying – or very creative guessing, neither of them could access the waste reality they were supposed to provide an account of. Reconfiguration When Nick voiced the uncertainties of the information to be reported in the flight account (discussed in the prior section), Simon asked him to only enter facts – only properly known data. That was it. They did not consider details in this conversation. Within the rationality of expert-based environmental management it would 110 2.3. Commanding Carbon be necessary that all the problems are clarified. This resembles Porter’s (1995) account of quantification: to count in political decision-making, numbers have to be construed as hard. Callon (1998a, 265) interprets Coase (1960) in this way: an internalisation of costs requires that all the agents who are part of an economic situation are clearly identified, they are able to negotiate and they know and agree about the externalities, overflowsZ in question. As long as all actors happily agree about overflows and other technicalities, they can negotiate prices, etc. However, if they are not at peace with each other, they are likely to struggle about the nitty-gritty (Callon 1998a, 266). Here is another option: Nick and Simon have been quite at peace. In their situation, neither was there a need to discuss the technicalities nor did they agree and share their knowledges about the situation. I read their relation this way: Simon wanted Nick to perform all consumption data as factual. Simon, through his reaction, took no responsibility for problems with the data produced by Nick. Thus, the boss was not committed to the facts produced for him. In effect, he partially redistributed the responsibility to account for the data from the inscription device to his engineer. In the relationship, constructed here, non-factual data did not have any right to be visible in the task-form. Such data was othered, excluded from counting. This provides an explanation for the comment in the flight task-form, depicted in Figure 2.6 (on page 105), eventually being deleted. Through Nick’s comment, overflows could be recognised by those who were to read the data inscribed by him. However, what is achieved, through the enactment of facts is a less vulnerable position of the actor. Fact-making, then, can be understood as a move in impression management (Goffman 1959). This would suggest that Simon told Nick how to perform well. And, this case lends itself to be read in that way: Nick would be judged by his boss as performing well if Nick a) completed filling task-forms while only submitting ‘facts’ and b) omitted declaring what was instrumental for the realisation of the said ‘facts’, i. e. if he enacted whatever kind of data, he inscribed into ESDR, as factual. While Simon was even further distanced from the imagined sources of carbon emissions, he was able to tell his engineer how to perform data. By that, he could intervene in the shaping of qualculations. He was able to command what ought to come with the figures, how to con-figure. Nevertheless, the re-configuration itself was performed by Nick. Modalisation Finally, Simon let me know that he had been in touch with my boss, the head of the EMS, Victoria. By telling me, as well as Victoria, that the data was not yet in a state which he considered appropriate, he distanced himself from the current status of data inscribed by Nick into ESDR. Furthermore, he signalled he would be improving the quality of the data after a little while. Interestingly, this approach indicated a further understanding of data: this time, he recognised that data was not static and that it could be improved. Thus, data, in this perspective, are not in-themselves-stable facts where all non-facts could simply be deleted. Rather, facts were informational entities which were not-yet-ready to be read by Victoria. By providing these explications, Simon modalised the data. He added information about their modality: Again following Goffman (1959), Simon was informing me about how 111 2. Data Construction in a Basement he had performed the data vis-à-vis Victoria, and by that enacted that performance towards me as well. Thus, the way Nick’s inscriptions of consumption was made to count vis-à-vis the HQ was altered. The boss did not simply sign off or not sign off the numbers (cf. Harper 1998, 224-225) but, rather, indicated a distancing from the data. He did not make the numbers, which had been worked out by Nick, his own and was not willing to defend them. Thus, again, Simon was not committing himself to the data which had been produced for him. He had, however, decisive agency in shaping his relation to data. Through a simple statement he was able to alter the modality of data. And he could easily extend this modality by letting others know about it. ♦ End of Section ♦ This section examined the distribution of agency to control consumption data and, therefore, carbon emissions. To conclude, I analytically differentiate the ways through which that data has been (not) controlled by Nick and his boss Simon in three segments, sketched in Figure 2.8 on the next page: the imagined sources of data, the qualculatory work and the extension of data to be read by the HQ. On the figure’s right hand, we find the actants which members considered sources of data, representing what they conceptualised as an independent external reality. Any ‘sourcing’ of data from that reality was implicating dependency of the data gatherer on the availability of sources. Therefore, the network which, partially, crystallised into accepted sources, like invoices, was very much in command: its relations were shaping what could be delivered to Nick, made available to him. A setback, for example, was that the informal waste disposal economy in that country did not provide the GCE with acceptable invoices. Nevertheless, this GCE’s agents did have access to other data fitting the KPIs. And that data was, eventually, read by the HQ as being authored under Simon’s responsibility. Thus, in terms of responsibility, the data passed Simon on its way from Nick to the readers at GFQ’s centre. However, Simon was not actually positioned between the data and the HQ. Rather, the latter accessed the data via Nick’s inscriptions in ESDR. I argue, therefore, that while Simon constituted a point of passage, he was not an obligatory passage point for the quantities inscribed in the database. In spite of that, we find that somehow Simon was present in the data, in the qualculatory work (see midst of Figure 2.8). To appreciate the quality of his presence, Figure 2.9 on page 115 serves as an orientation for us to conceptualise non-obligatory forms of passage points enacted around Nick’s qualculations. OPP To recall, an OPP is defined as a point through which data has to pass in order to move on. Actor 1 needs to pass on the data to Actor 2 for the data to travel further. Actor 2, at that moment, can translate and transform the data. Then, Actor 2 is established as an OPP. In the case discussed above, Nick (as Actor 1) did not need Simon (as Actor 2) to deliver data to the HQ. Nick only needed ESDR for that. For Nick, the database constituted an OPP while his superior did not. 112 Figure 2.8: Relations of Control over Data at a GCE 2.3. Commanding Carbon 113 2. Data Construction in a Basement Contingent Passage Point I consider a point through which data may (but does not necessarily) pass, a contingent point of passage. In the meeting, Simon’s encounter with paper, waste and travel data was contingent upon other actors (Actor 1, Nick or me) deciding (if I may employ the cognitivist take, pace Latour) to pass data to him. • Thus at a contingency decision point (A), Actor 1 can also keep the data in her hands without handing them over to Actor 2. Then, only Actor 1 can transform the data. However, if Actor 1 passes data to Actor 2, the latter can transform that data. • The moment at which Actor 2 receives, potentially transforms and passes on that data can be understood as the contingent passage point. At that point, data may or may not be changed.24 Despite data having been passed to Simon, he did not pass on the data to others on his own, but passed it back to Nick – with some remarks, orders. • Now, Nick had the chance to decide what to do with the data. This constitutes a further contingency decision point (B). He could keep the data as he received it from Simon – or he could ignore the changes/orders by the boss and keep the data as it was at the time of contingency decision point (A). This binary choice, of course, does not adequately capture the point. Actor 1, necessarily, had to interpret Actor 2’s performance. Nick had to make sense of Simon’s instructions. We found: not Simon himself shaped the data, but Nick did. Therefore, Simon sough to shape Nick. Thus, Simon’s ability to shape data was contingent on Nick’s acceptance of Simon’s authority: only if the formal hierarchy was implemented ‘correctly’, could the boss order Nick. At GFQ, working for the company presupposed at least a superficial acceptance of these hierarchies. Automatic Affirmation Point Finally, Actor 1 may directly pass on data in the intended direction, passing by Actor 2. In that meeting, this was the case for most of the data produced by Nick. Although Simon was responsible for that data, he let the data pass by, i. e. Nick could leave the office without being asked to account for further data. In that way, automatically, Simon signed off all the data absent from explication. Having understood the principle of these points, we need to make one step further, and underline their situatedness. None of these points existed in isolation. The differentiation provided in Figure 2.9 on the next page serves analytically distinctions, rather than claiming that any situation can be reduced to any of these points. The situation in which Simon was in was shaped by him being formally responsible for the data produced by Nick. We can, therefore, consider it straightforward for Simon 24 Note, in theory at this point data could be passed on to another actor, rather than (or not only) back to Actor 1. 114 2.3. Commanding Carbon Figure 2.9: Obligatory, Contingent and Automatic Affirmation Points of Passage Actor 2 Actor 1 t OP P • Contingent P assage P oint • Contingency Decision P oint A t Automatic Af − f irmation P oint • Contingency Decision P oint B t that he wanted to exclude from that data such qualifying elements which disqualified the quantifiers. This interpretation is strengthened by this observation: he asked the engineer to make available only facts to the HQ, rather than comments inscribed into ESDR which would subvert facticity. This had a significant consequence: Nick was to carry the responsibility for these (dis)qualifications in person, rather than placing them into GFQ’s carbon machinery. As we have seen above, the comment in the task-form, specifying some of the qualculationary practice were eventually deleted. And the effect of reducing the amount of qualifying disqualifiers was the performance of the GCE’s consumption reality more in line with the form which (so was the HQ interpreted) GFQ was desiring. However, while the task-forms were purified – converging with Latour’s (1987, 227) proposal that to serve a centre of calculation best, ones needs to produce immutable, combinable mobiles – data was also extended. Simon wanting Nick to ensure that only facts would be inscribed in ESDR, implied that the problematic qualifiers would be stored in the human – Nick. This extension of data storage from digital material to humans was repeated by Simon when letting the HQ (my boss Victoria and me) know that he considered the task-forms, themselves, not complete. Thus, the more complete information about the GCE’s environmental data was then distributed over several humans and non-humans (see left hand of Figure 2.8; cf. Lippert 2011a). In a way, then, Simon’s activity at the contingent passage point can be conceptualised as having passed on data fragments to several actors: to Nick the data-to-be-inscribed-into-ESDR, and to Victoria and me meta-data. In parallel, hence, data on this GCE’s consumption existed from then on. This constitutes an interesting friction in the way data was shaped: on the one hand, Simon asked for purification of the facts and enacted their hybridisation on the other. To sum up, there was no such thing as a single entity or location commanding the production of carbon emissions. Nick’s qualculatory work was informed and constrained by the availability of inscription devices which he could perform as legitimate sources. Precisely because these actants were not fitting to the EMS’s forms, Nick had to 115 2. Data Construction in a Basement constantly make practical compromises. And the finalisation of his task to gather data was, in that respect, contradicted and shaped at the same time by his boss’s instructions to ensure the facticity of data. Thus, the relations between Simon and data were added to the data. Checking data with the superior, hence, rendered the data more social, rather than less. While traces of uncertainty were deleted within ESDR, they proliferated amongst the human data storage system. When only reading data off ESDR the imaginary of control would appear. If Latour is right, then control over carbon was made even more unrealistic through Simon’s demands for getting rid of the traces of Nick’s practical compromises; for ‘every trace’ (ibid., 245) has to be summarised, rather than simply deleted, to come closer to the modern myth of control over ‘nature’. This book can, in this respect, be read as a modest contribution to support the non-ESDR storage of GFQ’s emissions. ♦ End of Chapter ♦ This chapter established reference points at two levels: theoretically and empirically. Using an interwoven narrative, I, first, introduced the theoretical apparatus of actornetwork theory (ANT) (broadly conceived) and illustrated how this books employs some of this theory-method toolbox’s analytical devices while, second, scrutinising carbon accounting practices on the ground. How to sum up? In terms of counting carbon emissions? How much emissions did this GCE emit – after all the work they put in discovering the fact? The fact is, quite simply, there is no singular fact. Well, what Nick, Simon (and me) achieved over the couple of days I visited them was to establish – about – one point five kilo tons. Of carbon emissions? Well the number could be linked to it. But it was quite uncertain. Just a couple of days after I had left them, their emissions count increased! By — one point five — grams. Literally. And another twenty days later the count decreased by 230 t. Thus, summing up through numbers clearly is not promising to understand anything about this GCE’s emissions. It seems much more apt, I argued, to zoom into such facts and study how they are done. I have no objections against summing up with numbers. But a quantifier in isolation does not do. We need also look at the qualifiers: what is the quantifier linked with? To answer such questions, I visited a subsidiary of GFQ in a Western Asian country. Studying this GCE’s agents who were supposed to ‘gather data’ for the database ESDR (which produced the numbers inscribed into the prior paragraph) illustrated key notions and ideas exercised by members. With Simon and his engineer Nick, this chapter undermines a conceptualisation of GFQ’s carbon emissions as existing out there, independently from humans. To reiterate: before I had arrived to visit Simon and Nick, the boss had asked his engineer to collect the consumption data which the HQ had requested. Then, I followed Nick, the agent who was to know the subsidiary’s consumption acts, count the quantities of goods and services utilised, and report them back to Simon and inscribe them into GFQ’s centralised environmental database. The chapter engaged in depth with the core of his practices, classifying consumption artefacts 116 into the accounting frame prescribed by the corporation’s EMS and when data was not available qualculating substitute data which would fit ESDR’s forms. Engaging with these practices allowed us to comprehend the emissions’ hinterland, i. e. the base needed to construct the emission facts. Typically, an ANT study would present a list of the elements which constituted that very base. However, as others have remarked before me (e. g. Strathern 1996; Suchman 2000; Mol 2002), when zooming in, one does not encounter an end to such a list. As a researcher, and as an author, one has to cut listing. For GCE’s members, amongst the most important items they needed for the job of providing an account of consumed goods and services were (what they considered) sources, like invoices as well as metering devices, inscription devices like pieces of paper or spreadsheets and ESDR and common to all of them many numbers, qualifiers and their contexts to make sense of the former. One problem with such a list is its fractal character. If one tries to briefly describe any of these elements, imagining it as a stabile entity, one likely fails – or does not give a detailed account. Take the task-form through which the database was collecting consumption data. For that form to exist and make sense it was dependent on very different kinds of entities: as a task to a human to complete filling the form, it was utterly relying on a human to enter data. And the task needed to be authored. Not just by a formal administrator who assigned the task-form to some worker. The form was connected with GFQ’s Information Technology (IT) administrators, software engineers, with web-browsers to render the form viewable, it required concepts and language. Furthermore, the task-form had very specific requirements on how it was to be completed. It was, thus, not separate from software-embedded as well as externally located scripts specifying the user what data to enter. One thing which the task required to be meaningful were numbers. Users were to enter numbers and select among restricted choices of qualifiers like the units in which the consumption category would be measured. Other modalities needed to be stripped off the data for the latter to be enterable into the task-form. Now, if we ask what numbers are and where they come from, yes, sometimes we are referred to seemingly independent sources. Often, however, sources did not exist or did not provide the required type of numbers. Then these entities had to be constructed. Simply, informally and without making a fuss about them. Reference realities then have been members’ imaginations or experiences. Establishing the quantity of consumed goods and services was involving structured qualculations. Numbers, categories and units were separated (rather than fused, as Robson (1992, 698) would have it) to allow for calculability. They were arranged and transformed in minds, in calculators or on paper. Through very particular generalising logics, members quantified past consumption for office blocks, even ones hundreds of kilometres away. To achieve such quantifications they arranged diverse elements, as just noted, say invoices, their local knowledge about prices or consumption patterns. This was part of doing the ontics (Verran 2010) of facts about quantities of consumption. 117 2. Data Construction in a Basement Elements reflecting consumption facts, thus, do not simply exist out there nor can they be unproblematically reconstructed. Quantities and qualities are neither given nor are they intrinsically separate. Rather, these elements are the effects of practices. And, correspondingly, when attending to these practices, ‘we start to discover alternative forms of materialisation’ (Law 2008, 11). Studying environmental accounting, then, should not restrict itself to studying the cognitive and formal models of its agents. ‘Instead of talking about subjects knowing objects we may then, as a next step, come to talk about enacting reality in practice’, Mol (2002, 50) proposes. Hence we turned to the work practices through which consumption facts have been constructed. Turning to the practices employed to construct the data, we identified mostly work which normally would stay invisible; it was to be invisible: articulation work. Because the elements presupposed by the data construction guidelines did not necessarily exist, members had to translate non-fitting entities into the EMS’s categories and into the task-forms’ fields. These translations were always partial. This was the case for both, classification as well as calculation cases: for the very purpose of translation was to manipulate entities to end up with entities which would fit better. Ballast was dropped on the way, histories stripped off. We could ascertain the normative and political relevancy of this work. The specific turns within the practices were sometimes contested. Such instances revealed precisely those normative and political implications of particular construction possibilities. Enacting one version of a consumption fact, rather than another, therefore implied that specific realities would be rendered invisible during the translation process. Those realities which were not accepted by the translation logics, which was partially related to the formal classification scheme, received another status: the status of being a monsterZ – an entity which does not fit the generalisation project. For calculations can be understood as mass translations, they are bound to result in enormous overflows of monsters. I illustrated this point, tracing back the situated practical necessity for translation practices that got rid off complexities to the field of mathematics. For those workers being responsible to ‘gather’ data, i. e. create through performative writing (Kalthoff 2005, 82), such practices of generating difference, of establishing irrelevancies and non-equalities, were required to get their work done – to achieve what would appear as simple, pure facts. The latter are the effects of situated practical compromises. The signs of consumption data are not speaking for themselves. They need to be interpreted (ibid. 2002, 34), and the reader has to imagine or trace back those practicalities and methods which constitute the facts’ hinterlands – because they have been bracketed out, deleted from view in order to have them forgotten (cf. Mol 2002, 160). Clearly, the Latin factum – something done – captures better the ontics of consumption than dare, Latin for something given. Consumption data is not data but fact. Data is not given but made. Although not strictly speaking having taken place in the basement, but in ESDR, consumption facts were entering a relationship with specific carbon conversion factors. 118 Whilst recognising their politics, this book does not follow the fact-making practices into the thicket of these factors. Much rather, we engage with what happens to the data enacted at these subsidiaries. When data meets conversion factors and, by that, takes the form of reportable emissions it does not relate to the consumption actually having taken place at the GCE, but to the qualculated fact prepared by a worker. And these qualculations may and do consist of various materials and references to all kinds of consumption, which are then summarised through a specific factor; even though the latter had been designed for seemingly clearly defined purposes which may or may not much overlap with the entities actually classified into the consumption category and, consequently, black-boxed into the fact. If we accept this kind of reality of fact-making then we do not wonder anymore about the encounter with a situation in which a qualculation is enacted anew. We saw that members sometimes recognised that some entities they utilised in their translations needed, themselves, to be opened, de- and re-constructed – not unfamiliar for engineers (Newman 1998). Enacting consumption facts, qualculating the representatives of the past, is always incomplete, never finished. Hence, the different versions of emission counts I mentioned earlier. At this GCE we encountered also an exercise of hierarchy in data construction which did not fit the neat top-down models with centralised decision making or even clearly identifiable OPPs. The paths which data actually were taking was highly contingent and partially very precarious. It was transported by heterogeneous carriers, which themselves formatted the data – rendered it as information. Some of these data transforming actants, like the engineer at the basement, had much more agency in shaping the data (which, ultimately, was read by the HQ) than others (e. g. the engineer’s superior). Ordering the ordering of data was not a simple task. Within the network of entities enacting each other, anybody else’s authority needed to be accepted by oneself for that governing to work out. The engineer was not forced to accept any data he read into invoices nor to accept his superior’s order to ensure the appearance of data as factual (by rendering all disqualifying modalities invisible). Even though prescriptions have been formulated, whether they were followed was an in-principle independent question. It is most noteworthy that there was no grand struggle involved. Most data was, anyway, automatically affirmed formally simply through the practice of bringing data into a room and nobody speaking about it. Nevertheless, the interventions by the superior, suggesting that the engineer better gets rid off the problematic modalities, raise another issue: this chapter conceptualises the order by the superior as having performative effects on two data-carrying actants. First, the data record within the database was cleaned up. And the humans involved (now, including you) knew perfectly that data has been deleted from the record. But this metainformation still exists – distributed among us. In order to carve out the book’s line of argument I have to point to this issue as something the humans have to struggle with, at least for the moment, for themselves (well, maybe, let’s meet up and see what we could do about it. . . ). It definitively raises a question for further research: 119 2. Data Construction in a Basement what happens to the knowledges about the silenced modalities within the workers who are part of environmental accounting? This is, however, not the route of investigation in this book. We follow the entities inscribed into the database. To follow that data is relevant because, as I establish below, these facts still perform – in wider political and market relations. And, if, as argued above, the purification practices make the political recognition of their ontics unlikely, then it is significant to trace the normative and political existence of that data until it is leaving the corporation. We also need to understand in more depth how the data produced in the subsidiary discussed above is related to GFQ’s carbon emissions overall in order to evaluate how promising carbon management can be. Definitively, this chapter cautions us in taking numbers in combination with a unit and a category – be it a gigaton of carbon emissions or 70,425 km of long-haul flights – at face value. Behind these signs, a load of qualifiers and assumptions have been hidden – precisely because revealing them would render carbon management impossible: calculability would be lost. For the HQ this was a real threat: not only came my colleague Elise across some of the uncertainties in this Western Asian GCE, but also my boss Victoria. While governing GFQ’s overall data collection, recognising that even at the level of operations at GCEs data was not completely clear must have been disturbing. Yet, for managing GFQ as a green company, it was vital that they could utilise consumption facts in their calculations. The following chapter, therefore, turns to this question: how the HQ related consumption data to the huge political issues like the environmental crisis, sustainable development and climate change. 120 CHAPTER Translating Data Into Sustainability In the prior chapter we have visited an engineer and his boss employed by a Western Asian subsidiary of GFQ. At this subsidiary we observed how they assembled data inscribed with the claim to represent several quantities of goods and services consumed by this GCE in the prior year. This data was, eventually, submitted to GFQ’s HQ. In this chapter we follow how such data was translated into so-called sustainability. For that I turn to the addressees of their data: GFQ’s environmental and sustainability managers. This chapter shows why they asked for such data. As will become evident, collecting data from largest GCEs was key for the corporation to enact itself as a rational agent of ecological modernisation. The role of this chapter is to trace the links between local practices all over GFQ and the global discourses of sustainable development, ecological modernisation and climate change. I show how the approach taken by this multinational is an instance of the widely spread take of turning environmental issues into accounting matter and translating the latter back into visionary materials, promising the green conduct of corporations – a circular movement. However, I argue that such a promise can only be produced through systematic ignorance of carbon emissions’ contingency and the overflowsZ produced in this circular translation exercise. In effect, carbon accounting moves an engagement with the environmental and social problems caused by corporate conduct out of hand, out of control. On my first day of formally working for GFQ’s environmental managers – a winterly monday with lots of snow and ice outside – Victoria introduced me to the Environmental Management System (EMS) and my tasks in it. After sketching the organisational status of the EMS, naming the GCEs which had been the forerunners within GFQ, she explained how her colleagues put the EMS into reality. Field Note Extract 3.a (‘Give me your data!’) The actual way it looks like is that they approach the facility managers and tell them: ‘give me your data!’ This stand indicates a hierarchical relationship between the EMS-Team and the facility managers. It was the EMS-Team which seemingly started the data collection process by 121 3 3. Translating Data Into Sustainability telling the GCEs to provide data to them. According to this statement, the EMS-Team was in a position to wait for the data. When I had first contact with Simon (the boss of Nick, encountered in the prior chapter), he announced having the data ready. Thus, he knew, he had to provide data to the EMS-Team. The data collection process was based on this shared understanding: GCEs are responsible for providing data. The construction of carbon emissions depended on data made available to the EMS-Team through the database ESDR. As part of her introduction to the workplace, Victoria continued pointing to some of the current problems of the EMS – which I discuss throughout this book – and to the cooperation with my new colleagues. She told me that the moment I was entering the field was a phase of change: the EMS was increasing the number of participating GCEs, yielding an even higher complexity; and an external auditor had criticised some of their practices. At the end of this introduction she handed over a couple of documents. One of these was a 86-page internal presentation, titled ‘EMS for GFQ’ (EMS4GFQ). This document emerged as an authoritative document in the field – it was the one Nick had used as a reference document, and EMS-Team members quoted from to point subsidiary agents to definitions which were to be utilised. In the succeeding section I sketch this document and, through that, show how GFQ performed the link between collecting data from GCEs to the grand discourse of sustainability. Afterwards I turn to actual practices of ‘data collection’. This is how they referred to the process reciprocal to the one sketched in the prior chapter: the EMS-Team figured itself as collecting data from the globally distributed subsidiaries of GFQ. These subsidiaries could be little companies with only a few hundred employees as well as own corporate groups – which, again, incorporated many subsidiaries. Studying data collection practices opens several questions. To provide the ground for engaging with these, the subsequent sections serve to introduce three coordinating techniques of the EMS: I make explicit how GFQ imagined the boundary of the environment which they accounted for; I show how they rendered their environmental accounts comparable and commensurable; and we see how members figured being able to provide a full account for GFQ even though they were never able to incorporate all of GFQ’s sites into the EMS. I conclude this chapter with a discussion of the technology through which GFQ’s impact on ‘the environment’ was made readable and viewable for both, internal and external audiences: GFQ’s Sustainable Development Report’s presentation of the employee carbon footprint. Using existing Foucauldian discourse analyses of sustainable development (Dingler 2003) and scrutinising members day-to-day discourse, language, motive, meanings, ways of acting (Clarke and Star 2008, 116), I argue that GFQ managed to translate the global discourses of sustainable development into the corporate EMS and to enact through that system so-called facts about GFQ’s environmental performance which were figured as carbon emissions. GFQ employed these emissions to provide publics with a vision of GFQ as a sustainable agent. Thus, the circle of translating the discourse into data and 122 3.1. Breaking Down Sustainable Development back was closed. What was lost on the way? This chapter details the realities which needed being ignored to enable neat facts to emerge which would be compatible with the requirements of presenting GFQ as a knowledgeable and reliable corporate citizen. 3.1 Breaking Down Sustainable Development This section opens by indicating key claims inscribed into GFQ’s official and publicly available Code of Conduct which linked the corporation to sustainable development. The focus of the analysis then swiftly shifts to a particular document which Victoria had handed to me: the MS PowerPoint presentation EMS4GFQ. To understand this document’s position, we need to note that this was an internal document. It was used only by environmental managers; and it was neither designed to be distributed to GFQ’s board of directors nor to external audiences.1 Hence, EMS4GFQ was not screened for political correctness. It intended to introduce GCEs’ environmental agents to the necessary background and specifics which they needed for accomplishing their tasks. Members of the HQ based EMS-Team enacted this document as the de-facto authoritative text to quote from when communicating with subsidiary agents. Within the relations among GFQ’s designated agents of ecological modernisation it was the official spokes entity for how the EMS worked. Outside of these agents’ relations, this organisational-technical document was not of interest for GFQ. If some arbitrary workers wondered whether GFQ was caring for the environment, then they could, for example, consult GFQ’s Sustainable Development Report, the Sustainable Development intranet portal or the corporation’s Code of Conduct. I use the latter text as a point of departure for this section’s analysis. The study of these documents seems promising because official language ‘tacitly [lays] down the dividing line between the thinkable and the unthinkable, thereby contributing towards the maintenance of the symbolic order from which it draws its authority’ (Bourdieu 1977, 21). Documents, thus, provide fragmental access to members’ discourses (Keller 2005). They are part of the ‘concrete and material ’ character of discourses (2011b, 48) which are performed in the midst of heterogeneous assemblages co-constituted by humans, non-humans and their practices. I received the Code of Conduct alongside the contract as a worker for GFQ. It was also a document publicly available; and it was presented to stake- and shareholders to provide evidence of GFQ’s transparency policy. The following quotes (anonymised) indicate how the corporation discursively situated itself. 1 As an exception to this, auditors were given access to the document. See Section 4.3. 123 3. Translating Data Into Sustainability [GFQ] is brought to life through the trust of our customers, shareholders, employees and public in the benefit and integrity of our group. [. . . ] Through our initiatives for the UN Global Compact programme and the acknowledgement of the OECD Guidelines for multinational corporations we integrate sustainability and social responsibility into our business. [. . . ] All statements of [GFQ] have to be complete, fair, definite, prompt and intelligible. Artefact 3.1.1: Extract from ‘Code of Conduct for Business Ethics and Compliance’ With this document, GFQ was enacted as a living creature which required being trusted as a condition for its existence. The corporation recognised several audiences, such as ‘customers, shareholders, employees and public’, all of which needed to exercise trust in the corporation’s practices. Especially, the authors of the document claimed, these audiences needed to rely on GFQ’s promise of providing ‘benefit’ and enacting ‘integrity’. Both latter concepts link to a number of grand claims: profit and advantage through GFQ’s existence was semiotically linked to the organisation’s honesty and strong moral principles; the Code of Conduct presented an internally consistent and uncorrupted corporation acting as a unified whole (Oxford Dictionary 2011a; Oxford Dictionary 2011b). The artefact’s last sentence provides prescriptions for GFQ’s own practices of text production. The image of integrity is continued in the terms ‘complete, fair, definite, prompt and intelligible’. These terms resemble the moral and technical connotations of ‘integrity’. Officially, GFQ figured itself as providing texts which were unabridged, legitimate, exact and comprehensible. At the same time, the corporation relates to other global players like the UN and the Organisation for Economic Co-operation and Development (OECD) and global discourses of sustainability and social responsibility. In this book, we are engaging with GFQ’s claim of integrating sustainability into its business by way of following the notion of ‘the environment’. This is were we jump within the Code of Conduct – to the section of nature and environment. Textually situated amongst issues like corruption, money laundering and terrorism, the reader would learn about the corporation’s interest in ensuring that its resources are well treated (i. e. not countering GFQ’s interests) as well as that natural resources are to be protected by their employees. EMS4GFQ linked to this paragraph of the Code of Conduct. By that we learn that the environmental management unit enacted itself as legitimised by GFQ’s general policy on how the corporation imagined its own conduct. Both documents overlapped in suggesting that globally all employees of GFQ were asked to reduce the ‘potential negative impact of internal operations on the environment’. EMS4GFQ proceeded on its next slide to GFQ’s cross sector plan of enhancing the EMS and of assessing GFQ’s total ‘internal greenhouse gases (GHG) emission footprint’. Victoria, as part of introducing me to the new job, explained how the EMS and her team’s activity of incorporating even more GCEs into the EMS were officially mandated by the board of directors. And so she also legitimised the corporation’s aim of reducing its carbon emissions by 25 % from 2006 to 2015. The reduction was also established in the presentation a couple of 124 3.1. Breaking Down Sustainable Development slides later.2 The move of EMS4GFQ from quoting the Code of Conduct to establishing an EMS and to assessing and reducing GFQ’s carbon emissions establishes a generative entry point for a critical analysis. GFQ figured itself as being able to reconcile environmental and sustainability interests with its interests as one of the largest multinationals worldwide. I argue that GFQ was not at all isolated with presenting such a claim. Much rather, this claim is wide spread. Here is how this claim links to the hegemonic discourses of sustainability and greening capitalism: Sustainable Development Normally ‘sustainable development’ is traced back to the report Our Common Future by Brundtland et al. (1987). However, already before the concept has been used. Dingler (2003, 215) identifies it in the World Conservation Strategy published in 1980 by United Nations Environment Programme (UNEP), the International Union for the Conservation of Nature (IUCN) and the World Wildlife Fund for Nature (WWF). Others have provided detailed discourse analysis of sustainable development.3 Therefore, here, it suffices to explain how GFQ could discursively link to this concept. To recap, following McNeill (2000), after publications like Silent Spring by Carson (1962) and The Limits to Growth by Meadows, Meadows, Randers, and Behrens III (1972) social and environmental movements as well as governmental authorities increasingly engaged with what was conceptualised as a ‘global ecological crisis’ (White 2006, 62). Radical analysts postulated fundamental changes of the political economy of the environment and society-nature relationships (e. g. Bookchin 1962; Enzensberger 1996). Global environmental risks were perceived as rather new and/or uncontrollable (Beck 1992; Skirbekk 1996). An analysis of the hundreds of years in which elites experimented with controlling local and global environments was not recognised widely (Glacken 1967; Thomas 1983; Grove 1996; Adams 2003; Clark and York 2005). In parallel to the increasingly recognised environmental crisis after World War II, colonial engagements were reframed as development politics. Modernisation theories suggested that Western techno-economic pathways were to be used as role models for the Global South; and dependency theories argued that the countries of the ‘periphery’ (Prebisch 1964 cited by Jacob 1997) could develop along alternative trajectories (not those of Western modernisation) (Eblinghaus and Stickler 1996, 22). In the 1980s, developmental and environmental discourses started to closely interweave. Our Common Future was a product of this trajectory; as was the Earth Summit in 1992. A product of the latter conference, Agenda 21, provides recipes for reconciling modernisation of the Global South, industrial development, spreading capitalist markets and saving earth. The document suggests that governments act: 2 Artefact 1.1 (on page 24) provides a view on the communication of the emission reduction aim. See especially Eblinghaus and Stickler (1996), Dingler (2003) and, for a full presentation of how I understand the relations of the discourses of sustainable development, ecological modernisation and environmental management, cf. Lippert (2010a, 2010b). Brief discussions of sustainable development’s etymology and struggles over defining a ‘concept’ provide e. g. Mebratu (1998), Brand (2004). 3 125 3. Translating Data Into Sustainability To promote and support policies, domestic and international, that make economic growth and environmental protection mutually supportive. (2.9.d) [And the authors reasoned:] The improvement of production systems through technologies and processes that utilise resources more efficiently and at the same time produce less wastes – achieving more with less – is an important pathway towards sustainability for business and industry. (United Nations 1992, 30.4) Growth was to be sustained by moving the limits which had been recognised some decades earlier (Dingler 2003; Næss 2006, 200). Efficiency gains were to be found through science and technology. Factor 4 by von Weizsäcker, Lovins, and Lovins (1997) argues that doubling the efficiency in production will result in twice the welfare. Business was able to welcome environmental regulation as reliable market conditions (Christoff 1996). Participation loomed everywhere. Sustainable development entailed the integration of NGOs into policy-making. As a result of the new discourse, about everything could be related to sustainability. Swyngedouw (2010b) sketches the diversity of ‘sustainabilities’; ranging from sustainable environments over sustainable companies and sustainable markets to sustainable poverty and loss. The complexity of performing projects as targeting all ‘three pillars’ of sustainable development – the social, the economic and the environment – is normally resolved through techno-managerial rather than social/political solutions (Blühdorn and Welsh 2007, 190). To design sustainable futures, old-fashioned development and capitalism were to be modernised. Næss (2006, 201) reads the discourse as proposing that simply hegemonic institutions have to be willing to learn; in the interaction of new technologies, better regulation by the nationstate and the ‘invisible hand’ of the market with e. g. consumers, pressure groups and companies sustainability is imagined as globally and peacefully emerging. This view is continuously reproduced; accordingly Huber (2008, 360) argues: Moving beyond its old-industrial stage, the modernisation of society now also entails ecological modernisation, , readaptation of industrial society within the global geo- and biosphere by modern means such as a scientific knowledge base and advanced technology in order to upgrade the earth’s carrying capacity and make development more sustainable. The discourse of rendering economic growth compatible with reducing environmental impacts, e. g. by reducing the amount of waste and increasing efficiency, is, thus, shared with GFQ’s Code of Conduct and the presentation EMS4GFQ. Both GFQ documents voiced that sustainability has to become part of the corporation’s internal operations. In the rationality of the discourse, economic and political actors only have to ensure their internal sustainability because one’s actors externality is another one’s internality. If only all render their practices profitable, green and socially just, then – summed up – sustainable development is truly taking place. 126 3.1. Breaking Down Sustainable Development Also institutionally, GFQ’s CSR unit was interacting with the sustainability discourse’s players. Regular meetings took place with UNEP, a global environmental NGO4 and leading ‘sustainability management’ academics. GFQ was part of publicly postulating action on climate change, interacting with governments, sending messages to the Kyoto Protocol follow-up conference at København (2009). Yet, like Agenda 21, the corporation was interested in continued growth as the leading business principle. Climate Change The discourse of sustainable development is nowadays seen as self-evidently connected to climate change (Ngwakwe 2012). Solutions to climate change need to be sustainable and sustainability politics need to engage with global warming. Swyngedouw’s (2010b) analysis suggests that the hegemonic discourse expects legal and technical instruments, like the Kyoto Protocol and ‘less carbon-intensive’ machinery, to bring about climatic sustainability. Climate change is understood as being caused by greenhouse gases, especially CO2 (IPCC 2007). Above, we have already seen that climate change is rendered as market failure (Stern 2008). It is taken-for-granted that nobody doubts that CO2 is the culprit.5 Hegemonic climate politics has transformed climate change into a question of the accumulation of legally recognised GHG; the latter are abbreviated as CO2 e, i. e. carbon equivalents. At GFQ’s HQ, EMS-Team members simply referred in shorthand to ‘carbon’, as do many, if not most, of those who engage with climate change and carbon markets in their everyday. Making different kinds of GHG as well as emission sources commensurable allows for models of market mechanisms to imagine the optimal, i. e. pareto-efficient, allocation of emissions (MacKenzie 2009a; Sullivan 2010, 115). At the same time, under the Kyoto Protocol emission trading allows corporations to reduce their carbon footprint by buying CERs, e. g. negative emissions sourced from so-called CDM projects in the Global South.6 In the EU, owners of installations which emit carbon beyond a legally determined threshold have to acquire CERs. While this policy addresses e. g. the cement industry or fossil-fuel based energy producers at large-scale, GFQ reacted to the climate change discourse as well. Not only has it become discursively straightforward for companies to publicise that they try reducing their emissions – for GFQ climate change also constituted a business risk. GFQ’s insurance branch was part of the global discourse of insurers who perceive climate change as leading to weather catastrophes, resulting in damages. If actors are insured against such damages then they receive compensation – expenditures for the insurer; damages of poor households are not immediately problematic for the insurer as these are less likely to be insured (Botzen, van den Bergh, and Bouwer 2010). Finally, GFQ was also acting as a broker and advisor in the carbon market. 4 Section 4.3 discusses in more depth the relation to this NGO. So-called climate change sceptics and deniers question merely in how far climate change is induced by human activity. There is evidence that the denial movement is heavily interwoven with conservative anti-environmentalist elites (Jacques, Dunlap, and Freeman 2008). 6 Böhm and Dabhi (2009), Ninan (2011), Bailey, Gouldson, and Newell (2011) can be of help to navigate the literature on carbon markets, the CDM and their problems. Lohmann (2009b), Blok (2010) link these discussions to Science and Technology Studies (STS). 5 127 3. Translating Data Into Sustainability Accordingly, within the internal universe of GFQ climate change was also a topic. The corporation had set its own carbon reduction target. An intranet portal of the CSR unit provided employees and GCEs with information on how GFQ reacted to climate change and the sustainability discourse. Artefact 3.1.2 allows an anonymised glance at this portal: a glossary providing information on the key concepts of the unit. ‘Environmental Management System (EMS)’ was one of these. The artefact links business, climate change, internal environmental protection and credibility. GFQ Artefact 3.1.2: Intranet Definition of EMS On the left hand we find a standard feature of GFQ’s intranet web configuration: a graph provided an up-to-date view on its stock quotation. The definition’s authors indicate GFQ’s claim to be best-in-class with respect to business practice of tackling global warming. The statement implies that such ‘leadership’ is necessary because the corporation’s ‘business is highly affected’. This links to a prior point I made: the financial sector performs itself as having a business in ensuring that clients do not experience damages and that GFQ’s investments are not threatened by weather catastrophes. Further, the authors suggest, the way the corporation addresses climate change needs to be ‘credible’. This resembles GFQ’s Code of Conduct. Credibility is imagined to be achieved by greening GFQ’s ‘own operations’. Thus, we learn, GFQ’s EMS is defined through the corporation’s practice of greening its internal business practices in order to perform well as a credible leader of businesses addressing climate change. The EMS emerges as a reaction to GFQ experiencing itself as being ‘highly affected’ by climate change. Imagining an EMS as a Solution I argue, the EMS, as a reaction to climate change, was discursively enacted as contributing to fighting climate change. This entails briefly reviewing two of the key strategies postulated to bring about sustainability. The sustainable development discourse is premised upon both, the need for change and the gradual quality of that change. Fundamental concerns about the ideals or instruments of sustainable development and the realisability of sustainability under capitalist or industrialist conditions are excluded (Pepper 2005). Much rather, the discourse emerged precisely through overcoming the postulation of any fundamental opposition. Internally, the ‘common sense’ (Gramsci (1971, 326) cited by Crehan (2002, 110-115)) of sustainable development is firmly rooted in trust in gradual change, sometimes entitled ‘reform’ (Prasad and Elmes 2005). Accordingly, step-by-step environmental sustainability is imagined as put into practice by increasing efficiency, innovating technology, managing environments globally, de-coupling economic growth from material resource 128 3.1. Breaking Down Sustainable Development use, dematerialising production as well as transforming economy through environmental economics. Within the discourse of sustainable development, attempts towards putting these aims into practice are seen as the only means to the desired form of progress. Radical strategies are not perceived as ‘pragmatic’ (Prasad and Elmes 2005). For the development of this book’s argument, the strategies of increasing efficiency and (global) environmental management are most relevant. (1) In terms of the efficiency aim two approaches are equally valid: increasing the output relative to a unit of input; or decreasing the input relative to a unit of output. The incentive to optimise a corporation’s efficiency can be understood as intrinsic to a competitive economy (Eblinghaus and Stickler 1996, 75-77). Thus, ‘eco-efficiency’ is seen as promising ‘winwin’ situations: both, the environment will be protected and the relative economic gain increased (Prasad and Elmes 2005, 848). Efficiency is to be brought about by improving production processes and its substances. Normally this requires expert knowledge (Dingler 2003, 244-246). A business’s operations have to be known in detail to be able to locate sites for improvements. (2) Global environmental management promises the rational administration of resource and energy flows on earth. Jänicke, Kunig, and Stitzel (1999, 143) suggest that, often, lay people are not able to recognise the required actions to bring about sustainability. Consequently, experts have to gather knowledge about the earth’s climate and ecosystem. This allows them to calculate e. g. carrying capacities or other thresholds, which can be communicated to globally distributed agents of change, such as nation-states or multinationals. For instance, during my field work GFQ recognised the 2◦ C aim proposed in the climate change discourse: science calculated that global warming below the threshold of 2◦ C is unlikely to lead to major catastrophes (Szerszynski 2010). To achieve sustainability, the role of the nation-state under scientific management of environments is reconfigured (Dingler 2003, 246f.). If only complete information, knowledge about all (natural) ‘laws’ and immanent dynamics were given, then the global system could be ruled by determinism; thus, internal processes and the effects of external influences have to be calculable, the system has to be scientifically analysable and manipulable (Dingler 2003, 247). These presumptions are inscribed into instruments like emission trading or the understanding that internal experts of a corporation may posses the knowledge required to manage the corporation’s impacts on the environment. As a policy, these strategies to render societal and business practices ecologically sustainable are usually addressed as ‘ecological modernisation’. This approach emphasises two moves: from end-of-pipe technologies (like a filter) to proactive technologies (like cleaning production processes themselves) and from regulating specific actors to subjecting actors to market mechanisms. Both, the innovation of technologies as well as the proper construction of markets require expert knowledge. Buttel (2000, 63-64) suggests that ecological modernisation is synonymous to sustainable development. The language of ecological modernisation would allow the conceptualisation of environmental success stories: any experience of win-win solutions is readily conceptualised as ‘best 129 3. Translating Data Into Sustainability practice’. Studies of the latter abound. Making success stories publicly available may set standards by influencing governmental regulation – in terms of both shaping and/or evading regulation (cf. Newton and Harte 1997, 89). At the same time, Hajer (1995, 26-29) observes as a key element of ecological modernisation policies that polluting actors have to carry the burden of proof in cases of environmental damage. And Mol and Sonnenfeld (2000b) suggest that ecological modernisation is actually taking place: their approach, Ecological Modernisation Theory (EMT), identifies that modern industrialism and capitalism manage to come close to sustainable development by means of ecologically modernising their processes (Gibbs 2000). Environmental standards would result in a ‘race to the top’ among both, corporations as well as nation-states (Drake, Purvis, Hunt, and Millard 2003, 165) if nature is appropriately internalised into the capitalist market. Mol (2010, 23) proposes such an ecological restructuring has indeed been taking place and explains it ‘as the growing autonomy, independence, or differentiation of an ecological rationality vis-à-vis other rationalities’. As evidence for this he quotes the ‘widespread emergence of environmental management systems in companies’ (ibid., 25). EMT argues that such institutional ‘changes have some permanency and would be difficult to reverse’ (ibid.). If we can identify such changes at GFQ, then we need to investigate what it is that gains permanency. He also proposes that financial services providers, like credit institutions and insurance companies become actors which spread socially the ecological restructuration. If EMT is right, then we can expect the major reconfiguration of societies and economies towards sustainable development to take place. After all, ‘everybody’ agrees that we need sustainable development; those who are not agreeing are virtually not considered somebody – radical green activists are criminalised and framed as eco-terrorists (Salter 2011, 216). Hegemonic actors have accepted that action and substantive change, not rhetoric, is needed (Blühdorn 2007, 252-253). However, Blühdorn questions whether these statements are not merely adding another layer of lipservice and spin doctoring. The risk is that hegemonic actors provide complex simulations of decisive action, rather than actual change. The point of this book, then, is to engage with a key technology which is supposed to render corporations environmentally sustainable; environmental management systems are best practice (Pojasek 2010). We need to investigate substantially the culture of the EMS’s agents: what are they doing, how are they achieving it. The prior chapter already indicated: there is more going on than superficial simulation. This discussion substantiated my claim that GFQ statements in its Code of Conduct and the presentation EMS4GFQ fit perfectly the discourses of sustainable development and ecological modernisation. Before continuing the actor-network theory (ANT) analysis commenced in the previous chapter, I now turn back to GFQ and how the corporation imagined its EMS. GFQ’s EMS To implement an EMS, the definition given as Artefact 3.1.2 (on page 128) was not useful. While it helped the reader to recognise that GFQ positioned itself in the sustainable development discourse, it was not providing much information about 130 3.1. Breaking Down Sustainable Development what or how such a system ought to do. Maybe this was the reason why Victoria asked me to propose a better definition of what an EMS is. She wanted me to provide an alternative to the definition shown as Artefact 3.1.2. A couple of days after she asked me for the definition I sent her an email with this definition: An Environmental Management System (EMS) refers to the quality loop of an organisation for continually improving its environmental performance. Within an EMS all activities are to be planned, carried out and documented in a systematic manner. It consists of defining the organisation’s environmental policy, setting up an environmental programme, implementing its measures, evaluating the success of measures, reporting the status of the organisation’s environmental implications and, finally, reviewing the environmental policy. Artefact 3.1.3: My Proposal for Defining ‘EMS’ for GFQ I had produced this text primarily based on my prior studies of environmental management and ecological modernisation.7 Thus, the direct source for this statement was my university education. At that stage of my field work, I did not have carried out the analysis which you are reading now. Providing this statement entailed two risks: on the one hand I might have not performed my task well. Maybe my prop