Among all the pictures of politics made and circulated every day there are very few of civil servants at work. This has something to do with conventions of anonymity and impartiality, preserving a distinction between government and administration. Our gaze is usually drawn instead to the theatre of politics, the performance enacted precisely in order that it be seen. But what goes on backstage, beneath the surface of appearances?
Guy Tillim’s photograph taken at City Hall in Lubumbashi shows a man at a desk, writing. The desk is covered in forms and files; there are more files on the shelf beside him, on the floor and presumably in the filing cabinet, too. There’s not much else: a battery offers a fragile-looking power supply, and there’s a chair for anyone who might want to come and consult this man, or perhaps bring him a new form or report, another document, more paper. This is government at a bare minimum, and that minimum comprises the writing and recording of actions, situations and decisions on paper.
Yet how should we read him, the man writing? There is some pathos in the picture, which implicitly juxtaposes the ambition of government with its rudimentary appearance here. But there is dignity, too: the man is calm, intent, not just doing something but ensuring it is done properly, keeping everything in order, maintaining the function of the system of which he is a part. Should we be more astonished by the conditions in which he works, or by the fact that he is there at all, and that this is what he is doing?
Towards the end of 1999, around 30 public health officials, researchers and representatives of interested organizations met in Gothenburg, Sweden to discuss Health Impact Assessment, then a new policy instrument being tried out in different contexts around the world. The meeting was part of a broader HIA initiative led by WHO’s European Centre for Health Policy.
The discussion was led by a paper which described some of the different models being explored in Germany, England, and Sweden as well as the US, and reviewed the health impact of the EU’s Common Agricultural Policy. The meeting in turn led to the production of a consensus paper, described as a ‘living document’. The paper was subsequently widely distributed and presented at both international and national-level meetings, and was translated into languages other than English.
Once people no longer sit round a table, it is the document which mediates between them. The paper or report does two things: first, it tells the group what it is by what it has done (it turns the meeting into a group), and second, it opens connections with other groups. For documents serve to coordinate behaviour as well as communicate information. They provide groups with a common language and vocabulary, helping them express to others what it is they are trying to do.
Minutes of meetings, reports, plans, statements, agreements and press releases are all ‘symbolic encodings’ which enable actions to be reproduced in time. This is a process of ‘externalization’, by which group activity produces objects which achieve an existence of their own. The object must be stable enough to be recognizable, but flexible enough to allow for further interpretation and adaptation. A key function of the document, then, is not necessarily to fix meaning, but to make continuing interpretation possible, to reproduce and extend the interactions it promotes.
Paul Cloutour was head of neighbourhood development for the city of Nantes 1989-2001, then director of urban policy 2001-2008 before taking responsibility for its citizen dialogue project. In a memoir, he takes due account of his relations with politicians, of the importance of innovation and risk-taking, of knowing how to ‘engage’ rather than to plan. But then he turns to ‘the power of writing’ in the work of policy, and reflects on the fluid and open-ended quality of what he describes as ‘A fuzzy occupation’:
If we not only define ourselves in writing, but in writing policy also define societal questions and how they might be addressed, it’s no wonder that beginning to write may appear as difficult as beginning to speak.
In Biegelbauer and Grießler’s study of law-making in Austria, officials spoke, like Paul Cloutour, of drafting legislation as a trade or craft. Though they had civil service guidelines they might refer to – the estimable Handbuch für Rechtssetzungstechnik – knowing how to write in practice was a matter of learning by doing, and of knowing what others had done.
Parliamentary committees ‘work through reports’, as Stewart says, but they don’t write them. The writing is done by a clerk, a civil servant writing in the name of a committee, or of its chair. Authorship, and the authority and responsibility associated with it, is separated from the work of writing. This is a difficult and deeply political task, for it is in the writing that a committee’s differences and uncertainties are exposed, and either evaded or overcome.
The relations between writer and reader are relations of power, at least in the way we ordinarily think of them. The writer appears to be the source or origin of what is written, the reader its passive recipient. Authorship seems to confer authority, and that authority is produced and carried by the written word of the document. Different forms of authority – religious, legal organizational and professional as well as governmental – may be established like this in similar ways.
Two features of the policy document further enhance its social and political significance: its abstraction and its plurality. When we speak, we invariably know who we are speaking to and where they are; oral communication happens in the moment of interaction. When we write, we frequently have no such indexicality on which to draw, and mean what we write to be valid in more than one instance: in this way, the language of the document comes to be cast in less definite, more abstract and generic terms.
Meanwhile, documents of whatever kind tend to cite, invoke, refer to or simply imitate others. A document acquires status by being recognizable as one of a kind: the authority of any single document is reinforced and maintained by the set or system of which it is a part. The document is a networked object.
In his study of the World Trade Organization’s Dispute Settlement Body, Nicolas Lamp explains what makes a document official, as well as how documents are placed in series and so provided with a context given by other documents.
But there is a sense in which the document makes its author official, too: since only its members and parts of its institutional apparatus such as committees and working groups are entitled to have statements issued as official documents, the documentary system serves to demarcate insiders from outsiders. In so doing, it validates what an author wants to say, giving it the status and authority of an official WTO communication, though it also holds that author to account, in the sense that a document can never be withdrawn, only added to, revised or corrected.
Max Weber begins his seminal account of bureaucracy by listing its principal characteristics. The third of these, after fixed jurisdiction and hierarchy, is the written document, or the file. By the same token, the great anarchist Michael Bakunin argued that authority might be overthrown by destroying the documents on which it rests. Weber dismissed that idea as naïve, though it’s significant that one of the first acts of protest following the fall of the Berlin Wall in 1989 was to destroy the files kept in the headquarters of state security.
The text is created in prose, forming a kind of narrative or account of an issue or problem and what might be done about it. It may take the form of a report or memorandum, perhaps a proposal. It has an author and an audience or readership, and its grammar is that of subject, verb and object. It begins top left and continues line by line, paragraph by paragraph and section by section, toward the bottom right.
The form uses space differently, separating areas of the page from one another by borders and boxes rather than the rhythms of communication in prose. For it seeks information, not sense: it requires that the reader become an author, completing it, filling it in, and its purpose is to render that authorial subject comparable with others. It deals in categories, operating in generic, abstract domains above the idiosyncrasies of individual cases. Its logic is less that of an author than of an administrative engineer; it seems to be addressed if nothing else to other forms.
What, then, are the properties and affordances of the file? They are cognitive and communicative, but like their individual components, irreducibly physical and material, too.
One of the functions of the file is to commute individual into collective agency. Matthew Hull’s study of bureaucracy in Islamabad was conducted in specific political circumstances in which taking (or being accorded) responsibility for an action puts an individual at risk. The movement of a file, its accretion of initials, actions and referrals, many of them recorded in the passive voice, is the way in which individual agency and responsibility recedes and blurs, and becomes superceded by institutional, organizational authority.
The file is heterogeneous, both physical (a folder) and abstract (a domain or issue which becomes a specifiable object of ministerial attention and responsibility). It brings a set of documents into relation with one another. It is unique: if it is to support any coherent or consistent action (including doing nothing) it matters that there is only one file in which documents relevant to a specific issue are collected. It is mobile, bringing together a package or bundle or box of papers which can be passed from one official to another, from subordinate to superior and vice-versa, from one department or unit to another. As it is annotated by officials as it crosses their desks, it becomes ‘a chronicle of its own production, a sedimentation of its own history’.
Cambrosio, Limoges and Pronovost show how biotechnology policy in Quebec is developed in and through the file or dossier, a native category which forms the ‘unit operation’ of ministerial activity. Its principal component is an inventory of the domain to which it refers, and this matters so much because it defines (governs) both the relationship between the documents contained in the dossier, and the relationship between the dossier and the world.
The dossier might contain notes or copies of related other previous government decisions, plans and commitments made by other officials, departments, ministries, as well as those of other governments and International Organizations. It may include information and data from government and other sources, and may make reference to generic government regulations and procedures. It sorts them even if only by listing them, and is itself already a kind of sorting: what is relevant and significant to the problem at hand is what is in the file, and what isn’t in the file is not.
The inventory reduces the complexity and fuzziness of the external world, and represents that world according to the internal constructs of the ministry. It describes the domain to which the dossier relates in terms and categories which make sense internally, to the government; civil servants are then able to work on this representation of the world, rather than on the world itself. They translate from the dossier into specific, local programmes and tasks, and then report on those initiatives in the terms it sets.
Of course, those responsible for it need to develop and maintain a reciprocal relationship between the account of the external world and the actual organization of that world: this is the purpose of a range of communications and consultations, as of the advisory boards and committees which serve as interfaces between internal and external actors.
Once a dossier exists, moreover, those outside government can make claims on it – but must do so in terms set by the dossier itself. Once an initiative is launched, external actors present themselves as appropriate counterparts, objects or recipients of it, imagining and organising themselves in such a way as to capture the attention of government. As a new domain is formed, its spokespersons are not delegated by pre-existing groups but give rise to a group by defining themselves or being defined as its spokespersons.
In each of these ways, the dossier or file is performative, meaning that it serves to create the reality it purports only to describe. The dossier and the domain develop in tandem, as the dossier gathers up a set of communications and at the same time becomes the medium through which parties communicate.
The major was concerned in his presentation to establish an authoritative definition of the situation, as were the Deputy Governor Stewart’s weekly ‘situation reports’ to Baghdad, which ‘imposed a structure on our actions – job schemes, the police, the council’.
The negotiated text
The peace talks at Rambouillet were prompted by a document (the draft accord), then marked by written accounts of the discussions which took place there. What was really going on was that the definition of the situation was being negotiated in the drafting of a document, the peace agreement. Participants moved or acted in turn, by making written comments on draft statements, such that the principal interaction of the meeting took place between documents as much as between groups or delegations.
What really mattered was what was written in the formal (documented) declaration – and then what mattered was not only to agree, but to sign, to give written evidence of agreement. The document seems essential to the way a meeting is realised, the way it and its participants acquire agency.
What the negotiators wanted was a peace agreement, and what they needed for that was a document, in relation to which their key protagonists might act.