The decision to embargo my thesis was made following a meeting with John, the head of the international donor agency country office in which I conducted my ethnographic fieldwork. We met at the doors of the central office and walked until we found somewhere we could sit and talk quietly. He was there briefly for the purposes of a conference. The meeting was within a month of my proposed submission date of my PhD thesis and John had been the gatekeeper whose permission permitted me to conduct an ethnographic study of the agency. My thesis had involved a study of civil servants within the international development donor agency and we had agreed prior to embarking on the fieldwork that I would share texts before submission. As I shared chapters, John had expressed a variety of concerns about factual accuracy, tone and my interpretations of the key results-based performance management processes we had agreed would be the focus of the study (which he summarised as ‘sensationalist nonsense’). Increasingly, he became concerned about the prominence of the office in the text as a whole and the evolving focus of my study. He stressed both the limits to the consent he had given and the risk of harm to himself, his staff and the agency as a whole, framed in terms of research ethics. For my part, I believed the tone was a fair reflection and that my focus on the underlying professional relations was crucial to explanation of results-based thinking and did not slip the bounds of my original permission.
Increasingly, it became apparent the problem was the very fact that I had accessed the office. As time passed during the writing process, we had been unable to come to a satisfactory shared position that squared my analysis and his concerns. It is worth stating that nothing I had written suggested wrongdoing or shortfalls in competence but simply the ordinary disjunctures between policy and practice. What was at stake was the public representation of ordinary bureaucratic procedures in terms other than the bureaucracy's own, at a time of intense press scrutiny and politically motivated critique. John stressed that I had a duty to respect the individuals, the officials, who were involved and their work. He represented strongly – and I felt plausibly – that if published the thesis might be fodder for a critical right-wing press and even if it was not so deployed, it might be sufficient for him to lose his job. He mentioned the possibility of legal action against me for any harm to his career. Following this meeting, and after discussions with my supervisors, I decided to ask my university to embargo my thesis.
My response to the decision was one of relief. Paul Rabinow and Anthony Stavrianakis describe a clearing of the air after a lightning flash when they decided to leave the field and the relationships of mockery, disinterest and domination that they had encountered in their fieldsite (Rabinow and Stavrianakis 2013: 40). For over a year, my writing process had been hemmed in by the demands of academic writing, by my responsibility to my material and by the obligations I owed to John and the office, manifested through John's powerfully stated objections. Our decision was to embargo the thesis, while leaving open the subsequent publication of articles or shorter pieces more flexible in style, less unremitting in form and less exposing to John and the office. We remade our agreement on the spot, both I believe intending to imbue it with legal authority. The moment was a break that created space for a thesis and subsequent published articles; but it was not quite a clean break since the obligations were reframed, not removed.
The break between desk and field is considered crucial to ethnographic writing: without it the desk collapses into the field (Mosse 2006: 937). Ethnographic work is situated alongside powerful systems of meaning present in the field, but is not of them (Clifford 1986: 2). Ethnographic writing must therefore be ‘anti-social’ – a rejection of the social writing practices of the field (Mosse 2006). Yet this rejection is increasingly difficult. Ethnographers increasingly publish and work in adjacent worlds to their participants; what they say can be googled; their emails are accessible; their offices can be visited. Objections to an ethnographic text from the field are not new to anthropology (Bosk 2006; see contributions to Brettell 1993; Mosse 2006) but we are increasingly living in an ‘age of objection’ (Simpson 2016). Perhaps the proper response for an ethnographer is simply to block out these objections – after all, being thick-skinned is a good attribute for ethnographers (Faubion 2009: 146). Yet it is not always easy to block out objections where the objectors are standing in your office, as with Edward Simpson, or are emailing you referencing research ethics and your own agreements in terms that demand to be taken seriously. Crucial to my difficulty in achieving a clean break was John's responsibility within his hierarchy for anything I wrote. Although heavily scrutinised from within the system, external ethnographic access to agencies like John's is unusual, the normal rule being ‘no admittance except on business!’ (Rottenburg 2009a: 60). The fact of access ratchets up the pressure on exit. The challenge becomes making a departure from the field that retains ethnographic poise and integrity – what Rabinow and Stavrianakis describe as Ausgangshaltung (2013: 42). In this article, I follow Rabinow and Stavrianakis (2013) in noting that physically exiting the field does not involve ending the engagement, but rather that the very process of exit – the stress, the negotiation – can be turned to productive analytical ends.
Motivated thus, I wish to turn attention to the organisational ethnographers’ predicament as a writer facing the differing demands the age of objection places on them. For ethnographers, the demands of the traditional or standard research ethics model are often particularly ill-fitting and must be contested or translated to work with anthropological norms and practices. When it comes to ethnographic writing practices, standard models of research ethics fail to acknowledge that the text itself (rather than the research) is often the cause of harm. This matters since, as anthropologists have acknowledged since Writing Culture, the text holds the key to methodological integrity. Ethnographic writing has conventions of style and genre that are at odds with bureaucratic styles, which – with one eye on objection – prioritise compliance, positivist and often quantified epistemologies, impersonal language and a denial of affect (‘emotive language’). Yet as universities themselves increasingly exert managerial and audit controls over research – through ethics and more recently data management regimes – ethnographies are increasingly subject to controls resembling those present in elite fieldsites like my own. Research participants object by deploying concepts from ethical regimes to discipline ethnographic writing practices, often by participants expert in similar kinds of regimes.
In my case, the demands of the ethnography and the ethical responsibilities owed to John presented a dilemma that played out in the arena of the text. I show how John's concerns and the worries about harm and consent were translated into bureaucratic ideas of risk and approval and that deprecated the interpretive, personal, debunking and critical tone of ethnographic writing. As ethnography is increasingly deployed to analyse the modern world, this dilemma is likely to become more pronounced. In this article, I analyse how the traditional research ethics model demands certain kinds of bureaucratic writing and suggest that just as the dilemma lies in the text, so do strategies for responding to the dilemma.
Ethnographic Obligations
What obligations bind ethnographic writing, and why do they matter? A good place to start are the research ethics that are based on the ‘standard’ biomedical model of research ethics. The applicability of this model to ethnographic study has been vigorously problematised: it fails to reflect increasingly common power dynamics where the researcher is researching powerful subjects; its insistence on ex ante ethical reviews cannot handle the ‘undesigned relationality’ (Bell 2019) or emerging objects that good fieldwork inevitably entails (Bell 2014: 515; Lederman 2006: 485; Van Den Hoonaard and Hamilton 2016; Wax 1977: 323); its idea of consent is likewise undermined by the same shifting and emergent ethnographic focus; and its notion of consent is further undermined by the difficulty in communicating ex ante to participants occupying a ‘very different conceptual world’ the violence intrinsic in objectifying them in academic writing (Wax 1995).
The ethical codes and review boards have been portrayed as audit regimes governing research practices (Pels 2000; Strathern 2000) – regimes that are increasingly reinforced through the institution of reputational risk management in universities as a secondary organising logic (Power et al 2009: 305–307). More recently, the introduction of data management regimes (notably the General Data Protection Regulation, GDPR) has instituted the reification and commodification of data as a further regime for the control of research (Pels et al 2018). These regimes are a poor fit for anthropology's relational, interpretive and collaborative episteme and have therefore been stoutly contested by anthropologists (see, for example, Berreman 2003; Bosk and De Vries 2004; Dilger et al 2019; Pels 1999, 2000; Simpson 2011; Sleeboom-Faulkner et al 2017). As Peter Pels writes, anthropologists have found themselves compelled to adopt a ‘trickster’ position situated between different epistemes and moralities (Pels 2000). The efforts to contest and resist inappropriate regimes of governance is a project that must be achieved and must be renewed continually.
What implications has this trickster position had on ethnographic writing practices? After all, most ethnographies seem to be able to make it through with careful attention to the participants, courtesy and care. To the deficiencies listed in the above paragraph, a further can be identified: it rests on an imaginary of harm and risk that is predicated on personal injury during the research process rather than the risks incurred through publication. Yet as David Mosse's account of his consultancy colleagues or Charles Bosk's clinicians or Tijo Salverda's businessfolk, it is during the writing process that the participants’ lives and actions are objectified, analysed according to an unfamiliar scheme and then made public (Bosk 2006; Mosse 2006; Salverda 2019). Reflecting on his encounter with a participant who broke into tears after reading the analysis, Bosk suggests that ‘informed consent’ is always incomplete: in ethnography there is always a betrayal, always deception. Ethnography takes the odd moments and offhand comments seriously and places them in a formal setting; its spirit is essentially debunking (2006: 213). Simpson puts it differently: ethnography is ‘a rather violent and abrupt process, which takes place slowly in a gradual and usually friendly fashion: pious fraudulence, professionalized duplicity, or ritualistic sanctimony’ (Simpson 2016: 115). None of that means there has been a breach of research ethics, however. As an ethnographer, it is possible to act better or worse (Wax 1995: 330), but unless there is a plausible indication of real professional and reputational harm then the ethnographer can act within the bounds provided by a liberal interpretation of informed consent. Indeed, perhaps some ruffled feathers are part of the price of good fieldwork and should be taken to be an indication that the interpretation had some substance (Bosk 2006: 214). This is the essence of the thick skin required.
Yet that does not quite bottom out the methodological implications of ethical rules, which manifest themselves textually in the details of anonymisation and confidentiality. Writing about Dublin intellectuals as a ‘well-established and self-protective bourgeoisie’, Mary Sheehan distinguished what was proper to write methodologically with the ‘dilemma of deciding what was safe to write’ (Sheehan 1993: 79). She mutes certain subjects – individuals’ sexual relationships or alcohol dependency – in her ethnographic writing. Bosk describes the process of reaching an accord with his participants: ‘It was a simple thing: A pseudonym here and there was changed to confuse gender and ethnicity. Some of the identifying details of the Nightingale Hospital were misstated. . . . I was asked to change a few rare diseases so patient confidentiality would not be compromised . . . We haggled a bit over some interpretive terms such as “mop-up service”.’ (2006: 211).
This kind of anonymisation or smudging of a text is not entirely innocent: if ethnography's crisis is still a crisis of representation then decisions over what to write matter. Sometimes the implication of a textual decision may be substantive: for example, if an ethnographer makes the argument that a donor office lives in a confected world called ‘Aidland’, disconnected from the specifics of the context and depoliticising as a matter of course, then the point should be made explicitly with reference to ethnographic material, not implied through a drafting decision to anonymise the country (a challenge I have encountered). Anonymisation of the context removes this possibility. It curtails the ability of the ethnographer to make a credible argument that the specificities of the fieldsite stand for a broader phenomenon. In insisting on anonymity, the picture is blurred and the story told loses its immediacy – problematising the extension of insights to other sites.
As anthropology focuses increasingly on elites, similar predicaments become more likely (Appel 2019; Gilbert and Sklair 2018; Ortner 2016; Souleles 2018). The question Simpson raises – ‘is anthropology legal?’ – becomes more pointed. Texts increasingly move in the same circles as the participants do. At the same time, anthropology is emerging from a period of methodological introversion to address the emergent objects and assemblages through which power is operated. Inevitably, that brings them to confront powerful organisations and networks, often through analyses that debunk and interpret the claims and accounts that appear on the surface – inevitably with risks for those individuals directly depicted. Moreover, while anthropology has traditionally eschewed prior ideological commitments or critical political-moral positions, increasingly there are calls for ethnographers to respond to the field in ways that permit a reflexive identification of the moral issues encountered in the field (Gilbert and Sklair 2018: 10). Growing ethnographic attention on assemblages of power and influence therefore confronts the proliferation of objection.
A Deeply Unhappy Relationship
Let me now turn to the detail of my interaction with my gatekeeper, John. The relationship between John and myself had become – in his words in an email to me – a deeply unhappy one. By the time of the meeting described in the opening paragraphs, it was a month before my scheduled submission date; for his part, he stated to me that as with Mosse's ethnography (2006) the ethnographic discussion of bureaucratic practices would have serious implications for his career and possibly his job, particularly in the climate of intense media and political scrutiny. The aftermath of the global financial crisis saw increased political scrutiny of the development sector and a proliferation of managerial initiatives with the aim of making development more effective. Results-based management audit regimes had been instituted as a key corporate performance tool across the sector. In this, John's agency was no exception, although aspects of the regime were contested within the agency itself on technical grounds. Nevertheless, the country office headed by John – which was responsible for managing a wide portfolio of grants across a number of development sectors – was subject to increased audit pressure and scrutiny by its own headquarters for the performance of its grants and contracts.
John was reasserting the cautious and above all compliant writing practices that pertained in the fieldsite – that is, in the donor agency – in an effort to protect it from the harm that he believed I posed. The growing concern with my presence and the initial permission to access was compounded by the risks attached to the text circling within the same environment as the participants. Towards the close of the meeting described in the opening paragraph, John and I therefore agreed on a tentative way forward that would govern future public texts that used any of the material collected in the office. On John's proposal, we wrote down on a notebook the following clauses that would govern any future publications, including the thesis:
Brendan will provide John drafts of any publications that he proposes to put in the public domain from data collected while in [the office]/with [officials].
John will provide detailed comments within a reasonable time highlighting areas that are considered inaccurate or potentially of reputational risk to John, [the office] or [the agency].
Brendan will review the comments – and accept or reject the proposed amendments.
Brendan will provide a revised complete draft showing where amendments have been made or rejected.
John will decide within a reasonable time, on the basis of the revised draft, whether to invoke the arbitration clause of the agreement.
If arbitration is invoked, we agree on an independent academic who knows [the agency] who agrees to arbitrate on any contested element of the thesis. Both parties have the right to argue their case with the arbitrator. The arbitrator will take into account accuracy, reputational risk to the subjects or [the agency], the integrity of the argument and academic control of material.
Both parties agree to abide by the decision of the arbiter. Arbitration voids any further cause between the Brendan and [John] relating to the material.
I know of no other case where this kind of rather odd agreement has been made. I by no means propose it as a way forward for future ethnographers: our arrival at this document was made in extremis. Had either of us been aware of the endpoint of the journey, neither John nor I would have started on it.
Donor agencies have typically remained closed to ethnographic analysis. The few exceptions are polemic (e.g. Goldman 2005) or rather guarded in the ethnographic materials they adduce (e.g. Mosse 2011). Many of the detailed ‘aidnographies’ shift from close ethnographic observation to deconstructing documents (Ferguson 1996 [1990]: 30–100; Murray Li 2007: 125–136), abstract descriptions (Mosse 2005: 21–26) or generalities (Rottenburg 2009b: 66–73, 80–88) when analysing donor agencies’ internal processes. As a practitioner, I had experienced how donor agencies can be highly influential in shaping development practice while remaining opaque to analysis. The donor agency acts as the hinge between the political and bureaucratic processes of their central office and those of the country in which they are working. As a site for ethnography, they offer a strategically crucial ‘foreshortened’ (Marcus 1998: 95) view of the networks and assemblages stretching both ‘home’ and outwards. I emailed several other heads of country offices from the same agency, but it is unsurprising that only one responded positively to my request. Although I continued to explore the possibility for adding a second site that would strengthen my analysis, I made no progress.
John's initial permission stated that I could spend time in the office provided specific concerns around reputational risk for the organisation and time requirements from the staff were met (acting similar to the deployment of reputational risk in universities as a secondary, reinforcing managerial logic; Power et al 2009). I emailed in response that I was prepared to offer bounds to the consent: first, to limit the data collected to specific office decision-making processes that would be identified with the office (we identified six immediately on my arrival); second, to agree restrictions in the areas or themes about which I can write: ‘[Broadly] I'd anticipate these focussing on: the sense-making practices around understanding the implications of the results agenda; and practices around developing new agendas like the “complexity agenda”’. Third, we agreed the right to see and comment on chapters. Provided I worked within these parameters, I would have use of and control over the material. Just before departure and in concluding our agreement, John emailed me saying that we ‘may have to agree on a form of arbitration’, which I agreed to without discussion – probably in error, but being deep in the throes of preparing for the fieldwork, I felt that any problems could be finessed textually through anonymisation. During my two months’ fieldwork, we did not further discuss the content of the thesis.
On departing the office, I had recommitted to sharing chapters in accordance with the original permission. After a significant period of reflection on the notes and further data collection (primarily historical), I shared my first chapters with John. The first introduced the office and the second analysed one of the six processes I had reviewed, the budget process. In his emailed response, John raised a number of objections that addressed both specific and structural elements in the text of the chapters. In the specific objections, John disagreed strongly with some of my interpretations; he took issue with the tone, suggesting that I was trying to make the ordinary bureaucratic discussions over performance and budget and project design sensationalist; he pointed out errors of fact; and he stressed terminological nuances that might have implications, if caught by – for example – a newspaper (for example, converting the term ‘budget cut’ into ‘budget reprofile’). In the structural objections, John raised concerns that the analysis revolved around his office rather than a deeper analysis of the ‘results agenda’ underpinning the results-based performance management processes. Both consent and the risk of harm were evoked. On the matter of consent, the two chapters had revealed the broader shape of the thesis and the centrality of the role of the office – alongside a significant historical chapter – in that ethnographic account. This was ‘significantly different’ from his original expectations, but not I felt the agreement which required for a proper analysis of the results-based management processes an analysis also of the underlying professional categories and relations. Further, he noted I was no longer focusing on the results agenda but on the operations of the office much more broadly. The risk of harm was significant. He was concerned that the details could be ‘exceedingly damaging to individuals’ and ‘presents a significant reputational risk to [the agency]’.
Having been closeted with my notes and analysis for months, I was (in retrospect, naively) dismayed by this reaction. Dwelling on John's highly critical responses, I found that I could manage many of the specific concerns through finessing the text – through anonymisation, tonal changes and greater care on wording. While these did not substantively change the analysis or my interpretation, they did weaken the immediacy of the analysis. They therefore curtailed the ability to make arguments that might credibly speak to broader processes, spiking the power of the ethnography to provoke analytical inferences with broader relevance. Some areas, however, I found I could not write about in a way that retained the integrity of the analysis. In particular, the presentation of my initial analysis of the results agenda in the budget chapter had incurred serious objections. It had become apparent to me that discussing the operation of the results agenda, the gaps between policy and practice and the manipulation of systems of control was impossible. These are necessarily hidden in any bureaucracy, to accommodate the contradictions in the official and unofficial scripts (Rottenburg 2009b: 67–68). Further, the objections regarding focus were hard to address. The office simply was central to my analysis.
The writing process was caught in a double-bind: I believed I could not address the results agenda without interpreting it in such a way as to incur harm; yet it was the results agenda on which my permission was predicated. I was faced with either risking my participants, compromising my analysis or pushing the limits of my consent. I sought to push the bounds of my consent by shifting focus from the results agenda to the deeper categories of technical expertise and their negotiation that interact with the audit processes and technologies of control, which acted at a more abstract level and could therefore be written about with less immediately worrisome implications for John. The office, however, remained the case study at the heart of my analysis and I continued to swap chapters. As I continued to send almost-finished chapters, the objections on focus continued to be voiced, now adding that I had moved from the focus on the results agenda and the scope of original consent. Facing the impasse, I acceded to a meeting and to the terms outlined above, and asked my university to embargo the thesis. I felt that the space for a text had been squeezed until there was no space left without compromising on the analysis or incurring potential harm.
Crucially, I believe that the risks John identified to himself and to the agency were plausible. The agency had come under increasingly strong scrutiny from a critical press and, if discovered, a thesis might be taken up and used as material for a hatchet job, with implications for the civil servants who had permitted it to happen. I felt responsibilities to John and to the staff. I was also worried by the threat of legal action, given my vulnerable position as a late-stage PhD candidate. I believed that although the form of a thesis afforded very limited options, I would be able to publish in shorter formats after the PhD. I was supported in these deliberations by my school and my supervisors. In retrospect, it is my belief that no ethnography of the results agenda with any critical content could be written that did not trigger the risks that John was so worried about. My very presence in the office triggered the risks and consent should never have been given in circumstances where harm was next to inevitable.
Writing Ethnographically, Ethically and Bureaucratically
The foregoing discussion noted that anthropologists have typically sought to contest the ‘standard’ models of research expressed through the codes of ethics, review processes and more recently data commodification. They have sought to translate the ideas into an ethics that will fit the undesigned relationality, co-production of data and interpretive epistemology that characterises ethnographic work – often necessitating a position that sits between epistemes. In this section, I show how my participants proposed framings of these ethical concepts that fit neatly with their own bureaucratic and managerial practices (characterised by audit and reputational risk governmentalities that echo those of the neoliberal university) and that sought to discipline my writing. I identify three reframings of ethical concepts in particular.
First, the ethical concept of ‘avoidance of harm’ was reframed by the idea of ‘reputational risk’ and was linked directly to the risks involved in publication of the text: a reputational risk that on one level applied to the agency as a whole and on another level (and in a closely related way) would fall on John and on the participants. Reputational risk and its implications in decisions were live matters of discussion within the fieldsite (as well as the university, as noted above; cf. Power et al 2009). Worries about negative press coverage had been a constant theme throughout my stay in the office. Over lunch, officials would joke: ‘when will you contact the newspapers about this?’ The political commitment to the agency within the country was understood to be fragile. The agency therefore took all due steps to ensure that across its reports and data it was meticulously compliant the policies and regulations governing their operation. Any analysis that probed the gaps between policy and practice, or the manipulation of the audit technologies, or indeed that suggested the processes of negotiation were in any way antagonistic between officials, were stoutly resisted on grounds of accuracy – they were ‘sensationalist nonsense’ – but they were also resisted on grounds of the risk that they incurred: they would generate significant personal and professional risks for people in the office.
Second, the original consent to the research was taken to indicate a scope of approval, contracted and agreed. That is, John framed the ethical idea of consent as approval for a defined scope for the thematic focus to the text. Drawing the boundaries to work through terms of reference was a commonplace activity within the office, whether employed in the procurement of consultancy services or the definition of a piece of analysis addressing a particular problem within the hierarchy. In my case, the scope of work was circumscribed by the commitments I had made up front – notably the study's focus on the results-based performance management processes. In that way, the concept from research ethics was translated into an idea with meaning in the field: the contractual terms of reference for a piece of work to which sign-off or approval had been given. As I note above, my thesis was caught in a double-bind between the original consent and the risk: it was impossible to write about the practice of these reporting and results technologies without risking harm; yet to shift from them was to depart from the areas for which I had consent.
Third, and closely linked, my agreement to give John prior sight of chapters was part of the initial consent. It was interpreted as setting out procedures of oversight and objection by which the foregoing boundaries and limits to the text could be policed (while permitting me to retain control within these bounds) – and therefore allowed John to make assessments of my respect for the initial consent and predictions of probable harm. The procedures of engagement on texts reproduced the dynamics of the hierarchy pertaining in the field-site and the bureaucracy. Readers may have observed that the only participant who appears in this article is John. While I invited consent from all key participants in my focal processes and ran chapters past other involved participants, only one of them (a key ally of John) had significant concerns with the texts. That is because the responsibility for my presence in the office was John's and the risk was his alone.
In short, the standard model of research ethics affords participants the opportunity to frame its key concepts to try to discipline ethnographic writing practices in a matter in keeping with their own writing practices. In this case, my research and my writing was to be carefully bounded in its scope, policed by concepts of reputational risk and subject to processes of scrutiny and audit. The translation is facilitated by the governing logics of the fieldsite following the ‘standard model’ research ethics. The ethnographer is urged to adopt bureaucratic writing practices – as noted above, meaning compliant, positivist and impersonal. Such a way of writing is strikingly contrary to ethnographic styles (open, interpretive, personal) and would present considerable methodological implications. Given the centrality of texts to the researcher, they go to the heart of the role of the ethnographic researcher. They demand that ethnographers navigate codes that are painfully at odds: to meet the demands of the academy or the participants; to avoid risk and danger to the participants, while decoding and recoding the field; to stay within the scope of consent while acknowledging the objects that were crucial to the field; to analyse with a mind freed from the field, while wishing to avoid crippling objections that would mire the text. They pushed towards careful selection of issues, a focus on more anodyne objects (Marzano 2016), and a process of blurring and anonymisation. Small surprise, then, that anthropology has consistently contested the standard model of ethics.
The Dilemma of Ethnographic Writing: The Limits to the Tricksters’ Dilemma?
Just write it and fix it up later. (Sheehan 1993: 79)
Readers may disagree with the outcomes from this negotiation; they may indeed be outraged at the limits allowed to be placed on academic freedom. They may ask whether too much control on the text was given up to gain access: was academic freedom given up too readily – did I allow myself to be gagged? How credible, really, was the threat to John's job – and how far was that threat a function of the decision to give me access in the first place? Other cases in this volume suggest embargoes where people's lives and limbs are on the line and the injunction against harm in the ‘standard model’ implies something similar. Was there no other avenue possible? Such decisions seem to shift attention from the academic worth or ethical compliance of the researcher and towards more personal elements: their thickness of skin (Faubion 2009; Marzano 2016), perhaps even their cowardice or laziness, as suggested by Rabinow and Stavrianakis (2013: 78). Visions of the heroic or noble anthropologist are difficult to square with those of Pels’ trickster. They provide little guidance on how to cut through the multiple and dense entanglements, constituted as much by ethical obligations as the political commitments.
My judgement was and remains that there was a significant risk to John's livelihood (sending a draft of this article to him, he noted that it would not be an obvious choice if he was offered the option of losing his job or an arm). Aside from threats of legal action, I had a formal ethical and a moral obligation to him and the donor agency closed the space for discussion. Without offering the limits to scope and process, I would have had no access whatsoever, yet they resulted in a double-bind. I believed that the agreement enabled me to make sure that publications would fairly balance the risk to John and my control of the argument. Although I embargoed the thesis, which as an extended document with all its requirements would have made proper anonymisation impossible, I have been able to publish shorter arguments that are more able to anonymise and narrow down the focus on objects. It placed a limit on my academic freedom to publish, but one that was subject to an arbitration that remained in academic hands. It was a response to the considerable pressures of the moment and the trail of decisions that had led us to that point.
Anne Alvesalo-Kuusi and David Whyte (2018) suggest one way to resolve the tensions: an approach to research ethics that proposes public interests should drive the interpretation of ethics (see also Marzano 2016). They ‘question the proposition that elites and the organisations they work for should be given the same protections as vulnerable research participants’ (2018: 147) and instead propose that public interest should drive the ethical rules. Yet the case for a public interest in anthropological analysis cannot easily rest on established discourses of value, precisely because ethnography questions the established discourses. Their value is the value of the awkward question, the insightful alternative perspective. Yet such a value cannot be won at the expense of harm to individuals: a danger that Alvesalo-Kuusi and Whyte acknowledge.
We may expect the dilemmas to continue as ethnography is increasingly deployed to stir bees’ nests and to address the elite assemblages and objects through which power is exercised and the world constructed. The dilemmas confronting the ethnographies identified here – those of Rabinow and Stavrianakis, Mosse, Simpson, Bosk, Sheehan, Salverda – are the real and inevitable prices for that. Yet access begets difficulties in exit that can be analytically fruitful: the contestation over writing practices reveals the audit regimes and the discourses within the site in a visceral and powerfully stated manner. It delineates the risks and fears that prevail. Conversely, as participants in the field translate concepts from research ethics and deploy them to bind writing practices, it also reveals the limits to the trickster position adopted by anthropologists. The necessity of anonymising and blurring the details of the fields risks reducing the credibility precisely as an exemplar of the broader processes and practices by which global capitalism is made up and conducted.
If the problem is textual in nature – in that the risks and tensions arise in connection to the text's circulation in the same circles as the participants – then perhaps the way to resolve the dilemma is textual too. It is not possible to write in both an academically satisfactory and a bureaucratic – that is, a compliant – fashion. The techniques for navigating texts through this domain are familiar enough: anonymisation, blurring of context, abstraction, changing of details. Freed from the obligation to make public the long-form thesis, I have been able to draft articles and to explore other avenues for publication. Being shorter, it is possible to tailor the level of specificity and anonymity. Yet the experience from my thesis is that this may not be sufficient: John observed that had I been able to secure access to a second office, the weight of the analysis would not have fallen on access to that office. Multi-sited ethnographies permit blurring more readily than concentrating on single sites. Methodological decisions may need to be taken anticipating the need to create the space for ambiguity in the text.
Acknowledgements
Dr Brendan Whitty would like to thank Dr Margaret Sleeboom-Faulkner and Dr James McMurray for organising the panel for which this article was originally prepared and to the two reviewers. He would like to thank Dr Ben Jones and Dr Arjan Verschoor as supervisors of the thesis that forms the basis for this article; the University of East Anglia School of International Development for funding my PhD research; my colleagues at the School of International Development for supporting me so well throughout; and particularly Dr John McDonagh and Prof Laura Camfield, two heads of school who helped me handle the fall-out of this. Acknowledgement also to John and to the other generous participants involved in this work – whilst the outcome was not what any of us wished, they were generous hosts and I was enriched by my time in the office.
References
Alvesalo-Kuusi, A. and D. Whyte 2018. ‘Researching the powerful: a call for the reconstruction of research ethics’, Sociological Research Online 23: 136–152.
Appel, H. 2019. To critique or not to critique? That is (perhaps not) the question . . . ’, Journal of Business Anthropology 8: 29–34.
Bell, K. 2014. ‘Resisting commensurability: against informed consent as an anthropological virtue’, American Anthropologist 116: 511–522.
Bell, K. 2019. ‘The “problem” of undesigned relationality: ethnographic fieldwork, dual roles and research ethics’, Ethnography 20: 8–26.
Berreman, G. D. 2003. Ethics versus ‘realism’ in anthropology: redux, in C. Fluehr-Lobban, (ed.), Ethics and the profession of anthropology: dialogue for ethically conscious practice, 51–83. Walnut Creek, CA: AltaMira.
Bosk, C. L. 2006. Irony, ethnography, and informed consent, in B. Hoffmaster (ed.), Bioethics in social context, 199–220. Philadelphia, PA: Temple University Press.
Bosk, C. L. and R. G. De Vries 2004. ‘Bureaucracies of mass deception: institutional review boards and the ethics of ethnographic research’, Annals of the American Academy of Politics and Social Science 595: 249–263.
Brettell, C. B. (ed.) 1993. When they read what we write. Westport, CT: Bergin & Garvey.
Clifford, J. 1986. Introduction: partial truths, in J. Clifford and G. E. Marcus (eds.), Writing culture: the poetics and politics of ethnography: a School of American Research advanced seminar, 1–26. Oakland, CA: University of California Press.
Dilger, H., P. Pels and M. Sleeboom-Faulkner 2019. ‘Guidelines for data management and scientific integrity in ethnography’, Ethnography 20: 3–7.
Faubion, J. D. 2009. The ethics of fieldwork as an ethics of connectivity, or the good anthropologist (isn't what she used to be), in J. D. Faubion and G. E. Marcus (eds.), Fieldwork is not what it used to be, 145–164. Ithaca, NY: Cornell University Press.
Ferguson, J. 1996 [1990]. The anti-politics machine: ‘development,’ depoliticization and bureaucratic power in Lesotho. Minneapolis, MN: University of Minnesota Press.
Gilbert, P. R. and J. Sklair 2018. ‘Introduction: Ethnographic engagements with global elites’, Focaal 81: 1–15.
Goldman, M. 2005. Imperial nature: the World Bank and struggles for social justice in the age of globalisation. Newhaven, CT: Yale University Press.
Lederman, R. 2006. ‘Introduction: anxious borders between work and life in a time of bureaucratic ethics regulation’, American Ethnologist 33: 477–481.
Marcus, G. E. 1998. Ethnography through thick and thin. Princeton: Princeton University Press.
Marzano, M. 2016. Uncomfortable truths, ethics, and qualitative research: escaping the dominance of informed consent, in W. Van Den Hoonaard and A. Hamilton (eds.), The ethics rupture: exploring alternatives to formal research-ethics review, 116–118. Toronto: University of Toronto Press.
Mosse, D. 2005. Cultivating development: an ethnography of aid policy and practice. London: Pluto Press.
Mosse, D. 2006. ‘Anti-social anthropology: objectivity, objection and the ethnography of public policy and professional communities’, Journal of the Royal Anthropological Institute 12: 935–956.
Mosse, D. 2011. Social analysis as corporate product: non-economists/anthropologists at work at the World Bank in Washington, DC, in D. Mosse (ed.), Adventures in aidland: the anthropology of professional international development, 81–102. New York: Berghahn Books.
Murray Li, T. 2007. The will to improve: governmentality, development and the practice of politics. Durham, NC: Duke University Press.
Ortner, S. B. 2016. ‘Dark anthropology and its others’, HAU: Journal of Ethnographic Theory 6: 47–73.
Pels, P. 1999. ‘Professions of duplexity: a prehistory of ethical codes in anthropology’, Current Anthropology 40: 101–136.
Pels, P. 2000. The tricksters’ dilemma: ethics and the technologies of the anthropological self, in M. Strathern (ed.), Audit cultures: anthropological studies in accountability, ethics and the academy, 135–172. London: Routledge.
Pels, P., I. Boog, J. Henrike Florusbosch, Z. Kripe, T. Minter, M. Postma, M. Sleeboom-Faulkner, B. Simpson, H. Dilger, M. Schönhuth, A. Von Poser, R. C. A. Castillo, R. Lederman and H. Richards-Rissetto 2018. ‘Data management in anthropology: the next phase in ethics governance?’, Social Anthropology 26: 391–413.
Power, M., T. Scheytt, K. Soin and K. Sahlin 2009. ‘Reputational risk as a logic of organizing in late modernity’, Organization Studies 30: 301–324.
Rabinow, P. and A. Stavrianakis 2013. Demands of the day: on the logic of anthropological inquiry. Chicago, IL: University of Chicago Press.
Rottenburg, R. 2009a. Far-fetched facts: a parable of development aid. Cambridge, MA: The MIT Press.
Rottenburg, R. 2009b. ‘Social and public experiments and new figurations of science and politics in postcolonial Africa’, Postcolonial Studies 12: 423–440.
Salverda, T. 2019. ‘Conflicting interpretations: on analyzing an agribusiness’ concerns about critique’, Journal of Business Anthropology 8: 4–24.
Sheehan, E. A. 1993. The student of culture and the ethnography of Irish intellectuals, in C. B. Brettell (ed.), When they read what we write: the politics of ethnography, 75–89. Westport, CT: Bergin & Garvey.
Simpson, B. 2011. ‘Ethical moments: future directions for ethical review and ethnography’, Journal of the Royal Anthropological Institute 17: 377–393.
Simpson, E. 2016. ‘Is anthropology legal? Earthquakes, blitzkrieg, and ethical futures’, Focaal 74: 113–128.
Sleeboom-Faulkner, M., B. Simpson, E. Burgos-Martinez and J. McMurray 2017. ‘The formalization of social-science research ethics’, HAU: Journal of Ethnographic Theory 7: 71–79.
Souleles, D. 2018. ‘How to study people who do not want to be studied: practical reflections on studying up’, PoLAR: Political and Legal Anthropology Review 41: 51–68.
Strathern, M. 2000. New accountabilities: anthropological studies in audit, ethics and the academy, in M. Strathern (ed.), Audit cultures: anthropological studies in accountability, ethics and the academy, 13–30. London: Routledge.
Van Den Hoonaard, W. C. and A. Hamilton 2016. Introduction. The ethics rupture summit in the context of current trends in research ethics review, in W. C. Van Den Hoonaard and A. Hamilton (eds.), The ethics rupture: exploring alternatives to formal research-ethics review, 5–22. Toronto: University of Toronto Press.
Wax, M. L. 1977. ‘On fieldworkers and those exposed to fieldwork: federal regulations and moral issues’, Human Organization 36: 321–328.
Wax, M. L. 1995. ‘Informed consent in applied research: a comment’, Human Organization 54: 330–331.