Revisiting the Case of Ethnography and International Relations

Abstract

This article revisits the debate on the role of ethnography in International Relations. It primarily does this by elucidating three points of tension in the literature on ethnography in International Relations. Firstly, it tackles the challenges related to ‘getting on’ with ethnography after the reflexive methodological developments that have taken place within anthropology since the 1980s. Secondly, it investigates how to overcome certain matters of scale and how to conceptualise the ‘international’ methodologically, or more specifically, ethnographically. When looking at issues that somehow exist and operate on the international scale, the ethnographic task of immersion in local scenes does sometimes seem like an ill-suited approach. However, I argue, this problematisation is dependent on a certain methodological understanding of what the international is. I attempt to formulate an alternative methodological approach that takes seriously the idea that international relations always can be accessed locally. This paper suggests that one of the main solutions to the obstacle of scale is methodologically abandon the imaginary of totalities as a higher level. In this way, ethnography can enable important understandings of social relations that exist across scales of local and global.

Repenser le rapport entre l’ethnographie et les relations internationales

Introduction

Ethnography is a method and methodology that is commonly associated with anthropology and is, in many ways, an intrinsic part of that discipline’s ‘way of working’. ¹ In International Relations (IR), a discipline whose methodology has traditionally been state-centric and has embraced problem-solving theories and rationalist perspectives, ethnography has a short and limited history. However, there have been attempts to work ethnographically in the study of IR, something that has since been identified as an ‘ethnographic turn’. Furthermore, in the past decade, there has been an increasing interest in methodological discussions on the relative values and challenges of using the ethnographic method and/or methodology to study IR. This paper aims to pick up and expand upon this discussion by engaging with some methodological perspectives and approaches that have thus far been lacking. With this in mind, my point of entry is to elucidate and attempt to tackle three points of tension in the literature on ethnography in IR that impose challenges to developments of the methodology.

Firstly, IR scholars attempting to use the ethnographic method as an instrument in developing empirically grounded understandings of international practices have been criticised for not considering the reflexive methodological developments that have taken place within anthropology since the 1980s. This criticism has primarily been posited by Wanda Vrasti who is critical of the way in which particularly feminist and constructivist scholars have adopted ‘a selective, instrumental and somewhat timid understanding of what ethnography is and does’. ² She partly blames this on poor interdisciplinary reading practices, as there has been an apparent lag between the literature from within anthropology and the ethnographic developments in IR. Certainly, it is clear that much of the ethnographic research in IR has been carried out without reference to the post-1980s critical reflections about what ethnography is and its effects. As I will discuss below, ethnography is not merely a tool for data gathering but rather a methodology that involves important ontological and epistemological commitments. However, I argue that there is value in moving on from such debates and gate-keeping practices and instead focus on how to develop an effective methodological framework for studying IR ethnographically.

Secondly, there is the matter of scale. Claims have been made that ethnography is not an appropriate way to conduct research in IR since the method’s intrinsic ‘sitedness’ and focus on local scenes entails an inability to capture world politics that operate on a ‘higher’ global level. ³ Ethnography, as a methodology that aims to study the particular, is seen to only provide understandings that are partial and of a different scale than IR. However, this problem of scale relies on methodological assumptions that I wish to untangle. The question is what the relation is between sitedness and totality, and how to deal with this relation methodologically. These methodological considerations need to take place in order to fully develop ethnography as a means of gaining insights in IR. In order to do so, as Vrasti suggests, it is necessary to start with in-depth engagement with the recent literature from anthropology. By doing this, I find that only by abandoning attempts at capturing and accounting for a totality in world politics, the project of doing ethnographic research in IR can progress.

Third, and relatedly, for IR researchers doing ethnography on the level of international politics and states it is seen as challenging to find an appropriate field-site to conduct research. ⁴ This problem is reliant on the idea that field-sites already exist as bounded locations or empirical units a priori. However, by engaging with some of the recent scholarship on ethnography in anthropology, I find that it has become apparent over the recent decades that ethnographers should ‘take up specific topics and issues (rather than simply people and places) as creating contexts for ethnographic study’. ⁵ It is also perceived as challenging to gain access to the highly regulated world of international politics. While it is true that various practices in IR happen behind closed doors, I argue that because collected knowledge is always partial and incomplete this should not be regarded as essentially problematic. Further, as many argue, there are also insights to be gained from these challenges. ⁶

The points made in this article are uncontroversial in fields more experienced with ethnography, such as Social Anthropology and Sociology. However, it is puzzling that in IR, especially in the mainstream of both European and American IR, the ethnographic methodology has received limited attention and remains underdeveloped. Although there has been some interest within the discipline in using ethnographic methods to gain different perspectives on IR and discuss methodological issues, this interest has neither been sustained nor highly visible. The points above aim to clarify the obstacles hindering the development of ethnography in IR, reintroduce certain debates and suggest potential paths forward.

The main contribution of this paper is therefore to open up space to reflect on some methodological issues in IR as it relates to ethnography. This speaks to recent renewed interest in methodological discussions in IR, and in social sciences generally, and calls for further developments. ⁷ Additionally, I will indicate the potential value this can have for critical theorising. Importantly, I argue that the methodological abandoning of both the idea of bounded localities and abstract totalities open up possibilities for how ethnography is regarded in IR and thus what it can achieve and contribute to theory, especially critical theory. Localised studies of global political practices, such as diplomacy, state intervention or humanitarianism, can provide important perspectives on how particular practices, structures and processes are maintained on a daily basis.

I do not wish to embed my argument in any particular theoretical perspective in IR, as I believe that the ethnographic methodology should not be limited to a particular paradigm. Iver Neumann defines IR’s subject matter as

political and social life that plays out in a setting where there is a plurality of polities. More specifically, the subject matter of IR is sovereign and suzerain relations between polities as they existed in the past, as they exist in the present and as they have been and are imagined to exist. ⁸

This is a fairly open definition, and it does not draw up either directions in theory or method.

Furthermore, my aim is in many ways to develop a critical methodology, following from these traditions, to counter the hegemony of positivist methodologies. ⁹ Furthermore, as an added note, I believe that the methodological choices outlined above can be beneficial to both feminist and post-colonial research in particular, as it allows for a focus on how global power is maintained on a local level without assuming that this level exists a priori.

This paper is divided into four sections. The latter three of these will discuss in turn the above outlined points of tension. The first section will provide some background that is relevant for my discussion and arguments. Here, I will begin by specifying how methods and methodologies are conceptualised and differentiated. Next, I will discuss how ethnography is conceptualised as a method, a methodology and a sensibility. This section will also examine and discuss the so-called ‘reflexive turn’ in anthropology and the effect that these challenges have had on what we imagine that ethnography does and what it can achieve. The second section will investigate and review the exchanges in the literature on IR and ethnography, and provide a discussion on some of the most prominent ethnographic works in IR. This section is divided into two parts, the first discussing the exchanges in broad terms while the latter is focused on the literature that seeks to study practice ethnographically. The third section will investigate the tensions between the local level and the global level in contemporary ethnographic scholarship. I will begin this section by investigating the multi-sited methodology and the underlying assumptions that enable it. Next, I proceed to provide a critique of the multi-sited approach, based primarily on its underlying holistic assumptions, and offer an alternative methodological perspective that abandons the imaginary of parts and totalities. The last section will provide a suggestion for how the methodology can be practiced. I will first discuss how to find and bound the field, before I discuss the challenges of accessing it.

Ethnography as Method and Methodology

It is common to make a distinction between ‘method’ and ‘methodology’ in social science. Generally, to implement a method ‘is to follow through a sequence of prespecified and regulated steps toward the realization of a determinate goal’. ¹⁰ It is often treated and thought of as a ‘recipe’ that needs to be followed in order to answer a certain question or test a particular hypothesis. ¹¹ As observed by Peter Katzenstein and Rudra Sil, much of the research in IR

has been founded on a positivist conception of social knowledge. Positivists generally gravitate toward a view of social inquiry in which patterns of human behaviour are presumed to reflect objective principles, laws, or regularities that exist above and beyond the subjective orientations of actors and scholars [. . .]. ¹²

These views and practices of methods in social science has come under increasing attack from critically minded scholars. As an example, Law ¹³ argues that ‘research methods’ do not only describe social realities, but also partake in creating them. Although positivism is always founded in the ideology of the researcher, the debate around different visions of the world were ‘neutralized by discussions of techniques of gathering and processing data, the search for shared methods of establishing universal truths, and an investment in the idea of the unity of social science’. ¹⁴ As a rejection of this ‘disciplining and hygenising’ function of method, critical scholars within IR have all to various degrees moved beyond thinking about methods to thinking about methodology. ¹⁵

While methods are seen as a tool for data gathering, methodology implies ‘a concern with the logical structure and procedure of scientific inquiry’. ¹⁶ Whereas methods are ways of collecting data, methodology is a system for making sense of this data. Questions of how to study the social world are necessarily entangled with considerations regarding what the social world consists of, what sorts of knowledge can be produced and what the effects of such knowledge is on reality. This embrace of methodology has therefore also necessarily involved an increasing focus on questions of epistemology and ontology. ¹⁷ In order to have reflections on methodological issues and questions, it must be entirely ‘separated from the tyranny of method and reframed as an overarching epistemological and meta-theoretical reflection’. ¹⁸ In sum, one can argue that there always needs to be both a method for data gathering and methodological reflection to make sense of the data collected using the method. However, methodological reflection on the metatheoretical implications of using a certain method needs to be the basis of both the choice of method, how to employ a certain method and how the researcher choses to interpret the collected data. Method and methodology are always deeply intertwined and cannot be separated.

Ethnography is a term used to describe a method, a methodology and sometimes a ‘sensibility’. The ethnographic research method typically involves ‘participating, overtly or covertly, in people’s daily lives for an extended period of time, watching what happens, listening to what is said, and/or asking questions through informal and formal interviews, collecting documents and artefacts’. ¹⁹ Strathern observes that ‘an ethnographic account is conventionally the description of a particular society and culture known to be based at some point on the experiences and observations of a fieldworker “who was there”’. ²⁰ Clifford, borrowing from Rosaldo and highlighting the importance of ‘being there’, refers to the ethnographic methodology as ‘deep hanging out’. ²¹ The term ‘participant observation’ refers to this way of doing research in which the researchers ‘join the flow of daily life while also taking notes on it’. ²² Because of the inherent unpredictability of these everyday encounters, the ethnographic method is regarded as a rather unpredictable and particularly un-parsimonious way of doing research. However, this lack of parsimony implies certain methodological commitments and is therefore not unintentional. As Wedeen observes, ‘ethnographers tend to view everything happening as a potential moment for evidence-gathering and/or rethinking the project’s premises’. ²³

These moments of re-evaluation based on observation and experience are essential to the ethnographic methodology and ‘sensibility’. It has been described as ‘an approach that cares – with the possible emotional engagement that implies – to glean the meanings that the people under study attribute to their social and political realities’. ²⁴ This is reflected in how ethnography is conceptualised as a methodology that aims to see things from other people’s point of view: ‘to grasp concepts that, for other people, are experience-near, and to do so well enough to place them in illuminating connection with experience-distant concepts theorists have fashioned to capture the general features of social life’. ²⁵ Ethnographers should therefore not have to adhere to positivist notions about ‘proper’ research, by specifying beforehand the parameters of the research – such as how many people will be talked to and for how long ²⁶ – as the implications of assuming what the ‘field’ looks like before entering it would involve a departure from the principles of ethnography as a methodology and a sensibility.

On the other hand, the ‘laxity’ of ethnographic research does not imply that there are no practical rules that should be followed. Ingold asserts that proper rigorous ethnographic work must adhere to important principles such as ‘long-term and open-ended commitment, generous attentiveness, relational depth, and sensitivity to context’. ²⁷ Reasonably, the methodological commitments of ethnographic research need to be reflected in the methods that are used. As Patrick Thaddeus Jackson observes, the ethnographic methodology ‘calls for some kind of data collection technique that respects the involvement of the researcher in and with the object researched’. ²⁸ Anthropologist Tim Ingold attempts to make sense of the tension between the philosophical and pragmatic dimensions of ethnographic theory by thinking about it as a ‘philosophy with people in it’: ‘an enterprise energised by the tension between speculative inquiry into what life could be like and a knowledge rooted in practical experience, of what life is like for people of particular times and places’. ²⁹

On a practical level, there seems to be wide agreement that in order for a project to be ethnographic it must include some form of fieldwork, preferably including participant observation. However, these aspects of the ethnographic method are usually not the most important defining characteristic of ethnography. Rather, the goal of ethnography is ‘capturing social meaning and ordinary activities’. ³⁰ How, exactly, this is done is subject to great variation among ethnographers and depends on the researcher’s own methodological reflections and circumstances. ³¹

There are however certain ethical and theoretical implications of ethnographically representing other people’s point of view, which became the subject of considerable debate in the late 1980s and early 1990s, both within and outside the discipline of anthropology. Put in simplified epistemological terms, these were confrontations with the positivism present in the belief that ethnographies were perfect representations, and reflections of the ethnographers’ complete understanding, of the ‘native’s point of view’. Geertz had earlier alluded to this problem when he asked: ‘if anthropological understanding does not stem, as we have been taught to believe, from some sort of extraordinary sensibility, an almost preternatural capacity to think, feel, and perceive like a native [. . .], then how is anthropological knowledge of the way natives think, feel, and perceive possible?’ ³² The ethnographer does not stand outside the context they are studying as a neutral observer, and further, they are never able to perfectly represent the lifeworlds of the people they are studying. ³³

Many of the most important early contributions to this debate were featured in the influential edited volume by Clifford and Marcus, Writing Culture. ³⁴ In the introduction, Clifford comments that until recently the process of writing had been ‘reduced to method: keeping good field notes, making accurate maps, “writing up” results’. ³⁵ He goes on to state that the mission of Writing Culture is to ‘crumble’ this ideology and highlight the performative function of writing. Further, the volume draws attention to ‘the historical predicaments of ethnography, the fact that it is always caught up in the invention, not the representation, of culture’. ³⁶ As Strathern observes, ethnographic accounts ‘do not refer to an independent reality which can be grasped by other means but create that sense of reality in the act of narration’. ³⁷ This reflexive moment in anthropology, which has since been labelled the ‘reflexive turn’ or the ‘crisis of representation’, spurred many within anthropology to abandon commitments to positivist assumptions and to reconceptualise what ethnographic research can achieve.

With the above debates in mind, the ethnographic approach must be regarded as essentially entangled in certain formulations of what is real and what knowledge of the real is. This means that as participant observation necessarily blurs the line between observer and observed, and thus between the researcher and the subject, and that this has implications on how ethnographic knowledge should be produced and presented. Further, I believe that reflecting on how ethnography represents its subject and the positionality of the researcher are fundamental aspects of its methodological commitments. After having provided an overview of the most important aspects of how ethnography is conceptualised and challenged, I will now investigate how this has been discussed in IR.

Ethnography in IR

Reflexivity has its own history within IR, where it has largely been inspired by critical and poststructuralist approaches to social science. ³⁸ While there has been some interest in ethnography within this post-positivist and reflexive corner of the IR discipline, ethnographers in IR have faced criticism for not addressing the issue of representation, as is done in anthropology. ³⁹ Further, the question of how to conduct ethnography in IR has not received significant attention and remains underdeveloped, especially within the mainstream of the discipline. Tensions exist between the traditions and practices of the IR discipline and ethnography, which have hindered the development of a framework for conducting ethnographies of the international.

These issues were explicitly discussed in the important and insightful exchange between Wanda Vrasti ⁴⁰ and Jason Rancatore ⁴¹ concerning the ‘strange case’ of the relation between the ethnographic methodology and the discipline of IR. I will now provide an overview of this debate. I also see it as important for the future of ethnography in IR to move beyond this debate. I argue that although there is both political and theoretical value in taking seriously the challenges surrounding representation, positionality and critical methodological reflection (this being, in fact, an important aspect of what I am arguing in this paper), ethnography can also provide a positive contribution to developments in the theoretical conceptualisations of IR’s object of study.

The purpose of bringing the ethnographic method into IR has primarily been to ‘direct greater attention to everyday practices and embodied actions, thereby countering the criticism of IR as a static and state-centric discipline ill-suited for grasping the complexities of political life’. ⁴² It was identified as a ‘turn’ by Vrasti, as the ethnographic method appeared ‘to have gained unprecedented momentum in IR over the past couple of decades’. ⁴³ Primary feminist, constructivist and post-colonial scholars are recognised as having been at the forefront of this ‘turn’. ⁴⁴ However, the reach and impact of the ethnographic methodology in IR have been rather limited. Therefore, referring to it as a ‘turn’ may be an overstatement. Instead, it should be placed in the context of the proliferation of ‘turns’ in IR in the 2000s, where the ‘overall trend is a shift in emphasis from epistemology and methodology towards ontology – from metatheoretical debates about the philosophical presuppositions of IR scholarship towards analysing the “stuff” that the world is made of’. ⁴⁵

The ethnographic method was seen as a way to expand our understanding of IR as it would enable researchers to get close to what the practitioners of global politics were actually doing, and more importantly, how they think about what they are doing. The ethnographic method was therefore seen as an effective alternative to ‘armchair analysis’ in IR theorising. Neumann describes this armchair mode of doing research as ‘text-based analyses of global politics that are not complimented by different kinds of contextual data from the field, data that may illuminate how foreign policy and global politics are experienced as lived practice’. ⁴⁶ In simplified terms, ethnography was seen as enabling scholars to instead provide understandings of how things actually are by basing analysis on observed practice.

However, these hopes reflected a rather under-developed understanding of what ethnography is and what it can achieve. Vrasti, in her criticism of the ‘ethnographic turn’ in IR, rather pessimistically declared that

Had ethnographically-minded IR scholars paid greater attention to the ways in which ethnography has been written and rewritten from anthropological quarters over the past two decades, it would have been clear to them that ethnography cannot accomplish the goals it was set out to realise. ⁴⁷

Without engaging with the problems of representation and positionality, and their epistemological implications, it is problematic to regard the ethnographic method as a way to provide ‘better’ and more critical understandings of world politics.

With this problem in mind, Vrasti goes on to observe three tendencies within ethnographic research in IR: (1) ethnography as reduced to signify a data-gathering technique (ethno-empiricism); (2) a way of writing aimed to provoke critical engagement (ethnografeel text); (3): a theoretical sensibility that aims to pay greater attention to every-day practices (ethnographilia). ⁴⁸ With regards to ethno-empiricism, Vrasti highlights works of feminist scholarship within IR that reduces ethnography ‘to a series of methodological choices’. ⁴⁹ Particularly, Carol Cohn’s ethnography ‘Sex and Death in the Rational World of Defense Intellectuals’ ⁵⁰ is criticised for implying a clear separation between the events she observed and participated in and her ability to afterwards objectively analyse these events. Here, Vrasti’s critique refers to the aforementioned reflexivity regarding the ethnographer’s ability to perfectly capture and represent the events and people’s worldviews.

The opposite tendency is observed in the discussion on the ethnografeel text. Here Vrasti highlights the work of another important feminist scholar, Cynthia Enloe. Enloe is criticised for contributing to ‘the tendency to sacrifice fieldwork experience on the altar of literary stylisation’. ⁵¹ While Enloe’s research did not include long-term fieldwork or participant observation, Vrasti argues that Enloe’s work is ethnographic in the sense that she prioritises research that is ‘experience-near’ in order to describe and understand militarised women’s lives. However, despite Vrasti’s convincing critique of Enloe for representing women’s perspectives without directly engaging with these women, it is perhaps unwarranted to criticise Enloe as an ethnographer, as she did not herself claim that her work was ethnographic. ⁵²

The last of Vrasti’s three categories, ethnographilia, is connected to constructivist scholar’s search for a methodology, particularly those who wish to study political practice in IR:

In the context of social constructivism, ethnography promises to lend methodological rigour to a theoretical approach that has always struggled to find a balance between agential and structural, subjective and objective, mid-level and grand-level theorising. In other words, ethnography can act as the missing methodological link that would make constructivism whole. ⁵³

However, Vrasti further claims that these hopes are in vain because IR constructivists’ selective understanding of ethnography reduces it ‘to its constitutive method, participant observation, and purposively ignores the by now common idea that ethnographic texts always interpret, distort and betray the social reality they seek to capture’. ⁵⁴ As previously stated, the ethnographic methodology is always grounded in certain epistemological and ontological commitments that relate to the ethnographer’s essential entanglement with the subject they are attempting to analyse. The ethnographic method can therefore not be reduced to securing a middle ground ⁵⁵ between, for example, a constructivist position that accepts a social ontology (the ‘constructedness’ of the social world), and a positivist epistemology that works on the assumption that there exists data-gathering techniques that can provide us with ‘seamless’ understandings of the social world. Below I will further elaborate on these points. While Vrasti is touching on crucial issues in the debate over ethnography in IR, I believe we have to move beyond the dichotomy of positivism versus reflexivity as a means to judge the value of ethnographic research.

An IR constructivist aiming to develop critical theory through methodological developments is Vincent Pouliot. In his article ‘Sobjectivism: Toward a Constructivist Methodology’ he attempts to define a constructivist methodology ‘that is specifically geared toward the constructivist style of reasoning’. ⁵⁶ According to Pouliot, this methodology needs to combine objective and subjective knowledge about international life, thus taking the aforementioned ‘middle ground’. While pouliot’s efforts in methodological developments do imply an acceptance of a particular rationalist stance, it does not imply that ethnography provides a way to access an ‘objective’ reality. ⁵⁷ Furthermore, Pouliot is attempting to develop a constructivist methodology, not a specifically ethnographic one. While he refers to the ethnographic methodology when stating that ‘sobjectivism’ should attempt to combine both ‘experience near’ and ‘experience distant’ concepts, and is explicitly influenced by Geertz, ethnography is here clearly conceptualised as one of many tools for data gathering that can be used. This is what Vrasti critiques Pouliot for and the tendency he represents: his suggestion that the ethnographic method can be separated from the ontological and epistemological commitments of the ethnographic methodology. Pouliot, by attempting to develop such an approach, allows for the ethnographer to stand outside the ‘reality’ that is being observed. One cannot both be inside a social world and outside it at the same time, according to Vrasti.

According to Rancatore, however, Vrasti does not take seriously the important distinction between method and methodology. ⁵⁸ While ethnographic methods such as participant observation can be regarded as tools for data-gathering, the ethnographic methodology provides ‘a philosophical support for their use’. ⁵⁹ Vrasti argues that the ethnographic methodology is always interpretative and, as method and methodology are always entangled, cannot be used to justify the reduction of ethnography to a data-gathering method. On the other hand, as Rancatore further argues, there can exist multiple ethnographic methodologies justifying the use of a particular method, and what Pouliot is attempting is ‘to offer a methodology for the purpose of getting on with the empirical work (employing research methods)’. ⁶⁰ Pouliot is developing a methodology that justifies the use of methods, such as the ethnographic one, as data-gathering techniques. Pouliot’s position is a result of an acceptance of the impossibility of a researcher to perfectly understand its subject through ethnographic research and a pragmatic effort to overcome this. Rather than assuming that ethnography can achieve a transcendental understanding of the social world, Pouliot pragmatically chooses to adapt a selective and instrumental version of the ethnographic method.

An example of this is the contribution of Pouliot’s monograph International Pecking Orders. ⁶¹ Although Pouliot does not explicitly brand his book as an ethnography, due to the methods used and the stated methodological goals of the project, it can justifiably be regarded as such. He describes the purpose of his book as ‘understanding the inner workings of multilateral diplomacy – the social theatre of world politics’. ⁶² In order to do so, he claims that there needs to be an ‘empirical redescription, as faithfully as possible, of what is actually happening on the ground from the practitioner’s point of view’ and that this requires ‘deep immersion and interpretation’. ⁶³ The book starts out with the observation that multilateral diplomacy, which is formally constituted by the principle of sovereign equality, ‘actually takes place on a deeply unlevelled playing field’. ⁶⁴ Pouliot is able to trace and investigate the process by which international hierarchy, or ‘pecking orders’, is sustained in and through multilateral organisations. Aside from being an important theoretical contribution to the study of international diplomacy, it is significant as a constructivist analysis that is methodologically grounded in everyday practices. The ethnographic method can be a valuable tool in IR theorising, because it permits scholars such as Pouliot to show and theorise about ‘what is really happening’ as opposed to text-book descriptions regarding how global politics ‘should work’.

Yet, these pragmatic choices do reflect certain problematic tendencies in IR as a discipline, or as Vrasti labels it, as a series of ‘disciplining acts’. ⁶⁵ Vrasti points out that the propensity of IR scholars not to take seriously the radical and reflexive developments within anthropology is revealing a disciplinary culture that only accepts interdisciplinary efforts that do not challenge the ontological imagination of the discipline. ⁶⁶ Furthermore, Jackson similarly observes that scholars taking the ‘reflexivist position’ are ‘segregated into a separate subsection’ and that their critical interventions are marginalised. ⁶⁷ Thus, consciously or unconsciously, scholars are encouraged to uphold certain ontological and epistemological positions in order to avoid such marginalisation. Vrasti thus concludes that ethnography can only be effective in IR if it is turned towards the discipline’s own practices and theories. ⁶⁸ By seeing things from the IR practitioner’s point of view, such an ethnography could uncover both the institutional practices and the types of knowledges that IR produces. Importantly however, beyond the clear necessity of this type of project, there still exists opportunities for the ethnographic methodology to illuminate and expand upon the subject of IR. I believe that one can ‘get on with empirical work’ by accepting an ethnographic methodology that is interpretive and reflexive. Moreover, the uniqueness of the ethnographic method is the situatedness of the researcher in the field, and the methodological consequences of these entanglements.

With regards to the ethnographic study of diplomacy, Iver Neuman’s work has been particularly important with regard to the development of critical theory that is ‘grounded’ in an ethnographic understanding of practice. Neuman makes the methodological choice not to reduce ethnography to a tool for data gathering, and further takes his positionality seriously. At Home with the Diplomats investigates the everyday practice of state diplomacy and shows that it differs from ‘the opinions of social scientists who have no personal experience with diplomats and diplomacy’. ⁶⁹ His analysis is largely based on his experiences working in the Norwegian Ministry of Foreign Affairs in the late 90s and early 2000s. Because of Neuman’s position as both a researcher and as a former practitioner of diplomacy, and his practice of textually positioning himself in the field throughout the book, this ethnography provides quite an interesting case. As observed by one reviewer: ‘the book is more a reflective memoir that positions the author as both object and subject of inquiry than an ethnography or a political analysis of diplomacy vis-a-vis international politics’. ⁷⁰ Neuman’s methodological choices are present throughout the text, as it is written in personal style that puts his position and entanglement in the field front and centre. He discusses the anxiety caused by this style of writing for a scholar such as himself, who is socialised in political science, as in this discipline writing should reflect ‘the idea that you are writing from nowhere, about objectively given stuff’. ⁷¹ Yet, by rejecting such conventions and by engaging with self-reflection and a reflexive style of writing, Neuman was able to provide rich ethnographic accounts, as well as interesting interpretive analyses, of everyday diplomatic practice without abandoning the important principles of the ethnographic methodology.

Despite the complaints about the lack of engagement with ethnography in IR, the study of IR has seen quite a few ethnographic studies of organisations looking at policy-making processes and organisational cultures. To name a few key works, Stacia Zabuski’s study of the European Space Agency, ⁷² Chris Shore, ⁷³ Marje Kuus ⁷⁴ and Rebecca Adler-Nissen and Alena Drieschova ⁷⁵ studies of the European Union, Deepak Nair’s ⁷⁶ studies of the Associations of Southeast Asian Nations (ASEAN) and Stephen Hopgood ⁷⁷ on Amnesty International. The ‘crisis of representation’ should not be regarded as an obstacle for conducting ethnographic research and developing an ethnographic methodology in IR. Rather, ethnography offers a valuable perspective and entry point to asking important theoretical and methodological questions regarding IR’s object of study – notably social relations on a global level. This is not a straightforward task however, as there seems to be considerable methodological obstacles to doing ethnographies of the international level.

The Local and the Global

A common topic in the literature on ethnographic methods, particularly outside anthropology, are the methodological challenges that arise when attempting to study processes that cannot be precisely located. IR theories have always been concerned with matters of scale, as the traditional objects of analysis, states and the international system, are fundamentally and always abstractions and going beyond local sites. This makes a theoretical tension between the unit level and the system-level; the particular and the general; between the local and the global, especially relevant with regards to the relationship between ethnography and IR. The problem that I am attempting to approach is how to conceptualise the ‘international’ methodologically, or more specifically, ethnographically. I argue that by choosing to abandon the imaginary of totalities such as the ‘locality’ or the ‘global level’, ethnography can enable important understandings of social relations that exist across these scales. This is based on the notion that sensibilities to local practices can in fact also inform theories of larger scales of interconnections and patterns.

Anthropologist Clifford Geertz wrote that the hope of anthropology is ‘to find in the little what eludes us in the large, to stumble upon general truths while sorting through special cases’. ⁷⁸ It is a common understanding that the purpose of ethnography is to pay attention to the particular in order to explain or understand ‘general truths’ such as the functioning of capitalism or the nature of rituals. ⁷⁹ However, IR has not traditionally shared anthropology’s interest in studying the particular. Kenneth Waltz famously warned against the ‘inductivist illusion’, a phrase he borrowed from anthropologist and structuralist Claude Levi Strauss, as ‘if we follow the inductivist route, we can only deal with pieces of the problem’. ⁸⁰ Within the traditional dominant paradigms in IR, theorising is at its best when it is founded in the system-level and deduced down from that, instead of inducing a totality from its parts.

Similarly, since the 1980s there has been an increasing concern in anthropology with the practice of ethnographic fieldwork in a ‘globalised’ world. The anthropologists Jean and John Comaroff addressed this unease when they asked: ‘Should we try to recognize where, in the particularity of the local, lurk social forces of larger scale, forces whose sociology demands attention if we are to make sense of the worlds we study without parochializing and, worse yet, exoticizing them’. ⁸¹ The problem has been that the simple study of a locality cannot possibly encompass the full scale of international or global relations. Studying a delimited local space or ‘a people’, as was common in traditional ethnographic accounts, was being critically examined as methodologically insufficient. As Arjun Appadurai argued, ‘locality’ is not something that is ‘out there’ to be discovered by the ethnographer but is in fact being partially produced by ethnographic representations. ⁸² Put differently, ‘the main object of ethnography is discovered to have always been, not in fact localities, but processes of “localization”’. ⁸³ Doing fieldwork on a single site was no longer seen as an adequate means of illuminating the ‘general truths’ in this changing context.

Rather, what came to be called ‘multi-sited ethnography’ emerged as a possible answer to the problem of how to adapt ethnographic research in a world that has somehow become less ‘local’. ⁸⁴ On a practical level, multi-sited ethnography entails conducting research in multiple fields that are conceived as being interconnected in some way. This methodology rests on several implicit and explicit theoretical positions. Important among these is the notion that cultural logics are always ‘multiply produced’ and that they are ‘at least partially constituted within the so-called system’. ⁸⁵ Because the idea that cultures are holistically contained within a delimited space was increasingly seen as a fallacy, multi-sited methods that allowed ethnographers to ‘follow movements of people, ideas, and objects, to trace and map complex networks’ were seen as more suited to understand social relations in the context of globalisation. ⁸⁶

These theoretical and methodological developments, as well as an increasing interest in globalisation, have meant that anthropologists have recently started paying attention to issues that have traditionally been in IR’s domain, particularly international political processes and global market relations. Lie, for instance, claims that his interest in ‘knowledge/power formations as constructed, consumed and articulated by and in the interface between the World Bank and Uganda’ required him to pay attention to ‘issues of state formation, sovereignty and globalisation’. ⁸⁷ In order to do so, Lie asserts that multi-sited fieldwork was required to ‘capture the relational aspect of power and, indeed, the partnership relation itself’. ⁸⁸ Similarly, in his anthropological study of peacebuilding and the formation of what he labels ‘franchised states’, Schia uses the multi-sited approach in order to ‘capture how words, concepts, and ideologies were being used and put into action, creating connections and links between various parts of peaceful building processes’. ⁸⁹ Long-term, multi-sited ethnographic fieldwork enables the analyst to identify the linkages between the ‘local’ and ‘global’ processes as they are constituted and imagined in practice.

The theoretical rethinking of ‘place and space’ in ethnographic research has thus opened up possibilities for how to study social relations on the international level. Indeed, the promise of the multi-sited approach was appealing to critically minded IR scholars who were asking questions regarding the linkages between everyday practices and the process of globalisation. ⁹⁰ Ethnographic methods started to be regarded by some in IR as a way to overcome state-centrism in making linkages between ‘local and global realities’. ⁹¹ Carol Cohn, influenced by Marcus ⁹² , used multi-sited ethnography in order to understand certain gendered US national security discourses. In her conclusion, she notes that

It is significant that over fifteen years, as I looked through a variety of windows, and listened to multiple local discourses and contextual permutations of national security discourses, I heard things in common, threads that could be pulled through; whether talking to generals or enlisted men, liberal strategists or a Secretary of Defense, certain continuities could be found. ⁹³

Furthermore, as was the case for Cohn’s work, multi-sited ethnography was seen by many as a way of ‘studying up’ – that is, shifting the focus from marginalised or ‘subaltern’ people to elites and decision-makers. ⁹⁴ For disciplines such as IR, where research has largely been focused on places of decision-making within the state and international organisations, the ‘multi-sited’ approach to ‘studying up’ emerged as a fitting means of doing ethnography.

When ethnographically ‘studying up’, such as in the case of research on power and global politics, there are, however, methodological problems involved in finding out what and where ‘the field’ is. This has been the cause of some anxiety for ethnographers concerned with politics and power. In the case of IR, claims have even been made that ‘the level of analysis encumbers attempts at ethnographic inquiry’. ⁹⁵ Because IR’s object of study is often seen as social relations that take place across polities, participant observation has to somehow take place on the abstract ‘international’ level. Thus, in a sense, it is difficult to gain insights into the core topics in IR, such as the causes of war and the practice of international cooperation, by relying on participant observation. As observed by Cris Shore, ‘there are many phenomena in the world that are not ‘easily captured by the ethnographer’s lens’. ⁹⁶ An example is research on organisations such as the European Union, which is ‘so large that it is difficult to grasp the whole entity from any singular or empirical perspective’. ⁹⁷ It is unclear what the EU is as a research object and where, exactly, it is located.

Furthermore, doubts about ethnographic efforts in IR and the following implications of abandoning attempts at structural explanations have been expressed in more critical corners of the discipline. Echoing the above concern, Koddenbrock argues that in IR as an ‘academic field tasked with studying world politics [. . .] systemic logics cannot be grasped by focusing exclusively on the minute details and their unstable assemblages’. ⁹⁸ Instead, what is needed in critical IR theorising is to ‘think big, to assert stability, totality and structure’. ⁹⁹ However, the multi-sited solution to this problem implies that doing fieldwork in multiple places within the EU somehow provides a more comprehensive understanding of its totality. Would this work for IR? The critique from IR about ‘thinking big’ relies on a presupposition that there is such a thing as big totalities in the first place. And this pre-conception is somewhat paradoxically shared with the multi-cited ethnographers.

There are certain problematic aspects of the assumptions underlying the multi-sited methodology. Anthropologist Mattei Candea has observed ‘a problematic reconfiguration of holism implicit (and sometimes explicit) in the multi-sited research sensibility – a suggestion that bursting out of our field-sites will enable us to provide an account of a totality “out there”’. ¹⁰⁰ There is an assumption that ‘any “global” entity is – must be, can only be – local in all its points’ and that ‘it follows that each “localized,” sited study is – must be, can only be – simultaneously a study of the “world-system”’. ¹⁰¹ In their investigations of the linkages between the local and the global, ‘multi-sited research replicated the implicit holistic assumption [. . .] that accounting for local ethnographic phenomena must involve locating them within an encompassing trans-local “system” located theoretically at a “higher” level’. ¹⁰² Candea argues that the world that Marcus and others imagine in their multi-sited approach is ‘woven of a single, many-stranded cloth (albeit with its knots, rips, and tears). By implication, therefore, a field-site is a contingent framing cut out of this seamless reality’. ¹⁰³

An illustrative case of this tendency can be found in the study of migration. Because of the way migrants are said to move between many different locations, multi-sited research has been seen as essential in understanding and tracking these movements. ¹⁰⁴ Further, terms like ‘transnational community’ have been used to label the diasporas, suggesting that all refugees had a sense that they were part of a global community beyond their locality. ¹⁰⁵ However, is it necessary to conceptualise a particular migrant community as being part of a wider transnational community? This is clear in anthropologist Ghassan Hage’s ethnography of several Lebanese refugee villages:

The thing that struck me when I began to globally locate the migrants from each of the villages I was researching is that there was no necessary strong sense of transnational community among them. Villagers certainly had a sense of belonging to a village whose members are globally spread around the world. But to say that the villagers formed a global community and that they had a communal sense of belonging to a transnational community of villagers was a bit stretching it in most cases. ¹⁰⁶

If we are to take seriously the fundamental aspect of the ethnographic sensibility, seeing things from other people’s point of view, ethnographers should refrain from creating connections between people that are not to some degree reflections of the world as imagined by the people themselves. On the other hand, referring back to Koddenbrock’s point about thinking in terms of totalities in IR, I consider this anthropological insight too weak. It is not necessarily so that people’s view of the world can disqualify IR’s view of the world. But there is perhaps another way in which anthropology makes a stronger point about the larger systemic totalities.

According to a recent anthropological trend, the existence of a ‘higher level’ should be questioned. In a story borrowed from a Buddhist parable that I encountered in several articles on ethnography and the problem of scale, ¹⁰⁷ three blind men who have been blind from birth and never encountered an elephant before are asked to each touch one part of an elephant. When the men are asked to describe what sort of animal the elephant is, each of them ends up providing a completely different description based on what part they were touching. ‘None manage to grasp the animal as a whole, yet all of them make reasoned deductions based on first-hand experience’. ¹⁰⁸ This parable attempts to convey the challenges of attempting to grasp the whole by looking at its parts, as in the case of the EU. However, the problem with such a methodological conceptualisation is that it assumes that the elephant exists in the first place. Or, put differently, what use is a description of an elephant that is not a description of its parts and their functions. Koddenbrock’s point is a deductive one in the sense that totalities have their functions and the exitance of parts deduced from the functions of the totality. However, as opposed to the blind men and the elephant, in the case of the ethnographic methodology there is no need to imagine that an ‘international’ systemic level even exists. According to Cook, Laidlaw and Mair, the idea that ‘accounting for or explaining “local” practices should consist of locating them within a wider system’ should be problematised. ¹⁰⁹ As an example, the so-called ‘world religions’ such as Christianity, Islam or Buddhism do not necessarily exist as coherent wholes. In the case of Buddhism: ‘we must be careful not to attribute explanatory power to an authentic, homogenous Buddhism’. ¹¹⁰ This is related to the common practice in social sciences to attribute explanatory power to ‘unseen forces exerted by a distinctive kind of unseen entity’. ¹¹¹ Entities such as ‘social structure’ becomes infused with much explanatory power and depth as it is alleged to exist on a ‘higher plane’ and is therefore seen as being intrinsically of a ‘wider scale’. However, they add, ‘we do not change the quantity or detail of the data we encounter merely by changing scale; we simply encounter different details’. ¹¹²

As a possible way to overcome this problematic bias towards totalities, drawing inspiration from Bruno Latour, some ethnographers have argued in favour of ‘flat description’. ¹¹³ If we eschew the notion that different levels of analysis exist a priori to our description of them, only this ‘flattening out’ can allow the analyst to understand ‘how the effect of there being different dimensions is generated and maintained’. ¹¹⁴ This is a methodological argument, not a theoretical one. These authors are not suggesting that the social world is flat, but rather that it is methodologically advantageous to study it as such. ¹¹⁵ ‘Flattening out’ the social world can enable IR scholars to ethnographically study how global politics exists, is maintained and practiced in particular localities. Put differently, rather than regarding localities as different sites of the global system as parts of an unseen whole, it can be beneficial to regard localities in their own right as different perspectives on how the global is practiced and maintained. Referring back to the earlier discussion on constructivist scholars’ attempts at developing a methodology that can overcome the struggle between different levels of analysis (the agent-structure problem primarily), I believe that such a methodological conceptualisation provides, or at least introduces, such an approach. In terms of Koddenbrock’s ‘thinking big’, this is still possible, but ethnography can never be deductive. Hence, systemic totalities need to be established each and every time as it is inducted from observations.

In a way, what I have attempted in this paper to provide an effective methodological way to approach a social world that does not present itself to the critical scholar as objective or easily comprehensible. The next chapter will discuss the more practical elements of this methodological reconceptualisation.

Practicing Ethnography

As observed by Merje Kuus, ‘it is one thing to recognise the analytical value of ethnographic fieldwork; it is quite another to do it’. ¹¹⁶ A problem that emerges within the multi-sited imaginary is how field-sites should be chosen. Candea argues that within the literature on multi-sited ethnography, this process of selecting a field-site is often omitted. ¹¹⁷ This is due to the implicit assumption that field-sites are ‘products of often conflicting political and epistemological processes “on the ground” and that it is this process which should be the object of ethnographic study’. ¹¹⁸ Part of the multi-sited methodology is the acknowledgement that while it ‘investigates and ethnographically constructs the lifeworlds of variously situated subjects, it also ethnographically constructs aspects of the system itself through the association and connections it suggests among sites’. ¹¹⁹ Marcus saw multi-sited ethnography as a revival of a certain understanding of constructivism, as these ethnographies ‘define their object of study through several different modes or techniques’ that ‘might be understood as practices of construction’. ¹²⁰ This is an explicitly emancipatory and deconstructive project that acknowledges the way that ‘sites’ and boundaries are politically and epistemically constituted, it is at odds with the idea that the properties of a ‘site’ should be discovered through open-ended research. In the words of Candea: ‘on the one hand [. . .] multi-sitedness highlights the construction and contingency of sites in a seamless world; on the other, the desire to leave “siting” to others, and to study “their” sites, seems to revive earlier notions of a site as a really existing entity out there, something to be discovered’. ¹²¹

What the ‘object of study’ is, whether it is bounded, unbounded or something else entirely, is always a consequence of an analytical decision. As previously asserted, this choice is a matter of methodological principle, however, not of an ‘a priori account of the world itself’. ¹²² Boundedness or unboundedness should not be represented as something that has been discovered to be ‘out there’, but rather as a consequence of the researcher’s practice of ‘bounding’. ¹²³ Candea therefore defends the traditional bounded field-site not because it is found to be bounded, but because all field-sites are in a sense arbitrary. Ethnographers need to embrace this fact and consciously choose the boundaries of the field without assuming that the field is already bounded. While the field-site as a place is arbitrary, the choice of a field with regards to the theoretical concerns motivating research is not. ¹²⁴ Put in more practical terms, asking a research question should precede choosing a field-site, as the latter is the practice of ‘constructing a field’ that will allow the ethnographer to answer the research question. ¹²⁵ If we follow the idea that all locations are arbitrary and always constructed to their logical end, there are no places that in-and-by their status as such offer more explanatory power than others outside the stated objective in the research question.

As previously argued, what is required is a process of ‘un-siting’ and consciously ‘re-siting’, not a process of finding multiple sites and the connections between them. As there are no such things as parts and wholes, ethnographers are only ever capable of making ‘partial connections’:

Certainty itself appears partial, information intermittent. An answer is another question, a connection a gap, a similarity a difference, and vice versa. Wherever we look we are left with the further knowledge that surface understanding conceals gaps and bumps. ¹²⁶

Therefore, it does in a sense not matter where you go but rather what understandings can be gained about a particular topic in a particular place.

As I have shown the practice of ethnographically ‘studying up’ has been viewed as a challenge both methodologically, but also with regards to the problem of access. Within anthropology, locations such as state institutions and international organisations had previously been left largely ignored, as the discipline has traditionally privileged small-scale societies as sites of knowledge. ¹²⁷ This turn was therefore met with some anxiety by anthropologists as ‘it became clear that studying up and “new” empirical scopes gave rise to practical and methodological challenges that might undermine the scholarly embodied ethnographic endeavour’. ¹²⁸ It was seen as more challenging to conduct traditional ethnographic fieldwork in the social spheres of the state, corporations and international organisations ‘as these sites have a political, bureaucratic and formal character that restrict access and limit what informants can say’. ¹²⁹

Tristen Naylor expresses the difficulties of doing ethnographic research involving participant observation in his study of G20 summits: ‘the largest challenges to this approach was gaining access to summits, a kind of problem that “looms large” in ethnography more generally’. ¹³⁰ . Similarly, Gusterson reflects on what he might have encountered during his fieldwork among nuclear weapons scientists if he was able to gain more access:

Although I spent as much time as possible simply ‘hanging out’ with Lab employees in church, in their homes, and on hikes, I sometimes wonder what I might have seen had I been allowed to come into the Lab day after day with my notebook and fade into the background. ¹³¹

There exists no particular or generalisable ways to overcome the challenge of access as circumstances vary. This is, as previously observed, part of the distinctiveness of ethnography as a method. As pointed out by Atkinson and Hammersley, ‘the discovery of obstacles to access, and perhaps of effective means of overcoming them, itself provides insights into the social organization of the setting or the orientations of the people being researched’. ¹³² Further, as previously touched on, it is important to acknowledge that the knowledge or ‘data’ that are gained through fieldwork is always partial and incomplete.

Furthermore, some scholars have challenged the idealised versions of ethnography as having to consist of ‘Malinowskian’ fieldwork and questioning the notions of ‘pure ethnography’ as stereotyped versions of what ethnography should be. Within IR, this view was posited by Rancatore in the aforementioned debate with Vrasti. ¹³³ Rancatore writes that it does not matter whether an ethnography is ‘real’ or not, in the sense of following idealised notions of what fieldwork is – the focus should rather ‘be on the kinds of explanations created and what their usefulness entails. What questions can now be drawn into the research frame? What do the results tell us about politics?’. ¹³⁴ Similarly, Lie in his contribution to the same debate, argues that although it is perhaps more difficult to achieve the classical ideals of ethnographic fieldwork when engaging with the state and international organisations, these restrictions can also have generative effects on research. ¹³⁵ Although accessing the field and participant observation might be limited by bureaucratic and political regulations, ‘turning challenges into opportunities is contingent on the researcher’s reflexivity and ability to adapt to new circumstances’. ¹³⁶ Thus, as ethnographers are leaving behind ‘outdated’ notions about the contained field-site, idealised notions about what fieldwork entails should also be renegotiated and reworked.

Conclusion

Even though IR’s object of study is by its very definition focused on phenomena that are conceived to exist on a larger scale and not easily located ‘on the ground’, there is a purpose in employing the ethnographic methodology when attempting to either explain what is happening in IR or developing empirically grounded theory. There is an ongoing debate in the IR discipline on what the international really is and how we can theorise it. ¹³⁷ In the methodological conceptualisation I have attempted to introduce in this article, IR can always be accessed locally, and this approach will necessarily lead to different kinds of theorising. As MacKay and Levin highlight, ethnography is not distinguished from positivist methodologies not ‘by its refusal to explain but by the kinds of explanations offered’. ¹³⁸ When investigating phenomena that the ethnographic researcher has grounded in a particular field, the detail that is found will necessarily be different but not of a somehow ‘lower’ explanatory value.

Lurking behind this conceptualisation, is the question ‘what if there is no international?’ This is a methodological question of course – a way of moving beyond the image of the elephant. The parts of the whole do not necessarily provide ‘lesser’ detail, just different detail. When attempting to gain insight into the practice of multilateral diplomacy for example, no location(s) can provide a ‘fuller’ or more comprehensive understanding of the phenomena than any other location. If we disregard the imaginary of bounded localities and totalities and global structures that exist ‘above’ the local, then any site is a potential place to discover the practices and functions of IR. As this site does not in itself exist ‘out there’ a priori to the researcher’s identification of it, and is in a sense always arbitrary, finding it is the matter of the ethnographer conceptually ‘bounding’ it in order to answer a particular research question.

‘The international’ is both present in and absent from local, situated lived-in worlds. I believe that approaching the ‘international’ realm ethnographically is possible, but that it will necessarily take a different form than traditional anthropological approaches. One common option is using the multi-sited methodology, which does especially make sense when attempting to understand the process and changes within international aid and development. Another option is to reimagine, methodologically, what the international means, in terms of totality: rather than regarding localities as different sites of the global system as if they are parts of an unseen whole, it can be beneficial to regard localities in their own right as different perspectives on how the global is practiced and maintained.