Reimagining Comparisons in International Relations through Reflexivity

Abstract

The article argues that International Relations, and especially those approaches that are informed by the epistemological and methodological premises of reflexivity, would benefit from a more diversified range of comparative methodologies other than those deriving from the work of J.S. Mill and more recent developments within the neopositivist canon. While discussions of methodology in International Relations have become open to a diversity of approaches in recent years, scholars have often been less prone to formulate explicit methodological guidance, especially in the form of practical guidance for alternative comparative research designs. Building on but further developing existing work on reflexivity and methodology, the article thus aims to open up methodological possibilities for reflexive IR by delineating three comparative strategies: defamiliarizing discursive comparison, contrapuntal comparison, and vernacular comparison. Each of the strategies is explained with reference to its theoretical and methodological background in existing scholarship, two key stages for its practical application, as well as examples. The article concludes by highlighting the importance and urgency of methodological innovation in IR––especially when it comes to approaches inspired by reflexivity.

Este artículo argumenta que las Relaciones Internacionales (RRII), y especialmente aquellos enfoques que se basan en las premisas epistemológicas y metodológicas de la reflexividad, se beneficiarían de una gama más diversificada de metodologías comparativas distintas de las derivadas del trabajo de J. S. Mill y de los desarrollos más recientes que han tenido lugar dentro del canon neopositivista. Si bien los debates relativos a la metodología en el campo de las RRII se han vuelto más abiertos con respecto a una diversidad de enfoques durante los últimos años, también es verdad que los académicos, con frecuencia, han sido menos propensos a formular una guía metodológica explícita, especialmente en forma de una orientación práctica que pueda servir para diseños alternativos de investigación comparativa. El artículo parte de la base de los trabajos existentes en materia de reflexividad y metodología, pero los desarrolla de manera más profunda, con el propósito de intentar abrir nuevas posibilidades metodológicas para las RRII reflexivas delineando, para ello, tres estrategias comparativas: la comparación discursiva desfamiliarizante, la comparación contrapuntística y la comparación vernácula. Explicamos cada una de estas estrategias con referencia a sus antecedentes teóricos y metodológicos en la literatura académica existente, con respecto a dos de las etapas clave para su aplicación práctica y, también, utilizamos algunos ejemplos. El artículo concluye destacando la importancia y la urgencia que tienen la innovación metodológica en el campo de las RRII, especialmente cuando se trata de enfoques inspirados en la reflexividad.

L'article affirme que les relations internationales, notamment les approches renseignées par les postulats épistémologiques et méthodologiques de réflexivité, tireraient parti d'une diversification des méthodologies comparatives, pour ne pas s'appuyer seulement sur celles dérivées du travail de J. S. Mill et de travaux plus récents au sein du canon néopositiviste. Bien que les discussions méthodologiques en RI s'ouvrent à diverses approches ces dernières années, les chercheurs se montrent souvent moins enclins à formuler des lignes directrices méthodologiques claires, surtout sous la forme de conseils pratiques pour adopter d'autres conceptions de recherche comparative. Se fondant sur des travaux existants sur la réflexivité et la méthodologie, mais en les approfondissant, l'article vise donc à ouvrir le champ des possibilités méthodologiques pour les RI réflexives en présentant trois stratégies comparatives : la comparaison discursive défamiliarisante, la comparaison en contrepoint et la comparaison vernaculaire. L'on explique chacune des stratégies par référence à son contexte théorique et méthodologique au sein de la recherche existante. L'on cite par ailleurs deux étapes clés pour son application pratique, ainsi que des exemples. L'article conclut en soulignant l'importance et l'urgence de l'innovation méthodologique en RI, notamment lorsqu'il s'agit des approches qui s'inspirent de la réflexivité.

Introduction

For a long time, the practice of comparison in IR has been predominantly linked to a limited range of comparative methods––chiefly though not exclusively derived from J.S. Mill's work––that are ill-suited to a large group of IR scholars whose approach to theory, methodology and knowledge production more generally is inspired by reflexivity as a research approach. Some researchers have responded to this by eschewing comparative methods as leading to a loss of necessary nuance (as noted by Boswell, Corbett, and Rhodes 2019, 2–3),¹ while those interested in comparing have found themselves in a double bind: because of the lack of alternative methodological guidance, by following conventional approaches to comparative research they risk producing research that is at the same time internally incoherent and always falling short of the benchmarks and standards set by dominant comparative approaches. Despite recent progress (as discussed below), this article maintains that IR needs a more explicit elaboration of comparative methodologies better suited for those scholars whose work is informed by reflexivity, and provides three strategies for comparison that can guide research practice (as well as teaching) and inform a more pluralist discipline.

In International Relations, reflexivity has come to define a broad and theoretically diverse range of scholarship, characterized by Alejandro (2020, 3) as serving an epistemological purpose of “dealing with the assessment of biases that limit the validity of knowledge claims”; and a socio-political one that “refers to the awareness that knowledge and discourse co-produce the socio-political world they refer to.” Researchers working with reflexivity commonly employ “modes of theorising” (Guzzini 2013) other than those linked to the hypothetico-deductive model . Rather than treating theory as external to the research process (either as the starting point for deriving hypotheses, or the theoretical generalization resulting from analysis) they see it as “the condition of possibility of knowledge” (Guzzini 2013, 531). Methodologically, this implies––for example––that “cases” are constructed through theorizing rather than selected on the basis of hypothesized relationships between variables. Moreover, reflexivity requires a critical assessment of the researcher's role in the process of knowledge production, and thus a constant dialogue between our analysis of the social world and the conditions that make the research possible. Comparison, if part of one's research plan, would thus require a fundamental rethinking––or reimagining, as the article's title suggests––along different methodological premises.

While reflexivity is increasingly written-about, featuring a diversity of definitions of the term itself and its implications for research (Hamati-Ataya 2013; Amoureux and Steele 2016; Knafo 2016),² there is still work to be done to fully develop its role as a research practice and as a methodological guide (Alejandro 2020; Alejandro and Knott 2022). The article seeks to harness the potential of reflexivity as a research practice that emphasizes the importance of making explicit research choices that are otherwise taken for granted. Sill, while centering reflexivity, I share with recent methodological work from political science and IR (Boswell, Corbett, and Rhodes 2019; Lai and Roccu 2019; Simmons and Smith 2021) the concern that comparison and case studies have been, for a long time, practiced and taught in IR from within a specific canon––mostly derived from Mill's work, especially in the United States and United Kingdom (Bennet and Elman 2007; Mahoney 2007).³ This is symptomatic of the long-standing tendency of mainstream IR to anchor itself methodologically to political science,⁴ and more specifically in this case to the field of “comparative politics”, which has only more recently started challenging the methodological status quo. Among such recent contributions that speak beyond disciplinary and methodological boundaries between IR and political science, Boswell, Corbett, and Rhodes (2019) suggest approaching the practice of interpretivist comparison through dilemmas, as an analytical route toward understanding what shapes beliefs, meanings and motivations of social actors across contexts (Boswell, Corbett, and Rhodes 2019, 41–2). Simmons and Smith (2021, 2–4) bring together an epistemologically diverse group of scholars to explore methodological possibilities beyond “controlled comparison.” While the argument engages with and builds on these works, it emphasizes the specific need for reflexivity in comparative work. A commitment to reflexivity entails, for the purpose of this article, both a critique of dominant approaches to comparison based on methodological/epistemological frictions, and an examination of the power relations and hierarchies that comparative methods have often reflected and upheld in our field. Reflexivist comparisons thus call for an awareness of the assumptions embedded in the construction of comparative frameworks and cases, and of the consequences, this has not only for our research findings, but for the people and places we research.

Therefore, I focus my contribution in this article on the development of three specific strategies, which are tailored to the needs of reflexivity as a research approach but flexible enough to be adapted and used through a range of fieldwork, archival, and other methods: comparison as a defamiliarizing discursive strategy, contrapuntal comparison and vernacular comparison. Building towards this, in the first section I review more established, dominant approaches to comparison and their frictions with reflexivity, and more recent developments in the literature that offer methodological grounding for alternative ways of comparing. This section also connects the paper to some of the literature on interpretivist approaches to the study of politics and IR, and emphasizes the conceptual links to a range of IR works that pave the way for methodologies informed by reflexivity. In the second section I unpack the key principles that underpin the comparative strategies proposed in the article, before introducing the three strategies: for each, I outline two steps for their practical application, and provide illustrative examples drawing on key questions and debates in IR literature. Comparison as a defamiliarising discursive strategy leverages the discursive use we make of comparison in order to create comparative frameworks that unpack and reexamine seemingly familiar connections. Contrapuntal comparison, building on Said's (1994) work and its uses in IR, juxtaposes narratives and perspectives, reaching for those that are left out of dominant frameworks, mapping historical connections across cases through transnational histories of colonialism and inequality, and subsequently reviewing how such connections transform our understanding of the cases and how they relate to the international. Lastly, vernacular comparisons look at how comparisons are constructed and used in the social world, as a way of grounding our understanding of political phenomena in the comparative frameworks of social actors. In laying out these three strategies, the article fulfils on the one hand the important task of providing practical advice for those practicing these methods in their research. On the other hand, it also equips researchers from across varied research traditions to better appreciate and interpret the contribution of reflexivist comparative research, on its own methodological terms.

Millian Methods and Their Legacy

The work of John Stuart Mill, especially his “Method of Difference” and “Method of Agreement”⁵ has been mainstreamed across large swathes of methodological scholarship and textbooks in US and UK IR (Bennett and Elman 2007; Mahoney 2007). Lijphart (1971) and Przeworski and Teune (1970) further developed Mill's approaches into the contemporary “most similar” and “most different” systems designs.⁶ Briefly, the most similar systems design is based on the comparison of cases similar in all respects except for a key explanatory (independent) variable. If the outcome (dependent variable) varies between the cases analyzed, then the cause can be traced back to the independent variable. In the most different systems design, cases differ in all respects except for the key explanatory factor. If the outcome is the same across the cases under scrutiny, this must then be due to the independent variable identified above.

These approaches are commonly applied to small-n comparisons for the purpose of establishing causal relationships where experimental and statistical methods are not applicable due to the small number of cases and large number of variables (della Porta 2008), while mimicking the same logic. The conception of causality applied here derives from Hume and Mill and responds to three criteria: constant conjunction (where two things that are causally linked repeatedly appear together); the logic of control, which assumes that you can isolate the effect of one variable onto another by holding other factors “constant”; and directional relationships, that is the possibility of clearly identifying cause and effect, where the former comes before the latter. The logic of control and specific case selection strategies are therefore harnessed for the purpose of approximating the logic of experimental research.⁷ In this approach, cases are understood as bounded, independent units––that is, they should not “affect each other with respect to the outcomes of concern” as otherwise, they would not be providing “independent evidence” that is needed to establish the presence of a causal relationship (Gerring and Cojocaru 2016, 19). Shilliam (2021, 99) calls this the assumption of “endogeneity.”

Mill himself found the application of his method to be ill-suited to the study of society and human relations:

Still less is this method applicable to a class of phenomena more complicated than even those of physiology, the phenomena of politics and history. There, Plurality of Causes exists in almost boundless excess, and effects are, for the most part, inextricably interwoven with one another. (Mill 1882, 324)

Addressing these problems (the plurality of causes, the difficulty in isolating effects) has led to further developments in comparative methods, such as Qualitative Comparative Analysis (QCA). This was developed to facilitate the analysis of more complex forms of causality than relationships between individual variables, looking for necessary and sufficient conditions by analyzing whether each of a set of conditions is present or absent (true or false) in each case (Ragin 1987). This is understood to be a way of accommodating more complex comparisons, involving a larger number of cases and factors by relying on Boolean logic, rather than on inferential statistical techniques, to arrive at causal inferences.

While there are some differences in the way the “canon” of comparative methods is taught, especially if we look beyond the United Kingdom and the United States, diversity is still usually leveraged for the purpose of causal inference and generalization of the kind outlined above. Comparative methods are in fact often taught by referring to the work of Emile Durkheim and Max Weber in continental Europe (see Morlino 2018).⁸ The distinction between the two is what informs Della Porta's (2008) account of “variable-based” (Durkheimian) and “case-based” (Weberian) approaches to comparison, with the latter focusing on understanding cases as “interpretable whole(s)” (204) and seeking to grasp “relations among constituent parts of a unit, a case” (205) rather than focusing on the relationships between variables. This approach is, according to Della Porta (2008, 204–7), used both by scholars who seek to interpret cases for their own worth and for the purpose of causal analysis, and leads to explanations that are more historically grounded and generalizations that are limited until further research is conducted on more cases (see also Mahoney and Rueschemeyer 2003). Other authors go further in outlining Weber's approach to the comparative method as one that is animated by a different approach to generalization, one that uses cases to outline the ideal-typical features of social phenomena in its varied manifestations.⁹ Just as often, scholars explain the comparative method with reference to Alexis de Tocqueville's strategies in Democracy in America. Marradi (2004), after discussing the meaning of comparison in its everyday as well as academic uses, provides a critical overview of the scope of various types of comparative research that is more diverse than most contemporary textbooks. He classifies academic uses of comparison on a continuum that goes from the most nomothetic (aimed at identifying general laws) to the most idiographic (concerned with particularity), with Tocqueville and Weber representing different varieties of idiographic approaches. Throughout the chapter, he is clear that the causal purpose of comparative research is often taken for granted but should not be, as comparison can also be used for other purposes: among other those cited here, reconstructing the logic of a system, highlighting contrasts, outlining the phases of a process of change.¹⁰ These variations have not been explored to their full potential. Some of this literature still relies on assumptions about generalization as a goal and implicit quality criterion (even when such generalization is to be achieved through more contextualized means and incremental approaches), and about cases as independent units, which reflexivity can help unpack. Moreover, practical guides to applying comparative research strategies are still mostly developed with reference to Milian methods and subsequent developments like QCA. Overall, then, existing debates have constrained the potential uses of comparison, and exhibit specific frictions with reflexivity that new comparative strategies ought to overcome.

Frictions with Reflexivity

Two such frictions with reflexivity are explored in more detail here. One of these has to do with the logic of control, which assumes you can isolate the effect of one variable on another by holding all other factors “constant.” This responds to the goal of finding relationships that are as general or universally applicable as possible. One of the implications of the logic of control for comparative research is that case selection, which is premised on the researcher's knowledge of a set of relevant key features about the cases, should take place at the stage of research design because inferences made about what causes differences or similarities among the cases hinge on case selection strategy. The researcher's point of view is also assumed to be neutral and unchanged throughout the research process––for instance, when moving from the analysis of one case to another.

Research inspired by reflexivity tends instead to see data collection as fine-tuning the theoretical lens through which cases are defined and analyzed––as theory is part of the research process, rather than being external to it (Guzzini 2013), and the data itself is also the product of positionalities that evolve over time. Whereas under the logic of control cases are understood as a fundamentally independent and bounded,¹¹ relational thinking that is common in much reflexivist IR would posit that cases cannot exist fully independently from one another.

A second friction of mainstream comparative methods for reflexivity in IR has to do with the issue of methodological nationalism as the tendency to isolate what happens within a polity from what happens outside of it. While not inevitable, methodological nationalism is in a sense a consequence of assuming cases to be bounded units providing independent evidence of causal relationships (Gerring and Cojocaru 2016, 19). Tilly (1997, 46) already argued that globalization posed a challenge to conventional approaches to comparison, as states could no longer assumed to be independent rather than interdependent units. Other scholars have gone further in their critique, pointing out that taking the nation-state as “the self-evident container of political, cultural, and economic relations” is not just an intellectual mistake but a “social process” rooted in and reinforced by the practices and institutions of the modern nation-state (Goswami 2002, 794). Methodological nationalism, therefore, risks emphasizing exemplary or deplorable features of states as artificially bounded units while removing them from the international and transnational processes within which they have been historically entangled, privileging a view of history from the perspective of nation- and state-making as opposed to, for example, colonialism and imperialism (Stoler 2001, 847). When trying to explain phenomena such as democratization (Roccu 2019) or revolutions (Lawson 2019), methodological nationalism tends to overlook how the international acts through or is constitutive of domestic polities, and in turn how the domestic shapes the international (Roccu 2019; Riofrancos 2021). Reflexivity would caution us against unreflexively taking things like states as units of analysis without considering the socially constructed nature of such categories and their boundaries,¹² thus leading us to see comparative frameworks not as objective frameworks of evaluation, but as a construct that reveals as much about the assumptions and worldview informing the comparison as they do about the phenomenon and cases under scrutiny.

Moving beyond the “Canon” in Comparative Research

This section highlights three steps through which scholarship in IR and its cognate fields has been supporting the development of comparative approaches that move beyond some of the rigidities of the Millian legacy outlined in the previous section. From a theoretical-methodological point of view, debates in IR have increasingly highlighted the existence of more varied forms of theorizing (Guzzini 2013), concept development (Berenskoetter 2017), and also causation (Kurki 2008; Humphreys 2017) and explanation (see Jackson 2017 on the difference between causal claim and explanation), than accounted for by traditional empirical and hypothetico-deductive models of social science. Guzzini (2013, 543), for example, argues that much IR theory engages in “ontological theorising,” that is (that is, theorizing the key phenomena of international politics). Methodologically this type of work calls not for Humean causal inference, but constitutive explanations, which aim to account for the properties that make something what it is, its underlying mechanisms and meaning-making processes. Critical realists have also forcefully argued for understanding the social world as a relatively open system (Patomaki 2010) in which “we cannot expect to find the kinds of regularities which scientists seek to isolate through experiment” (Humphreys 2017, 660). As Jackson (2017, 691), notes, in this context “a causal explanation is not likely to look anything like a linear combination of discrete variables”; it will rather be a holistic causal story (Kurki 2008, 286).

This body of work highlights that the use of cases and comparison in research cannot be bound to using control to produce causal inference, and points at the need for the kind of strategies suggested in this article. Lerner's (2022) work, for example, explores how collective trauma constitutes many aspects of international politics and should be understood as “an ontological condition of international life” (Lerner 2022, 12). In line with the Weberian ideal-type conceptual approach of the book, the three cases selected––from India, Israel and the United States––do not respond to any Millian logic of controlled comparison, as the author leverages their diversity to illustrate the scope of his argument rather than for causal inference. His discussion of case selection adopts the terms “deviant” and crucial” (Lerner 2022, 18), while showing in fact how the cases are constructed as such through the research, rather than “found” and selected because of these features.

A second step in moving beyond the canon of comparative research has been the rethinking of the purposes of comparison and our understanding of how cases should be treated. Much of this progress comes from methodological scholarship in political science and IR, which is usually labeled as interpretivist because of its focus on the social construction of “truths” about the political world, and the belief that these can “only be accessed, or co-generated, through interactions between researcher and researched as they seek to interpret those events and make those interpretations legible to each other” (Schwartz-Shea and Yanow 2013, 4). Just like this article, Boswell, Corbett, and Rhodes (2019) set out to respond to the limitations of positivist approaches to comparison as well as the concern that interpretivists might not have use for it. Among the varied uses for comparison they highlight in the book, they argue that comparison can provide “decentred” explanations of the social world, for example, unpacking the underpinning beliefs and meanings that guide individual behavior across contexts (Boswell, Corbett, and Rhodes 2019, 4). Comparison is used to question settled or taken-for-granted explanations by introducing contrasting perspectives, rather than for hypothesis testing and causal inference. Case selection is thus a reflexive process that begins with provisional decisions about studying intrinsically interesting cases, which are adapted as the research progresses, and as the researcher's approach to the project as a whole evolves (Boswell, Corbett, and Rhodes 2019, 57–9). Recognizing the constructed nature of the case is also part of Simmons and Smith's (2021, 12) effort to overcome some of the limitations of controlled comparisons by expanding our notion of what it means to compare. In this volume, Soss (2021, 85) reflects in particular on the difference between studying a case (selected on the basis of key features, as in mainstream approaches) and instead the more interpretivist approach of “casing a study” as an “ongoing research activity designed to advance insight, understanding and explanation.” Comparison, as reframed in this interpretivist tradition, thus embraces complex specificity (as opposed to parsimony) while advancing conceptual and theoretical understandings of the phenomena under scrutiny (Boswell, Corbett, and Rhodes 2019, 151).

Lastly, this call for innovation has led to some debate regarding how comparison can be practiced differently and what strategies best respond to the need for greater reflexivity. Some researchers have responded to this by developing approaches to comparison variously defined as “relational.” Parkinson (2021, 157) argues for using a relational rather than a controlled framework in order to analyse how cases (in her research, armed groups specifically) change over time as their roles and the relations among each other change. “Encompassing” (Schwedler 2021, 176) or “incorporated” (McMichael 1990, 2000) comparisons see cases as part of a relational framework, in this case linking them to a larger process or structure that helps explain them.¹³Riofrancos (2021), also echoing some of the arguments in Lai and Roccu (2019), emphasizes how even in studies of single cases there is a process of co-constitution at play between a specific unit at the center of our analysis, and the general political phenomenon we are researching: “particular sites offer a view of the general not because they are cases of it, but because they are constitutive of it” (Riofrancos 2021, 108). The latter contributions by Lai and Roccu (2019) and Riofrancos (2021) have been explicitly informed by earlier work by Michael Burawoy (1998, 2009a, b). Intervening in debates around comparative ethnography (Burawoy 2017) and through his own original research, Burawoy suggested comparing cases (such as the factories where he carried out ethnographic work, in Hungary, Russia, and the United States) not as part of a controlled comparison, but on the basis of the macro-structures or processes within which they were embedded (capitalism and socialism; the transformations internal to capitalism and post-socialist transitions) (Burawoy 2009b). Cases are thus constructed as either different stages of a process, or as belonging to different macro-structures, allowing for a more contextual analysis of how regimes of power operate across cases. Building on these contributions, this article argues that more work can be done to operationalize comparative strategies informed by reflexivity as a guide for research practice––this is achieved by focusing more specifically on the practical steps and outcomes of the three comparative strategies discussed below.

Principles of Reflexivist Comparisons

In this section, I outline some key principles that can inform a reflexivist approach to comparison. While some of these are already embedded in reflexivist methodological traditions in IR (as will become clear in the discussion below), the focus here is developing their relevance for comparative methodologies in particular. They are based on premises developed from the previous section, including a reflexivist discomfort with the logic of control; overcoming a conception of cases as bounded coherent units, and avoiding methodological nationalism or “groupism”; using comparison for purposes other than establishing causality (and when adopting a causal purpose, not linking this to a Humean conception of cause-effect relationships); and finally a commitment to research practices that try to overcome, and not reinforce, power hierarchies and relationships. Drawing on an interdisciplinary set of sources from the humanities alongside the social sciences, this section discusses how to incorporate reflexivity in comparative research by presenting three principles that inform this kind of methodological approach.

First, reflexivist comparisons are informed by a closer relationship between concepts and context, and theory and history, which are not seen as independent, or “external” to one another. Berenskoetter (2017, 160) writes that understanding the relationships between concept and context––that is the frame of environment within which a concept is embedded––is essential in order to understand why concepts are understood and interpreted in a certain way. For researchers working within a reflexivist framework, one of the implications is the need to historicize and contextualize theoretical constructs (Berenskoetter 2017, 161–72). Building on an understanding of theory and history as deeply interlinked, reflexivist comparisons also address concerns around the unreflexive use of abstraction in IR that over time has contributed to reinforcing specific hierarchies of knowledge, for instance, according to Krishna (2001) an “amnesia” on the question of race. Abstraction is in fact part and parcel of our ways of thinking and making sense of the world, but is also “never innocent of power: the precise strategies and methods of abstraction in each instance decide what aspect of a limitless reality are brought into sharp focus and what aspects are, literally, left out of the picture” (Krishna 2001, 403). Reflexivist comparison would therefore require a kind of theoretical reflection that is in dialogue with (rather than abstracting from) the context where our research is situated, and that explicitly reflect on what issues or aspects of a phenomenon have historically been left out of comparative frameworks, while also subjecting our frameworks to “critical self-reflection” (Blaney and Inayatullah 2002, 126). Similarly, concepts are not elaborated and then “applied” to specific cases, but often situated in a specific context, adapted to circumstances, and imbued in power relations that should be explicated rather than elided (Schaffer 2015). The implication for comparative methodology is that cases are constitutive of theoretical frameworks, rather than descending from them through processes of case selection. As the following section will show, this leads to comparative strategies that place the construction of cases and frameworks at the forefront.

Second, when it comes to the question of defining cases and the practice of comparison as a method, reflexivity does away with the notion of cases as independent, bounded units. This also means moving past unreflexive “compare and contrast” strategies and tackling the question of commensurability and incommensurability directly (Steinmetz 2004; Stanford Friedman 2011). This speaks to questions raised by debates in comparative ethnography (Abramson and Gong 2020), which have grappled with the uncertainty in making judgments about the comparability of certain settings, and trade-offs involved in the choice between in-depth nuanced findings and greater breadth potentially afforded by studying multiple sites.¹⁴ One way of approaching this is by focusing on relationality instead, understood here as a “research sensibility that emphasizes connection and interrelatedness amongst individuals, communities, histories, and knowledge, as well as the worlds––both past and present––in which these are rooted” and moving away from approaching power and knowledge as “singular, fixed objects of analysis” to explore how they are constituted through “layers of knowledge, history, and material practice” (Tucker 2018, 227). For comparison, this means that cases are in fact porous, historically situated configurations constructed through the mapping of these connections. Defining comparison as relational can also help us shift the focus from privileging a nationally-focused perspective to one where we look comparatively at the multiple ways in which transnational forms of rule––like imperialism––connected countries and contexts at various levels, “through circuits of knowledge production, governing practices, and indirect as well as direct connections in the political rationalities” (Stoler 2001, 831) that informed that rule. Comparative frameworks, therefore, are temporarily held in place and reflexively examined, rather than being set at the start of the research process.

Lastly, reflexivity entails a critical commitment to examining and redressing conditions of inequality, including those that are part of and give rise to our research. It thus prompts a rethinking of the criteria or objectives used to evaluate knowledge produced through comparisons: instead of nomothetic generalization, it could favor a more productive kind of “communicative generalisation” (Cornish 2020) informed by a reciprocal relationship with the research communities we work with as scholars. Here generalization can be produced through one's ability to communicate to multiple audiences (“Could I make my argument in front of the people I have engaged with as part of the research, as well as in front of my scholarly community?”) or by engaging with the perspectives of multiple speakers, giving rise to a less abstract and more nuanced account of the phenomenon under scrutiny (Cornish 2020, 90–2). It could also be framed as a logic of translation (Cheesman 2021, 76), that is as a route to provincialize (Chakrabarty 2007) universal categories that are otherwise imposed across diverse contexts. This kind of logic is also applied to theorizing: rather than the kind of abstraction devoid of context and history mentioned above, a reflexivist approach to comparison calls for a kind of theorising that conceptualizes without reifying (Berenskoetter 2017, 161), that works towards theoretical innovation while maintaining complex specificity (Boswell, Corbett and Rhodes 2019, 151), or highlighting how a certain theoretical lens is “relevant beyond a specific scenario or era” (Lerner 2022, 18). The elements described above are not exhaustive of the methodological implications of reflexivity as a whole. Rather, they are highlighted here because they have crucial implications for how we approach comparison as a methodology from the point of view of reflexivist IR. These elements thus inform the discussion of the three compartive stragies discussed in the following section––comparison as a defamiliarising discursive strategy, contrapuntal comparison and vernacular comparison––which are in turn intended to guide research practice while leaving scope for adaptations and further strategies to be developed in the future.

Reimagining Comparison: Three Strategies

Here I develop three strategies for comparative research in IR that fall outside the Millian framework, and that might be of practical use to scholars whose research is informed by reflexivity. The strategies build on existing literature, both methodological and substantive, but to this literature, I contribute an explicit elaboration of methodological strategies for comparing cases, something that is commonly left implicit in substantive literature based on comparative research, or absent from methods books. For each of the three strategies proposed below, I develop and explain two stages through which they are applied: first, the construction of comparative frameworks; and second a reflexive review of the terms of the comparison. The terminology of the “reflexive review” was developed by Alejandro and Knott (2022) to discuss linguistic reflexivity in the context of doing literature reviews. Here, the reflexive review of the terms of the comparison takes on an additional different meaning, alongside the terminological one, that is one materially and methodologically related to reviewing the conditions (terms) of the comparison. While they are explained in this order, in fact, reflexive review occurs in different ways at different stages of the research process. Comparisons are commonly treated as a methodology (that is, a research strategy informed by specific ontological and epistemological commitments) more than as a method (tools and practices for gathering and analyzing data), in methodological texts that draw this kind of distinction. In discussing how reflexivity can be put into practice when doing comparative research, this section also puts more emphasis on the research strategy informing a specific project than on specific methods, which might vary depending on the type of project. This means that the two steps discussed here––the reflexive review and construction of comparative frameworks––result in an indication of what kind of cases form part of the research, why, and how they are related to each other. As mentioned above, this must be complemented by further specification of methods and other elements of the research design that are outside of the scope of this article.

Comparison as a Defamiliarizing Discursive Strategy

Comparison as a defamiliarizing discursive strategy leverages for methodological purposes the use we make of comparison in speech and in writing, as we identify similarities or differences among phenomena to make ourselves understood and make an argument. Anderson (2016), in reviewing his scholarly production on nationalism, argued that in spite of what his methodological training had taught him, comparison should be understood first and foremost as a discursive strategy. He argued that the unusual, “surprising” comparisons of Imagined Communities (Anderson 1983) were meant to challenge the widespread assumption that nationalism originated in Europe and that “other” nationalisms should thus be compared to the European standard.¹⁵ Reflecting the key point made in the previous section about the close relationship between theory/concepts and history/context, comparison as a discursive strategy, therefore, puts emphasis on the constructed nature of comparative frameworks, but also on their performative function––as they make historical and social processes into “cases” whose categorization potentially shapes not only the process of knowing about them but also policy interventions.

Here in this article I specifically propose comparison as a discursive strategy with the intent of defamiliarizing ourselves with the things we think we know, or that we assume to be commensurable. Comparison as a methodology always implies an ability to assess commensurability among the phenomena under investigation, and to set standards according to which these commensurable phenomena will be compared. A reflexive approach requires considering the politics of comparison, and acknowledging that the way we use comparison, assess commensurability, and set standards reflects specific assumptions. Comparison as a defamiliarizing discursive strategy thus aims to challenge dominant assumptions by deconstructing and “making strange” what is otherwise taken as familiar. In doing so, it is meant to help “unravel[s] the self-other opposition that reproduces systems of epistemological dominance” (Stanford Friedman 2011, 759), by questioning their positioning as the standard against which other cases are compared.

Comparison as a defamiliarizing discursive strategy––as a methodology and not a method or technique––is meant to guide the construction of cases and frameworks adopted in the comparison. It requires complementary methods to carry out the comparison, gather and analyze data. The outcome of this strategy (see Table 1 for a summary of the steps and outcomes of the three strategies) is a comparative framework that turns a seemingly familiar understanding of the relationship between cases into a question that requires elucidation through an analysis of the discursive frameworks within which they are embedded. This approach can potentially appeal to scholars who are theoretically and methodologically interested in the study of discourse but can also constitute a way for others to incorporate a discursive dimension in a research project that relies on other kinds of methods as well. This comparative strategy can be broken down into two steps––the construction of a defamiliarizing comparative framework and a reflexive review of the terms of the comparison––which are explained here with reference to a brief example on transitional justice processes in Bosnia and Herzegovina (BiH; from Bosna i Hercegovina) and the United States. Transitional justice––as a set of mechanisms or initiatives to deal with the consequences of mass violence––has developed, since the conception of the term in the historical context of the 1990s,¹⁶ as something to be done in countries affected by war or coming out of authoritarian regimes. In the post-Cold War World, this has usually meant countries in the Global South or East (Müller 2020; Trubina et al. 2020)––such as Bosnia and Herzegovina. More recently, calls for transitional justice mechanisms like truth commissions for the crimes of settler colonialism, racial violence and slavery have led to the expansion of the field of transitional justice to the Global North, especially, but not exclusively, settler colonial countries like the United States.

The first step––the construction of the comparative framework––proceeds by defamiliarizing our understanding of the cases and their relationship, through an explicit discussion of the discursive construction of these cases. In this example, in this step, one would thus discuss the “surprising” effect of the comparison––due to the fact that countries like the United States are usually compared with their Western “peers” (on the exclusion of the United States case from transitional justice scholarship see Posthumus and Zvobgo 2021, 513–5), and Bosnia and Herzegovina usually associated with other post-war states that are framed as “ethnically divided.” Having made explicit the implicit understanding of unequal positioning of the two cases in the field of transitional justice, the construction of a defamiliarizing framework would move past the appearance of established knowledge in setting up the comparison. A common assumption––based also on the published literature on the subject––might be that BiH is the place where transitional justice has happened more, compared to the United States. However, if we analyze the politics of truth commissions as a transitional justice tool, this comparative strategy helps us deconstruct superficial understandings of the cases. In fact, the United States has actually had a number of truth commissions at state and local level, dealing with various aspects of racial and settler colonial violence (Posthumus and Zvobgo 2021). On the other hand, in Bosnia and Herzegovina (as in other former Yugoslav countries), attempts to create state-backed truth commissions were stifled (Dragovic-Soso 2016), truth commissions have been used for revisionist projects (Barton Hronešová 2022), and so generally failed. This could then constitute the comparative backdrop of researching transitional justice in the United States and BiH.

The second step in the defamiliarizing discursive comparison is a reflexive review of the terms of the comparison, which here consists of analyzing and reconstructing the discursive frameworks within which the cases are embedded. Similar to other methodological literature on reflexivity (Alejandro and Knott 2022), this also calls for attention to language and meaning as part of our reflexive practice. Here, “terms” of the comparison refers both to terms as words (terminological issues) and to terms as the conditions of the comparison. The implication of this for the example at hand would be discussing and questioning the categorical “boxes” within which Bosnian and US politics are normally understood: the former as a “post-war country” and the latter as an “established” or “liberal” democracy. Some relevant questions might include: when does a country stop being considered post-war? And what kind of post- is relevant to the US experience with truth commissions (does the post- refer to the Civil War, segregation, violence against indigenous people of North America?). Alternatively, and in light of political developments (including, for example, the conservative attack on abortion rights and the rise of far-right violence) would a review of the terms of the comparison recast the United States as transitioning away from liberal democracy (in contrast to the common framing of transitional justice being carried out in countries transitioning towards it)? Lastly, as reflexive researchers, we would also ask what the potential practical/political consequences of these reconstructed comparative frameworks would be, for example, in terms of supporting specific agendas or justice demands from certain groups. For this comparative strategy as for the other two discussed below, the reflexive review of the terms of the comparison is a necessary step to avoid falling into the trap of reifying these reconstructed comparative frameworks, which would go against the constructivist understanding of the social world espoused here.

Contrapuntal Comparison

Contrapuntal comparison draws from Said's writings on contrapuntal reading (Said 1994) and post- and de-colonial scholarship in IR informed by his thought. Contrapuntal reading was developed by Said as a method for recovering the presence of alternative perspectives in literary works that would otherwise be read exclusively from the context of the dominant culture within which they were produced. To read contrapuntally is therefore to “read back” and identify––for example––the presence of empire and the plantation economy in Jane Austen's Mansfield Park, as the empire is seldom mentioned, but underpins the ability of the English estate to exist and function as it does in the novel (Said 1994; Wilson 1994). As a form of comparison, contrapuntal reading can juxtapose narratives and perspectives, reaching for those historically oppressed or marginalized that are left out of dominant frameworks, maintaining a “simultaneous awareness both of the metropolitan history that is narrated and of those other histories against which (and together with which) the dominating discourse acts” (Said 1994, 51).

Contrapuntal comparison echoes Said's call for “worlding” texts, that is “historicising them (…) paying attention to the hierarchies and the power knowledge nexus embedded in them” (Chowdhry 2007, 105) or “revealing their entanglements with power” (Chowdhry 2007, 110), by applying it to the study of international politics. As highlighted by Bilgin (2016), contrapuntal reading is also ultimately about “connectedness”: making sense of connections between the things we study, and their intertwined and overlapping histories. Krishna (2001) argues that contrapuntal reading can help overcome IR's problem of “abstraction” that makes race substantially invisible to its theoretical paradigms, by removing such paradigms from the historical circumstances of their development. Abstraction is also a methodological tendency, one which reflexivity counters by promoting a closer relationship between theory and history, as discussed above. Contrapuntal comparison thus responds to a relational approach as a “research sensibility that emphasizes connection and interrelatedness amongst individuals, communities, histories, and knowledge, as well as the worlds––both past and present––in which these are rooted” (Tucker 2018, 227), and responding to calls for IR methodologies to follow “colonial and decolonial struggles across multiple (ontolological, epistemological, and geographical) sites” (Tucker 2018, 227–8).

This type of comparison links together cases through their mapping in a transnational field of colonial histories and inequalities, producing comparative accounts of political phenomena that are sensitive to the historical and structural power imbalances that constitute the international system. This approach might appeal thus beyond post- and de-colonial traditions that already pay attention to the contemporary implications of colonial history, offering a way for reflexive scholars of other theoretical inclinations within reflexive IR to embed analyses of these historical dynamics within their comparative frameworks. This strategy also consists of two key steps. First, we build contrapuntal comparative frameworks iteratively, through the research process, by mapping the historical connections of the cases. For example, “generations” of theories of revolution have conventionally privileged European perspectives, even when the cases analyzed and compared are not necessarily European revolutions (Beck 2018). Conventional comparative frameworks create standards for comparing revolutions on a “most similar” and “most different” basis, which does not tell us much about revolutions as transnational phenomena to be understood through an “inter-social” (rather than merely “international”) account of their anatomies (Lawson 2019). A comparative project following a contrapuntal strategy as the one developed here would potentially generate knowledge by comparing, for instance, the French Revolution with the Haitian one as a way of mapping the “transnational fields of contention” (Lawson 2019, 25) within which they existed and upon which they acted. As Bhambra (2016, 8) notes, the writing out of Haiti from a large part of revolutions literature “structures and distorts hegemonic accounts” of phenomena such as the creation of a European identity centered around ideas of equality and freedom, and its construction of a mirroring other.” Contrapuntal comparison would help us challenge such distortions by presenting a more nuanced, and relational, account of political phenomena such as revolutions.

In its second step, contrapuntal comparison is carried out through both a reflection on the categories we are temporarily adopting for the purpose of our analysis, and a review of how the connections we mapped transform our understanding of the cases and how they relate to the international. For instance, we might ask what a contrapuntal comparison of the French and Algerian Revolution would look like. The use of the term “revolution” to characterize the anti-colonial uprising against the French in Algeria is far from universally accepted––in fact, France only formally recognized the events as a “war” (as opposed to military “operations” or “events”) in 1999 (Vince 2020, 2). Therefore, a reflexive review of our contrapuntal comparison of Haiti and France would involve asking how the term “revolution” has been conceptualized and applied in comparative research, and what kind of exclusions that might have generated. This strategy could also be used to reflect on other relations of colonial exploitation (and violent insurgence against it) within which Haiti and France were imbricated, or look at the Haitian revolution in relation to subsequent movements in Latin America that were inspired or directly aided by it (Lawson 2019, 24–5). Taken together, the construction of contrapuntal frameworks and their review would therefore generate knowledge that is not “generalisable” in a nomothetic sense, but that expands our mapping of the phenomenon of interest by adding depth (a more comprehensive and “interconnected” historical context) and breadth (developing an understanding that centers links and relations, rather than “isolating” factors).

Vernacular Comparison

The third strategy proposed here is vernacular comparison. Rather than being just a methodology, here the comparison forms part of what is being studied as well as part of our methodological reasoning. One of the problems identified in this article is that the parameters of mainstream comparative methods often lead researchers to face the problem of methodological nationalism. Cases are isolated on the basis of a somewhat neat separation between what is endogenous and exogenous to them. According to Brubaker (2003), this is a problem that can be scaled down to the comparison of other kinds of units––such as ethnic, national, and religious groups––and that fundamentally shapes the kind of findings one might derive from such comparisons.¹⁷ Reflexivity prompts us to go beyond a critique of methodological nationalism or “groupism” to consider how our research can overcome this kind of rigidity that sits uncomfortably with the social world we are observing. While the article has already made clear that reflexive comparative frameworks are always held in place temporarily and subject to reflexive review, vernacular comparison offers a chance to ground our comparative frameworks in their social, vernacular use. It seeks to understand how comparisons are used, constructed and contested in the social world as a way of developing a comparative understanding of a phenomenon. Comparative frameworks therefore result from the research, where participants close to the political phenomenon we are investigating construct and compare as part of everyday discourse.

Vernacular comparison produces a comparative framework in relation to a political phenomenon that is grounded in its vernacular understanding and experience. It could be of interest to IR scholars studying communities of practice and policy-making as well as those working on everyday IR or the interplay of local and international agencies, as in the field of transitional justice and peacebuilding in the example above. Vernacular comparison also proceeds by constructing vernacular comparative frameworks as a first step. In this case, the process is a grounded one guided by similar methodological assumptions as Charmaz (2006) proposed for her version of grounded theory, where we begin with inductive data that are nonetheless imbued with some pre-existing theoretical assumptions (as data are co-constructed by researchers and participants; and always informed by some theoretical assumptions). This first step would thus consist of a grounded analysis of vernacular comparative frameworks. We can apply this here to the example of transitional justice research that has focused on the use of deliberative mechanisms as part of the process of dealing with the past, such as the consultations for the establishment of RECOM, a Regional Commission for the former Yugoslavia––(Kostovicova 2017). In this context, one potential use of vernacular comparison might be to look at deliberative processes such as RECOM to assess how participants constructed specific comparisons in order to make sense of the need for or aims of specific transitional justice mechanisms. One might look, for example, at how comparisons between the former Yugoslavia and South Africa––in relation to the Truth and Reconciliation Commission (TRC)––are leveraged and contested in RECOM consultation meetings, what kind of elements are compared, and how issues of commensurability are addressed. This kind of analysis––a vernacular comparison––thus leads to the construction of a vernacular comparative framework, one that is grounded rather than imposed from the outside, and that could prompt a rethinking of established assumptions about RECOM and the former Yugoslavia and the TRC and South Africa as “cases.”

The review of the terms of the comparison––the second step of the comparative process––is partly already embedded in their construction, as vernacular comparison draws from socially constructed understandings of what we compare. However, the second step of the reflexive review also takes a further analytical leap, by reviewing the terms of the comparison through theory construction, by making explicit the theoretical value of the comparison and refining our understanding of key concepts building on our new data. This contributes to the process of “elucidation” of concepts in interpretivist research (Schaffer 2015), where the aim of the research is often to shed light on how shared understandings are created, reproduced, imposed, disrupted or changed (Schaffer 2015, 7). If we refer again to the example of deliberative mechanisms for dealing with the past, in this phase of the comparison we could look at what the comparison of the former Yugoslavia and South Africa tells us about conceptions of transitional justice (and about justice more broadly), truth, or reconciliation that are held by participants to the RECOM meetings, which often included local community representatives as well as Non Governmental Organisations, victims’ associations, and veterans. For example, it might be theoretically insightful to note how the goal of “reconciliation” that was very prominent in the setting up of the South African TRC was overshadowed, in the RECOM discussions and the comparisons drawn therein, by the question of reparations. Confirming previous findings about the contested nature of reconciliation as a post-war project in the former Yugoslav region, this also allows for further, and more grounded, theoretical elaboration of the reparative dimension of justice after war. Constructing a specific representation of the South African TRC, this review also raises questions about the possible use of vernacular comparison in debates about transitional justice in South Africa, thus opening up further avenues for research. Lastly, the reflexive review of the terms of the comparison can also involve an effort at “communicative generalisation” (Cornish 2020) where findings are brought back to the communities that informed the research, and their relevance is discussed with them.

To conclude the discussion of these three strategies, it is important to note that for research inspired by reflexivity, comparative frameworks are always artificially and temporarily fixed, and thus subject to ongoing scrutiny and review. Even when they are grounded in social discourses and popular use, they should not be naturalized but seen as constituted socially, reflecting intersubjective understandings rather than an accurate representation of “fixed” identities. Moreover, it should be further stressed that the examples are only illustrative, and that comparative strategies are more akin to a methodology (a strategy for doing research) than to methods (the specific tools applied to gather and analyze data). Therefore, the steps and examples presented here are instructive but general enough to allow flexibility in their application, as they are meant to be adapted to specific projects and methods. These three strategies, sharing a commitment to reflexivity as a research approach, are thus meant to help in the research design of comparative studies, which is not confined in this case to “case selection” and the development of theory, but constitutes an iterative process in which cases and their comparison are built throughout the research.

Conclusion

In this article, I have drawn on recent debates on reflexivity, methodology and research practice in IR and beyond, in order to set out three strategies for comparison that can be more consistent with, and appealing to, a broader range of IR research. Instead of rejecting comparison as a whole based on the challenges posed by a specific approach to comparative research, this article maintains that IR would be enriched by the development of strategies of comparison that are unmoored from the assumptions of Humean causality, the logic of control, and generalisability in the neopositivist sense. After all, it is difficult to reject comparison outright because it is, at some level, engrained in human thinking: it is a “mode of cognition” underpinning figures of thought or speech such as metaphor, metonym, and allegory, and underpins self-other relations that are often premised on a reflection on the nature of similarity and difference (Stanford Friedman, 2011, 756).

The article has thus first of all examined how comparative methods have been mainstreamed in IR, and reviewed alternative frameworks of comparison and debates that have taken place outside the field of IR as well. Two issues in particular stood out. First, debates in other social sciences demonstrate that it is not necessary to concede to any one specific epistemology the power to define comparative research based on their own standards. Second, while the article has found greater diversity in comparative methods than commonly portrayed by methodological texts in Politics and IR, it has also underscored the need for an explicitly reflexivist engagement with the question of comparison. The article thus calls for IR to open itself to broader methodological influences in its rethinking of comparison as a concept and research practice, and shows how reflexivity can provide not only the theoretical background but also the practical guidance necessary to practice comparative research differently.

The key argument I have advanced in this article is that IR could greatly benefit from the explicit development of comparative methodological strategies that differ from the mainstream Milliam framework; and that are more reflexive in nature and thus appeal to a growing crowd of IR scholars whose work is inspired by reflexivity. Reflexivist comparisons are historically informed and situated rather than relying on abstracted frameworks. They question established frameworks of comparison, including through a contrapuntal approach. They approach comparison relationally and through juxtaposition rather than seeing cases as bounded units, and make questions about commensurability and incommensurability explicit in the research process. Reflexivist research accepts and makes use of the constructed nature of comparison, including in its discursive dimension. Lastly, reflexivity also entails a commitment to challenging conditions of inequality that structure the research process itself, and thus leads us to rethink our routes toward knowledge contribution/production alongside and with the research communities we inhabit.

The three strategies I propose in this article––defamiliarizing discursive comparison, contrapuntal comparison and vernacular comparison––respond to clear needs in the IR literature for methodologies aimed at recovering histories and perspectives that have been marginalized in IR (Agathangelou and Ling 2004; Chowdhry 2007). While not all reflexivist research projects will require a comparative methodology, those who do might find one of these three strategies applicable or adaptable to their research design. Far from wanting to complicate and thus foreclose the use of comparison beyond mainstream IR, the goal of this article is to open up its use and make it easier to practice reflexivity while doing comparative research. The purpose of the article is also not to make comparison an exclusively theoretical or abstract exercise. If anything, the principles and strategies outlined above, and reflexivity itself as an approach, make it impossible to divorce theoretical reflection from historical context or ideas from the material conditions that gave rise to them and structure them. Overall, the development of these comparative principles and strategies makes an important contribution to literature on IR methodologies, responding to the call to make reflexivity not just part of our theoretical commitments but something we consistently practice through our research choices and in our teaching.

Acknowledgement

I would like to thank the reviewers and editors for their helpful feedback and constructive engagement with the manuscript, as well as Adam Lerner, Roberto Roccu and Laura Sjoberg for feedback on earlier drafts.

Footnotes

References