Regardless of approach and whether it is explicitly acknowledged as such, time and temporality play critical roles throughout all areas of the discipline of International Relations (IR). This is both because of the critical, constitutive aspect of time and temporality in international political practice, but also because IR is, as Felix Berenskoetter has argued, a ‘future oriented enterprise’: the study of international politics attempts to explain and understand past events largely to better predict and/or approach the future (Reference Berenskoetter2011). In international security, for instance, debates about history and its usage are vibrant, productive and ongoing, but issues of time and temporality are typically treated as a subset of these debates rather than its own independent issue. Yet, time and its representation are inescapable whenever one engages the past to better understand the future. Scholars, whether they acknowledge it or not, are always already addressing issues of time and its representation. While the actors (as well as structures) at work are indelibly marked by time, they simultaneously operate with their own representations, understandings, and experiences of time. These representations can be distinct or overlapping – individual, shared, or something else entirely – but they form a constitutive part of political life. How IR understands and represents time inextricably shapes the knowledge claims it produces.
This article seeks to further understand how IR approaches issues of time and temporality, to better identify the particular representation and understanding of time employed and to show how it shapes and constrains scholarship. What can we learn by placing temporality front and center rather than evaluating it as an aspect of methodological and epistemological debates about the proper use of history? (Alker Reference Alker1984; Fioretos Reference Fioretos2011). I argue a deeper investigation and appreciation of the way in which temporal assumptions shape IR theory would better enable scholars to explain and understand important events in IR and the relationships that exist between international actors. In particular, attention to issues of scope, temporal context, contingency, and event-based temporalities can allow IR to better approach the processes at the heart of international political practice, the nature of ongoing relationships, and the way important events have systemic effects. Awareness of the constitutive role of temporality can produce important insights into existing epistemological commitments about what constitutes ‘good’ theory, which would, importantly, result in a heightened disciplinary focus on temporality as its own area of inquiry.
Issues of temporality have been engaged primarily through debates about the proper utilization of history and historical methods. This article argues the implications of how we consider the past and its relationship to the present and future has broad implications well beyond the question of whether and how one chooses to employ historical methods. To do this, I turn to the treatment of temporality within sociology and social theory to develop insights that have more widespread applicability (Alker Reference Alker1984; Lustick Reference Lustick1996; Kratochwil Reference Kratochwil2006; Fioretos Reference Fioretos2011; Patomaki Reference Park2011). Recent moves such as the so-called ‘practice turn’ and the emphasis toward reflexive scholarship in IR borrow directly from sociological approaches and interrogate what it means to conduct social inquiry and this article argues questions of temporality in IR will benefit from a similar engagement (Pouliot Reference Peters, Pierre and King2008; Bauer et al. Reference Bauer2009; Friedrichs and Kratochwil Reference Friedrichs and Kratochwil2009; Adler and Pouliot Reference Adler and Pouliot2011; Bigo Reference Bigo2011; Madsen Reference Madsen2011; Amoureux and Steele Reference Amoureux and Steele2015).
To be sure, temporal issues in their own right have become an increasingly vibrant area of study in IR. Nevertheless, these critiques have been mainly substantive examinations of political practice and remain relatively insular to this point (Hutchings Reference Jackson2008; Hom and Steele Reference Hom and Steele2010; Solomon Reference Snyder2013). This article argues that each set of approaches in IR – be they statistical, critical, rational, or otherwise – resolve issues of time and temporality largely for themselves rather than treating it as a separate issue that affects IR scholarship as a whole. Work in comparative politics, for example, explores the relationship of temporal assumptions to the causal mechanisms of political processes (see, e.g., Gryzmala-Busse Reference Grzymala-Busse2011). As well, there have been self-conscious efforts to incorporate particular aspects of temporal fluidity and structural change into statistical analysis utilizing Bayesian approaches. While these approaches are important to build upon, they remain narrow in their implications and consequently are not yet part of a broader debate on temporality’s role in the discipline (see Western and Kleykamp Reference Western and Kleykamp2004; Park 2010). When ideas about temporality have been brought directly into IR, the focus has typically been on theorizing specific institutional development. This article broadens this idea to foundational concepts such as war and security because they are no less constituted by representations of time and the temporality/temporalities of political life (Fioretos Reference Fioretos2011).
This article ultimately argues for conceptualizing temporality – both the temporality/temporalities intrinsic to international politics as well as the temporal assumptions informing scholarly knowledge production – as a stand-alone issue. This would do two things. First, it de-naturalizes the common representation of time as linear, neutral, and unitary, and articulates how conceptions of time are systems of meaning that are contingent, complex, and constructed in practice. Second, it better enables intellectual space for conversation across these traditions to develop a more robust understanding of the ways in which temporality shapes the study and practice of IR. All scholars, regardless of method or question, inevitably address time in their work and thus are implicitly offering a conception of and approach to time. Regardless of how one resolves the question ‘what is time?’ as scholars of political and social experience we are explaining and predicting actors who operate with contextual understandings of it that shape, constrain, and inform their experience and actions. Scholars using approaches as divergent as Bayesian analysis, Deleuzian psychoanalysis, computational approaches, and post-structural theory have all sought to recognize the complexity of time and temporal experience (see, e.g., Buthe Reference Buthe2002; Hutchings Reference Jackson2008; Lundborg Reference Lundborg2012). Much of this conversation remains intellectually segregated, however, and this article argues that conceptualizing temporality as its own issue provides a language for scholars in disparate traditions to jointly inform and engage each other. In particular, this would expose how the dominant conception of temporality based on clock-time itself warrants investigation and interrogation, rather than simply finding ways to include more ‘time’ into scholarly work. By directly enabling a more robust appreciation for the role of time and temporality in scholarship and political practice, I argue this creates a space for epistemological openness that fosters methodological and intellectual pluralism. Time itself may be beyond the scope of IR’s field of inquiry, but the temporalities informing both actors and scholars is not.Footnote 1
This article will be structured around three questions: what does it mean to critically engage representations of time?; what are the problems this poses for IR scholarship?; and finally, what could we do differently if we fully accept the implications of temporality? In the first section, I identify existing critiques of the particular formulation of time IR shares with other social sciences developed by theorists such as Barbara Adam, Pierre Bourdieu, and Andrew Abbott. These works show how time is generally understood and represented. The second section applies these ideas to show how this understanding limits scholarship. Focusing on IR’s representation of time, I show how it shapes and constrains the scope and effectiveness of theory, complicates predictions and privileges certain questions over others. In the final section, I outline some of the issues that become explicit when utilizing an approach that takes temporality seriously, such as scope, contingency, temporal context, and event-based temporalities. This final section will demonstrate that some of the problems we as scholars encounter may not be inevitable consequences of the difficulty of theorizing complex international events, but arise from the temporal assumptions embedded in our work.
Time, temporality, and social science
Given the inextricable relevance of time to social and political life, as well as the temporal issues foundational in constructing and articulating theoretical claims about international politics, one might assume that temporality already occupies a prominent position in IR. If trade flows, processes of globalization, armed conflict and war are all issues that only come into meaning across time, then one can reasonably argue there should already be an explicit focus on the issue within IR.Footnote 2
The argument I develop in this section for foregrounding temporality and more fully temporalizing IR theory builds upon two separate sets of investigations and debates about temporality in IR, as well as in political science/sociology more broadly. Part of this article’s goal is to bring together these disparate strands of scholarship to show how temporal concerns exist throughout the discipline, and regardless of approach or area of inquiry, IR scholarship as a whole can benefit from treating temporal concerns more carefully and self-consciously. First, the case I make here for treating temporality as a stand-alone issue builds on arguments that are self-conceived as operating within positivist traditions. Debates about causality and historical methods in comparative politics and IR emerged most recently with Timothy Buthe’s call to ‘Take Temporality Seriously’ which argued for the value of an approach that treated time in a more complex and nuanced manner (Reference Buthe2002). While acknowledging its intellectual importance, critics argued Buthe’s approach would be too complex or contextually inaccurate to incorporate into existing IR debates. For example, Scott de Marchi, Tulia Falleti, and Julia Lynch questioned the workability of an emphasis on time and temporality. Anna Gryzmala-Busse’s work on time and causality took this discussion a step further by offering a set of methods in response to these critiques that demonstrated specifically how scholars could more effectively model and predict events based on a more complex understanding of temporality’s role in politics. Others argued extant techniques either already incorporated time into their work or at least retained the potential to do so if it were necessary. Most notably, Jong Hee Park and Bruce Western argued that statistical methods such as Bayesian analyses of ‘changepoints’ already can (and do) fully incorporate time although they admittedly leave the representation and conception of ‘time’ at work uninterrogated.Footnote 3
Second, my argument builds upon a critical tradition of IR that argues contingency, context, and indeterminacy are inevitable parts of political and social life. In these approaches, time – just as other important concepts like sovereignty and the state – is simply a set of intersubjective, shared meanings and therefore, it is argued, our attention should focus on the processes of meaning production and social construction that shape our understandings of it (Hutchings Reference Jackson2008; Hom Reference Hom and Steele2010; Lundborg Reference Lundborg2012; Solomon Reference Snyder2013). Hutchings’ work in particular reminds us that ‘time, as a (intersubjective, public) category through which our experience of, and action in, the world is organized, has complex and multiple meanings’ (Reference Jackson2008, 5). World politics has its own concept of time, she argues, and ‘shapes narratives of political time,’ the very same narratives that IR scholars utilize in their investigations of international politics (Hutchings Reference Jackson2008, 5). In what follows, I bring these two strands of thinking together to develop a critique of the predominant approaches to understanding time currently operating within IR and illustrate the value of thinking critically about time.
Representing time in social scientific inquiry
Burgeoning scholarship on temporality in IR – particularly from this second, critically informed area of thought – has begun to denaturalize the common representation of time that is based on the classical conception of time as clock-time (Hutchings Reference Jackson2008; Hom Reference Hom and Steele2010; Hom and Steele Reference Hom and Steele2010; Patomaki Reference Park2011; Solomon Reference Snyder2013). Hutchings’ work on world-time and the distinction between chronos and kairos in temporal orderings of world politics, Hom’s tracing of the role of time in the production of sovereignty, and Solomon’s analysis of the temporal aspects of subjectivity in particular show how temporal understandings and assumptions are embedded in foundational concepts utilized in IR and international security (Hutchings Reference Jackson2008; Hom Reference Hom and Steele2010; Solomon Reference Snyder2013). Just as IR has engaged in the reflexive analysis of what were once commonly understood and uncontested ideas like the state, security, and power – these works on temporality have similarly begun to show how conceptions of temporality that arise from dominant understandings of time are also products of social process and/or political practice. This is not to argue that classical understandings of time are meaningless, but rather to highlight that a particular conception of temporality, both for the scholar and the actor(s) under investigation, is hegemonically reproduced whenever the conception of ‘time’ is utilized in a way such that its meaning is ‘self-evident’ (Bourdieu Reference Bourdieu1977).
IR primarily employs the conception of clock-time that animates traditional scientific (and social scientific) inquiry, a conception of time that sees it as ‘present everywhere, the same everywhere, independent of anything we do. It carries no descriptive label and has no need to advertise or to repudiate that label. When seen as this uniform background, time is quantifiable. Its measurable segments are exactly the same length, one segment coming after another in a single direction’ (Dimock Reference Dimock2002, 911). This classical, scientific understanding of time derives primarily from classical physics and Newtonian mechanics and continues to remain influential both in terms of our ‘common-sense understanding of the world’ and ‘our assumptions of social scientists’ (Adam Reference Adam1990, 48). Time is represented by the clock: measurable, natural, and independent of human experience. This metaphorical representation of time as ‘clock-time,’ Adam argues, ‘incorporates recurring cycles as well as the linear, unidirectional flow of time; duration as well as instants; and lastly, a spatiotemporal representation of time’ (Reference Adam1990, 54).
Clock-time imagines not only a spatiotemporal representation of time as a linear flow occurring at a singular rate, but also views time as ‘unitary’ and ‘neutral,’ which implicates assumptions at work in IR (as a social science) in a way that devalues the importance of temporal context. Representing time as unitary refers to understanding it as possessing a singularity that places it beyond the realm of debate or discourse. The names attached to a section of time or particular points on the timeline can be debated – a particular point can be referred to as ‘February,’ ‘Winter,’ or ‘second semester’ – but the actual time that attaches to it cannot. Even if different societies or analysts perceive temporal context differently, a hegemonic understanding of time still structures the relations of those perceived temporal contexts where natural time remains ontologically privileged. According to this view, it still ‘really is’ 2015, regardless of what anyone chooses to call it.Footnote 4
The classical view of time also holds that time is neutral and possesses no independent explanatory power on its own. Because of this neutrality, theoretical pronouncements can be time-invariant, just as they are spatially universal and potentially apply to any location in the international sphere. Theoretically speaking, then, it is logical to conclude that an identical event can occur two days, two years, or even two centuries apart. This is because there is nothing intrinsic to the passing of time that prevents the possibility of events occurring exactly as they happened before. If it were possible to derive an actual law that arms races result in interstate conflict, for example, then that law would extend into the indefinite future (and by extension, the indefinite past), much as the law of gravity holds across all time and space. Context may matter, but temporal context does not.
The underlying conception of time at work here is important not insofar as whether it accurately depicts time ‘as it really is,’ but because this understanding of time shapes the temporal assumptions of international events in practice and the theoretical claims and predictions scholars make. This classical view of time encourages a view of reality that strives for laws, theories, and hypotheses that apply across time and are generalizable. If time is linear, neutral, and unitary, then laws and theories should govern behavior regardless of what point on the timeline the law is applied to. Theory should explain motion regardless of time – it should almost literally be time-less.
Social science’s a-temporal view of reality
Given the stated desire of many social scientists (especially post-behavioral turn) to explain social reality in an ‘objective’ fashion, this uncritical acceptance of classical time is both logical and necessary. Bourdieu’s concept of practice has become increasingly popular within IR overall at least partially as a way of engaging the empirical without uncritically accepting objectivist social science. The concept of practice also places special emphasis on the contextual position of agents (Adler and Pouliot Reference Adler and Pouliot2011). Bourdieu’s critique of ‘objective’ social science argues that the scholarly move to theorize social reality replaces the actor’s conceptions of time and temporality with the time of the analyst. This reification of the analyst’s temporal understanding is a de-temporalizing move:
Science has a time which is not that of practice. For the analyst, time no longer counts: not only because – as has often been repeated since Max Weber – arriving post festum, he cannot be in any uncertainty as to what may happen, but also because he has the time to totalize, i.e to overcome the effects of time. Scientific practice is so ‘detemporalized’ that it tends to exclude even the idea of what it excludes: because science is possible only in a relation to time which opposed to that of practice, it tends to ignore time and, in doing so, to reify practices … The detemporalizing effect (visible in the synoptic apprehension that diagrams make possible) that science produces when it forgets the transformation it imposes on practices inscribed in the current of time, i.e. detotalized, simply by totalizing them, is never more pernicious than when exerted on practices defined by the fact that their temporal structure, direction, and rhythm are constitutive of their meaning (Bourdieu Reference Bourdieu1977, 9).
Bourdieu’s claim about the constitutive importance of temporality is especially important in the context of IR. Wars, battles, trade flows, international migration and the like are all constituted by actors operating within temporal spheres that may be shared, overlapping, or even universal, but are unlikely to be shared by the observer/theorist. Modeling behavior across time – natural time – requires a model that applies regardless of temporal context. Temporal context, however, is not simply a set of circumstances to be captured by variables, but constitutive of an actor’s experience and behavior. Divorcing it removes an inextricable element of behavioral choice in practice because temporality shapes the very meaning and idea of actor’s actions in the moment. Factors like how far back in history scholars look when searching for parallel examples to the issue at hand, whether they believe they are in a time of ‘crisis,’ and how far in the future they think/care about all matter deeply for actors deciding in the moment. Without deliberate analysis on the issue, it is unlikely these assumptions will automatically match up with the ex post facto assumptions of scholars.
Echoing the critique of objectivist knowledge that Bourdieu advances, Adam argues, ‘sociality then is located in that process of adjustment and not in its result, in the being part of the old and the new at the same time. Sociality could therefore be described as the dynamic meeting, the interpenetration of continuity and change, of conservation and revolution’ (Adam Reference Adam1990, 40; Patomaki Reference Park2011). For the actor acting ‘in the moment’ time is an inescapable element of their reality, ‘as it unfolds, time is a vital part of the social game, and the urgency of events dictates not just what people do, but the timing of their actions. There is an antinomy between the time of the leisured observer and of the pressured social actor, the rhythm and tempo of the action being an integral part of its meaning’ (Gosden Reference Gosden1994, 121). This de-temporalization operates at a deep level, where ‘the very instruments of communication – writing and speech – impose a linear sequence on events which may originally have been more confused and layered’ (Gosden Reference Gosden1994). Kratochwil highlights this problem when discussing IR’s relationship with history, arguing, ‘what is often forgotten is that history is always remembered from a certain situation in the present. For while things past now have relevance … in this sense history is an encounter with the self rather than simply a storehouse of data’ (Reference Kratochwil2006, 15). This last point bears underlining, as the social and political events that constitute international politics operate at the dynamic intersection of persons, agents, institutions, and structures. Temporality and temporal dynamics are intrinsic to their meaning, both in the moment and after the fact. Context and historical texture is important to be sure, but temporal conceptions operate at a level beyond simple contextualization both for the scholar and the actors under investigation.
The process of analytical representation itself also mediates the experience of the actor in a way that can undermine the analyst’s ability to access – that is, to represent – experience in the same way as the agent who falls under their gaze. One can utilize future events to infer causality, purpose and a particular meaning to actions that may have been much more complicated in the moment than when juxtaposed with history’s understanding – think of the inter-war period in Europe, actors acting during what would become known as the ‘Cuban Missile Crisis’ or the actions of a state beginning a war the analyst knows will last decades but those acting in the moment do not. Put simply, an event’s meaning for the scholar often relies on its relationship to events past and future to which the actors who constitute the event itself do not, and in some cases cannot, have access. Attention to the way in which actors act in time – the way in which they temporalize their actions and behavior – is important to attend to in order to better approach future events.
Security studies as social science
Security studies scholarship is self-consciously methodologically pluralistic and so there is no defined, single rulebook for what constitutes ‘good scholarship,’ nor is there a page to turn to identify how to treat time and what assumptions should be made about temporal dynamics. The resolution of these issues is mostly implicit and decided by scholars themselves. Even so, in practice there are some shared understandings at work that occur (mostly) throughout the discipline even if their meaning is not discrete, immediately approachable or something to which all scholars subscribe.
One of the central aspects of this scholarship – as Krause and Williams argue, one of its ‘dominant conceptions’ – involves Walt’s notion that security studies is about the generation of ‘cumulative knowledge of the use of force’ where ‘the field must follow the standard canons of scientific research’ (Walt Reference Walt1991, 222; Krause and Williams Reference Krause and Williams1997, 37):
The history of security studies follows the model of a particular understanding of the growth of scientific knowledge: one of a linear progression through time that ultimately yields a form of knowledge beyond time and history. While progress toward knowledge may be frustratingly variable, the goal to which it aspires is not. To understand, in this view, is to work within the strictures of a particular conception of science and knowledge: the search for timeless, objective, causal laws that govern human phenomena (Walt Reference Walt1991, 222).
One of the crucial points of distinction in security studies is precisely this view of science as a search for ‘timeless’ and ‘objective’ laws, as it reveals the assumptions at work throughout IR.
Krause and Williams would be the first to admit that not all security scholars share these commitments – they are articulating a ‘dominant’ vision precisely so they can more effectively examine alternative, critical approaches to security – but even they concede that the conception of security studies as searching for ‘timeless’ and ‘objective laws’ is ‘dominant.’ Although critical theorists and non-positivists notably disagree regarding achievability, for scholars not working in those traditions many would agree that searching for time-less laws is an ideal to be strived for, even if ultimately impossible. Patrick Jackson observes, ‘testable hypotheses and general claims are thus portrayed as almost unquestionable goals of IR scholarship’ and are so unquestionable that they ‘hardly even’ need ‘the label of “science” to distinguish them from alternatives’ (Reference Jackson and Nexon2010, 7). Regardless of whether we can precisely measure just how ‘dominant’ these assumptions are in practice, these ideas form central concepts and commitments around which the scholar must position themselves and as Jackson observes, this frequently turns on the question of whether a claim is ‘scientific’ and ‘verifiable’ in a ‘disciplining’ manner (Reference Jackson and Nexon2010, 9–10). One of the important functions of this particular valuing of ‘time-less theory’ operates at the level of epistemological commitments regarding what constitutes ‘good theory’ or ‘good scholarship.’
These ideas play out in each of the major schools of thought. In (American) security studies, realism remains highly influential, if not dominant. One observer notes that ‘liberals are more careful about temporal sequence of cases than are realists’ and his claim provides some insight into the structure of the discipline itself (Cederman Reference Cederman2001, 18–19). Generally speaking, realism places special emphasis on structural explanations and resists agent-level theorizations. Waltz’ and Mearsheimer’s variants, for instance, explicitly value macro-level explanations that operate independent of historical and/or temporal context – a point on which international historians have critiqued them (Waltz 1979; Schroeder Reference Sambanis1994; Spiro Reference Solomon1994; Mearsheimer Reference Mearsheimer2001). Similarly, constructivists privilege structure in their theories of action, and how one resolves the ‘agent-structure’ debate still retains influence for those debating norms and identity (Finnemore Reference Finnemore1996; Wendt 1987, 1998, Reference Wendt1999). Liberals may indeed be relatively more temporally inclined because of their focus on institutions that appear more directly affected by issues like sequencing. That said, each school of thought, despite their differences, ultimately values time-less theory as an ideal and only when this is rendered impossible by the empirics do temporal concerns come back in (Fioretos Reference Fioretos2011).
To be sure, some IR scholars do deal with certain temporal dynamics in a more nuanced manner then depicted by Walt et al., particularly in the area of sequencing. Cederman’s work on learning and the democratic peace, for example, is quite careful about taking sequence into account (Reference Cederman2001)(Cederman 2002). Keck and Sikkink’s notion of norm cascades similarly accounts for sequence in norm development and diffusion (Reference Keck and Sikkink1998). Post-structural theorists explicitly commit to processual theorizing and historical-genealogical accounts in their understanding of the state and security (Campbell Reference Campbell1992; Weber Reference Weber1998). On the whole, however, these projects that incorporate temporal dynamics, contingency, and sequencing explicitly position themselves as offering accounts of the world that differ from – or are critical of – dominant understandings of security, norms, and the democratic peace. Equally important, they leave the dominant representation of time unquestioned, as well as the epistemological values based upon that representation.
Even when projects are attuned to certain temporal dynamics, however, they are rarely explicitly articulated as such, and consequently advancements in the way IR conceptualizes time and temporality overall have been slow to develop. The implications of this silence are broad. Without sufficient theorizing of temporal dynamics, for instance, moving from explanation to prediction and/or relating insights from the past to the present becomes increasingly complicated. Indeed, bringing ideas that resonate when engaged with the past – no matter how the past is constructed or understood – into the present and/or future is noble, important, and ultimately, difficult work. Temporality as a concept provides a shared language for thinking through the basic question of relating the past to the present and future which lies at the heart of much IR work.
Other negative implications arise from the current treatment of temporality. While ideas such as sequencing or duration may be emphasized and this is to the good, rarely are they linked back to the temporal assumptions that arise from the particular representation of time that undergirds theory production and evaluation. Even as ‘time’ is more fully incorporated, it remains unquestioned. Issues like epistemological values about generalizability, contingency in models or the determinacy of claims may be negotiated in practice, but the role understandings of time and temporality play in maintaining those commitments remain under-theorized because these are not understood as outgrowths of temporal assumptions. Foregrounding temporality would better enable scholars to theorize and produce knowledge in a manner that begins by privileging ideas like contingency and sequencing from the outset (as critical theorists have done), rather than bringing it in at the end when other approaches are less successful.
Overall, temporality plays a foundational role independent of how much ‘time’ is included because of the work it does in epistemological concepts such as the explicit valuing of theory that explains more of the past than less or the belief that a theory, even one attuned to sequence, for instance, should ideally apply regardless of temporal location. These complications are facets of a linked issue that arises from a particular representation of time – IR’s temporality – not simply from complexities that may or may not arise on a case-by-case basis. The way time is represented and the assumptions we make about temporal dynamics deeply affect the questions we ask, and how we understand context, make predictions, value competing explanations and approach international events.
Temporality in security studies: a critical engagement
Representations of time, temporal assumptions and clock-time are already topics of interest to certain scholars but to demonstrate temporality’s relevance across the discipline necessitates showing that it negatively affects predictive models and theory construction. Scholars are always admittedly engaged in a process of simplifying events; some form of reduction is an inevitable feature of scholarship that analyzes the international sphere. Thus, general critiques of social science applied to security studies have had to show why they are something other than an inevitable cost of doing business or risk being dismissed entirely (Mearsheimer Reference Mearsheimer1994–1995; Wendt Reference Wendt1995). Challenges from international historians to realist claims regarding the past have run headlong into this concern (Schroeder Reference Sambanis1994). Temporality and temporal concerns are different because the way in which temporal assumptions shape theory development and modeling acts as a constraint on the effectiveness of explaining and understanding the politics under investigation.
What’s lost in a-temporal scholarship?
Prediction plays a central role in security studies and as a result claims that explain behavior across time are valued based on how much of the past (and consequently, present and future) they explain. Models that employ cross-temporal comparison and possess widespread applicability regardless of time period are privileged over those that do not. A model that only explains a narrow slice of the past will almost always be de-emphasized when compared with one that possesses broad temporal applicability. This is not to say that narrow slices of the past are necessarily better than broader views, but rather the breadth with which we approach the past should be actively theorized in light of the particular project’s goals instead of assumed into existence by scholarly preference.
While scholars (for the most part) will admit that the natural science ideal of truly ‘timeless’ and universally valid theory may never be achieved, it still presents a limit case toward which theory builders orient themselves, their projects, and their evaluation of proposed theoretical models. If I advance a claim that democratic regimes seek to deescalate military crises and base that claim on a causal logic arguing the requirement for public support deemphasizes using force in a crisis, then that logic – along with its predictive potential – is evaluated by its ability to explain events past, present, and future. How does this hold in cases of border disputes? Can this explain US actions in World War I? World War II? What about the US occupation of Iraq? Ultimately scholars want to know whether the claim is robust enough to hold into the future and this turns on whether it holds across the past.
Epistemological commitments at work throughout this example highlight the value of cross-temporal validity and the ideal of time-invariant claims. The past, present, and future are linked in a way such that the theory that can explain the behavior of agents at the majority of points in the past is assumed to be most likely to explain the future (immediate or otherwise). This idea builds on the assumption that the future is simply the next point in a continuous flow of time. Barring an intervening variable or a story regarding institutional change, learning, path dependency, or some other process, the presumption is that the future will replicate the past, because there is nothing meaningful about the temporal location (the point in time) of the event. Undergirding this idea of past/present/future is an assumption that because time is neutral and linear, events and/or actions in the past are no different than those occurring in the present or future. They are simply located at a different point on the timeline. Identifying this assumption is important because it has implications for the development of theory and knowledge. Under the extant view, temporal location is a marker with no explanatory impact or relevance on its own. Theories build from an understanding of political space as one that is unaffected by temporal dynamics. While incorporating temporal dynamics does not make prediction impossible, it does emphasize the need to narrow focus and prediction parameters to those appropriate to the particular work in practice.
Important implications follow from this spatiotemporal representation. For example, the past becomes understood as a ‘time-less present’ (Barcham Reference Barcham2000, 139). Manuia Barcham characterizes this de-temporalizing move as one that:
leads to the necessarily synchronic prediction that bodies (be they concrete or abstract [emphasis added] singular or plural) exist in an ahistorical essentialism wherein reality is collapsed into a timeless present such that what is now is the same as what was, which in turn is the same as what will be … (Barcham Reference Barcham2000).
De-temporalizing events by treating time as empty privileges a viewpoint that sees the past, present, and future as equivalent. Present and future may be outside the bounds of study due to the status of the observer, but there is nothing inherently meaningful about either. They simply reside to the right of the observer on the timeline because past/present/future are all a continuation of the same. Treating time as ‘empty’ operates as an assumption rather than as a tested or proven idea, and part of the point of conceptualizing temporality as a stand-alone issue is to better enable discussion about whether this assumption is a good one.
Projects utilizing large-n methodologies, for example, categorize events as a collection of data points representing a particular international behavior across time. Models are tested against this mass of data so that scholars can develop a corpus of knowledge in an objective manner. Treating international behavior – wars, battles, treaties, negotiations – as data that can be located on points of a timeline and aggregated initially brackets motion, sequence, and structural dynamics, even if it eventually brings these issues back in. Whether an event occurred in 1815, 1915, or 2015, it can potentially be coded as an independently occurring event operating in empty time (and space). To go back to the democratic logic hypothetical, if there were 20 instances of war initiation from 1900 to 2015 and the public played a dampening role in 12 of these cases, a scholar can make a robust claim that public mediation plays a significant role. The order these cases occurred, may – but does not have to – hold relevance, because the author can always argue that in 60% of cases in the data universe the causal logic held. Regardless of whether the data tested are coded points in a database or case studies, at first, time falls out almost completely. The past begins as a terrain to be explored and theorized. Only later is it potentially understood as a dynamic process that is complex and potentially discontinuous, and even then only if the initial claim is challenged.
This hypothetical may appear to be a primitive and unfair representation of quantitative analysis, and substantively speaking it may be. Nonetheless, it illustrates baseline epistemological assumptions that remain in place so long as the representation of time in IR remains unchallenged. Large-n projects can and often do (to a certain extent) incorporate different types of temporal dynamics. Jong Hee Park’s work on diversionary conflicts, for example, illustrates how quantitative analysis can move toward addressing temporal complexity through his utilization of ‘changepoints’ in Bayesian statistical analysis to identify structural shifts that implicate the overall model’s claims – in his case, the question of Presidential uses of force as a diversionary tactic (2010). His model identifies 1942 as the critical point, which approximates the qualitative argument others make regarding the 7 December 1941 attack on Pearl Harbor as a tipping point for American uses of force. While Park’s work points in a valuable direction, it is also somewhat of an exception in explicitly theorizing temporal disjunctures via statistical investigation rather than inserting them ex post facto. Yet, while Park brings in one aspect of temporal complexity, it still leaves foundational assumptions about its representation of time unquestioned. Changepoints are an important and useful concept and unlikely to be the exception but the rule.Footnote 5 Greater awareness of temporality would better normalize concepts like this. Changepoint analysis, for example, could be better informed by directly engaging qualitative approaches, and vice versa. A shared language of temporality could better enable this cross-approach conversation.
The point here is not that large-n projects are always as simple as the hypothetical I offered, nor am I arguing that time is ignored altogether – objectivist scholars concern themselves with time as well. Acknowledging that timing matters, however, is not quite the same thing as fully accepting the implications of temporal theorizing or moving toward a greater investigation of temporality’s role in IR’s claims. It still leaves extant the dominant representation of time as natural, essential, and unitary, which has important epistemological and methodological implications. One specific implication is that it devalues including temporal dynamics that are based upon the practices of the agents and structures under investigation vs. those based on scholarly assumptions.
Objectivist scholars, as well as many others, are the first to admit that their models are simplistic representations of complex events, but all simplifications are not created equal. If temporality is an inextricable element of events like war or international trade flows, it is potentially counterproductive to first privilege models that lack temporal dynamics, only to bring them back in later. Temporally dynamic models are also currently de-emphasized because they are epistemologically evaluated from a perspective with particular temporal commitments that put more nuanced models at a disadvantage. Because of the value of parsimony, a complex model that incorporates temporal dynamics from the outset will not be preferred over a simpler model if they possess equal explanatory or predictive power. The danger is that this privileging of parsimony comes at the expense of seeking out temporally detailed and sophisticated explanations that could potentially be more useful when making the specific predictions that animate the field in the first place.
William Sewell’s articulation of the problem of ‘events’ for social science and history provides useful insight into some of the difficulties that can arise from over-emphasizing cross-temporal comparison. He argues events have the potential to disrupt the seemingly natural and linear flow of history that animates mainstream temporality because ‘events’ are not just occurrences or interactions among static entities that exist across time, but sites of discontinuity that potentially rupture social systems. This ‘rupture’ poses a potentially radical problem for the extant epistemological assumptions of the social scientist and cannot always be solved by simply adding temporal context after the fact.Footnote 6 If one posits inter-subjective understandings of social reality, then events can radically reshape the entities themselves, not merely their attributes. If subjects are subjects because of their relational properties, then events that alter these relational components also alter their very existence as subjects (Jarvis Reference Jervis2009; Solomon Reference Snyder2013). Models of behavior based on previous understandings of agent and structure might literally be comparing apples to oranges due to relational shifts that alter the makeup of these agents. Consequently, ‘events’ radically disrupt assumptions of cross-temporal continuity of agents upon which these analyses are built. As well, eventful temporalities may also complicate the causal relationships that analysts have posited/discovered (Sewell Reference Schroeder1996).
If one accepts the transformative potential of new events it also becomes apparent that there is a built in theoretical time lag produced by its temporal commitments. Theory becomes intrinsically conservative because new events (occurrences/emergent agents/etc.) are inevitably epistemologically devalued. If time is linear, neutral, and spatiotemporally empty, then events that occur near the line of the present can almost never truly disrupt a robust finding. This is true almost regardless of how transformative an event is because a new event can only constitute a comparatively small slice of the existing timeline, and therefore never displaces a robust finding until sufficient time has passed such that history has shifted the balance in favor of the new reality. If Sewell’s view of events as ‘sequences of ruptures that effect transformations of structure’ possesses merit, then this epistemological commitment disarms the ability to theorize the discontinuous ruptures that are usually most in need of academic inquiry, such as the end of the Cold War, the French Revolution, outbreaks of great power war or massive ethnic conflicts. Hutchings argues, ‘rather than time being understood as distinct from the events that unfold within it, it is the ongoing emergence of events that is equivalent to time … the future is always ambiguous and so, in another sense, is the past’ where it is ‘bound up with an onward continuous movement in which the significance of past events shifts according to an unpredictable pattern of actualization – in this sense the past is being re-written constantly in the present and future’ (Reference Jackson2008, 57). This is not to say that every time there is a shift in IR one must throw out existing theories. Rather, the current manner in which temporality intersects with epistemological commitments in IR shifts the balance too far in favor of established theories. It maintains a rigidity in approach that makes these crises and discontinuities especially difficult for scholars to adapt extant theories. It also does not fully recognize the implications of the constant ‘re-writing’ of the past that is intrinsic to time because even those political structures that seem most rigid are inherently contingent, or in Hutchings words, ‘the time of politics is associated with indeterminism’ (Reference Jackson2008, 54). Of course, not all structures are prone to radical change at every moment, nor do all structural changes necessitate new theoretical claims. However, it does mean scholars should be more attentive to the potential inapplicability of theoretical claims based on new changes (structural or otherwise).Footnote 7
The understanding of events provides perhaps the clearest example where temporality shapes and constrains IR work. War and armed conflicts are a foundational aspect of security scholarship, but at its core, war is an event, and therefore, the way we understand events like war have a profound effect on the discipline as a whole. Put simply, IR is fundamentally about relationships, which are an ongoing series of interactions understood and interpreted across time.
The concept of a war itself provides a good example of how foregrounding temporality can inform IR scholarship on the nature of events in particular. David Weberman demonstrates the ‘non-fixity’ of the past by showing how past events are defined by their relationship to the future and therefore as the future changes, the reality of that past event alters as well (Reference Weberman1997). Much like Wendt’s argument that IR drives by ‘looking in the rearview mirror’ Weberman goes beyond observing that present understandings of the past change, to argue that the past itself changes as the future unfolds (Reference Wendt2001). While this runs against the intuitive idea that the past is done and cannot be undone, the finer one’s analytical focus gets the more the fluidity of events – especially their beginning and end – becomes apparent. Weberman highlights Yitzhak Rabin’s assassination as example. At ~10 a.m., Rabin is shot, but does not actually die for another 3 hours. At what point would we say that Rabin is assassinated? And how would we characterize the ‘in-between’ hours if the event itself does not even become an assassination until enough time has passed for Rabin to perish from his wounds?Footnote 8
Answering the question may appear relevant only metaphysically. The important ‘causal logic;’ involved appears clear – Amir shot Rabin for ostensibly political reasons and that shooting was fatal. Most events in IR, however, do not follow so obvious a causal logic. Take for instance, identifying the cause of a war. If one were to evaluate the role of ethnic motivations in explaining the onset of a civil war it would require assessing the motivations for a conflict alongside its trajectory into war (Fearon and Laitin Reference Fearon and Laitin2003). The common understanding (coding) of civil wars relies on a metric stating a conflict becomes a civil war only after there have been more than 1000 deaths in total with greater than 100 casualties on each side in order to distinguish war from genocide or massacre (Harff Reference Hom2003; Sambanis 2004a, Reference Sambanisb). This approach is similar to Weberman’s because it is explicitly backward looking. Just as is the case with the assassination example, even once the shooting has begun, future events – the 1000th battle death – determine whether it becomes a civil war.Footnote 9
If an inextricable element of the logic of causality relies on the idea that causes must precede effects, then how one codes the onset of civil war will have a significant impact on what one views as its ‘motivation’ or proximate causes. To continue the example, one could code the point at which the 1000th death occurs and then identify the origin of that campaign by identifying the first politically motivated attack that resulted in death. If political motivations are present at this point, but not ethnic ones, then this model could conclude that ethnic motivations are irrelevant. However, what if political unrest is the language that these insurgent groups cloak themselves in to create a political agenda sufficient for initial group formation, and then widen their agenda to include ethnic concerns for recruitment and retention? If the other group counters with ethnic mobilization, then it is possible that ethnic motivations did not exist at the ‘onset’ of civil war, but may have been instrumental to its escalation.
While some might argue that this is just an intervening variable, things become more complicated if one recognizes how the future affects the existence of a civil war. Rabin’s cause of death appears relatively clearly linked to the shooting – but for the action of Amir, he would have lived. The same argument, however, could apply for the ‘intervening variable,’ but for the move to mobilize along ethnic lines, there would not have been 1000 deaths and thus no civil war. In this case would we say that ethnic motivations did or did not ‘cause’ the civil war? Without ethnicity the initial conflict might not have escalated to the point where it became a civil war, but if it was not the reason the insurgents fired the first shot of the ‘war’ then it seems difficult to assert that ethnicity ‘caused’ the conflict.
The implications of this example go well beyond the issue of civil wars. State–state wars, insurgencies, and armed conflict in general similarly have elements in process that constitute their existence, and therefore have the potential to implicate what we say caused or did not cause an event. Wars are a particular example of this, but are by no means the only one. Most aspects of IR that scholars seek to explain and understand are constituted by process, be they states, institutions, armed conflict, terrorism, or economic trade and investment. Conflict and cooperation – the focus of much scholarship – are, at root, relationships across time. Efforts made to better understand how temporality shapes and constrains our current understandings may indeed complicate current IR work, but it will also open up new areas of investigation that potentially more effectively capture the inextricable dynamics intrinsic to these relationships.
The inevitable contingency surrounding whether a political event even becomes understood as a meaningful event in the future – as is the case with all social events – demonstrates one of the many ways in which temporality is not simply a marker or idea external to an event and actor’s behaviors, but intrinsic to the social and political event itself. This potentiality demonstrates why temporality matters, but perhaps more importantly, it also illustrates how the non-fixity of the past implicates research for even the most self-professedly ‘objective’ of methods. This example demonstrates how the particular temporal environment of the areas we inquire into could be at odds with the temporal environment of the observer, but also that that difference matters. As such, it illustrates the implications of Hutchings’ understanding of temporality as ‘unfolding,’ and in particular, the importance of the non-homogeneity of time, where it is ‘constantly differentiating with multiple possibilities within it’ (Reference Jackson2008, 57).
Time and (the) discipline: international relations
Temporal assumptions based on the common, shared understanding of time complicate the theoretical analysis of international events in discernible ways. IR, though, is not the only discipline to employ this particular representation of time. Sociology has engaged in similar practices and this section utilizes Andrew Abbott’s analysis of representations of time in sociology as a heuristic for IR. I turn to Abbott’s work both as an example of how to treat temporality as its own issue, rather than as a detail embedded within other substantive or methodological debates, but also because his work operates as a lens for examining and debating the effects of particular understandings of time on scholarly work.
Abbott offers an important sociological perspective on temporality that differs from the usual historiographical approach (Schroeder Reference Sambanis1994; Spiro Reference Solomon1994; Lustick Reference Lustick1996; Kratochwil Reference Kratochwil2006; Fioretos Reference Fioretos2011; Patomaki Reference Park2011). While temporality and temporal assumptions cannot easily be disaggregated from issues of history and historiography, I argue here that attending to the temporal dimension on its own remains important. Leaving it as one facet of debates about history or statistical techniques misses the foundational role temporality plays for all scholarship, regardless of methodology. Assumptions regarding time and its representation operate at a level beyond history and historiography, and Abbott’s perspective is particularly valuable in showing how temporality operates both above and below historical context when investigating socio-political phenomena.
Utilizing this sociological perspective does not, however, obviate the need for these debates nor articulate the issues as wholly separable. While the way we approach history shapes our understanding of temporality, the way we approach time also shapes our understanding of history. A temporality based on common representations of time as neutral and unitary encourages a view of history as a repository of data to be mined for discrete and objective data points. This view has been ably criticized as counterproductive: ‘scholarship that regards history as a tool to be bent to the will of theory would constitute a creaking orthodoxy’ (Hall and Kratchowil 1993, 419). Lustick reminds us that scholars ‘must recognize background narratives are constructed, not discovered,’ highlighting the sociality inherent in any reading of history and the indeterminacy of so-called ‘facts.’ Much of the criticism of treating history as a set of extant data to be explored reflects Carl Becker’s foundational work questioning the meaning and existence of a ‘historical fact’ (Lustick Reference Lustick1996, 613). One example Becker himself employs is the ‘simple fact’ of Caesar crossing the Rubicon, stating, ‘the simple fact turns out not to be a simple fact at all. It is the statement that is simple – a simple generalization of a thousand and one facts’ (Becker Reference Becker1955, 329). His example is a particularly important one for IR generally and security studies in particular. Think of the statement, ‘Country X won a battle against Country Y during war Z on day ABC’ – each day, each action, and each historical statement, when applied to an undertaking reliant on the interactions of so many individuals, institutions, and actors is composed of a minimum of thousands of such ‘historical facts.’ Foregrounding temporality provides an important addition to these reflexive debates on history in IR because it reverses the arrows (see, e.g., Schroeder Reference Sambanis1994; Berger Reference Berger1997; Williams Reference Williams2005; Bennett and Elman Reference Bennett and Elman2006; Fioretos Reference Fioretos2011). It is not just how we approach history that shapes our temporality; the particular understanding of history as objective data is supported and encouraged by temporal assumptions entirely separate from the choices we make regarding historical methodology.
Abbott, time, and the practice of social science
Andrew Abbott’s analysis of temporality in sociology shows how his discipline’s temporal assumptions – assumptions IR largely shares – manifest in a worldview he identifies as ‘General Linear Reality’ (GLR). GLR is a useful heuristic for thinking about IR’s treatment of time and temporality. Abbott’s critique and identification of GLR in some ways tracks along traditional critiques of positivist approaches, but issues of temporality operate at a level separate from debates about positivistic approaches for three reasons. First, the argument advanced here is that the predominant temporality in IR is one that treats time itself in a positivistic manner. Time is treated in an ‘objective’ way even though the meaning and experience of time is shared, socially constructed, and fluid. Second, much of the debate about temporality has revolved around the uneasy relationship between science and history. Yet, questions abound about how to link the two, whether there is a science/history binary, and if so, which side of the divide should be privileged. Foregrounding temporality as a stand-alone issue could help displace that binary by encouraging a view where issues of time and temporality are treated as part and parcel of scholarly inquiry rather than forced back into questions on the scientific use of history or vice versa.Footnote 10 By reflexively engaging these commitments as issues of temporality rather than only as subsets of larger debates IR can more easily assess and alter the temporal commitments at work and the implications they have substantively, methodologically, and epistemologically. Finally, and perhaps most importantly, the issue of temporality can operate as a point of pluralistic inquiry – all scholars engage in some sort of temporal theorizing regardless of whether it is explicit or implicit. Conceptualizing temporality as an explicit set of concerns rather than as something only certain approaches value creates an epistemological space available to scholars regardless of their position on positivist approaches.
Bringing in insights from a sociological perspective also opens space for different insights into extant work. For instance, if a gap exists in our understanding of a particular relationship – for example, between groups involved in an insurgency and terrorist attacks – the gap may not be the product of poor scholarship or a lack of imagination, but rather may arise from epistemological assumptions that inform our work – assumptions that can be interrogated, questioned, and modified. If a category of analysis used by scholars has difficulty matching up to its practice in the present – such as war, insurgency, terrorism, or other uses of force – this may not be a function of poor scholarly decisions, but a reflection of the broader political fields that scholars seek to understand.Footnote 11
Representations of concepts like war or terrorism are contested and powerful, and shaped by contemporary discourses in the present. At the same time, it is equally important to realize that disjunctures between categories of analysis and contemporary practice may be inevitable. Observing them is not an indication that a particular category cannot be compared across time, but just that when it is scholars must be attuned to the question of why certain comparisons resonate and others do not. Clausewitz’ personal notion of war may be vastly different than contemporary policymakers, for instance, but it also could be functionally identical, given the way collective memory and interpretation operate in the temporal imaginary of contemporary practitioners of military strategy. Temporal awareness does not preclude cross-temporal comparison, but asking why certain issues are comparable despite centuries passing leaves us better equipped to understand and employ them effectively, especially when we turn to making claims about the future.
Three aspects of GLR are especially resonant when applied to IR. First, objects of inquiry are understood as fixed entities with attributes. While attributes may shift over time, the passage of time does not necessarily affect the entity itself. Second, variables possess univocal meaning. There is an assumption against providing a variable with ‘contradictory effects’ because it would complicate the falsifiability of the model, as well as potentially introduce endogeneity. Finally, there is a commitment to the validity of the model’s predictions independent of context.Footnote 12
Abbott sees a central element of GLR as the assumption that entities are radically separable from their attributes. In the same way that many constructivists separate identity from the corporate entity – that is, ideas from the material substrate – this claim assumes that entities are the same regardless of where we find them on the timeline (Wendt Reference Wendt1999). Their attributes may change, but their ‘thing-ness’ does not (Jackson and Nexon Reference Jackson and Nexon1999). Animating this assumption is the idea that one can only devise theories and make predictions if the comparisons (across time) are speaking to the same entity. Scholars must be careful not to compare apples and oranges; robust theory predicting interstate conflict cannot be based on observed behavior if ‘states’ are not always states across time. A theory that sought to explain great power aggression by states in the 20th century, for example, would have difficulty if it uncritically treated ancient empires, colonial powers and modern states under the same heading as the same entity – unless, of course, there was good theoretical reason for believing all these could be coded as such. For rational choice, formal modelers and large-n studies, this goes without saying. Similarly for many constructivists, while the identity of the state may shift over time, its status as a state qua state (and thus the entity/attribute dichotomy) is retained. A state may gain power or lose power, build up arms or draw down, even shift identities from revisionist to status quo, but these are attributional changes, and therefore do not call into question its status as a state itself. One of the areas where this separation of entities and attributes is explicitly called into question is in post-structural theorizing. Performative theories of the state articulate a processual relationship between security and state subjectivity that calls into question the entity/attribute dichotomy (Ashley and Walker Reference Ashley and Walker1990, 1993; Campbell 1992; Weber Reference Weber1998). One of the concerns levied against this type of theorizing, however, is that it lacks the ability to make predictions, in some ways, precisely because the notion of process is so temporally bound (Kurki Reference Kurki2006). If entity and attribute are inextricable, comparing entities with varying attributes across time appears impossible.
GLR also privileges univocal meaning – the assumption against providing a variable with ‘contradictory effects.’ Empirically speaking, however, a single attribute can possess multiple meanings, and Abbott cites psychoanalytic theory as showing how anxiety can produce passivity or rage depending on context (Abbott Reference Abbott2001, 49). While it is certainly the case that IR and security studies have been adept at utilizing dual pathways and a nuanced treatment of variables, the epistemological commitment to falsification ultimately complicates these projects (King et al. Reference King, Keohane and Verba1994; Van Evera Reference Van Evera1997). Most theorists would express concern that multiple attributes for a variable within a logic renders it difficult to falsify, for example, if a logic predicts states will capitulate or escalate conflict depending on context, empirical validity becomes more difficult to demonstrate or refute.
Approaches that follow rationalist paradigms tend to privilege univocal meaning in the IR context. Threats stimulate internal or external balancing (Layne Reference Layne2003; Pape 2005). Greater transparency in information across states encourages or discourages conflict (Finel and Lord Reference Finel and Lord2000). Even the utilization of logics of appropriateness in contrast to consequentialist logics still lean toward univocal meaning: there is an appropriate way for states to act based on their identity as a democratic actor or adherent to the nuclear non-proliferation regime (Finnemore Reference Finnemore1996; Tannenwald Reference Spivak, Williams and Chrisman1999). Works that employ polyvocal meaning exist, but still must position themselves with respect to this presumption. Path dependency, for instance, includes the possibility of polyvocality, but still sees resistance from those who view it as either teleological or simply thick explanation (Peters et al. Reference Paul, Larson and Wohlforth2005; Bennett Reference Bennett and Elman2006). Similarly, emotionality has become increasingly relevant in explaining behavior in the security sphere; yet, factors such as status, respect, state ‘anger,’ and other ideas still must retain commitments to singular outcomes or articulate ideas similarly (Murray Reference Murray2010; Hall Reference Hamati-Ataya2011; Paul et al. Reference Patomäki2014).
Finally, independence of context appears to be one of the most powerful and widely held notions that Abbott calls into question. He sees this commitment as necessitating ‘the causal meaning of a given attribute cannot, in general, depend on its context in either space or time. Its effect does not change as other variables change around it, nor is its causal effect redefined by its own past’ (Abbott Reference Abbott2001, 59). In the context of IR this assumption appears largely as is. If one posits causal statements that empire formation results in over-expansion, then one can unproblematically assert evidence for this causal logic from ancient times, medieval times, as well as modern times (Snyder 1991, Reference Snyder2003). So long as the ‘important’ elements line up, the point in time at which these entities are located is of no causal significance. There is nothing intrinsic to temporal context that precludes a causal logic from retaining its explanatory potential. Constructivists and others may problematize this claim by invoking sociocultural context, but this questioning does not arise from the temporal location of these entities – these critiques are only useful if they can articulate a particular socio-political reason why disparate temporal locations matter.
Independence of context arises from sociology’s temporality and representation of time (according to Abbott), but some argue that the alternative – a focus on time – reductively and narrowly focuses on the time period an event took place to the exclusion of other factors. Falleti and Lynch responded to Buthe’s idea of ‘periodization,’ where the past is bracketed into periods defined by ‘critical junctures,’ – that is, post World War II – with the claim that a focus on time-periods eliminates context (Falleti and Lynch 2009). They argue that context is layered and occurring at differential rates that time-based brackets may not capture. Focusing on important moments in one layer (IR) may obscure what could be more relevant and important factors (state formation and production, for instance) in other layers. What Gryzmala-Busse and others argue in response, and I further claim here, is that foregrounding temporality actually makes context more contextual. Appropriately investigating the temporality/temporalities of the ongoing, layered historical processes intrinsically involved in international political actions would potentially be an important element of any attempt to make cross-historical causal claims. The argument for temporality here is not about assigning time-periods of investigation ex post facto, but utilizing the temporal understandings of the actor to assess what they consider important.
If IR seeks to understand practitioners, be they individual, collective, or otherwise, then investigating break points and time periods in the respective political imaginaries does provide insight. Westphalia, the inter-war period, post-Cold War, and post-2001 all could (and in some cases have) serve as important points of distinction. Moreover, as Park’s analysis importantly reminds us, these points may be determined in practice and not simply through the scholar’s understanding of events. Why we as scholars choose common points of demarcation should be subject to more rigorous temporal and reflexive analysis to better understand whether and why these points correspond to shifts in observable behavior rather than doing it through scholarly fiat. Doing so could open IR to new areas of inquiry, both geographically and temporally.
The question of alternative: thinking differently about temporality
Ultimately the goal of this article is to defend the value of thinking differently about temporality so to develop a more robust inquiry into conceptions of time across the discipline. Perhaps most importantly, this article seeks to demonstrate that the representations of time that dominate the discipline are not natural, essential, or objective. The implication is that even for those who claim to treat time in a nuanced manner, they leave its representation unquestioned. By largely accepting this notion of time, IR models are built upon a notion of clock-time that remains at odds with the political and social experience of the actors under investigation. The first step that needs to be taken, then, is an acceptance that the dominant manner in which we conceptualize, use, represent and understand time is a particular construct potentially at odds with our area of inquiry. Using ‘time’ better or taking time seriously is to the good, but ultimately incomplete. Recognizing the constructed nature of that representation of time in our own work and in political practice requires a substantial rethinking of how we represent time and understand its influence on our work, not simply finding ways to shoehorn in increasingly complicated ideas based on spatiotemporally constituted ideas of a neutral, unitary past.
Rather than dictate how temporality should be approached or understood from one viewpoint, this article can only begin the process by offering concepts to reflect on in future scholarship. Regardless of methodology or approach, greater inquiry into four issue areas could prove valuable – scope, contingency, events, and temporal context.Footnote 13 In terms of scope, scholarship should move to reflexively examine the temporal boundaries at play in the arguments and theories IR offers. Accounting for temporality would make explicit the temporal conditions to which the theorizing applies, rather than leaving these unstated or simply willed into existence by convention or authorial prerogative. Just as many projects are highly specific about where and how the theoretical pronouncements apply (e.g., democratic regimes, autocracies, microstates, systemic structures, empires, Asian powers) scholarship should more explicitly open itself to investigating when and in what context the pronouncements, predictions, comparisons, or lessons arising from it can be utilized in the future. This could take the form of bracketing claims to a particular period and while this certainly occurs in the status quo, especially given the move toward mid-level theory of late, foregrounding temporality in IR would go well beyond. For instance, limiting theories to those with shared temporal processes or actively theorizing the predictive scope of the claims advanced, including the temporality of the future in question (e.g. is it likely to unfold linearly, cyclically, or something else entirely). This does not necessarily preclude ‘timeless’ pronouncements, but it does mean that those claims would have to be defended rather than assumed. For instance, there may be good reasons to believe that arms races do (or do not) result in conflict regardless of temporal context, but the warrant behind a claim with such a broad temporal scope needs to be more directly defended.
Second, foregrounding temporality in IR would promote deeper investigation into the contingency of international politics and the relationships between the actors, events, and structures that constitute it. Projects engaged in theorizing security inherently address the contingency of political events, but recognizing the role of temporality would build it into all aspects of the project, beginning with the ontological recognition that entities are not necessarily solid or static across time. Approaches that foreground temporality could posit entities as products of processes and incorporate the contingency that relational understandings of international political life require (Thelan Reference Tannenwald1999; Capoccia and Keleman Reference Capoccia and Kelemen2007; Gryzmala-Busse Reference Grzymala-Busse2011; Patomaki Reference Park2011). Doing so would place the burden on the scholar to demonstrate why there is continuity or apparent stability in the entities or agents themselves, rather than simply the robustness of the particular claim being advanced. It is important to note that this move would not presume some sort of radical fluidity or inevitable indeterminacy that precludes making statements that possess meaning across any amount of time. After all, there are many entities within international politics that are stable across time. What it does require, however, is greater attentiveness and research into why stability may operate across time, as well as why certain events and/or institutions are understood to be stable across time. Both of these concerns could enable a better understanding of the role of structures in international politics. It would also leave open the possibility that theoretical models, agents, and structures can be altered dramatically by new events.
A third aspect of this move would entail greater attention to temporal context(s), meaning scholars would more directly theorize their understanding of how future events relate to the past. Similar to contingency, the generalizability of explanatory models would not simply be assumed, but deliberately stated with potential boundaries, scenarios of applicability, and imagined situations where the prediction is more or less likely to hold. This approach would replace the binary spectrum of more or less generalizable. Rather than beginning with simple explanations and then moving toward more complex, temporally contingent ideas if these theories do not work, models that rely upon temporal context would be placed on equal footing from the outset. This would promote the investigation of temporal dynamics throughout each phase of scholarly knowledge development. International events are inevitably complex; adding explicit temporal dimensions does not necessarily make them more so.
Some could object that this heightened focus on temporality would make an already difficult task unnecessarily complex. Scott de Marchi has argued that more temporally attentive approaches could render scholarly work needlessly complex and ultimately unworkable. De Marchi uses the example of constructing facial recognition software to demonstrate why time as a ‘feature’ is problematic because ‘one must strive to reduce the dimensionality of parameter spaces when one confronts complex problems; otherwise, one never has enough data to determine whether a model captures something essential [emphasis added] about a problem or only some nonsystematic component of the sample’ (Reference De Marchi2005, 53, emphasis mine).
I agree the fear of over-complexity is legitimate, but not sufficient to warrant continuing to treat temporality as a background issue for four reasons. First, the presumption that time is simply a ‘variable’ to be added is emblematic of the positivist approach that treats time as just another element of an observation. However, temporality – especially in the context of international security with its focus on massive socio-political events – is no mere variable. Political life can only be understood in relation to time and only through the interaction of processes across time do politics even come into being. Second, the idea of models capturing something ‘essential’ about a political ‘problem’ becomes complicated once one accepts the foundational aspect of processes in political life. ‘Essential’ qualities are non-existent; qualities come into being through, because of and within time, and this is especially the case in international politics. Third, the idea that ‘parameter spacing’ could become too large relies upon the notion of ‘time-less theory,’ as those ‘parameters’ could conceivably be limited through temporal scope conditions – that is, consciously limiting the space examined in terms of time. Finally, there is the issue that in the case of IR generally, but international security in particular, there are frequently problems with sample size – great power wars, nuclear weapons development – where small-n analysis is necessitated by the empirics and a focus on temporal aspects may be especially important, even if it would be inappropriate in other contexts with plentiful data points.
All of this is not to say that temporality and temporal dynamics should always come first and foremost in scholarly work. After all, IR is a problem-driven discipline that focuses on generating useful knowledge about issues with very real implications such as political violence, poverty, and forced migration. Foregrounding temporality is a move in service of that end, not an end in and of itself. While focusing on questions of time and temporality to the exclusion of political practices is problematic, it is also the case that there is currently a bias in IR against theorizing these questions. There is certainly a danger in over-correcting, but refusing to begin the process for fear of tumbling down a slippery slope of scholarly navel-gazing, or alternatively because the potential effects are too far-reaching, is equally dangerous. Foregrounding temporality would bring a critical aspect of politics and its practices back into the frame, and if this were to have any effect, it would be to make IR’s empirical research more empirical, not less.
Finally, accepting the importance of temporality would move IR toward recognizing an eventful temporality as the norm rather than the exception. Incorporating an eventful temporality would effectively call into question some of the commitments and epistemological assumptions that currently dominate the field. For example, highlighting the temporal scope of the claims made – both in terms of prediction and explanation – or identifying specifically how temporal location affects the contingency of the claims, could become part of the work itself rather than potential criticisms in need of addressing after the fact.Footnote 14 As well, reconceptualizing the commitment to epistemological assumptions about time-invariant theory or theoretical claims that rely on history as objective data would allow for a more contextualized approach to theorizing the central feature of security studies – war, armed conflict, and the use of force (Quadagno and Knapp 1992; Wagner-Pacifici 2010).
Privileging an eventful temporality like that offered by Sewell and other historical sociologists would also encourage a more nuanced approach to the historical event itself. Recent work on Deleuze and Lundborg’s application of the concept of the ‘pure event’ highlights that not only could the relationship between events come under further theorization, but the event itself could be more fully understood as a temporally complex occurrence. Events like ‘9/11’ operate as ‘ambiguous events that seem to lack a straightforward connection to a present state of affairs’ because they do not possess a ‘clear and present being’ but rather operate in an ‘ambiguous process of becoming’ (Lundborg Reference Lundborg2012, 4 and see also Jarvis Reference Jervis2009). As the ethnic conflict example demonstrates, one need not accept Deleuzeian thought in its entirety to recognize the manner in which events like wars or terrorist attacks operate at the ongoing intersection and temporal overlapping of past and future. This is both for the ‘moment’ itself as well as our ex post facto interpretations and representations of these events – for example, the end of the Cold War, ‘9/11,’ and World War II.
In addition to reemphasizing the transformative potential of events, foregrounding temporality would direct attention to the ways in which IR and security studies narrates events of the past, present, and future. Recognizing this narrative aspect of theory construction, testing, and development provides further potential for a more productive approach to temporality. Regardless of methodological approach, IR’s study of international security parallels history in that scholars narrate a series of events in a way they find intelligible and offers insight into future events. Quantitative scholars may employ data points, formal modelers may offer decision trees and qualitative/historical scholars might utilize case studies, but each are involved in articulating events in the past, present, and future and theorizing how they relate to each other. In the broadest sense of the term they each offer a narrative that makes sense of these events. The debates surrounding narrative methods offers insight into engaging temporal assumptions and can be used as a starting point to de-privilege some of the unstated temporal assumptions at play in understanding war and international conflict.
De-privileging dominant temporal understandings by incorporating insights from narrative methodology would also emphasize process and relationships, rather than entities and attributes. Recognizing the narrative dimension involves making
processes the fundamental building blocks of sociological analyses. For them social reality happens in sequences of actions located within constraining or enabling structures. It is a matter of particular social actors, in particular social places, at particular social times. In the context of contemporary empirical practice, such a conception is revolutionary. Our normal methods parse social reality into fixed entities with variable qualities. They attribute causality to variables … variables do things, not social actors. Stories disappear … Contingent narrative is impossible (Abbott Reference Abbott1992, 428).
One of the most important insights that can be derived from this branch of IR scholarship is that recognizing the narrative element of scholarly practice enables a shift in epistemological approach. In the words of one of its advocates, ‘narrative … is not about words versus numbers or complexity versus formalization. It is, instead, the portrayal of social phenomena as temporally ordered, sequential, unfolding and open-ended stories fraught with conjunctures and contingency’ (Griffin Reference Griffin1993, 407). Importantly, well-told stories possess an open-ness and recognition of the contingency of events that is best employed with a relaxation of some of the dominant epistemological and temporal assumptions along with active theorizing of how the particular events under investigation relate (Bates et al. 1998).
Perhaps most importantly, choosing to incorporate temporality would open up new spaces for analytical investigation. Even if the temporal critique offered here is correct and its implications widespread, there are certainly many claims in IR that do possess significant correspondence with international political life. One question raised by this critique is ‘why is it that these theories seem to work, even though they under-theorize the role of temporality?’ For instance, how is it and perhaps more importantly, why is it that the Roman Empire, Napolean’s military experiences and Britain in the 19th century can be utilized to meaningfully analyze the US role in the world in the 21st century?Footnote 15 Historical memory might be more central to theoretical predictions than otherwise assumed (Berger Reference Berger1997; Rasmussen 2003). These are very much open questions, but examining the why involved – rather than relying on an epistemological assumption that they do until proven otherwise – will allow for a better understanding of what issues can be bracketed, and which cannot.
Conclusion
Foregrounding temporality in IR would encourage treating time and its interpretation not just as a marker on the spatiotemporally constituted timeline, but as an intrinsic and important aspect of social and political life regardless of approach, methodology, or school of thought. While scholars necessarily simplify their representations of socio-political life, this article argued because temporality forms such a critical and constitutive aspect of political life, scholars must continually and explicitly interrogate whether it is captured in an effective manner and whether or not the means by which we produce this knowledge misses something important.
While the development of an approach that incorporates a more effective understanding of temporality is valuable, this article also seeks to create an awareness of the importance of temporality and raise it to the level of an independent issue in broader epistemological debates within the discipline. Currently, debates about epistemological assumptions in IR and security studies are not explicitly focused on the temporal dimension of scholarship and the insights into time and temporality that do exist remain relatively segregated across IR. Even for those scholars who do more fully incorporate ‘time,’ they largely leave its representation unquestioned. Part of the value of treating temporality as a stand-alone issue in IR is that it would better enable an investigation of issues that remain largely implicit. Moreover, conceptualizing temporality as a concern provides an umbrella and epistemologically open space for discussion of an issue that implicates scholarship across approaches and substantive areas of inquiry.
Although it is beyond the scope of this article to fully investigate all of the potential implications of bringing temporality into the larger discussion surrounding IR’s epistemological commitments, one area where this could have specific impact is in relation to the heightened moves toward reflexivity in IR theory and security studies (Brigg and Bleiker Reference Brigg and Bleiker2010; Hamati-Ataya Reference Hamati-Ataya2010; Hamati-Ataya 2011; Hom and Steele Reference Hom and Steele2010; Eagleton-Pierce Reference Eagleton-Pierce2011; Amoureux and Steele Reference Amoureux and Steele2015). Situating scholarly work and political practice in a temporal context could destabilize some of the objectivist commitments that de-emphasize reflection, as Berenskoetter has argued (Reference Berenskoetter2011). Other disciplines – most notably post-colonial studies – have turned their analysis back upon themselves in actively theorizing how their work and representations function and shape the political practices under investigation (Spivak et al. Reference Spiro1994). IR scholarship does not occur in a vacuum and viewing scholarship as particular interventions at particular moments, rather than mimetically representing the world ‘as it really is,’ could assist investigations of the political implications of security studies in the present as well as provide insight into the continuing theory/policy gap (Lepgold and Nincic Reference Lepgold and Nincic2000; Bleiker Reference Bleiker2011).
Acknowledgements
The author would like to thank Andrew Hom, Kimberly Hutchings, Patrick Jackson, Adria Lawrence, George Lawson, Charles Lipson, Patchen Markell, Candice Moore, Michelle Murray, Jade Schiff, William Sewell, Duncan Snidal, Ron Suny, Alexander Wendt, and participants at the University of Chicago PIPES workshop, as well as anonymous reviewers and the editors at International Theory.
