Neorealism, neoclassical realism and the problem(s) of history

Following scholarship on IR’s ‘historical turn’ as well as on neorealism and neoclassical realism, this article finds fault particularly in neorealism’s implicit reliance on the historically contingent but incompletely conceptualised transmission of systemic factors into state behaviour. Instead, it suggests that neoclassical realism (NCR) is well-suited to leveraging ‘history’ in systematic and general explanation. This article interrogates two routes towards a historically sensitive NCR (intervening variables and structural modifiers), and how they enable different operationalisations of ‘history’ as a sequence of events, cognitive tool or collective narrative. The first route suggests history underpins concepts and variables currently used by neoclassical realists. Here, history is more easily operationalised and allows a clearer view at learning and emulation processes. It is also more clearly scoped, and therefore less ‘costly’ in terms of paradigmatic distinctiveness. The second route, in which history modifies structural incentives and constraints, is more theoretically challenging especially in terms of differentiating NCR from constructivist approaches, but lends itself to theorising systemic change. Both routes provide fruitful avenues for realist theorising, can serve to emancipate NCR from neorealism in IR and foster cross-paradigmatic dialog. Examining how ‘history’ can be leveraged in realism allows interrogating how other ‘mainstream’, positivist approaches can and should leverage historical contingency, context and evidence to explain international processes and outcomes.


Introduction
In much of mainstream IR, and particularly in neorealist and neoliberal schools of thought, interests and behaviours are assumed to be predetermined and universal through time and space. History and historical analysis are then relegated to 'footholds' for validating more general theories of foreign policy and international politics. 1 Of the major '-isms', only constructivism is 'propelled towards accounts of time and place specificity, context and change'. 2 The 'historical turn' in IR scholarship has therefore largely been a turn with a direction: away from 'scientific' theorising, and especially from neorealism, in which Schroeder famously diagnosed 'an attitude toward history not uncommon among scholars of many kinds: an unconscious disdain for it, a disregard of its complexity and subtleties and the problem of doing it well or using it wisely; an unexamined assumption that its lessons and insights lie on the surface for anyone to pick up, so that one can go at history like a looter at an archaeological site, indifferent to context and deeper meaning, concerned only with taking what can be immediately used or sold'. 3 Perhaps inevitably, then, the insight that historical context matters has turned scholars, including realists themselves, towards more historically literate research. More recently, neoclassical realists have suggested that (the sum of) state behaviour(s) accumulates over the longer term to produce 'systemic outcomes', and even 'reshape the structure of the international system'. 4 This was a risky move. Not only was it potentially self-defeating because it anchored systemic outcomes in unit-level behaviour (what Waltz called 'reductionist'), it also hinted at the possibility of an iterative relationship between structures and agents reminiscent of ('thin') constructivist work and thus challenged paradigmatic boundaries. It provides a starting point to specifying the relationship between history and realist theories of state behaviour and international politics.
In this article, I interrogate venues for a historically literate realism that introduce an 'understanding of the contingent, disruptive, constitutive impact of local events, particularities and discontinuities' all the while retaining aspirations to systematically theorising international phenomena. 5 I am less interested here in the 'true' nature of history, that is, what it is, which constitutes the wider 'problem of history'. 6 Different conceptualisations and understandings of history, alternatively as an (1) objective context (a series of events), (2) individual memory and/or practical knowledge based on worldviews and value judgements or (3) social recollection of collective past experience, for example, in the form of cultural narrative, have permeated the philosophy and discipline of history and the wider social sciences. 7 I am more interested here in how these different conceptualisations of how history works have been and can be leveraged theoretically and methodologically to inform realist theory, and by extension explanatory theory in IR. They 'shape and shove' the theoretical process in particular ways. 8 More specifically for the purposes of this article, I argue that they provide fruitful avenues of inquiry within and at the paradigmatic borders of realism. As I explore below, much of (neo)realism tends, because of its desire to glean from historical example universal rules of international politics, towards the first conceptualisation and to treating history like a pool of data. This 'naturalistic' inclination opened neorealism to the historicist challenge in the first place. 9 Finding a theoretical place for the latter two conceptualisations in realist approaches is therefore likely more challenging, but need not require a return to those classical realists sceptical of the 'scientific' study of history. 10 Rather, it is in neoclassical realism (NCR) and its 'intervening variables' and

(Neo)realism and the problem(s) of history
Neorealists and neoclassical realists conventionally share assumptions on the ahistorical importance of anarchy, systemic conditions and the balance of power. They suggest that incentives and constraints derived from an international system organised by the structural rule of anarchy drive state behaviour. 'Structure' here describes a set of macrosocial arrangements, 14 also referred to as the 'rules of the [international] game'. 15 Realists assume that, interwoven with the material world, an inescapable anarchical system produces a variety of ever-similar pressures -incentives and opportunities, threats and constraints -which condition a state's national interests. An anarchical system defined in terms of power and the corresponding uncertainty over how other states choose to employ their power provides the overriding interest for states for their own survival, security and welfare. Therefore, states are always interested in and trying to assess their position in the international system and their own as well as other states' relative capabilities to deduce the level of threat they face. 16 These relative power capabilities consist of different factors, to various degrees deducible or 'measurable' by outside observers (including decision-makers). 17 States react to (changes in) other states' power, assumed intentions and actual behaviour (or how they chose to employ their capabilities) through some form of hedging, balancing or bandwagoning. While most states are on average able to correctly assess these systemic stimuli, those that do not fail to prosper and risk their survival over the long run. 18 This produces macro-patterns, especially balances of power, which do not rely on specific contexts, periods or historical understandings. Rather, they logically derive from an unchanging security dilemma. This theoretical formulation is characterised by inescapable 'recurrence and repetition', by a cyclical theory of history. 19 Mearsheimer notes, 'international politics has always been a ruthless and dangerous business, and it is likely to remain that way'. 20 These arguments are often formulated in a 'scientific', axiomatic manner, which reflects (much of) neorealism's ontological and epistemological commitments. It relates to a particular, 'naturalistic' understanding of theory, history and (social) science that developed out of 19th and 20th century debates in the philosophy of history. 21 It intersects with what Vaughan-Williams calls the 'familiar narrative' of how IR developed in the post-WWII era. The 'behaviouralist revolution' kicked off a gradual development towards privileging 'structure and space over time and context' in political science, away from idiographic to more nomothetic conceptions of history that value generalisation combined with explicit, parsimonious assumptions and scope conditions. 22 History becomes exogenous and superfluous to analysis, useful primarily to illustrate the inescapable rules of world politics, support hypothesis formulation for historical cases and inform current policy in the form of 'lessons'. 23 Hobson and Lawson call this mode of historical inquiry 'history without historicism'. 24 This development, Vaughan-William argues, is exemplified by neorealism, as 'focused on the supposedly timeless regularities of the state and states system instead of the contingencies of life'. 25 It was directed against a more historically oriented intellectual predecessor in classical realism. 26 Classical realist writings do not provide an analytical framework in the sense neorealists desire: they are better understood as 'philosophical insights about human nature, power and politics' rather than a parsimonious theoretical structure with predictive capacity. 27 In terms of knowledge production, this development invites the challenge that (neo)realism's supposedly universal insights are simplbased on some histories to the exclusion of others. 28 The failure to adequately contextualise key concepts like the state, power or anarchy, leads to 'subliminal eurocentrism'. 29 In turn, realism's disciplinary dominance is suggested to stifle alternative lenses that interrogate historically divergent state behaviour.
Three aspects complicate the oft-posited linkage between neorealism and ahistoricity, however. Firstly, recent research has questioned whether the predominant interpretation of neorealism as embedded in quasi-positivist, 'scientific' epistemology corresponds to the leading neorealists' scholarship. From one side, this had led to multiple attempts to pin various (neo)realists down to testable propositions, on which they were then found not to deliver. 30 From another, it has been argued that, contrary to theoretical shorthand, for example Waltz was not a positivist in any coherent sense. 31 Waltz himself repeatedly defended his theory on the grounds that it does not seek to predict and cannot be 'falsified' by testing against specific evidence. 32 Whether this argument is representative of other neorealists is debateable, of course. 33 Rather, secondly, many contemporary realists ascribe to 'soft' positivism (although notable exceptions exist), that is, to testing theory against empirical evidence with the goal of generalisation combined with an awareness of social science's inherent limitations. This relates to subjectivity, interpretation, unpredictability and research ethics which make it difficult to evaluate objects of research based on agreed standards and methods. Therefore, objects of research are generally assumed to be similar enough to discern generalisable patterns. 34 This softer positivism allows neorealists to contextualise their results, 35 and to 'fill in' their general propositions with historical analysis. 36 For example, Gilpin argued that across different historical cases foreign policy choice was not in fact predetermined by the possibility of conflict but rather its contextually dependent likelihood, and consequently included 'a causal role for ideas, institutions and domestic politics'. 37 This is not fully historicist, of course, in that the underlying theory remains anchored in generalist assumptions (as intended) -but it presents a first step to treating contingency and contextuality as more than 'irrational' exceptions. Where following this path further reduces realism's 'scientific credibility', it also allows realists to appreciate contextual, agential and ideational facets of state behaviour more fully. 38 Thirdly, neorealism's assumptions are grounded in a perspective on learning and strategic failure based on preceding example, and thus not fully ahistorical (or at least atemporal) themselves. This relates to, for example, emulation of successful state behaviour, which over the long term and on average results in functional similarity of the system's units. Recent research also questions the predominant interpretation of neorealist scholarship that suggests stricter, determining effects of anarchy on patterns of state behaviour. Rather, structure constrains behaviour and makes specific outcomes more likely. It does so by functioning as a 'selector' 39 which rewards some foreign policy choices and punishes others over the long run and on average. 40 Any decision-maker who refuses to obey the selfhelp logic embedded in the system risks the survival of the state they represent. Decision-makers therefore seek to perceive systemic conditions, deduce the national interest and pursue policy accordingly. In turn, state behaviour may also be generative of structure. 41 It must remain unclear, however, how this 'transmission belt' operates to induce accurate appraisal and how failure would incur compliance, without considering the chronological order in which neorealism seems to imply (but not explicate) such dynamics operate. In addition to recurrence and repetition, trial-and-error requires iteration, that is, the progression of time and use of historical examples for learning and emulation. Neorealism relies on history without explicitly analysing the linkage, that is, the transmission belt, between systemic constraints and state behaviour that could further elucidate how its lawlike regularities emanate out of past practice and experience.
Both critics of neorealism as well as neorealists themselves seeking to defend their propositions have already highlighted that at least in the short term and under specific conditions, other things matter more than systemic factors in producing state behaviour; and that history can tell us what these things are or have been -including so that leaders can learn from them and change policy. Rethinking the use that history, however conceived, may have in realist approaches therefore seems timely as well as theoretically appropriate. In most of its contemporary varieties, neorealism remains (intentionally) geared against fuller historicisation precisely because it seeks to remain at the highest level of generality (what Hobson and Lawson call a 'mega-macro' mode of history). 42 In the below, I argue that other forms of realist theorising, such as NCR, are more open to such an endeavour.

Neoclassical realism: bringing history back in?
In contrast to (much of) neorealism, NCR explicitly analyses the translation of external stimuli into state behaviour. 43 While characterised by considerable divergence, NCR can broadly be characterised by an 'outside-in' approach to the analysis of state behaviour and international politics. 44 Its initial gambit and best-known contribution has been the attempt 'to explain variations of foreign policy over time and space' by supplementing neorealism with unit-level variables that mediate the impact of systemic stimuli. 45 Some neoclassical realists deviate from neorealist 'baselines' primarily to explain historical 'mistakes' in which systemic drivers failed to generate expected behaviour ('type I'). Others started developing more generalisable statements about the regular interaction of systemic and domestic factors in the production of state behaviour ('type II'). Recently, yet others have turned to interrogating patterns of international relations as an output of state behaviours ('type III'). 46 All three types start from the less strict understanding of the constraints systemic conditions impose on state behaviour developed above. States are thus assumed to have 'considerable latitude' in defining their interests because structure 'merely sets parameters' for state behaviour. 47 Systemic conditions are neither obvious nor specific enough to guide decision-makers to only one possible course of action. 48 Decision-makers have limited access to relevant information, 49 and are uncertain about which policies are appropriate in response to a given scenario. 50 In their choices, they are limited by domestic political incentives and constraints, which are captured as intervening variables. 51 Different neoclassical realists employ widely different intervening variables. 52 To overcome the limitations an ahistorical neorealism imposes on the study of state behaviour, many neoclassical realists return to writers like Machiavelli, Carr and Morgenthau. 53 Here, they quickly re-discover historical sensitivity, and insights on the importance of context and contingency. For Machiavelli, while fortune and necessity narrow the range of alternatives, they still require statesmen to apply their wisdom to grasp contextual opportunities. Indeed, part of the point of his writings is to outline notable historical instances of prudent and imprudent judgement. 54 For Carr, the relation between the imperatives of the international system and resulting state behaviour is not simply deterministic: there is still 'something which [the decision-maker] can think and do, and [. . .] this thought and action are neither mechanical nor meaningless'. 55 They depend on a leader's understanding of history, morality and their ability to persuade. Historians and leaders are both products of their own places, times and generations, which in turn decides what 'facts of the past' they determine into 'facts of history'. 56 Morgenthau acknowledges the importance of historical knowledge and experience for ethics, evaluation and political judgement, 57 and as a 'source for change'. 58 Taking into consideration historical experience distinguishes politics from the mere application of violence. 59 Morgenthau in turn is influenced by Weberian thought, where history shapes the relationship between interests and ideas. 60 One way to historicise realism, then, may be to return to its classical realist roots and ditch the 'neo'.
For most neoclassical realists, however, the aim in harking back to classical realist writing is to combine the historicity of classical realism with the systematisation of neorealism. 61 Classical realists do not develop a systematic theory of international relations in the sense neorealists aspire to, nor did (or do) they seek to. Neoclassical realists usually posit that there are things in neorealism that are worthwhile preserving: its intuition that systemic factors prevail in driving state behaviour, primarily, as well as its commitments to systematic causal logic, generalisability, empirical measurement and even prediction. 62 Evidently, neoclassical realists have been and are open to critiquing and deviating from neorealism. In most NCR analyses, some version of neorealism functions as a 'baseline', to the extent that NCR scholarship is often categorised by the degree to which it diverts from neorealist assumptions and expectations. NCR is therefore often charged with being a supplementary theory, a 'neo-neorealism' -both empirically, where it explains phenomena neorealists cannot, and theoretically, where it strives to 'rescue' neorealism from criticism. 63 In response, neoclassical realists have moved further away from NCR's initial, 'type I' gambit, that is, supplementing neorealism in explanations of state behaviour by adding unit-level variables. They have instead tried to outline how NCR can stand on its own as a theory of international politics. 64 This 'type III' NCR revolves around the production of systemic outcomes as an accumulation of unit behaviour (foreign policies) across the system. It sees structure as an 'emergent property' generated by the 'interaction and arrangement of the units' 65 and has brought renewed focus on the specification of systemic drivers and outcomes. On the one hand, this has involved thinking about relative and changing levels of threat and permissiveness, including in regional sub-systems, as well as about a variety of so-called structural modifiers. 66 On the other, it has renewed interest in the production, or perhaps accumulation, of behavioural patterns (such as balances of power) and the causal mechanisms involved across different units, such as socialisation or emulation. 67 This 'type III' direction in neoclassical realist theorising is likely to further increase contestation over its adherence to paradigmatic boundaries, and its overlap especially with (realist) constructivism. 68 In Guzzini's terms, trading in concepts usually associated with other paradigms purchases better explanations but incurs costs in terms of distinctiveness (also see below). 69 Of course, suggestions that realists 'cannibalise' other paradigms and hoover up 'their' concepts at times rely on an artificial straitjacket that ignores classical realist roots in particular (as Guzzini also discusses). 70 It also threatens to police theoretical progress in unintended ways: openness is theoretically productive in ways that adherence to paradigmatic boundaries usually derived from mid-range theory expectations is not. 71 In particular, it has opened new venues of systematic theoretical inquiry into the drivers and outcomes of state behaviour, at the same time as it has forced neoclassical realists to carefully justify their selection of concepts and cases. 72 For example, Guzzini lauds Wohlforth's engagement with 'power' precisely because he acknowledges realist 'indeterminacy' for any specific historical context. 73 For the purposes of this article, this openness invites pushing and probing paradigmatic boundaries further.
Can realism be historicised to allow for an inquiry into state behaviour that takes history seriously while retaining aspirations to generalisation across cases based on a structural-systemic logic, that is, one closer to what Hobson and Lawson call a 'midpoint' mode of history? 74 If so, where would history be at work theoretically? Following NCR's two theoretical gambits (towards 'type II' and 'type III' research), I suggest that NCR offers two possible routes of inclusion: one, at the unit level, and another at the international level. Both come with benefits and costs; both raise new theoretical and empirical questions worthy of further consideration.

History at the unit level
As explained, NCR uses unit-level variables to explain, variously, distortions from neorealist baseline expectations (type I), general dynamics of foreign policy and grand strategic choice (type II) and patterns of state behaviour across the international system (type III).
A state's national interests are derived from material reality, that is, the relative distribution of capabilities in an anarchical system. However, the appraisal of these systemic conditions and their translation into interests is fraught with difficulty. Historical knowledge and experience generate those concepts that guide decision-makers in their interpretation of the reality of the international system and the deduction of interests. History would then be a reservoir of 'knowledge structures' 75 that help to order complex situations, and 'specify national interests amidst conditions of uncertainty'. 76 Across different decision-makers and governments, and with changing circumstances, this historical knowledge or imagination finds different discursive iterations, that is, be expressed in different analogies or metaphors for the purpose of historically informed reasoning. Specific sets of historical knowledge and experience serve as an existing domestic structure that mediates which concepts, perceptions or strategic narratives decision-makers or domestic publics can easily understand, articulate and are thus likely to adopt.
Indeed, historical experience underlies intervening variables neoclassical realists readily employ already, such as collective identity, strategic culture or nationalism. 77 These variables interrogate the role of worldviews which influence foreign policy and which are usually rooted in socio-cultural, political and economic experiences and contexts. 78 History would intervene in such a way at the domestic level because it would be channelled through a specific individual or collective perspective to inform threat perception, decision-making dynamics and bureaucratic processes. This fits with the second and third conceptualisations of 'history' (see above), as subjectively interpreted, selectively narrated and instrumentally used -rather than objective or 'factual'. Historical narratives are shaped as much (if not more) by language, culture, ideology, individual beliefs and cognition as by historical events. 79 Decision-makers interpret history or tap into existing, collective historical narratives per their own predilections, experiences, beliefs and socio-cultural background. This implies that 'history' is a reservoir from which perceptions, persuasion and programmatic principles are drawn (or onto which they are projected) in the foreign policy process.
This also provides a rationale for how realists can provide (historically informed) foreign policy advise. 80 As such endeavours would inevitably be subjective, these scholars help tackle the endemic uncertainty about systemic stimuli with the help of history, analogical reasoning and so forth. I am not suggesting that this is a particularly innovative insight. Rather, if history is theorised at the unit level (e.g. in how it affects individual perception, ideas and strategic cultures differently in different domestic settings) such scholarly activism is less afflicted by the contradictions neorealist policy advise countenances. 81 Treating history as something that affects predominantly unit-level variables raises questions as to the nature of a 'variable' within social scientific frameworks, as well as to the philosophical nature of 'history'. 82 Can history itself be treated as a variable? If so, would this be 'generalisable', in the sense of comparison across empirical examples, to not resort to either generic statements about some historical experience or artificially 'freezing' the course of history in specific instances? For one, history's causal role seems to fall outside what neoclassical realists usually conceive of as 'variables', notwithstanding the considerable range of different understandings. In particular, history and historical processes as developed above underlie other unit-level variables rather than playing the moderating, complementing or even primary causal role unit-level factors such as perceptions, strategic culture or state capacity play elsewhere in neoclassical realist scholarship. 83 Here, thinking about 'history' (rather than a specific vehicle of historical content, so to speak, for example, analogies) as a variable creates more theoretical problems than opportunities. Instead, history seems to serve as a latent factor, a basis from which more active variables (such as perceptions, beliefs, ideology or strategic culture) can be drawn. In that case, the question remains whether history is playing much of a causal or theoretical role itself, and what its added value might be compared to other unit-level variables. For one, thinking about history as characterising existing variables can help consolidate NCR's ever-proliferating variables, and thus neoclassical realist claims to general theory. 84 Such ordering aside, an answer might be found in explanations of change and learning, and the variety of domestic pathways to socialisation and emulation under conditions of anarchy. 85 Neorealism suggests that systemic incentives compel states to adapt internally through either innovation or imitation of successful other states, which leads to convergence of functions and behaviour over time. 86 The same incentives encourage socialisation into common international practices. 87 Failure to do either leads to punishment and ultimately threatens state survival. 88 This suggests a process of learning over time: the specific dynamics of emulation and socialisation are historically contingent. They operate domestically through existing procedures, mechanisms and ideas in a pathdependent fashion. 89 While interrogating learning poses challenges as regards distinctiveness which I elaborate below, it allows NCR to answer questions regarding change and progress within the constraints of an anarchical system that neorealism cannot. Relating chosen analogies and metaphors back to specific historical experience, for example, can identify why, faced with an ambiguous international environment, leaders in one country debate a crisis in terms of 'Vietnam' and 'Munich', while others resort to the Iranian Revolution or 9/11, and what this difference means for resultant behaviour. 90 Vocabularies available to decision-makers matter, both for how they think about international problems as well as for how scholars approach textual records. 91 This gives clues as to the origins of, and changes over time in, threat perception or the likelihood of decisional outcomes in intra-elite bargaining. This in turn allows neoclassical realists to explore forms of change and learning within the constraints of an anarchical system that neorealists have been reluctant to theorise. 92

History and the international environment
Limiting the use of 'history' to something that happens within, or can be tethered to, the boundaries of a state or individual is to artificially narrow much of the interlinked, interactive and trans-/supra-/sub-national nature of historical processes. Insofar as this is a useful analytical abstraction (e.g. because specific historical narratives or analogies are of interest for specific foreign policy decisions) it is defendable. Alternatively, to avoid theoretical problems inherent in treating 'history' as operating at the unit level, 'history' could alternatively be introduced as a structural modifier. Intuitively, the number, character and density of relations in a (sub)system is shaped by history. Rather than a process occurring at the unit level, history would inform the salience of systemic stimuli, that is, the extent to which the exogenously given, independent variable determines state behaviour.
The intuition of structural modifiers emanates out of a critique of (Waltz' treatment of) 'structure'. 93 The notion that only the third element of structure, namely the relative distribution of capabilities, is ultimately causal (as an independent variable) is precisely why neorealism is ahistorical, after all. 94 Per Buzan, systemic factors exist that 'not only affect the ability and the willingness of units to interact, but also determine what types of levels of interactions are both possible and desired'. 95 Buzan calls these factors 'interaction capacity' and identifies only two of them: the evolution (and diffusion) of technological capabilities and shared international norms and organisation. Both of Buzan's factors have a temporal dimension in that they presumably change with time. They are not structural but systemic, which means they 'condition the significance of structure and the meaning of the term system itself'. 96 They pull and push the units in the system in more specific ways than the distribution of capabilities alone would because they change the salience of threats and opportunities. Per Snyder, they 'modify the effects of the more basic structural elements on the interaction process'. 97 Van Evera gets at a similar notion in calling these factors the 'fine-grained structures of power' that affect the intensity of the security dilemma. 98 History, conceptualised as past interactions (i.e. objective contexts) and/or shared memories and experiences, conditions the significance of the distribution of power. It affects the units' interactions, rather than each unit individually, for example, by making the threat from rising power more salient in specific situations than others.
The challenge in including history in such a way relates to the systemic nature of structural modifiers. Snyder specifically argues that structural modifiers affect all units in the system in a similar manner, that is, the presence or absence of military technology (e.g. nuclear weapons) has system-wide effects. Some neoclassical realists have challenged this notion. They suggest that there is neither a theoretical necessity for system-wide effects nor much empirical grounds for it. In fact, modifiers considered by Buzan, Snyder and Van Evera such as technology, offense-defence balance and specifically geography are regional or even relationship-specific. The impact of structural modifiers may be 'limited to particular sectors or regions within the system, categories of units or pairs of units'. 99 Akin to geography, then, historical events and experiences 'create constraints and provide opportunities for some units and for patterns of strategic interaction within the given structure of the system'. 100 Historical experiences, for example, of war, past cooperation or broken trust, affect the environment in which states operate and thus their foreign policy choices.
Following this logic, states can change the way they interact through learning from experience -within the causal frame of an anarchical system defined by relative distributions of power. This fits empirically with the observation that cooperation and institutionalisation are possible, although limited, within an anarchical system. In addition to changing interaction capacity in terms of technology or norms, then, history may play a modifying causal role in shaping the experience of anarchy. For example, historical events or experiences may have raised the (perceived) costs of warfare, which replaced security competition with more peaceful forms of competition between great powers. It could explain the shift from a focus on territorial acquisition and military capabilities, to competing with other states by diplomatic means (perhaps limited to some sub-systems). This paves a way for NCR to take seriously history, whether understood as sequences of events or shared experiences, as an influential factor in pulling and pushing the system's units and making specific behavioural outcomes more or less likely. This model faces (at least) two theoretical challenges, however. Firstly, neoclassical realists tend to share with neorealists a loosely materialist ontology. 101 Conventionally, structural modifiers therefore have a material base in technology, geography or powerful actors promoting norms and institutions. What makes NCR realist is predominantly the shared assumption that state behaviour is a product of 'environmental compulsion'. 102 'History' is not easily traceable to a material base. Where it modifies NCR's independent variable, it may therefore blur paradigmatic boundaries with non-realist approaches. 103 In turn, it risks underplaying the contribution NCR offers in differentiation from both classical realism as well as various constructivist, historical and institutionalist approaches.
Secondly, many neoclassical realists seek to evaluate 'the causal impact of specific hypothesised independent variables (IVs) and intervening variables (IIVs) on the dependent variables (DVs)'. 104 This commitment to discernible variables, and to the independence and temporal priority of cause over effect implies a Humean concept of causality. 105 Here, causality refers to the 'difference between the systematic component of observations made when the explanatory variables take one value and the systematic component of comparable observations when the explanatory variables takes on another value'. 106 Where history affects the likelihood of specific outcomes, which in turn provide the independent variable for subsequent state behaviour, it creates iterative relationships of the kind usually associated with constructivism.
Indeed, some neoclassical realists already consider the possibility of such iterative causal relationships between systemic and unit-level factors over time, for example, in 'type III' NCR. 107 Ripsman, Taliaferro and Lobell argue that different strategic choices, influenced by systemic conditions under anarchy as well as unit-level intervening variables (which are causally linked with the independent variable), accumulate to affect relative power and international outcomes. 108 Over time, other states react to these new systemic conditions, as all units constantly re-assess the relative distribution of capabilities. Here, history provides both cause and effect of a causal path over time, which, while intuitive, is also problematic within the framework neoclassical realists conventionally uphold. More so, in the long term, systemic 'outcomes and the policies and grand strategies of the principal units themselves can help reshape international structure'. 109 Consequently, the authors argue, 'structural change has its roots in the individual investment decisions of the great powers, their decisions to pursue a grand strategy of restraint or one of overextension and the particular domestic constraints and opportunities of particular great powers'. 110 This supposedly positions NCR as a dynamic (rather than static) theory. Almost in the same breath, however, the authors limit this dynamism, presumably to clarify paradigmatic boundaries: structure retains 'the dominant influence over the range of systemic outcomes that are possible' and states do 'not determine the structure of the international system' through their policies, but rather have an 'impact' on it. 111 It remains underspecified how these arguments can causally hold at the same time. 112 More recently, it has been suggested that a solution might lie in clarifying what kinds of systemic outcomes 'type III' NCR seeks to explain -patterns of behaviour such as balances of power, broadly 'observable political phenomena' resulting from state interaction, 113 or changes in the system's character itself. 114 Here, thinking about history as a structural modifier helps explicitly problematise aspects of 'type III' NCR that are already present, rather than create new ones. Indeed, operating at the paradigmatic margins, so to speak, need not create only problems: for one, it can be generative of new insights and questions, and spur the theorising process. 115 To what extent NCR is tied to materialism and positivism is a matter of much productive debate, for example, with some neoclassical realists espousing more relational and even interpretivist frameworks. 116 This type of intra-and cross-paradigmatic debate is to be expected where a theoretical approach expands in scope all the while seeking to clarify conceptual underpinnings. 117 It points to venues for further conceptualisation and research into NCR as capable of exploring change of systems as well as change within systems. 118

How would we know, and why does it matter?
Whether 'history', conceptualised alternatively as objective context, individual memory or social recollection of past experience, is meaningful in NCR is not least also a methodological question, and thus requires methodological choices. What types of history get included, who gets to decide and what level of evidential threshold is required to suggest 'history' rather than any other, possibly more meaningful or more precisely operationalised, factor has affected or guided state behaviour? This speaks to the possibility that historicising NCR incentivises creative research designs and scholarly collaboration across disciplines, paradigms and area specialisations. It presents an opportunity for scholars to contribute new puzzles, cases and concepts to a still-growing school of thought, and thereby recapture some of the broader realist tradition's dynamism. 119 While NCR is not geared towards any particular method, the above suggests two fruitful directions for neoclassical realist theorising.
Firstly, neoclassical realists can engage other theories comparatively, for example, in terms of coherence or empirical accuracy. Neoclassical realists frequently do so by referencing neorealist baselines from which their own explanations deviate. Reflecting on how 'history' operates at the unit level or regarding the international environment opens to the approach a set of new, alternative 'baselines' for comparison. Broadening its comparative base, so to speak, can help emancipate NCR as a stand-alone approach (rather than remaining tethered to neorealism as its supplementary extension). Thinking about history predominantly at the unit level lends itself to in-depth analyses of a state's grand strategy formation and foreign policy decision-making, and thus usefully combines previous neoclassical realist analyses with those, for example, from different variants of foreign policy analysis. Treating past interactions as a structural modifier seems geared to understanding sub-system patterns of behaviour, such as the formation of regional alliances or differential treatment between some units of the system. It thus overlaps with the English school or historical institutionalism in short-term explanation but rejects the notion that primary institutions alter how anarchy conditions state behaviour. Here, Grieco's 'collaborative challenges' may be a suitable methodological option. 120 While erring towards a mode of 'history without historicism' within a (soft) positivist framework, such comparative theory-testing can yield intriguing results and be productive of new approaches. 121 Secondly, if the aim instead is to formulate a neoclassical realist theory of international politics ('type III'), and thereby return to 'grand theorising' (as opposed to the mid-level theorising more associated with comparative empirical testing), clarifying NCR's relationship to history is pertinent. 122 Ripsman, Taliaferro and Lobell's arguments on the difference between neorealism as a static and NCR as a dynamic theory capable of illuminating systems change points in the right direction. However, their formulation leaves unclear where NCR might differ from, for example, constructivist approaches to international politics. If accumulated policies, especially of great powers, can alter the structure of the system, 123 that is, the rules that govern the units' behaviour, this comes close to an anarchy (re-)produced socially by the system's units. This type of NCR moves ever closer to a 'structurationist' approach to historical analysis, where the past becomes, as Reus-Smit suggests, a realm of 'difference, of variation, of insights into the present drawn from contrasts and contingent connections'. 124 Of course, especially the conceptual 'long-term' may well be a (perhaps unusual but) theoretically desirable meeting point for neoclassical realists and (realist) constructivists. 125 Not thinking seriously and carefully about history explicitly and at the level of first principles and core concepts would then constitute a wasted opportunity to showcase NCR's analytical richness and diversity at the same time as neoclassical realists further specify their theories' assumptions, scope conditions and objects of analysis. A neoclassical theory of international politics has the potential to account for patterns of behaviour in an anarchical system all the while taking seriously both the insights different contexts provide and the forces that give rise to these contexts. As suggested above, even though it is unlikely to provide a generative theory for the system's underlying principles. NCR is open to investigating systemic change, emulation, learning and socialisation in ways neorealism is not. The aspiration NCR retains to generalisation within systemic framework avoids the 'butterfly effect' Lawson identifies in some IR scholarship that 'fetishises the particular and the exceptional, failing to see how historical events are part of broader processes, sequences and plots'. 126 A thusly reformulated NCR may find considerable overlap in works associated with (realist) constructivism, historical sociology and the English School. Where clear paradigmatic boundaries may be desirable from the perspective of comparative theory testing as outlined above, such overlap and blurry boundaries can also make for more fruitful scholarly contestation and dialog in terms of the theorising process itself. To the extent that it re-introduces aspects of historicism and even reflexivity into realism, this enables dialog less encumbered by the hierarchical, epistemological or methodological posturing and paradigmatic 'policing' diagnosed elsewhere. 127 Ultimately, the aim of such dialog is then not to accurately determine all relevant causes and processes, nor to subsume all approaches under one synthesised theoretical umbrella. 128 Rather, in the case of NCR, it concerns presenting coherent and systematic designs to investigate when and how systemic drivers interact with contextual factors to produce (patterns of) state behaviour. Within these bounds, the benefits in terms of theoretical innovation of thinking seriously and carefully about history may well be worth the costs in terms of distinctiveness. It is likely that neoclassical realists will continue to push against paradigmatic boundaries, and in turn try to re-assert distinctiveness (or else be reminded that they should). In this way, as Sterling-Folker outlines, contestation and dialog about NCR in particular mirrors similar tensions within larger disciplinary discussions about the role and nature of theory and how to evaluate its progress. 129 Thinking about history in the defined ways NCR suggests thus helps clarify different ways of understanding and conceptualising historical processes and narratives more broadly, and facilitates ways in which differently-minded and situated scholars can converse about key concepts, methods and empirical analysis. 130

Conclusion
This article mapped out the potential for historicising realist approaches to foreign policy analysis and international relations. Much of neorealism has explicitly espoused social scientific reasoning and (soft) positivism to correct the presumably eclectic and untheoretical historicism especially of classical realist writing. This standard narrative is complicated by three observations: that neorealist commitments to positivism are not straightforward, that they have been and are open to contextualising their empirical analyses, and that they make use of concepts such as 'socialisation' and 'emulation' in their theoretical mechanisms without fully interrogating them theoretically. I argued that NCR can correct for some of this theoretical awkwardness by offering two routes of theorising (with) 'history': at the unit level, where history underlies existing intervening variables, or systemically, where history affects the international environment states operate in.
Both routes may traverse multiple different conceptualisations of history based in the wider philosophy of history -as objective fact, as individual experience or as collectively shared and intersubjectively narrated. This emancipates neoclassical realism from the narrower understanding of history still prevalent in neorealism. And both routes produce theoretical and methodological challenges (at different levels of magnitude) respectively for NCR. In particular, they pose difficult questions in terms NCR's theoretical boundaries and distinctiveness vis-à-vis classical realism, constructivism, historical institutionalism and analytically eclectic approaches. However, they also enable neoclassical realists to explicitly interrogate the role of history in the dynamics of decisionmaking and strategic interaction, and thus contribute, should they so desire, to contextualising the ways states react to systemic constraints and in turn produce international outcomes. In addition, historicising NCR presents an opportunity for scholars to contribute new empirical puzzles, questions and concepts. Collaboration that complements a (neoclassical) realist framework with regional and historical expertise has beneficial methodological as well as more broadly disciplinary consequences.
Finally, thinking about ways in which to operationalise 'history', however conceived, within loosely positivist frameworks such as NCR has benefits for other IR frameworks, including 'thin' constructivism and historical institutionalism. For one, it helps clarify different understandings of what history is and does, and how it informs ways of explaining international politics. In turn, investigating increasingly blurry paradigmatic boundaries allows for a dialog characterised less by epistemological posturing or enforced synthesis, and more by inter-paradigmatic and even interdisciplinary collaboration. This can help pave the way towards a historically literate, systematic study of state behaviour and international relations.