Kenneth Waltz is not a neorealist (and why that matters)

Introduction

Any theory covers some matters and leaves other matters aside. ¹

Faced with scepticism about the status of grand theory in International Relations (IR), scholars are re-evaluating Kenneth Waltz’s contribution to theoretical debates in the field. Readers of Waltz have variously recast his work as structural functionalist, scientific realist and classical realist in liberal clothing. ² We contribute to this re-evaluation by systematically assembling misreadings of Waltz that continue to occur across all of IR’s schools — that his theory is positivist, rationalist and materialist — and offering a synthetic account of his main contributions to IR theory. By linking Theory of International Politics (TIP) to Man, the State, and War (MSW) and his other writings, we show that Waltz offers IR scholars a coherent vision of the worth and method of grand theory construction that is uniquely ‘international’. ³ Aiming at more than exegesis, we argue that scholars should revisit Waltz’s approach to theory construction. By embracing a practical, perspectival approach in his ‘images’, and forwarding a theory of international politics that is more attentive to social construction than is commonly recognised, Waltz offers a role for theory that is relevant to present-day discussions — particularly on how normative concerns can guide grand theorising. The theoretical horizon of Waltz’s scholarship, often understood to be transcended, has therefore yet to be fully explored. ⁴

We proceed in five sections. In our first, we present and refute each of the three ‘misreadings’ with reference to Waltz’s own explicit statements. By his own account, Waltz shares none of the core assumptions of the dominant neorealist school of IR that followed in his wake — despite being labelled its ‘father’. ⁵

In the second section, we argue that misreadings of Waltz may stem from his sometimes opaque language. Waltz uses positivist language to explain his theory in TIP, despite ultimately claiming that positivism is inappropriate for a theory of international politics. Although Waltz is not a positivist and rejects predictive theory testing, he does provide some basis for those who mistakenly treat him as one. In our third and fourth sections, we clarify Waltz’s theory of international politics by returning to MSW’s grounding in political theory and methodology of ‘images’. We argue that MSW and TIP are part of the same theoretical project: the concepts developed in the former continue in the latter. This understanding of Waltz unifies his work into a coherent approach to world political theorising. ⁶ Unlike existing readings of Waltz, ours emphasises the social content of his approach. ⁷

In our fifth and final section, we propose three benefits of our re-examination. First, Waltz’s understanding of grand theory construction can be both instructive and practical. Researchers can benefit from Waltz’s example by scrutinising how he grounds theoretical inquiries in political theory, ensuring that even at an abstract level, they remain connected to pressing practical concerns. Second, Waltz’s ‘images’ offer an underappreciated and effective framework for the causal role of ideas by emphasising the perspective of both subject and scholar. They are therefore different than ‘levels of analysis’ — the empirical division of politics into distinct individual, domestic and international interactions. Third, getting Waltz right has pedagogical value for graduate training in IR. By foregrounding the link between normative and explanatory theory, Waltz shows how an issue (the possibility of peace) can be related to a set of methodological needs and analytical commitments. This offers an unusually synthetic vision of what makes world politics — and the study thereof — special.

Waltz: Lost in translation

Misrepresentations of prominent thinkers are not unusual in IR (Welch, 2003). What is unusual about the misrepresentations — or misreadings — of Waltz is the unprecedentedly important position that he occupies in the discipline. Most of the major contributions to IR since TIP’s publication have begun from its premises, seeking either to improve their explanatory power or to criticise them. As Ken Booth (2011: 5) notes, TIP has become the ‘magnetic north by which students of relations between states navigated their different journeys to explanation and understanding’.

We argue that, with a few exceptions, IR scholars have misrepresented TIP’s premises, building their own contributions on an inaccurate account of Waltz’s theory of international politics. These misrepresentations fall into three categories: those that construe him as a positivist; those that contend that his neorealism needs rational actors; and those that criticise his conception of capabilities as crudely materialist. These are the ‘core assumptions’ of neorealism, which form the basis of a large research paradigm that has spawned several offshoot schools. ⁸ Mirroring the profound influence of TIP on IR at large, these misrepresentations appear in all of its research paradigms, crop up in contributions that are critical of mainstream approaches and are repeated in introductory IR textbooks. In his many acerbic responses to these readings, however, Waltz has denied that he or his writings are positivist, rational-choice or materialist. Waltz’s own approach to world politics — by his own presentation — appears to share none of the core assumptions of his supposedly chief contribution to IR, neorealism.

In the last decade, a growing number of scholars have pushed back against some of these misreadings: Jackson (2011: 114) argues from a philosophy of science standpoint that TIP is better characterised as ‘monist’ than positivist; Waever (2011) and Polansky (2016) show how TIP employs Weberian or Hobbesian ideal-types; Buzan and Albert (2010) have corrected Waltz’s misapplication of Durkheim’s ‘differentiation’; and Bessner and Guilhot (2015) argue that TIP recasts classical realism in liberal clothing. Rosenberg (2013, 2016) has suggested that Waltz proposes a chastened theory of ‘the international’ that is bound too closely to political science categories. While we are indebted to these analyses and build on them here, to our knowledge, misrepresentations of Waltz — which remain prodigious — have not been systematically compiled and examined anywhere, nor studiously contrasted with his own statements. ⁹ We take up this task in the following.

Why Waltz’s readers are confused

The preceding discussion raises a question: why do many IR scholars make claims about Waltz that are at odds with Waltz’s own explicit remarks? One possible explanation is that the dominance of positivist methods in American IR, and the institutional status of IR in North America as a subfield of political science, ²⁵ has obliged almost everyone in it to understand Waltz through a positivist lens. In other words, Waltz’s theory is understood as a collection of predictive, law-like relationships between independent and dependent variables either because post-TIP scholars lack the vocabulary to discuss Waltz in non-positivist terms or because they have ignored Waltz’s non-positivist core. ²⁶ Waltz (1982: 680–681) suggested this himself in response to one review of TIP.

Waltz’s misrepresentations may also stem from his methodological ambiguity, however. Waltz (1979: 5–6) does not state a consistent position himself about the relationship between his theory and the empirical search to confirm or disconfirm predictions derived from it (Buzan et al., 1993). On the one hand, he is adamant that law-like statements relating variables to one another do not constitute theories: ‘Rather than being mere collections of laws, theories are statements that explain them … Theories are qualitatively different from laws. Laws identify invariant or probable associations. Theories show why those associations obtain’ (Waltz, 1979: 5). Waltz clarifies this further:

Of purported laws, we ask: ‘Are they true?’ Of theories, we ask: ‘How great is their explanatory power?’ … Laws are ‘facts of observation’; theories are ‘speculative processes introduced to explain them’. Experimental results are permanent; theories, however well supported, may not last … Laws remain, theories come and go. (Waltz, 1979: 6)

On the other hand, Waltz ends TIP’s first chapter — which sets forth his methodology — with the seemingly self-contradictory claim that theories are validated through hypothesis testing:

In order to test a theory, one must do the following: 1 State the theory being tested. 2 Infer hypotheses from it. 3 Subject the hypotheses to experimental or observational tests. 4 In taking steps two and three, use the definitions of terms found in the theory being tested. 5 Eliminate or control perturbing variables not included in the theory under test. 6 Devise a number of distinct and demanding tests. 7 If a test is not passed, ask whether the theory flunks completely, needs repair and restatement, or requires a narrowing of the scope of its explanatory claims. (Waltz, 1979: 13)

This view of theory testing, which philosophers of science call ‘hypothetico-deduction’, depends on a ‘nomological’ or ‘covering-law’ view of theory (Jackson, 2011: 41–71). One explanation is that, for Waltz, the value of a theory or ‘model’ is its ability to generate accurate predictions, regardless of whether it contains unrealistic or absurd assumptions. Such a view is at odds with Waltz’s remarks elsewhere, however: ‘In one sense a model represents a theory. In another sense a model pictures reality while simplifying it, say, through omission or through reduction of scale. If such a model departs too far from reality, it becomes useless’ (Waltz, 1979: 7, emphasis added). In other words, Waltz first claims that theory exists to provide ‘explanatory value’ rather than generate predictions, but then identifies prediction as the test of a good theory. This apparent inconsistency in methodology may explain some of the confusion among Waltz’s readers. As we argue later, there is a way to resolve it.

Second, Waltz also does not directly define what he means by ‘explanatory value’. While he often asserts that the explanatory value of his theory lies neither in a theory of foreign policy nor with predictive laws, he does not offer a clear definition of what his theory should do (Waltz, 1979: 122; 1982: 680; 1988: 618; 1996: 57; 1997: 916; 2000: 27). On the one hand, he claims that theory should be about locating the most relevant details of a certain ‘causal realm’ — of a domain of phenomena with distinct, irreducible features. In the first chapter of TIP, Waltz (1979: 8) writes:

A theory is a picture, mentally formed, of a bounded realm or domain of activity. A theory is a depiction of the organization of a domain and of the connections among its parts … A theory indicates that some factors are more important than others and specifies relations among them … Theory isolates one realm from all others in order to deal with it intellectually.

However, he does not explain why we seek to do this, leaving unsaid what function theories are supposed to perform for the theorist, and thus how we might evaluate our own work. In fact, the only explicit statement that Waltz makes on the function of theory in TIP beyond the term ‘explanatory value’ refers to its predictive power. In the fourth chapter of TIP, he writes:

A theory has explanatory and predictive power. A theory also has elegance. Elegance in social-science theories means that explanations and predictions will be general … It tells one what to expect and why to expect it. Within a system, a theory explains recurrences and repetitions, not change. (Waltz, 1979: 69, emphasis added)

This has clearly implied to many readers, against Waltz’s expressed designs, that theories should accurately predict historical patterns over the longue durée and relate those patterns back to some specific set of causal relations — exaggerated for the purposes of the theory, but ultimately based on something real. As noted, Waltz firmly opposes this elsewhere. However, it does leave readers of TIP with two puzzles. First, it is not clear at what level of historical granularity prediction is supposed to lie such that a failure of prediction would serve as a disconfirmation as opposed to a simple misapplication of the theory. Second, it is not clear how ‘explanatory power’ is to be measured as something distinct from prediction in the first place because it is never given specific definition, instead appearing together with, even if nominally distinguished from, prediction. Both of these mysteries arguably stem from the absence in TIP of a clear research question or problem that would orient possible answers.

Third, and compounding the preceding points, Waltz draws heavily on the language of rationalism but deviates from rationalist assumptions in crucial ways. At a number of points throughout TIP, Waltz (1979: 71, 55–56, 105) claims to be offering an IR analogue to a theory of the firm and to oligopolistic competition. However, Waltz is not interested in a model that solely generates accurate predictions, as previously discussed, since he also wants his theory to offer a recognisable image of reality. Moreover, Waltz does not assume that actors are instrumentally rational — meaning that actors in his theory need not seek efficient means to ends or have a consistent set of preferences (Waltz, 1979: 118, 122; 1986: 330). Instead, Waltz (1979: 74) argues that the structure of the international system disposes actors to be rational in this way: ‘Structure affects behavior within the system … through socialization of the actors and through competition among them’. Indeed, Waltz (1979: 77, emphasis added) seems to explicitly recognise a culturally or socially constructed quality to the way in which structures influence agents: ‘Competition spurs the actors to accommodate their ways to the socially most acceptable and successful practices’. The significance of this deviation from rationalism should not be overlooked. Waltz’s theory, rather than assuming actor rationality, instead seeks to explain actor rationality as a historically evident tendency.

Waltz’s interventions into the field are thus very different from those of neorealism as normally understood. Neorealism is, as Legro and Moravcsik rightly deduce, positivist, rationalist and materialist. It is unsurprising that many IR scholars make inaccurate statements about TIP because, in that book, Waltz leaves opaque or unclear many key premises on the nature of his theory himself. ²⁷ The spectacular success of TIP is thus all the more surprising. A clearer reading of Waltz may not have animated as vibrant a set of conversations as it did. As we will show in the following, however, a close reading of MSW and TIP together reveals a Waltz who is more coherent than suggested by either the alternative explanations or the neorealist school that grew in his wake.

Man, the state and Waltz’s images

Waltz’s division of international relations into three ‘images’ in MSW has become an enduring way of thinking in IR. Unfortunately, what Waltz means by image appears to be frequently misunderstood: their perspectival character is conflated with real or empirically distinguishable loci of causal action, typified by ‘levels-of-analysis’ approaches (e.g. Buzan, 1995; Buzan et al., 1993: 22; Senese and Vasquez, 2008). For Waltz, images are heuristic standpoints: they focus our attention on some phenomena that we would not otherwise see clearly but, in the process, necessarily obscure other phenomena. The three images depict differently oriented worlds and anchor different causal accounts of world politics — the featured processes that occur within them focus our attention on different kinds of structures or actors.

Indeed, the importance of the difference between ‘levels of analysis’ and images for Waltz’s project cannot be overstated. The division of politics into levels of analysis is a (philosophical) realist or empiricist attempt to refer to specific causal relationships within a layered institutional arrangement. According to levels of analysis, the international differs from the domestic because it contains relationships between sovereigns under conditions of anarchy. A given political outcome either is or is not due to such relationships, and does or does not belong in the international level. Waltz’s images are ideal-types: they are simplified world-pictures that portray particular perspectives on politics. They do not correspond to the world and are not confirmed or disconfirmed on a case-by-case basis. Instead, they clarify causal dynamics that should be treated as always present, and that, in the third image, define the international as such, even if these dynamics are sometimes invisible in our empirically messy reality. The international structure — as an observer-independent reality — defines the level of analysis; the image, conversely, defines the ‘international’. Waltz elsewhere connects images with ‘theory’ in TIP; understanding the former — and the importance of the ‘third image’ to world politics — is thus crucial for making sense of the latter.

An ‘images’ approach to world politics operates by placing our vantage point — our ability to see a ‘nexus of important causes’ — in a particular place, where it necessarily distorts other loci (Waltz, 1959: 12, 227, 230). This move is necessary because, according to Waltz, world politics, in all their empirical complexity, cannot be understood all at once, and clear theories cannot result from attempts to do so. The process of ‘imaging’ is a necessary first step for clearly theorising any one realm of world political activity. In the 2001 preface to MSW, Waltz (1959: ix) elaborates the connection between images and theory — and, by extension, between MSW and TIP — as such:

The word ‘image’ suggests that one forms a picture in the mind; it suggests that one views the world in a certain way.^[28] ‘Image’ is an apt term both because one cannot ‘see’ international politics directly, no matter how hard one looks, and because developing a theory requires one to depict a pertinent realm of activity. To say ‘image’ also suggests that in order to explain international outcomes one has to filter some elements out of one’s view in order to concentrate on the presumably fundamental ones. In relating the first and second images to the third, I viewed the third image as ‘the framework of state action’ and ‘as a theory of the conditioning effects of the state system itself’. Explaining international outcomes requires one to examine the situations of states, as well as their individual characteristics.

What I then called ‘the state system’, I later defined more precisely as the structure of the international political system.

Images are not identical to theories; they instead lay the ‘foundations’ for them. An image may give rise to theories that propose contradictory propositions: ‘there is no one prescription for each image’ (Waltz, 1959: 13). Images are the presuppositions upon which theories rest.

MSW did not propose a theory of international politics, but it did clarify what would count as one. It would place the ‘locus of causes’ of international outcomes in the third image alone: ‘a theory of the conditioning effects of the state system itself’ (Waltz, 1959: 231). ²⁹ A theory of ‘international politics’ is thus not a theory of foreign policy, which would proceed from the unit-level characteristics of the second image. A theory of ‘international politics’ views its whole realm of observation from within the perspective of the system taken by itself (‘international politics’), absent the perspectives of the states’ internal characteristics or the nature and behaviour of man. The strength of this view is to be found in its ability to explain international outcomes in ways that the other two images cannot; it uniquely locates important causes within the structure of the international system, to the necessary exclusion of the other images (Waltz, 1959: ix, 230–232).

The main argument of MSW rests on the importance of the third image to an overall understanding of world politics — a field of political activity much greater than, but which encompasses, international politics. ³⁰ First- and second-image theories can, according to Waltz, grant insight into the outbreak of some wars between states. ³¹ They thus have bearing on world politics but, as Waltz argues throughout MSW, we are unable to explain on those bases alone the persistence of war over time, or the effects of the state system on it. A world politics full of ‘good’ or ‘reformed’ people and ‘good’ or ‘reformed’ states, even if we could achieve some clarity about those conditions and then realise them on earth, would still face the prospect of war because of the distinct dynamics of international politics (Waltz, 1959: 51, 72–79, 83ff, 108, 112–114, 125, 218–219). Second-image arguments, for example, hold that because the ‘internal conditions’ of states ‘determine external behaviour’, ‘defects in states cause war among them’ (Waltz, 1959: 83). Even if people or states could be perfected, however, such perfectibility would be insufficient to eliminate war:

Solutions for the problem of war based upon the pattern of either the first or the second image must assume the possibility of perfection in the conflicting units. Perfection being impossible for states as for men, the liberal system can at most produce an approximation to world peace. … if conflict arises not only from defects in the subjects but also from the quality of the relations among them, it may be that no amount of improvement in the individual subjects would be sufficient to produce harmony in anarchy. (Waltz, 1959: 119, emphasis added)

The third image thus captures an essentially inescapable dimension of world politics: the ‘international’, understood as a locus of causes independent of the other images. We can better explain war by paying ‘constant attention to the external pressures’ to which states are subject — pressures that the third image makes ‘clear’ (Waltz, 1959: 218, 222). This is because war is a constant possibility ‘in a world in which there are two or more states each seeking to promote a set of interests and having no agency above them upon which they can rely for protection’ (Waltz, 1959: 227). The existence of anarchy as the structural condition between states means that war is a permanent possibility, regardless of changes in the other two images. This is summed up in the ‘proposition that wars occur because there is nothing to prevent them’ (Waltz, 1959: 232). The move to a third-image view thus fully shows us that war will still happen if world politics is populated by ‘good’ people and/or ‘good’ states, ‘so long as we operate within a nation-state system’ (Waltz, 1959: 230–231). It does so by focusing our attentions on the heretofore neglected ‘conditioning effects of the state system itself’ (Waltz, 1959: 231).

MSW does not rest all of its explanatory weight on the third image. Mirroring its dialectical structure, MSW encourages the clarification of each image in isolation and a comparison of their explanations (Waltz, 1959: 230). Theories from different images can (and should) be compared in assessing ‘the causal weight of forces identified by one or another of the three images’ (Waltz, 1959: xi). This is because an image taken alone will tend to ‘distort’ the others:

all three images are a part of nature. So fundamental are man, the state, and the state system in any attempt to understand international relations that seldom does an analyst, however wedded to one image, entirely overlook the other two. Still, emphasis on one image may distort one’s interpretation of the others. (Waltz, 1959: 161)

While isolating the third image is necessary for understanding the persistence of war, it is not sufficient to understand world politics. The third image may inadequately explain outcomes that require the ‘help’ of the other images: ‘some combination’ of them may be ‘required for an accurate understanding’ of world politics (Waltz, 1959: 14). Waltz’s (1967) longest book, though not a ‘theory of foreign policy’ in the style of TIP, represents his most thorough effort to explain world politics from a second-image perspective. The need for other images is especially true of individual instances of war, where a systemic approach must remain silent on the precise reasons that one state attacks another. The structure of world politics does not ‘directly cause state A to attack state B’ and its explanation ‘does not hinge on accidental causes’ (Waltz, 1959: 231–232). It instead points out that ‘any accident can bring about a war’ (Waltz, 1959: 231). Scholars explaining a war must therefore ‘look to motivation and circumstance in order to explain individual acts’: ‘location, size, power, interest, type of government, past history and tradition — each of which will influence the actions of both states’ (Waltz, 1959: 232). Some states, Waltz notes, will be especially war-prone ‘by virtue of their internal conditions’ (Waltz, 1959: 232) — in other words, their regime type — a variable thought to not matter in Waltzian neorealism. Indeed, according to Waltz (1959: 231–232, emphasis added):

These special reasons become the immediate, or efficient, causes of war. These immediate causes of war are contained in the first and second images.… Variations in [them] are important, indeed crucial, in the making and breaking of periods of peace — the immediate causes of every war must be either the acts of individuals or the acts of states.

The preceding discussion demonstrates how images differ from levels of analysis that empirically separate world politics into domestic and systemic levels and seek the causes of war in one or the other. Images clarify a unique set of causal relationships, and can be compared heuristically to explain world political situations — as Waltz did in his empirical work — but do not alone explain world politics. ³²

Following the publications of MSW and TIP, Waltz consistently restated these points: a third-image theory does not explain individual instances of war or their causal mechanisms, as neorealists are wont to do. ³³ As Waltz (1959: ix) states in the 2001 preface to MSW, the book, ‘strictly speaking … did not present a theory of international politics’. This task was left to TIP. MSW ‘did, however, lay the foundation for one’ by clarifying the purpose of images, and how they are intended to interact: comparatively, but discretely.

Three noteworthy themes emerge from MSW. First, while TIP’s major research question is opaque, MSW’s is not. In MSW, Waltz (1959: 1) opens by asking a clear question: ‘are there ways of decreasing incidence of war, of increasing the chances of peace?’ This grounds MSW in a normative question and draws our attention to the variety of extant answers, provided chiefly by the history of political philosophy. Waltz thus encourages us to compare his treatment of the problem of war with those of the political philosophers who precede him (Hassner, 1997: 25–26; Pangle and Ahrensdorf, 1999: 240, 257). Second, Waltz’s own answer to the question — the ‘third image’ — tells us how relations between states are unlike other kinds of social relations, and thus directs readers to another, though more strictly analytical, question: ‘What defines international politics as such?’ Third, and last, Waltz’s methodological comments in MSW are simpler, because more simply put, than those in TIP. By using the perspectival language of ‘images’, reminiscent of Weber’s ideal-types, Waltz indicates the non-predictive character of his approach. We turn to these themes in the following.

How to read Kenneth Waltz

The preceding account of MSW clarifies Waltz’s structural theory: TIP is a continuation of the project that Waltz began in MSW, rather than a deviation from or revision of it. Waltz’s views on theorising international politics, as something distinct from foreign policy (and as only part of world politics in general), carry forward from MSW. However, the lack of a clear research problem in TIP makes it hard to determine what, other than predictive success, would establish his theory’s explanatory power — unless TIP is treated as a sequel to MSW, as we do, whereupon it becomes clearer.

Two key research questions that Waltz addresses in MSW help clarify TIP. The first asks: ‘Can we have perpetual peace in the future?’ — to which Waltz answers ‘no’ because of the dynamics of the third image. The second asks: ‘What constitutes international politics as such?’, which Waltz answers by separating the images and considering how they interact. By recognising how these questions also underpin TIP, parts of the later text become less confusing or stand out as more important.

First, Waltz signals in TIP that what a structural theory of international politics actually entails is as important as any specific proposition about polarity or balancing. This stands in marked contrast to much neorealist scholarship. As noted earlier, subsequent neorealism examines the specific outcomes of world politics, and has little in the way of ‘meta-theory’ — arguments about theory construction. ³⁴ At several points throughout TIP, Waltz grapples with the question of how structures exist as independent social things, possessing their own autonomous power to influence state behaviour. In one instructive place, he writes:

[S]tructures limit and mold agents and agencies and point them in ways that tend toward a common quality of outcomes even though the efforts and aims of agents and agencies vary. Structures do not work their effects directly. Structures do not act as agents and agencies do. How then can structural forces be understood? How can one think of structural causes as being more than vague social propensities or ill-defined political tendencies? Agents and agencies act; systems as wholes do not. But the actions of agents and agencies are affected by the system’s structure … The effects are produced in two ways: through socialization of the actors and through competition among them. (Waltz, 1979: 74)

As Goddard and Nexon (2005) argue, this view of structure is associated with ‘structural functionalism’: social systems condition agents, through the inculcation of particular ‘cultural norms’, to engage in actions that maintain those systems. ³⁵ That Waltz devotes the first two chapters of TIP and many other passages to these meta-theoretical issues makes sense if TIP’s aim, even its primary aim, is to define what makes the study of international politics unique — to carve out a distinct space for ‘international relations’ by separating it from the study of foreign policy and to justify this separation on the basis of theory. ³⁶ This suggests that Waltz sought to establish the conditions for future IR scholars to continue that project on solid meta-theoretical footing, rather than offer a specific and complete explanation of historical patterns in state behaviour. ³⁷ Indeed, Waltz’s post-1979 statements about TIP consistently centre on his definition of ‘theory’, and the subsequent elucidation of a ‘theory of international politics’, in precisely these terms (e.g. Waltz, 1986, 1996, 1997, 2003).

Second, as a further attempt to address the key questions of MSW, TIP rests on a social-constructivist view of structure that has been ignored in rationalist readings that emphasise Waltz’s economic metaphors. In explaining how structures cause world political outcomes, Waltz (1979: 76–77, emphases added) clearly identifies something more than preference maximisation:

The first way in which structures work their effects is through a process of socialization that limits and molds behavior. The second way is through competition … Socialization encourages similarities of attributes and of behavior. So does competition. Competition generates an order, the units of which adjust their relations through their autonomous decisions and acts … Competition spurs the actors to accommodate their ways to the socially most acceptable and successful practices.

As this quote shows, Waltz identifies a key ‘cultural’ ³⁸ or social dimension: the presence of norms of acceptable behaviour both oblige other actors to learn those norms and point to a repository of cultural scripts and practices that actors must learn in order to survive. By remaining agnostic about the motives behind states’ adherence to ‘socially most acceptable’ practices, Waltz actually anticipates recent work by critical constructivists that makes a similar point (Krebs and Jackson, 2007). Moreover, Waltz forwards a conception of the co-constitution of structure and agency in IR nearly a decade before Wendt (1987) did the same. This demonstrates further how Waltz intends to develop not just an elegant theory of international politics, but a new way for IR scholars to theorise politics in general. ³⁹

Finally, readers of TIP can now evaluate the success or failure of Waltz’s structural theory. A full evaluation is beyond the scope of this article, but we can offer an overview of what would qualify. We argue, along with Goddard and Nexon (2005) and Jackson (2011: 149–151), that Waltz’s theory stands or falls on matters of conceptual coherence far more than on empirical accuracy, consistent with Waltz’s own statements across several decades (e.g. Waltz, 1982: 681; 1997: 916; 2003). This is because Waltz (1979: 80) views structure purely conceptually in the first place: ‘Since structure is an abstraction, it cannot be defined by enumerating material characteristics of the system. It must instead be defined by the arrangement of the system’s parts and by the principle of that arrangement’. ⁴⁰ While Waltz (e.g. 1979: 127) professes some desire to explain broad-scale historical patterns, his critics would need to adduce a leviathan of empirical evidence indicating that international politics bear no resemblance to Waltz’s depiction. Rather, theory evaluation should focus on whether Waltz’s now-familiar formula of ordering principle, unit differentiation and distribution of capabilities allows us to effectively describe international politics as a unique causal domain distinct from the particular preferences and decisions of individual national communities (e.g. Deudney, 2011; Polansky, 2016).

This sort of evaluation has been rare so far, we argue, because IR scholars — especially mainstream, American ones — have tended to either misread or ignore the reason Waltz separates the international from ‘unit-level’ characteristics such as foreign policy, conflating the former with ‘world politics’ broadly understood or adopting a levels-of-analysis approach that mashes together international with domestic and thus reduces one to the other. Read alongside MSW, TIP’s limits and uses become clear: it is a theory of international politics that can be applied to world politics in a general sense, but which will often give incomplete explanations, particularly of single cases — hence Waltz’s consistent statements to this effect. ⁴¹ The image theorising in MSW thus clarifies what Goddard and Nexon (2005: 24) note about TIP: Waltz’s ‘international system … is an analytic category designed to capture single, irreducible facets of what is, in reality, multifaceted action and order’.

What to do with (properly read) Waltz

Reading Waltz as a meta-theorist or social constructivist is valuable for thinking about world politics in three major ways. First, Waltz offers methodological guidance by his example as a grand theorist in the methodological sense: he defines the fundamental problems of the field and establishes the scope and terms of scholarly attempts to solve those problems, weaving together basic theoretical commitments and general propositions about what generates international politics as such. As some have recently noted, this kind of grand theorising is increasingly neglected in mainstream American IR in favour of ‘mid-range’ hypothesis testing, a neglect due in large part to scholars’ hostility towards paradigmatic thinking, now synonymous with fruitless methodological debate (Mearsheimer and Walt, 2013). Waltz demonstrates how — and when — this hostility is unfair. By orienting his project around problems of political theory first and foremost, he draws on political theory in answering enduring questions of value and possibility as they pertain to the conduct, and investigation, of all politics (Pangle and Ahrensdorf, 1999: 257).

This problem-centred approach does not take the form of an analytically promiscuous eclecticism, but rather of systematic and holistic theoretical world-building. The lesson he thus offers is on the methodological foundations of grand theory construction. Waltz shows how starting with political theory gives IR scholars a basis for thinking big without concurrently building paradigmatic silos, as well as the methodological virtues of starting with political-theoretic problems, rather than philosophical boundaries.

Second, Waltz’s images offer practical guidance for analysing cultural or ideational variables that do not neatly fit within established boundaries between the personal, domestic political and international. Properly understood, an image is a perspective on the field of world politics that emphasises a particular arrangement of causal interactions, focusing on them specifically to clarify what makes them distinct. One implication of this is that notable uses of Waltz’s concept of ‘image’ miss its ideal-typical qualities. For example, Gourevitch’s (1978) famous ‘second-image reversed’ focuses not on which causal orientation appears most important, but on specific causal mechanisms through which international politics influence domestic politics. This conflates images with levels of analysis — the empirical isolation of cause–effect processes inside their institutional positions — and erases the perspectival and ideal-typical character of Waltz’s approach. Conversely, a proper use of images shows that prominent levels-of-analysis approaches fail to adequately represent what distinguishes international politics as such. For example, Putnam’s (1988) ‘two-level’ game connects international political outcomes to domestic political factors by showing how foreign policy decisions depend on domestic ratification. Putnam’s explanation is thoroughly third image, however, because it does not include variation in possible domestic institutional arrangements. It instead notes how local audiences contribute causally to international political outcomes, that is, when people act as states (and hence are subject to international mechanisms of socialisation and competition). It thus belongs in the third image.

Waltz consistently employed his images to both show why theories developed in one perspective should not be mixed with others, and demonstrate how images heuristically clarify world political scenarios for analysts and policymakers. Waltz argues that the close competition of bipolarity, for example, tends to clarify capabilities and security interests. However, polarity alone does not determine state behaviour, nor does it guarantee that states will act on their security interests (e.g. Waltz, 1964: 904–906). Indeed, Waltz’s various criticisms of US foreign policy, written before and after the Cold War, would be misguided if the third image determined state behaviour. Rather, he uses the third image to describe the structural situation faced by the US and clarify what a prudent American foreign policy could look like. ⁴² The third image thus has a practical, heuristic role for understanding the possibilities of action, but does not causally explain why a specific action occurs. Elsewhere, Waltz evaluated foreign policy from within the second image alone, comparing policymakers’ goals with their results (Waltz, 1967). ⁴³

These examples show the flexibility of an images approach: images do not align neatly with empirical domains, but instead can encompass transnational or subnational forces so long as these are relevant to a given nexus of causes. Hence, institutions not normally included in the ‘international’, such as diplomatic academies or other sources of ‘strategic culture’, can be placed within a third-image context and studied as part of the system that selects for certain kinds of state behaviours and reinforces anarchic structural conditions. Once again, Waltz’s approach is more sympathetic to constructivist research interests than is normally assumed, in a way that can inform future realist theorising at the grand or macro scope.

Third, and finally, approaching Waltz as we suggest offers pedagogical benefits beyond ‘getting him right’. IR survey courses typically start with Waltz and then continue through the ‘isms’ — neorealism, neoliberalism, then constructivism — as a way of presenting the fundamental debates of the discipline. TIP is the most widely assigned publication in doctoral IR training seminars at US universities (Colgan, 2016: 7). Yet, what is taught and what is published is disconnected, with ‘middle-range’ non-paradigmatic research comprising the bulk of top journal publications, and with the largest portion of IR scholars globally refusing to identify with any paradigm at all. ⁴⁴ Part of why this is, we argue, is because realism and liberalism are not taught as philosophically meaningful contemporary approaches in IR theory. An awareness of the philosophical roots of Waltz would better educate future IR scholars about the links between normative and explanatory theory in realism and liberalism. Waltz’s project is instructive about the ends of IR theory, and about the essential problems of political theory that orient realism and liberalism alike. MSW and TIP, when read in tandem, relate the possibility of peace to a set of methodological needs and analytical commitments, together constituting an unusually synthetic view of what makes international politics — and the study thereof — special. If students of world politics are to begin their surveys of the field with Waltz, let them grapple with the full scope of Waltz’s ambitions and puzzles. Doing so will lead to a greater appreciation of why and how the study of world politics exists in the first place.

Conclusions

We have argued that linking TIP with MSW clarifies both texts’ aims, Waltz’s many post-TIP rejoinders and Waltz’s own contribution to IR theory. This reading separates Waltz’s theory from the neorealism that grew in his wake, and the consequent debates that have occupied much of academic IR since TIP’s publication. We have also argued that such an understanding is useful not just to scholars interested in ‘getting Waltz right’, but also for those who research world politics writ large. By highlighting the normative and meta-theoretic concerns that Waltz emphasised in his writings, including (and perhaps foremost) in TIP, we show how revisiting Waltz clarifies the methodological and pedagogical worth of grand theorising in IR.

We do not intend to co-opt Waltz to the side of an ‘ism’, refashioning TIP into ammunition for a paradigm war. Indeed, many of the premises of these wars — the sharp exclusion of ideas or culture in favour of materialism or rationalism, for example — cannot be found in Waltz. Despite its ‘realism’, Waltz’s theory accommodates the variety of desires, hopes and goals that animate world politics by avoiding rationalist assumptions and allowing a variety of state behaviours to coexist — even if the theory has less to say, by design, about specific events.

For the same reasons, Waltz shows us how to build ‘grand theory’ in a post-paradigmatic IR desired by some scholars. MSW is framed by a pressing normative concern: the effects of war on human life and the ‘apparent disproportion’ between humankind’s ‘peace wish’ to eliminate war and the ‘fleeting’ results of those efforts. It analyses both causal explanations and the prescriptions of thinkers across the modern era (Waltz, 1959: 2–12). The normative concerns of MSW thus justify its methodological commitments. TIP also draws the reader’s attention to the normative character of political life and the limits that the international system imposes on concerns such as order and justice (e.g. Waltz, 1979: 111–114). Perhaps most crucially, Waltz establishes a distinct basis for IR as a discipline: it is not that international politics are empirically separate from other kinds, but that enquiries into them require a unique perspective — one that should not be forced together with perspectives on other domains of social life — hence Waltz’s (1959: ix) preference for images over ‘levels of analysis’.

The first aim of this article has been to save Waltz from his sectarian descendants. However, revisiting Waltz should also include critical evaluations of his project — so that the reading we advance is not the last Waltz. Here, we return to the problem of how to test Waltz’s theory of international politics. If prediction is out, and Waltz is so heavily invested in articulating a social theory and methodological orientation, how should we evaluate his project? While Waltz does make claims concerning patterns of state behaviour across history, we argue that simply ascertaining whether or not these patterns obtain is not an appropriate test of his work. More important than the specific empirical content of his theory is the programmatic fertility of his approach: whether it offers a sound basis for maintaining IR as a research tradition in possession of distinct and tractable objects of study. TIP is the face, as Marlowe might have written, that launched a thousand (author)ships — for better or worse. However, as we have shown, there are features of Waltz’s work that have not received due appreciation from most scholars in the field. Recent re-evaluations of Waltz — including, we hope, this one — promise a fuller picture of whether his emergent, perspectival, ideal-typical approach to grand theory still has a place in current discussions. Our odds are that it does.