The epistemology of international politics: offensive realism and the Neorealist Scientific Research Program

Abstract

What follows is a much-needed appraisal of the current state of scientificity and progressiveness in the study of international politics. Imre Lakatos’ Methodology of Scientific Research Programs will provide the epistemological grounds. Three broad claims will be made and supported: (i) Waltz’s neorealism is not an isolated theory of international politics, but a proper scientific research program; (ii) Mearsheimer’s offensive realism too is not an isolated theory, but a set of theoretical amendments to the Neorealist Scientific Research Program’s protective belt; and (iii) the offensive-realist theoryshift is both theoretically and empirically progressive.

Epistemology; Neorealism; Offensive realism; Research programs; Science

Introduction

Any worthy scientific field of research must periodically assess the extent of theoretical and empirical progress put forth by its relevant body of work. Yet, this is rarely done by International Relations (IR) scholars, despite the field’s decades-long claim to scientific status. This article aims to partially rectify this situation by addressing, in precise and rigorous epistemological terms, IR’s arguably most successful theoretical output – Kenneth Waltz’s neorealism – and one of the most influential and debated theories in recent years – John Mearsheimer’s offensive realism. It will be argued that the full extent of Waltz’s contribution to IR theory can only be appreciated not as an isolated theory of international politics, but as a proper scientific research program (SRP). It will also be demonstrated that Mearsheimer’s offensive realism is not an independent, alternative theory, but instead a development within the Neorealist Scientific Research Program (NSRP), which in turn can be seen as both theoretically and empirically progressive.

Imre Lakatos’ Methodology of Scientific Research Programs (MSRP) is the epistemological anchor for the following analysis. While a handful of alternative epistemologies were also available, Lakatos’ seems particularly useful for appraising developments within IR theory. As noted by Elman and Elman (2003), IR scholars are inclined to view their own work, as well as their rivals’, not as isolated theories, but as families of theories that share some basic premises or assumptions. Also, there is a decent level of tolerance for the coexistence of multiple, competing theories, as well as considerable tenacity in defending each scholar’s preferable one. All of these, of course, are fairly consistent with Lakatos’ depiction of the scientific process (Lakatos 1970a).

This article’s endeavor is warranted not only due to the neglect of epistemological self-reflection by IR scholars, but also because some of the most notorious attempts have been less than adequate. When Vasquez (1997) sought to demolish neorealism by depicting it as a supposedly degenerating research program, for example, he failed on two grounds: by completely misapplying Lakatos’s methodology and failing to devise a proper SRP from Waltz’s work.

This significant epistemological feud started when Elman and Elman (1995) countered Paul Schroeder’s criticism of neorealism, grounded on the claim that evidence of interstate power-balancing was historically scarce (Schroeder 1994). The Elmans’ contention, based on Lakatos, was twofold: (i) even if that is the case, it might not be enough to undermine the NSRP completely; and (ii) an SRP might still be improved and, in any case, it can only be discarded if a better one comes along (Elman and Elman 1995). This argument was enough for Vasquez to officially declare, also based on Lakatos, the degenerating state of neorealism as an SRP, due to neorealists’ inability to deal with the “balancing anomaly”, having to resort to supposedly contradictory propositions about state rational behavior (Vasquez 1997).

However, as later pointed out by the Elmans themselves, Vasquez’s claim was built in total disregard to Lakatos’ metrics. For once, he came up with four indicators of degeneration, which are at best incomplete – and at worst distorted and inconsistent – versions of Lakatos’ criteria for scientificity and scientific progress (Elman and Elman 1997, 923-4). Furthermore, Vasquez reduced the NSRP to the sole proposition that “states balance power”. This also flagrantly disregards all the elements that, according to Lakatos, comprise an SRP and guide its development (Elman and Elman 1997, 924-5).

It should also be noted that the claims put forth in this article are by no means self-evident and uncontroversial. It is our understanding that a rigorous and thorough attempt at codifying neorealism into a coherent SRP has not yet been made¹. Of course, it does not mean that an SRP was not in place in a more intuitive or practical sense. After all, one of the primary purposes of the MSRP as an analytical tool is to guide the rational reconstruction of previous scientific undertakings (Lakatos 1970b), even to a point in time when the methodology did not exist (Zahar 1973). Instead, it does mean that neorealism’s scientificity extent, as well as its current state of scientific vigor, could not perhaps be fully and adequately grasped. Also, the article’s second and third main claims – that offensive realism is a theoryshift within the NSRP and that it is theoretically and empirically progressive – are novel to the IR literature. For instance, according to Randall Schweller, a prominent realist scholar, Waltz’s and Mearsheimer’s contributions to IR theory brought about two distinct and competing SRPs, both at a theoretical dead end due to suffering “from too much success” and having said “everything that can be usefully said about the theory” (Schweller 2003, 345).

The first section of the article will lay down the MSRP and its criteria for scientific appraisal. The following section will frame neorealism as an SRP. Then, we will elaborate on how offensive realism is a theoretical modification within the NSRP, insofar as it amended the program’s protective belt of auxiliary hypothesis, while entirely respecting its hard core and positive heuristic. Finally, in the closing remarks, it will be argued that the offensive realism theoryshift is theoretically and empirically progressive, attesting to neorealism’s continuing relevance as a thriving scientific approach to international politics.

Lakatos and the Methodology of Scientific Research Programs

Lakatos’ MSRP derived directly from the main epistemological contest of the previous century, between Karl Popper’s falsificationist program and Thomas Kuhn’s sociological approach to science. Popper’s proficuous work was a reaction to verificationism, the dominant philosophical approach to science during the earlier part of the 20th century – particularly backed by the neopositivists of the so-called Circle of Vienna (Neurath et al. 1986). For neopositivists, the demarcation line between scientific and non-scientific (i.e., metaphysical) propositions rested on the formers’ provability through direct confrontation with empirical data. Popper had a problem with this inductionist reasoning because a proposition could never be logically and definitely proved from any series of confirming experiments (Popper 1962). Not even a fall back to probabilism – that is, high probability as a substitute for the unachievable criterion of definite proof – was good enough for Popper.

Popper’s response to this logical conundrum was his falsificationist program, according to which even though a proposition could never be definitely proven right, it could at least definitely (and deductively) be proven wrong. A new demarcation line was born: a scientific proposition must be refutable, that is, it should be able to specify upfront its potential falsifiers in a transparent manner. A non-scientific proposition would be one wholly protected against refutation. Hence the quintessential “all swans are white” as a model for scientific propositions, since a single observation of a non-white swan would be enough to put it down on the ground (Alves-Mazzotti and Gewandsznajder 1999).

The most contentious reaction to Popper’s falsificationism came from Kuhn (1962). According to him, the idea of science as a continuous process of critical refutation of theories through crucial experiments, followed by their replacement with new unrefuted (but refutable) theories, is naïve and inconsistent with the actual history of science. For Kuhn, this critical inclination, a cornerstone of Popper’s model, was an exception seen only in “extraordinary” or “revolutionary” moments in science. The regular state of affairs, dubbed “normal science” by Kuhn, would show the scientific community’s dogmatic adherence to a paradigm, with considerable staying power even in the face of mounting counterevidence. Kuhn also could not specify any rational and objective criteria that might cause revolution and a shift to a different paradigm. To Lakatos, it was akin to a religious conversion (Lakatos 1970a, 155).

According to Lakatos, Popper’s initial naïve ideas on falsificationism derived from two main basic assumptions: (i) it is possible to draw a clear line between theoretical and factual or empirical propositions, and (ii) an attested factual or empirical proposition is true, that is, proved from facts (Lakatos 1970a, 97-8). Lakatos, however, criticized this view, inherited from the naturalistic approach to the scientific method, and believed there is no infallible way to separate pure factual from theoretical propositions². The conclusion is obvious: not only scientific propositions cannot be definitely proved from facts, but they also cannot be truthfully falsified by them. Nevertheless, Kuhn’s solution based on dogmatic paradigmaticism was not enough for Lakatos, for he believed it irrevocably infused the scientific process with an element of irrationality.

Popper tried to rescue his falsificationist program by amending it with some methodological guidelines. Now fully aware that pure factual propositions do not exist, scientists must decide which propositions should form a set of unproblematic background knowledge that, for all intents and purposes, will function as an “empirical basis”. A proper scientific proposition, then, should be liable to be contrasted to and refuted by this empirical basis. At this point, all aspirations to absolute truth had to be renounced, since the content of the “empirical basis” would be the product of methodological – and therefore arbitrary – ruling.

Nonetheless, Lakatos believed that Popper’s methodological falsificationism still fell short of solving its naïve and unrealistic depiction of how science works. In particular, two recurrent and prominent features of science remained untackled: (i) theoretical tests are not usually two-sided in the sense of opposing a theory and the empirical data, but “are – at least – three-cornered fights between rival theories and experiments” and (ii) “some of the most interesting experiments result, prima facie, in confirmation rather than in falsification” (Lakatos 1970a, 115). Furthermore, Kuhn was right on at least one point: actual scientists fight hard to protect and save their theories against disconfirming evidence, up to a degree that would appear irrational to a true Popperian. An epistemology that provided rational criteria for theory rejection and scientific progress was needed, but it also had to be consistent with the basic traits of real-world science. To Lakatos, this epistemology was a sophisticated version of methodological falsificationism, which became the cornerstone of his MSRP. First, a new demarcation criterion had to be introduced:

For the naïve falsificationist any theory which can be interpreted as experimentally falsifiable is ‘acceptable’ or ‘scientific’. For the sophisticated falsificationist a theory is ‘acceptable’ or ‘scientific’ only if it has corroborated excess empirical content over its predecessor (or rival), that is, only if it leads to the discovery of novel facts. This condition can be analyzed into two clauses: that the new theory has excess empirical content (‘acceptability1’) and that some of this excess content is verified (‘acceptability2’). The first clause can be checked instantly by a priori logical analysis; the second can be checked only empirically and this may take an indefinite time (Lakatos 1970a, 116).

The new demarcation criterion accepted that empirical “facts” alone are not enough to bring down a theory, not only because they are inherently fallible but mainly because scientists can (and do) modify their theories to try to save them from anomalies. The point here is to search for progress and rationality in how this is done. When a new theory is created to save or substitute for another theory that did not fare well empirically, does it have excess empirical content over its predecessor? In other words, does it allow us to know/explain more of the world than we could before? If it does, the new theory is scientific; if it does not, it should be dismissed for being dishonest pseudoscience. From this demarcation – or acceptability – criterion follows a falsification one:

For the naïve falsificationist a theory is falsified by a (‘fortified’) ‘observational’ statement which conflicts with it (or which he decides to interpret as conflicting with it). For the sophisticated falsificationist a scientific theory T is falsified if and only if another theory T’ has been proposed with the following characteristics: (1) T’ has excess empirical content over T: that is, it predicts novel facts, that is, facts improbable in the light of, or even forbidden, by T; (2) T’ explains the previous success of T, that is, all the unrefuted content of T is included (within the limits of observational error) in the content of T’; and (3) some of the excess content of T’ is corroborated (Lakatos 1970a, 116).

Following the demarcation and falsification criteria stated above, one conclusion is clear: scientificity cannot be observed in isolated theories, irrespective of how they fare against the empirical basis, but in sequences of theories. A theory can only be dubbed scientific in relation to a prior or rival one, and only if it has excessive content (i.e., if it brings novel facts to light). Additionally, a theory can only be eliminated or falsified³ if a better theory comes along (i.e., a theory that contains excessive content that is – at least partially – empirically corroborated). In sum, the basic unit of science is no longer the theory but rather the genealogies of theories Lakatos named SRPs.

According to Lakatos’ methodology, an SRP comprises four main parts: a hard core, a negative heuristic, a positive heuristic and a protective belt. The hard core is made of inviolable assumptions. Its content is protected by the negative heuristic, which is “a set of propositions that say that this content [the hard core’s] cannot be directly challenged or tested” (Elman and Elman 2003, 26). The creation of the hard core is an arbitrary, methodological decision. In Newtonian physics SRP, for example, the hard core was composed of Newton’s three laws of dynamics and his law of gravitation (Lakatos 1970a, 133). Its negative heuristic forbade any theoretical development within the SRP from violating the content of any of these laws.

Contrary to the hard core, the protective belt is specifically designed to be modified and adjusted by the introduction of new auxiliary hypotheses, in order to protect the hard core from anomalous observations. The protective belt directly absorbs the blows from the empirical basis and tries to accommodate them. However, this process is not done in a disorderly fashion but instead with guidance from the SRP’s positive heuristic, which “consists of a partially articulated set of suggestions or hints on how to change, develop the ’refutable variants’ of the research-programme, how to modify, sophisticate, the ‘refutable’ protective belt” (Lakatos 1970a, 135). The positive heuristic indicates that, contrary to what a naïve falsificationist was led to believe, the development of an SRP involves expecting anomalies (or falsifiers) and planning how to deal with them. The positive heuristic is what allows for the relative autonomy of theoretical science and the internal rational reconstruction of a particular scientific field (Lakatos 1970b).

To summarize, a scientific theory is theoretically progressive, meaning it must predict new facts that were not covered – or were even forbidden – by the previous theory. Lakatos called ad hoc1 the theoretical amendments meant only to rescue a theory from anomalies without predicting any new facts. An SRP is considered progressive if its theories-chain is both theoretically and empirically progressive, meaning that at least some of the new facts predicted are actually corroborated⁴. Lakatos called ad hoc2 the theoretical amendments that, despite being scientific, cannot pass the empirical test. Finally, a third, heuristic criterion must be observed. A theory might predict novel, empirically corroborated facts, and still fail to reflect the SRP’s positive heuristic, resulting in an odd and incoherent mess. In Lakatos’ words, “one may achieve such ‘progress’ with a patched up, arbitrary series of disconnected theories. Good scientists will not find such makeshift progress satisfactory; they may even reject it as not genuinely scientific” (Lakatos 1970a, 175). Theoretical amendments that pass the novelty and empirical tests (that are not ad hoc1 or ad hoc2), but nonetheless fail the heuristic test, are called ad hoc3.

An SRP that systematically fails to produce non-ad hoc theoretical change to deal with empirical challenges has entered a degenerating stage and risks being superseded by a superior, more progressive one. This replacement happens in an inter-programmatic theoryshift, in which the previous program’s hard core is modified or a whole new one is created, resulting in a new SRP. Note that, to be progressive, this theoryshift needs only to avoid being ad hoc1 and ad hoc2. The heuristic criterion does not apply since the new SRP is bound to have its own positive heuristic, whereas an intra-programmatic theoryshift (within an SRP) must respect all three ad-hocness criteria.

A final note on the MSRP regards the meaning of “novel fact”, probably the most important concept of Lakatos’ methodology. There is a wide controversy on the matter and at least four main interpretations (Elman and Elman 2003, 33-40). Of these, the heuristic novelty interpretation, often called Zahar/Lakatos3, seems to be the most consistent with the original spirit of Lakatos’ epistemology. Accordingly, the anomalous facts that led to the creation of the new theory cannot count as novel, for they “provide little or no evidential support for the theory, since the theory was specifically designed to deal with the facts” (Zahar 1973, 102-3). Hence, a new theory must predict novel facts unrelated to the problem-solving context that provoked the theoryshift in the first place. If it cannot, it should be considered ad hoc1 and unscientific.

The Neorealist Scientific Research Program

In a seminal treatise, Waltz (1959) notoriously pointed out that the causes of international phenomena, most importantly war, could be searched in the characteristics of individuals, states, or the international system. The then most prominent and successful realist theory of international politics, Hans Morgenthau’s classical realism, drew on the first level of analysis (human nature) to explain international power competition, expansion, and war (Morgenthau 1954). In the book that presented neorealism to the world, Waltz’s main contention was that the striking recurrence of international phenomena through time, despite enormous variations of the internal attributes of states and previous political units, strongly suggested that a relevant part of the explanation must be found outside of states, that is, in the international system (Waltz 1979). According to Waltz, a system comprises interacting units and a political structure. The challenge, therefore, was to conceive of a political structure apart from the units and sort out how the former influences the latter’s behavior.

In neorealism, states are taken to be the units of the modern international system. The essential feature of the international political structure is anarchy, its ordering principle, which essentially means “the absence of a central monopoly of legitimate force” (Waltz 1988, 618). The other relevant attribute of the international political structure is how resources and capabilities are distributed among states, which in turn tells us the number of great powers or the system’s polarity. Anarchy sets the general tone and incentives of the system, while polarity dictates how the main units might cope with them.

Only two theoretical assumptions are needed to grasp how the international political structure shapes state behavior according to Waltz: states are unitary⁵ actors that want to survive as independent political units (Waltz 1979; 1988). Take these theoretical assumptions, add the anarchical nature of international politics, and the logical conclusions are: (i) the international system is a self-help one, in which states have to provide for their basic needs and security in particular; and (ii) states must pay careful attention to the balance of power among them since excessive relative power might breed aggression in a world devoid of institutionalized central protection⁶. International politics are, for neorealism, essentially balance-of-power politics, and the international system takes on the explanatory role for its inherently competitive and conflictual nature.

Nevertheless, why should states fiercely compete and fear each other so much if they are survival-oriented? The answer lies in the security dilemma, a theoretical notion best described in a classic article by John Herz (1950). The security dilemma is the tragic condition that makes self-help measures to increase one’s own security appear threatening and hostile to others, who, in turn, are compelled to take measures of their own, creating a spiraling and self-fulfilling dynamic. According to Waltz, these self-help measures include increasing and improving one’s internal capabilities, especially military assets (internal balancing), and forming alliances with weaker states to balance the power of stronger ones (external balancing)⁷ (Waltz 1979; 1988).

Furthermore, Waltz suggests that states’ concerns with their power position are of a conservative nature, meaning they are encouraged to maintain their international power position rather than to strive to maximize it. Balance-of-power logic means that power-maximizing behavior will ultimately be defeated by an overwhelming counterbalancing coalition, thereby rendering the greedy state’s security worse. Security maximization is the goal, not power maximization (Waltz 1988, 616).

As long as the international political structure remains anarchic, structural change can only happen in the distribution of capabilities among the major states of the system. There are two distinctive and analytically relevant possibilities, following Waltz: multipolarity, a system comprised of three or more great powers, and bipolarity, a system comprised of only two. A proper structural theory must be able to account for differences in the system’s dynamics as a result of different structural conditions. On this point, Waltz’s main assertion is that multipolar international systems tend to be more war-prone and unstable (more likely to change) than bipolar international systems⁸.

The above admittedly very summarized description of Waltz’s neorealism is, nonetheless, sufficient to subsidize an attempt at outlining what might be the NSRP. For the NSRP’s hard core, the following content is proposed:

(HC-1): Human groups, who act as the decisive and independent political units, are the main actors of international politics⁹;

(HC-2): States – or, for that matter, the decisive human groups that comprise the political units of a given international system – are taken to be unitary actors;

(HC-3): The ultimate goal of states – same caveat as above – is to maximize their prospects of survival as independent political units; and

(HC-4): The broad lines of international politics, particularly in relation to great power politics and its most recurrent aspects, are predominantly explained by the essential elements of the international political structure (i.e., anarchy and the underlying distribution of power).

The NSRP’s negative heuristic prohibits any theoretical amendments from violating any of the hard core’s propositions. If they do, an inter-programmatic theoryshift has taken place. A noteworthy absence in the hard core is the “rational-actor” assumption, which has become something of a truism when talking about neorealism¹⁰. However, Waltz explicitly rejected the rational-actor assumption, first because it is a unit-level – not a structure-level – variable, but also because he believed the scope of a systemic theory is to explain structural incentives, not predict how or how well specific actors will respond to these incentives (Waltz 1988; 1996). Hence, it seems unfair and inappropriate to include this assumption as part of the NSRP’s hard core, which does not mean it cannot be counted as an auxiliary hypothesis worked around the program’s protective belt to allow for the explanation of actor-specific decision-making (Elman 1996). That is precisely one of offensive realism’s theoretical innovations, as stated below (Mearsheimer 2009).

However central to his general argumentative scheme, the additional assumptions or hypothesis worked out by Waltz belong to the NSRP’s protective belt and are liable to modification. These include, among others: (i) states have a conservative, status quo bias¹¹; (ii) the security dilemma is an inescapable reality under anarchy and leads to power and security competition, crisis, and war; (iii) balancing – external and/or internal – is the expected political phenomenon in the face of concentrations of power; and (iv) the anarchic international system can be either multipolar or bipolar, with multipolar systems being more war-prone and unstable than bipolar systems.

Finally, Elman & Elman’s take on the NSRP’s positive heuristic makes sense and is probably adequate. They believe that the program’s positive heuristic states that neorealist theories must make predictions strictly about international political outcomes – or systemwide dynamics – such as the recurrence of balancing or multipolarity’s relative instability (Elman and Elman 2003, 27, 27-8). This heuristic criterion is in keeping with Waltz’s adamant defense of neorealism as an international politics (and not foreign policy) theoretical domain (Waltz 1988; 1996).

Offensive realism: an intra-programmatic theoryshift

Mearsheimer’s theory of international politics sets out to answer two general and important questions: “what causes states to compete for power?” and “how much power do states want?” (Mearsheimer 2001, 22). The answers are provided deductively from five basic assumptions about international life:

1. “The international system is anarchic”;

2. “Great powers inherently possess some offensive military capability”;

3. “States can never be certain about other states’ intentions”;

4. “Survival is the primary goal of great powers”; and

5. “Great powers are rational actors” (Mearsheimer 2001, 30-1).

Assumptions (i) and (iv) are, of course, directly taken from the NSRP’s hard-core propositions (HC-4 and HC-3, respectively), while assumptions (ii) and (iii) are truisms working as fairly implicit premises and a backdrop to Waltz’s logic. Also, hard-core propositions HC-1 and HC-2 are entirely preserved in Mearsheimer’s theory since it is admittedly state-centered (great power-centered, even) and takes states to be unitary actors that only differ in their relative power. The only odd piece out is (v) the rational-actor assumption. This proposition is the first theoretical amendment to the NSRP’s protective belt and has some important analytical corollaries. First and foremost, the rational-actor assumption allows the theorist to make predictions about the behavior of specific states in specific situations, not based on who the states are internally but on their international relative power position¹². In Popperian terms, this assumption also increases the theory’s stock of “potential falsifiers” because now great powers’ anomalous behavior can be seen as counterevidence. This argument will be explored further in the closing section of the article.

Back to the two guiding questions, the answer to the first one is a faithful restatement of Waltz’s reasoning: interstate power competition is a direct result of the anarchic, self-help nature of the international political structure and the ensuing security dilemma (Mearsheimer 2001, 29-54). The answer to the second question, in turn, brings about another amendment to the NSRP’s protective belt, a major one that makes offensive realism worthy of its name: great powers have an unlimited desire for power and aim to become the most powerful state in the system. In “Mearsheimer’s world” (Snyder 2002), security maximization is power maximization.

Every great power, in offensive-realist logic, wishes to achieve hegemony in the international system, meaning it would be the only true great power left and face no serious threat to its survival. Nevertheless, according to Mearsheimer, global hegemony has not been feasible in the past and is unlikely to be feasible in the future due to a set of assumptions about the material – logistical, tactical, and strategic – constraints on the offensive potential of military power¹³.

The first one is the stopping power of water, the enormously debilitating effect that oceans impose on a state’s force projection capabilities (Mearsheimer 2001, 114). This constraint would not be such a problem to a great power’s ability to expand were it not for the second assumption, the primacy of land power. Mearsheimer seriously doubts the potential that sea and air power – the two components of a state’s military might that are not severely hindered by the stopping power of water – have to achieve major political goals, much less territorial aggrandizement (Mearsheimer 2001, 83-128). The last material assumption is the unlikelihood that any great power will ever achieve nuclear superiority¹⁴. Therefore, fear of retaliation from a nuclear power tends to remain a powerful constraint on any great power’s expansionist intent (Mearsheimer 2001, 128-33).

If global hegemony is out of reach, then offensive realism expects great powers to try to dominate their own land-locked region, that is, become regional hegemons. That is done mainly through war fighting, great powers’ predominant strategy for gaining relative power¹⁵. According to Mearsheimer, a great power will be revisionist (power maximizer) until regional hegemony is reached, and from there onwards, it will acquire a status quo bias and strive to maintain its advantageous power position. An established regional hegemon would also be encouraged to export its military might to another region to help weaker local powers contain an aspiring regional hegemon, thereby avoiding the emergence of a global peer competitor¹⁶. This process is called offshore balancing (Mearsheimer 2001, 234).

Here we can discern a theoretical amendment to the NSRP’s protective belt. The basic unit of analysis is shifted from the systemwide (global) balance of power to regional balances of power, which are relatively and logically independent. In addition, another consequential theoretical amendment can be found in the possible polar configurations these regional balances of power can admit. First, a region can be hegemonic, dominated by only one great power. Second, Mearsheimer believes that not only the number of great powers is relevant to polarity but so is how evenly power is distributed among them. That results in the following possibilities: bipolarity (a region comprised of two great powers evenly balanced), balanced multipolarity (a region comprised of three or more great powers evenly balanced) and unbalanced multipolarity (a region comprised of three or more great powers, with one of them – a potential hegemon – being particularly stronger than the others) (Mearsheimer 2001).

A final assumption worthy of mention, also amending the NSRP’s protective belt, deals with great powers’ preferences in the face of rising powers. Offensive realism claims that there are two alternatives, balancing (directly engaging the threat to defeat and contain it) and buck-passing (trying to pass the buck of containment, at least partially, to other great powers), assuming that buck-passing is the preferable strategy¹⁷. The reason is simple: though riskier, buck-passing aims to achieve the same result as balancing – that is, to check an ascending threat – but on the cheap, thereby leaving the buck-passer in a stronger position vis-à-vis both the threat and the other unfortunate buck-catching great powers (Mearsheimer 2001, 157-62). Balancing is expected when buck-passing is either not possible due to geographic proximity with the threat or the absence of buck-catchers or when the enemy is so strong that there is a real chance buck-passing will fail and result in a successful bid for regional hegemony.

Conclusion: theoretical and empirical progress in the Neorealist Scientific Research Program

Offensive realism is an intra-programmatic theoryshift within the NSRP, for it offers a set of theoretical amendments to the program’s protective belt, while being entirely consistent with its hard core and positive heuristic. But is it scientific (predicts novel facts) and progressive (empirically corroborated)¹⁸?

It is fair to say that the theoryshift towards offensive realism was empirically provoked by the recurrence of revisionist or power-maximizing state behavior, which is inconsistent with Waltz’s expectations (based on auxiliary hypotheses). Revisionism had already been widely reported as an intrinsic, non-accidental feature of international politics (Snyder 1991; Zakaria 1992; Huntington 1993; Schweller 1994). Eric Labs, for example, in an incisive and methodologically rigorous empirical study, showed that great powers tend to expand their war aims during wartime when they perceive military and systemic opportunities (Labs 1997).

Power maximization was consistent with classical realism, of course, but its unit-level (human nature) explanatory variable made it inconsistent with the NSRP’s hard core (Morgenthau 1954). The same was true for neoclassical theories of foreign policy that combined unit- and system-level variables, such as Zakaria’s state-centered realism (Zakaria 1998) and Schweller’s balance of interests theory (Schweller 1997). Not to mention social-constructivist approaches that relied on notions like “predatory identities” (Wendt 1992). Offensive realism brought revisionism consistently into neorealism’s structural domain.

With the power-maximization and the rational-actor assumptions, offensive realism allowed for the rational reconstruction, borrowing a term from Lakatos, of centuries of great power politics and hegemonic wars: the Habsburg’s struggle for European domination in the 16th and 17th centuries (Kissinger 1994); Napoleonic France’s bid for European hegemony in the early 19th century; the United States’ path to regional hegemony in the Western Hemisphere throughout the 19th century (Mearsheimer 2001, 238-52); Germany’s bid for European hegemony twice, in World War I and World War II (Mearsheimer 2001, 181-90); Imperial Japan’s bid for East Asian hegemony in the 1930s and during World War II (Mearsheimer 2001, 172-181); and, finally, the Cold War as a massive power and security competition to prevent the Soviet Union from conquering Eurasia (Mearsheimer 2001, 190-202).

Offensive realism’s set of assumptions that lead to an analytical shift to regional balances of power explains a notable anomaly for Waltz’s original power-balancing hypothesis: the alliance between Western European states and the United States, the stronger power, against the Soviet Union during the Cold War¹⁹. Within offensive realism’s rationale, however, it is easily explained²⁰: the European post-war structural configuration was an unbalanced multipolarity, with the Soviet Union as a formidable potential hegemon, and the United States, though stronger, entered the stage as a conservative offshore balancer to help contain and deter an eventual Soviet bid for regional hegemony.

In terms of polarity, additional novel theoretical – and empirically corroborated – predictions are suggested. Offensive realism is in full agreement with Waltz regarding the virtues of bipolarity: it is the least war-prone and unstable power distribution due to strong incentives to balancing (mainly internally), which breeds equilibrium and simplicity (lower likelihood of miscalculations of power and resolve). The recent Cold War is the most notable example. Balanced multipolarity is more war-prone and unstable because there is a strong incentive to buck-passing while balancing, when it does happen, is predominantly external and inherently less effective²¹ (Mearsheimer 2001). However, the most war-prone and unstable power distribution is unbalanced multipolarity, a structural configuration unaccounted for by Waltz. The problem here is the presence of the potential hegemon, which has the incentives to reach for regional domination, raising fear and the security dilemma’s intensity through the roof²². As a result, offensive realism offers a more fine-grained explanation of the war-proneness and instability of multipolar systems. It helps explain, for example, why the first half of the 20th century was significantly more unstable and witnessed two World Wars as compared to the previous century (1815-1914). Europe was multipolar in both periods, but of the worst (unbalanced) kind prior to 1914 and 1939.

The theoretical notion of the offshore balancer sheds an interesting light on great power behavior. For example, it helps explain England’s secular part on the European balance-of-power game. England’s insular position, and the resulting constraints on its (land) power projection, meant that it was never a candidate for regional hegemony in Europe and had to play the offshore-balancer role. England’s case also reinforces Mearsheimer’s claim that offshore balancers are in a perfect position to be the quintessential buck-passers or balancers of last resort (Mearsheimer 2001, 261-4). On this matter, though, the spotlight is reserved for the United States, the “poster child for offensive realism” (Layne 2005). To Mearsheimer, it is as if the United States itself wrote the offensive-realist handbook: in the 19th century, it focused on achieving regional hegemony in the Western Hemisphere; then turned conservative and acted as an offshore balancer in World War I and World War II, not before trying to pass the buck to its regional allies for as long as it could; and after World War II, when buck-passing was unfeasible, it promptly engaged the Soviet Union in a fierce power and security competition to prevent the latter from dominating the Eurasian mainland (Mearsheimer 2001, 238-61).

But if the United States has been the poster child for offensive realism for most of its independent history, it is only ironic that it has become a thorn in the theory’s side since 1990. When the Cold War ended, and before offensive realism came into being, Mearsheimer (1990) famously predicted the end of the North Atlantic Treaty Organization (NATO) and an American political and military withdrawal from overseas. Offensive-realist logic was later even more incisive in this regard: if the power distribution in other regions is balanced, the structural incentive for the United States is to step back and pass the stabilizing buck to regional great powers (Mearsheimer 2001, 386-92). That, of course, has yet to happen, as the United States is to this day seriously committed (politically and militarily) to the defense and stability of key regions of the globe (Mendes 2019).

Still, when considering offensive realism and post-Cold War international politics, an innovative theoretical study by Eugenio Diniz is worth some consideration (Diniz 2006). The author set out to combine offensive-realist logic with unipolar politics dynamics, as seen in Wohlforth’s (1999) seminal work, to derive theoretical expectations regarding the relevance of multilateral relations for the lone pole and the remaining states in the system. Pending empirical corroboration of Diniz’s expectations, there is great potential for theoretical innovation since Mearsheimer is notoriously pessimist about any consequential role for multilateral organizations and arrangements (Diniz 2006, 505-6). An important caveat, though, is that Diniz’s discussion demanded significant adaptations of offensive realism, especially in terms of the conceptual differences between hegemonic and unipolar systems and in applying the theory’s (inter)regional approach to a global balance of power, to a point in which it is questionable if Mearsheimer’s offensive realism is indeed the analytical starting point.

In any case, the United States’ continued large-scale military presence overseas, particularly in Europe, is an anomaly to offensive realism, and there is no way around it²³. Inertia can only count for so much (Mearsheimer 1998), and dropping the rational-actor assumption would be a degenerating option while unit-level explanations violate the NSRP’s hard core. Nevertheless, the almost certain power shift on course in East Asia, from a balanced to a (China-centered) unbalanced multipolarity, if accompanied by a redirection and concentration of the United States’ external balancing efforts, would be a refreshing lift of offensive realism’s shoulder. In fact, things seemed to be moving this way in President Obama’s second term, when he announced a “pivot to Asia” and his administration’s Defense Strategic Guidance stated that “while the U.S. military will continue to contribute to security globally, we will of necessity rebalance toward the Asia-Pacific region” (Department of Defense of United States of America 2012, 8). Now the future is unclear, however, especially considering the Russian 2022 war in Ukraine and the recrudescence of great power politics in Europe²⁴.

Independently on how the future of the American Pacifier unfolds, and luckily to offensive realism, Lakatos’ methodology is not naïve falsificationism. In fact, Lakatos readily recognized that every theory, even the best, is born amidst an “ocean of anomalies” (Lakatos 1970a, 135). Maybe the way offensive realism winds up dealing with current anomalies, if they persist, will yield more theoretical and empirical progress to the NSRP. The main point, though, is that it already is a successful and progressive theoryshift, attesting to neorealism’s continuing relevance as a scientifically rigorous research program in IR theory.

References

• Alves-Mazzotti, A., F. Gewandsznajder. O método nas ciências naturais e sociais. São Paulo: Pioneira Thomson Learning, 1999.
• Brooks, S. “Dueling realisms.” International Organization 51, no. 3 (1997): 445-477. doi: https://doi.org/10.1162/002081897550429
  » https://doi.org/10.1162/002081897550429
• Department of Defense of United States of America. Sustaining U.S. global leadership: priorities for 21th century defense. Washington, 2012.
• Diniz, E. “Relacionamentos multilaterais na unipolaridade: uma discussão teórica realista.” Contexto Internacional 28, no. 2 (2006): 505-565.
• Elman, C. “Horses for courses: why not neorealist theories of foreign policy?” Security Studies 6, no. 1 (1996): 7-53. doi: https://doi.org/10.1080/09636419608429297
  » https://doi.org/10.1080/09636419608429297
• Elman, C., M. Elman. “Correspondence: history vs. neo-realism: a second look.” International Security 20, no. 1 (1995): 182-193. doi: https://doi.org/10.2307/2539222
  » https://doi.org/10.2307/2539222
• Elman, C., M. Elman. “Lakatos and neorealism: a reply to Vasquez.” American Political Science Review 91, no. 4 (1997): 923-926. doi: https://doi.org/10.2307/2952175
  » https://doi.org/10.2307/2952175
• Elman, C., M. Elman. “Lessons from Lakatos.” In Progress in International Relations Theory: Appraising the Field, edited by C. Elman and M. Elman, 21-70, Cambridge: MIT, 2003.
• Fischer, M. “Feudal Europe, 800-1300: communal discourse and conflictual practices.” International Organization 46, no. 2 (1992): 427-466. doi: https://doi.org/10.1017/S0020818300027776
  » https://doi.org/10.1017/S0020818300027776
• Glaser, C. “Realists as optimists: cooperation as self-help.” International Security 19, no. 3 (1994): 50-90. doi: https://doi.org/10.2307/2539079
  » https://doi.org/10.2307/2539079
• Herz, John. “Idealist internationalism and the security dilemma.” World Politics 2, no. 2 (1950): 157-180. doi: https://doi.org/10.2307/2009187
  » https://doi.org/10.2307/2009187
• Huntington, S. “Why primacy matters.” International Security 17, no. 4 (1993): 68-83. doi: https://doi.org/10.2307/2539022
  » https://doi.org/10.2307/2539022
• Keohane, R., L. Martin. “Institutional Theory as a Research Program.” In Progress in International Relations Theory: Appraising the Field, edited by C. Elman, and M. Elman, 71-108, Cambridge: MIT, 2003.
• Kissinger, H. Diplomacy. New York: Simon & Schuster, 1994.
• Kuhn, T. The structure of scientific revolutions. Chicago: University of Chicago, 1962
• Labs, E. “Beyond victory: offensive realism and the expansion of war aims.” Security Studies 6, no. 4 (1997): 1-49. doi: https://doi.org/10.1080/09636419708429321
  » https://doi.org/10.1080/09636419708429321
• Lakatos, I. “Falsification and the methodology of scientific research programmes.” In Criticism and the growth of knowledge, edited by I. Lakatos and A. Musgrave, 91-196, Cambridge: Cambridge University, 1970a.
• Lakatos, I. “The history of science and its rational reconstruction.” In PSA 1970: Boston studies in the philosophy of science, edited by R. C. Buck, and R. S. Cohen, 91-135, Dordrecht: Reidel, 1970b.
• Layne, C. “The ‘poster child for offensive realism’: america as a global hegemon.” Security Studies 12, no. 2 (2004): 120-164.
• Levy, J. “Case studies: types, designs, and logics of inference.” Conflict Management and Peace Science 25, no. 1 (2008): 1-18. doi: https://doi.org/10.1080/07388940701860318
  » https://doi.org/10.1080/07388940701860318
• Mearsheimer, J. “Back to the future: instability in Europe after the cold war.” International Security 15, no. 1 (1990): 3-54. doi: https://doi.org/10.2307/2538981
  » https://doi.org/10.2307/2538981
• Mearsheimer, J. “Reckless states and realism.” International Relations 23, no. 2 (2009): 241-256. doi: https://doi.org/10.1177/0047117809104637
  » https://doi.org/10.1177/0047117809104637
• Mearsheimer, J. “The false promise of international institutions.” International Security 19, no. 3 (1995): 4-49. doi: https://doi.org/10.2307/2539078
  » https://doi.org/10.2307/2539078
• Mearsheimer, J. J. “The future of America’s continental commitment.” In No end to alliance: the United States and Western Europe: past, present and future, edited by G. Lundestad, 221-242, New York: St. Martin’s, 1998.
• Mearsheimer, John J. The tragedy of great powers politics. New York: Norton & Company, 2001.
• Mendes, F. P. “Building commitment: the US military presence in Europe during the cold war.” Revista da Escola de Guerra Naval 25, no. 3 (2019): 587-611.
• Morgenthau, H. Politics among nations: the struggle for power and peace. New York: Knopf, 1954.
• Neurath, O., H. Kahn, and R. Carnap. “A concepção científica do mundo: o círculo de Viena.” Cadernos de História e Filosofia da Ciência 10 (1986): 65-76.
• Popper, K. Conjectures and refutations: the growth of scientific knowledge. New York: Basic, 1962.
• Posen, B. “The security dilemma and ethnic conflict.” Survival 35, no. 1 (1993): 27-47. doi: https://doi.org/10.1080/00396339308442672
  » https://doi.org/10.1080/00396339308442672
• Rosecrance, R. “War and peace.” World Politics 55, no. 1 (2002): 136-166. doi: https://doi.org/10.1353/wp.2003.0007
  » https://doi.org/10.1353/wp.2003.0007
• Schmitt, C. The concept of the political. Chicago: University of Chicago, 1996.
• Schroeder, P. “Historical reality vs. neo-realist theory.” International Security 19, no. 1 (1994): 108-148. doi: https://doi.org/10.2307/2539150
  » https://doi.org/10.2307/2539150
• Schweller, R. “Bandwagoning for profit: bringing the revisionist state back in.” International Security 19, no. 1 (1994): 72-107. doi: https://doi.org/10.2307/2539149
  » https://doi.org/10.2307/2539149
• Schweller, R. “The progressiveness of neoclassical realism.” In Progress in international relations theory: appraising the field, edited by C. Elman and M. Elman, 311-48, Cambridge: MIT, 2003.
• Schweller, R. Deadly imbalances: tripolarity and hitler’s strategy of world conquest. New York: Columbia University, 1997.
• Snyder, G. “Mearsheimer’s world: offensive realism and the struggle for security.” International Security 27, no. 1 (2002): 149-173.
• Snyder, J. Myths of empire: domestic politics and international ambition. Ithaca: Cornell University, 1991.
• Taliaferro, J. “Security seeking under anarchy: defensive realism revisited.” International Security 25, no. 3 (2000): 128-161.
• Vasquez, J. “The realist paradigm and degenerative versus progressive research programs: an appraisal on neotraditional research on Waltz’s balancing proposition.” American Political Science Review 91, no. 4 (1997): 899-912. doi: https://doi.org/10.2307/2952172
  » https://doi.org/10.2307/2952172
• Walt, S. The origins of alliances. Ithaca: Cornell University, 1987.
• Waltz, K. Theory of International Politics. New York: McGraw-Hill, 1979.
• Waltz, K. “International politics is not foreign policy.” Security Studies 6, no. 1 (1996): 54-57. doi: https://doi.org/10.1080/09636419608429298
  » https://doi.org/10.1080/09636419608429298
• Waltz, K. “The origins of war in neorealist theory.” Journal of Interdisciplinary History 18, no. 4 (1988): 615-628. doi: https://doi.org/10.2307/204817
  » https://doi.org/10.2307/204817
• Waltz, K. “The stability of a bipolar world.” Daedalus 93, no. 3 (1964): 881-909.
• Waltz, K. Man, the state and war. New York: Columbia University, 1959.
• Wendt, A. “Anarchy is what states make of it: the social construction of power politics.” International Organization 46, no. 2 (1992): 391-425.
• Wohlforth, W. “The stability of a unipolar world.” International Security 24, no. 1 (1999): 5-41.
• Zahar, E. “Why did Einstein’s programme supersede Lorentz’s?” British Journal for the Philosophy of Science 24, no. 2 (1973): 95-123.
• Zakaria, F. “Realism and domestic politics: a review essay.” International Security 17, no. 1 (1992): 177-198. doi: https://doi.org/10.2307/2539162
  » https://doi.org/10.2307/2539162
• Zakaria, F. From wealth to power: the unusual origin of America’s global role. Princeton: Princeton University, 1998.

• 1
  Examples of scattered, less rigorous attempts include Vasquez (1997), Elman and Elman (1997; 2003) and Keohane and Martin (2003).
• 2
  Galileo’s observation of the moon’s surface aided by the brand-new technology of telescopes comes to mind, rejected by many as it was because they simply did not buy the underlying optical theory (Lakatos 1970a, 98). On the most basic level, even “direct” visual observations are not technically factual, for they depend on optical (theoretical) propositions about the human eye.
• 3
  The idea of “falsification” does not make much sense at this point anymore, and keeping it is perhaps only justified by Lakatos’ willingness to remain in Popper’s orbit.
• 4
  This could take some time, as Lakatos so frequently reminded, particularly if new methods or technologies are required for experimentation. The 100-years gap between Einstein’s prediction of gravitational waves and their actual measurement, when the technology finally became available, comes to mind. Luckily for Einstein, many of his boldest new predictions could be tested much sooner (Zahar 1973).
• 5
  The “unitary actor” assumption is rather born out of methodological necessity, after all a structural model of international politics must abstract from any internal attributes of states.
• 6
  This also leads one to expect much lower levels of cooperation among states, compared to those obtained in hierarchically integrated systems. Anarchy discourages interdependence and encourages self-sufficiency, and there are the additional possibilities of cheating and uneven distribution of gains, resulting in a deteriorating position in the international balance of power. See Waltz (1979) and Mearsheimer (1995).
• 7
  The main alternative to external balancing, according to Waltz, is bandwagoning, meaning to ally with the stronger power in order to profit from the spoils of its conquests. This behavior is not encouraged by the international system, according to neorealist reasoning, because it is contrary to balance-of-power logic. See Waltz (1979) and Walt (1987).
• 8
  This tendency is due mainly to differences in the balancing dynamics of the two systems (predominantly external in multipolarity, predominantly internal in bipolarity), resulting in varying prospects for equilibrium, miscalculation and deterrence (Waltz 1979; 1988). For an earlier, less theoretically rigorous version of the argument, see Waltz (1964).
• 9
  This assumption goes back to realism as a political-philosophical view. See Schmitt (1996). It also allows for a welcome and appropriate theoretical application of neorealism beyond the modern nation-state framework. See Fischer (1992) and Posen (1993).
• 10
  For example, Elman and Elman (1997, 924) suggest the rational-actor assumption as a hard-core proposition for the NSRP.
• 11
  It might be argued that this assumption is linked to a distinctive neorealist theory, dubbed defensive realism (Glaser 1994; Brooks 1997; Taliaferro 2000). Following this article’s proposition, it would mean that Waltz created both an SRP and – on the protective belt – a particular theory within it.
• 12
  Hence, Rosecrance’s claim that Mearsheimer created a theory of foreign policy is unwarranted (Rosecrance 2002). If the behavior is motivated (explained) by the state’s international relative power position, then it is international political behavior, not foreign policy, and a reflection of the system’s dynamics. Mearsheimer’s theory is consistent with the NSRP’s positive heuristic.
• 13
  These assumptions should indeed be seen as a third set of theoretical amendments to the NSRP’s protective belt, for they lay the material background for the scope and extent of power competition among the great powers. As will become clear, if these assumptions are relaxed, or if any of them becomes materially obsolete, markedly different theoretical predictions would ensue.
• 14
  Nuclear superiority is achieved when a state has nuclear monopoly or when it has no reason to fear nuclear retaliation, either because it can fully destroy the adversary’s nuclear arsenal on a first strike, or because it has developed a very reliable missile and air defense system.
• 15
  Minor and less effective relative power-gaining strategies include blackmail (coercion), bait and bleed (make two other powers fight and weaken each other) and bloodletting (contribute to the protraction of a war that is already happening between two other powers) (Mearsheimer 2001, 147-55).
• 16
  Layne (2005) sees this as a contradiction in Mearsheimer’s logic: if the stopping power of water is a fact, why should regional hegemons care if there are other hegemons in different regions of the world? The inherent uncertainty about the future, specially about the result of arms and technological races, however, may be invoked to save Mearsheimer from this apparent contradiction. The material constraints on power projection might be technologically overcome in the future.
• 17
  Bandwagoning is dismissed for being against balance-of-power logic and is only theoretically acceptable for a great power in very specific conditions (Mearsheimer 2001, 162-5).
• 18
  The bulk of this article is admittedly (meta)theoretical. Its length is barely sufficient for a proper discussion of the epistemological and theoretical aspects of the argument. The empirical side of the argument, pertaining to offensive realism’s excessive corroborated content and the NSRP’s overall progressiveness, therefore, can only be addressed in a preliminary and illustrative manner. References will be made to existing empirical studies that are consistent with some of offensive realism’s novel predictions, which can be assessed and judged on their own terms. But for the most part, what follows should be viewed as a plausibility probe (Levy 2008) and a research agenda setting, rather than an exhaustive and definitive appraisal of empirical success.
• 19
  Stephen Walt’s (1987)balance of threat theory, created to deal with this anomaly, depended on assumptions that are problematic to the NSRP’s hard core, such as perception of intentions and the offense-defense balance.
• 20
  This point is made, also in Lakatosian epistemological terms, by Diniz (2006, 554).
• 21
  Offensive realism’s assumption that buck-passing is preferable to balancing makes the inefficiencies and tardiness of external balancing more consistent with neorealism. This fact should also be counted as an improvement of the NSRP’s explanatory power.
• 22
  In unbalanced multipolarities, buck-passing is riskier and therefore less advised. Nevertheless, it is still tempting, especially in earlier stages, which in turn might lower the prospects for deterrence and increase the likelihood of war.
• 23
  It is consistent with Wohlforth’s (1999) theoretical take on unipolar politics, though, which is mostly done on neorealist grounds. Wohlforth’s “unipolarity theory”, however, would still have to pass the adhocness test, that is, it must show excessive explanatory power beyond its factual raison d’être: the United States’ post-Cold War unipolar moment.
• 24
  Even though it can hardly be claimed that, as of 2022, Russia is again a potential hegemon in an unbalanced multipolar Europe.

Publication Dates

• Publication in this collection
  06 Jan 2023
• Date of issue
  2022

History

• Received
  21 Sept 2022
• Accepted
  26 Oct 2022