Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 18 June 2021

Political Ontology

Abstract and Keywords

This article examines the relevance of ontology to political analysis. It explains that as political science became more reflexive and less confident that it was before, ontological concerns have increasingly come to the fore. It stresses the fact that no political analysis has ever been ontologically neutral. It discusses the concept of political ontology, the status of ontological claims, and ontological disputes in political analysis. It also highlights the consistent disparity between the often tacit and normalized analytical assumptions of existing mainstream approaches to political analysis and those which emerge from sustained ontological reflection.

Keywords: ontology, political analysis, political science, political ontology, ontological disputes, analytical assumptions

The problems of pure philosophical ontology have seemed so deep or confused that philosophers who concentrate primarily on the concept of being as such have acquired an occasionally deserved reputation for obscurity and even incoherence.

(Jacquette 2002, xi)

The terms “political” and “ontology” have, until recently, rarely gone together and, given the above comments, it might seem desirable to maintain that separation. Political scientists, for the most part, have tended to leave ontological issues to philosophers and to those social scientists less encumbered by substantive empirical concerns. Yet as the discipline has become more reflexive and perhaps rather less confident than once it was at the ease with which it might claim a scientific license for the knowledge it generates, so ontological concerns have increasingly come to the fore. In addressing such issues, as I shall argue, political analysts have no so much moved into novel terrain as acknowledged, reflected upon, challenged, and, in some cases, rethought the tacit assumptions on which their analytical enterprises were always premised. No political analysis has ever been ontologically neutral; rather fewer political analysts are prepared to proceed today on the basis of this once unacknowledged and unchallenged presumption.

Consequently, however tempting it might well be to leave ontology to others, that option may not be available to us. The principal aim of the present chapter is to explain why this is so. The argument is, in essence, simple. Ontological assumptions (relating to the nature of the political reality that is the focus of our analytical attentions) are logically antecedent to the epistemological and methodological choices more usually identified as the source of paradigmatic divergence in political science (p. 461) (cf. King, Keohane, and Verba 1994; Monroe 2004). Two points almost immediately follow from this. First, often unacknowledged ontological choices underpin major theoretical disputes within political analysis. Second, whilst such disagreements are likely to be manifest in epistemological and methodological choices, these are merely epiphenomena of more ultimately determinate ontological assumptions. Accordingly, they cannot be fully appreciated in the absence of sustained ontological reflection and debate.

This is all very well in the abstract, but it remains decidedly abstract. The second challenge of this chapter is to demonstrate that “ontology matters” in substantive terms. This may sound like a tall order. However, it is in fact rather more straightforward that might be assumed. First, we might note that political ontology is intimately associated with adjudicating the categories to which legitimate appeal might be made in political analysis. As Charles Tilly and Robert E. Goodin note, “ontological choices concern the sorts of social entities whose consistent existence analysts can reasonably assume” (2006). In other words, whether we choose to conduct our analysis in terms of identities, individuals, social collectivities, states, regimes, systems, or some combination of the above, reflects a prior set of ontological choices and assumptions—most obviously about the character, nature, and, indeed, “reality” of each as ontological entities and (potential) dramatis personae on the political stage.

Second, even where we can agree upon common categories of actors, mechanisms, or processes to which legitimate appeal can be made, ontological choices affect substantively the content of our theories about such entities (and hence our expectations about how the political drama will unfold). A shared commitment to ontological individualism (the view that human individuals are the sole, unique, and ultimate constituents of social reality to which all else is reducible) is no guarantee of a common approach to political analysis, far less to a common account of a specific political drama or context. The substantive content of our ontological individualism will vary dramatically if we regard actors to be self-serving instrumental utility maximizers, on the one hand, or altruistic communitarians, on the other, just as our view of the strategies appropriate to the emancipation of women will vary significantly depending on our (ontological) view as to the biological and/or social character of seemingly “essential” gender differences (compare, for instance, Brownmiller 1975; Daly 1978; Elshtain 1981; Wolf 1993; Young 1990). In these, and innumerable other ways, our ontological choices—whether acknowledged or unacknowledged—have profound epistemological, methodological, and practical political consequences.

Given this, it is pleasing to be able to report that contemporary political analysts are rather more reflexive, ontologically, than many of their immediate predecessors. Representative of contemporary trends in this respect is Alexander Wendt. Ontology, he suggests,

is not something that most international relations (IR) scholars spend much time thinking about. Nor should they. The primary task of IR social science is to help to understand world politics, not to ruminate about issues more properly the concern of philosophers. Yet even the most empirically minded students of international politics must “do” ontology.

(1999, 370)

(p. 462) In the brief survey that follows, my aim is to indicate in outline form what “doing” political ontology entails. But it is first important to establish, in somewhat greater detail, what it is and why it is important.

1 Political Ontology: What Is It?

Most standard philosophical treatments of ontology differentiate between two, albeit closely related, senses of the term.1 The first, and more abstract, is concerned with the nature of “being” itself—what is it to exist, whether (and, if so, why) there exists something rather than nothing, and whether (and, if so, why) there exists one logically contingent actual world. The second sense of the term is concerned with the (specific) set of assumptions made about the nature, essence, and characteristics (in short, the reality) of an object or set of objects of analytical inquiry. However ethereal such issues may nonetheless seem, political analysts have principally concerned themselves with the latter, philosophically more prosaic, set of concerns. In Benton and Craib’s (2001) terms, political ontology is a “regional ontology.” This chapter replicates that focus.

Thus, whilst ontology is defined, literally, as the ‘science’ or ‘philosophy’ of being, within political analysis it has tended to be defined in more narrow and specific terms. Norman Blaikie’s definition is here representative. Ontology, he suggests, “refers to the claims or assumptions that a particular approach to social [or, by extension, political] enquiry makes about the nature of social [or political] reality—claims about what exists, what it looks like, what units make it up and how these units interact with one another” (1993, 6). Ontology relates to being, to what is, to what exists, to the constituent units of reality; political ontology, by extension, relates to political being, to what is politically, to what exists politically, and to the units that comprise political reality.

The analyst’s ontological position is, then, her answer to the question: What is the nature of the social and political reality to be investigated? Alternatively, what exists that we might acquire knowledge of? As this already implies, ontology logically precedes epistemology. However put, these are rather significant questions whose answers may determine, to a considerable extent, the content of the political analysis we are likely to engage in and, indeed, what we regard as an (adequate) political explanation. Thus, for “ontological atomists,” convinced in Hobbesian terms that “basic human needs, capacities and motivations arise in each individual without regard to any specific feature of social groups or social interactions” (Fay 1996, 31), there can be no appeal in political explanation to social interactions, processes or structures. For “ontological structuralists,” by contrast, it is the appeal to human needs and capacities that is ruled inadmissible in the court of political analysis. Similarly, for (p. 463) those convinced of a separation of appearance and reality—such that we cannot trust our senses to reveal to us that which is real as distinct from that which merely presents itself to us as if it were real—political analysis is likely to be a rather more complex and methodologically exacting process than for those prepared to accept that reality presents itself to us in a direct and unmediated fashion.

Working from this simple definition, a great variety of issues of political ontology can be identified. Adapting Uskali Mäki’s thoughtful (and pioneering) reflections on economic ontology (2001, 3; see also Mäki 2002, 15–22) to the political realm, we might identify all of the following as ontological questions:

What is the polity made of? What are its constituents and how do they hang together? What kinds of general principles govern its functioning, and its change? Are they causal principles and, if so, what is the nature of political causation? What drives political actors and what mental capacities do they possess? Do individual preferences and social institutions exist, and in what sense? Are (any of) these things historically and culturally invariant universals, or are they relative to context?

Such questions readily establish a simple analytical agenda for political ontology. They also serve to indicate that no political analysis can proceed in the absence of assumptions about political ontology. That such assumptions are rarely explicit hardly makes them less consequential. Presented more thematically, amongst the ontological issues on which political analysts formulate consequential assumptions are the following:
  1. 1. The relationship between structure and agency, context, and conduct.

  2. 2. The extent of the causal and/or constitutive role of ideas in the determination of political outcomes.

  3. 3. The extent to which social and political systems exhibit organic qualities or are reducible in all characteristics to the sum of their constituent units/parts.

  4. 4. The (dualistic or dialectical) relationship between mind and body.

  5. 5. The nature of the human (political) subject and its behavioural motivations.

  6. 6. The extent to which causal dynamics are culturally/contextually specific or generalizable.

  7. 7. The respective characteristics of the objects of the natural and social sciences.

  8. 8. Perhaps most fundamentally of all, the extent (if any) of the separation of appearance and reality—the extent to which the social and political world presents itself to us as really it is such that what is real is observable.

Whilst interest in, and reflexivity with respect to, such ontological issues has certainly risen considerably in recent years, coverage of such issues is very uneven. Indeed, it is really only some of these issues—principally the first, second, third, and, to some extent, the fifth—that have prompted sustained ontological reflection to date.2 It is on these issues that this chapter will concentrate principally.

(p. 464) The crucial point, for now, to note about each of these issues is that none of them can be resolved empirically. Ultimately, no amount of empirical evidence can refute the (ontological) claims of the atomist or the structuralist; neither can it confirm or reject the assumption that there is no separation of appearance and reality.3 This is all rather disconcerting and perhaps explains the characteristic reluctance of political analysts to venture into debate on, and thereby to lay bare, their ontological assumptions. For to acknowledge an ontological dependence, and hence a reliance upon assumptions that are in principle untestable, may be seen to undermine the rightly cherished and long-fought-for authority of the analyst and the analytical traditions in which her contribution is constructed. Yet, on any sustained reflection, silence is not a very attractive option either. For, whether we like it or not, and whether we choose to acknowledge it or not, we make ontological assumptions—in Wendt’s terms, we “do” ontology. These assumptions profoundly shape our approach to political analysis and cannot simply be justified by appeal to an evidential base. It is to the consequences of such choices that we now turn.

2 … And Why Is It Important?

However significant they may be in their own terms, ontological assumptions find themselves increasingly the subject of the political analyst’s attentions largely for their epistemological and methodological consequences.

Again it is important to be precise about our terminology, for confusions abound in the literature.4

Epistemology, again defined literally, is the “science” or “philosophy” of knowledge. In Blaikie’s terms, it refers “to the claims or assumptions made about the ways in which it is possible to gain knowledge of reality” (1993, 6–7). In short, if the ontologist asks “what exists to be known?”, then the epistemologist asks “what are the conditions of acquiring knowledge of that which exists?” Epistemology concerns itself with such (p. 465) issues as the degree of certainty we might legitimately claim for the conclusions we are tempted to draw from our analyses, the extent to which specific knowledge claims might be generalized beyond the immediate context in which our observations were made, and, in general terms, how we might adjudicate and defend a preference between contending political explanations. As this indicates, epistemological assumptions are invariably ontologically loaded—whether knowledge is transferable between different settings for political analysis and hence whether we can legitimately generalize between “cases” (an epistemological consideration) depends on (prior) assumptions about the ontological specificity of such settings.

Yet the implications of ontological choices are not confined to epistemology; they are also methodological.

Methodology relates to the choice of analytical strategy and research design which underpins substantive research. Although methodology establishes the principles which might guide the choice of method, it should not be confused with the methods and techniques of research themselves. Indeed, methodologists frequently draw the distinction between the two, emphasizing the extent of the gulf between what they regard as established methodological principles and perhaps equally well-established methodological practices. What they invariably fail to do is to acknowledge and reflect upon the ontological dependence of methodological choices. For our purposes methodology is best understood as the means by which we reflect upon the methods appropriate to realize fully our potential to acquire knowledge of that which exists.

What this brief discussion hopefully serves to demonstrate is that ontology, epistemology, and methodology, though closely related, are irreducible. Ontology relates to the nature of the social and political world, epistemology to what we can know about it, and methodology to how we might go about acquiring that knowledge.

As this perhaps already serves to indicate, their relationship is also directional—ontology logically precedes epistemology which logically precedes methodology (see also Archer 1998; Bhaskar 1989, 49; Gilbert 1989, 440; though cf. Smith 1990, 18). We cannot know what we are capable of knowing (epistemology) until such time as we have settled on (a set of assumptions about) the nature of the context in which that knowledge must be acquired (ontology). Similarly, we cannot decide upon an appropriate set of strategies for interrogating political processes (methodology) until we have settled upon the limits of our capacity to acquire knowledge of such processes (epistemology) and, indeed, the nature of such processes themselves (ontology).

Political Ontology

Fig. 23.1. The directional dependence of ontology, epistemology, and methodology: the case of postmodernism

Source: Hay (2002: 227).

The directional dependence of this relationship is presented schematically and illustrated with respect to postmodernism in Figure 23.1. As this already serves to indicate, to suggest that ontological consideration are both irreducible and logically prior to those of epistemology is most definitely not to suggest that they are unrelated. The degree of confidence that we might have for the claims we make about political phenomena, for instance, is likely to vary significantly depending on our view of the relationship between the ideas we formulate on the one hand and the political referents of those ideas, on the other. In this way, our ontology may shape our epistemology; moreover, both are likely to have methodological implications. If we are happy to conceive of ourselves as disinterested and dispassionate observers of an external (p. 466) (political) reality existing independently of our conceptions of it, then we are likely to be rather more confident epistemologically than if we are prepared to concede that: (1) we are, at best, partisan participant observers; (2) that there is no neutral vantagepoint from which the political can be viewed objectively; and that (3) the ideas we fashion of the political context we inhabit influence our behavior and hence the unfolding dynamics of that political context.5 Such ontological assumptions and their epistemological implications are, in turn, likely to influence significantly the type of evidence we consider and the techniques we deploy to interrogate that evidence. If, for instance, we are keen to acknowledge (ontologically) an independent causal role for ideas in determining the developmental trajectory of political institutions, then we are likely to devote our methodological energies to gauging the understandings of political subjects. If, by contrast, we see ideas as merely epiphenomenal of ultimately determinant material bases (for instance, the self-interest of the actors who hold such ideas), then our methodological attentions will be focused elsewhere.

(p. 467) 3 The Status of Ontological Claims

Given the sheer volume of literature devoted in recent years to questions of ontology (principally the structure–agency and material–ideational relationships) in political science and international relations, it might be tempting to assume that the need for a series of reflections on this question is relatively undisputed. The reality, however, it somewhat different. For even in sociology, perhaps the natural home of reflection on such issues, there are dissenting voices. In making the case for the centrality of such concerns to political analysis it is perhaps appropriate that we first deal with the potential objections. Among the most vociferous of critics of the “craze” for abstract ontological reflection is Steve Fuller. His central argument is simply stated:

Given the supposedly abortive attempts at solving the structure–agency problem, one is tempted to conclude that sociologists are not smart enough to solve the problem or that the problem itself is spurious.

(Fuller 1998, 104)

The case is certainly well made, and might be extended to almost all ontological reflection within the social sciences. There would seem to be little to be gained by political analysts in following their sociological forebears into an ontological cul-de-sac of obfuscation and meaningless abstraction.

Yet Fuller’s remarks are not quite as devastating as they might first appear. For, in certain crucial respects, they reveal a systematic, if widespread, misinterpretation of the nature of ontological disputes of this kind. In this respect they prove quite useful in helping us establish what is—and what is not—at stake in debates about the relative significance of ideational and material, or structural and agential factors. Put most simply, ontological issues such as these are not “problems” to which there is, or can be, definitive solutions.

To appeal to the issue of structure and agency, for instance, as a “problem” with a potential “solution” is effectively to claim that the issue is an empirical one that can be resolved definitively. Yet, claims as to the relative significance of structural and agential factors are founded on ontological assumptions as to the nature of a social and political reality. To insist that such claims can be resolved by appeal to the evidence is, then, to conflate the empirical and the ontological. To put this in more practical and prosaic terms, any given and agreed set of empirical observations can be accounted for in more or less agential, more or less structural terms. We might, for instance, agree on the precise chain of events leading up to the French Revolution of 1789 whilst disagreeing vehemently over the relative significance of structural and agential factors in the explanation of the event itself. Evidence alone is not ontologically discriminating, though it is often presented as such.6

(p. 468) Two important implications follow directly from the above discussion. First, if the relative significance of structural and agential, ideational, and material factors cannot be established empirically, then we must seek to avoid all claims which suggest that it might. Sadly, such claims are commonplace. Even Wendt himself, doyen both of the “structure–agency problematique” and of constructivism in international relations theory, is not above such conceptual confusions. Consider the following passage from an otherwise exemplary discussion co-written with Ian Shapiro:

The differences among … “realist” models of agency and structure—and among them and their individualist and holist rivals—are differences about where the important causal mechanisms lie in social life. As such, we can settle them only by wrestling with the empirical merits of their claims about human agency and social structure … These are in substantial part empirical questions.

(Wendt and Shapiro 1997, 181, emphasis added)

Wendt and Shapiro are surely right to note that ontological differences such as those between, say, more agency-centered and more structure-centered accounts, tend to resolve themselves into differences about where to look for and, indeed, what counts as important causal mechanisms in the first place. This implies that ontology precedes epistemology. Such a view is entirely consistent with the argument of the previous section—we must decide what exists out there to know about (ontology) before we can consider what knowledge we might acquire of it (epistemology), let alone how we might go about acquiring that knowledge (methodology). Yet having noted this, Wendt and Shapiro almost immediately abandon the logic it implies, suggesting that we might choose between contending ontologies on the basis of what we observe empirically. Surely this now implies that epistemology precedes ontology. If our ontology informs where we look for causal mechanisms and what we see in the first place (as they contend), then how can we rely upon what we observe to adjudicate between contending ontologies?

Wendt and Shapiro’s confusion is further compounded in the passage which immediately follows, in which a Popperian logic of falisifiability is invoked:

The advocates of individualism, structuralism and structuration theory have all done a poor job of specifying the conditions under which their claims about the relationship of agency and social structure would be falsified. (Wendt and Shapiro 1997, 181)

Here again we see direct appeal to the possibility of an epistemological refutation of ontological propositions. The point is that, as ontological positions, individualism, structuralism, and structuration theory cannot be falsified—our preference between them has to be adjudicated differently. A similar conflation underpins Wendt’s recent prescriptive suggestion that “ontology talk is necessary, but we should also be looking for ways to translate it into propositions that might be adjudicated empirically” (1999, 37). If only this were possible. When, as Wendt himself notes, ontological sensitivities inform what is “seen” in the first place and, for (philosophical) realists like himself, provide the key to peering through the mists of the ephemeral and the superficial to the structured reality beneath, the idea that ontological claims as to what exists can (p. 469) be adjudicated empirically is rendered deeply suspect. Quite simply, perspectives on the question of structure and agency, or any other ontological issue for that matter, cannot be falsified—for they make no necessary empirical claim. It is for precisely this reason that logical positivists (like Popper) reject as meaningless ontological claims such as those upon which realism and structuration theory are premised.7

It is important, then, that we avoid claiming empirical license for ontological claims and assumptions. Yet arguably more important still is that we resist the temptation to present positions on, say, the structure–agency question as universal solutions for all social scientific dilemmas. In particular, social ontologies cannot be brought in to resolve substantive empirical disputes. Giddens’ structuration theory can no more tell me who will win the next US presidential election than the theory of predestination can tell me whether my train will arrive on time tomorrow. The latter might be able to tell me that the movements of trains is etched into the archaeology of historical time itself, just as the structuration theorist might tell me the next US presidential election will be won and lost in the interaction between political actors and the context in which they find themselves. Neither is likely to be of much practical use to me, nor is it likely to provide much consolation if my train is late and my preferred candidate loses. It is important, then, that we do not expect too much from “solutions” to ontological “problems.”

4 Ontological Disputes in Political Analysis

Of all issues in political ontology, it is the related though by no means interchangeable (see Pettit 1993) questions of the relationship between individuals and social collectivities and between structure and agency that have undoubtedly attracted the most sustained attention and reflection over the longest period of time. A rather more recent set of concerns relates to the question of the relationship between the material and the ideational as (related or independent) dimensions of political reality. In the brief sections which follow, I consider each set of issues in turn.

4.1 The Individual–Group Relationship

In political analysis and the philosophy of the social sciences more broadly there is no more hardy perennial than the question of the relationship between individuals and (p. 470) social collectivities or groups (see Fay 1996, ch. 3; Gilbert 1989; Hollis 1994; Pettit 1993; Ryan 1970, ch. 8). Can collective actors (states, political parties, social movements, classes, and so forth) realistically, or indeed just usefully, be said to exist? If so, do they exhibit organic qualities, such that their character or nature is not simply reducible to the aggregation of the constituent units (generally individual actors) from which they are forged? Are such entities (if that is indeed what they are) appropriate subjects of political analysis and, if so, what if any behavioural characteristics can be attributed to them?

These and other related ontological questions have divided political analysts, and will no doubt continue to divide political analysts, as they have divided philosophers, for centuries. Generally speaking the controversy they have generated has seen protagonists resolve themselves with one of two mutually exclusive positions. These are usually labeled “individualism” and “holism” and they are often defined in mutually antagonistic terms. As Margaret Gilbert explains, ontological individualism is simply the doctrine that “social groups are nothing over and above the individuals who are their members” (1989, 428). It tends to be associated with a further, analytical set of claims, namely that what she terms “everyday collectivity concepts” (states, classes, parties, and other groups) “are analysable without remainder in terms of concepts other than collectivity concepts, in particular, in terms of the concept of an individual person, his [sic] goals, beliefs and so on” (1989, 434–5).

Holism, by contrast, is invariably understood as the simple denial of individualism, the doctrine that “social groups exist in their own right” (1989, 428) or, in Brian Fay’s more applied terms, that “the theories which explain social phenomena are not reducible to theories about the individuals which perform them” (1996, 50). In its more extreme variants, however, holism is less a belief in the organic nature of social and political reality than the dogmatic assertion that the task of social and political analysts is exclusively to document the (causal) role of social, i.e. holistic, phenomena, processes, and dynamics (cf. Ryan 1970, 172). In this form, holism, though very much in vogue in the 1970s, is now little more than a term of abuse within contemporary political science. It might be tempting, then, to see the dispute having been resolved in favor of individualism. This, however, would be too rash an inference to draw. For although most analytical routes in political science today lead from individualism, many make considerable concessions, as we shall see, to holism.

The dispute, as already indicated, is a timeless one, with perhaps the most eloquent defender of ontological (and, indeed, methodological) individualism being John Stuart Mill:

The laws of the phenomena of society are, and can be, nothing but the laws of the actions and passions of human being united together in the social state… Men [sic] are not, when brought together, converted into another kind of substance, with different properties… Human beings in society have no properties but those which are derived from, and may be resolved into, the laws of nature of individual man. (1970 [1843], 573; cited in Hollis 1994, 10)

Though, as is often noted, Mill was by no means consistent in keeping to the strictures of such an individualism and can be found at various times on the other side of the (p. 471) fence, he is a seemingly obligatory first citation for those asserting or defending their (ontological) individualism.

Unremarkably, the most dogged contemporary defence of individualism is found in rational choice theory. Jon Elster is characteristically incisive in claiming that “the elementary unit of social life is the individual human action.” Consequently “to explain social institutions and social change is to show how they arise as the result of the action and interaction of individuals” (1989, 13). What is unusual about this comment is that, unlike most rational choice theory, it seeks to present and defend individualism in ontological, rather than in more narrowly methodological, terms. Yet even rational choice theory, resolutely committed as it remains to methodological individualism, has made significant concessions to the organic qualities of social and political collectivities identified by holists. Indeed, in this respect, the developmental trajectory of rational choice in recent years is suggestive of something of an emerging ontological consensus amongst political analysts. Two points might here be made. First, whilst there have always been those who have presented rational choice theory in such terms (most notably, Friedman 1953, 14–15), many more contemporary rational choice theorists seem prepared to accept the ontological irrealism of rational choice assumptions, defending such premises in terms of their analytical utility not their correspondence to an external reality (for a more sustained discussion, see Hay 2004). Second, the move by many rational choice theorists, particularly so-called rational choice institutionalists, from an absolute towards a “bounded,” i.e. contextdependent, conception of rationality significantly qualifies and arguably violates any purist defence of ontological or, indeed, methodological individualism. For, put simply, if the stylized rational actor’s utility- and/or preference-function is a product of her context, role, or systemic function (as in much contemporary rational choice institutionalism), then to explain her behavior or to predict the consequences of her behavior in terms of such a utility/preference-function is no longer to subscribe to a methodological individualism.

As this perhaps suggests, however seemingly entrenched holism and individualism have, on occasions, become, a commonsense ground between such antagonistic extremes exists and is inhabited by a growing number of political analysts. Such a position accepts, ontologically, the following: (1) that a social whole is “not merely the sum of its parts”; (2) that there are “holistic properties” of such social wholes; (3) that these “can sensibly be said to belong to the whole and not to any of the parts”; and yet (4) that dismantle the whole and we are left with the parts and “not them and some mysterious property which formerly held the whole thing together” (Ryan 1970, 181).

4.2 The Structure–Agency Relationship

No less classical or disputed an issue in the philosophy of the social sciences is the question of the structure–agency relationship. Though closely related, it is by no means irreducible to the question of the relationship between groups and individuals and has been far more hotly contested than the latter in recent years.

(p. 472) Though space does not permit a detailed review of the literature, the key trends can nonetheless be established relatively simply (for more sustained discussion see Hay 1995; 2002, 89–134):

  • The proliferation of interest in the relationship between structure and agency has in fact been remarkably consensual, with scholars in political science and international relations rounding on both structuralist and intentionalist tendencies.8

  • In so doing they have come to champion a range of perspectives from social theory, notably Giddens’ (1984) structuration theory and the critical realists’ strategic-relational approach (Bhaskar 1979; 1989; Jessop 1990; 1996).

  • What each of these perspectives shares is the attempt to explore the dynamic interplay of structure and agency.

In short, and almost without exception, those who have reflected in a sustained fashion upon the question of structure and agency have done so with an increasing sense of frustration at the tacit intentionalism or, more usually, structuralism of existing mainstream approaches to political analysis. In particular they have found structuralism lurking in some apparently unlikely places. Chief amongst these is rational choice theory.

As a perspective which emphasizes the rationality exhibited by conscious and reflective actors in the process of making choices, it is difficult to imagine an approach that is seemingly more attentive to agency. However, impressions can be deceptive. For, within any rational choice model, we know one thing above all: that the actor will behave rationally, maximizing his or her personal utility. Consequently, any rational actor in a given context will choose precisely the same (optimal) course of action. Actors are essentially interchangeable (Tsebelis 1990, 43). Moreover, where there is more than one optimal course of action (where, in short, there are multiple equilibria), we can expect actors’ behavior to be distributed predictably between—and only between—such optima. What this implies is that the agent’s “choice” is rendered predictable (and, in the absence of multiple equilbria, entirely predictable) given the context. The implications of this are clear. We need know nothing about the actor to predict the outcome of political behavior. For it is independent of the actor in question. Indeed, it is precisely this which gives rational choice modes of explanation their (much cherished) predictive capacity.

In short, it is only the substitution of a fixed preference function for an indeterminate actor that allows a spurious and naturalist notion of prediction to be retained in rational choice (see also Hay 2004). Render the analytical assumptions about the individual actor more complex and realistic (by recognizing some element of contingency) and rational choice models become indeterminate.

This raises a final and important point, something of a leitmotif of this chapter. The rise of political ontology has increasingly led to a series of challenges to naturalism (a belief in the possibility of a unity of method between the natural and social sciences) (p. 473) and to naturalistic political science more specifically. The above paragraphs provide but one example. As they suggest, rational choice theory can deliver a naturalist science of politics only by virtue of the implausible (ontological) assumptions it makes about the universally instrumental, self-serving, and utility-maximizing character of human conduct. These serve, in effect, to empty agency of any content such that the actor becomes a mere relay for delivering a series of imperatives inherent in the context itself. In short, a naturalist science of politics is only possible if we assume what we elsewhere deny—that all actors, in any given context, will act in a manner rendered predictable (in many cases fully determinate) by the context in which they find themselves. Soften the assumptions, or even the universality of the assumptions, and the fragile edifice of naturalism crumbles. With it must go the universal pretensions of much rational choice theory and, indeed, the very possibility of a predictive science of the political.

4.3 The Ideational–Material Relationship

Very similar themes emerge in the burgeoning literature on the relationship between the ideational and the material and the extent to which ideas may be accorded a causal and/or constitutive role in the determination of political outcomes.9 Here, once more, the key question relates to the limits of naturalism. In particular, it is suggested, the existence of an irredeemably cognitive dimension to the social and political world for which there is no direct equivalent or analogue in the natural world, presents profound ontological impediments to a naturalist social science.

Once again there has been a considerable degree of harmony and consensus amongst those who have addressed these issues in ontological terms. The result is a convergence upon, and consolidation of, a position usually labeled constructivism in international relations theory, and usually seen as a development of historical institutionalism in political science (for a useful review see Blyth 2003). It defines itself in opposition to the materialist and naturalist rump of mainstream political science and international relations.

Like the qualified materialism of many contemporary rational choice institutionalists and neo-realists,10 constructivists start from the recognition that we cannot hope to understand political behavior without understanding the ideas actors hold about the environment in which they find themselves. Yet here the materialists and the constructivists part company, with the latter refusing to see such ideas as themselves reducible to ultimately determinant material factors (such as contextually given interests). Consequently, they accord ideas an independent causal role in political explanation. Nonetheless, whilst it is important not simply to reduce the ideational to a reflection, say, of underlying material interests, it is equally important not to subscribe to a voluntarist idealism in which political outcomes might be read off, more (p. 474) or less directly, from the desires, motivations, and cognitions of the immediate actors themselves. What is required, instead, is a recognition of the complex interaction of material and ideational factors. Political outcomes are, in short, neither a simple reflection of actors’ intentions and understandings nor of the contexts which give rise to such intentions and understandings. Rather, they are a product of the impact of the strategies actors devise as means to realize their intentions upon a context which favors certain strategies over others and does so irrespective of the intentions of the actors themselves.

Constructivism is, however, a broad church, encompassing a diverse range of positions. At the idealist end of the spectrum we find varieties of “thick” constructivism keen to privilege the constitutive role of ideas whilst not entirely denying the significance of material factors. At the other end of the spectrum we find varieties of critical realism whose rather “thinner” constructivism tends to emphasize instead the constraints the material world places on such discursive constructions.11 What each of these positions shares, however, is a complex or dialectical view of the relationship between the ideational and the material and a rejection of the possibility of a naturalist social science.

5 Conclusion

As the previous sections have sought to demonstrate, the proliferation of literature on political ontology in recent years has produced (or perhaps reflected and reinforced) a remarkable consensus. The vast majority of authors who have interrogated systematically the relationships between structure and agency and the material and the ideational as ontological issues, have, for instance, come subsequently to promote a post-naturalist, post-positivist approach to social and political analysis premised upon the acknowledgement of the dynamic interplay of structure and agency and material and ideational factors. In so doing they have pointed to a consistent disparity between the often tacit and normalized analytical assumptions of existing mainstream approaches to political analysis and those which emerge from sustained ontological reflection.

In particular they have challenged the often parsimonious and self-confessedly unrealistic analytical assumptions which invariably make naturalist approaches to political science possible. This is undoubtedly a useful exercise and has already given rise to genuinely novel approaches to political analysis and a series of important insights (the contributions of the new constructivist–institutionalist synthesis being a case in point). Yet it can be taken too far. In one sense it is unremarkable that political ontologists, interested principally in the extent to which the complexity and (p. 475) contingency of the “real world” of social and political interaction might be captured, encourage us to choose complex, credible, and realistic analytical assumptions. Yet this is not a costless move. Simple, elegant, and parsimonious analytical assumptions are unlikely to satisfy the political ontologist, but this may not be sufficient reason to jettison them. However unrealistic they may be, they have an appeal and can certainly be defended in the kind of pragmatic terms that are unlikely to feature prominently in the ontologist’s deliberations. Here, as elsewhere, clear trade-offs are involved. Political ontology can certainly help us to appreciate what is at stake in such choices, providing something of a counterbalance to the mainstream’s characteristic silence on its most central assumptions, but it cannot be allowed to dictate such choices alone.


Adler, E. 1997. Seizing the middle ground: constructivism in world politics. European Journal of International Relations, 3: 319–63.Find this resource:

Archer, M. S. 1998. Social theory and the analysis of society. Pp. 69–85 in Knowing the Social World, ed. T. May and M. Williams. Buckingham: Open University Press.Find this resource:

Benton, T. and Craib, I. 2001. Philosophy of Social Science: The Philosophical Foundations of Social Thought. Buckingham: Open University Press.Find this resource:

Bhaskar, R. 1979. The Limits of Naturalism. Brighton: Harvester Wheatsheaf.Find this resource:

——1989. Reclaiming Reality. London: Verso.Find this resource:

Blaikie, N. 1993. Approaches to Social Enquiry. Cambridge: Polity.Find this resource:

Blyth, M. 2003. Structures do not come with an instruction sheet: interests, ideas and progress in political science. Perspectives on Politics, 1: 695–706.Find this resource:

Brownmiller, S. 1975. Against Our Will: Men, Women and Rape. London: Secker and Warburg.Find this resource:

Carlsnaes, W. 1992. The agent–structure problem in foreign policy analysis. International Studies Quarterly, 6: 245–70.Find this resource:

Cerny, P. G. 1990. The Changing Architecture of Politics: Structure, Agency and the Future of the State. London: Sage.Find this resource:

Christiansen, T., Jørgensen, K. E., and Wiener, A. 2001. Introduction. Pp. 1–21 in The Social Construction of Europe, ed. T. Christiansen et al. London: Sage.Find this resource:

Daly, M. 1978. Gyn/Ecology: The Metaethics of Radical Feminism. Boston: Beacon Press.Find this resource:

Denzau, A. T. and North, D. C. 1994. Shared mental models: ideologies and institutions. Kyklos, 47: 3–31.Find this resource:

Dessler, D. 1989. What’s at stake in the agent-structure debate? International Organization, 43: 441–73.Find this resource:

Elshtain, J. B. 1981. Public Man, Private Woman: Women in Social and Political Thought. Princeton, NJ: Princeton University Press.Find this resource:

Elster, J. 1989. Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press.Find this resource:

Fay, B. 1996. Contemporary Philosophy of Social Science: A Multicultural Approach. Oxford: Blackwell.Find this resource:

Friedman, M. 1953. Essays in Positive Economics. Chicago: University of Chicago Press.Find this resource:

Fuller, S. 1998. From content to context: a social epistemology of the structure–agency craze. Pp. 92–117 in What is Social Theory? The Philosophical Debates, ed. A. Sica. Oxford: Blackwell.Find this resource:

(p. 476) Giddens, A. 1984. The Constitution of Society. Cambridge: Polity.Find this resource:

Gilbert, M. 1989. On Social Facts. Princeton, NJ: Princeton University Press.Find this resource:

Goldstein, J. and Keohane, R. O. (eds.) 1993. Ideas and Foreign Policy: Beliefs, Institutions and Political Change. Ithaca, NY: Cornell University Press.Find this resource:

Grossmann, R. 1992. The Existence of the World: An Introduction to Ontology. London: Routledge.Find this resource:

Hay, C. 1995. Structure and agency. Pp. 189–206 in Theory and Methods in Political Science, ed. D. Marsh and G. Stoker. London: Macmillan.Find this resource:

——. 2002. Political Analysis. Basingstoke: Palgrave.Find this resource:

——2004. Theory, stylised heuristic or self-fulfilling prophecy? The status of rational choice theory in public administration. Public Administration, 82: 39–61.Find this resource:

Hollis, M. 1994. The Philosophy of Social Science. Cambridge: Cambridge University Press.Find this resource:

Honderich, T. (ed.) 1995. The Oxford Companion to Philosophy. Oxford: Oxford University Press.Find this resource:

Jacquette, D. 2002. Ontology. Chesham: Acumen.Find this resource:

Jessop, B. 1990. State Theory: Putting Capitalist States in their Place. Cambridge: Polity.Find this resource:

——1996. Interpretative sociology and the dialectic of structure and agency. Theory, Culture and Society, 13 (1): 119–28.Find this resource:

King, G., Keohane, R. O., and Verba, S. (eds.) 1994. Designing Social Inquiry: Scientific Inference in Qualitative Research. Princeton, NJ: Princeton University Press.Find this resource:

Kenny, M. and Smith, M. J. 1997. (Mis)Understanding Blair. Political Quarterly, 68: 220–30.Find this resource:

Mäki, U. 2001. Economic ontology: what? why? how? Pp. 3–14 in The Economic Worldview: Studies in the Ontology of Economics, ed. U. Mäki. Cambridge: Cambridge University Press.Find this resource:

——2002. The dismal queen of the social sciences. Pp. 3–32 in Fact and Fiction in Economics: Models, Realism and Social Construction, ed. U. Mäki. Cambridge: Cambridge University Press.Find this resource:

Marsh, D. and Furlong, P. 2002. A skin not a sweater: ontology and epistemology in political science. Pp. 17–41 in Theory and Methods in Political Science, ed. D. Marsh and G. Stoker, 2nd edn. Basingstoke: Palgrave.Find this resource:

——and Stoker, G. 2002. Introduction. Pp. 1–16 in Theory and Methods in Political Science, ed. D. Marsh and G. Stoker, 2nd edn. Basingstoke: Palgrave.Find this resource:

Mill, J. S. 1970. A System of Logic. London: Longman; originally published 1843.Find this resource:

Monroe, K. R. (ed.) 2004. Perestroika, Methodological Pluralism, Governance and Diversity in Contemporary American Political Science. New Haven, Conn.: Yale University Press.Find this resource:

North, D. C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press.Find this resource:

Pettit, P. 1993. The Common Mind: An Essay on Psychology, Society and Politics. Oxford: Oxford University Press.Find this resource:

Ryan, R. 1970. The Philosophy of the Social Sciences. Basingstoke: Macmillan.Find this resource:

Schmitt, F. F. 2003. Socialising metaphysics: an introduction. Pp. 1–37 in Socialising Metaphysics, ed. F. F. Schmitt. Boulder, Colo.: Rowman and Littlefield.Find this resource:

Smith, M. J. 1998. Reconceptualising the British state: theoretical and empirical challenges to central government. Public Administration, 76: 45–72.Find this resource:

——1999. The Core Executive in Britain. London: Macmillan.Find this resource:

Smith, S. 1990. Positivism and beyond. Pp. 11–46 in International Theory: Positivism and Beyond, ed. S. Smith, K. Booth, and M. Zalewski. Cambridge: Cambridge University Press.Find this resource:

Suganami, H. 1999. Agents, structures, narratives. European Journal of International Relations, 5: 365–86.Find this resource:

(p. 477) Tilly, C. and Goodin, R. E. 2006. It depends. Pp. 3–32 in The Oxford Handbook of Contextual Political Analysis., ed. R. E Goodin and C. Tilly. Oxford: Oxford University Press.Find this resource:

Tsebelis, G. 1990. Nested Games: Rational Choice in Comparative Politics. Berkeley: University of California Press.Find this resource:

Wendt, A. 1999. Social Theory of International Politics. Cambridge: Cambridge University Press.Find this resource:

——and Shapiro, I. 1997. The misunderstood promise of realist social theory. Pp. 166– 90 in Contemporary Empirical Political Theory, ed. K. R. Monroe. Berkeley: University of California Press.Find this resource:

Wolf, N. 1993. Fire With Fire: The New Female Power and How to Use It. New York: Fawcett Combine.Find this resource:

Young, I. M. 1990. Throwing Like a Girl and Other Essays in Feminist Philosophy and Social Theory. Bloomington: Indiana University Press.Find this resource:


(2) Whilst the appropriate preference function(s) and behavioral assumptions that we should adopt in, for instance, game-theoretic modeling has been a focus of considerable attention, the vast majority of that reflection has failed to acknowledge the ontological character of the issue.

(3) For, clearly, what counts as evidence in the first place depends on one’s view of the relationship between that which is observed and experienced, on the one hand, and that which is real, on the other. Where the (archetypal) pluralist sees an open and democratic decision-making process, the (similarly archetypal) elite theorist sees the work of covert agenda-setting processes behind the scenes, and the (no less archetypal) Marxist, evidence of preference-shaping ideological indoctrination.

(4) In the much-lauded second edition of their highly respected and influential text on Theory and Methods in Political Science, for instance, the editors and contributors display a marked lack of consistency in defining ontology and epistemology. Given that theirs is practically the only entry-level introduction to these topics currently available to students of political science, this is all the more tragic. Thus, in their introductory essay, David Marsh and Gerry Stoker suggest, quite remarkably, that “ontology is concerned with what we can know and epistemology with how we can know it” (2002, 11). Yet in the first substantive chapter of the volume, David Marsh, this time with Paul Furlong, defines ontology (correctly) as “a theory of being” and suggests that epistemology relates to “what we can know about the world” and (more problematically) “how we can know it” (2002, 18–19). Of these, only the second definition of ontology is entirely unproblematic.

(5) To suggest that our ideas influence our conduct and that our conduct has, in turn, the capacity to reshape our environment is not, of course, to insist that it necessarily does so in any given setting over any particular time-horizon. It is to suggest, however, that insofar as conduct serves to shape and reshape a given political landscape, the ideas held by actors about that context are crucial to any understanding of such a process of political change (see also Rueschemeyer, in Oxford Handbook of Contextual Political Analysis, ed. R. E. Goodin and C. Tilly 2006. Oxford: Oxford University Press).

(6) This is largely because the process of presenting evidence is invariably one which situates it ontologically (with respect to often tacit ontological assumptions, such as the extent, if any, of a separation of appearance and reality).

(7) However tempting this strategy may seem, however, it does not provide an escape from ontological issues and choices. For, as indicated earlier, whether we choose to acknowledge them or nor, political analysis necessarily proceeds on the basis of ontological assumptions.

(8) See, for instance, Adler 1997; Carlsnaes 1992; Cerny 1990; Dessler 1989; Kenny and Smith 1997; Smith 1998; 1999; Suganami 1999; Wendt 1987. For a review, see Hay 1995.

(9) See Hay (2002, 194–215) for a more sustained discussion.

(11) For a variety of different positions within this spectrum compare the various contributions to Christiansen, Jørgensen, and Wiener (2001).