Abstract
This article evaluates Mark Bevir and Rod Rhodes’ interpretive approach to political analysis by examining their account of social change. Their work is significant because they have endeavoured to construct a distinctive approach which strikes a productive balance between philosophical reflection and analytical attention to the empirical domain. Moreover, in elaborating their approach, Bevir and Rhodes wage a war along a number of fronts: positivism, institutionalism and post-structuralism, and so an analysis of their work enables us to take stock of a range of contemporary methods and approaches. In considering their underlying presuppositions and commitments, we pay particular attention to their account of human agency and its relationship to social structures and political power.
While we agree with much of their critique of positivism, naturalism, realism and institutionalism, we argue that Bevir and Rhodes risk overplaying the role of interpreting the individual beliefs and desires of relevant actors to the detriment of a wider net of social practices and logics. Moreover, in challenging their understanding of the post-structuralist approach to political analysis we develop its resources to enrich the possibilities of a critical interpretivism, moving beyond concepts like tradition, dilemma and situated agency. Put more precisely, we argue that the radical contingency of social structures and human agency – their structural incompleteness – discloses new ways of understanding both their character and their mutual intertwining. For example, we develop the categories of lack, dislocation and political identification to think human agency and its relation to wider social structures. More broadly, we argue that an approach developed around different sorts of logics – social, political and fantasmatic – goes some way to steering a different course between a pure thick descriptivism that focuses principally on individual beliefs and desires on the one hand, and a concern with causal laws and mechanisms on the other.
We can understand and explain practices and actions adequately only by reference to the beliefs and desires of the relevant actors. Hence to study political life adequately we have to engage in the interpretation of the beliefs and desires of those we study (Bevir and Rhodes, 2003, p. 18).
Recent discussions about the character, purpose and methods of political science have highlighted a number of problems in law-like and positivist accounts of political science. Critics justly raise questions about the latter's strong commitments to predictive modelling, causal explanation, the separation of fact and value, and the splitting of description and explanation from critique and evaluation (see Flyvbjerg, 2001; Schram and Caterino, 2006; Topper, 2005). The debates have also brought a renewed interest in alternative approaches to political analysis, such as those put forward by critical realists, 1 neo-positivists, 2 post-structuralists and postmodernists, 3 as well as interpretivists. In this article we focus on Mark Bevir and R. A. W. Rhodes’ version of the latter approach.
In general, interpretivism's protagonists draw upon a variety of theoretical traditions to articulate a distinctive critique and response to the positivist mainstream. Indeed, there has been a proliferation of interpretive approaches in the social and political sciences, including developments in the history of ideas; the study of political culture and institutions; analyses of public policy; the exploration of international relations; not to mention the analysis of political ideologies, discourses and media representations. 4 But while interpretivism comes in many forms, its proponents are unified in their opposition to the causal law model, and their desire to understand the meanings of human actions and practices in particular contexts. Indeed, the common denominator of this approach is a critique of the naturalistic ideal of causal explanation, including the pre-eminence accorded to the role of prediction as the key criterion for discriminating among competing explanatory hypotheses. But equally important is the adoption of a perspective that seeks ‘from within’ to make intelligible the meanings and reasons social agents give for their actions and practices. This approach contrasts with the objectivistic standpoint of the positivist mainstream, which adopts a stance ‘outside’ the social and political phenomena it seeks to explain. Expressed more positively, those within the hermeneutical tradition focus on the practice of interpreting the self-interpretations of social actors in particular historical contexts. Contextualised self-interpretations thus become the key component of the hermeneutical logic of explanation in the social sciences.
We evaluate the hermeneutical challenge to the law-like model by examining the work of Bevir and Rhodes, which has endeavoured to construct a distinctive interpretivist approach to political analysis. Their work is significant because it tends to balance philosophical reflection with much greater analytical attention to the empirical domain. They thus seek to connect general hermeneutical insights about the human and social sciences through a more systematic analysis of concrete political phenomena. Their work is important not least because it seeks to develop and justify the philosophical presuppositions of interpretivism in the context of applying their particular approach to a series of concrete issues in British governance and policy making. Moreover, in elaborating their approach, Bevir and Rhodes wage a war along a number of fronts: positivism, institutionalism and post-structuralism, and so an analysis of their work enables us to take stock of a range of contemporary methods and approaches. In so doing, our review focuses initially on the basic arguments presented in Governance Stories, which gathers together a number of papers reflecting on the nature and role of interpretation, analysing the history of various traditions of political science and models of government and governance, as well as conducting ethnographies of several governance regimes and practices (Bevir and Rhodes, 2006a). We also refer to a related set of their writings, in order to contextualise and broaden the relevance of our intervention.
It is important to stress, however, that our concern is not so much with an evaluation of their empirical studies and conclusions. This is a central focus of Martin Smith's intervention, which queries whether concepts like beliefs and tradition are adequate to the task of explaining various governance practices, and the precise role attributed to myths like the ‘Westminster model’ in accounting for the key institutions of British government (Smith, 2008). While these issues are clearly significant, they are not our primary concern in this article. Instead, we seek to burrow beneath the specific empirical claims presented in this book in order to consider their underlying presuppositions and commitments, and to contextualise their overall approach. We thus concentrate on their general interpretivist approach to political analysis, paying particular attention to their account of human agency and its relationship to social structures and political power.
In a nutshell, we claim first that Bevir and Rhodes overplay the role of interpreting the individual beliefs and desires of relevant actors to the detriment of a wider net of social practices and logics, and second that they not only provide a caricature of ‘the’ post-structuralist account of the so-called ‘structure-agency’ debate, but that in so doing they miss an opportunity to take advantage of an important set of conceptual and theoretical resources with which to address this thorny issue in contemporary political analysis. By contrast, we draw upon a range of post-structuralist thinkers to enrich the possibilities of a critical interpretivism. Put more precisely, we argue that the radical contingency of social structures and human agency – their structural incompleteness – discloses new ways of understanding both their character and their mutual intertwining. We suggest that an approach developed around different sorts of logics – social, political and fantasmatic – can go some way to steering a different course between a pure thick descriptivism that focuses principally on individual beliefs and desires on the one hand, and a concern with causal laws and mechanisms on the other (see Glynos and Howarth, 2007). In elaborating these claims, we begin by setting out the basic contours of Bevir and Rhodes’ approach.
Analytical Interpretivism: Tradition, Dilemmas, Situated Agency
Bevir and Rhodes present us with a relatively clear and parsimonious model of critical explanation in political science. Drawing upon and modifying Donald Davidson's picture of human action, their first task is to ‘unpack’ the beliefs and desires of individual actors in an effort to explain the latter's intentions – and hence their reasons – for acting the ways they did. Indeed, the description and analysis of such intentions and reasons furnish the means for explaining the causes of their actions (Bevir and Rhodes, 2005, p. 177; see also Bevir 2006). The second task is to construct ‘narratives’ which ‘point to the beliefs and desires that cause the actions’ (Bevir and Rhodes, 2005, p. 177). Rival narratives are then compared and evaluated by determining which ‘best fit’ the available facts. This opens the way for a critical evaluation of the different stories and, more importantly, prepares the way for a critique of the actual beliefs, practices and traditions so interpreted (Bevir and Rhodes, 2005, p. 183).
This emphasis on what they call a ‘narrative form of explanation’, or the construction of narratives in political science, is analogous for them to the role of theory building and theory testing in the natural sciences (Bevir and Rhodes, 2003, p. 20). 5 In this conception a fact is ‘a statement, typically about a piece of evidence, which nearly everyone in the relevant community would accept as true’, and narratives ‘explain shared facts by postulating significant relationships, connections, or similarities between them’ (Bevir and Rhodes, 2006a, p. 28). The upshot of these thoughts is to problematise the evaluation of rival stories by simply assessing whether they do or do not fit the facts. Instead, in a way that is reminiscent of Taylor's device of ‘perspicuous contrast’, Bevir and Rhodes advocate the comparison of ‘bundles of narratives’ by assessing ‘their success in relating facts to one another, highlighting similarities and differences, and exploring continuities and disjunctions’ (Bevir and Rhodes, 2006a, p. 28).
The more general theoretical point that emerges is that they share with other hermeneuticists such as Charles Taylor and Peter Winch the belief that we need to move beyond self-interpretations to the practice of interpreting self-interpretations. For example, they point out how ‘interviewees can be self-serving and misleading’ and that ‘the validity and reliability of “facts”’ thereby attained ‘can always be disputed’ (Bevir and Rhodes, 2003, p. 193). The belief is also evident when they argue that the case for interpretivism ‘should not be confused with the claim that political scientists must accept actors’ own accounts of their beliefs’ (Bevir and Rhodes, 2005, p. 178). In short, interpretivists (of various hues) agree that interpreting self-interpretations yields contextualised self-interpretations as the prime means of elucidating social and political phenomena. But Bevir and Rhodes’ work is especially attractive because it holds out the promise of furnishing us with a set of concepts with which to explain social and political change beyond the self-interpretations of the actors themselves. As we shall see, their concepts of dilemma, tradition and situated agency seek to move us in precisely this direction.
To begin, the notion of a tradition refers ‘to the initial human condition of being thrown into a world already inhabited not only by people, but also by their definite beliefs’. In other words, understood as ‘a set of understandings someone acquires as an initial web of beliefs during a process of socialization’ (Bevir, 1999, p. 200), traditions are ‘contingent entities’ which ‘people produce by their own activities’ (Bevir, 1999, p. 203). This means that the notion of a tradition incorporates both inherited beliefs and transformed beliefs, ‘as they are handed down from generation to generation’ (Bevir and Rhodes, 2005, p. 173). And while the notion of tradition is evocative of social structures ‘in which individuals are born’, and ‘which then acts as the background of their beliefs and actions’, individuals can ‘adapt, develop, and reject much of this inheritance’ (Bevir and Rhodes, 2005, p. 173). The concept of tradition therefore makes possible an analysis of individuals who ‘inherit’ – and are thus inevitably influenced by – ‘a set of beliefs that forms the background to their later reasoning’, but who can nevertheless via their ‘local reasoning’ transform such beliefs over time (Bevir and Rhodes, 2005, p. 176).
The concept of ‘tradition’ is intimately connected to the idea of a ‘dilemma’. The latter is an explanatory concept that helps us better account for processes of social and political change. It relates to circumstances in which new beliefs enter a system of meaning, ‘where any new belief, merely by virtue of being adopted, poses a question of the web of beliefs into which it is inserted’ (Bevir, 1999, p. 243). Consequently, dilemmas arise all the time. In the context of natural science practice, for example, dilemmas are not just rare anomalies that prompt epochal changes to scientists’ webs of belief, but they are also ‘the concerns that prompt scientists to extend prevailing theories during a period of normal science, and even the trivial puzzles that lead all of us to adopt new beliefs all the time in our everyday existence’ (Bevir, 1999, p. 229). Dilemmas are deemed to occur when there is a mismatch or conflict between ‘perceptions of failings’ (in an existing pattern of government for example) and the existing beliefs of individuals, which ‘push them to reconsider beliefs and the intellectual tradition that informs those beliefs’ (Bevir and Rhodes, 2006b, p. 98).
It is this interplay between traditions and dilemmas that Bevir and Rhodes use to explore the question of social change. This is because change occurs when agents respond to novel ideas or problems. It is a result of people's ability to adopt beliefs and perform actions through a reasoning that is embedded in the tradition they inherit (Bevir and Rhodes, 2005, p. 173). Historians and researchers for their part must ‘make sense of the way people modify their web of beliefs by portraying the new beliefs as responses to dilemmas confronting the old ones’ (Bevir, 1999, p. 263). In short, the emergence of novel ideas or problems in a particular tradition offers opportunities for individuals to reflect and act in different ways, and the resultant social practices can have the effect of altering institutions and traditions. For example, evolutionary theory may provoke a dilemma for Christians who interpret the Bible literally, leading them to reject evolutionary theory or to change their beliefs (Bevir, 1999, p. 230). Or one might try to account for changes in civil service Acts, Codes and practices by appealing to the dilemma posed to our traditional understandings of public service ethos by new ideas about public management, open competition, efficiency and accountability (Bevir and Rhodes, 2003, pp. 134–5).
As a result, while critical of aggregate concepts and abstractions such as ‘structure’ or ‘ideology’, Bevir and Rhodes are not averse to acknowledging their importance in explaining political action and social change. However, the key question they must address is how to connect traditions and dilemmas with individual beliefs and desires, and thus to explain change. It is here that Bevir and Rhodes insist on the importance of what they call ‘situated agency’, which they develop by establishing a contrast between their position and their particular gloss of post-structuralist critiques of autonomy. While they endorse post-structuralist critiques of strong human autonomy, which do away with any form of structural constraint or empowerment, they are opposed to what they present as the complete abandonment of agency by post-structuralist theorists such as Foucault and Laclau. Laclau, for instance, explores the meanings of beliefs and actions of individuals within discursive contexts, in which the latter are understood ‘as the relations between … semantic units’. According to Bevir and Rhodes, however, this ‘allows almost no room for human agency’ indeed, for post-structuralists in general individuals are ‘mere effects of discourse’ (Bevir and Rhodes, 2006b, p. 91, p. 105). The basic claim here, then, is that post-structuralists reify entities such as discourses or ideologies – cultural schemas or social structures – thus precluding any understanding of human action and agency. As Bevir and Rhodes put it, post-structuralists are mistaken ‘to conceive of traditions as reified quasi-structures that somehow determine the beliefs that people come to hold’ (Bevir and Rhodes, 2005, p. 180). Instead, their alternative categories of traditions, dilemmas and situated agency entail neither the valuing of order as such, nor the idea that structures constitute insurmountable obstacles to change, but are compatible with a form of methodological individualism in which agents can alter their beliefs and traditions. As Bevir puts it, ‘Individuals are agents capable both of modifying traditions and of migrating across traditions’ (Bevir, 1999, p. 197).
Structure, Agency and Power
The crucial question, of course, is whether or not Bevir and Rhodes’ notions of ‘tradition’, ‘dilemma’ and ‘situated agency’ enable them to steer a safe passage between an ontological monism on the one hand, which essentialises social structures or social agents, or an ontological dualism that separates structures and agents into an unstable compromise formation on the other. We can approach this issue by examining Bevir and Rhodes’ critique of post-structuralist accounts of subjectivity and discourse. Of course, much has been written about post-structuralist political theory, especially its various understandings of the relationship between structure and agency. But since a good chunk of this seems almost wilfully to misrepresent or even caricature the post-structuralist position, it is worth starting our account by situating Bevir and Rhodes’ intervention in relation to this broader context.
At least three kinds of criticism can be identified. First, there are those who present post-structuralism (or postmodernism, with which it is often used interchangeably) as a revamped form of structuralism in which human agency is reduced to reified discursive structures. Second, there are those who accuse post-structuralists of a kind of super-voluntarism, in which agents can act or construct without any structural constraint at all. Finally, there are some who claim that post-structuralists deconstruct the structure/agency distinction as purely arbitrary and without value. For instance, in a well-known textbook on political analysis, Stuart McAnulla offers the following characterisation:
For the postmodernist there is no point in attempting to establish the ‘real’ relationship between structure and agency. Any understanding we have of the issue is viewed as one constructed in the language and discourse we use … There is no ‘structure’ and ‘agency’ which exists ‘out there’ to discover; they are merely concepts within a discourse through which we apprehend and construct the world around us (McAnulla, 2002, pp. 282–3). 6
While it would be difficult to see how our understanding of the issue could avoid being constructed in the language and discourse we use – how can it not be the case that ‘structure’ and ‘agency’ are concepts through which we apprehend and construct the world around us? – the charge is relatively clear: post-structuralists dissolve the issue of structure and agency into a tissue of linguistic differences that can be moulded at will.
Clearly, Bevir and Rhodes’ allegations fall more into the first class of criticisms. They charge post-structuralists with abandoning a concern with agency in the name of ‘reified’ cultural structures or ideologies, which turn individual agents into the mere effects of discourse. Instead, for Bevir and Rhodes, while individuals are always ‘thrown’, ‘situated’ or ‘socialised’ in particular traditions, they have the capacity to act in a reasonable fashion to transform the tradition in which they find themselves. The first problem with this account is that it tends to encourage a far too mechanistic way of linking self-interpretations to background contexts, and overemphasises the individual's creativity in coping with dilemmas by recourse to various strands of tradition. As Bevir and Rhodes put it, ‘individuals are not autonomous since they always experience and reason against the background of tradition, but they are agents who have the capacity to modify and reject any aspect of a tradition’ (Bevir and Rhodes, 2004, p. 16, emphasis added). Yet this raises a key question: why might some aspects of traditions exert greater appeal than others, or why might some aspects resist modification? This suggests a potentially promising line of thought which appears to escape the critical scope attributed by Bevir and Rhodes to their narrative interpretations. For the critical aspect of their explanations appears to be centred primarily upon unmasking ‘the partiality of a political interpretation’ and ‘the contingency of traditions and their interpretations’ (Bevir and Rhodes, 2004, p. 16). This reveals that the relationship between agency and structure is not sufficiently theorised, and lacks a sufficiently worked out conception of power and ideology (see also Finlayson, 2004). Indeed, we feel that this is partly because Bevir and Rhodes are overenthusiastic in their rejection of the ontological – as opposed to a simply contextual – import of structure, leading to an underdeveloped social ontology. 7
The second problem with this account – and the same holds for the other two kinds of criticism we mentioned – is that it completely misrepresents the post-structuralist account of structure and agency, at least as it has been developed by theorists such as Ernesto Laclau, as well as those who have sought to develop this approach further. To avoid further confusion and misrepresentation, it is important to reiterate what post-structuralists have actually said in developing their account of subjectivity, structure and power. In general, their strategy is first to deconstruct the structure/agency opposition, in which one element is prioritised over the other, while simultaneously seeking to re-inscribe this binary division into a new conceptual infrastructure (to use Derrida's terminology) that gives each element its due consideration within a distinctive ontological framework.
Here is how Laclau actually puts his position, taking care to contrast it with purely structuralist or voluntarist conceptions:
Let us assume we fully accept the structuralist vision: I am a product of structures; there is nothing in me with a separate substantiality from the discourses making me up; a total determinism governs my actions … But the question immediately arises: what happens if the structure I am determined by does not manage to constitute itself, if a radical outside – which does not share a common measure or foundation with the inside of the structure – dislocates it? The structure will obviously not be able to determine me, not because I have an essence independent from the structure, but because the structure has failed to constitute itself fully and thus to constitute me as a subject as well … The freedom thus won in relation to the structure is therefore a traumatic fact initially: I am condemned to be free, not because I have no structural identity as the existentialists assert, but because I have a failed structural identity (Laclau, 1990, p. 44, emphases in original).
And as against the view that the question of structure and agency is simply a matter of language and subjective construction, he goes on to argue that ‘The opposition between a society that is completely determined in structural terms and another that is entirely the creation of social agents is not an opposition between different conceptions of the social, but is inscribed in social reality itself’. Indeed, this complex dialectic follows from the fact that ‘the subject exists because of dislocations in the structure’ (Laclau, 1990, p. 60, emphases added).
Given structurally incomplete sets of social relations, then, subjects become active political subjects when they can no longer ‘go on’ in their normal, routinised fashion. That is to say, by articulating aspects of existentialism and psychoanalysis with post-structuralism – via theorists like Heidegger, Lacan and Žižek – it is possible for post-structuralists to sketch a movement away from the idea of subjectivity as simply a certain position within a discourse to the idea of a ‘radical subject’ that is constitutively incomplete and split. This conception of subject is predicated on at least three notions:lack, dislocation and identification, each of which enables us to distinguish subjectivity from identity, while also enabling us to address the importance of agency. To begin, the subject is marked by an internal lack or impossibility, which is only disclosed in certain circumstances. We may conceptualise these conditions as dislocatory events, as these are social situations where it is no longer clear how the subject is to ‘go on’ – how it is to follow the rules, for instance, or engage in its routinised practices. Lack is thus revealed when identities are exposed to situations where the contingency or the undecidability of dislocated social structures is made visible. It is in these situations of structural failure that we see the emergence of radical subjectivity, as subjects are literally compelled to identify with new objects and discourses to fill the lack made visible by a dislocatory event. As Laclau puts it, ‘the subject is nothing but the gap or distance between the undecidability of the structure and the moment of decision’ (Laclau, 1990, p. 30). Finally, this concept of radical subjectivity holds at both the individual and collective levels, as well as for non-political and political events and processes. I become a subject when I mislay my paper five minutes before the workshop is about to begin, and I scramble around in a panic trying to find it. But black schoolchildren and youths in Soweto in June 1976 became radical political subjects when they took on the might of the South African police and army with the demand to end Bantu Education in the name of ‘Black Power!’ and ‘Amandla!’.
It is clear, then, that post-structuralists such as Laclau do allow ‘room for human agency’. Indeed, one might be forgiven for thinking that Laclau concedes a too powerful – even decisionistic – role to human subjectivity and agency in constituting social structures, for (following Derrida) he sometimes alludes to Kierkegaard's claim that the moment of decision is akin to ‘the moment of madness’, where the taking of a decision is analogous to ‘impersonating God’, where God is that being ‘who has not to give [an] account of his actions before any tribunal of reason, because He is the source of rationality’ (Laclau, 1996, p. 56). But it is equally important to remember that for Laclau ‘all decision is taken within a certain structural context’ (Laclau, 2004, p. 322). And since the structure is never completely dislocated, human beings are ‘mortal gods’ – ‘those who have to fill the gaps resulting from the absence of God on Earth’– who can only simulate God; the ‘madness’ of our decisions always falls short of the omniscience of the latter (Laclau, 1996, p. 56). Instead of total indeterminacy, social actors (whether individual or collective) are always partially situated in a particular social context, in which their ‘decisions’ involve the foreclosure of some political options. This means ‘the madness of the decision is … as all madness, a regulated one’ (Laclau, 1996, p. 57; see also Laclau 2004). In short, there is no avoiding the concrete analysis of a particular historical conjuncture in order to explore the specific reasons and conditions in which the radical contingency of social structures and relations is made visible, thus offering new possibilities for action and identification. 8 Equally, it is important to see how these new acts of constitution become sedimented and routinised, so that their ultimate contingency is covered over; or conversely, why and how the transformation of institutions and social structures has been prevented.
Viewed against this background, Bevir and Rhodes’ (mis)representation of post-structuralism presents us with a series of classical ‘either/ors’: either language is constituted in a ‘bottom-up’ fashion on the basis of individual utterances and meanings, or it is a fully fledged and all-determining structure that fixes meaning from above, thus functioning to ‘avoid all appeals to human agency’ (Bevir and Rhodes, 2003, p. 22). Either the words and concepts of our language represent the world, or there is no representation and reference at all. 9 . Yet it is not clear why one should accept such loaded choices. After all, the whole point of post-structuralist theories of language and human subjectivity is to problematise the idea of a fully present subject and a fully constituted linguistic structure, so that we are not forced to choose between, on the one hand, an atomistic and referential conception of language, which is composed of individual utterances that refer to states of affairs, and on the other hand a concept of language as an all-pervasive substance which is completely closed and thus altogether external to subjectivity (compare Derrida, 1976; 1978). Following Heidegger and Wittgenstein, and as developed by Cavell and Derrida, we argue that linguistic signs and grammar have a different, more complex and intimate relationship with ‘reality’ (Mulhall, 2001, pp. 93–7, pp. 240–3).
Thus the post-structuralist position rejects both intentionalist and structuralist responses to the structure-agency dichotomy, and seeks instead to radicalise dialectical accounts by problematising the residual dualism in Giddens’ structuration theory and Bhaskar's critical realism (compare Hay, 1995; 2002, ch. 3). In other words, rather than prioritising totalised and determining social structures, on the one hand, or fully constituted subjects on the other, post-structuralists argue that we should begin by accepting that social agents always find themselves immersed in or ‘thrown into’ a system of meaningful practices, or what Lacan has called a ‘symbolic order’, which both shapes their identity and structures their practices. Crucially, however, these structures are ontologically incomplete entities that can never fully determine the identity of agents, or their ability to act. Indeed, it is in the ‘space’ or ‘gap’ of social structures, as they are rendered visible in moments of crisis and dislocation, that the notion of a political or radical subject can emerge through what might be called particular ‘acts of identification’. And as these identifications are understood as decisions between different objects – newly available beliefs, ideologies or discourses – some of which are excluded or repressed, while those that are identified with can never be fully actualised, then any form of identification is doomed to fall short of its full promise. Social structures and different forms of life are thus marked by gaps and fissures, as they are forged by political decisions and exclusions that repress certain possibilities. And this commitment to the ‘primacy of politics’ also brings with it the idea that social structures and forms of life are always constituted by exercises of power, and are always marked by certain forms of exclusion.
So while we accept with Bevir and Rhodes the general proposition that subjects are ‘thrown’ into worlds not of their choosing, but have the capacity under certain conditions to act differently, the key question is how best to account for these conditions, and for this – we argue – one needs a fully developed and convincing ontology which Bevir and Rhodes lack. From this point of view, it is worth stressing that our notion of discourse is not an epistemological concept, as Jessop and other critical realists imply, but an ontological category that is based on the centrality of articulatory social practices. And the articulatory element of a social practice captures its iterable and connecting dimension, in which each repetition of a practice discloses (however marginally) the possibility of its alteration, and in which each practice involves the linking together of heterogeneous elements. In turn, these dimensions presuppose the ultimate contingency of social relations. However, in clarifying and developing this ontological stance, we must not forget the need to relate this to the concrete tasks of social and political analysis. For example, we must be able to link this ontology explicitly to our explanations of the constitution and reproduction of the social relations into which subjects have been thrown, and to our accounts of the way in which they are gripped by certain discourses and ideologies, while also accounting for the way in which the grip of various identifications can be loosened (compare Hay, 2006). It is here that the category of hegemony is important, because it seeks to account both for the way dominant structures cover over their political emergence, as well as the exclusions and repressions by which they are engendered. Hegemony speaks directly in this regard to the analysis of various relations of domination. But the logic of hegemony also captures a particular kind of political practice that involves the linking of diverse demands into a new project that can challenge existing regimes and practices in the name of something new. These reflections carry important explanatory and critical implications.
Logics of Critical Explanation
In reflecting on the conditions for advancing convincing critical explanations we could, for example, advance a nexus of logics with which to explain problematised phenomena critically in a way that respects their singularity, but that does not restrict itself to just reporting or describing that singularity. Here the notion of logic is understood in a very specific way, as capturing the point, rules and ontological preconditions of a practice or regime of practices, and as opposed to causal laws, mechanisms and contextualised self-interpretations. One could thus develop a typology of such logics: the discernment of social logics would enable us to characterise practices or regimes by setting out the rules informing the practice and kinds of entities populating it, while political logics would allow us to account for their historical emergence and formation by focusing on the conflicts and contestations surrounding their constitution. Finally, fantasmatic logics would furnish us with the means to explain the way subjects are gripped or held by a practice or regime of practices. Taken together, logics would by no means be reducible to the empirical phenomena they are designed to elucidate, though nor would they be accorded a fully transcendental role and function.
As we have emphasised in this article, of central ontological importance from a post-structuralist point of view is a commitment to the radical contingency of social objectivity, whether in the form of structures, agents or institutions, which in our view has important ramifications for our understanding of social change, political subjectivity and the overall structuring of social relations. This means that our ontological perspective is not just concerned with a detailed listing of the different sorts of entities in the world – what is in the world – which, of course, is always relative to particular situations and practices, but it also raises prior considerations about how entities are in our social worlds and that they are the way they are, which admits of more general reflection. Indeed, in terms of political analysis, this perspective enables us to highlight the constructed and political character of social objectivity, and then to articulate a connected series of concepts that can help us to analyse social relations and processes, while remaining faithful to our ontological commitments. This is precisely what the above typology of logics would aim to accomplish. In sum, then, in contrast to the ‘thin’ social ontology offered by Bevir and Rhodes, the post-structuralist account of the structure/agency dilemma not only seeks to address the question of social change, but enables us explicitly to unfold a social ontology consisting of a framework of logics (at least in the way we understand this term) and a nuanced account of subjectivity with which to construct critical explanations.
Footnotes
This article is based on a paper presented to the PSA Conference, University of Bath in April 2007, and draws from our book Logics of Critical Explanation (Routledge, 2007). We thank Martin Smith, Rod Rhodes, Mark Bevir and Jacob Torfing for their comments on an earlier version of this article.
2
See Elster (1989; 
.
4
See Bevir (1999; 2000; 2005); Bevir and Rhodes (2003); Buzan and Wæver (2003); Campbell (1992; 1998); Dillon (1996); Fischer (2003); Freeden (1996); Hajer (1995); Norval (1996); Stavrakakis (1999; 2005; 2007); Torfing (1998); Yanow (2003); .
5
Pre-empting mainstream charges of relativism and subjectivism that this view provokes, Bevir and Rhodes begin by rejecting what they term the logics of vindication and refutation advocated by proponents of verification and falsification, respectively. In their view, both positions are seen to ‘ground objectivity or truth in confrontations with basic facts’, thus presupposing the idea that we have ‘pure experiences’ of the external world (Bevir and Rhodes, 2005, p. 182). This leads them to reject the idea of truth as certainty, but not to dispute a commitment to objectivity (Bevir and Rhodes, 2005, p. 183). Instead, they argue that the latter involves a logic of evaluation ‘by comparing rival stories using reasonable criteria’ (Bevir and Rhodes, 2005, p. 183).
6
In a similar move, Jessop claims that post-structuralist discourse theorists are ‘empty realists’, whose ontological presuppositions allow them to ‘discursively turn base metal into gold [if only!] or else convince those laughing at the emperor's fine new clothes that he really was wearing them’ (Jessop, 1990, p. 295), thus implying that they are able to construct and deconstruct the world as they please. (And of course, it goes without saying that they can construct and deconstruct the world as they please – as is the prerogative of any social scientist – although their resultant constructions are not likely to pass muster unless they lead to the development of convincing narratives and explanations that can convince relevant communities of scholars of their plausibility and verisimilitude.) And he goes on to claim that post-structuralist discourse theorists can say ‘nothing meaningful about the nature of entities apart from the fact that they exist’ (Jessop, 1990, p. 295). Here again there is much confusion, for the point that Laclau makes about the ‘discursive construction’ of entities is equivalent to the basic hermeneutical claim that the meaning of any particular object is relative to a particular context or situation (both for the researching subject and the subjects that are researched). A spherical object such as a bundle of newspapers held together by string, or a piece of foam rubber, is a football in the context of a particular rule-governed practice; in other words, its meaning and significance is relative to a particular set of meaningful practices. In addition, with respect to ontological questions, it is worth reiterating that post-structuralists are in fact realists both in the sense that they affirm the existence of a reality that is independent of thought – that the world is not a product of our ideas and language (i.e. if human beings were subtracted from the world tomorrow we would still have good reason to believe that other things such as footballs would continue to exist) – but more importantly because of its commitment to the view that our conceptions of things and entities do not exhaust their meaning or being. In other words, the symbolic orders in which we encounter things and objects are always vulnerable to the effects of what Lacan calls ‘the Real’ – that which cannot be symbolically represented, but which can dislocate – which can in certain circumstances render objects contingent and unstable. Things can acquire different meanings and functions in different historical contexts and situations, though this does not mean that they do not exist, and nor does it mean that we cannot say anything about these things within certain relatively sedimented social contexts.
7
Although we do not subscribe to McAnulla's critical realist ontology as a solution to this deficit, a similar point is made by him concerning Bevir and Rhodes’ ‘neglect of ontology’ and their ‘procedural individualism’ (McAnulla, 2006). Elsewhere, we argue that an ontology which foregrounds the radical negativity and contingency of social relations opens up a way of affirming the continued importance of structure without reifying it. This, in turn, opens up a way of thinking about the relationship between subject and structure along a range of dimensions, among which feature the political and ideological dimensions, thus making possible both a normative and ethical critique of social and political forms (see Glynos and Howarth, 2007).
8
These arguments have been developed by post-structuralists in a range of different historical contexts. For example, in a critique of mainstream theories of democratic transition and democratisation, focusing particularly on the South Africa case, Howarth has argued that transitions are invariably marked by ‘dislocated social structures’, in which ‘subjects are literally compelled to become collective political agents intent on reconstituting a new order within which identities can be stabilized’ (Howarth, 1998, p. 201). But he goes on to distinguish between decisions that are taken within more or less sedimented sets of social relations (such as regimes and institutions), and decisions that are taken about social structures themselves (see also Howarth, 2000a; 2000b; . Many more statements of this sort could be mobilised to confirm our point.
9
A central problem with Bevir and Rhodes’ hermeneutical approach is the theory of language and subjectivity that underpins it. At the heart of their account, as we have noted, is an effort to interpret the beliefs of social agents, understood as ‘individual viewpoints’. This is achieved by locating beliefs within a wider system of meanings, and by analysing the intentions of the authors that speak or write them. The difficulty with this account is that it effectively bypasses the complexities of representation and subjectivity in the name of a problematic model of communication, wherein a sovereign agent first forms ideas and beliefs, and then expresses them in a linguistic utterance or written statement (see Derrida, 1976; 1978; . In short, in this model, messages are transmitted and received by sovereign subjects, and language is relegated to the role of a transparent medium of such communicative practices.
Admittedly, they situate agents and authors within particular intellectual traditions and belief systems, thus weakening their complete sovereignty, but they do not really question the capacity of language to convey meanings and beliefs in a transparent and unproblematic fashion, nor do they problematise the ‘presence’ of both the ‘speaking’ and ‘receiving’ subject in formulating and transmitting ideas. The closest they come to compromising this aspect of language and subjectivity is in their discussion of the role of rhetoric, where people's expressed and actual beliefs differ (Bevir and Rhodes, 2005, p. 179). But even here rhetoric is viewed as an instrument of communication, which can be rationally and consciously deployed by calculating agents. ‘When people use a rhetorical pattern’, say Bevir and Rhodes, ‘they do so because they believe it will get a suitable response to their ideas’ (Bevir and Rhodes, 2005, p. 179, emphasis added). Thus the social scientist ‘can explain people's choice of rhetoric by identifying their relevant beliefs and preferences’, situating the latter within their appropriate language games, and then relating these language games to wider belief systems and intellectual traditions. What we see here is a prioritisation of an instrumental over what might be termed a constitutive conception of rhetoric. In this latter view, ‘rhetorical patterns’ are not just chosen, but actually shape our language use and meaning in non-conscious ways (see Howarth and Griggs, 2007). Of course certain tropes may be consciously employed for strategic reasons, but even when they are they often come to structure our way of thinking in a constitutive way. Copernicus’ concentric and geocentric model, or the ‘war on terror’ metaphor, would be clear examples in which the constitutive rather than instrumental dimension of rhetoric is evident. From this point of view, not only do such tropes introduce unintended and unanticipated connotative consequences, but this ‘overflowing’ of possibilities is best seen as an intrinsic part of language and rhetoric.