Extending interpretivism: Articulating the practice dimension in bevir and rhodes’s differentiated polity model

Philosophical Background Rod Rhodes is an influential political scientist who was one of the first scholars to argue, in the early 1990s, that the government was a differentiated and fragmented polity, and who explored the concept of policy networks - which, since then, has become the new orthodoxy in policy analysis (Rhodes, 1997). Mark Bevir is a political theorist whose work shows strong historical leanings (Bevir, 1999). Both of their backgrounds are clearly visible in their articulation of interpretivism. One of the delights of BR’s interpretivism is its philosophical sophistication. Modernist empiricism, with its adherence to the correspondence theory of truth, the fact-value dichotomy, and universalist ideals of knowledge, operates as if a century of by now widely accepted philosophy has simply bypassed it. Smith is probably right that BR are too quick in depicting most of Anglo-Saxon political science as ‘modernist empiricist’, but the more important observation is that empiricism is not restricted to the relatively small number of political scientists who study rational choice theories. Empiricism with its appeal to ‘data’, ‘methods’, its reification of core disciplinary concepts such as ‘state, ‘institution’, or ‘power’, the taken-for-granted epistemological authority of quantitative forms of analysis (with the ensuing emphasis on reliability and validity), and the claim to (quasi-) causal explanations on the one hand, and its dismissal of qualitative and interpretive methods as ‘soft’, ‘anecdotal’, and ‘subjective’ on the other - is much more pervasive, both in politics and the policy sciences (Bevir and Rhodes, 2010: 43). The core of Bevir and Rhodes’s philosophy is antifoundationalism, which they describe as ‘any epistemology that rejects appeals to a basic ground or foundation of knowledge in either pure experience or pure reason’ (2010: 43). Instead, knowledge is always perspectivist. Whatever we perceive as facts are always facts under a particular description that organises our observations (Fay, 1996: 74). For that reason, knowledge is also provisional; there is no final arbiter (brute facts, pure reason, or the right method) for determining the truth of a statement, and what is considered true today might require revision tomorrow. Although anti-foundationalism is no longer a controversial or outlandish claim in philosophy, as BR rightly point out (2010: 43), it is not always obvious what its implications are for social science research. Anti-foundationalism often leads to accusations of relativism or an ‘anything goes’ epistemological position. Marsh, for example, claims that the notion of ‘truth’ does not apply to interpretations and chides BR for suggesting so (2011: 36). But this shows only that Marsh implicitly adheres to a naïve realist correspondence theory of truth, in which each representation has an exact correspondence somewhere out there in the world (Allen, 1993: 9-10; Wagenaar, 2011: 59). (This is yet another example of the sway that empiricism holds over social scientists of all stripes.) It would go too far to discuss the thorny subject of the philosophy of truth here. Let me just say that after Nietzsche ‘Truth’ has become ‘truth’: multifaceted, theoretically loaded, and embedded in historically situated language games and ordinary practices. However, the situation that reality is not the final arbiter of the truth-value of our statements, and that different theoretical propositions may do an equally good job in explaining a particular slice of the world, does not mean that there is no fact to the matter. Theoretical explanations are hooked up with the world, and some better than others. This connection between theoretical proposition and the world does not occur through some unspecified, language-mediated correspondence with brute reality, but through practice, as well as the resistances we encounter when we act upon the world. Truth moves about in a landscape of action. This explains why most people, when provided with the necessary information, are in a relatively good position to judge which statements do a better job in explaining a particular part of reality with which they are familiar. I will return to this important point later. Although fears of relativism in interpretivism are misguided, I am afraid that BR, with their idealist conception of interpretivism, have not been particularly successful in laying to rest such fears. But before we get there we need a better understanding of the practical implications of BR’s anti-foundationalism. They articulate two of them, which they call ‘meaning holism’ and ‘antirepresentationalism’. Both hang together. Concepts such as ‘state’, ‘voter’, ‘unemployment’, or ‘climate’, they argue, are not linear representations of objects ‘out there’ in the world. Because, as we saw, our observations always appear to us ‘under a particular prior description’ their meaning or sensefulness, our very ability to recognise them as an observation of something - an office of the state, unemployment statistics, a young offender - hinges on these observations being embedded in a web of concepts. As BR put it: Meaning holism implies that our concepts are not simply given to us by the world as it is. Rather, we build them in part by drawing on our prior theories in an attempt to categorize, explain, and narrate our experiences.