Breaking Ranks: The primacy of epistemology
by Marc Geddes
Within the Politics Department at Sheffield University, questions about ontology and epistemology can almost always be traced back to the post-positivist research ethic of Professor Colin Hay. Although he has written authoritatively on the subject, his word has become – in some quarters, at least – irrefutable.  This account of ontology and epistemology is taught as though it is fact, rather than contention, which has possibly limited the reflexivity within our Department, and perhaps even affected the discipline more widely. Whilst the account may well be true, this is not certain – and to preclude other options would do a disservice to the teaching of political analysis. So, what exactly is the point of contention?
It is usually assumed that ontology comes first, and epistemology comes second; that the latter is dependent on the former. Ontology, conceived on its broadest level, is ‘a branch of philosophy that is concerned with the nature of what exists’, and in the social sciences concerns the central question of ‘what is the nature of social reality?’. Epistemology ‘is a theory of how human beings come to have knowledge of the world around them’, or ‘a philosophical grounding for establishing what kinds of knowledge are possible – what can be known – and criteria for deciding how knowledge can be judged as being both adequate and legitimate’.  In this sense, then, ontology is about ‘what exists’ and epistemology is about ‘how we can know what exists’.  From those definitions, it is clear that we must acknowledge the difficulty in discussing ontological and epistemological assumptions separately – they are necessarily entwined. At this point, a number of scholars make the presumption that ontological principles precede those of epistemology. To make the argument that ‘reality’ exists without us knowing about it is not necessarily problematic. ‘Reality’ – whatever this is – does exist, regardless of our knowledge of it. However, many scholars then go on to argue that we can make ‘ontological assumptions’, which precede our ‘epistemological principles’. To do this is to conflate the nature of reality with our knowledge of reality; it is to argue that we can know about something without principles about knowledge. It is, essentially, putting the cart before the horse. Scholars, Hay included, often underestimate the linkage between the two domains of philosophy. To go one step further, however, one might ask how far these two realms of thought are linked, or: is it possible to separate existence from what we believe existence entails?
The consequences for how we answer this question determines much of our relationship with social reality. Indeed, the ultimate consequence would be to question ‘reality’ in its entirety, and collapse ontology within the field of epistemology.  If we accept that humans are ‘self-interpreting’ animals (an obvious point), then epistemological principles arguably become as important (if not more so) than ontological assumptions. This has crucial implications for how we study politics in particular, and ‘the social’ more widely: it suggests that we have to explore the ‘perceived’ reality of groups and that we cannot study an ‘objective’ reality. Thomas Kuhn – to this day controversial – has made a similar case for the natural sciences. He argues that science is conducted within competing (and mutually exclusive) ‘paradigms’. When paradigms change (usually through ‘revolutionary’ processes), the world itself changes:
Led by a new paradigm, scientists adopt new instruments and look in new places. Even more important, during revolutions scientists see new and different things when looking with familiar instruments in places they have looked before. 
Kuhn’s thesis is that theories or ideas (i.e., paradigms) lead to particular observations, which inverts the positivist understanding that observations lead to general theories. One evident example is the shift from Newtonian mechanics to quantum mechanics in physics. Taken together, the Kuhnian analysis of the natural sciences, in addition to the primacy of epistemology in social sciences, both indicate a preference for a hermeneutic understanding of social existence located in an ‘idealist’ framework.
Some political scholars are apprehensive about this analysis because it could lead to a relativist view of the world. Furthermore, they argue that if this were the case, then ‘relative knowledge’ has little to no practical value to making policy-relevant knowledge. If we can’t say anything that’s ‘objectively true’, then can we say anything at all? This is a damning verdict. However, the problem with this argument is that it defines ‘objective truth’ as ‘absolute certainty’. This is not the case in the social sciences and, as demonstrated by Kuhn’s analysis, the same even applies in the natural sciences. Knowledge claims, truths and facts are usually accepted following some sort of social negotiation: peer review, inter-subjective reasoning, and so on. Knowledge, then, is dependent on the idea of ‘shared facts’ rather than ‘given facts’. This position, although radical, does not inhibit socially valuable knowledge; it merely asserts that our claims to it must be based on a hermeneutic understanding of reality. This departs from our traditional understanding of ‘objectivity’ towards one where ‘[o]bjectivity arises from using agreed facts to criticise and compare rival interpretations. A fact is a statement, typically about a piece of evidence, which nearly everyone in a given community would accept as true’.  It is possible to make the case that an anti-foundationalist epistemology can still make policy-relevant knowledge – by denying ‘absolute certainty’, we hope to eliminate dogma, not the pursuit of knowledge nor the belief that you could be ‘right’ or ‘wrong’. In other words, this epistemology asks us to be more humble about our claims, and to accept the possibility that you (or I) might very well be wrong about something.
Where does this leave us? Most importantly, it suggests that the study of political phenomena must be grounded in an analysis of truth claims and shared facts of the actors’ that are being studied, and less attention ought to be paid on the ontological commitment of the author of the study. As Mark Bevir and R.A.W. Rhodes point out: ‘[W]hen people disagree, they disagree precisely about what they think they know is real, and so their various statements about being [ontology] are all ones that can be analysed as ones about what we know [epistemology]’.  This anti-foundationalist framework is at odds with an analysis that believes in the separation between perception and reality. Perhaps there is the case that a real world of objects does exist; however, these are only realised once we have a framework for understanding that world. In this way, ontology and epistemology do not precede one another. There is no ‘line’ of dependence. Rather, it is better to understand the two domains as inextricably inter-dependent. On an abstract philosophical level, they both depend on each other. When it comes to the study of the political, the primacy of epistemology is clear to understand the actions, beliefs and behaviour of political actors (as the Bevir/Rhodes quote has sought to show). For that reason, it is perhaps a little perilous that so many students (and, indeed, some scholars) usually accept an epistemological dependence on ontology as fact rather than contention; but it is even more dangerous that so many students fail to understand the consequences that flow from making such an assumption.
This analysis forms part of my enquiry on theories of knowledge in the social sciences, which constitutes the foundation of my approach towards the study of parliament in the UK.
 See, among others, C. Hay (2002) Political Analysis: A critical introduction, Basingstoke: Palgrave Macmillan.
 N. Blaikie (2007) Approaches to Social Enquiry, second edition, Cambridge: Polity, p.13 and p.18.
 Hay (2006) ‘Political Ontology’, in R. Goodin and C. Tilly (eds.) The Oxford Handbook of Contextual Political Analysis, Oxford: OUP, p.80 and p.83.
 R. Trigg (2001) Understanding Social Science, Oxford: Blackwell, p.22.
 T. Kuhn (1962) The Structure of Scientific Revolution, Chicago: Chicago University Press, p.110.
 M. Bevir and R.A.W. Rhodes (2005) ‘Interpretation and its Others’, Australian Journal of Political Science 40:2, p.183.
 M. Bevir and R.A.W. Rhodes (2006) ‘Disaggregating Structures as an Agenda for Critical Realism: A reply to McAnulla’ British Politics 1:1, pp.398-9. The methodological implications are, of course, quite complex as a result. That’s a matter for another time though.