'Is a Science of Politics Possible and Desirable', Advanced Political Analysis POL6225, Colin Hay, Word Count: 3148

Is a science of politics possible and desirable? In order to consider the possibility and desirability of a science of politics we must consider the question in three different ways. Firstly, we must understand what a science of politics would look like; secondly, we must question the ontological possibility of a science of politics; thirdly, we must consider the implications of a science of politics in terms of its ability to take account of structure and agency, the relationship between ideational and material factors; and its ability to take account of change and continuity. Firstly, we must define what is meant by a 'science of politics'. In order to answer the question it is necessary to establish what a 'science' of politics is. The 'scientific' approach is derived from the Cartesian revolution, an attempt to establish a 'secular science' 'neutral on questions of meaning and value' through liberating 'reason and knowledge' away from authority, basing knowledge on 'innate human characteristics'1. As a result, a science of politics contains 'two rival traditions of scientific inquiry'; either rationalist or empirical 2. The rationalist understanding of the world is predicated on Cartesian logic that 'through reflection' 'of general axioms' there 'might be derived an understanding of the underlying and unobservable structures which' 'constituted the reality of the natural order'3. This rests on deductive the logic of moving from general principles to specific conclusions on the assumption that if the logical premises are true, the conclusion must also be. This is in contrast with the empirical model, which states 'there is no appearance reality dichotomy'4. Instead it claims that 'the world presents itself to us in a direct' 'unmediated way through our senses'5. Empiricist models are therefore predicated on inductive logic by proceeding from specific instances to generalisable laws. Both rationalist and empiricist understandings of the world therefore rest upon an assumption of establishing generalisable laws; rationalism through general laws which individual phenomena are then explained by; empiricism through individual instances used to construct general laws. As a result of such a focus upon universal laws, scientific approaches to politics, can be said to have a parsimonious approach to explanation. Parsimonious explanations seek to establish explanations of the greatest number of instances predicated on the least number of variables. Being predicated on an assumption of either moving from or towards generalisable laws, the scientific approach to politics reflects such an attitude. Furthermore, both approaches seek to render human behaviour predictable in that they highlight consistencies over time. In order to critique such a scientific approach to politics, we must first identify whether the ontological position of rationalism and empiricism is possible. The scientific approach to politics could be seen to be flawed on an ontological level. Firstly, in seeking to create a science of politics, political analysts depend on the ability to analyse objective phenomena. This is seen in both the specific instances used to validate conclusions reached by deductive reasoning, or the instances which used to create general laws within inductive reasoning. This is problematic within explanations in politics, as these are dependant upon ontological assumptions which are not verifiable by empirical data. Ontology is 'the science or philosophy of being', and represents the claims that the individual makes as to what exists and the way that it is structured6. Such assumptions are not empirically verifiable, as 'different analysts bringing different ontological assumptions' will 'interpret the empirical sequence of events' differently' 7. This is
1 2 3 4 5 6 7 Hay, Colin, 'Political Analysis: A Critical Introduction' (Palgrave, Basingstoke, 2002) p.76 Ibid. p.77 Ibid. p.77 Ibid. p.78 Ibid. p.78 Ibid. p.61 Hay, Colin, 'King Canute and the 'Problem' of Structure and Agency: On Times, Tides and Heresthetics' in

Political Studies (57(2)2009) p.262

'Is a Science of Politics Possible and Desirable', Advanced Political Analysis POL6225, Colin Hay, Word Count: 3148 because in order to explain phenomena we must first decide of what is this an instance? What What is its nature?8. At the point of interpretation, it is therefore impossible to adjudicate between competing accounts of phenomena, and therefore to establish objective phenomena. Additionally, the scientific approach is dependant on the ability of the individual to be situated outside of the object of study in order to achieve an objective understanding of the object of analysis. This is necessitated by the desire to create a value neutral science, free of subjectivity. However, 'theory is always for someone and for some purpose'9. A theorist, as an individual who inevitably operates within one or more contexts, is always placed in a subjective position governed and constrained by their place within social, historical and economic relations. The difficulty in attempting to create a science of politics, without the reflexive awareness of the individual's own subject position, is that the individual brings their own ideological bases for coming to study into their own analysis. What Cox describes as 'problem solving theory', i.e. seeking to provide 'a guide to help solve the problems posed within the terms of the particular perspective which was the point of departure', results in the reinforcement of 'particular national, sectional, or class interests', as the assumptions of the theorist are not critically analysed 10. This would appear to make a science of politics both impossible, in that it is impossible to achieve the objective position, and undesirable, in that it leads simply to the recreation of existing social relations. Such an argument though could be seen to have limited utility. Pleasants notes that if 'getting the facts right will not help us resolve long-standing ontological disputes, then 'why should political scientists' 'be bothered about resolving longstanding ontological disputes?'11. This appears to misunderstand the ontological argument, in that the claim is not for the end of all investigation, but greater reflexivity and awareness on behalf of the theorist. Whilst the striving for a pure science of politics can clearly lead to problems in regards to claiming objectivity, this should not be taken to an extent whereby all knowledge and evidence is written off as simply the perspective of an individual. Instead political analysis should be reflexive and aware of its own limitations. The potential for a science of politics should also be adjudicated on the effects ontological assumptions have on its ability to deal with the most fundamental problems of politics. One problem is the structure-agency debate. The adoption of a scientific approach to politics results in structuralist explanations, which could be false, and also undesirable in rendering all human action pre-determined. Structure is 'the setting within which social, political and economic events occur and acquire meaning', and explanations focus on the way context defines individual behaviour12. This can be contrasted with voluntarist explanations, which focus on the 'capacity of an actor to act consciously' in order 'to attempt to realise his or her intentions'13. Structural explanations have a closer tie to the natural sciences in sharing a pursuit of constructing generalisable laws. Structuralism attempts to take generalised circumstances of individuals or groups, and explain behaviour as a result of them, and thus allow generalisable laws. This is seen in behaviouralism's move from individual instances to the creation of general laws, and rationalist approaches which move from general laws to the explanation of specific instances. This approach is problematic though if the differences between the natural and behavioural sciences are taken into account, in that the object of analysis within the natural sciences is incapable of exercising agency, and therefore diverging from the expectations created by laws. The same however is not true of the subject of analysis in politics, for instance within rational choice theory. Rational choice theory, establishes a relationship between context and behaviour,
8 Ibid p.262 9 Cox, Robert W, 'Social Forces, States and World Orders: Beyond International Relations Theory' in Millenium Journal of International Studies (10(2)1981) p.128 10 Cox pp.129-130 11Pleasants, N, Structure, agency and ontological confusion: Response to Hay, Political Studies, (57(4)2009) p.887 12 Political Analysis p.94 13 Ibid. p.94

'Is a Science of Politics Possible and Desirable', Advanced Political Analysis POL6225, Colin Hay, Word Count: 3148 where explicating the context of the individual, makes it is possible to determine the action the individual will take. This is made possible by the assumptions that all individuals are rational self interested utility maximisers, and that there is only one rational course of action for an individual in a given context, and thus creating a general law. Several problems can be highlighted with the dominance of structural explanations. Firstly, this fails to take account of agency of the individual, in that 'change results from' a 'process which is operative at the structural level, meaning that human agents are never granted the autonomy to have any independent effect upon it'14. This however provides an alternative ontological account, and cannot be a devastating criticism, as 'such ontological differences cannot be resolved or adjudicated empirically since no interpretation of the evidence is itself ontologically innocent' 15. It would however appear to be undesirable when 'a passive agent is advanced because social determinism is endorsed' due to all action being derived from structure16. Furthermore, it can be suggested agents must be able to affect structures, as otherwise it is impossible for those structures to come about. If 'action is epiphenomenal then logically structure must predate it', 'yet because action is not held to create it, then the sources of structure are located elsewhere' but 'social structures are never admitted to have social origins'17. Finally, if all individuals are within such structures, as structuralist approaches would claim, the question has to be asked 'if we are indeed all simply expressions of the structures we bear, how could we hope to know?'18. However, limitations exist to these criticisms. Firstly, most rational choice theorists concede that people are not rational, it is an exercise in parsimony to create models, from which much can be learned if we are aware of the limitations of such models. Secondly, many of the criticisms of scientific structuralism, are charges 'to which any genuine structuralist could quite happily plead guilty', and as ontological questions there is no basis for a decision of empirical validity19. Whilst problems exist with a scientific approach to politics, the parsimony and insight granted by taking analytical short cuts, may indicate a need to maintain some elements of a science within politics, but only with an awareness of the limits of such scientific ambition. The account of the relationship between structure and agency within a naturalist understanding of politics also suggests a unique nature to the relationship between ideational and material resources within a science of politics. Scientific understandings of politics can be claimed to struggle to take account of ideas within politics. This occurs in two ways: firstly, an inability to account for ideas held by individual actors; secondly, an inability to recognise ways in which theorists engage with ideas. As noted previously, the tie to the natural sciences means a science of politics is incapable of recognising agency. Political science, having to account for behaviour in some way, therefore uses 'the concept of material interest' to 'make possible a naturalist science of politics' 'in that they 'serve to render the behaviour of actors predictable given the context within which it occurs'20. Naturalist approaches thus make two assumptions. '(1) that actors' behaviour is a simple reflection of their self-interest' and '(2) that such interests are' 'given by' 'their material context' 21. These assumptions are problematic on two levels. Firstly on the level of individual action, even if it were accepted that individuals always seek to act in a 'rational' way, and that interests are reducible to material factors,
14 Archer, Margaret S, 'Social Theory and the Analysis of Society', in eds. Mat, Tim and Williams, Malcolm,

Knowing the social world (Open University Press, Buckingham, 1998) pp.75-76 15 King Canute p.262
16 17 18 19 20 Archer p.75 Archer p.76 Political Analysis p.109 Political Analysis p.107 Hay, Colin, Ideas and the Construction of Interests, in eds. D. Bland and R. Cox, Ideas and Politics in Social Science Research (Princeton University Press, Princeton, 2011) p.71 21 Ibid. p.72

'Is a Science of Politics Possible and Desirable', Advanced Political Analysis POL6225, Colin Hay, Word Count: 3148 doubtful in itself, such an account appears problematic. This assumes 'actors are blessed with perfect information', otherwise the behaviour of individuals would be unpredictable, as their 'level of rationality' would be dependant upon their 'perception of such interests rather than those interests themselves22. Furthermore, naturalism rests on the assumption that in a given context, only one course of action can be pursued in the interests of actors. This could be seen to be flawed in two ways. Firstly, individuals are likely to have many contextual influences. The possibility of such intersections renders it likely that the individual will have multiple interests that they can pursue. Even if it is accepted that the individual has a single interest, 'there is invariably more than one optimal strategy for the maximisation of an actor's' interest 23. As a result ideas play an important role in that they must either inform the decision as to which context is dominant, but also between competing paths if an individual were to have a single context. Moreover, ideas are important for explanation. The assumption of naturalism is that it is possible to establish a descriptive understanding of human political behaviour. It can however be claimed that in doing so, a naturalist approach ignores the the fact that 'the social or discursive construction' of a subject 'may have an effect on political and economic dynamics independently of the emprics' of the subject itself 24. What this means is that rather than being a closed system, the way analysts conceptualise a system, even if it fails to reflect the 'objective reality' of the system, can create the behaviour that is being described, by causing individual actors to behave as if it did. This indicates a need to find a role for ideas within political explanations in a way which is impossible from a scientific approach. However, such a critique remains limited in scope in two ways . The first is that the question of whether individual's interests can be tied entirely to material factors is ultimately an ontological question, and therefore whilst analytical questions can be asked, it is a matter of preference. Perhaps more persuasively, most rational choice theorists would argue that they do accept that in many ways people do not always act in a rational way. It is therefore in the accounting of ideas that we see one of the biggest weaknesses of a scientific approach to politics. However, this does not render approaches such as rational choice theory useless, but instead argues for an awareness of their weaknesses. Finally though, we can look to the issue of change within politics to examine the effects of a scientific approach to politics. Finally, scientific approaches to politics could be seen to favour synchronic or comparative static analyses of society, damaging conceptualisations of change. As previously noted, scientific approaches to studying politics establish generalisable rules for political conduct, as through this activity we can achieve goals such as a predictive science. What this means is that for any causal mechanism or conclusion to be 'scientific' in nature, analysts must 'resist history' and other parochial factors 'as they deny the influence of particular times and history' in order to render general laws possible25. Naturalists must therefore operate under one of two assumptions. Analysts are forced to adopt a synchronic perspective, or accept that generalisable laws are a simplifying assumption which allow greater understanding of a more complex reality. The problem with synchronic analysis is that whilst offering a description of the conditions of a particular period, 'it can tell us nothing about the process or even the extent of change' with reference to either how the present situation emerged, or is changing26. Furthermore, it can be questioned whether synchronic models offer an effective form of political analysis given the ontological challenges posed by advocates of a diachronic approach. It could be suggested that 'rather than the development of the system in question' being 'split up into phases during which its structure and form can be assumed static',
22 23 24 25 Ibid. p.73 Ibid. p.73 Political Analysis p.202 Tilly, Charles, 'Why and How History Matters' in ed. Goodin, Robert E, The Oxford Handbook of Political Science (Oxford University Press, Oxford, 2009) 26 Political Analysis p.145

'Is a Science of Politics Possible and Desirable', Advanced Political Analysis POL6225, Colin Hay, Word Count: 3148 change is continuous, and therefore we recognise nothing of value when we isolate a certain period27. Such a critique though has its limitations. Again the problems associated with a naturalist perspective exist only in so far as the assumptions that underlie analysis go unrecognised. As long as the ontological fallacy at the heart of naturalist approaches is recognised, then such an approach has much to offer in parsimony and utility. By contrast diachronic analysis, by regarding change as a constant process, is 'a laborious and time-consuming exercise.' Finally though, it would appear again we are often dealing with ontological questions in such cases. If we take the model of social change brought by Marxism of revolutionary change, this is not an unrecognised ontological assumption, but rather a conscious choice to reject models of diachronic analysis. In its inability to take account of change, a naturalistic approach to politics appears neither desirable nor possible. The inability to take account of change appears to render a science of politics impossible at the point at which it fails to explain one of the most fundamental aspects of politics. The problems of adopting a scientific approach to politics appear to be under three categories. The first is ontological questions at the heart of political analysis. These questions, such as the actual nature of social change, appear to be impossible to reconcile given the inability to answer ontological questions with the use of empirical evidence. However, on some of these ontological questions it is possible to make judgements about the desirability of a political science, such as the fatalism that the dependency on structuralist explanations that is rendered necessary by a scientific approach. Secondly, are issues of awareness of assumptions made by scientific approaches to politics. Whilst it is accepted by many that a diachronic analysis of politics more accurately represents the true nature of change within politics, in the name of parsimony and effective explanation, it can be beneficial to adopt a synchronic model. What is of importance is that the nature of these assumptions is made clear by political analysts in order for others to be aware of the limitations of such work, and for it therefore to be only accepted as an accurate model to the extent that those assumptions hold true. Finally, there are certain problems with a scientific approach to politics which appear ultimately irredeemable. The inability to gain an Archimedean point outside of the system, for instance, appears to render many of the conclusions derived by a scientific approach to politics difficult to sustain, at the point at which a scientific model fails to take account of its own inherently subjective position in deriving theory.

27 Ibid. p.148

'Is a Science of Politics Possible and Desirable', Advanced Political Analysis POL6225, Colin Hay, Word Count: 3148 Bibliography Archer, Margaret S, 'Social Theory and the Analysis of Society', in eds. Mat, Tim and Williams, Malcolm, Knowing the social world (Open University Press, Buckingham, 1998) Cox, Robert W, 'Social Forces, States and World Orders: Beyond International Relations Theory' in Millenium Journal of International Studies (Volume 10, no. 2, 1981) p.128 Hay, Colin, Ideas and the Construction of Interests, in eds. D. Bland and R. Cox, Ideas and Politics in Social Science Research (Princeton University Press, Princeton, 2011) p.71 Hay, Colin, 'King Canute and the 'Problem' of Structure and Agency: On Times, Tides and Heresthetics' in Political Studies (Volume 57, no.2, 2009) p.262 Hay, Colin, 'Political Analysis: A Critical Introduction' (Palgrave, Basingstoke, 2002) Pleasants, N, Structure, agency and ontological confusion: Response to Hay, Political Studies, (Volume 57, number 4, 2009) Tilly, Charles, 'Why and How History Matters' in ed. Goodin, Robert E, The Oxford Handbook of Political Science (Oxford University Press, Oxford, 2009)