Editorial to special issue on rational choice and political power

The literature on power is something of a wormhole between the worlds of political philosophy and political science. It ranges over empirical strategies for measuring power through to the debate surrounding its normative implications. Empirical political scientists engage with Marx’s moral imperatives and normative philosophers pick holes in survey methodology. There are few literatures that are so splendidly cross-disciplinary in nature – and few books that captures it better than Keith Dowding’s Rational Choice and Political Power. The contributions to this special issue, designed to engage with the book’s recently published second edition, all reflect this. From Greg Whitfield’s article on the normative project of public reason, through Pamela Pansardi’s reflections on conceptual analysis, to Jim Johnson’s critique of Dowding’s use of formal modelling for empirical explanation, the issue sports a wide range of perspectives that stand testament to the book’s breadth.

While diverse, they all touch on the major challenge Dowding tackles for traversing both the empirical and normative domains: distinguishing what is the case from what one wants it to be. It is easy, for example, to say a political party is powerful when you want the party to be held to account and powerless when you do not. Along these lines, Dowding (2019, p. 173) suggests in the new Postscript that ‘the original idea for RCPP came from my critique of Steven Lukes’s three dimensions of power. I thought the collective action problem could explain everything which he thought required adding dimensions of power’. Workers may not push for legislation in their interest because they are crippled by collective action problems, the critique goes, not because a power elite is controlling them. Marx’s prophecy of a Communist revolution assumes that once the proletariat wakes up to their real interest individual workers will revolt and smash the means of production. But where violence takes effort and in the short term runs the risk of incarceration, it is presuming a lot to think large groups can coordinate in the face of free-riding incentives. While we might rightly want to hold capitalists responsible for their privilege, their power is underlain by the collective action problem faced by the many.

Steven Lukes responds in the issue with his article ‘Power and Rational Choice’. He agrees that we do not need to impute the power of outsiders to explain a group’s lack of power, calling it the ‘paranoid fallacy’, and questions ‘just how far [he and Dowding] do part ways’ (Lukes, 2021). There are cases though that do appear to stir rival judgements between the two. One of the more contentious is in what sense sexual or racial bias constitutes an exercise of power. When an individual unconsciously gains from this kind of bias, such as when a male mechanic is judged more authoritative than a female engineer purely because they are male, Dowding would describe the judgement as the male mechanic’s luck that he might then use as a source of power. Lukes argues this would overlook the mechanisms that sustain and reproduce the power structure. When patients immediately and unconsciously presume the white medical intern is their doctor rather than a person of colour, they are perpetuating a similar kind of racial bias. Perpetuating prejudice, even when it is unintentional and non-strategic, Lukes (2005) argues, is to exercise power.

Included at the end of the issue is Dowding’s response to all the contributions in ‘Individually Lucky, Collectively Powerful’. His response to Lukes is to agree that racial and sexual bias is implicated in the exercise of power when it is used ‘to create advantage’ (Dowding, 2021). When prejudice exists towards social types, it gives people the resource to use it against individuals of that type. Dowding argues that token behaviour that unconsciously reproduces privilege and prejudice is – while blame-worthy – not necessarily a form of ‘power over’. To suggest otherwise would be to remove an important sense of the victims’ agency – their ability to fight back. Readers may be interested to compare his comments here with his recent book It’s The Government, Stupid, which similarly argues that individual responsibility can only be assigned comparatively. Behaviour is never the absolute responsibility of individuals but dependent upon the incentive structure. We need to look at those who set the incentives across societies – broadly speaking governments – and not individuals themselves for the comparative amount of homelessness, obesity, gambling, and drug-addiction (Dowding 2020).

The points of disagreement and agreement between Lukes and Dowding should help clarify some of the methodological debate that boils under the surface of the power literature. Lukes favours interpretative methods in political science for identifying the third dimension of power whereas Dowding favours rational choice and revealed preference theory for modelling coordination problems. While the two articles suggest the authors do not think it is an all or nothing affair – far from it – Lukes suggests their disagreement is ultimately ‘not about how the world works but about how to describe and understand it.’ He argues ‘part of what is involved in assessing the extent of the power of an individual’ is a normative judgement about the consequences of that assessment for the interests of those affected. The extension of ‘the powerful’ must therefore be determined by both empirical and normative considerations simultaneously, in tension with a number of Dowding’s (see also 2016b) claims.

In ‘Why we do need a Concept of Power’, Pamela Pansardi pushes further on the point. She notes that ‘while we can legitimately distinguish between the appraisive and the descriptive aspect of a concept … it is impossible to deny that also the descriptive aspect implies some kind of evaluation’ (Pansardi, 2021). She criticises Dowding’s take on conceptual analysis, which appears to imply the reverse. The new Postscript to Rational Choice and Political Power suggests the particular definition of power should not matter for the book’s substantive argument, only the proxies he selects for measuring what he nominally calls power. The proxies for social power are (at least for individuals) resources like wealth and votes that can be used to cover the costs of threats, offers, legitimate authority, and information. Pansardi argues the relation between resources and power is simply too vague to capture all the normative issues at play, like the evaluation of society and the assignment of social responsibility (see also Morriss 2002). She claims that shirking a normatively sensitive definition of power in favour of more empirically tractable measures is ‘surrenderist’. My article ‘Power Obsessed’ might be read as a counterpoint to Pansardi’s here (Bosworth, 2021). In his response, Dowding clarifies his developing view on conceptual analysis with a discussion of pluralism, vagueness, and incoherence. He defends the subscript gambit, which he employed in the original edition, for dealing with conceptual disagreement over power. If you disagree with his definition of power, he proposes you define it your way with a subscript ‘power_1ʹ and leave his analysis to stand for what he calls ‘power_Dowding’.

In Jim Johnson’s (2021) paper for the issue he critiques the use of game theory in Rational Choice and Political Power. His critique also works as an exposition of his and Dowding’s (related but independent) projects on modelling in the social sciences (see Johnson 2014, 2019, 2020, Dowding 2016a, Dowding and Miller, 2018) which should be of interest to philosophers of social science in addition to power buffs. Johnson suggests Dowding has come to endorse what he calls ‘the standard rationale’, the view that we should ‘use formal models to deduce predictions that, treated as empirical hypotheses, can be tested against evidence derived from the “real world”’. This is the wrong way to think about models, Johnson argues. Game-theoretic inferences should instead be treated as fables for decidedly conceptual and normative purposes, not so much as empirical predictors. In his response, Dowding elaborates his view that a good model – the entire model – must be somehow analogous to the world. If the games imply certain consequences that do not track the empirical data, they are not fit to serve even as fables. Dowding suggests models are important for sidestepping the problem of induction (Goodman 1958) in the sense that they are supposed to model mechanisms to explain why we can expect generalisations from past data to persevere or not. He defends a kind of scientific realism (Dennett 1998, Ladyman and Ross 2007), which he suggests deflates the distinction Johnson makes between abstract theory and concerete observation. Dowding suggests stories and empirical prediction must work hand in hand, in a different manner to that of Johnson.

Gregory Whitfield’s paper ‘Two Puzzles for Shared-Reason Accounts of Persuasion’ (2021) addresses Dowding’s (2019, p. 183–7; 2016a, 2018) recent attempt to clarify the place of persuasion within the RCPP framework. Somebody intentionally persuading another will be exercising power, but whether or not that power is manipulative or not depends on the persuader’s intention. The ascription of power itself might be normatively neutral. This can be contrasted with Lukes’s position, which implies the unintentional reproduction of linguistic bias constitutes an exercise of coercive power. Whitfield also notes that Dowding presumes the consensus approach to public reason, which is controversial amongst political liberals committed to convergence public reason. The paper raises a number of issues with Dowding’s analysis, particularly with his definition of non-coercive persuasion as persuasion by means of shared reasons. Like Lukes, Whitfield is suspicious of the reliance on intention for an account of manipulation. Dowding responds with an analysis of the problems Whitfield raises in his paper and links it back to Whitfield’s (2020) project on manipulation.

I think the contributions to this special issue all demonstrate that the arguments in Rational Choice and Political Power are still as important as ever. The book continues to serve as essential reading for those who find themselves lost in between the empirical and normative in social science and philosophy, with a distinctive way of reconciling the two. This symposium provides a guide to ways in which the approach may be developed and refined further within the literature.

Disclosure statement

No potential conflict of interest was reported by the author(s).