Why Normative Behaviourism Fails

Abstract

Starting from Jonathan Floyd’s contrast between ‘mentalism’ and ‘behaviourism’, I argue that, in general, we cannot make sense of a person’s behaviour without also understanding the thinking behind it. Floyd claims that ‘mentalist’ political philosophy is undercut by inconsistency and disagreement in people’s political judgements, but the evidence suggests otherwise. The public does not divide up into rival camps as philosophers do, but instead are mostly value pluralists. Conversely, it is unclear what justificatory role is played by the kinds of observed behaviour – insurrection and crime – singled out by Floyd. Failure to resist a political system cannot tell us whether people endorse the principles it embodies. So the system Floyd wishes to defend – ‘social-liberal-democracy’ – cannot be vindicated as ‘the people’s choice’ simply by producing evidence that it provokes less resistance than alternative systems, past and present.

Jonathan Floyd’s ‘normative behaviourism’ offers a deep challenge to more conventional ways of doing political philosophy. ¹ Since my verdict on this method is going to be critical, it may be helpful to begin by noting the following two key points on which Floyd and I agree:

1. The primary aim of political philosophy is practical. It should generate recommendations as to how we should order our collective life. This contrasts, for example, with the view that its primary aim is the analysis of political concepts, or the view that its aim is to depict an ideal society, whether or not this has any practical relevance to what we should do here and now.

2. Although political philosophy is normative, in the above sense, it is also fact-dependent. Its principles are grounded by facts about the world, though not entailed by them. Significant new discoveries – for example, about how people’s beliefs are formed or about how social and political institutions work – should cause us to revise our principles. This brings political philosophy into close contact with empirical social science. ²

These points of agreement are significant. Our disagreement emerges when we ask what should count as relevant facts in the context of point 2. Floyd’s main distinction here is between ‘mentalism’ and ‘behaviourism’. For mentalists, he claims, the relevant facts are facts about what people think. Principles are to be grounded in the beliefs people have and the judgements they make, for example, what they believe and say about social justice. There are different versions of mentalism depending on the discovery procedure to be used for identifying the relevant beliefs, but the common factor is the reliance on thought itself.

Normative behaviourists, by contrast, hold that the relevant facts are facts about what people do rather than about what they think. Analogously to mentalism, there could be different version of behaviourism depending on the segment of behaviour that is taken to be relevant for political philosophy. Floyd picks out two specific forms of behaviour, insurrection and crime, partly on the grounds that people who engage in them are running real risks to their own interests, and in this commentary, I will focus on the former. But first, I want to scrutinise the mentalism/behaviourism contrast itself.

One might first be sceptical that for fact-dependent political philosophy, any one kind of fact has privileged status. We may need to appeal to facts of many different kinds in constructing a political argument. For example, in developing a theory of political authority, we may cite evidence about what people believe to be legitimate, about their behaviour when confronted by coercive law, about the long-term consequences of adopting particular institutions, such as representative democracy, and so forth. It seems arbitrary to declare in advance that only facts of a particular type should count.

Drawing a hard line between people’s thought and their behaviour is also problematic. A simple example makes the point. Suppose I see someone driving a car off the end of the pier into the sea. What have I witnessed: was it a tragic accident, a suicide attempt, or a stunt for the next Bond film? To understand what I have seen, I need to know the driver’s intention, and if driving off the pier was intentional, the reason he was doing it. The physical behaviour by itself tells me very little: it needs to be interpreted by reference to ‘thought’.

Floyd might accept this last point, and continue to claim that what was distinctive about normative behaviourism was its focus on action as opposed to opinion or intuition, even while conceding that action was ‘thoughtful’, so to speak. And indeed, when he comes to deploy normative behaviourism in the last part of his book to ground the political system that he labels social-liberal-democracy, he makes an assumption about the meaning of the behaviour whose incidence he tracks. Consider insurrection in particular: this is not formally defined, but it presumably refers to large numbers of people engaging in violent protest to get rid of their present government. Although Floyd (2017: 179–180) admits that their reasons for taking up arms might be very different on different occasions, he assumes that insurrection constitutes a rejection of the form of government (and the principles it represents) itself. But this is a risky assumption to make: an insurrection might be fuelled by economic grievances, by the desire to replace one autocrat by another, and so forth. We cannot learn very much just by watching people storm the presidential palace. One obvious solution would be to ask them why they were doing what they were doing: but if we did that, we would be crossing the line that is supposed to separate behaviourism from mentalism.

What then does Floyd find so unsatisfactory about the mentalism that he claims has been the dominant approach within political philosophy up to now? He uses the term very broadly so as to include anyone who uses some version of ‘reflective equilibrium’ in which normative principles are tested by comparing them with the pre-theoretical judgements that people make on the issues to which the principles are supposed to apply. Whose judgements are going to count? Floyd mainly discusses versions of mentalism for which the relevant data are the philosopher’s own intuitions about relevant topics, although he does also consider the use of experimental and other evidence about what people in general think. His overall argument is that in either case such evidence is not of good enough quality to be of any use in political philosophy. It merely reproduces the problems that political philosophy is tasked with resolving, of which the two main ones are inconsistency and disagreement in judgement.

Floyd claims that, even in the case of a single person, the judgements they would make in response to questions of the kind that interest political philosophers – such as whether children have the right to inherit their parents’ wealth – will, taken together, form an inconsistent set. So, there is no coherent series of ‘considered judgements’ available to use in the search for reflective equilibrium. The example of incoherence that he gives on p. 135, however, is hypothetical only; and Floyd is ready to concede that further reflection might allow the internal inconsistency to be removed. The bigger problem, therefore, is disagreement between persons, which Floyd presents as ineradicable. If so, that does appear to cause problems for political philosophy, which aims to speak to all citizens, at least, and not just to its own creator plus anyone else who happens to share her beliefs. As Rawls (2001: 31) would put it, it seeks to achieve not just wide but general reflective equilibrium where ‘the same conception [of justice] is affirmed in everyone’s considered judgements’. But do people really disagree as radically in their political beliefs and judgements as Floyd supposes?

It is tempting to conclude that they do by casting them in a drama scripted by political philosophers, and assuming that the public world is made up of squabbling utilitarians, Rawlsians, Nozickians and so on. Floyd imagines people being asked to specify their ideal political order. He continues as follows:

Some will demand a state in which every citizen is guaranteed a minimal standard of welfare; others that no great inequalities are permitted between different citizens, regardless of choices of lifestyle or the varying market value of different careers. Some will demand of an ideal state that it maximises aggregate happiness; others that it guarantees the core freedoms of exchange and contract, come what may …. (Floyd, 2017: 129).

These positions are of course instantly recognisable to anyone schooled in political philosophy. But what evidence is there that the public at large would divide along these lines when asked Floyd’s question?

We do in fact have a great deal of relevant evidence, obtained through experiments, surveys and so forth, about what people think about the issues that interest political philosophers, such as issues of social justice. ³ This does not reveal full agreement on such questions, but nor, however, does it show people dividing up into rival camps in the way that Floyd supposes. Instead most people are value pluralists: they are likely to see some merit in each of the principles represented above – the social minimum, limits on inequality, efficiency, freedom of exchange. Presented with practical choices, they will look for solutions that accommodate to some degree each of these values. Of course, their intuitions are also often inconsistent with one another, especially when presented with dilemmas crafted so as to offer a stark choice between two of these values (and therefore unlikely to be encountered in real life). In such cases, the answers they give will be influenced by framing effects that highlight one horn of the dilemma at the expense of the other.

But this is no reason for political philosophers to despair. It gives them the opportunity to develop a theory that best accommodates the complete set of judgements that empirical research reveals. If the public were already in full agreement on normative questions (first-order), ⁴ political philosophy would be redundant; if there were radical across-the-board disagreement, it would be impossible, as Floyd suggests. Incomplete agreement allows the philosopher to investigate the factors that may explain it, and propose her own resolution.

That concludes my short defence of ‘mentalism’ against Floyd’s critique (I have said already that I do not believe political philosophy can rely solely on evidence about what people think, so this is not meant to be an argument for mentalism pure and simple). I turn next to his proposed alternative, normative behaviourism, which simplifying somewhat takes the following form. First, we test principles by seeing them as embodied in real-world political systems. Then we examine how people react when living under these systems. In particular, we look at when they engage in resistance behaviour – insurrection and crime being the two main examples. Contrariwise, lack of resistance to a particular system is taken to signal that this is the system that they prefer to live under. The principles themselves are then justified by this form of popular acceptance.

There is some ambiguity, however, about the exact form of the justification that is being provided here. One version, which might be called ‘realist’, asserts that lack of resistance demonstrates that people endorse the system they are living under, and thereby render it legitimate. This in turn justifies the principles embodied in the system. The claim, in other words, is that a set of political principles is valid if a political system underwritten by those principles is regarded as legitimate by the people who live under it, as demonstrated by their compliance with its rules. A second version is utilitarian: people accept the system because it satisfies their various preferences, both private and public, better than any other system they can see on offer, and so the principles it embodies are justified because they are instrumental to the general happiness. The latter interpretation yields a more conventional reading of Floyd’s position than he might wish for, but it appears to be supported by several of his remarks about the significance of behaviour. ⁵

For either form of justification, there is the danger of an endogeneity problem – the worry that the prevailing political system may shape people’s beliefs and preferences and thereby their behaviour. Some of the claims that Floyd makes strike me as deeply ahistorical. For example, he describes social-liberal-democracy as ‘the people’s choice’, corresponding to what he calls at one point ‘foundational human preferences’:

When we say that the vast majority of people prefer social-liberal-democracy, what we really mean is that the reactions people give to different regime types, including social-liberal-democracy, demonstrate indirectly that the latter is the preferred political system (Floyd, 2017: 238).

But who count as ‘people’ here? Is it really all of humanity from the hunter-gatherers onwards? Is the claim that human history is one long series of experiments with different political systems until people finally realise that social-liberal-democracy was what they had been looking for all along? No one could seriously believe that. In particular, we know that liberal democracy has economic and social preconditions. It developed in technologically advanced, prosperous societies whose populations enjoyed a relatively high level of literacy and we have also learnt by experience that it does not transplant easily to places where those conditions are absent. To be fair to Floyd, he concedes that economic prosperity may be a significant factor in explaining why people in democratic societies do not engage in insurrection – they have quite a lot to lose by revolting – but he then sets that explanation aside on the grounds that this is not a matter of choice; it is not open to us to choose economic prosperity in the way that we can choose to support (or reject) political institutions (Floyd, 2017: 194). But if it is indeed the case that a certain level of economic development is a precondition for liberal democracy, then all we can say on the basis of evidence about its stability is that once people are economically prosperous, they are disinclined to overthrow their government. We cannot infer that they are committed to social-liberal-democracy as a matter of principle.

This illustrates the general point made above that we have to interpret people’s behaviour before we can draw any normative conclusions from it. The mere fact that people acquiesce in a political system, as opposed to rising in opposition, by itself tells us rather little about the grounds of their acquiescence. It might be indoctrination, it might be fear, it might be simple complacency – or it might indeed be a belief in the system’s legitimacy, but to reach that conclusion we would need to find out more about their thinking.

The evidence that Floyd (2017: 185–189) cites in support of his general claim that ‘liberal democracies produce less insurrection than any other political system attempted in human history’ is evidence that compares the performance of democracies and autocracies in the world today. But liberal democracy is still relatively young in world-historical terms, and it may be too soon to judge that over the long historical stretch it will prove less insurrection-prone than other systems. In particular, hierarchically organised peasant societies across the world proved to be long-lived and relatively stable: peasant revolts mainly occurred when the reciprocal relationship between lord and peasant was disrupted by the introduction of new forms of agriculture (Moore, 1969: ch. 9). Although the threat of repression was certainly always present, these societies were largely stabilised by indoctrination and simple inertia.

This may also be the point at which to consider modern China, which seems to provide an alternative contemporary model that is at least as stable as liberal democracy in terms of its ability to gain popular acquiescence. Floyd (2017: 250) seems convinced that ‘China will only survive as a cohesive, peaceful political society by granting stronger individual rights, greater freedom of speech, and full universal suffrage’. This view, also the hope of many liberals, seems less convincing today than it did when Floyd was writing in 2016. Although many local protest events occur in China each year, it appears that the protesters and the authorities have reached a modus vivendi whereby the policing of protests is relatively light so long as the protestors do not overstep the mark and mount a direct challenge to Communist Party rule. Indeed ‘the regime has even encouraged narrowly targeted protests to identify social grievances, to monitor lower levels of government and to remedy the weakness of its political system’ (Li, 2019). This illustrates how authoritarian regimes can render themselves insurrection-proof in the medium term at least, by channelling opposition in such a way as to leave the pillars of the regime secure.

The upshot is that normative behaviourism does not offer a promising way to justify social-liberal-democracy. There is a slight whiff of ‘the end of history’ in Floyd’s claim that it is the political system that almost everyone will prefer once they have tried it. It may also make us too complacent about the dangers of what is sometimes called ‘democratic backsliding’ in established democracies, though that expression too has unfortunate teleological connotations. Floyd and I may in the end disagree about the reach of political philosophy itself – whether it can provide a justification for political principles (and through them for a political system) that transcends time and place. ⁶ Perhaps we have to start with political beliefs that are more local, and be satisfied if we can work them up into a coherent political theory that can guide us here and now.