1 Why Normative Behaviourism Does Not Improve Political Realism Eva Erman & Niklas Möller Forthcoming in Res Publica *** According to normative behaviourism, political theorists should ground their principles in behaviour rather than in thoughts as is done in mainstream political theory. Focusing on ‘real actions’ of ‘real people’, normative behaviourism turns facts about observable patterns of behaviour into grounds for specific normative political principles. For this reason, this way of doing normative political theory has strong realist credentials, given its methods, values and ambitions. In fact, Jonathan Floyd argues that it is an improvement of political realism since it solves two problems that allegedly face many realists, namely, the legitimacy problem, i.e., how we should distinguish genuine acceptance of a political system from false acceptance, and the institutional problem, i.e., how we should translate political principles into viable political institutions. In this paper, we make two claims. First, normative behaviourism does not solve the legitimacy problem encountered by realists, because its solution rest on a flawed distinction between foundational principles and ‘principles that matter’, together with a problematic use of a Humean internal reasons approach. Second, normative behaviourism does not solve the institutional problem encountered by realists, because its solution is in fact much more unfeasible than realist prescriptions, since feasibility is interpreted as mere possibility. We wind up our analysis by showing that normative behaviourism encounters new problems that realist approaches typically do not face. First, normative behaviourism is a kind of closet utilitarianism, but with a more 2 problematic value measure, which rests on universal principles of the kind that realists usually reject. Second, by arguing for democracy and equality, normative behaviourism runs the risk of coming too demanding and unattractive for realists, who carefully separate democracy from political legitimacy, both conceptually and normatively, and argue that the latter does not require the former. We conclude that despite several aspects superficially attractive to the realist project, normative behaviourism fails in its attempt to supply an improved version of realism. The paper is structured as follows. In the first section, we explain the main features of normative behaviourism and what makes it a member of the ‘realist family’, according to its proponents (I). The second section refutes normative behaviourism’s alleged solution to the legitimacy problem (II), while the third section does the same with regard to the institutional problem (III). In the fourth section, we wind up by bringing up two new problems that normative behaviourism encounters, that typically do not emerge for realists (IV). I. Normative behaviourism as a member of the realist family The methodological pivot that sets normative behaviourism apart from mainstream political theory, according to its proponents, is its focus on manifested political preferences, defined as the normative preferences exposed through particular patterns of behaviour, rather than the conventional ‘mainstream’ method of identifying normative preferences exposed through particular patterns of thought, such as intuitions, value commitments, principles or considered judgements, which are relied upon to justify normative theories (Floyd 2017: 168). Floyd specifically contends that two behavioural patterns serve as distinguishing markers between a good (legitimate) and a bad (illegitimate) political system: the prevalence of insurrection and crime in a society. He posits that participation in insurrection and crime, given its inherent high personal risk, is a course of action people generally choose only when they find their existing life 3 intolerable (Floyd 2017: 168–169). Consequently, political systems causing fewer instances of such behaviour are more defensible than those leading to more, and certainly, the system resulting in the minimum of such behaviour is, without doubt, more justifiable than any other (Floyd 2017: 169). Due to its reliance on patterns of thoughts, mainstream political theory is destined to fail, according to Floyd, because reasonable people diverge on the fundamentals of political normativity (Floyd 2017: 121). By contrast, normative behaviourism’s reliance on patterns of behaviour is more promising, since they converge and are stable over time. By examining our reactions to changes in our political surroundings, we can discern which systems we are less inclined towards, which ones we favour more, and perhaps even identify the one that suits us most effectively. From this, we may conclude that there exists at least one method to justify political principles, namely, the principles embodied by the most fitting system, that is, the system with least insurrection and crime (Floyd 2017: 167). Normative behaviourism shares many features with realism and is therefore described as a member of the ‘realist family’. The starting-point for both approaches is the idea that we can learn about what we ought to do now and in the future by looking at the past, such as what political systems have had the least amount of insurrection and crime (normative behaviourism) and what has historically and over time been the main stabilizing features of politics (realism). Both approaches start out from the idea that political theories should ground their political principles in ‘real actions’ of ‘real people’, thereby justifying principles with actions, rather than ground their political principles in ‘hypothetical actions’ of ‘imaginary people’, justifying actions with principles, as in mainstream political theory (Floyd 2022: 2). Hence, normative behaviourism shares Geuss’ famous realist point of departure, according to which political theory must start from and be concerned with the way the political, economic, and social institutions actually operate in a society at a given time, rather than with how people ideally or rationally ought to be or act (Geuss 2008). In a similar fashion, it “begins with what 4 actually motivates people in the real world, and then asks, empirically speaking, just which set of principles, in practice, best caters to them” (Floyd 2022: 6; Galston 2010, 2017). Furthermore, normative behaviourism is committed to a focus on ‘realist’ values, most notably the values of order, stability and legitimacy. Both approaches understand politics at the general level as unchangeable, characterized by contestation, deep disagreement and a struggle for power and therefore regard the goal of politics as order and stability, but that this must take a legitimate form (Floyd 2022; Sleat 2018; Hall 2017). For normative behaviourism, this goal is achieved by identifying those principles that in practice produces least discontent in society. This entails that focus is directed at “whatever principles are perceived as most legitimate by those who have to live with them” (Floyd 2022: 7). Finally, normative behaviourism shares with realism the ambition of offering clear political prescriptions (Jubb 2019; Sleat 2018). Since it aspires to understand people as they are, rather than as they ought to be, it moves from behaviour to institutions and then to principles; in other words, from ‘is’ to ‘ought’ rather than from ‘ought’ to ‘is’. For this reason, Floyd argues, normative behaviourism never ends up with principles that are not action-guiding, which is a problem facing mainstream political theory, asking how ideal principles may be applied to concrete practices (Floyd 2022: 8). Despite these overlapping commitments, Floyd insists that normative behaviourism advances realism since it is equipped to solve several problems facing realism. In this paper, we focus on two of these problems: the legitimacy problem and the institutional problem. II. Rebuttal of the solution to the legitimacy problem In a nutshell, the legitimacy problem concerns how we should distinguish legitimate acceptance of a political system from acceptance based on false consciousness or 5 coercion. When responding to this problem, it is important to avoid the problems encountered both by the ideal route taken by mainstream political theorists, which typically advocate majoritarian democratic procedures, and the realist route, which typically define as legitimate whatever is accepted as legitimate by the people in question (Floyd 2020: 145). The ideal route has the problem of grounding such procedures since reasonable people disagree even about fundamental values (Floyd 2017: 121). The realist route has the problem of distinguishing authentic acceptance from acceptance based on false consciousness or coercion (Floyd 2020: 145). Normative behaviourism adopts the realist strategy of basing authentic legitimacy on perceived legitimacy. Therefore, just as realism, it has to differentiate legitimate acceptance from false or coerced acceptance. However, it does so, not by asking what the majority would say about the suggested political principles, but by observing how people behave under different systems in history. If more people turn to crime and insurrection under a given regime than under alternative feasible regimes, that regime is illegitimate, even if the majority at present perceive of it as legitimate. This is so because actions speak louder than thoughts (Floyd 2020: 146): while realists start their normative investigation with facts about politics, focusing on politics’ constitutive features, such as deep disagreement, normative behaviourism only counts facts about certain behavioural trends (with regard to crime and insurrection) when assessing governing political principles and their relative legitimacy (Floyd 2020: 146). This approach to the legitimacy problem encounters two problems for normative behaviourism, according to Floyd. One problem is epistemic and concerns lack of access to proper information in authoritarian states, whose leaders often hide statistics about crime and insurrection. Another problem is scalar and concerns how many people would have to turn to crime and insurrection before a set of principles of legitimacy should be abandoned. To respond to the first problem, normative behaviourism adopts a comparative approach in assessing these principles. The aim is to root for the political system that produces the least amount of these two behaviours. To respond to the second 6 problem, normative behaviourism adopts a long-term outlook and focuses on large numbers of people (Floyd 2020: 146). II.I. Facing a dilemma Floyd’s responses to these two problems, however, reveal tensions within normative behaviourism that, upon further investigation, presents a dilemma for the account: either it faces an analogue problem of distinguishing authentic acceptance from acceptance based on false consciousness or coercion as do political realism, or it solves it by adding value premises, in effect becoming a target of his own attack on both idealism and realism. A fundamental tactic for an account belonging to the realist family is, as Floyd puts it, to ground authentic legitimacy on perceived legitimacy. The realist tactic, however, struggles to distinguish authentic acceptance from acceptance based on either coercion or false consciousness. This then encourages attempts to distinguish between reasonable and unreasonable acceptance, and ultimately between what a reasonable and an unreasonable person could accept, which means, in parallel with the ideal response, that rather than rising above what appears to be intractable valueconflict, realists can fall right back into it (2020: 145). Floyd’s strategy for avoiding this pitfall, as mentioned above, is to use as the condition on authentic legitimacy, not what people say – or indeed, think – about a system of government, but their actual behaviour. Floyd envisages cases where a majority of people take their rule to be legitimate for all kinds of problematic reasons, such as ignorance of alternative systems, or that they take their leader to have a divine right to power. On his account, however, this would not entail that the rule is legitimate. “[I]f more people find themselves turning to insurrection and crime under a given regime than under alternative feasible regimes, then that regime is indicted, even if at present the majority declares it legitimate” (2020: 145-46). 7 Floyd thus argues that normative behaviourism is an improvement of realism since although it is similarly grounded in actual politics – how things are rather than how we wish them to be – it avoids turning to values in order to distinguish authentic legitimacy from perceived legitimacy and thus becoming ‘closet-idealism’ or ‘closetmoralism’. In other words, since according to normative behaviourism, what people (doxastically) take to be legitimate is clearly distinct from what is legitimate, the account is not forced to take on the kind of value commitments that tempts realists when they want to distinguish what is accepted from what is acceptable (that is, if they want a normative account of legitimacy rather than merely a sociological one), for example, by arguing that in order for a political order to be legitimate, the agreement (Horton 2010) or willing consent (Bellamy 2010) must be perceived as free (Newey 2010), and cannot rely on means that are too tyrannical (Horton), be coerced (Williams 2005) or be a result of total deception (Horton 2010). In one sense, looking at actual behaviour clearly solves the legitimacy problem when people say or think one thing but do another. The overall success of Floyd’s strategy, however, is a function of how the behavioural aspect is interpreted in more detail. And on a closer look, it seems as if Floyd is facing a dilemma. On the first horn of the dilemma, the condition for legitimacy is interpreted as a function of people’s present behaviour. On this interpretation, you can ‘read off’ the legitimacy of the system of rule in question directly by looking at its current level of insurrection and crime. If that level is less than other feasible systems of government, it is legitimate. Otherwise, it is not. This ‘direct’ interpretation of the legitimacy condition for normative behaviourism seems natural given Floyd’s ambition to go from ‘is to ought’, reading off what is right from people’s actual behaviour (without sneaking in any value premises on the way). It also fits best with the above quote that “more people find themselves turning to insurrection and crime under a given regime”. However, this interpretation does not avoid the traditional realist problem of distinguishing what is accepted from what is 8 acceptable. For it is reasonable to assume that false consciousness or coercion may not only have doxastic consequences, making us believe that the rule is legitimate, but also affect our behaviour in various situations. By falsely believing that the ruler has divine power, or at least is blessed by someone who has, we may refrain from rebelling in cases where we would otherwise turn to insurrection and crime. Or (which is arguably even worse), what if a power is simply too coercive to make rebellious action feasible, such that even if we both believe that the power is illegitimate and definitely would act on that belief by acts of insurrection and crime were the coercion less severe, we are now unable to do so? Perhaps the punishment is simply too harsh, or the control too rigid for us to succeed even if we tried? It thus seems – on the first horn of the dilemma – as if the ‘direct’ reading of the legitimacy condition in fact fares worse than basing legitimacy directly on what people think is legitimate, since both cases where we act upon false consciousness and cases where we would like to act but cannot, become legitimate. Assume, say, that the society of Orwell’s 1984 displayed minimal levels of insurrection and crime; it would then mean that it is in fact legitimate. Even worse for the realist project, accepting such a low-bar condition on legitimacy would not only legitimate fictional nightmares states, but also present-world dictatorships or authoritarian regimes of which many oil-states make out the paradigm. And that would be much too big a bullet to bite. Floyd’s solution to this potential problem is to go for historical track records. Rather than looking directly at present behaviour, normative behaviourism looks for large numbers and long-term trends. This means that we cannot justify a new experimented-with set of principles on the basis of a few years of success in a single context. We can however, at least sometimes, reject such experiments, by noting that all the key elements of an earlier failed system are present within it. In these cases we know that the ‘new’ experiment will ultimately fail, at least once temporary environmental conditions have been removed (2020: 146). 9 In other words, on Floyd’s official program, what determines whether or not a system of government is legitimate on normative behaviourism is not the present level of crime and insurrection compared to alternative orders, or even the level of crime and insurrection during the entire regime of the present order (‘a few years of success in a single context’). In order for the power to be legitimate, the principles governing the society should “produce less insurrection than any other kind of political system in recorded history” (2020: 140).1 With this move away from present levels of crime and insurrection to historical track records, Floyd gets a contrast class that enables him to reject what is likely to be temporary situations of behaviour satisfaction. As Floyd puts it: “If its historical record is poor, then its political present – perhaps buoyed by natural-resource money, or steep but unsustainable economic growth, or a charismatic leader whose star will eventually wane – is comparatively irrelevant” (2020: 146). But we ask: irrelevant according to whom, and, more importantly, on what grounds? Floyd’s quick conclusion might seem natural: of course, with gold and sweet talk it might be possible to keep people satisfied for a while, but eventually the spell will break, and so, looking at (dis)satisfaction over time is the solution. However, this is still just our pre-theoretical intuitions talking, and so we must ask ourselves how the political present becomes “comparatively irrelevant”, according to the normative behaviourist account. Pondering this question, it seems clear that steering away from the first horn resulting from the ‘direct’ interpretation of the legitimacy condition of normative behaviourism, places Floyd firmly in front of the second horn: his account becomes reliant on the same mainstream ‘moralist’ value commitments he argued threatens 1 The last being his claim about social-liberal democracy – the system of rule that Floyd argues, as a matter of empirical fact, to be the legitimate one. 10 traditional realist accounts when they want to avoid conflating perceived legitimacy with actual legitimacy. This is so since the strategy used in order to avoid the first horn of the dilemma is grounded on a number of implicit value commitments that are just as sensitive to “intractable value-conflict” (2020: 145) as Floyd accuses the realist solutions to be. This becomes evident when highlighting the fundamental basis of Floyd’s account: reading off what is right from people’s behaviour rather than their mental states. If actions really speak louder than words, if what counts is people’s actual behaviour in actual situations rather than their mental states – typically concerning hypothetical situations, thought examples and intuitions – it seems to entail, as in the first horn of the dilemma discussed above, that the legitimacy of a political rule is determined by people’s actual satisfaction (or rather, lack of dissatisfaction). Floyd’s preferred condition, however, relies on the value commitment that large numbers and long-term historical trends determine the legitimacy status of a political rule. This implies that no current rule is legitimate unless the principles that it embodies have a proven historical record, which is the paradigm of a conservative value commitment.2 Moreover, by rejecting a given system of rule by referring to “key elements of an earlier failed system”, we are relying on value commitments to the effect that some elements (judged as ‘key elements’) trump all other elements, however well-supported by good reasons and positive data, even including very low present levels of crime and insurrection. The upshot of Floyd’s strategy is thus that his account, no less than realists trying to establish authentic acceptance, also “encourages attempts to distinguish between reasonable and unreasonable acceptance, and ultimately between what a reasonable and an unreasonable person could accept” (2020: 145). Hence, rather than improving political realism by solving the legitimacy problem, normative behaviourism inherits exactly the 2 Floyd indeed acknowledges that one of the potential objections to his account is “that normative behaviourism is excessively conservative, given that it apparently only justifies the historically-tested status quo” (2020: 148, 2023). Note here, though, that our point in not that the commitment of justifying the historically tested is problematic, but that it is, no doubt, a value commitment. 11 same ‘problem’3 of relying on value commitments and thus fails to “escape the kind of value-conflict it wants to transcend” (2020: 145). In the rest of this section, we trace these challenges facing normative behaviourism to its fundamental thesis that the account is grounded in behavioural facts rather than values and explain why this thesis is flawed. II.II. The role and status of fundamental principles In his writings, Floyd repeatedly and emphatically states that his account is grounded by facts rather than values, and spends a considerable effort to defend this thesis (Floyd 2017, 2020). We take this defence to have two related parts, which we treat in this subsection and the next. Floyd’s first defence is aimed directly at G. A. Cohen, who famously has argued that facts cannot ultimately ground principles (Cohen 2003, 2008). Cohen argues that no fact of any kind can ground a normative principle, unless there is a further normative principle which explains why this is so. Ultimately, Cohen holds, there will always be a normative principle that is not grounded on facts. Floyd takes Cohen’s argument to be the most important objection to his account (2017: 170-171, 2020: 148), and questions both its relevance and truth. Before turning to Floyd’s argument, let us rehearse the thrust of Cohen’s line of argument.4 Cohen distinguishes between two kinds of principles: those that are dependent on facts and those that are fact-independent. On his account, “a principle can reflect or respond to a fact only because it is also a response to a principle that is not a response to a fact” (Cohen 2003: 214). The idea is that if we have a principle P that is sensitive to facts F, there is another “more ultimate principle that explains why F supports P” (Cohen 2003: 3 We use scare quotes here since we do not share Floyd’s premise that our actual disagreement in value questions is a problem, neither a justificatory or a methodological one. We will not pursue that line of argument here, though, since our claim in the paper is simply that Floyd’s account does not improve on political realism. 4 See Erman and Möller 2023. For an extended analysis, see Erman and Möller 2016, 2017, 2019. 12 218). Cohen exemplifies with the principle ‘we should keep our promises’ (A), which we assume depends on the fact that only then can people pursue their projects. Principle (A), Cohen argues, is true only because there is a further principle (B), ‘we should help people pursue their projects’ (Cohen 2003: 216–17). This further is also true – if true it is – only because of a further fact, namely that people can achieve happiness only if they pursue their projects. But the latter fact is explanatory salient only if there is yet another principle (C) that states ‘we should promote people’s happiness’. For a utilitarian hedonist, we have now reached a principle which is true regardless of any further facts; it is a fact-free normative principle. If we are not utilitarians, however, the regress continues at least one more step, since there must be some fact which explains in virtue of what we should promote people’s happiness. But eventually, we reach some normative principle which holds regardless of any further fact. Floyd’s first line of defence accepts, for the sake of argument, the truth of Cohen’s claim, but argues that it is irrelevant: “facts matter more than ultimate principles when justifying either moral or political principles. This is because ultimate principles are so abstract, due to their fact-independence, as to be virtually useless, normatively speaking” (2020: 149). Even if there are fact-free ultimate principles, Floyd continues, “it will still be the case that, as far as political philosophy is concerned, fact-dependent principles are where the action is” (2020: 150). The kind of principles of interest in political philosophy, Floyd argues, are those which pass tests like a reasonable rejection test, so if we show “that only one set of principles passes that test, then it will generally be considered that one has provided all the grounding one needs” (2020: 150). Unfortunately, Floyd’s line of argument here seems to misunderstand the relevance of Cohen’s argument. Of course, as the history of moral and political philosophy has showed, an ultimate grounding principle might be abstract: Kant’s categorical imperative (in any of its versions), or respecting the autonomy of each person, or maximizing the well-being of sentient beings; these are all examples of fundamental principles that are abstract, in need of both interpretation and further (typically factual) 13 premises in order for us to apply them in practice. But although I cannot always – and perhaps not even typically – directly ‘read off’ the answer to whether I should, say, tell a lie in a certain situation from a set of ultimate grounding principles, the correctness of the actual – intermediate – principle, which I might use in order to figure this out (e.g. “lie only in cases where telling the truth would only hurt the recipient”), will depend on some underlying principle, which, and this is Cohen’s point, will ultimately be a fact-free principle, i.e. one that is correct regardless of some further fact. In other words, even if the principles which we typically deal with in political philosophy are less abstract than any fact-free grounding principles would be, the potential correctness of these political principles are ultimately explained by some fact-free principles. This means that Floyd’s claim that behavioural facts (crime and insurrection) determine the legitimacy of a rule, that historical records rather than present levels are decisive, and that key elements of earlier failed system may invalidate a present system of rule, are all values that are either themselves fact-free, or are ultimately explained by a principle that is so. That such principles may be ‘further down the explanatory line’ is neither here nor there. If the foundation is not solid, the whole building falls apart regardless. Hence, Floyd’s argument against the relevance of the claim that facts cannot ground principles is unsuccessful, but what about his objection against its truth? Floyd’s main argument here is that “principles, in order to be justified for a particular kind of creature (e.g. a human being), have to be grounded in facts about that creature” (2020: 152) Sheep and bears are factually different from humans, and thus the principles that apply to them do not apply to us, and vice versa. That means, Floyd argues, “that for a principle to be justified, it has to be grounded in facts about the kind of creature it is intended for” (2020: 151). It might be facts about how a certain kind of creature thinks, how it behaves, or what it feels, but regardless, it must be some set of facts grounding our principles. “This means,” Floyd concludes, “that there are no ultimate principles; only ultimate facts” (Floyd 2020: 151). 14 Again, Floyd’s argument misses its mark. He is definitely right in claiming that facts play a role in the grounding of our principles. That we should not hit the person next to us in the head is typically justified by facts such as that it would hurt her or be against her will. And more generally, moral and political principles are typically justified by reference to all kinds of empirical facts about humans, such that we are rational beings, that we feel things, and that we have certain wills and desires. But this is perfectly in line with Cohen’s argument. Let us take the example of a principle such as ‘you should not hit a person on the head.’ That principle, let us say, is grounded in the fact that hitting a person hurt them. Well, this just means that the truth of that principle is dependent on another principle, ‘you should not hurt people.’ Now – according to the now familiar process – either that principle depends on a further fact, or it is true regardless of factual circumstances. Let us say that we give a number of further explanations for why hurting people is bad (such that it is detriment to our health, makes us sad, etc). Eventually, the list of reasons will run out, and we will be left with a principle stating that ‘Given that hurting people do …[the full list of factual reasons], it is wrong to hurt people.’ Finally, since we have assumed that all relevant facts about why hurting people is bad is included in ‘…’, this principle is fact-free, i.e. it is valid regardless of what the ‘ultimate facts’ are. In other words, the fact that the kind of creature we are gives rise to moral principles which should otherwise not hold, cannot escape the fact that ultimately, when we reach rock-bottom, our principles will be fact-free. That the level of crime and insurrection determines the legitimacy of a rule and that historical records trump current records, if indeed that it the case, are either fact-free or are explained by principles which are. In sum, normative behaviourism does not in fact move from ‘is to ought’, as is proclaimed, but from ‘ought to is’, as any account in normative theory. 15 II.III. The role and status of reasons internalism Floyd’s second defence of the thesis that normative behaviourism is grounded in facts rather than values utilizes the idea of reasons internalism, i.e., the idea that in order for a person to have a reason to act in a certain way, some motivational fact must be true about her (cf. Darwall 1983; Schroeder 2007). As a consequence, Floyd argues, “imperatives are, when properly understood, hypothetical imperatives, given that they depend upon contingent features of the motivational make-up of the creatures for whom they are intended” (Floyd 2020: 151). More specifically, Floyd endorses a version of the Humean theory of reason in which it is necessary, in order for someone to have a reason to act in a certain way, that she has a desire or will that would be served by this act (Floyd 2017: 172-173). Introducing this idea, Floyd gives the parallel cases of a recommendation for a bear to go fishing in a river full of salmon, and the recommendation for humans to choose a social-liberal democratic rule (i.e. the kind of rule he argues fulfils his condition of minimizing crime and insurrection). “Why”, he asks rhetorically about the former, “would this recommendation have no normative force if we can observe both that the bear would like to eat salmon and that fishing in the river would enable it to do so?” (Floyd 2017: 172). And why, he asks about the latter, “would this recommendation have no normative force if we can observe both (1) that humans want to avoid these things [famine, war, disease, subjugation, and instability] and (2) that building this political system would enable them to do so?” (Floyd 2017: 172). Floyd finds it obvious that both of these cases do have normative force, and that they do so since the recommendations involve satisfying the needs and desires of the bear and the humans, respectively. Discussing a variety of choice situations – from choosing a restaurant to a partner or a career – he argues that being reasonable “means nothing more or less than consistent faithfulness to one’s pre-existing needs and desires, 16 including their relative strengths and the connections between them” (Floyd 2017: 173).5 He continues: Reason is thus about logic in the sense that it is about being consistent with and faithful to my pre-existing needs and desires. Eating mud, simply because you hate it, is illogical, and thus unreasonable. Eating cheese, simply because you like it, is logical, and thus reasonable – provided, again, that there are no further, stronger, and countervailing needs or desires. And so it goes on (Floyd 2017: 173). Since it is a factual matter what our desires are, Floyd argues, what we have reason to do is grounded in facts. Hence, facts ground principles, and we can go from ‘is’ to ‘ought’ rather than the other way around. The quick way of objecting to the above argument is of course to simply apply Cohen’s point to Floyd’s argument. That is, even if Floyd is right, it would still be the case that what ultimately explains the normative force of the recommendation for a bear to go fishing in a river full of salmon, and the recommendation for humans to choose a socialliberal democratic rule, would be the (fact-free) principle that we should do what, and only what, would serve our “pre-existing needs and desires”. That is a value commitment that Floyd, and everyone convinced by his account, must hold, in just the same way as mainstream political theorists must hold on to their respective basic value commitments. However, since adhering to internalist accounts of reason in general and the Humean theory of reason in particular is popular in realist accounts in general, it may be valuable to look more closely at Floyd’s utilisation of Humean theory of reason, and how it coheres with his account. 5 On an exegetical note, it should be mentioned that there are several ways to interpret the ’need’ part of Floyd’s reasons-claim here, some of which are in tension with the Humean theory of reason, since what we desire and what we need are different things. However, since Floyd explicitly sides with “Hume’s famous dictum about reason always being a slave to the passions” (Floyd 2017: 175) and every example Floyd brings forward involves only the desire part, we take it to be clear that Floyd has the Humean theory of reasons in mind here. 17 To begin with, it is important to distinguish the idea that human needs and desires are typical – and legitimate – sources of normativity, from the idea that for something to be a reason to act, it must contribute to the satisfaction of a desire. The former is commonplace in contemporary political theory. Hence, it is rather uncontroversial to hold, at least prima facie, that the bear both desires and needs food such as salmon gives it (or constitutes) a reason to go fishing, just as you do not have a reason to eat mud but a reason to eat cheese, if you hate the former and love the latter. What is not uncontroversial, of course, is the latter claim that a person’s (or indeed a creature’s) reasons can be reduced to whatever contributes to fulfilling its desires and needs. The main problem for the controversial view that “faithfulness to one’s preexisting needs and desires” is necessary for someone to have a reason to act in a certain way, is that it has some extremely repugnant conclusions. Surely, eating cheese, simply because you like it, is reasonable. But what about people who happen to desire awful things? To take an infamous example, imagine a person who desires to torture babies. That desire might be direct, in the sense that the person in question gets some direct satisfaction from the torture, or it might be indirect, in the sense that the person believes that babies love being tortured, together with the normative belief that you should make people happy. Regardless, it seems highly problematic to say that the person has a reason to torture babies. And it seems equally problematic that a person has no reason at all, say, to defuse a bomb that would kill an entire skyscraper full of people, simply because she has no desire to do so. In other words, it seems highly problematic to hold the general view, as Floyd does, that “[m]y feeling unhappy, for example, is only a reason for you to help me if you care at some level about my happiness. Sure, I could say you should care, but that would only bite if I could find some other commitment underpinning it you do already hold” (Floyd 2023: 492). While Floyd might be right that causally (or motivationally) speaking, a norm only has ‘bite’ if it has some committable underpinning, as a view on normative reasons suitable for moral or political theory, it is problematic indeed. Consequently, it is 18 not, as Floyd writes: “because you think you would save an imaginary child drowning in a pond that you ought to accept a particular principle of justice. It is because you think you would turn a runaway trolley this way or that way that you should accept a particular principle of warfare” (Floyd 2023: 491; italics in original). So while we agree that eating cheese, simply because you like it, is perfectly reasonable, we insist that torturing babies simply because you like it, if that happens to be the case, is not reasonable, even if you have “no further, stronger, and countervailing needs or desires” (Floyd 2017: 173). One way out of this ‘repugnant conclusion’ would be to argue that while reasons ultimately are grounded in desires, what matters is not people’s actual desires (“one’s pre-existing needs and desires”) but people’s desires under some suitable counterfactual circumstances. The most popular suggestions here include some version of having full information (e.g. Smith 1994), being rational (e.g. Korsgaard 1996), having a coherent belief-set (e.g. Brandt 1979), or being emphatic (e.g. Korsgaard 1996; McDowell 1979). On such suggestions, the aspiring child-torturer should arguably not have reasons to act on his desires, since his fully rational, informed and emphatic version would not have any child-torturing desires. Whether suggestions as the ones above really avoid all versions of what we above called the repugnant conclusion is a matter of controversy, but it seems clear that they stand a much better chance to do so. However, what is also clear is that they are highly idealized versions of reasons internalism.6 Consequently, were Floyd to adopt such an interpretation of his Humean theory of reason, his account would also in this respect be just as ‘idealistic’ as the ‘ideal’ accounts against which he contrasts himself (as well as accuses other political realist accounts to fall back to). Indeed, Floyd’s enterprise of staking his account on a specific idea on what constitutes a reason seems, even if ignoring its problematic entailments, to be against his overall aim to avoid ‘disagreement about 6 Indeed, in their overview of reasons internalism, Lord and Plunkett (2017: 334) call such versions ‘idealized internalism’. 19 thoughts’ that in his view haunts mainstream political theory. As mentioned above, reason internalism as such is highly controversial, and his ‘actualist’ account even more so. Hence, even if he were to falsify Cohen’s thesis with his Humean theory of reason (which he does not), it would be at the cost of going against his own methodological criteria. III. Rebuttal of the solution to the institutional problem So far we have treated our first claim, i.e., that normative behaviourism does not solve the legitimacy problem. Let us now move to our second claim, i.e., that normative behaviourism does not solve the institutional problem either. The institutional problem concerns how we should translate political principles into feasible political institutions. It starts with two complaints against mainstream political theory made by realists: that its political principles are typically too abstract and therefore too indeterminate to generate workable institutional prescriptions, and that its political principles typically entail unfeasible institutions, violating the proviso ‘ought-implies-can’. Realists themselves, however, have not come up with a viable solution to this problem either, according to proponents of normative behaviourism. In Floyd’s view, the appropriate way of addressing this problem is by starting out from proven institutions and thereafter move to abstract principles. In other words, moving ‘from-is-to-ought’, rather than the other way around, normative behaviourism begins with a particular set of institutions that generates the least insurrection and crime in practice, and then argues that whatever principles these institutions express are justified. It thereby offers both determinate and feasible institutional proposals, something which allegedly neither mainstream political theory nor realism does (Floyd 2020: 147). As we demonstrated in the previous section, since normative behaviourism does not in fact move from ‘is-to-ought’, as is proclaimed, but from ‘ought-to-is’, as any account in normative theory, the proposed solution to the institutional problem starts from a false 20 premise. However, would the account solve this problem if the justification would have been successful? This is what we explore in the current section. Let us begin with Floyd’s general requirement for any successful account in political theory, notably, that “a convincing and meaningful answer” to political philosophy’s organizing question – i.e. how should we live? – must fulfil two criteria: (1) it must be politically determinate (“in terms of its principles and institutions” (2017: 229), and (2) it must be “rationally compelling to all those who would have to live with it” (2017: 166). We see no support for why normative behaviourism would fare better with regard to fulfilling these two criteria than mainstream political theory, such as Rawls’ theory. Recall that the main argument for why the former offers a more convincing answer than the latter is that the latter relies on agreements on ‘patterns of thoughts’ whereas the former relies on ‘patterns of behaviour’, and while people disagree about thoughts, they converge in their behavioural patterns. However, we don’t find this inference persuasive. The role of agreement that Floyd assigns to mainstream political philosophy rests on a misunderstanding. Indeed, not even Rawls’ concept of ‘overlapping consensus’ relies on actual consensus for the justification of his theory. Therefore, it is an open question which of the two approaches best fulfils the two criteria. With regard to the first criterion, it surely seems as if Rawls’ two principles of justice are far more determinate than Floyd’s, since Floyd does not formulate any substantive principles but simply stresses that the best (most legitimate) political system is ‘socialliberal democracy’, incorporating two basic values: the rule by the people (democratic decision-making) and equality. Since largely all contemporary political theories incorporate some version of these values into their accounts, this answer can hardly count as “politically determinate”. With regard to the second criterion, we cannot see why Rawls’ two principles would be less “rationally compelling” to those who would have to live with them than socialliberal democracy. Floyd puts forward four arguments for why people “should be convinced by social-liberal democracy” as an answer to the organizing question. It is 21 convincing for those who care about security, who prioritize the avoidance of cruelty, who want to see people flourish, and who find social-liberal democracy less unbearable than any other attempted system (2017: 229-231). However, it is not self-evident that these values are more rationally compelling than Rawls’ principles. Now, while it is wholly unclear why normative behaviourism would fulfil the two criteria of a sound political theory better than mainstream accounts generally, it does seem as if the criteria are connected to feasibility concerns, such that a satisfactory fulfilment of them would constitute a possible solution to the institutional problem. A determinate principle is typically more practically applicable than an indeterminate one, and likewise, a principle which is rationally compelling to those who are supposed to abide by it, raises the likelihood of it being successfully adopted and implemented. As we saw above, Floyd’s strategy here is to start out from institutions that in fact generate the least insurrection and crime, and then make explicit the principles regulating them (which hence are considered justified). By doing so, these institutions become both determinate and feasible, according to Floyd, and more so than proposals suggested by realists (2020: 147). However, we cannot see why normative behaviourism offers a convincing response to the institutional problem and in what sense it improves political realism. There are two reasons for this. First, if a feasible political system is a system in which its institutions have proven that they are legitimate (social-liberal democracy), i.e., have generated the least insurrection and crime, then the question of feasibility becomes redundant or irrelevant, since such a system is already up and running, as it were. While that would entail a solution to the institutional problem where there is no such problem, it will give no guidance for any other societies. Second, if we were to apply the principles regulating such a social-liberal democracy (e.g., the Danish system, which is Floyd’s example) to a country like Zimbabwe, it seems likely that this is less feasible than applying any principle of legitimacy suggested by realists. Floyd’s own argument is that in situations where we 22 lack the resources needed to implement his suggested principles, such as in a country like Zimbabwe, there is no evidence that introducing egalitarian-liberal-democracy overnight is impossible. History is full of societies which, after standing still for centuries, transformed their economic and political arrangements at relative speed. Think, for example, of some of the consolidated democracies of East Asia (Floyd 2020: 147-48, our italics). But if this is the argument, it holds for virtually any mainstream account too: as long as it is not proven to be impossible (and thus does not violate the proviso ‘ought implies can’), it is feasible. Moreover, and more importantly for our purposes, it seems as if realists propose more feasible prescriptions after all, since they typically reject the idea that there is a single set of criteria that all legitimate political systems must meet. Instead, realists emphasize that legitimacy involves delivering on values and managing conflicts in ways that are contextually sensitive and acceptable to the subjects of political power. While they maintain that some forms of power or authority are illegitimate, such as those that are based on unnecessary coercion or deception, political legitimacy typically requires something much weaker than the principles regulating a social-liberal democratic system. For Bernard Williams, for example, what is required is that the exercise of political power is justified in ways that the subjects of that power can understand and potentially accept (Williams 2005). Hence, rather than proposing demanding (sometimes unfeasible) principles requiring democratic rule and equality, it suffices that the system can provide acceptable reasons for its actions, which involves being responsive to the values, interests and understandings of its subjects. Even less demanding and thus more feasible for societies like Zimbabwe, Raymond Geuss’ view is that a legitimate system is one that can navigate and manage conflicts effectively within their particular contexts (Geuss 2008). A third realist example is John Dunn, who argues that that legitimate 23 political systems are those that can effectively deliver on the values that they promise to their citizens. This could mean a promise of democratic control, as demanded by normative behaviourism, but it could also mean something much less ambitious, such as protection against poverty (Dunn 2005). The point here is simply that many realist proposals, such as the three mentioned here, offer far more viable solutions to the institutional problem than normative behaviourism. IV. Winding up: new problems facing normative behaviourism Let us wind up our analysis by bringing up two new problems facing normative behaviourism, which realist approaches typically do not face. We argue, first, that normative behaviourism is a kind of closet utilitarianism, but with a more problematic value measure that rests on universal principles which realists reject, and second, that normative behaviourism is both too demanding and normatively unattractive for realists. Normative behaviourism and utilitarianism are structurally very similar. For normative behaviourism, a fundamental premise is that dissatisfaction among citizens in society is bad, while for utilitarianism a fundamental value premise is that well-being, in terms of desire-fulfilment, happiness, or preference-fulfilment, has non-instrumental value and is good. Connected to these values are some fundamental principles. For normative behaviourism, those societies are best whose principles generate as low level of dissatisfaction among citizens as possible ”with both the political system and the ways of life it facilitates” (Floyd 2017: 189), while for utilitarianism, roughly put, those societies are best whose principles generate as much well-being as possible. Apart from similar value premises and principles, both approaches have similar measurements and hence empirical implications. According to normative behaviourism, the amount of dissatisfaction among citizens is measured in terms of the level of crime and the level of insurrection, while utilitarians typically measure the maximation of well-being (or something similar). On one influential view, what matters for well-being is the overall 24 level of preference-satisfaction in a person’s life as a whole (Crisp 2021). Lastly, with regard to output, for both approaches it becomes mainly an empirical question which kind of society the theories suggest (Erman and Möller 2023). In sum, both approaches focus on actions, behaviour and well-being. For normative behaviourism, the justification of principles roughly equals “political commitment X numbers, with numbers meaning the percentage of people who commit a certain type of action under a certain type of political system” (Floyd 2017: 169), while for utilitarianism, the justification of principles roughly equals the sum of well-being over all individuals. Yet, Floyd refuses to recognize these structural similarities and the accusation of being a closet utilitarian (Erman and Möller 2023). On his view, utilitarianism fails because the justification of “any particular metric of happiness” would have to rely on “convincing reasons for tracking one metric over another”, about which people disagree (2020: 154). This is not the case for normative behaviourism, because “real actions really do ground political principles” (Floyd 2023: 492). However, this is simply not true. As we saw above in our discussion of the legitimacy problem, no actions can by themselves ground any principles. Just because normative behaviourism is grounded in a negative and non-utopian value (low levels of dissatisfaction), rather than, say a positive and utopian value (e.g. total equality or distributive justice), it is no less a value premise. As pointed out by Judith Shklar, who grounds her ‘negative’ account in a ‘fear of cruelty’, a liberal political theory cannot rest on “this or any other naturalistic fallacy” (Shklar 1989: 30). To become “a principle of political morality”, she argues, liberalism of fear needs to be recognized as “a necessary condition of the dignity of persons” (1989: 30). Similarly, low levels of dissatisfaction must be recognized as necessary condition for a good (legitimate) society on the normative behaviourist account. A value premise, in other words. In sum, normative behaviourism encounters a new problem that realism does not have, since realists would wholeheartedly reject the substantive universal principles of the kind that normative behaviourism embraces (see Modood 2023). Realists are 25 typically contextualists, at the most defending thin universal procedural principles as exemplified earlier. Also a second problem materializes for normative behaviourism in relation to realism, namely, that its prescriptions become both too demanding and normatively unattractive for realists. For sure, realists share with normative behaviourists the worry that the ‘status quo bias’ may be a major obstacle, since both wish to ground their accounts in facts while at the same time envisage radical departures from our current state of affairs (Prinz and Rossi 2017; Rossi 2016; Floyd 2022: 9). However, to recommend something (social-liberal democracy) “for ‘here’ based on what happened over ‘there’, or even ‘anywhere’ in the past” (Floyd 2022: 14). seem to go against realism in several respects. First, one of the key features of realism is to carefully distinguish, both conceptually and normatively, political legitimacy from democracy. It would simply be too demanding in most contexts for political legitimacy to require democracy and thus too context-insensitive (see Baderin 2023). Second, it would also be highly unattractive, since those who call themselves ‘radical realists’ see capitalism as the main obstacle for a legitimate political order. They typically reject or question this economic system, which is both praised and protected in all social-liberal democratic systems (we have witnessed so far in history). As noted above, realists are often contextualists and would not accept any substantial and context-insensitive universal principles of any kind. By contrast, Floyd is not contextualist with regard to the justification of the correct principles – since these are justified from low levels of crime and insurrection – but at the most with regard to the implementation of those principles (feasibility). In sum, we conclude that normative behaviourism fails in its attempt to supply an improved version of realism and encounters new problems that realists typically do not face. 26 References Baderin, Alice. 2023. “Behaviour and Thoughts: For a Pluralistic Model of Empirically Informed Political Philosophy.” Political Studies Review 21(3): 476–482. Bellamy, Richard. 2010. “Dirty Hands and Clean Gloves: Liberal Ideals and Real Politics.” European Journal of Political Theory 9: 412–30. Brandt, Richard. 1979. A Theory of the Good and the Right. New York: Clarendon Press. Cohen, G. A. 2003. “Facts and Principles.” Philosophy & Public Affairs 31 (3): 211–245. Cohen, G. A. 2008. Rescuing Justice and Equality. Cambridge, MA: Harvard University Press. Crisp, Roger. 2021. “Well-Being.” In The Stanford Encyclopedia of Philosophy ed. E. Zalta (Winter 2021 Edition). Available at: https://plato.stanford.edu/archives/win2021/entries/well-being/ Darwall, Stephen. 1983. Impartial Reason, Ithaca: Cornell University Press. Dunn, John. 2005. Setting the People Free: The Story of Democracy. Princeton University Press. Erman, Eva and Niklas Möller. 2016. “What Distinguishes the Practice-Dependent Approach to Justice?” Philosophy & Social Criticism 42 (1): 3–23. Erman, Eva and Niklas Möller. 2017. “Practice-Dependence and Epistemic Uncertainty.” Journal of Global Ethics 13 (2): 187–205. Erman, Eva and Niklas Möller. 2019. “How Practices Do Not Matter.” Critical Review of International Social and Political Philosophy 22 (1): 103–118. Erman, Eva and Niklas Möller. 2023. “Does Normative Behaviourism Offer an Alternative Methodology in Political Theory?” Political Studies Review 21(3): 454-461. Floyd, Jonathan. 2017. Is Political Philosophy Impossible? Thoughts and Behaviour in Normative Political Theory. Cambridge: Cambridge University Press. 27 Floyd, Jonathan. 2020. “Normative behaviourism as a solution to four problems in realism and non-ideal theory.” Critical Review of International Social and Political Philosophy 23(2): 137–162 Floyd, Jonathan. 2022. “Can real actions justify realist principles? Normative behaviourism as a member of the realist family.” Critical Review of International Social and Political Philosophy (Online First): 1-20. Floyd, Jonathan. 2023. “Normative Behaviourism: A Reply.” Political Studies Review 21(3): 490-500. Galston, William. 2010. “Realism in political theory.” European Journal of Political Theory 9(4): 385–411. Galston, William. 2017. “Anger, humiliation, and political theory: Bringing the darker passions back.” In Matt Sleat ed. Politics Recovered: Realist Thought in Theory and Practice, 93–113. Columbia University Press. Geuss, Raymond. 2008. Philosophy and Real Politics. Princeton University Press. Hall, Edward. 2017. “How to do realistic political theory (and why you might want to).” European Journal of Political Theory 16(3): 283–303. Horton, John. 2010. “Realism, Liberal Moralism and a Political Theory of Modus Vivendi.” European Journal of Political Theory 9: 431–48. Jubb, Robert. 2019. “On what a distinctively political normativity is.” Political Studies Review 17(4): 360–369. Korsgaard, Christine. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. Lord, Errol and David Plunkett. 2017. ”Reasons Internalism.” In The Routledge Handbook of Metaethics, ed. T. McPherson and D. Plunkett. London: Taylor & Francis: 324339. McDowell, John. 1979. "Virtue and Reason." Monist 62(3): 331-350. Modood, Tariq. 2023. “Empirical and Normative Without a Universal Human Nature.” Political Studies Review 21(3): p. 462–468. 28 Newey, Glen. 2010. “Two Dogmas of Liberalism.” European Journal of Political Theory 9: 449–65. Rossi, Enzo. 2016. “Can realism move beyond a methodenstreit?” Political Theory 44(3): 410–420. Prinz, Janosch and Enzo Rossi. 2017. “Political realism as ideology critique.” Critical Review of International Social and Political Philosophy 20(3): 348–365. Schroeder, Mark. 2007. Slaves of The Passions. Oxford: Oxford University Press. Shklar, Judith. 1989. “The liberalism of fear.” In Nancy Rosenblum ed. Liberalism and the moral life. Harvard University Press, pp 21–38. Sleat, Matt. 2018. “Politics recovered: On the revival of realism in contemporary political theory.” In Matt Sleat ed. Politics Recovered: Realist Thought in Theory and Practice. Columbia University Press, 1–26. Smith, Michael. 1994. The Moral Problem. Oxford: Blackwell. Williams, Bernard. 2005. In the beginning was the deed: Realism and moralism in political argument. Princeton University Press.