Political Normativity… All-Things-Considered

Abstract

The idea of a distinctively political normativity came under sustained fire lately. Here I formulate, test, and reject a moderate and promising way of conceiving it. According to this conception, political normativity is akin to the kind of normativity at play in all-things-considered judgments, i.e., those judgments that weight together all the relevant reasons to determine what practical rationality as such requires to do. I argue that even when we try to conceive political normativity in this all-things-considered way, and even when we do not concede from the get-go that moral reasons necessarily trump or override normative reasons of a different kind, political normativity is still reducible to morality, because the peculiar content of all-things-considered political oughts can be explained by the interplay of general moral principles and contextual facts that do not obtain exclusively in political scenarios. If my arguments are correct, I provide political realists with one more reason to withdraw from the metaethical battle over the idea of a distinctively political normativity and show that the moralist approach is defensible against a prima facie promising, but ultimately untenable, alternative.

1 Introduction

Since Williams (2005) insisted on the need to emancipate political theory from the status of applied moral philosophy, several scholars attempted to describe what a genuinely political way of doing political theory may look like. In doing so, some of these political realists – as they have come to be known - converged toward the contention that there is a distinctive political normativity.

This idea came under sustained fire lately. In a recent and influential paper, for instance, Leader Maynard and Worsnip (2018) reconstructed and criticized several possible profiles this idea could take, and none of them seem to survive scrutiny.

Realists’ replies to this and other attacks - such as those of Erman and Möller (2021, 2022) for instance - took various forms. Some of them attempted to recast realist political theory as different from the strictly analytical procedures common among contemporary political philosophers, advocating a more discursive approach (Jubb 2019). Others tried to retain some grip over the notion of political normativity, dropping the claim that such a kind of normativity is sui generis (i.e., on a par with, say, prudential or epistemic normativity), while defending its distinctiveness with respect to morality.

In the latter group, two positions securing some form of distinctiveness for political normativity proved particularly popular. The first operates by excluding morality from the notion of political normativity altogether. The second allows morality into political normativity, but aims at securing distinctiveness by claiming that the latter remains nonetheless irreducible to the former (Rossi and Sleat 2014).

The first position is quite radical but mostly methodological. It holds that realists can and should tap only into non-moral normative resources to formulate critical normative judgments about politics. On this view, political normativity is understood as encompassing only epistemic considerations - as radical realists would have it (Aytac and Rossi 2023) - or perhaps a mixture of epistemic and instrumental considerations, as others have suggested (Burelli and Destri 2022). Either way, morality allegedly has no place in political normativity. I will not consider this position here.

The second position is more moderate and has significant meta-normative implications that I intend to explore. It admits morality as one source (among many) of political normativity, yet holds that the specific circumstances, concerns, pressures, and constraints of political action make the latter irreducible to the former. Philp’s (2010) dismissals of liberal theory as a benchmark for political action because of its inattentiveness to the practicalities of political choices, Newey’s (2010) claim that moral considerations do not always trump considerations of other kinds, and more recent contentions from other political realists are in line with this moderate position - as Sleat (2022) recently pointed out.

In this paper, I will test a reasonable and apparently promising hypothesis about how to flesh out this second, moderate position into a determinate conception of political normativity. The hypothesis consists of two thesis:

• 1. All-things-considered thesis: Political normativity is akin to the kind of normativity at play in all-things-considered judgments – those judgments through which, after having weighed different kinds of normative considerations, one finally determines what one ought to do not from a moral point of view, a prudential point of view, or the point of view of any specific normative domain, but what one ought to do simpliciter, full stop. On this conception, what one ought to do politically is simply what one ought to do all-things-considered (hereafter ATC).

• 2. No moral rationalism thesis: In politics, what one ought to do ATC is not necessarily what one morally ought to do or is morally allowed to do. In other words, moral considerations are relevant inputs to political deliberation, but they do not automatically determine its outputs.

The reason for the twofold nature of this hypothesis is that the All-things-considered thesis alone is not enough to substantiate a distinctively realist conception of political normativity. Indeed, many moralists embrace moral rationalism, according to which there is no contradiction between the verdicts of morality and those of ATC rationality: if you morally ought to ψ, then you ATC ought to ψ, and the fact that you are ATC justified in ψ-ing just means that you are also morally justified in ψ-ing¹. Without the No moral rationalism thesis, a moralist could embrace the All-things considered thesis and appeal to moral rationalism to reduce political normativity, so conceived, to morality. Hence the twofold nature of the hypothesis.

There are three reasons supporting the plausibility of the hypothesis above and making it worthy of assessment. First, although popular, moral rationalism is not universally accepted. Some philosophers reject it, and all its many variants face important objections (see Gert 2013 for a review)². Second, even if moral rationalism is true, it would need to be true a priori in order to make the hypothesis implausible from the start, for only then it would be impossible for anyone to ever rationally discard moral verdicts. And it is highly debatable that moral rationalism can be true in this sense. Sure, there are a priori arguments in favor of it (e.g., Smith 1993). But I agree with Dorsey (2016, p. 41–65) and others (e.g., Stroud 1998) both in seeing arguments of this sort as dangerously circular and in taking the validity of moral rationalism as a matter of substantive, first-order normative inquiry. Finally, even conceding that moral rationalism might be true as a matter of first-order normative inquiry for private moral agents, it might still be false for political ones, i.e., for agents acting in a political capacity as public officials, citizens, and so on. Perhaps the peculiar concerns, constraints and circumstances of politics really make moral considerations more easily outweighed by other considerations, as Sleat and other moderate realists seem to suggest. Assessing whether this is the case is precisely what I am after here.

Unfortunately for the moderate realists, I will show that this is not the case. Indeed, in what follows I claim that the hypothesis turns out false because the No moral rationalism thesis is ultimately implausible, and I do this by showing that moral rationalism seems to hold even in political hard cases that apparently put a strain on it. This means that political normativity, even when understood in this moderate, ATC fashion, is most likely still reducible to morality. To show this, I will proceed as follows. In Sect. 2 I clear the ground by elucidating the conceptual toolkit I will rely on. In Sect. 3 I spell out the hypothesis more extensively. In Sect. 4, I present my strategy to test it and discuss the negative results.

2 Metatheoretical Premises

The first notion in need of elucidation is that of normative considerations. I shall use this term interchangeably with that of normative reasons, i.e., considerations that count in favor of an option (an action, an attitude, etc.). Quite uncontroversially, I shall assume that reasons vary, besides in their kind (more on this in a second), also in their weight. The weightier a reason is, the stronger its support for the relevant option.

The second notion is that of deliberation, which I understand as specular to that of justification. I take deliberation as the assessment process by which one weighs all the relevant normative considerations at play, thereby reaching a verdict or judgment about what to do. Deliberation thus represents a particular kind of decision, namely one supported by the balance of reasons or, in other words, a justified one. Mutatis mutandis, the term justification denotes here the reasons-based defense of a deliberative output, which succeeds only as far as it manages to make the relevant negative attitudes toward the deliberating agent inappropriate³. I shall assume that a full justification establishes that a deliberative output like, say, an action, is at least permissible (after all, required and supererogatory actions are also permissible ones). But I shall not take any more specific position on the issue of how the balance of reasons can affect the specific deontic statuses of deliberative outputs. The issue is hotly debated in metaethics, and I want my argument to hold for as many positions as possible. As far as I can see, nothing major in my argument hangs on these underlying metaethical issues, anyway⁴.

The final notion to clarify is that of a normative domain, normative standpoint, or kind of normativity (I shall use these labels interchangeably): the Prudential, the Moral, and the Epistemic are obvious examples. There are two main ways of telling normative domains apart. One approach focuses on the content of the relevant normative considerations, i.e., on the differences ‘in what they enjoin agents to do, which agents and situations they apply to, and which considerations they treat as relevant’ (Leader Maynard and Worsnip 2018, p. 761-2). What makes a consideration a moral, a prudential, or an aesthetic one, on a content-based approach? Here I submit, following Dorsey (2016, p. 37), that each normative domain has its content determined by paradigmatic concerns and values in virtue of which the relevant considerations hold. Self-interest, for instance, is the paradigmatic concern of the Prudential; Truth that of the Epistemic; Beauty, style and so on are among those of the Aesthetic. In the case of the Moral, I take everyone’s interests as the paradigmatic concern, and I understand both the notion of ‘everyone’ and that of ‘interests’ broadly. The range of subjects whose interests matter morally may be extended to include the agent herself, non-human animals, and even inanimate objects such as the environment, nature in general, and so on. As for the notion of interests, I understand it as encompassing not only one’s welfare and well-being but also one’s autonomy, capabilities, etc.⁵

Another way to distinguish normative domains is to focus on the sort of reactions triggered by violations of domain-specific normative considerations. On this psychological approach, distinguishing between different normative domains requires looking at the ‘particular attitudes and emotions each kind of normative judgment is constitutively associated with’ (Leader Maynard and Worsnip 2018, p. 762). Morality, for instance, is constitutively related to what Stephen Darwall has called second-personal reactive attitudes like blame, guilt, and resentment (Darwall 2006, 2013): these are precisely the reactions which moral justification protect from. Beyond morality, there is a bit of controversy over the reactions constitutively associated to each normative domain. Some think that there are epistemic and aesthetic forms of blame, while others (like me) are more inclined to think that actions prudentially, epistemically, or aesthetically unjustified all give rise to third-personal reactions, i.e., various forms of contempt and disesteem⁶. In my view, when someone acts imprudently, fails to believe what one ought to believe, or makes an aesthetically poor choice, the appropriate reactions have much more in common with one another than they have with moral blame and resentment: they lack the distinctive ‘sting’ of moral blame and its distinctive connection to accountability (Hieronimy 2004, p. 116). At any rate, nothing major - for my argument’s purposes - hangs on how we call and categorize these domain-specific reactions. All that matters is that we are generally able to tell them apart.

A few words now on the notion of all-things-considered deliberation and justification. The remarks above should suffice to give the reader a rough idea of how deliberation and justification work within a normative domain. Agents seeking a morally, aesthetically, epistemically, or prudentially justified course of action select the relevant considerations within that domain, weigh them against one another, and find out what is the verdict from that point of view. However, in our daily lives, virtually any choice we are called to make requires weighing together reasons of different kinds, striking a balance between them to figure out what we rationally ought to do all-things-considered.

ATC rationality is thus not a normative domain among others, on a par with prudence, morality, etc., but rather an overarching one. In other words, both the character and the content of all-things-considered rationality depends, as it were, on the things considered in any given case. From the psychological point of view, I contend that when an agent behaves in an ATC unjustified way, the appropriate reactive attitudes he faces are those pertaining to the kind of normative considerations that should have come out on top from an ATC point of view. Correspondingly, I assume that when an agent behaves in an ATC justified way, all the domain-specific reactive attitudes pertaining to the reasons at stake are thereby made inappropriate. Consider the following example. Imagine that in a cold winter night a friend of yours shows up with a horrible but very warm jacket – the only appropriate jacket he has for such weather conditions. Assuming prudential reasons outweigh aesthetic ones in cases like this, and that no other relevant considerations apply, your friend is behaving rationally or, in other ways, in an ATC justified way. On my view, this entails that the aesthetic reactive attitudes that her outfit of choice would otherwise give rise to (be them a form of second-personal blame or third-personal disesteem) are thereby unfitting. Similarly, the content of what, in each situation, is ATC justified depends on what kind of consideration, among the ones at play, is weightier. In the case above, the content of the final ATC ‘ought’ is prudential, and that ought is grounded in self-interest.

Is there really such a thing as an ATC normative verdict? A few philosophers argued there is no such thing by questioning our capacity to combine different normative inputs in a non-arbitrary way (Tiffany 2007), by claiming that attempts to explain the supreme authority of the ATC point of view involve contradictions (Copp 1997), or unpalatable normative commitments (Baker 2018). I lack the space to address these skeptical concerns here⁷. But for the purposes of this paper, the following contention should do. Notwithstanding the objective difficulty of weighing different normative considerations together and of explaining how one can do so, everyone should have first-hand experience with cases in which a deliberation resulted justified in an ATC sense. Easy cases in which reasons pertaining to different normative domains supported incompatible options, but in which the losses internal to one domain vastly outweigh the gains in another, immediately come to mind. At any rate, there is no need for ATC judgments to be easy and widespread to avoid skepticism. It suffices that they are possible, and I think it is relatively safe to assume they are.

Let me now explain how to assemble the ideas elucidated so far into a testable hypothesis about political normativity.

3 The Hypothesis

Aristotle notoriously categorized politics as an ‘architectonic’ science:

‘As there are many actions, arts, and sciences, their ends too are many; the end of the medical art is health, that of ship-building a vessel, that of strategy victory […]. But where such arts fall under a single capacity – as bridle-making […] falls under the art of riding, and this and every military action under strategy […] – […] the ends of the architectonic arts are to be preferred to the subordinate ends’ (Aristotle 2012, 1094a6-18).

Politics, for Aristotle, is the architectonic science par excellence, not only because its mastering requires commanding many other sciences, but also because its end is an all-important one. The political philosopher is thus ‘the architect of the end, with a view to which we call one thing bad and another good without qualification’ (Aristotle 2012, 1152b1-3, emphasis mine).

The hypothesis presented before follows Aristotle in an important respect, but parts ways with him in another. It follows Aristotle in insisting that politics is an architectonic and overarching normative domain, in that it requires to mediate between competing considerations and pressures and to aim for what is good, or rational, ‘without qualification’. It distances itself from Aristotle by rejecting his specific conception of the end of politics as the prosperity of one’s polity, as this would entail – and purely by means of presupposition – an unduly narrow conception of political normativity⁸. The prosperity of one’s polity is certainly a factor to consider in political deliberation, but more universalistic moral considerations about – for instance – the prosperity of other polities also plausibly matter, to some degree. Therefore, I refuse to assume that national interests necessarily override other normative considerations in ATC deliberations. I shall concede though that, in certain cases, they may.

My hypothesis can be further refined by nesting it in the realists’ claim – shared by Philp (2010) and Bellamy (2010) among others – according to which the difference between moral and political normativity is a substantive one, to be found in the different content of the normative judgments applying to private morality, on the one hand, and politics, on the other (Leader Maynard and Worsnip 2018, p. 777). For Philp, the difference is that politics is more consequentialist than private morality: in ticking-bombs scenarios, a politician might be required to get his hands dirty to save lives, whereas a private moral agent might not. And the same may apply to other sorts of prima facie immoral actions, like lying. According to my hypothesis, this sort of differences may indeed obtain, and its two constituent theses offer a rough account of why they may.

Leader Maynard and Worsnip consider and reject this general realist claim, but an examination of their main objection illuminates both how my hypothesis can be expressed in its terms – i.e., in terms of content - and how it can bypass such an objection.

Their objection highlights an alleged weakness in the content-based approach to distinguishing normative domains: on their view, a content-based approach leads to a virtually infinite multiplication of normative kinds. Sure - they argue - agents in different roles such as doctors, journalists, parents and even politicians ought to do and enjoin to do things that private moral agents ought to abstain from. In this limited sense, we might speak of medical normativity, parental normativity, and so on. ‘But […] only an idiosyncratic construal of ‘morality’ or ‘ethics’ would exclude such domains on these grounds […].’ (2018, p.761).

For this reason, they propose to rely on the psychological approach for distinguishing normative domains. On such an approach, they argue, the differences in content between politics and private morality can be subsumed under the umbrella of morality. A politician might ought to do or enjoin to do things that a private moral agent ought to abstain from, but as long as it is fitting to morally blame or resent the politician for his failure to comply with the relevant principles, those principles still count as moral. For instance, if we blame and resent the politician for not lying in certain circumstances, then we are treating him as morally obligated to do so. If we do not blame or resent him for that, then we are taking him as morally entitled to lie. In both cases, the relevant principles applying are moral in character.

The objection is successful against Philp’s and perhaps other ways of expressing the difference between moral and political normativity in terms of content. But there is one that I think can avoid it. As I mentioned in Sect. 2, instead of focusing on the actions one enjoins to do under a given principle of a certain sort, we can see the content of each normative domain as determined by paradigmatic concerns and values in virtue of which domain-specific considerations hold. And we saw in that section how easy it is to identify standard normative domains in such a way. The Prudential, for instance, emerges as a normative domain primarily concerned with self-interest; The Epistemic as primarily concerned with truth; The Moral as concerned with everyone’s interest (as broadly defined previously), and so on.

Were the realist to take up this axiological version of the content-based approach, the proliferation of normative domains lamented by Maynard Leader and Worsnip would be limited, and the realist would find himself in the position of making an important concession without surrendering to their objection. On such a view, indeed, the realist may concede that the principles and obligations applying to political agents are moral in character. But this concession need not be a capitulation, for the realist may still maintain that the content of such principles and obligations may not be necessarily moral (i.e., not grounded on everyone’s interest).

The realist may do so by advancing two specific claims, tightly related to the two theses my hypothesis consists of. First, and as for the All-things-considered thesis, the realist can contend that political agents are under an obligation to act in an ATC justified way and that this obligation is moral in character. This means that, when political agents fail to act in an ATC justified way, not only they become appropriate targets for the reactive attitudes specific to the kind of normative considerations they failed to appreciate, but also of moral blame and resentment. On this view, it is political agents’ moral duty to deliberate in an ATC perspective. Second, and as for the no moral rationalism thesis, the realist may then suggest that, in contrast with private morality, political judgments answering the question ‘what is to be done ATC?’ might be non-moral in content, and hence that in political contexts moral justification and ATC justification may come apart. For instance, in political scenarios normative considerations with a non-moral content (such as prudential considerations based on self-interest) might trump considerations with a moral content (based on everyone’s interest), and thus come out on top of ATC deliberation.

When we join these two claims, we are presented with the possibility that a political agent might have an obligation, moral in character, with a potentially non-moral content. Such a political agent might thus be ATC justified in doing something immoral. Is this hypothesis plausible? Let me now proceed to test whether this is the case, starting from some considerations about to how to do so.

4 Testing the Hypothesis

In the specific sense seen above, political normativity might be akin to ATC normativity. But kinship relations come in degrees, and where to put the threshold of distinctiveness is far from obvious. Matt Sleat, in defending the moderate account of political normativity my hypothesis substantiates, has the same concern. ‘We might still want to ask just how different political thought needs to be in order to be meaningfully distinctive as political thought if full autonomy is not the aim. But then it is not clear that there is going to be any independent threshold that needs to be met and against which we can assess such distinctiveness’ (Sleat 2022, p. 472).

A virtue of my hypothesis is that it poses a quite clear scalar benchmark against which to assess the distinctiveness of its underlying conception of political normativity. Remember: the hypothesis is that, for political agents but not for private moral agents, the content of ATC deliberative outputs might not necessarily be moral, even when considerations with a moral content are among the relevant inputs. Therefore, the more kinds of normative considerations with a non-moral content can trump those with a moral one in political cases, the more successful the hypothesis proves to be.

The general argumentative strategy I shall follow to test the hypothesis and evaluate its performance is the following.

• 1) The first step is to identify a few paradigmatic, non-moral normative domains besides the Moral. Although perhaps more domains might be pointed out starting from the content-based approach, I shall limit my attention to the only prima facie ‘full-blooded’ ones I can think of, namely the Aesthetic, the Prudential, and the Instrumental (for reasons I shall provide in Sect. 4.3, instrumental normativity can be treated jointly with epistemic normativity). We all have, I think, an intuitive grasp of their characteristic concerns and values, and thus of their content⁹.

• 2) Then, I will do my best to present pairs of analogous cases - one involving a private moral agent and another involving a political agent - in which there appear to be significant differences in the content of the ATC normative requirements these agents respectively face¹⁰. I will try to do so for each of the non-moral domains identified above. More specifically, I will present pairs of analogous cases in which it appears that for political agents, but not private moral agents:

  • i) considerations with a non-moral content (e.g., aesthetic considerations grounded in beauty, prudential considerations grounded in self-interest, etc.) retain a normative priority over moral ones, and in which:

  • ii) political agents may be legitimately resented and blamed for failing to realize the priority of non-moral considerations and thus to act in an ATC justified way (remember: the hypothesis holds that the final ATC ought for political agents might be non-moral in content, but it concedes that political agents’ duties to act in an ATC justified way is moral in character).

As mentioned before, the more kinds of normative considerations with a non-moral content can trump over those with a moral content in political cases, the more prima facie successful the hypothesis will turn out to be.

• 3) Finally, I will double-check the reliability of these results. I will control their validity by assessing whether they can be explained away without presupposing the sort of asymmetry between morality and politics embedded in the hypothesis. In other words, the task here is to check whether there are plausible and more parsimonious explanations for the apparent differences in the content of ATC oughts of political and private moral agents. One way to fulfill this reductive explanatory task is by appealing to:

  • i) standard general principles with a moral content, i.e., principles plausibly grounded in the protection of everyone’s interests, and.

  • ii) empirical contextual facts affecting the application of such principles.

The more prima facie differences can be explained in this or other reductive ways, the less compelling the results will be, and the less plausible the hypothesis becomes. Indeed, if all such apparent differences can be redescribed as ultimately having a moral content, the hypothesis will be completely falsified, for its second constitutive thesis will turn out to be false: moral rationalism would remain plausible even in political contexts, even when assessed through a content-based approach.

I shall start by examining the role of aesthetic considerations (4.1), then proceed to prudential one (4.2), and finish by examining epistemic and instrumental considerations (4.3).

4.1 Aesthetic Normativity

One may dismiss aesthetic considerations as irrelevant for deliberation as such (be it political or moral) on the ground that such considerations are primarily reasons for attitudes: they offer reasons to like or appreciate, but not much else. However, an increasing number of scholars agree that aesthetic considerations can provide reasons for actions too¹¹. There seems to be, for instance, aesthetic considerations pertaining to artistic integrity to play all the notes in a sonata, to pay some form of respect to Shakespeare’s text when adapting his plays, to preserve works of art, and so on.

Generally speaking, we do not (and should not) care much about political agents’ responsiveness to aesthetic considerations. Although their elegance or artistic sensitivity might add to their charm and political success, these are normally irrelevant traits for their value as political agents. There are, however, prima facie plausible cases supporting the idea that political agents may be blamed and resented for failing to respond to at least some specific aesthetic considerations. Consider, for instance, the following case of aesthetic preservation:

Marco: Marco is the newly elected mayor of Beautiville. He won by promising to fight the housing crisis (a crisis that the city administration so far failed to mitigate) and to support a variety of social housing projects. However, it turns out, there is no more available lots within the municipality borders. Also, the laws of the State make it impossible to expropriate and repurpose private buildings. The construction of social housing projects in Beautiville thus requires demolishing buildings owned by the municipality itself. But all such buildings are beautiful and meaningful parts of the city architectonical heritage. Marco eventually decides to tear down some of them to deliver on its electoral promise and to correct for the administration long-standing neglect of the social needs of the community.

Let’s assume, as it seems prima facie plausible to do, that considerations with an aesthetic content (i.e., considerations concerning the beauty and meaningfulness of the buildings to be demolished) are here weightier than considerations with a moral content (i.e., considerations concerning the plight of those suffering from the housing crisis, the expectations of Marco promisees, and so on). Sure, Marco not only made a promise; he may also have a reparative duty to fulfill qua institutional representative of the administration, namely a duty to relieve the pain of those that suffered from the housing crisis that the municipality allowed for so long¹². However, Marco might nonetheless be legitimately blamed or resented for doing what morality demands here, namely, to clear the ground for new social housing projects. The aesthetic considerations in favor of preserving the old buildings, grounded on their beauty and historical relevance, might be strong enough to justify the blame. So it seems to be the case that Marco, as a political agent, failed to fulfill an obligation (moral in character) to act in a way that is ATC justified. And it further seems to be the case that the content of such obligation is aesthetic rather than moral.

However, the possibility of cases like this one is insufficient to lend credibility to the hypothesis, for two reasons. First, because there appear to be no asymmetry here between political and private moral agents. It is indeed easy to construct cases identical to the one above involving private moral agents, as consideration of aesthetic preservation are relevant for them in just the same way. Take, for instance, the following one.

Luca: Luca is a private art collector who promised to foster female literacy in Eastern Africa by devolving 1 million dollars to an NGO, but he runs out of funds before fulfilling his promise. The only way for him to raise that money is by selling some of the masterpieces he owns to Mahmood, the son of a crazy sheik who will burn them just to flex his lavish lifestyle in front of his Instagram followers. After reflection, Luca decides to sell part of his collection to Mahmood.

On the apparently plausible assumption that aesthetic considerations are here weightier than moral ones, Luca seems to be just as blameworthy as Marco. Therefore, the political context in which the latter deliberates does not appear to make any significant difference.

Second, and more instructively, it is debatable whether consideration of aesthetic preservation are purely aesthetic in content. Indeed, it seems that considerations pertaining to the preservation of beautiful and historically significant artifacts are also inherently moral from a content-based perspective, meaning that they can hardly retain their weight without presupposing a reference to the interests and well-being of others (as opposed to simple considerations about beauty and other aesthetic qualities).

One may resist this point by drawing attention to cases of apparently ‘pure’ aesthetic obligations, such as the following one proposed by Eaton (2008). Suppose that, at some point in the past, a layer of soot has been laid over a masterpiece and that a mediocre artist painted one of his minor works on top of it. You may suppose, as Eaton suggests, that there is an aesthetic obligation to restore the canvass so to bring the masterpiece back to light. And you may further think that this obligation is purely aesthetic in content, in the sense that it seems to depend only on the superior aesthetic qualities of the masterpiece. However, as Matheson and Milam have noted, on reflection it is quite hard to retain this intuition, for the obligation to restore seems here to be grounded more in the importance of making the masterpiece potentially accessible to viewers (a consideration that is moral in content, since it presupposes a reference to the interests and well-being of others), than in its superior aesthetic quality per se. Notice, indeed, that regardless of whether the restoration is carried out or not, the masterpiece would exist anyway (Matheson and Milam 2021). So, what else besides the legitimate interests of other human beings (a paradigmatically moral concern) could ever ground the obligation to bring it back to light?

In synthesis, from a content-based approach, aesthetic normative considerations in general can hardly feature among those that a political agent might be morally blameworthy for discounting in an ATC perspective. Some aesthetic considerations – i.e., considerations of aesthetic preservation – seem to be an exception, but they do not establish any asymmetry between private moral agents and political ones, and it is debatable whether their content is actually aesthetic, for they seem ultimately grounded in the legitimate interests of others – a paradigmatic moral concern.

4.2 Prudential Normativity

Let me now move to the domain of prudential normativity, understood as primarily concerned with self-interest. As in the case of aesthetic normativity, it is dubious that political agents might ever be blamed or resented for failing to respond properly to all prudential considerations. Even from a purely strategic perspective on politics - a perspective totally unconcerned with the normative status of the goals that are politically pursued - it is hard to deny that a good political agent is an agent that effectively pursues such goals even when doing so entails risks and prudential costs for herself. A disposition to sacrifice seems an undeniable characteristic of great political leaders of all times, and probably of good citizens too.

However, there is a peculiar and derivative kind of prudential considerations that a political agent might be, at least apparently, legitimately blamed for being insufficiently responsive to. I am referring to considerations pertaining to what would be prudentially good for the political community within and for which a political agents operates. Leader Maynard and Worsnip consider and reject a similar position in their paper, but they frame it in terms of collective instrumental normativity:

‘One might wonder […] whether political normativity can be construed as a special kind of collective instrumental normativity. On such a view, there are certain ends that we have collectively as a political community, and political normativity concerns the realization of these ends. It is surely right that we can hypothetically talk about what would be rational from the point of view of advancing ends that a collective shares. However, in order for these collective ends to have any normative authority over the actions of political actors […] there must be a claim that such actors are obligated to take into account the ends of the community as a whole […] And that claim takes us beyond the domain of instrumental normativity’ (Leader Maynard and Worsnip 2018, p. 780).

Leader Maynard’s and Worsnip’s contention against this view is, ultimately, that it is parasitic. Instrumental rationality is only formally normative: it gives us reasons to do certain things, but only as far as we have full-blooded, authoritative reasons to pursue the ends those things are in service of. And what else could give us any full-blooded reasons to pursue the given ends of a political community, if not morality?

However, one may dodge this objection by reframing the position, as I suggest, in terms of collective prudential oughts. On this construal, just as there are ends and goals that would be prudentially good for a single agent to pursue, so there might be ends and goals that would be, objectively, in the interest of a given polity to achieve (call this the polity’s ‘national interests’¹³). As the prudential interests of a single agent can conflict with the interests of other individuals, so the national interests of a given polity can clash with the interests and needs of other political communities. And under the plausible assumption (which I will question in a second) that morality is in good part other-regarding and egalitarian in structure, it is not unplausible to describe the second conflict, too, as one between prudential and moral considerations, thereby avoiding the criticism. It may not be from morality, but from a derivative and collective sort of prudence, that the instrumental considerations pertaining to the achievement of national interests inherit their normative authority.

Recall that I am not trying to defend this position to claim that it offers, in and for itself, a suitable view of political normativity. The hypothesis I am examining here is wider and does not hold national interests as the supreme or the only normative compass of political ATC deliberation. It simply grants that such interests are relevant inputs of such deliberation, and that they may be relevant enough to outweigh or trump the moral interests of other individuals and polities in certain circumstances.

The contraposition of the following two cases suggests that this might be case – and hence that political agents, in contrast with private moral agents, might fall under an obligation, moral in character, to prioritize prudential considerations:

Laura: Laura is a not-so-junior academic and an activist for gender equality. Her window of opportunity for getting a tenured position is rapidly closing. One day, a few hours before she is supposed to give the final speech at a rally against the pay-gap at her university, which she promised to give, she gets a last-minute invitation to an interview for an assistant professor position – one of the last she’s likely to get, as she knows. But the interview can only take place at the time of her speech. Reluctantly, she decides to forgo the interview and to join his fellow campaigners.

Bjorn: Bjorn is the president of Prudentialand, a northern European country not much exposed to the threat of climate change. He must decide whether to ratify an ambitious, international climate action agreement. The agreement promises to significantly slow down the rising sea levels, thus avoiding the flooding of lowlands in third-world countries and the displacement of roughly 50 million people, opening new opportunities for even more decisive international action in the future. However, the agreement will also impact Prudentialand’s economy, as 50.000 Prudentials will lose their jobs. Reluctantly, Bjorn decides to ratify the agreement.

In both cases, the relevant agent deliberates in favor of the considerations with a moral content, discounting the prudential ones. But there seems to be a significant asymmetry between Laura and Bjorn. Bjorn might be the appropriate target of his fellow citizens’ blame and resentment, a fact that might be interpreted as the result of him having an obligation with a moral character but a prudential content, namely an obligation to foster the interests of his polity. On the other hand, it is hard to fathom how Laura might blame or resent herself for her choice, or feeling guilty for it, and thus how she might have a similar obligation. If this is indeed the case, the hypothesis would be corroborated.

But there are several ways to explain away this apparent difference and to reconfigure both cases as internal to the moral domain. One of these is to focus on Laura’s case and to appeal to moral duties toward oneself. On such a view, morality sometimes requires agents to foster their own good, for instance – as Kant has it - by developing their talents (Kant 2011 [1797], p. 440-6). Plausibly, Laura has a moral obligation to foster her talents by getting the position and perhaps, given the stakes of the case, this obligation may trump her promissory one to give the speech¹⁴. If this was the case, she should feel guilty and blame herself for her deliberation. Hence, there would be no substantial asymmetry between private moral agents and political ones.

A second way, inspired by Scheffler, works by posing agent-centered moral prerogatives (Scheffler 1994). On such a view, morality always permits moral agents to grant some extra weight to those considerations that retain a particular value from their own point of view, allowing them to override the requirement of bringing about the best possible outcome (when morality is understood in consequentialist terms) and, perhaps, to even derogate to promissory duties (when morality is understood in deontological terms). Starting from this view, one may claim that Laura acted in a morally admirable way, but that she went beyond what others could reasonably expect from her. We could explain the residual disesteem she might feel for herself, notwithstanding her morally worthy decision and the praise coming with it, in the following terms: she did something morally supererogatory when she could have done (morally speaking) something for herself. Following the same line of reasoning, one may posit similar prerogatives in the political case and account for the blame Bjorn may be targeted with by appealing to the fact that, in contrast with Laura, the cost of Bjorn’s supererogatory choice does not fall only on himself, but on others too. Hence, his fellow citizens might reasonably blame or resent him, in pretty much the same way as a private moral agent (say, a wife) could resent someone (say, her husband) for deciding to donate some of her own money to charity.

The third way to explain away the disanalogy without presupposing significant differences in the content of private morality vis a vis political normativity is by questioning the premises that frame the pursual of national interests as prudential and thus non-moral in content. One may indeed object to this construal by claiming that it only makes sense on the background of a peculiar conception of the content of morality, namely a strictly impartial and egalitarian one. Among moral philosophers, those that take morality to be an impersonal and strictly impartial business do tend to frame the pursual of one’s own projects and the extra weight typically granted, in deliberation, to the interests and well-being of one’s inner circle, as non-moral or extra moral¹⁵. But many other authors think, on various grounds, that the very content of morality itself requires or at least allows some degree of partiality.

A similar theoretical contrast is visible among political philosophers. Some of them – typically, cosmopolitans of some sorts – deny the existence of special moral obligations toward local forms of political organizations and reject the moral significance of political or cultural belongings. And they do so on the assumption that all human beings are (or can and should be) treated as citizens of a single community. Many others, however, see these allegiances as part and parcel of morality and as having primal moral significance. Rawls for instance - and many of his followers - notoriously thought that citizens have significant duties of justice toward one another in virtue of the basic social structure they share and the expectations of reciprocity their interactions – on the background of such structure - give rise to. On a Rawlsian view, we do not owe much to foreigners, and certainly not as much as we owe to our fellow citizens (Rawls 1999). Bjorn’s case, on such a view, does not represent a contrast between consideration with a prudential and moral content, respectively, but a contrast between two competing considerations, both with a moral content¹⁶.

4.3 Epistemic and Instrumental Normativity

So, the blame and resentment raised by political agents’ failures to respond properly to certain aesthetic and prudential considerations does not lend much support to the hypothesis. In each of these cases, we can offer plausible and more parsimonious explanations for these reactions by appealing only to considerations with a moral content. The availability of these reductive moral explanations deprives the no moral rationalism thesis of much of its appeal.

But what about epistemic and instrumental normative considerations? For the purposes of this paper, we can treat them jointly. After all, efficacy is partly, if not entirely, an epistemic matter: it requires knowing many truths about the situations in which one is called to operate and how one’s actions are likely to play out in practice. Perhaps, these considerations stand a better chance to validate the hypothesis, at least in part.

In examining this case, it is important to be clear about what the hypothesis does and does not entail. Leader Maynard and Worsnip offer a useful remark to this end:

‘In claiming that political normativity is a kind of moral normativity, moralists are not arguing that either good political decision-making or good political theories involve only moral theorizing. Most obviously, a failure to grasp relevant empirical information or to understand how one’s decision will play out in practice will likely lead to bad political decision making. But that is true of moral decision-making generally: someone with good moral instincts can make bad moral decisions for such reasons’ (2018: 765).

I agree. Clearly, lack of responsiveness to epistemic and instrumental considerations can result in bad decision-making, be it moral or political. The hypothesis suggests something different, however, namely that political agents might be blameworthy for such unresponsiveness in a way or to a degree that private moral agents are not. In other terms, the hypothesis suggests that whereas a private moral agent’s factual ignorance about relevant empirical information and the foreseeable consequence of her action is generally excusing or justifying, a political agent may not deflect blame and resentment on this ground. ‘I did not know’ or ‘I did not foresee that’ may be potential excuses or justifications for the former, but not for the latter.

To be sure, factual ignorance does not always protect private moral agents from blame and resentment either. Sometimes such an ignorance – which we can understand as the result of an insufficient responsiveness to epistemic and instrumental considerations – is itself culpable, and thus blameworthy. Drawing on Strawsonian ‘quality of will’ accounts, an agent is blameworthy for an action performed out of ignorance if such an ignorance is itself the result of an improper regard or insufficient concern for another person’s significant interests (Shoemaker 2013). For instance, you may not know that a sociopathic baker has poisoned the slice of cake you are handing me, and thus be excused for contributing to my murder. You had no reasons to check. But the doctor that gives only a superficial look at my clinical history and that kills me by prescribing a medication I am allergic to is not off the hook in virtue of his ignorance. As his ignorance was the product of an unduly superficiality in performing his job, his ignorance is morally faulty and the doctor blameworthy for his action¹⁷.

But factual ignorance can nonetheless be generally excusing for private moral agents, especially when they deal with hard cases in which doing the right thing requires complex reasoning and the handling of many complicated pieces of information. Although the degree to which one is responsible for her own epistemic incompetence is an object of debate, it seems safe to maintain that, in cases of this sort, an epistemically naïve, unprepared, or simply stupid private moral agent can normally refer to such incompetence to deflect or mitigate blame. If her ignorance is the result of her own limited capacities but her intentions were good, her wrongdoing is not the product of a lack of goodwill.

For political agents, however, this does not seem to be the case. Compare the following two cases:

Marty: Marty is trying to figure out whether he should keep eating industrially farmed meat or not. He takes up this deliberative task at the best of his capacities and in good faith. He is not biased in favor of meat-eating and he is not deliberately insulating himself from relevant sources of information: every time he finds articles, social media posts and opinion pieces on the subject, he tries to metabolize their content. However, because of his limited competence, education, and intellectual capacities, he fails to be persuaded by the existing evidence that animals are sentient beings capable of experiencing emotional and physical pain. Because of this epistemic failure, he fails to realize that humans ought not to impose unnecessary suffering on animals and that he thus ought to abstain from consuming factory-farmed meat.

Liz: Liz is the newly elected prime minister of Toryland and she is about to pass into law her ambitious tax reform. Along with her minister of finances, she worked on the reform to the best of her capacities. She has quite libertarian allegiances, but she is not unduly biased toward any given economic doctrine. However, she is not persuaded, or paused, by the arguments put forward by a group of economists claiming that her plan will negatively impact employment and private savings. Neither she nor her economic advisors are experienced, capable, or savvy enough to appreciate the evidence these economists bring to support their predictions. She thus takes advantage of her solid majority, pass the reform into law, and the skeptics’ predictions, unfortunately, prove correct.

There is a notable difference between these two cases. In Marty’s case, moral blame and resentment do not strike as appropriate reactions. If we embrace the quality of the will account of blameworthiness, it seems that his good will – and the actions expressing it, e.g., his getting informed on the issue – are enough to protect him from these reactive attitudes. If he is not responsible for his own epistemic shortcomings and cognitive limitations, he is plausibly excused for his action. To the contrary, in Liz’s case, blame and resentment does not seem as out of place. In her case, the epistemic failure and the limitations it is caused by do not seem to excuse her – at least, not to a comparable degree.

Both cases involve insufficient responsiveness to epistemic considerations (i.e., insufficient responsiveness to evidence), but the point holds for instrumental considerations too. As Leader Maynard and Worsnip argue, it is true that non-moral skills are needed to achieve moral ends in the private sphere, just as they are necessary to achieve ends in the political one – in other terms, responsiveness to non-moral considerations, such as instrumental ones, is necessary to both morally and politically good actions. However, whereas a private moral agent can fail to be effective in the pursual of a good end – even bringing about dramatically bad consequences in attempting to bringing it about – and still be morally justified or excused if his will and intentions are good, when it comes to politics this sort of failures does not seem to protect the agent from blame and resentment. In this respect, efficacy – a value pertaining to the domain of instrumental rationality – seems more intimately related to the concept of good political agent than it is to the concept of a good moral agent. A good will, in politics, may not matter much.

It is now time to ask whether we can explain away this apparent difference in content between private morality and political normativity. Is it due to distinctive features of politics, to some distinctively political contextual facts? Or we can account for it simply by appealing to general moral principles plus some contextual facts that happen to obtain even outside of politics?

If the former was the correct answer, we would have a way of grounding the claim that the content of political normativity is irreducibly different from standard morality, as it would include instrumental and epistemic considerations in its core. But I do not think this is the case. In fact, I believe there is a very plausible reductive moral explanation for such a difference, one that does not appeal to distinctive features of political situations, but only to general moral principles and to specific facts and circumstances characterizing several contexts beyond political ones. These facts are the following ones.

• 1) In political contexts, deliberations have high stakes - in facts, the stakes get higher as one climbs up the political ladder (e.g., from city-level to national-level).

• 2) In political contexts, the relevant knowledge for deliberation is vast and complex (again, the vaster and the more complex the higher the political level).

• 3) In political contexts, deliberation is collectively binding, and its effects reverberates even on individuals that did not agree with its result (the exemption being international treaties).

The gist of the putative, reductive moral explanation is that in situations with very high stakes, whose handling depend on the understanding of vast and complex bodies of knowledge, and the outcome of which will be binding even for those that will not agree with how the situation will be handled, those freely taking up a position in the decision-making process inevitably generate legitimate expectations about their fittingness for such a position, thereby incurring – willingly or not – in a moral obligation ‘to deliver’, so to speak. Statements of ignorance such as “I did not know” or “I could not make it work”, in such situations, do not typically justify nor excuse. And when ignorance does excuse, in these contexts, the standards for its non-culpability are set much higher than they normally are in private situations¹⁸.

Notice that none of the above-mentioned facts is specific to politics as such. Many agents in non-political contexts face deliberations with high stakes, in which the mastery of vast and complex bodies of knowledge is necessary, and in which results are going to be binding even for those that disagree. Take, as an example, CEOs and other top managers in the private sector. Some of their decisions have very high stakes, as they can affect the livelihood of employees and investors, and certainly require a rather firm grasp on several bodies of knowledge about markets, economic trends, and so on. Moreover, the outcome of such decisions affects employees and investors regardless of whether they agree with the managers’ deliberation or not. When such agents display an insufficient responsiveness to normative considerations that are non-moral in content, but rather epistemic and instrumental, they are, too, appropriate target for blame and resentment. Politics, once again, does not seem to make much of a difference.

5 Conclusion

In conclusion, the hypothesis does not come out of the testing in good shape. All the political cases in which considerations with an apparently non-moral content claim a superiority over considerations with a moral one can be recast as cases of conflict between competing moral considerations, and explained away by way of reference to general moral principles and to contextual facts that are recurring in political scenarios, but not distinctively political in any relevant sense. In other words, we can account for the specific content of political principles without leaving the moral domain and without presupposing that politics, as a domain of human action, is peculiar in any interesting sense. And this shows that the no moral rationalism thesis, on which the distinctiveness and plausibility of the hypothesis depend, has not much going for it.

Beyond this, there is a valuable lesson to learn from the testing of this hypothesis, which concerns the peculiarity of the Moral as a normative domain. Morality, as we saw, has quite a tendency to encroach on other normative domains, making it hard to steer clear from it as realists would like to do. Aesthetic, prudential, and epistemic considerations can quite easily gain significant moral salience, and this is not as true the other way around. While prudential verdicts regarding self-interest, aesthetic verdicts concerning beauty, and epistemic verdicts grounded on truth do not take moral considerations as relevant, moral verdicts do take considerations regarding such things as relevant, at least in all the occasions (undoubtedly, very frequent) in which humans can be said to have an interest in their underlying concerns.