Non-Conceptual Normative Pluralism and the Dualism of Practical Reason

Abstract

According to normative pluralists there are no truths about what one ought simpliciter to do, only truths about what one ought to do according to some normative system or standpoint. In contrast with conceptual normative pluralists who argue for this conclusion on the basis that the concept of an ought simpliciter is somehow defective, non-conceptual normative pluralists defend this conclusion on first-order grounds. Non-conceptual normative pluralism has recently received a book-length defence by Mathea Slåttholm Sagdahl. In this article I critique Sagdahl's defence of non-conceptual normative pluralism. Firstly, I challenge Sagdahl's attempt to motivate non-conceptual normative pluralism by appealing to Henry Sidgwick's work on the dualism of practical reason. Secondly, I criticise her response to the most prominent argument against non-conceptual normative pluralism in the philosophical literature, the argument from nominal-notable comparisons. Thirdly, I explore the possibility of a view similar to Sagdahl's position but which accepts the ought simpliciter.

1. Introduction

According to normative pluralism there are no truths about what one ought simpliciter to do, only truths about what one ought to do according to some normative system or standpoint such as morality or prudence (Brown 2023: 1–2; Case 2016: 2–3; Copp 2021: 417; Sagdahl 2022: 7 & 33). Normative pluralism divides into conceptual and non-conceptual varieties. Conceptual normative pluralism (‘CP’) holds that the concept of an ought simpliciter is confused, incoherent, or in some other way defective.¹ Non-conceptual normative pluralism (‘NCP’) shares a commitment to the claim that there are no truths about one ought simpliciter to do but denies that this is because the very concept of an ought simpliciter is defective. NCP has recently received an impressive book-length defence by Mathea Slåttholm Sagdahl (2022). Sagdahl (2022: 33) contends that concept of an ought simpliciter is ‘coherent but empty’.

In this article I assess several key elements of Sagdahl's defence of NCP, in particular, her positive case for the view which draws on Henry Sidgwick's (1874/1884/1907) work on the dualism of practical reason. I argue that Sagdahl does not provide a compelling case for NCP. I also argue that she does not provide a successful response to the most prominent objection to NCP in the philosophical literature, the argument from nominal-notable comparisons. Finally, I explore the possibility of a view similar to Sagdahl's position in many respects, but which accepts the ought simpliciter.

2. Sagdahl and Sidgwick

Sagdahl attempts to justify normative pluralism in a way that avoids an ‘unrestricted pluralism’, according to which any source of standards (e.g., medieval chivalry or imprudence) provides us with genuine reasons and oughts.² ‘Some standards’, Sagdahl contends, ‘are too bad or crazy to qualify as making genuine normative claims on us’ (2022: 50). Sagdahl's strategy for avoiding unrestricted pluralism is to reject what she calls the ‘top-down’ approach to normativity (a position characteristic of CP) which takes reasons and oughts to ‘flow from’ or be ‘issued by’ a source of standards (2022: 52–53). Rather, Sagdahl adopts a ‘bottom up’ approach according to which ‘rather than starting with principles, codes or standpoints and taking oughts and reasons to be issued by them, we should use reasons as our basic unit of analysis’ (2022: 53), where ‘being a reason is not a mere formal feature of some system but some substantive relation of what we truly ought to do’ (2022: 55).

As Sagdahl (2022: 56–57) recognises, her ‘bottom up’ approach to normativity does not obviously lend itself to normative pluralism. How do we arrive at normative pluralism on this approach to normativity? Sagdahl's answer is that ‘there are different sets of reasons, and these sets are mutually incommensurable. It is not possible to force these sets of reasons into a single unified picture of the way in which reasons count against or in favour of certain actions’ (2022: 57). By ‘incommensurable’ Sagdahl means such reasons are noncomparable; it is not possible to order them according to any comparative relation such as strength or weight (2022: 24–26).³

Before I explain Sagdahl's case for the conclusion there are different sets of noncomparable reasons, I want to emphasise that Sagdahl's claims about reasons appear perfectly compatible with the existence of ‘unqualified’, ‘plain’, or simpliciter reasons.⁴ After all, on Sagdahl's view, reasons are not fundamentally code or standpoint relative; reasons are not reasons-according-to-standpoint S (cf. Copp 2021: 417–18). Rather, moral and prudential reasons are just different kinds of a single type of thing, namely, what she labels ‘genuine normative reasons’ (2022: 165). Sagdahl (2022: 3) appears to think that genuine normative reasons can be classed as moral or prudential because, very roughly speaking, prudential reasons concern what would make your own life go well while moral reasons ‘also concern the welfare and good lives of others.’⁵

Sagdahl's defence of the thesis that there are distinct sets of noncomparable reasons draws on Sidgwick's work on the dualism of practical reason, which Sidgwick conceives of as concerning a conflict between utilitarianism and egoism. These theories, as developed by Sidgwick, tell us, respectively, that we ought to do what would bring about the greatest general or universal happiness (understood in terms of pleasure) and that we ought to do what would bring about our own greatest happiness. Sidgwick thought that we are justified accepting both theories, leading him to despair about the possibility of arriving at a consistent theory of practical reason.⁶ What Sagdahl finds attractive in Sidgwick's work is the idea that morality and prudence are two distinct sources of reasons which generate oughts that may conflict with each other and that there is no way to resolve this conflict (2022: 63–64). Sagdahl finds a way of defending these claims in Sidgwick's work that she uses to motivate the thesis that moral and prudential reasons are noncomparable (‘the noncomparability thesis’). This is the idea that we can occupy two different ‘points of view’ or ‘standpoints’ and from these two standpoints different actions are recommended or required. Sagdahl (2022: 64) approvingly quotes Sidgwick's remarks in the third edition of the Methods of Ethics (1884: 402) that the dualism of practical reason:

consists in the inevitable twofold conception of a human individual as a whole in himself, and a part of a larger whole. There is something different which is reasonable for him to desire when he considers himself as an independent unit, and something again which he must recognise as reasonably to be desired, when he takes the point of view of a larger whole.⁷

Importantly, Sagdahl (2022: 65–66) explains that she understands standpoints as allowing us to appreciate the normative significance of certain facts, stating that ‘it is not these perspectives themselves that are fundamental but rather the facts that underlie them’ (2022: 66).⁸

An important complication for Sagdahl's use of standpoints to defend the noncomparability thesis is that – at least as Sagdahl interprets the notion of a standpoint – we can adopt a variety of other standpoints, such as those of our family, our nation, and our species. Why don't facts about our membership of these groups – or these ‘identities’ as Sagdahl (2022: 70) also puts it – generate noncomparable reasons and oughts?⁹ What's so special about the personal point of view and the general (universal) point of view? Sagdahl's (2022: 71–72 & 75–76) answer to this question involves (tentatively) appealing to David Brink's (1997) work which explores a Kantian route for arriving at something like the dualism of practical reason. Brink argues that Kant may be committed to something like the dualism of practical reason in roughly the following way: Kant holds that moral requirements apply to us in so far as we are rational agents; they concern what we must will or value qua rational agent – i.e., qua creatures who are capable of deliberative self-governance (1997: 274–75). Consequently, such requirements are reason-giving (1997: 280–81). (I don't think that this conclusion follows but I'll ignore this issue for the moment.) But, as well as essentially being rational agents, we are essentially particular rational agents; we are rational agents numerically distinct from other rational agents. Brink (1997: 288) contends that:

Given that there are a plurality of purely rational agents, there must be requirements concerned with my own agency that apply to me just in so far as I am a particular rational agent, independently of my contingent interests and desires, just as Kant believes there are requirements of impartial concern that apply to me simply in so far as I am a rational agent.

Brink (1997: 288) labels these ‘requirements of categorical prudence’ and suggests that they are requirements to promote one's own rational agency.

The first point to note about Brink's Kantian argument¹⁰ is that it is not obvious that a requirement to promote one's own rational agency will correspond to what we intuitively think of as prudence. To see this, consider certain cases of suicide. Suppose I choose to end my own life because I am in chronic pain. This looks like an example of doing something which benefits me but frustrates my own agency. Secondly, and more significantly, Brink's argument relies on a controversial constitutivist approach to normativity – one that I find unconvincing for reasons explained by David Enoch (2006, 2011, 2021), among others. Thirdly, Brink does not really make a case for the idea that we must adopt the end of promoting our own rational agency separate from the end of promoting rational agency as such. Brink just asserts that there must be requirements of categorical prudence because we are particular rational agents, but he does not argue that we, qua particular rational agents, must value our own rational agency as such (as opposed to rational agency generally). Perhaps a case for this conclusion can be made. In fact, I think it is much easier to make a case for this conclusion than the conclusion we must value rational agency generally,¹¹ but Brink does not make it. In so far as Sagdahl attempts to privilege the personal standpoint and the universal standpoint using Brink's work, she inherits all the problems I have explained here. Drawing on Brink's Kantian argument also strikes me as overly committal, requiring that one buy into a wholesale Kantian approach to ethics. I raise this point in part because Sagdahl (2022: 68) worries that Sidgwick's own way of defending the noncomparability thesis is overly committed to a particular view of the contents of morality and prudence.¹²

Before I move on to consider Sagdahl's response to a prominent objection to normative pluralism, I want to discuss Derek Parfit's (2011: Ch. 6) reply to what he takes to be Sidgwick's argument for the noncomparability thesis. I do this because Parfit's discussion of Sidgwick involves a development of the notion of a point of view which is arguably distinct from Sagdahl's identity-based interpretation of the idea. I want to show that this alternative interpretation is no more promising for defending the noncomparability thesis.

Parfit suggests that Sidgwick's argument for noncomparability has the following form: to compare moral and prudential reasons we would need to be capable of adopting a third, neutral point of view separate from the personal of view and the impartial (universal) point of view, but there is no such point of view. Parfit's (2011: 135) reply is that ‘To be able to compare partial and impartial reasons, we don't need to have some third, neutral point of view. We can compare these two kinds of reasons from our actual, personal point of view’. Parfit (2011: 40) tells us that ‘We have an impartial point of view when we are considering possible events that would affect or involve people who are all strangers to us.’ Moreover, even when ‘our actual point of view is not impartial’ we can think about events ‘from an imagined impartial point of view’. I am broadly sympathetic with Parfit's response to Sidgwick, although I would put things differently: We can think about the reasons we would have for (e.g.,) choices in some situation when those choices would only affect strangers. Similarly, we can consider the same situation from a partial point of view where those affected by our choices are ourselves and people that we care about. (That is to say, we hold fixed all details of the situation which are not related to the identity of the people involved.) When we think about a situation in these two different ways, we may discover different reasons for choosing between various options. Compare, for instance, choosing between distributing some benefit between two agents A and B when they are strangers and when you are one of A and B. If you are, say, A, you have a reason to distribute the good to A given by the fact that doing so would benefit you that you do not have when A and B are strangers (Fletcher 2021: 21–22). This does not mean that those reasons which we appreciate when we adopt a partial point of view are not reasons when we are considering a situation from an impartial point of view or vice versa; these points of view do not create reasons. We can ask a question about the relative strength of those reasons which we appreciate from the impartial point of view and those we can appreciate from the partial point of view. I do not see how adding some third point of view is necessary or even helpful for answering this question, at least given the assumption that a point of view is just a device for helping us to appreciate reasons that we in fact have.

3. The argument from nominal-notable comparisons

Sagdahl (2022: Ch. 5) responds to what is probably the most prominent challenge to normative pluralism in the philosophical literature, namely, ‘the argument from nominal-notable comparisons’. This argument is focused on cases where you could do some significant moral good at a small prudential cost or vice versa. In such cases it seems that you ought simpliciter to choose the notable good (Crisp 2006: 132; Dorsey 2013: 119–20; Hills 2010: 31; Parfit 2011: 135–36). For example, consider (a slightly modified version of) Roger Crisp's (2006: 132) Two Doors case:

You are confronted by two doors. If you pass through door A, you will experience a minor twinge in your leg, but nothing further will happen. If you pass through door B, you will not experience the twinge, but some other person, a stranger and out of sight, will suffer an extremely painful electric shock.

In Two Doors it seems that the following is true: you prudentially ought to pass through door B, you morally ought to pass through door A, and you also ought simpliciter to pass through door A. There is a conflict between the recommendations of prudence and morality and the moral reason in favour of going through door A (that doing so will avoid another person suffering serious harm) outweighs the prudential reason for going through door B (that passing through door B would mean you avoid some discomfort).

Before I continue, I should note a feature of Sagdahl's discussion of nominal-notable cases: her discussion of such cases is framed in terms of conflicting moral and prudential requirements (or a moral or prudential requirement to φ and a prudential or moral permission to ψ) rather than conflicting moral and prudential oughts.¹³ Sagdahl (2022: 17–19) tells us that moral and prudential requirements, like moral and prudential oughts, are explained by moral and prudential reasons. However, she does not explain how she thinks of the relationship between moral or prudential oughts and moral or prudential requirements, although she does appear to commit herself to the view that moral and prudential requirements entail moral and prudential oughts.¹⁴ Sagdahl (2022: 116) contends that nominal-notable cases need to involve a nominal reason ‘sufficient to ground a normative requirement or permission’ because otherwise this nominal prudential or moral reason could be outweighed by some other prudential or moral reason. However, there is no danger of this if we operate with a nominal prudential or moral reason which grounds a prudential or moral ought. Despite this, for the sake of continuity with Sagdahl's own discussion, I will follow her in framing nominal-notable cases using requirements from this point onwards.

A second feature of Sagdahl's discussion of nominal-notable cases I need to register is that Sagdahl's discussion of such cases proceeds in terms of whether there is something that one ‘ought all things considered’ to do in such cases. Sagdahl (2022: 35) thinks that she can accept that there is an all things considered ought on a particular understanding of the idea, specifically, a ‘quantificational’ understanding of ‘ought all things considered’ according to which one ought all things considered to φ when all relevant normative standpoints agree that one ought to φ. In the context of the argument from nominal-notable comparisons, what is at issue between Sagdahl and her opponent is whether there are cases that show that when one of morality or prudence requires you to φ and the other requires (or permits) you to not φ, there is a fact about what you ought to do – where the relevant ought takes into account both morality and prudence but is authoritative over the verdicts of morality and prudence (cf. Sagdahl 2022: 116). I use ‘ought simpliciter’ to refer to this ought. This terminological choice allows me to avoid – as Sagdahl is forced to do – having to distinguish between different ways of understanding the notion of an all things considered ought. In her discussion of the argument from nominal-notable comparisons, Sagdahl (2022: 125) distinguishes the pluralist-friendly quantificational sense of ought all things considered from what ‘one ought all things considered to do, in the overarching normative sense’ (2022: 125).¹⁵

Sagdahl's (2022: 117–23) first response to the argument from nominal-notable comparisons – focused on the variant involving cases where prudence is the nominal good such as Two Doors – is to suggest that such cases, in so far as they in fact involve a conflict between morality and prudence, may not be possible. This is because it would be imprudent to pursue a nominal prudential benefit at a notable moral cost; choosing the nominal prudential benefit, and having the moral vices that would lead one to be the sort of person who would choose the nominal prudential benefit, are likely to lead one to have an overall prudentially worse life. In other words, prudence would not require or permit you take the nominal prudential good but rather would require you to take the moral option. Sagdahl (2022: 124–25) suggests that if both morality and prudence require you to φ then you ought all things considered to φ in her quantificational sense, allowing her to capture the verdict that there is something that you ought all things considered to do in the relevant cases (when properly interpreted to reveal that they do not involve a conflict between morality and prudence).

Sagdahl recognises that how convincing one finds her case for the conclusion that prudence would require one to take the notable moral good in the relevant cases will depend on one's views about the nature of prudence. Sagdahl (2022: 123) acknowledges that this conclusion looks less plausible if one accepts a view like hedonism. This is correct, but the point also extends to desire-satisfaction theories. I am sympathetic to hedonism or a desire-satisfaction theory. However, I cannot defend these views here.¹⁶ Moreover, Sagdahl herself does little to motivate the sorts of ‘more sophisticated’ (2022: 123) theories of prudence that she favours. Consequently, I will limit myself to observing that this is a point at which one's sympathy for Sagdahl's discussion will depend significantly on one's background commitments.

Sagdahl's second response to nominal (prudential)-notable (moral) cases is to offer a way to debunk the intuition that there is something that one ought simpliciter to do in such cases. Sagdahl (tentatively) suggests that cases where prudence and morality conflict for an agent in this way have features that should make us doubt this intuitive verdict. She (2022: 125–26) suggests that for such a case to arise – i.e., in order for a nominal (prudential)-notable (moral) case which genuinely involves a conflict between prudence and morality to be possible – the relevant agent would be psychologically unfamiliar to us because they have an ‘exceptional motivational structure’, and they would be in circumstances very different from those we are accustomed to. The relevant agent would be such that they:

(1) do not have the same interest in the concerns of ordinary human being, such as an interest in love, respect, and genuine relationships; (2) do not suffer from guilt; (3) do not have an interest in cultivating moral virtues; and (4) are in an exceptional circumstance where the risk of damaging sanctions are absolutely minimal.

To return to Crisp's Two Doors case described above, (1)–(4) are supposed to make us doubt the reliability of the judgement that the agent ought simpliciter to take the moral option – i.e., to go through door A (which results in a minor twinge in the agent's leg and no harm to anyone else). However, if we suppose the truth of (1)–(4) should make us doubt the reliability of the judgement that the agent ought simpliciter to take the moral option, the truth of (1)–(4) may also call into question the reliability of the judgement that the agent has sufficient reason to take the moral option in Two Doors. Sagdahl is committed to this latter verdict about the Two Doors case. Sagdahl (2022: 165–66) states that on her view we have sufficient reason to act in accordance with moral requirements and sufficient reason to act in accordance with prudential requirements. Both options are, as she also puts it (2022: 164), ‘rationally eligible’. We can stipulate that in the Two Doors case morality requires you go through door A while prudence requires you to go through door B.¹⁷ To illustrate this claim about the way Sagdahl's debunking response may threaten her own view, I will draw on Sagdahl's own discussion of how (1)–(4) threaten the reliability of judgements such as the judgement that the agent ought simpliciter to go through door A but apply her points to the judgement that the agent has sufficient reason to go through door A. To paraphrase Sagdahl (2022: 125–27), we need to recognise that, if (1)–(4) obtain, the judgement that the agent has sufficient reason to go through door A may be distorted by the following factors: (i) our values, which do not align with those of the imagined agent in the case, making it hard to put ourselves in their shoes; (ii) our inability to really imagine there being no risk to the agent of sanctions for going through door B; and (iii) moral outrage towards any agent who would go through door B. My point here is that (i)–(iii) could be used to try to debunk the intuition that the agent has sufficient reason to go through door A, not only the intuition that the agent ought simpliciter to go through door A.

Sagdahl (2022: 136–37) also discusses nominal-notable cases where prudence is the notable good requiring you to φ and morality is the nominal good requiring you to ψ. Sagdahl dismisses the possibility of such cases because she suggests that morality plausibly contains agent-centred prerogatives which allow an agent to prioritise her own well-being. Consequently, it will not be the case that the agent is morally required to choose the nominal moral good. However, it is plausible that morality is not always sensitive to personal cost. Consider the following case: Tom receives a great deal of pleasure from surreptitiously spying on Grace, deeply desires to spy on Grace, and spying on Grace is one of his central life projects which provides him with a sense of meaning and purpose. Tom's spying on Grace seems impermissible despite the prudential cost of ceasing and the impermissibility of his action does not seem to change if the cost to him of ceasing to spy on her increases. Cost does not always make a difference to the moral status of actions (Holden 2023: 68–69). If we can identify a case where cost does not make a moral difference, then we may be able to identify a nominal-notable case with the relevant structure.¹⁸

Sagdahl (2022: 138–39) considers a kind of nominal (moral)-notable (prudential) case which she does think may be possible. Suppose that morality permits you to take either the nominal moral good or the notable prudential good while prudence requires that you take the notable prudential good. Commenting on cases with this structure, Sagdahl (2022: 139) says that ‘an agent who cares both about morality's demands and about his or her own interests…need have no qualms about taking the prudential option and disregarding the nominal moral reason, as the agent is all things considered justified in doing so.’ I presume ‘justified’ here is being used as a synonym for ‘permitted’. As Sagdahl notes (2022: 139) she can employ a quantificational understanding of ‘all things permitted’ to say than an agent is all things permitted to take the prudentially required option in this case because both morality and prudence permit taking this option. However, her view cannot capture what I take to be the overwhelmingly plausible verdict that you ought to take the prudentially required option and, consequently, that you would make a serious mistake if you took the other option.

A powerful way of rejecting the noncomparability thesis related to the argument from nominal-notable comparisons (especially the variant considered in the previous paragraph) is Spencer Case's (2016: 13–14) ‘concurrence argument’. Suppose that one has three options: Option 1 is morally better than option 2 but prudentially worse, option 2 is prudentially better than option 1 but morally worse, and option 3 is just as morally good as option 1 and just as prudentially good as option 2. Case suggests that it is obvious that one ought to choose option 3 and that the ‘ought’ here is the ought simpliciter. Sagdahl's chief response to Case's concurrence argument¹⁹ draws on her first reply to nominal (prudential)-notable (moral) comparison arguments; she (2022: 140) suggests that if ‘morality and prudence are entangled’ in the way she thinks ‘such that there are always notable moral reasons to act prudently and notable prudential reasons to act morally’, then Case's concurrence scenario ‘may, after all, be impossible, as both standpoints would, in fact, uniquely favour [option] 3’. However, once again, one's sympathy with this response will depend on one's sympathy for Sagdahl's views about the correct theory of prudence.²⁰

4. Noncomparability, sufficient reason, and the ought Simpliciter

My final challenge to Sagdahl's case for NCP concerns Sagdahl's move from the noncomparability thesis to the conclusion that there are no truths about what agents ought simpliciter to do. Why not accept the noncomparability thesis but hold that what we ought simpliciter to do is either what prudence tells us we ought to do or what morality tells us we ought to do, but neither option individually? Buridan's donkey cases mean that everyone should agree that there are instances where one ought to (φ or ψ) but not φ or ψ individually (cf. Sagdahl 2022: 99). (The normative pluralist can read the ‘ought’ in the previous sentence as a qualified ought within a single normative standpoint like prudence.) I see no in principle barrier to making sense of this view which combines noncomparability and acceptance of the ought simpliciter on Sagdahl's picture of the normative domain, given the points I made in Section (2) about her view being consistent with the existence of unqualified or simpliciter reasons. Sagdahl (2022: 183) considers something in the vicinity of this view under the label ‘weak monism’ but frames it as holding that we have a set of options which are all plain (simpliciter) permitted. She (2022: 186) says that its ‘mysterious’ how we can go from qualified oughts to unqualified permissions. I do not find it mysterious that we can go from qualified oughts to an unqualified disjunctive ought if qualified oughts are ultimately explained by unqualified reasons.

This view I am proposing has the advantage of cohering well with the plausible claim that there are cases in which, when morality (or prudence) does not recommend any available option, one ought simpliciter to φ because of the prudential (or moral) reasons in favour of φ-ing. For example, take a case where there are no moral reasons concerning which option you should take but prudence favours one of the options more than the others.²¹ Sagdahl thinks that she can capture the judgement that you ought all things considered to φ in such cases by using her ‘quantificational’ understanding of the all things considered ought (introduced in Section (3) above). I am not convinced that this is correct because this is not a case where ‘standpoints are in agreement’ (2022: 35); morality does not ‘agree’ that you ought to φ, rather it is silent on this issue.

What is the relationship between the view I am suggesting here and the position that one has sufficient reason to do what you morally ought to do or what you prudentially ought to do? (A view which is attributed to Sidgwick by some philosophers.)²² I think that it is plausible that if one has sufficient reason to φ and sufficient reason to ψ, one ought simpliciter to (φ or ψ). Sagdahl might object that the ought simpliciter must resolve conflicts between morality and prudence (cf. Sagdahl 2022: 143 & 216). However, it is plausible that what is essential to the idea of an ought simpliciter is that it is such that you would be making a mistake if you fail to comply with it, that if you ought simpliciter to φ it is advisable to φ, and that it is the ought which features in deliberation about what to do.²³ All of this can be true of an ought which does not settle conflicts between morality and prudence.

5. Conclusion

To conclude, I have argued that Sagdahl has not offered a convincing case for NCP. However, I highly recommend her insightful and rigorous discussion to anyone interested in these issues.