A new moral sentimentalism

Abstract

This paper argues for a novel sentimentalist realist metaethical theory, according to which moral wrongness is analyzed in terms of the sentiments one has most reason to have. As opposed to standard sentimentalist views, the theory does not employ sentiments that are had in response to morally wrong action, but rather sentiments that antecedently dispose people to refrain from immoral behavior, specifically the sentiments of compassion and respect.

Keywords:

• Sentimentalism
• moral emotion
• compassion
• respect
• moral reasons

Acknowledgments

For helpful discussion and comments on earlier versions of this paper, I thank Mark Schroeder, Jonathan Dancy, Michael Slote, Justin D’Arms, and Daniel Jacobson; participants at the 2014 Pittsburgh Area Philosophy Colloquium, 2014 Georgetown University Philosophy Conference on Emotions and Emotionality, 2012 Pacific Division Meeting of the APA, and 2012 Fall Meeting of the Indiana Philosophical Association; and two anonymous reviewers and one anonymous editor for this journal.

Notes

1. Here, ‘sentiments’ refers to a host of attitudes that bear a family resemblance, such as anger, joy, guilt, shame, love, sadness, jealousy, fear, pride, disgust, respect, and compassion. Sentiments are (at least) mostly, but need not be entirely, coextensive with the category of emotions.

2. I assume for the purposes of this paper that there is indeed a property of moral wrongness.

3. My thoughts about elucidating the nature of a property have been influenced by the work of Ralph Wedgwood (2007) and Mark Schroeder (2005, 2007a, 2007b), who themselves draw on the work of Kit Fine, and in Schroeder’s case, Jeffery King (e.g. Fine [1994] and King [1998, 2002]). See these works for more detailed views on property analysis.

4. This is to be distinguished from conceptual or semantic analysis, in which the concept or meaning of the relevant term is analyzed.

5. This (rough but useful) understanding of property analysis will allow us to appreciate the full force of an important circularity problem for prominent sentimentalist views, which I discuss in the following section.

6. See e.g. Lewis (1989) and Brower (1993) for defenses of dispositionalism. McDowell (1985), Wiggins (1987), and Johnston (1989) defend versions of standard normative sentimentalism. Gibbard (1992) defends a similar view, although he treats normative language as noncognitive. Recent critique can be found in D’Arms and Jacobson (1994, 2000a, 2006).

7. In rough form, compassion can be understood as involving at least (1) some sort of emotional pain or distress directed at (what one believes to be, or perceives, or imagines as) the misfortune of another, and (2) an intrinsic desire to alleviate that misfortune. I shall not commit here to a full-fledged theory of compassion, nor do my arguments require any such commitment (for work on the nature of compassion, see e.g. Blum 1980; Nussbaum 2001; Cates 2003; Deigh 2004; Weber 2004; Cannon 2005; Crisp 2008).

8. Of course, acts of these kinds may very well be callous; but nothing implies that they must be. Theft, for example, may result in something less than significant harm to another, even if it does result in some harm. More generally, an act that results in a decrease in well-being need not result in significant harm, but may result merely in annoyance, inconvenience, discomfort, or dissatisfaction.

9. The type of respect relevant to morality and thus to my analysis of moral wrongness is a form of recognition respect, as opposed to appraisal respect (cf. Darwall 1977). Appraisal respect involves thinking highly of someone; recognition respect, on the other hand, involves due consideration and the disposition to refrain from behavior one would otherwise engage in. The type of respect relevant to morality involves due consideration of other people’s desires and interests, and the disposition to curtail selfish behavior. It is what goes under the famous label ‘respect for persons.’ It assigns to the desires and interests of others a baseline level of importance – when I have respect for someone, the desires and interests of that person matter to me, and do so because they matter (or should matter) to the person whose desires and interests they are.

10. Talk of degrees (or levels, amounts, etc.) of compassion and respect is simply a way of describing the fact that one can have more or less compassion or respect.

11. In Section 3.3, I explain how having a particular degree of compassion and respect can indeed rule out certain behavior.

12. Just what sort of ‘ought’ is at issue in PNS will be discussed in Section 3.4.

13. As I take it, the explanatory power of a theory T vis-à-vis an explanandum S is a matter of how probable S is given T. If T entails S then the probability of S given T is 1, and since that probability is maximally high the explanatory power of T vis-à-vis S will also be maximally high. But we should be sure not to conflate explanatory power and overall explanatory value – part of what makes an explanation a good one overall is its explanatory power, but other factors are relevant as well, e.g. the explanation’s simplicity.

14. In Section 3.7, I respond to an objection according to which compassion and respect are to be analyzed in moral terms after all.

15. Although describing wrong-kind reasons in this way is common, it may be only roughly accurate; see e.g. Hieronymi (2005) and Schroeder (2012).

16. It might be thought that the attitude of respect is not a sentiment, properly construed, because it lacks an affective dimension, and thus that PNS is not properly or fully a sentimentalist theory. In response, I would suggest that respect indeed has an affective component – a distinctive ‘feel’ – even if it is subtle; furthermore, an affect-requirement for being a sentiment is controversial. In any case, whether respect is properly classified as a sentiment is not important for my purposes – what matters is that the attitude of respect answers to reasons. If that criterion is satisfied, then I am happy to accept that mine is a hybrid theory that analyzes moral properties partially in terms of the sentiment of compassion and partially in terms of the non-sentimental attitude of respect.

17. I argue for this sort of view of moral reasons in Vogelstein (2011).

18. It is worth noting that some philosophers have attempted to identify some desire or set of desires that everyone necessarily has, and which is necessarily satisfied by refraining from immoral action (e.g. Korsgaard 1996a, 1996b; Schroeder 2007a). Evaluating these proposals would take us too far afield for present purposes; suffice it to say that they have not enjoyed wide acceptance.

19. Conway (2001) argues that the notion that one should not have compassion towards those who do wrong (or that such compassion should be mitigated) is a modern Western cultural bias and thus fails to support the contention that the norms intrinsic to compassion require anything other than universal compassion for all beings that are capable of suffering misfortune.

20. Indeed, it may simply be impossible to have all the sentiments that one has most right-kind reason to have to the degree to which one has most reason to have them; e.g. perhaps one cannot have both the degree of indignation one has most reason to have towards a wrongdoer and the degree of compassion one has most reason to have for him.

21. It might be thought that this response is inconsistent with the idea that (recognition) respect assigns to people’s desires and interests a baseline level of importance – a gloss which could suggest that the reasons for having respect must exist unconditionally and thus irrespective of whether the person themselves has a sufficient modicum of compassion or respect (I thank an anonymous reviewer for raising this issue). This thought, however, would be mistaken – although it represents one coherent view (and a view which, I should note, supports my previous response to the objection at issue), it is likewise coherent to suppose that the relevant baseline is not fixed, but might be determined by such things as the degree of compassion or respect that the person themselves has.

22. One might worry that a different circularity problem emerges from the plausible notion that reasons to have compassion and respect can be moral reasons. Specifically: given that we have reasons to refrain from doing what’s morally wrong, we have moral reasons to have compassion and respect, because having those sentiments will help us avoid doing what’s morally wrong; and if those are some of the right-kind reasons at issue in PNS, then, it could be argued, circularity looms. Note, however, that such reasons would not be right-kind reasons for having compassion and respect, i.e. reasons of fit – because they derive merely from the instrumental value of having those sentiments (albeit instrumentality vis-à-vis avoiding morally wrong behavior), as opposed to being reasons that speak to the fittingness of the sentiments themselves, they are wrong-kind reasons, and thus would not be included in PNS. I thank an anonymous editor for raising this issue.

23. One might object to the suggestion that there can be a precise degree to which one has compassion or respect. Of course, there is no unit of measure or precise method for measuring the degree or level of a person’s compassion or respect, or of any sentiment (although there might be in principle). But if compassion and respect indeed come in degrees, i.e. if a person can have more or less compassion or respect, then the idea that such sentiments come in precise degrees should not be problematic, even if it is limited in its application; indeed, it is difficult to see what the plausible alternative would be – if degrees of compassion and respect are not precise, then they are vague or indeterminate, and the claim that compassion and respect can only be had to indeterminate degrees of strength seems far more problematic than the claim that such degrees are precise.

24. These are rough characterizations. For detailed accounts of subjective reasons, see Vogelstein (2012), Whiting (2014), and Sylvan (2015).