![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
Miriam Schoenfield argues that moral realism and moral vagueness imply ontic vagueness. In particular, she argues that neither shifty nor rigid semantic accounts of vagueness can provide a satisfactory explanation of moral vagueness for moral realists. This paper constitutes a response. I argue that Schoenfield's argument against the shifty semantic account presupposes that moral indeterminacies can, in fact, be resolved determinately by crunching through linguistic data. I provide different reasons for rejecting this assumption. Furthermore, I argue that Schoenfield's rejection of the rigid semantic account is based on a presupposition that ultimately implies the very same claim that is under dispute: the vagueness of moral predicates in imperfect languages persists in the perfect language, as well.
Acknowledgements
I am grateful to Hannes Leitgeb for discussion and comments and to the anonymous referee for very helpful comments. This work was supported by the H2020 Marie Skłodowska-Curie Actions [grant agreement number 675415].
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 This is called tolerance intuition, which traces back to (Wright Citation1975, Citation1976). In general, three main features are associated to phenomena of vagueness: borderline cases, blurry boundaries and tolerance. See Keefe (Citation2000) and Williamson (Citation1994) for discussion.
2 See Shafer-Landau (Citation1995) for earlier discussion on moral realism and vagueness. Also, see Dougherty (Citation2017) and Schoenfield (Citation2016, 262–3) for more examples of vagueness in moral issues.
3 ‘Moral vagueness’ simply refers to the vagueness of moral terms, in particular moral predicates.
4 See Lewis (Citation1982) and Williamson (Citation1994) for semantic and epistemic accounts of vagueness, respectively.
5 However, as we shall see later, my defense of the semantic view ultimately undermines Schoenfield's main objection to Williamson's epistemic account. Thus, at least a Williamsonian epistemicism can consistently explain moral vagueness for moral realists.
6 See Dougherty (Citation2014, 352–72 and 358).
7 See Barnes (Citation2014, 339–62 and 339). Also, see Abasnezhad and Jenkins (Citation2018) for a recent discussion of Barnes’ theory of ontic vagueness and Abasnezhad and Hosseini (Citation2014) for an alternative theory of ontic vagueness.
8 See Fine (Citation1975), Lewis (Citation1986) and Keefe (Citation2000).
9 Here ‘P’ stands for a vague moral predicate, which has some clear cases of positive/negative application and some borderline cases for which it is indeterminate whether P. We can think that ‘P’ refers to ‘is morally permissible’. Furthermore, ‘application of predicate’ means its linguistic usage here.
10 Informally, according to the shifty view, the truth value of ‘John is tall’ depends to the way ‘tall’ is used in a community, and hence the truth value may change from one community to the other. According to the rigid view, however, the truth value does not change from one community to the other.
11 For instance, whether an abortion at 19 weeks is morally permissible.
12 More precisely, Schoenfield thinks that it is unacceptable for moral realists to say that Cheryl can find out, for instance, whether some abortion, for which she was initially uncertain, is in fact determinately permissible by collecting linguistic data; Schoenfield (Citation2016, 266).
13 Schoenfield (Citation2016, 266).
14 This seems to be a reasonable requirement for being a competent speaker. If someone feels conflicted about whether a person with the height of 195 cm is tall (or a person with the height 160 cm is non-tall) in our normal community, then she does not seem to be a competent English speaker. Alternatively, one could appeal to a theory that suggests most people are mistaken about moral facts while competent about the usual practice of how moral terms are used. However, appealing to such a theory will undermine the idea that collecting linguistic data are helpful in resolving moral indeterminacies. I am grateful to the referee for pointing this out.
15 Indeed, gathering any kind of data does not seem to help settle borderline cases. If John is a borderline case of tallness, even finding out that John's height is 177.65746 cm does not help to settle whether John is tall.
16 Otherwise, if most people choose, for instance, ‘yes’, then it shows that rules governing the application of P in this community is determined about whether a is P and it is Cheryl who is not competent enough regarding how the language works.
17 In fact, deciding on whether a is P based on the linguistic data, itself, seems to be subject to vagueness. In particular, it is indeterminate what the minimum amount of margin between two options should be, in order to suffice determinately choosing one option over the other.
18 In fact, if the problem of borderline cases could be settled in this way, vagueness would not be such a challenging issue.
19 Generally speaking, the linguistic data regarding a predicate does not always confirm linguistic usage of the predicate. While the linguistic data is the result of surveys at a certain point in time, the linguistic usage of the predicate is about the general and unwritten rules of application of the predicate over a long period.
20 We now know that Pluto is not actually a planet.
21 For detailed discussion on indeterminacy of meaning, see Davidson (Citation2001), in particular ‘The Inscrutability of Reference’ and ‘On the Very Idea of a Conceptual Scheme’.
22 I should emphasize that, here, I only aimed to show that Schoenfield's argument against the shifty semantics is unsuccessful. Providing a comprehensive shifty semantic account of moral vagueness that can explain moral deliberation for moral realists is another matter and subject to another paper. I am grateful to the referee for pressing on this.
23 As mentioned earlier, rejecting Schoenfield's argument against shifty semantic account results her argument against epistemic account of moral vagueness to fail too. In particular, she argues that Williamson's epistemicism cannot explain moral vagueness for moral realists since it would imply commitment to shifty semantic account. Now that it has been demonstrated that it is not a problem for moral realists to accept the shifty semantics, it should not be a problem to accept Williamson's epistemicism, either.
24 See Lewis (Citation1983).
25 See Wedgwood (Citation2001).
26 See Schoenfield (Citation2016, 270–1) for details.
27 Following Schoenfield, I shift the focus from ‘permissible’ to ‘better than’, simply because Wedgwood uses ‘better than’ in developing CRS.
28 See Wedgwood (Citation2001, 18).
29 Having a correct preference means having a preference that is permitted by the goals of practical reason.
30 Here, I agree with Schoenfield, so I will not go through of the details of the argument. See Schoenfield (Citation2016, 272–3).
31 Schoenfield does not consider a third option, where the intermediacy in vague moral terms is the result of both the indeterminacy in which inference rules of practical reason constitute the conceptual role for ‘better than’ and also the indeterminacy about which relations make the inference rules of practical reason valid. In particular, it could be the case that among rules r1, … , rn, it is indeterminate which one constitutes the conceptual role for ‘better than’. Also for each ri, there are relation si 1, … , si m such that it is indeterminate which one makes ri valid.
32 Schoenfield (Citation2016, 273).
33 This is because according to CRS, the moral properties/relations are identical to the properties/relations that make certain rules of practical reason valid.
34 Note that the second conjunct is simply reiteration of A, given moral realism and CRS.
35 It is easier to see the falsity of B, given the semantic account of vagueness, when reiteration of A in B is removed:
B') If it is indeterminate which properties and relations make the inference rules for practical reason valid, then the perfect language will contain predicates that lack precise application conditions.
36 While it is the case that amputating a person's arm to save a billion lives is better than not doing the amputation and letting a billion people die and that amputating a person's arm to save another's life is not better than not doing the amputation and letting the person die, it is indeterminate whether amputating a person's arm to save hundred lives is better than not doing the amputation and letting hundred people die.
37 Here, R is the inference from ‘it is better for me to do x than y at t’ to forming the preference for doing x over y at t.
38 By ‘slight change’, I mean a change from ‘indeterminacy’ to positive (or negative) classification. Alternatively, classifying a positive case as a negative case seems to be a major change.
39 S1 … Sn agree on all positive and negative cases of ‘better than’ and only disagree on its borderline cases.
40 Such explanation makes the antecedent of B true and its consequent false.