Conceptual role semantics for moral terms

Philosophical Review

Volumn 110, Issue 1, 2001, Pages 1-30

  Wedgwood, Ralph ^a  

^a MASSACHUSETTS INSTITUTE OF TECHNOLOGY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 34248742643     PISSN: 00318108     EISSN: 15581470     Source Type: Journal    
DOI: 10.1215/00318108-110-1-1     Document Type: Review

References (68)

1. For noncognitivism, see especially Simon Blackburn, Essays in Quasi-Realism (New York: Oxford University Press, 1993), essays 6-11.
2. and Allan Gibbard, Wise Choices, Apt Feelings (Oxford: Oxford University Press, Clarendon Press, 1990). I use the term 'statement' as a neutral term for any speech act that is performed by an apparently sincere utterance of an indicative sentence. For cognitivists, moral statements are assertions or purported statements of fact; for noncognitivists, moral statements are speech acts of some other kind, such as expressions of emotion or prescriptions.
3. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977), 15.
4. For an example of reductive naturalism, see Frank Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford: Oxford University Press, Clarendon Press, 1998), chap. 6. Such a reductive account of moral properties could be correct even if it is not in any sense an a priori or conceptual truth, just as 'Heat = molecular motion' may give a correct reductive account of heat even though it is not in any sense an a priori or conceptual truth. The reduction of facts about heat to facts about molecular motion (and so also of thermodynamics to the statistical mechanics of molecular motion) is often regarded as the paradigm case of a successful reduction;
5. see Ansgar Beckermann, "Supervenience, Emergence, and Reduction," in Emergence or Reduction? ed. Ansgar Beckermann, Hans Flohr, andjaegwon Kim (Berlin: de Gruyter, 1992), 94-118.
6. The view that moral properties are nonnatural properties is especially associated with G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903). Another important metaphysical issue is the extent to which moral facts and properties are essentially dependent on our subjective responses to them;
7. for a discussion of this issue, see my "The Essence of Response-Dependence," European Review of Philosophy 3 (1998): 31-54.
8. Noncognitivists will be committed to accepting this if they also accept strong supervenience, S5 modal logic, and the idea that properties are closed under Boolean operations, as I explain in 'The Price of Non-Reductive Moral Realism," Ethical Theory and Moral Practice 2 (1999): 199-215.
9. We can also put the point in terms of "moral facts." If it is the function of moral statements to state truths or facts, then the meaning of a moral statement must determine some condition such that the statement is true just in case that condition obtains. Mackie's error theory must claim that such conditions never actually obtain, while reductive naturalism claims that such conditions can be specified in wholly nonmoral naturalistic terms.
10. Such metaphysical accounts of the nature of a moral property need not follow immediately from such an account of the meaning of the relevant moral term. They may follow only from the conjunction of this semantic account with certain further truths (whether empirical or a priori) that do not directly follow from the semantic account alone. Still, as I have argued, the relevant notion of a "moral property" must be interpreted as the notion of "the property determined by the meaning of the relevant moral term as the referent of the term in question." So, the metaphysical account of the nature of this property must be based, at least in part, on the semantic account of the meaning of the relevant term.
11. See Paul Ziff, Semantic Analysis (Ithaca: Cornell University Press, 1960), chap. 6.
12. and Judith Thomson, "The Right and the Good," fournal of Philosophy 94 (1997): 273-98.
13. In effect, an account of what it is for a term to mean cow may employ both the term 'meaning' and the term 'cow', but it may not itself employ the phrase 'means cow' or any variant thereof. That is, I aim to conform to a constraint analogous to Christopher Peacocke's constraint, within his theory of concepts, that an adequate account of a concept should take "the A(C) form";
14. see his A Study of Concepts (Cambridge: MIT Press, 1992), 5-16.
15. Compare Saul Kripke's insistence that a theory of reference should not be "circular," in Naming and Necessity (Cambridge: Harvard University Press, 1980), 68-70.
16. For an argument against noncognitivism, see my "Non-Cognitivism, Truth and Logic," Philosophical Studies86 (1997): 73-91.
17. For example, Gilbert Harman regards conceptual role semantics as a rival to truth-conditional semantics;
18. see his "(Non-Solipsistic) Conceptual Role Semantics," in his Reasoning, Meaning and Mind (Oxford: Oxford University Press, Clarendon Press, 1999). Hartry Field, on the other hand, has advocated a "two-factor view" according to which a term's meaning involves both its conceptual role and its contribution to truth conditions as two mutually independent elements;
19. see his "Logic, Meaning and Conceptual Role," Journal of Philosophy 69 (1977): 379-408. In contrast to these approaches, I shall propose that it is precisely the conceptual role of moral terms that fixes the reference or semantic value of those terms;
20. compare Ned Block, "Functional Role and Truth Conditions," Proceedings of the Aristotelian Society, supp. vol. 61 (1987): 157-81.
21. This account of the logical constants owes much to Christopher Peacocke;
22. see especially his "Understanding Logical Constants: A Realist's Account," Proceedings of the British Academy 63 (1987): 153-99. In his work on the logical constants, Peacocke aims to show how the insights of conceptual role semanticists like Gilbert Harman, and intuitionists like Dag Prawitz, can be combined with a realist truth-conditional semantics;
23. see Harman, 'The Meaning of the Logical Constants," in Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, ed. Ernest Lepore (Oxford: Blackwell, 1986), 125-34
24. and Prawitz, "Proofs and the Meaning and Completeness of the Logical Constants," in Essays on Mathematical and Philosophical Logic, ed. Jaako Hintikka, Ilkka Niiniluoto, and Esa Saarinen (Dordrecht: Reidel, 1979), 25-40. I have noted the main points where I diverge from Peacocke in notes 17 and 18 below.
25. To "reject" a sentence is to disbelieve the proposition that the sentence expresses-that is, to be in the state that would be expressed by denying the proposition in question. More importantly, for our purposes, it is correct to "reject" a sentence if and only if the sentence is not true. Thus, acceptance and rejection are opposites, in the sense that it cannot be correct simultaneously to accept and reject any sentence.
26. Some philosophers believe that it is impossible for someone to violate the rules of disjunction-introduction without failing (at least on that occasion) to understand the term 'or'. But the claim that this is impossible is a very bold claim. Frege's claim, that it would be a "hitherto unknown type of madness," seems more apt;
27. see The Basic Laws of Arithmetic, trans. Montgomery Furth (Berkeley: University of California Press, 1964), 14.
28. This focus on the basic rules governing the use of a term parallels Harman's idea that a concept's content-determining conceptual role is determined by what the concept "immediately implies" or "immediately excludes" ("The Meaning of the Logical Constants," 130-32).
29. and Peacocke's idea that possession of a concept requires finding certain patterns of reasoning "primitively compelling" ("Understanding Logical Constants," 154-56).
30. It may hold quite generally, as Michael Dummett has argued, that a term's meaning precisely consists of whatever speakers must grasp to have a full understanding of the term;
31. see Dummett, "What is a Theory of Meaning? (I)," in his The Seas of Language (Oxford: Oxford University Press, Clarendon Press, 1993), esp. 3-4. But I do not need to rely on this general Dummettian claim here. All that I need is that a full understanding of logical constants and moral terms consists in some sort of grasp of whatever determines their meaning.
32. I strongly suspect that the notion of "mastering a rule" can only be explained in partly normative terms. For this reason, I do not accept Peacocke's claim that the possession conditions for concepts should be "non- normative" (A Study of Concepts, 139);
33. on this point, I agree with Alan Millar's "Critical Notice" of Peacocke's A Study of Concepts, Mind 103 (1994): 73-82.
34. Robert Brandom offers an explicitly normative version of conceptual role semantics, in his Making It Explicit (Cambridge: Harvard University Press, 1994). However, my approach is not committed to Brandom's claim that normative notions are prior to intentional notions (nor to his claim that the notion of inference is prior to the notion of representation, nor to his radical holism about concepts).
35. This idea simplifies and generalizes some of the points that Peacocke makes about the existential quantifier, in "Understanding Logical Constants," 160-62.
36. This semantic value can then justify all the other rules of inference (such as the elimination rules) that apply to the use of the term. Since the term's semantic value is determined by the "basic rules" that apply to the term, it is plausible that all the other rules applying to the use of A are justified, at least in part, on the basis of these basic rules.
37. This general idea has been suggested before (though not, so far as I know, developed in detail). For example, Harman suggested that "what makes something the concept of danger is in part the way in which the concept is involved in thoughts that affect action in certain ways" ("(Non-Solipsistic) Conceptual Role Semantics," 219).
38. Compare also Brandom, Making It Explicit, 245-49.
39. Davidson, Essays on Actions and Events (Oxford: Oxford University Press, Clarendon Press, 1980), 40, 99. Note that this is not how Davidson himself uses the phrase 'all things considered better than': for him, it signals a certain sort of conditional or prima facie judgment (39).
40. Compare Judith Thomson, "The Right and the Good," 275-76.
41. Compare the idea that Philippa Foot considers, that there is not only a "subscripted 'ought'" (like 'ought from a moral point of view', or "the 'ought' operating within the system of etiquette or some system of institutional rules"), but also a "free unsubscripted 'ought'";
42. see her Virtues and Vices (Oxford: Blackwell, 1978), 169-70. Foot is skeptical of this idea;
43. the idea is defended in Gavin Lawrence, 'The Rationality of Morality," in Virtues and Reasons, ed. Rosalind Hursthouse, Gavin Lawrence, and Warren Quinn (Oxford: Oxford University Press, Clarendon Press, 1995), 89-147, esp. 118.
44. Derek Parfit uses the term 'rational' in this way, especially in Reasons and Persons (Oxford: Oxford University Press, 1984). (As noted above, I prefer to reserve the term 'irrational' for what may in a broad sense be called "internal incoherence.") In more recent work, Parfit prefers the term 'what there is most reason to do';
45. see his "Reasons and Motivation," Proceedings of the Aristotelian Society, supp. vol. 71 (1997): 98-130.
46. Gibbard uses the term 'advisable' (Wise Choices, Apt Feelings, 18-19, 89).
47. Strictly speaking, I will be using 'x 'y' and 't'as schematic letters that can take the place of either object-language terms or their metalinguistic equivalents.
48. Although it is irrational for anyone who understands the term B to violate this rule, such irrationality is obviously not a "hitherto unknown type of madness" (see note 14 above);
49. on the contrary, it is all too common. This difference, between the rules of rationality built into the meaning of the logical constants and those built into the meaning of moral terms, certainly calls for explanation. But it does not seem to be an objection to my conceptual role moral semantics.
50. What exactly is irrational about preferences that are reflexive or symmetric in this way? Suppose that the purpose of intentions is to control behavior: if one fails to behave on a given occasion as one intended to behave on that occasion, then one has necessarily made a mistake, either in one's intentions or in one's behavior. So intentions that are inconsistent-that is, intentions that make it impossible for one to act as one intends to act-guarantee one's making some such mistake, and are therefore irrational. This makes it plausible that a set of conditional intentions-intentions to act in certain ways if certain conditions obtain-that guarantees that, if those conditions obtained, it would be impossible to act as one intended to act in those conditions is also irrational. This then would explain why reflexive or symmetric preferences are irrational; for as I am using the term, a "preference" is just a conditional intention- specifically, the intention to do x and not y if one does either.
51. This is the theory that Parfit argues against in Reasons and Persons, part 2.
52. It may be that "eudaimonists," such as Aristode and Aquinas, do not recognize two distinct concepts, a concept of what promotes one's interests more, and a concept of what is (all things considered) a better thing to do, but see only one concept here. If so, these philosophers would not distinguish between the rule prescribed by the "self-interest theory of rationality" and the rule of practical reasoning that I have discussed above. Eudaimonism of this sort is clearly compatible with the semantics that I am proposing for moral terms.
53. See Virtues and Vices, xii. Foot also makes a similar point about the terms 'good' and 'harm': "it is quite impossible to call anything you like good or harm" (Virtues and Vices, 120).
54. This is Parfit's example of "Future-Tuesday-Indifference" (Reasons and Persons, 124).
55. As Garrett Cullity and Berys Gaut put it, neoHumean and Kantian theories hold a "constructivist" view of the relation between practical reason and the good, while Aristotelian theories hold a "recognitional" view of this relation;
56. see their "Introduction" to Ethics and Practical Reason, ed. Cullity and Gaut (Oxford: Oxford University Press, Clarendon Press, 1997), 1-27, esp. 4.
57. For instance, although Frank Jackson and Philip Pettit's official account of moral terms is an example of noncircular conceptual analysis, they also give an account of what it is for someone to have a "non-intellectual moral belief," in terms of the characteristic dispositions of beliefs of that sort;
58. see Jackson and Pettit, "Moral Functionalism and Moral Motivation," Philosophical Quarterly 45 (1995): 20-40, esp. 35-38. Since the dispositions that they appeal to are all rational dispositions, this account of moral belief could be adapted to the rule-based version of conceptual role semantics that I have been working with here.
59. For noncognitivism, see the works cited in note 1 above. For the causal theory of reference, see Richard N. Boyd, "How to Be a Moral Realist," in Essays on Moral Realism, ed. Geoffrey Sayre-McCord (Ithaca: Cornell University Press, 1988), 181-228. For noncircular conceptual analysis.
60. see Jackson and Pettit, "Moral Functionalism and Moral Motivation.
61. and David Lewis, "Dispositional Theories of Value," Proceedings of the Aristotelian Society, supp. vol. 63 (1989): 113-37. Not all previous approaches to moral terms fall neatly into one of these three categories.
62. For example, Peter Railton's ideas seem to fuse elements of both the causal theory and noncircular conceptual analysis; see his "Naturalism and Normativity," Social Philosophy and Policy 7 (1989): 151-74.
63. Moore argued for this stronger claim, when he said "propositions about the good are all of them synthetic and never analytic" (Principia Ethica, 7).
64. Recently, many philosophers have offered conceptual analyses that are explicitly acknowledged to be circular; see, for example, David Wiggins, "A Sensible Subjectivism?" in his Needs, Values, Truth, 3d ed. (Oxford: Oxford University Press, Clarendon Press, 1998).
65. and Michael Smith, "Response- Dependence without Reduction," European Review of Philosophy 3 (1998): 85-108. Since these circular conceptual analyses rely on the meaning that moral terms have in the object language, they do not in my view give a satisfactory semantics for moral terms.
66. Peacocke, "Can Possession Conditions Individuate Concepts?" Philosophy and Phenomenological Research 56 (1996): 433-60, esp. 451-52.
67. This claim is much weaker than the claim of noncognitivists like A. J. Ayer, that moral disputes "may always break down" without any irrationality or nonmoral error or ignorance on either side (that strong noncognitivist claim was effectively attacked by Foot in Virtues and Vices, 96-109). It is also weaker than the claim that these moral disagreements are "rationally irresoluble," in the sense that no amount of rational dialogue would ever lead to the dispute's being settled. It is just the claim that there are some moral questions on which it is possible for at least minimally rational, conceptually unconfused thinkers to disagree, despite having all the same relevant nonmoral beliefs.
68. Moore, Principia Ethica, 5-21. I should emphasize that I am not endorsing the stronger conclusions that Moore draws there (such as that "propositions about the good are all of them synthetic," or that goodness is an absolutely "simple," nonnatural property).