psychologistic semantics and moral truth

Psychologistic Semantics and Moral TruthAuthor(s): Terence HorganSource: Philosophical Studies: An International Journal for Philosophy in the AnalyticTradition, Vol. 52, No. 3 (Nov., 1987), pp. 357-370Published by: SpringerStable URL: http://www.jstor.org/stable/4319925 .

Accessed: 28/06/2014 19:01

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Philosophical Studies: AnInternational Journal for Philosophy in the Analytic Tradition.

http://www.jstor.org

This content downloaded from 193.142.30.55 on Sat, 28 Jun 2014 19:01:21 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=springer

http://www.jstor.org/stable/4319925?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


TERENCE HORGAN

PSYCHOLOGISTIC SEMANTICS AND MORAL TRUTH

(Received 16 September, 1986)

1. In this paper I shall outline an account of moral truth which differs in significant ways from the various traditional accounts in meta-ethics, and I shall argue that the potential virtues of this new approach make it worthy of further examination and elaboration. The proposal falls within what I call "psychologistic semantics," a broad orientation toward the workings of natural language which treats the notions of truth and reference rather differently than they have traditionally been dealt with in philosophy of language. First I shall summarize the leading ideas of psychologistic semantics, and its main attractions. Then I shall explain the new conception of moral truth which emerges within this framework, and I shall point out some important advantages this conception has over the various standard options in meta-ethics. Finally I shall list a number of salient questions which need to be answered by any well-developed version of the general approach I shall describe.

2. The dominant recent conception of truth and reference in philosophy of language, which I shall call referential semantics, has two main components. First is the contention that for natural languages (as well as for the formal languages typically studied by logicians), the notion of truth is to be characterized via a Tarski-style recursive truth definition; such a definition will employ semantical concepts like the notion of a singular term's denoting an object, and the notion of a predicate's applying to an object. (Denotation and applicability may be conveniently lumped together under the blanket term 'reference'.) Second is the contention that the notion of reference should itself be understood in a robustly realist way - i.e., as involving genuine, direct, word-world relations. (So-called "causal theories of reference" constitute one sort of attempt to provide a robustly realist account of reference.) 1

Psychologistic semantics contrasts sharply with referential semantics.

Philosophical Studies 52 (1987) 357-370. C 1987 by D. Reidel Publishing Company.



358 TERENCE HORGAN

The leading ideas are the following.2 First, truth is a normative notion: when we call a sentence true, we are expressing the normative judgment that it is correctly assertible. Second, this is not pragmatism, because correct assertibility is to be distinguished from warranted assertibility -

even ideal warranted assertibility. Third, standards of correct assertibility are not monolithic; rather, they vary somewhat from one context to another, depending upon the specific purposes our discourse is serving at a given time. Fourth, correct assertibility is ordinarily a joint product of two factors: (i) the relevant assertibility-norms, in a given context of discourse; and (ii) the actual behavior of various genuine entities in the world. Fifth - and here the contrast with referential semantics emerges quite strongly - our discourse often employs standards of correct assertibility under which a sentence can count as correctly assertible (and hence as true) even if it contains singular terms that do not pick out any actual constituent of the world, and even if no actual constitutents of the world answer to the sentence's unnegated existential-quantifier expressions.

Let me clarify these points, in particular the last one, with an example. Under psychologistic semantics, the truth of a sentence like

(B) Beethoven's fifth symphony has four movements

probably does not consist in the fact that some actual entity in the world answering to the term 'Beethoven's fifth symphony' also answers to the predicate 'has four movements'. Rather, (B) is true because the actual behavior of various genuine entities in the world, none of which need answer to the term 'Beethoven's fifth symphony', suffice to render (B) correctly assertible under the relevant assertibility norms. Especially germane here is the behavior by Beethoven which we would call "composing his fifth symphony." But a considerably wider range of behaviors by actual entities is relevant too: in particular, Beethoven's earlier behavior in virtue of which his later behavior counts as "composing his fifth symphony"; and also the broad range of human behaviors and practices in virtue of which such behavior counts as "composing a symphony" in the first place.4

Returning now to my sketch of the leading ideas of psychologistic semantics, several additional points should be added to the five already mentioned. Sixth, assertibility-norms normally link a sentence's correct



PSYCHOLOGISTIC SEMANTICS AND MORAL TRUTH 359

assertibility quite closely to the kinds of evidential conditions which competent speakers would regard as strong grounds for the sentence; thus, a sentence normally counts as prima facie correctly assertible if one's evidence is sufficiently relevantly similar to what one considers prototypical evidence for such a sentence. (We all know quite well, for instance, what kinds of evidence are relevant to the truth of sentence (B) above. And (B) is undoubtedly correctly assertible, given our ordinary standards of correct assertibility for symphony talk - whether or not there really exists, either in space-time or in "Plato's heaven," an actual entity to which the term 'Beethoven's fifth symphony' bears a referential relation.) Seventh, the linkages between evidential conditions and correct assertibility normally are defeasible: this defeasibility rests in part upon a background of other sentences which themselves are assumed, in a given context, to be correctly assertible. (As Quine and Duhem stressed long ago, our sentences really face the tribunal of empirical evidence collectively, not singly.) Eighth, this interanimation of sentences, vis-a-vis the role of evidential conditions in correct assertibility, generates a whole spectrum of ways that a sentence's correct assertibility can depend on the behavior of actual entities in the world. (At one end of the spectrum would be sentences whose assertibility norms, in a given context of usage, coincide with those laid down by referential semantics; under these norms a sentence can be true only if some unique actual entity answers to each of its singular terms, and at least one actual entity answers to each of its unnegated existential-quantifier expressions. At the other end of the spectrum would be sentences whose assertibility norms, in a given context, are such that those sentences are sanctioned as correctly assertible by the norms alone, independently of the behavior of actual objects in the world. And various intermediate points would be occupied by sentences whose correct assertibility, in a given context, does depend upon the behavior of actual objects in the world, but where this dependence does not consist in the direct "correspondence" described by referential semantics.) Ninth, the task of uncovering the various assertibility norms which we actually employ, in various communicative contexts and for various kinds of discourse, is largely a task for the empirical sciences, in particular for the cognitive psychology of human linguistic competence; hence the term "psychologistic semantics." Tenth, the ontological



360 TERENCE HORGAN

commitments embodied in our scientific and non-scientific discourse depend not merely upon the range of our quantifier expressions (pace Quine), but also upon the operative assertibility norms; the entities to which we are ontologically committed are those entities (and only those entities) which we quantify over when we are using the assertibility norms laid down by referential semantics.

Under psychologistic semantics, ontology depends in large part upon what the empirical facts turn out to be about the operative assertibility norms in our discourse. Still, I think it is plausible to conjecture that under a correct and fully worked out psychologistic semantics, it would turn out that the ontological commitments embodied in our language are very sparse indeed. Specifically, I suspect that the only genuine ontological commitments are to physical objects - where by 'physical object' I mean what Quine means when he says that a physical object is "simply the whole four-dimensional material content, however sporadic and heterogeneous, of some portion of space-time."5 But my sub- sequent discussion will be largely independent of this nominalistic, minimalistic, materialistic hunch about ontology.

What about truths of pure mathematics, and their relation to ontology? Under psychologistic semantics, it seems very plausible to suppose that the assertibility norms of pure mathematics fall at or near the extreme "conventionalist" end of the above described spectrum of ways that language can be related to the world. I.e., it is plausible to suppose that the correct assertibility of statements we count true in mathematics is entirely - or almost entirely - a function of the relevant assertibility norms, independently of the behavior of actual objects in the world. This is a strongly anti-platonistic conception of mathematical truth.

Since I consider it largely an empirical question what assertibility norms we employ in various contexts of discourse, I also think that the contest between referential semantics and psychologistic semantics is largely an empirical one. But let me now mention two advantages of psychologistic semantics, over against referential semantics, which are especially germane to traditional questions that arise in philosophy generally and also in meta-ethics. One major advantage is epistemological, the other ontological.




Psychologistic semantics holds out the promise of a tractable epistemology, particularly within the framework of a naturalistic conception of human beings as complex physico-chemical systems (the "granola conception," as I'll call it). Consider, for instance, our knowledge of statement (B) above. Philosophers who are concerned about ontological questions in aesthetics, and who approach these questions within the framework of standard referential semantics, have been much exercised by the ontological status of musical works of art. If we begin by assuming that a sentence like (B) cannot be true unless some real entity answers to the term 'Beethoven's fifth symphony', then we are apt to suppose (as many have) that this term denotes a complex, internally structured universal - an abstract entity which exists eternally, and is not part of the spatiotemporal causal nexus. But once we suppose this, it becomes very hard to see how mere humans could ever have knowledge about symphonies, such as the knowledge expressed by (B). For, we cannot come into any sort of causal contact with these putative entities, but rather can only causally interact with those concrete things we call "performances of Beethoven's fifth symphony," "copies of the score of Beethoven's fifth symphony," etc. One is tempted to say, of course, that we know the symphony itself via our knowledge that concreta "token" it or "describe" it. But this only pushes the epistemological problem back a step. For, how could we know that a given event tokens a certain abstract entity, or that a given manuscript describes it, unless we could somehow directly compare the event or the manuscript with the abstract entity itself? Yet it is most implausible - especially given the granola conception of human beings - to suppose that we can have some sort of quasi-perceptual cognitive communion with an entity which has no spatio-temporal location and does not causally interact with anything. And if someone replies by saying that our performance-instances and score-copies are instances of, or copies of, a particular abstract universal because we stipulate it so, rather than because we somehow directly compare those concrete things with the putative abstract entity, then again the epistemological puzzles are merely pushed back. For, now it becomes very hard to see either (i) how we can justifiably claim to know that there really exist such putative abstract entities as symphony-types at all; or (ii) how it could



362 TERENCE HORGAN

be the case that our stipulative acts could link a term like 'Beethoven's fifth symphony' to one specific such complex universal, rather than to some other one that is isomorphic to the first.

These highly vexing epistemological problems do not arise under psychologistic semantics, given the plausible assumption that the assertibility norms for symphony talk are norms under which a sentence like (B) can be correctly assertible even though no actual entity answers to the term 'Beethoven's fifth symphony'. Similarly for a wide variety of other kinds of discourse. (Talk about numbers, sets, and other mathematical entities is another especially salient example.) Thus one major philosophical advantage of psychologistic semantics, vis-a-vis referential semantics, is its greater potential for rendering our knowledge claims justifiable under the granola conception of human beings.

A second advantage is ontological. In ontology, as in theory con- struction generally, there is desirability in parsimony. The most valuable kind of parsimony comes not merely from minimizing the number of distinct kinds of entities one posits, but also from a metaphysical scheme in which the relational network connecting various kinds of entities to one another is not needlessly complex and baroque. Thus it is highly desirable, if possible, to make do with an ontology in which (i) the only entities posited are ones that belong to the spatio-temporal causal nexus; and (ii) the causal nexus itself is not construed as containing a baroque synchronic hierarchy in which sui generis entities at "higher" ontological levels (corporations, universities, nations, and the like) supervene upon lower-level entities of the kind described in the natural sciences. Together these two desiderata render highly attractive a minimalist materialist ontology, such as an ontology con- sisting of Quinean physical objects and nothing else. Yet it is most implausible to suppose either that putative higher-level entities can be reductively identified with entities posited by such a minimalist ontology, or that talk about the "irreducible" entities can be systema- tically paraphrased into a more austere idiom which avoids apparent reference to those putative entities.6 Hence the attractiveness for ontology of psychologistic semantics, over against referential semantics. For the former can, while the latter evidently cannot, allow for genuine truth without ontological commitment - even when reductive identi-




fications are not forthcoming, and even when the true sentences about higher-level entities cannot be "paraphrased away."

3. Psychological semantics affords an account of moral truth different from the traditional meta-ethical alternatives. I shall call this account psychologistic moral realism (henceforth PMR, for short). The key features are as follows. First, there is indeed genuine moral truth: viz., correct assertibility, under contextually appropriate assertibility norms, of sentences containing moral vocabulary. Second, such sentences are not equivalent in meaning to sentences free of moral vocabulary; and the former are not "reducible" to the latter in any other philosophically interesting sense either. Third, despite their irreducibility, moral statements also are not "transcendentally true" - where transcendental truth might mean either (i) that these statements depend for their truth upon the existence, as part of the "furniture of the world," of such entities as sui generis moral properties or sui generis moral facts; or at any rate (ii) that the correct assertibility of these statements depends upon some- thing more than merely (a) assertibility norms that are grounded in actual linguistic practices and the purposes underlying those practices, and (b) the naturalistically describable behavior of physical objects.

PMR obviously differs from the various forms of noncognitivism, such as emotivism and prescriptivism. For, noncognitivist theories claim that moral sentences really make no assertions, and hence that they are not literally true or false at all. Noncognitivist theories are prima facie highly implausible and unintuitive, because moral talk does after all have the grammatical form of declarative truth-stating discourse, and also because in our ordinary practice we regard such talk as being no less truth-stating than other kinds of discourse that employs declarative sentences. PMR is therefore preferable to noncognitivism, because PMR accords better with the way we use moral language and with our practice of assigning truth values to moral sentences.

PMR also differs from various forms of meta-ethical "naturalism" -

i.e., positions which claim that sentences containing moral vocabulary are equivalent in meaning to, or otherwise interestingly reducible to, sentences that lack any moral or normative vocabulary. These reductive accounts have never been very plausible, because as Moore's celebrated



364 TERENCE HORGAN

"open question argument" makes clear, there seems to be a palpable meaning gap between moral language (and other normative language) on the one hand, and non-moral, non-normative language on the other. PMR is preferable to meta-ethical naturalism, because PMR respects the conceptual autonomy of moral language and other normative discourse, vis-a-vis non-moral, non-normative language.

Among the traditional approaches to moral truth, meta-ethical "non-naturalism" is perhaps the one that shows the most respect for the conceptual autonomy of moral talk, and for the fact that we regard such talk as statement-making and as possessing truth value. Yet the articulated versions of non-naturalism, such as Moore's, face severe epistemological and metaphysical problems of the kind I discussed in the preceding section. Epistemologically, it is very difficult to explain convincingly how we could gain cognitive access to transcendent "non- natural" moral properties or facts - or, more generally, to sui generis transcendent moral truth. (Moore himself, like other non-naturalists before him, posited a special quasi-perceptual faculty of moral intuition; but this is implausible, especially under the granola conception of human beings.) Metaphysically, sui generis transcendent moral truth is very hard to integrate into a materialistic, naturalistic, world-picture.

Nowadays it is fashionable for defenders of 'moral realism" to try dealing with these problems by appealing to the idea that moral properties, moral facts, or moral truths are supervenient upon non-moral ones. They typically argue, via a form of "innocence by association," that supervenience renders moral truth no more problematic than, say, truths about human mentality - which also are commonly held to be supervenient upon physico-chemical truths.7 But this way of arguing is quite dubious, because supervenience should be acknowledged as a feature of moral discourse under any meta-ethical account of moral language. The non-cognitivist R. M. Hare, who evidently introduced the word 'supervenience' into the philosophical lexicon, insisted on this point - although for him supervenience was a contraint upon proper moral talk rather than a feature of moral properties or moral truth, since he denied that there are moral properties and denied that moral talk has truth value.8 Even more to the point, Moore himself also insisted on the supervenience of the moral upon the non-moral, even though he did not (as far as I know) actually use the word




'supervenience'.9 So mere appeal to supervenience removes none of the epistemological or metaphysical difficulties of objective non-naturalist meta-ethical theories like Moore's.10

PMR is preferable to such theories, precisely because it appears capable of avoiding these difficulties. Moral epistemology becomes potentially quite tractable under PMR, even assuming the granola conception of human beings. For, knowledge of moral truth becomes essentially a matter of mastering the assertibility norms of moral discourse, together with the capacity for cognitive access to the kinds of non-transcendent evidential situations which, under relevant assertibility norms, make for correct assertibility of specific moral claims. And moral truth is no longer metaphysically mysterious, or at odds with a naturalistic, materialistic, ontology. For, even though moral truth is autonomous and non-reducible, nevertheless it is not transcendent either. Rather, like other irreducible kinds of truth (e.g., truth about corporations, or nations, or symphonies), moral truth is ultimately a product of two non-transcendent factors - viz., (i) the operative assertibility norms, and (ii) the actual behavior of the actual, non- transcendent, entities in the world.

4. I shall conclude by listing a number of issues that will require further investigation, as part of any attempt to work out the details of PMR within the framework of psychologistic semantics. Some of these issues need examination within psychologistic semantics anyway, but are especially salient with respect to moral language.

(i) The nature of, and emergence of, assertibility norms for moral language. Psychologistic semantics requires a general account of the nature of assertibility norms, and of their emergence and maintenance within collective social practice. Developing such an account is not purely aprioristic work for philosophers alone; on the contrary, it probably will depend, in part, upon empirical developments in dis- ciplines like cognitive psycholinguistics. An adequate account should do two things simultaneously; (a) explaining how norms of correct assertibility emerge from actual linguistic practice; and (b) explaining how individuals, and even whole cultures, can be mistaken about what is correctly assertible under the norms governing their own discourse, where the mistakes result not from any defects in linguistic competence,



366 TERENCE HORGAN

but rather from factors like inadequate epistemic contact with world. Also, an adequate account of assertibility norms presumably will advert not merely to actual linguistic practice as a major partial source of normativity, but also to the underlying purposes of various linguistic practices as another important partial source.

Questions about the nature and emergence of assertibility norms are especially intriguing with respect to moral language and other normative language, because normativity now enters at two levels: at the metalinguistic level, where assertibility norms are operative just as much as they are for any other kinds of discourse; and at the object-language level, since the sentences in question themselves have moral content or some other kind of normative content.

We need ways of understanding normativity in general, including the normativity governing correct language use, which render normativity non-mysterious, relative to materialistic metphysics and the granola conception of human beings, and which do so without attempting any reduction of normativity to the non-normative.'1 Non-reductive explanatory accounts, as they develop, also are likely to shed light on the interplay between metalinguistic assertibility norms and those object- language sentences which themselves employ moral terms or other normative terms.

(ii) Ways that the behavior of objects in the world contributes to moral truth. I suggested earlier that in the case of sentences like (B), the world's contribution to correct assertibility is considerably more subtle, indirect, and holistic than it would be under referential semantics. The truth of (B) does not depend upon there existing, as part of the world's furnishings, some actual entity which answers to the term 'Beethoven's fifth symphony' and which satisfies the predicate 'has four movements'. Rather, as I said, it depends upon a wide variety of behavior of actual entities in the world - including the activity by Beethoven we call "symphony composing," and the various behaviors and practices in the culture which make it the case that activities like Beethoven's actually count as symphony composing.

The question that arises for PMR concerns the ways that goings-on in the world contribute to the correct assertibility of various kinds of moral talk. In the case of statements like (B), we already have a fairly good idea what kinds of events are particularly germane for the




sentence's correct assertibility, and why. But there is no denying that the comparable question is more elusive for moral language especially, and to some extent for normative language in general. We need a clearer conception of the various ways that the world itself contributes to the correct assertibility of moral and other sentences; and this kind of understanding is likely to go hand in hand with a clearer understanding of assertibility norms for such sentences.

(iii) The issue of bivalence. The thesis of semantic bivalence is the claim that every sentence is either true or false. Under psychologistic semantics, it is quite likely that certain sentences, under contextually appropriate assertibility norms, will be such that neither they nor their negations are correctly assertible. Such sentences will be neither true nor false, under the operative norms. One reason to expect failures of bivalence is the prevalance - indeed the ubiquity - of vagueness in our language. Often when vague terms are in use, we expect genuine "in-between" cases where neither a sentence nor its negation are correctly assertible. Everyone's favorite example is the fellow who has rather little hair on his head, but who has enough whisps here and there that no large spots are totally bare. It may well be neither true nor false, of such a person, that he is bald.

An important question for PMR is whether there are more substantial failures of bivalence in moral language (or in other kinds of normative language), over and above the relatively harmless in-between cases of the kind we sometimes encounter with garden-variety vagueness. Certain rather pressing moral problems might well turn out to be subject to moral indeterminacy. Questions about the morality of abortion, in certain kinds of circumstances, are obvious candidates for such indeter- minate status, especially since the predicate 'is a person' is itself palpably vague. A newly fertilized human egg cell is not a person, but a pre-natal infant two weeks from term arguably is a person. Hence some sentences asserting, for instance, the moral permissibility of abortions in specific kinds of circumstances during certain intermediate points of pregnancy, may be neither true nor false.'2

(iv) The question of methodology of normative ethics. Epistemology, among its other goals, seeks to explain why some particular methods of inquiry are more appropriate than others, in various subject areas. As I argued earlier, psychologistic semantics holds out encouraging



368 TERENCE HORGAN

prospects for making epistemology more tractable, especially in relation to those kinds of knowledge which seem most recalcitrant under traditional referential semantics - such as mathematical knowledge, which has always been very hard to reconcile with the platonistic conception of mathematical truth towards which we seem driven under referential semantics.

Moral inquiry too has always seemed rather elusive, especially in comparison to the more directly empirical kinds of inquiry typical in the sciences. In making specific moral judgments, and in also in pur- suing normative ethics, we tend to use methods such as (a) relying upon our intuitions concerning cases that seem clear-cut pre-theoretically; (b) dealing with difficult cases by pointing out seemingly important similarities with, and differences from, cases that seem intuitively more clear-cut; (c) employing "wide reflective equilibrium" to effect as good an overall fit as possible between our general normative principles and our pre-reflective intuitive judgments; etc. PMR has the potential to be informative about why such procedures are generally appropriate, and also about their potential pitfalls; for, as I stressed earlier, psychologistic semantics seeks to wed truth (i.e., correct assertibility) more closely to our actual evidential standards than does traditional referential semantics. A credible account of moral truth should shed light on the epistemic appropriateness of various methods of inquiry in normative ethics and in moral decisionmaking.

There is, then, serious work to be done before it can be said with confidence that psychologistic semantics in general, and psychologistic moral realism in particular, are viable positions. But the attractions of both the general program for semantics and the specific sub-program for explaining moral truth make the two projects worthy of serious philosophical attention."3

NOTES

I Referential semantics is epitomized by Hartry Field's well known paper 'Tarski's theory of truth', Journal of Philosophy 69 (1972): 347-375. Some philosophers, including Field himself, have held that although referential semantics would be an essential part of a full-fledged semantic theory for natural language, such a theory would include other components as well. See, for instance, Field's 'Logic, meaning and conceptual role', Journal of Philosophy 74 (1977): 379-409. Psychologistic semantics,




as described below, is neutral about the matter of supplementary components of semantic theory. 2 Psychologistic semantics is described in more detail in my papers 'Psychologism, semantics, and ontology', Nouis 20 (1986): 21-31; and 'Truth and ontology', Philo- sophical Papers 15 (1986): 1-21. There appear to be some substantial similarities between my own position and certain recent discussions of language and the world by philosophers who declare themselves opposed to metaphysical realism. I consider myself a metaphysical realist, despite my rejection of referential semantics. However, it may well be that much of what I say here could be endorsed by these anti-realists. Important recent anti-realist works include Michael Dummett, 'What is a theory of meaning?', in S. Guttenplan (ed.), Mind and Language (Oxford: Oxford University Press, 1975), and 'What is a theory of meaning? (II)', in G. Evans and J. McDowell (eds.), Essays in Semantics (Oxford: Oxford University Press, 1976); Nelson Goodman, Ways of Worldmaking (Indianapolis: Hackett, 1978); Hilary Putnam, Reason, Truth, and History (Cambridge, UK: Cambridge University Press, 1981), and Realism and Reason: Philosophical Papers, Volume 3 (Cambridge, UK: Cambridge University Press, 1983); Richard Rorty, Philosophy and the Mirror of Nature (Princeton: Princeton University Press, 1979), and Consequences of Pragmatism (Minneapolis: University of Minnesota Press, 1982). 3 Depending upon how much gets packed into the notion of idealization, it might turn out that truth in my sense could not diverge from ideal warranted assertibility. But even if this is so, I think that my notion of truth is different, since it is not radically epistemic. (I suspect that the notion of "ideal" warranted assertibility covertly re-introduces the idea that truth depends upon how things really, metaphysically, are in the world.) I Here I should add an important proviso: I doubt that there is some actual entity in the world answering to the term 'Beethoven' either. The problem is that singular terms purportedly referring to persons (and, indeed, most other singular terms as well) are somewhat vague, whereas genuine entities in the world are not vague. Quine describes well this vagueness problem in his 'Events and reification', in E. LePore and B. McLaughlin (eds.), Actions and Events: Perspectives on the Philosophy of Donald Davidson (Oxford: Basil Blackwell, 1985). He writes: "Physical objects are well individuated, being identical if and only if spatiotemporally coextensive. Yet is has been felt that physical objects, bodies in particular, are poorly individuated. Who can aspire to a precise intermolecular demarcation of a desk? Countless minutely divergent aggregates of molecules have equal claims to being my desk. True enough, but this circumstance attests only to the vagueness of the term 'desk', or 'my desk', and not to that of 'physical object'.... Vagueness of boundaries has sparked philosophical discussion in the case of desks because of their false air of precision. Mountains meanwhile are taken in stride; the thought of demarcating a mountain does not arise. At bottom the two cases really are alike. ... The mountain is no particular physical object; any one of a vast number would serve. The desk is to be viewed similarly; the cases differ only in degree" (pp. 167-168).

Beethoven too is to be viewed similarly, in my view - for the very reasons Quine gives concerning the desk. I W. V. 0. Quine, Philosophy of Logic (Englewood Cliffs: Prentice-Hall, 1970), p. 30. 6 I explain the problems with the reductive identification strategy, and with the elimina- tive paraphrase strategy, in my two papers cited in Note 2. ' I owe to Mark Timmons the phrase "innocence by association." For examples of such arguments, see David 0. Brink, 'Moral realism and sceptical arguments from disagree- ment and queerness', Australasian Journal of Philosophy 62 (1984): 111-125; and Nicholas Sturgeon, 'Moral explanations', in D. Copp and D. Zimmerman (eds.), Moral- ity, Reason, and Truth (Totowa: Rowman & Allanheld, 1985). In fairness to these authors, it should be pointed out that they recognize the need to do more than merely



370 TERENCE HORGAN

plead innocence by association. For instance, in 'Moral explanations' Sturgeon argues that moral facts are explanatory vis-d-vis human behavior, by virtue of their supervenience upon non-moral facts. Timmons and I present a critique of this argument in our paper 'Moral epiphenomenalism', in preparation. 8 R. M. Hare, The Language of Morals (Oxford: Oxford University Press, 1952). 9 Moore wrote that "if a given thing possesses any kind of intrinsic value in a certain degree, then not only must that same thing possess it, under all circumstances, in the same degree, but also anything exactly like it, must, under all circumstances, possess it in exactly the same degree. Or to put it in the corresponding negative form: It is impossible that of two exactly similar things one should possess it and the other not, or that one should possess it in one degree, and the other in a different one" [his emphasis]. Quoted from G. E. Moore, 'The conception of instrinsic value', in his Philosophical Studies (New York: Harcourt, Brace & Co., 1922), p. 261. 10 For this reason, among others, the thesis that all truths are supervenient upon the truths of physics does not by itself constitute an adequate version of metaphysical materialism. I discuss the point at greater length in my 'Supervenience and cosmic hermeneutics', Spindel Conference Supplement on Supervenience, Southern Journal of Philosophy 22 (1984): 19-38. 11 A very suggestive work of the sort I have in mind is Edna Ullman-Margalit's The Emergence of Norms (Oxford: Oxford University Press, 1977). 12 I am indebted to Ned Hettinger for bringing to my attention the relevance of the abortion issue to questions about bivalence in morals. 13 This paper was written during a summer seminar entitled "Philosophical Problems about Truth and Reality," directed by Hilary Putnam and sponsored by the National Endowment for the Humanities. I am grateful to the Endowment for its support, and I thank Putnam and Mark Timmons for helpful comments and discussion.

Department of Philosophy, Memphis State University, Memphis, TN38152, U.S.A.



psychologistic semantics and moral truth

Documents