1 These distinctions are loosely adapted from Pettit's, Phillip ‘Realism and Response Dependence,’ Mind 100 (1991) 588–95Google Scholar.

2 It should be noted that Zemach has an unorthodox notion of property (143). However, this view does not appear to substantially affect the distinctions made here.

3 Note, however, that one of the arguments which we classify as an argument for metaphysical realists might also appeal to scientific realists (see section III.2 below).

4 What Zemach says about these arguments is that they show the ‘objective reality’ of aesthetic properties; however, he does not clarify what ‘objective’ means (56). At some points, it sounds as though the arguments indeed are meant to defend ontological primary property realism. One of them is presented in a section entitled ‘The Primacy of Aesthetic Properties’ (64). Also, Zemach describes a different argument, which clearly is an argument for primary property realism, as suggesting that ‘beauty is an objective feature of reality’ (111). At other points, however, it sounds as though the arguments offered to scientific realists are only meant to combat descriptive anti-realism (or what Zemach calls non-cognitivism), as when he describes the view ‘that there are no aesthetic properties in reality’ as the view that ‘“elegant” and “gaudy” should not be given a semantic interpretation’ (66).

5 ‘Truth and Beauty,’ The Philosophical Forum 18 (1986), 21. We refer to this article throughout our discussion, as sections of it are relevant to understanding Zemach's positions in Real Beauty.

6 In describing metaphysical realists, he remarks that ‘none of the proofs [for scientific realists] will be acceptable to realists who reject scientific realism. Metaphysical realists argue that …. Reality (das Ding an sich, the noumenon) is as it is, and we cannot assume that the theory that pleases us the most describes it more accurately than a theory that pleases us not at all’ (67). This implies that scientific realists, in contrast, do believe that a certain account of the world as it is in itself (i.e. of the noumenon) does describe that world accurately.

7 The other argument is this: since theoretical accounts of art that employ aesthetic terms are successful theories, a scientific realist should accept those theories and the reality of the entities that they posit (i.e. aesthetic properties) (56-61).

8 An earlier presentation of this argument is discussed by Goldman, Alan in ‘Realism About Aesthetic Properties,’ Journal of Aesthetics and Art Criticism 51 (1993) 31–7CrossRefGoogle Scholar.

9 Zemach extends this claim not only to scientific theories, but to argumentative assertions in general. See 68-70, and section III.2 below.

10 Premise two of the above argument, combined with premise four, yields this strong claim.

11 ‘Truth and Beauty,’ 30

12 See, for example, McAllister, James Beauty and Revolution in Science (New York: Cornell University Press 1996), 9–12Google Scholar.

13 ‘Truth and Beauty,’ 30-1. In the last sentence of this passage, it is unclear what ‘powerful’ means: it may mean simply ‘having predictive power.’ If so, however, it is still unclear to what ‘explanations’ is referring: Naive Realism itself or those scientific theories which may be justified by empirical observation once Naive Realism has been adopted. If the former, then Zemach is claiming that Naive Realism is adopted, in part, because it has predictive power; if the latter, then he is claiming that Naive Realism is adopted in part because it produces or licenses theories which have predictive power. What the former claim could mean is not at all clear to us; in either event, Zemach would be clearly begging the question at issue by calling predictive power an ‘aesthetic constraint,’ since that it is such is what he is attempting to prove. On the other hand, however, ‘powerful’ might mean ‘having explanatory power'; this would remove the circularity from the argument. In any case, since the inclusion of such a criterion in the argument does not affect our criticisms, we do not include the notion of being ‘powerful’ in our reconstruction of Zemach’s argument for the strong claim.

14 ‘Truth and Beauty,’ 32

15 In ‘Truth and Beauty,’ Zemach defines ‘Naive Realism’ as ‘a philosophical doctrine … according to which the world really is as it appears to be’ (30). As noted, he says that it is ‘the most basic contention of Naive Realism that seeming to see a pointer move is, prima facie, a good reason for believing that a pointer moves’ (31). This seems different from the sense of Naive Realism which Zemach dismisses in Real Beauty as ‘not a live option in metaphysics’ (53). There it is defined as the belief ‘that the human parochial image of nature is an accurate replica of the real world as it is in itself,’ and described as ‘quite unreasonable’ (53). This sense of Naive Realism is stronger than the former: its claim seems to be that all or most of the properties we experience are similar to some primary (i.e. noumenal) property. The former version claims only that some of the phenomenal properties we experience can provide justification for claims about primary properties. It is this weaker sense which Zemach employs in his argument here.

16 Though it does not bear directly on the success of his argument, one might also contest Zemach's claim that ‘of all possible belief-regulating laws, [Naive Realism] is at once the simplest and the one richest in colorful detail’ (‘Truth and Beauty,’ 31). Why, for example, is not Berkelyean Idealism more beautiful than Naive Realism? Though it has less variety, or richness, in Zemach's terminology (i.e. no noumena behind phenomena), could it not be said to be simpler, in not positing a relation between phenomena and noumena? Though he claims that aesthetic criteria are the sole tools for theory selection, Zemach offers little insight into how the actual evaluation of two theories, such as these, in terms of their aesthetic properties, could be approached with any theoretical rigor. A more in-depth treatment of such issues can be found in McAllister, Beauty and Revolution in Science, 105-24Google Scholar.

17 That epistemically good scientific theories must be beautiful is claimed by Sullivan, J.W.N. in ‘The Justification of the Scientific Method,’ Athenaeum (1919)Google Scholar. Kivy, Peter argues for the weaker thesis that certain types of beauty, though not necessary for epistemic virtue, always increase epistemic virtue when present in a theory (‘Science and Aesthetic Appreciation,’ Midwest Studies in Philosophy 26 [1991] 180–95)CrossRefGoogle Scholar. Historical evidence bearing on the use of aesthetic criteria in scientific theory evaluation is reviewed by McAllister, Beauty and Revolution in Science, 39–60Google Scholar.

18 Such an option would not be open, of course, to an eliminative materialist, who would insist that phenomenal terms fail to denote phenomenal entities, since there are no such entities. But scientific realists need not be eliminative materialists; see, for example, the discussion by Hung, Edwin in The Nature of Science: Problems and Perspectives (Belmont, CA: Wadsworth 1997), 212–9Google Scholar.

19 On such a secondary property view, ‘X is beautiful’ will mean only something like: ‘my phenomenal experience of X has the property of beauty,’ or ‘my experience of X is of a certain sort.’ Though he ultimately accepts a similar view as one element of his aesthetic realism (see §IV), Zemach offers some arguments which appear to criticize the view that aesthetic properties are merely secondary or tertiary properties: for example, he argues that aesthetic experiences cannot be what aesthetic sentences are about (30-2). However, these arguments fail to block the secondary property approach. One of his arguments is that such views cannot explain negative aesthetic appraisal; however, this criticism applies only to one theory of aesthetic experience (Clive Bell's). All a secondary property theorist need do here is posit a secondary property of ugliness. Zemach's second argument is that although such approaches should describe aesthetic experience in nonaesthetic terms, they fail to do so. However, such a non-circularity constraint only applies to ontological anti-realist approaches, which try to show that there are no distinctively aesthetic properties at all; a secondary property view need not be such a view. Of course, there are other criticisms of the view that aesthetic properties are secondary properties: see, for instance, Goldman, Alan Aesthetic Value (Boulder: Westview Press 1995), 18Google Scholar.

20 One might dispute this, but Zemach himself is careful not to make his argument here depend on the falsity of this claim. He writes that ‘a metaphysical realist may doubt whether aesthetic sentences can be true of reality, because our aesthetic judgments are not about noumena: the features we aesthetically evaluate are all phenomenal … Nevertheless, even if the noumenon we see as a green tree is entirely unlike our phenomenal green tree, it may still be pretty, sublime, repulsive, or gaudy, as we could verify had God chosen to let us know the world as it is’ (68).

21 Zemach clearly takes this to be an argument for primary property realism, since he refers to it as one of ‘several arguments that some tertiary properties are also primary’ (109). The full argument is given on 68-70.

22 Though supervenience is an important element of Zemach’s aesthetic theory, we do not discuss his use of this notion here. For recent discussion of aesthetic supervenience see Levinson, Jerrold ‘Aesthetic Supervenience,’ in Music, Art, and Metaphysics (New York, NY: Cornell University Press 1990)Google Scholar; Currie, Gregory ‘Supervenience, Essentialism and Aesthetic Properties,’ Philosophical Studies 58 (1990) 243–57CrossRefGoogle Scholar; Bender, John ‘Realism, Supervenience, and Irresolvable Aesthetic Disputes,’ Journal of Aesthetics and Art Criticism 54 (1996) 371–81CrossRefGoogle Scholar; and Eaton, Marcia ‘The Intrinsic, Non-supervenient Nature of Aesthetic Properties,’ Journal of Aesthetics and Art Criticism 52 (1994) 383–97CrossRefGoogle Scholar.

23 The fallacy resulting from the use of this hypothetical God-like observer is similar to the one occurring in the argument for ontological primary property realism discussed above (§III.2).

24 We thank Bernard Linsky, Patrick McGivern, and Alex Rueger for helpful discussions and comments on the manuscript. G.G.P. was supported by a Social Sciences and Humanities Research Council of Canada Doctoral Scholarship. A.A.C. was supported by a Social Sciences and Humanities Research Council of Canada research grant.