1 Walton even tries to bring abstract painting and music in range of his analysis of representational art (see secs. 1.8 and 8.6).

2 As Walton observes, this suggestion is not unprecedented, and in fact the germ of Walton“s theory can be found in remarks of Ernst Gombrich in his well-known essay, “Meditations on a Hobby Horse,” in Meditations on a Hobby Horse and Other Essays on the Theory of Art (London: Phaidon Publishers, 1963).

3 The terminology is idiosyncratic, and perhaps unfortunate, for the following reason. Mental states such as desiring, fearing, imagining and thelike are often said to have intentional objects, at least, where this is a way of describing part of their contents, to wit, the objects at which they aim, whether or not such things exist in reality. Now the content of such states and the content of representations clearly have a lot in common, but Walton has thus barred himself from characterizing the internal content of artistic representations in terms of objects in this sense.

4 John Barth stories, Escher prints, Steinberg drawings and Picasso's Bull's Head are offered as examples here.

5 This chapter also deals illuminatingly with matters of style, narration and point of view in verbal and non-verbal representations-matters which cannot be pursued here for lack of space

6 The reason for my qualifying “almost” is this. Many critical observations about characters in fictions, especially those of a comparative and cross-work sort, seem most perspicuously to call for recognition ofa sense of characters as abstract aspects of novels—as opposed to fictional persons—which are created by their authors in the act of creating their works, and which thus exist much as does any cultural object of an abstract sort. Of course, to recognize characters in this sense is not to admit into one's ontology Meinongian inexistents. (For treatments along these lines, see van Inwagen, Peter, “Creatures of Fiction,” American Philosophical Quarterly, 14 [1977]: 299–308Google Scholar; Lamarque, Peter, “Fiction and Reality,” in Philosophy and Fiction, edited by Lamarque, P. [Aberdeen: Aberdeen University Press, 1983], pp. 52–72Google Scholar; and Currie, The Nature of Fiction, who calls such abstract entities “roles”.)

7 As it turns out, works are sometimes interpreted in such fashion that confirmation of the make-believe theory is the upshot, though without theinterpretiveness of the descriptions offered being signalled—that is, without acknowledging that such works or our engagement with them might plausibly be described in ways not so congenial to makebelieve theory. But this seems a pardonable excess for one in the grip of a theory as fruitful as Walton's.

8 See Wollheim, Richard, Art and Its Objects (Cambridge: Cambridge University Press, 1980)CrossRefGoogle Scholar, supplementary essay 5, and Painting as an Art (Princeton: Princeton University Press, 1987), chap. 2. In a forthcomingessay, Walton tries to show that Wollheim's notion of seeing in should in fact be analyzed in terms of imagined seeing (see Kendall Walton, “Seeing-in and Seeing Fictionally,” in A. Savile and J. Hopkins, eds.,Psychoanalysis, Mind and Art ‘Blackwell, forthcoming’), butI cannot address that here.

9 It is not open to Walton to appeal to an expansive notion of imagining broad enough to encompass the sort of perceptual experience I am claiming is independent of, and logically prior to, imagining as ordinarily understood. For he admits in the opening chapter of Mimesis that, despairing of defining it explicitly, he has no choice but to rely on our pre-analytic understanding of imagining to fill out the account. Walton does, however, manage tocontrast usefully imagining—negatively—with believing, entertaining, supposing, imaging and so on.

10 See sec. 1.8.

11 Note also that I have little option but to see it that way; presumably, whether to imagine it or not is something there should be more discretion about.

12 See sec. 7.4.

13 See Currie, The Nature of Fiction, chap. 1, for such a proposal. This view, note, can easily be extended to non-verbal fictions, unlikeothers criticized by Walton to be mentioned shortly.

14 Though I certainly agree with Walton that a fictional text is not adequately characterized as one that is non-asserted, I am less comfortable than he is that a text's being fiction (i.e., something it is prescribed we make-believe) and its being in portions asserted are perfectly compatible. The problem I see is this. While on a sentence-by-sentence basis, there is perhaps no logical or pragmatic incoherence in simultaneously intending your hearer to believe P, and bidding him to imagine P, but at a higher level, making a literary fiction in clear-cut cases (i.e., leaving “new journalistic” and other clearly borderline modes aside) seems to involve a governing or framing intention for a text as a whole—that it be taken imaginatively, and that it be understood as related, in particular, by a narrator necessarily non-identical with the author. I am not sure that such a governing intention does not preclude an assertoric stance between the author of a fictional work and individual statements of the text, even those occurring in fairly essay-like stretches.

15 This follows my suggestion regarding musical works in “What a Musical Work Is,” Journal of Philosophy, 77 (1980): 5–28CrossRefGoogle Scholar (reprinted in Jerrold Levinson, Music, Art, and Metaphysics [Ithaca: Cornell University Press, 1990]).

16 Related criticisms of Walton on this point are to be found in Neill, Alex, ”Fear, Fiction, and Make-Believe,” Journal of Aesthetics and Art Criticism, 49, 1 (1991): 47–56CrossRefGoogle Scholar, and Currie, The Nature of Fiction, chap. 5. Neill claims that Walton forgets that Charles's perspective on his game is an internal one ”… and thus that he make-believes that he is threatened by the slime rather than merely believes that it is make-believe that he is threatened” (p. 52), whileCurrie alleges that Walton's theory here and elsewhere overlooks or blurs“thedifference between the operator and the attitude senses of makebelieve” (p. 210), favouring the former when the latter is phenomenologicallymore appropriate and explanatory.

17 Discussed briefly in my note, “The Place of Real Emotion in Response to Fictions,” Journal of Aesthetics and Art Criticism, 48, 1 (1990): 79–80CrossRefGoogle Scholar. What I claimed there is that parallel emotions for real individuals often potentiate make-believe emotions for related characters, and make them seem, on reflection or recollection, more or other than they strictly are. It is possible my explanation of this phenomenon involves ascribing confusion to the appreciator, though of an aesthetically harmless sort, as to what really transpires in his experience during the phase of maximum involvement. I remain uncertain whether this is so, and whether if so, it renders the explanation implausible.

18 These Darwinian responses would be akin to what Noel Carroll, in his stimulating The Philosophy of Horror (New York: Routledge, 1990), recognizes under the rubric of shock reactions, though of broader scope.Carroll makes the important observation that horror is not reducible tonon-cognitive shock, but he fails to see that something of that more primitiveorder may be standardly involved in the way horror cinema works on us. (See my review of Carroll's book, which treats of some of the same issuesas Walton's, in Journal of Aesthetic and Art Criticism, 49, 3 [1991].Google Scholar) The Darwinian responses I speak of are also akin to the “gut feelings” Walton considers at one point (p. 198), but what he says against them as a wholly adequate explanation of Charles's behaviour at the horror movie does not, I think, invalidate the sort of secondary role I have in mind for them here.

19 See Walton, Kendall, “Transparent Pictures,” Critical Inquiry, 11 (1984): 246–77.CrossRefGoogle Scholar

20 I would further conjecture that a postulate of Darwinian (or corporeal) proto-belief states is perhaps most plausible where it is a question of threatening or attractive things, i.e., things which are “interesting to the will,” in Schopenhauer's sense.

21 Thus, a given instance of emotional response to a fiction might very well involve all of the following, to different degrees: (a) make-believeemotion of a sustained sort toward a given character; (b) moments of irrational, real emotion toward a character, when at peak points one loses one's control of self and surroundings; (c) occasions of real emotion toward the representation per se, especially one whose power is recognized; (d) flashes of involuntary, “Darwinian” emotion toward entities one “Darwinianly” believes to be present; (e) activation of real emotions toward existent individuals or kinds related to a given character, which partiallyfuse with and colour the makebelieve emotion sustained toward the character; and possibly even (f) phases of real emotion directed at the idea or thought ofa character (see Carroll, Horror, chap. 2). In addition to (a), Walton is willing to credit only (b) and (c). But (d) and (e), I suspect, have more potent ancillary explanatory value than either of those.