Aesthetic properties as value-justifying

The phenomenon of aesthetic properties remains deeply puzzling – at least to me. But apparently also to the aesthetic community at large. There is no agreement about either their existence or their nature, whether they are subjective or objective, whether they apply to art works only or to the world at large, and so forth. We have interesting but not quite convincing arguments for their reality 1. and we have doubtful attempts to say what sort of properties they are. 2. We have various standard stories, but to every page of those stories we have dissents. One chief symptom of the mystery is the striking absence of any detailed analysis or elucidation of or negotiating disagreements concerning particular aesthetic properties. Our references to them are mainly one-liners, so to speak, not serious and sustained efforts to see into them. In an effort to make some slight headway in lessening the mystery I am proposing what seems to me a promising but admittedly speculative characterization of a considerable number of them. I want to see how far I can ride this leading idea, in hopes of pushing back the obscurity at least a little.


I start with some ground rules. The aesthetic properties of immediate interest to me are not what Eddy Zemach calls pure aesthetic properties, namely those that are purely evaluative, but those that are totally or mainly descriptive. I am encouraged by Sibley's distinction between the descriptive and evaluative use of aesthetic predicates to believe that in many cases the evaluative implications of an aesthetic predicate can be neutralized, so that the term can refer to a purely descriptive aesthetic property. Further, I assume that aesthetic properties of all sorts can be straightforwardly asserted, whatever other uses their predicates may also serve, such as inviting others to share the speaker's perceptions or preferences. Thirdly, I exclude from consideration properties of actual psychological effect, such as frighteningness or comfortingness, which I take to be subjective relations not properties of objects. For simplicity in the present discussion I will use “aesthetic property” exclusively for those that are wholly descriptive, unless otherwise indicated.

Finally, I assume that a principal constraint on aesthetic properties (so delimited) is that they provide reasons in support of aesthetic judgments. This may be expressed in terms of their relation to purely evaluative aesthetic properties. For simplicity I will refer to the positively evaluative aesthetic properties as “beauty” and the negatively evaluative ones as “unbeauty.” So my chief question becomes, what must be the case for aesthetic properties to confer beauty on an object in the sense of being beauty- constituting , which is equivalent to their being (part of) what is beautiful about the object, which in turn is for them to be beautiful properties of the object. 3. Such beauty claims are not limited to overall beauty; rather the normal claim will posit beauty with respect to some property: beauty with respect to harmony, organization, expression, etc. I shall call this justificatory constraint the core requirement any aesthetic property must meet.


A first-stage task is to determine what traits aesthetic properties must have in order to have the possibility of satisfying the core requirement.

One widely held view concerning aesthetic properties is that they are response-dependent properties where the response does not reduce to a simple sensory response, for instance an experience of color, tone, feel, odor, or taste. 4. Does this limitation follow from the core requirement? Pace Eaton (Eaton 1994) I wish to support the view that it does, making use of ideas in Guy Sircello's underappreciated A New Theory of Beauty. Citing the sensory color of anything as beauty-constituting 5. must always fail because even if the yellow color is implicated in its beauty, its justificatory force comes by way of a degree property which it possesses. The yellow is beautifully vivid, radiant, lemony or mellow. Vivacity, radiance, tartness, mellowness are candidates for constituting beauty because they can be present to an eminent degree. Yellowness is not because it cannot be present to an eminent degree, not being a property of degree at all. Beauty requires a superlative (beautifully vivid implies highly vivid), hence only a degree property can be beauty-constituting.

For this reason some of Eaton's other examples, e.g. of a picture's representing Aunt Mabel, are also disqualified by the core requirement. A picture cannot be beautiful merely because it represents Aunt Mabel or Perseus rescuing Andromeda or any other subject. It can be representationally beautiful only by dint of representing her, or him, beautifully. And that will take a property of degree, say, representing Aunt Mabel insightfully or Perseus' rescue dramatically. To vary the expression, it will require the representation to be beautifully insightful or beautifully dramatic.

Sircello is also on the right track in holding that degree-properties like hotness or largeness when construed quantitatively cannot be beauty-constituting in and of themselves. For nothing can be beautifully tall or swift-footed where the height is construed in meters or the swiftness in feet per second. For this reason a building cannot be beautifully proportioned because it embodies golden sections, in spite of all the popular aesthetic mythology down the centuries. There is no such thing as being beautifully compliant with a rule, even if the golden section is in its way a beautiful mathematical relationship. Similarly a bull cannot be beautifully large, nor a miniature whippet beautifully small. What seem at first sight quantitative beauties turn out to be qualitative ones. The miniature whippet is beautifully delicate or doll-like, the beautiful bull beautifully monumental or masterful-looking. The Palladio villa is not beautifully golden-section-compliant but beautifully harmonious in respect of its proportions. 6.

Thus far Sircello's restriction of beauty-constituting properties to properties of qualitative degree, as he calls them, is sound. It states a necessary condition for any beauty-constituting property. Sircello also makes another valid point to which I alluded in passing a moment ago. For a property of qualitative degree to be beauty-constituting it must be present to a high degree. The same is true mutatis mutandis of properties that are ugliness-constituting. 7.

Yet not all properties of qualitative degree are beauty-/unbeauty-constituting when present to a high degree, because some are quite powerless to justify an aesthetic judgment. One such is an example used by Eaton for another purpose: shaped like a duck . This is certainly a property of qualitative degree, yet nothing can be beautiful or unbeautiful simply because it possesses this property to a high degree. The reason is, I think, illuminating. The problem lies in the fact that a thing can be shaped exactly like a duck. Hence it cannot be beauty-constituting. If it could, everything would be beautiful of its specific kind in being shaped precisely as it is. And surely nothing can be beautifully shaped merely because it is shaped in exactly the way it is: that property is trivially possessed, being a consequence of self-identity, and no beauty-constituting property could be in this way trivial.


The mention of resemblance brings to the fore what I want to suggest is a central trait of aesthetic properties (under the limitation imposed at the beginning). All aesthetic properties involve, I believe, a certain sort of resemblance. That resemblance plays some role is generally admitted in the case of some. It certainly is true of Scruton's class of aesthetic properties of comparison: “a writer's style [described] as bloated or masculine, a colour as warm or cold, a piece of music as architectural.” (Scruton 1974, 30-31). Comparison is seemingly also involved in several of Scruton's other classes: those of technical accomplishment (being balanced, economical, skilful); the “human” properties, as Beardsley calls them, of mental and emotional life (being joyful, agitated, intelligent), a category which in my view includes expressive properties of art works. Similarly most of Alan Goldman's formal, emotional, behavioral, evocative, representational and second-order perceptual terms can readily be construed as involving comparisons.

Such resemblances are subject to a distinctive constraint: the most eminent degree of resemblance to which they can attain never amounts to qualitative identity. Why so? I think the answer is that the resemblances link discrepant categories. Thus an undoubted aesthetic property, the swiftness of a line in the graphic domain, call it graphic or linear swiftness (as in the following images of the swan, the maid and the forest 8.) is such that there is no meaningful way to raise the likeness to the point of identity or even indiscernibility relative to any of the types of real swiftness that supply relevant aesthetic resonances (pen strokes, hurled projectiles, etc.). Similarly the resemblances that give meaning to hot pink and cool jazz, are resemblances across categories . 9. In consequence, the idea of altering the color until it attains an exact equivalence to a given degree of heat is empty, absurd. Thus also the ideas of visual designs being exactly as balanced as a balance scale is, or a musical composition being exactly as sad as a person is, are conceptually outlandish. Rather in all these cases the likeness is a matter of engendering resonances or echoes from other domains than that of the object to which they are ascribed. The property in the resemblant domain is one that as a whole does not and cannot hold literally in the domain to which the present object belongs. Thus metaphor has been frequently invoked as shedding light on the character of aesthetic properties. 10. The lines that cannot be literally be swift are nonetheless as-if swift. In a classic formulation, swift lines relate to other lines as swift movements relate to other movements. 11.

Aesthetic resemblance is also, evidently, a resemblance of incommensurables. Such incommensurability largely accounts, I believe, for the elusiveness of aesthetic properties of which Pettit (1983), for one, makes so much. I think it also helps explain how aesthetic properties justify aesthetic judgments. More of this in a moment.


First, more needs to be said about the ontology of the resemblance. Evidently there is some literal, which is to say physical, resemblance between the lines in the drawings just shown, and the manifold sorts of motions that constitute the resemblant-referents of the aesthetic predicate (falling stars, hurled missiles, etc.). But it is only the tracks or paths of these motions that literally resemble anything in the drawn lines, just as it is only certain components of a sad behavior-syndrome that music can literally resemble. No wonder, because these resemblances hold within a single category wide enough to embrace line and star or missile track, or musical and human movement.

But these literal resemblances can't possibly suffice for the aesthetic effect, since to every track of a swift motion there is a track of a slow one that is precisely identical to the former. So no such literal resemblance can explain the quality in the domain of lines that corresponds to swiftness in the domain of motions. 12.

The aesthetic effect expressed by calling a line swift, in contrast, links the line with properties of movement itself, the speed of the pen stroke or projectile, as well as with its energy and much more. That is, the linkage extends to what cannot be literally resemblant. Or the resemblance imputed is more robust than the difference of categories permits. Thus, taken in its full character, no such resemblance can actually obtain, precisely because of the categorial discrepancy. What does obtain is an impression of such a resemblance. The line impresses us as being swift or slow in ways or to an extent to which it cannot possibly be swift or slow. The resemblance is thus merely intentional – part of the intentional content of the impression. 13.


If this model is to be applicable to all aesthetic properties, many apparent counterexamples must be shown compliant with it. Will it fit gracefulness, shorn of its common evaluative aspect? I think so, but I have no time to argue the point. 14. Instead I wish to move to something of even more moment. Is the fact of a property of qualitative degree being beauty-constituting assured by its phenomenological side being an impression of cross-categorial resemblances? Is the (extreme) “swiftness” of the lines in the drawing of the swan necessarily beauty-constituting? Is the (extreme) “hotness” of a pink or “coolness” of a blue? Can't we easily imagine merit-neutral cases, or even ones that are ugliness-constituting?

To these questions there are answers of two sorts. The most accessible explanation of why the resemblance is beauty-constituting is that such resonances, as I have called them, are inherently attractive, wondrous, so to speak. Our capacity to form such impressions gives a wondrous richness to our experience. The sensory property when invested with the resemblance is to that extent admirable, regardless of any other merit. This constitutes a default beauty independent of any further grace, as general in its way as Kant's free beauty.

But second, there is a much more material beauty in the particular case because of the sort of graphic swiftness of the lines, eminently in the case of the swan. It is a decisive, firmly controlled and yet exuberant swiftness. The full ensemble of resemblances picks up on and reinforces the alert and almost martial air of the depicted bird, which is proud and forward (if crossed, froward!). We are likely to feel that these qualities are admirable ones, such as are worthy of being welcomed and enjoyed. Thus it becomes unsurprising that we are able to find beauty in the lines, participating as they do in this larger web of agreeable cross-categorial resemblances.

This sort of explanation has a number of advantages:

1. It offers a way to separate non-aesthetic from aesthetic properties, the base from the supervenient 15.. The base consists of the intrinsic properties of the object -- the curvature, thinness, evenness, etc. of the lines, some of them forming a swan-representation, others a surround of a specific sort, and so forth -- plus the intrinsic properties in the large class of resemblant properties in the other domains, for instance the properties of actually moving things, inanimate and animate; plus such literal resemblances as hold between the image and the referents. The phenomenological side of the supervenient property is the impression of more or less “fanciful” resemblances, the resemblances that defy measure and exact ranking, are typically somewhat nebulous, impressionistic, or hyperbolic. The aesthetic property is precisely the capacity of that base to engender a condensed crowd of such resemblance-impressions. To be sure the experience is not clearly articulated. The resemblances are not held in sharp focus. A quotient of fuzziness is essential. But when we search for the secret of the line's swiftness nothing else can be found to elucidate it. Since the role of the imagination is larger in aesthetic than in ordinary perception, the almost universally cited requirement of a special sort of perception is readily accommodated without recourse (so far forth) to an occult faculty.

2. It explains the particularity of aesthetic properties so often embraced, notably by Sibley (1973/2001). Any variation in the merit-responsible base will vary the ensemble of cross-categorial resemblances, hence the particular merit-constituting, aesthetic property.

Yet more constructive, in my opinion, is the fact that this conception suggests a way to analyze or elucidate particular aesthetic properties. To unfold the content of such a property is to tease out the cross-categorial resemblances. Even if this can never be completely accomplished, even if our impressions always precede and outrun our account, and even if the prominence or reclusiveness of given strands of resemblance cannot be sharply assessed, we do find, I think, many of the resemblances accessible and nameable as well as gratifying to apprehend. Our aesthetic cognition can be rendered somewhat less nebulous without the nebula of the whole ever being entirely resolved. 16.

3. The present conception also offers, I think, the possibility of aesthetic perceptions being subject to valid criteria of accuracy/inaccuracy of perception. For first, the impressions of cross-categorial resemblance are founded upon real or literal resemblances even though these are stretched beyond the literal in the ultimate impression. Thus it is possible to weed out many perverse attributions because there is no resemblance-base for them. For instance if a person ascribed cruelty to the lines. There is simply no resemblance in the lines to any known aspect of cruelty. Second, where the available comparison base offers a handle for a resemblance-impression and it is not taken up by a viewer, we can ascribe a deficiency of imagination. The viewer's experience may be valid as far as it goes, but may not go far enough. And more constructively we can validly put more reliance on percipients who have a greater capacity to detect cross-categorial resemblances than do others. Our optimal aesthetic property discriminators must qualify in this regard, and such convergence among them as we can find will serve as our best standard for the reality of the aesthetic property. 17.

4. The variation in perceptiveness of aesthetic properties in different cultures can be readily accommodated without its impugning the reality of the properties. 18.


All that having been said, it remains to explain the positive and negative evaluative force of different aesthetic properties. Does the present conception suggest a way to do that? Why are some aesthetic properties beauty-constituting and others unbeauty-constituting whereas the base properties by themselves are not (they are at most beauty-responsible)? I accept the idea that to be beauty-constituting is deeply connected with basic human needs, particularly cognitive needs. Unity is beauty-constituting because (or when) it makes patterns easy to cognize, whether perceptually or conceptually, in relation to a given level of complexity or disorder. Variety or complexity connects with cognitive challenges, that is, with types of intelligibility that are hard won and therefore all the more gratifying. Intensity connects with clarity, other things being equal. This is recognized by virtually all. In aesthetic cognition, as here conceived, unity, variety and intensity pertain to the aesthetic suggestiveness (the cross-categorial suggestiveness) of objects or events. A special challenge is posed by the distinctive phenomenological “density” of resonances; and when the object is organized in such a way as to enable us to rise to that challenge, the value is positive. When the challenge cannot be met because of the disorganization of the object, because the resemblances awakened by its different strands are disorderly, too little various, or lacking in intensity, the result is unbeauty, and at the extreme, ugliness. 19.

Well and good, but how do the resemblances I am making so much of come into it? They do so, I suggest, in considerably increasing the harmony or conflict in our cognitive systems compared to what the base properties do. It is not only the physical and merely sensory character of the line that is in view, but the vastly richer nebula of aesthetic resemblances, the potential for satisfaction and dissatisfaction is much enhanced. It is not that the potential is absent in the subaesthetic. Ordinary perception of the world provides a measure of the same sort of satisfaction and dissatisfaction even without aesthetic resemblances coming into play. In normal perception we appreciate the intelligibility of the world and depreciate the many momentary and the relatively few enduring unintelligibilities that afflict us. The difference, so far as I can see, is one of subtlety or playfulness or a distinctive sort of depth that can be found in aesthetic perception.


1. Zemach and Carroll justify the objectivity of aesthetic properties in part by the argument that so much agreement exists among qualified observers under optimizing conditions that aesthetic predicates, being learned ostensively, cannot be reasonably regarded as misused. Levinson's considerably more nuanced realism is similarly based. Even if this argument is sound, there remains the puzzlement about what the content of such terms is, that is, what complex property a given aesthetic predicate designates. It is the content I am most baffled by and think we stand to gain most from shedding light on. Only if we become clearer about that can we understand why aesthetic properties validly sustain evaluative judgments.

2. Zemach's proposal, similar in scope but not in content to mine, is that aesthetic properties are secondary properties ‘tinged by desire.' This interesting thesis is not easy to work out in detail, but in a forthcoming paper I try my best to do it. I believe that the thesis must be modified in a number of ways before it will survive objections.

3. Sibley's articles from the classic 1959 ‘Aesthetic Concepts' on are replete with references to likeness-references of one sort or another. The prevalent association of aesthetic concepts with ways things look has an evident connection with resemblance. Metaphorical uses, evidently resemblance-based, are central to Nelson Goodman's theory of the aesthetic. My reliance on resemblance is not intended to deny Goodman's assertion, “Metaphor permeates all discourse, ordinary and special….” Goodman 1968, 80. I am contending that a special sort of likeness is central to (descriptive) aesthetic predicates, but at the same time I believe that aesthetic properties figure pervasively in “discourse, ordinary and special.” See also Berys Gaut 1997.

4. One might wonder whether there is always as great a crowd of resemblances underwriting an aesthetic property as the preceding account supposes. What of hot pink? Is that not just a resemblance of a certain highly saturated and light shade toward the blue side of middle red, on the one hand, and felt hotness: the pink looks the way the heat feels. But I think the more aesthetic is the impression, the more other resemblances come into play: the sense of vivaciousness, daringness, flamboyance, and who knows what else. The question is not merely what a given impression contains or foregrounds but what the impression suggests, and I think it suggests more. Realize also that the idea of the aesthetic property of the object transcends actual impressions; it includes those that are latent in the disposition, i.e., all those that are realizable under optimizing conditions.

As to ostensibly non-comparison properties such as grace or awkwardness, I can make these suggestions. Gracefulness of bodily posture is never, I think, attributed without the suggestion of a resemblance to grace of movement, grace of gesture, and manifold other domains (e.g. gentility). They in turn contain resonances from their counterpart domains. There are also anthropomophic suggestions, for instance, of different parts of the body acting in willing concert with others. Similarly awkwardness of movement contains resonances of an anthromorphic (and more generally biomorphic) sort, of different systems perversely acting contrary to each other, defeating the whole.

5. One may hope that different species of a general aesthetic property, say gracefulness, can be delineated by teasing out the dominant resemblance-resonances on the phenomenological level. Is there any other way to get inside these perceptual concepts?

6. Clearly variations in ascription of an aesthetic property will be affected by what impressions of cross-categorial resemblance a person has. This is often taken as a reason for denying the possibility of true supervenience, or true objectivity by any other means, e.g., by a causally reliable relationship even under optimizing conditions. In the controversy between Goldman, Vaida, et al and Levinson, Carroll, et al on this point, I will just say that I side with the latter. At the very least, the evidence for the negative is by no means decisive. Also I believe that for my purposes the best working hypothesis is the realist one. We are, I believe, more likely to discover things about the content of aesthetic properties if we have not given up hope of realism being ultimately vindicated.

7. One trait of many, perhaps all, aesthetic properties is vagueness, and it is of the first importance to try to establish what level of discriminability a given aesthetic property admits. In my research into Mondrian's asymmetric balance I find reason to reject claims that the balance is perfectly precise, exact, exquisite. The level of discrimination among Mondrian experts simply does not sustain any such claim. This underlines a very frequent defect of aesthetic attributions: egregious exaggeration. See “Mondrian's Balance” on my personal website: .

8. Cross-categorial resemblances must not be confused with synesthesia. The latter is strikingly independent of resemblance between the trigger and the response. For example, letters and numerals are perceived as colored without there being any evident resemblance.

9. The most plausible explanation of cultural differences in aesthetic perceptiveness is in terms of specialization. Cultures specialize in a limited range of aesthetic effects, the ranges differing from culture to culture. Overall two cultures of interestingly different sensitivities may be equally perceptive. But there is no reason to rule out overall differences as well, as where one culture cultivates many sensitivities whereas the other cultivates only a few or even none to a comparable degree. Realism about aesthetic properties is challenged by such diversity only if cultures of comparable sensitivity with respect to a given range differ in the aesthetic properties they ascribe within that range.

10. When this paper was presented at the ASA Eastern Division meeting on April 9, 2005, the commentator, Carol Gould, raised the question of whether resemblance, being symmetrical, fits the case of aesthetic properties. Applying this to our examples, we might doubt that a color being perceived as hot implies that hot things are perceived as hot-colored, or that a line being perceived as swift implies that swift things are perceived as line-like. I doubt that the impression of A's resembling B must carry within it the impression of B's resembling A. The intentionality of impressions seems to disqualify this inference. However, the impressions do seem to go together: just as highly saturated and light reds are perceived as hot, warm surfaces can be perceived as more like warm colors than they are like greens or blues and cannot be perceived contrariwise. If this is correct then a de facto symmetry obtains.


Brady, Emily and Levinson, Jerrold 2001.E. Brady and J. Levinson, eds. Aesthetic Concepts , Oxford : Oxford University Press, 2001.

Carroll, Noël 1999. Philosophy of art: a contemporary introduction. London and New York : Routledge, 1999.

Cohen, Ted, 2001. “Sibley and the wonder of aesthetic language,” in Brady et al.

Eaton, Marcia 1994. “The intrinsic, non-supervenient nature of aesthetic properties,” JAAC 52:4 Fall 1994 383-397.

Gaut, Berys 1997. “Metaphor and the understanding of art,” Proceedings of the Aristotelian Society, 1997, 223-241.

Goldman, Alan 1993. “Realism about aesthetic properties,” JAAC 51:1 Winter 1993.

_______1995. Aesthetic Value, Boulder , Col. :Westview Press, 1995.

Levinson, Jerrold 1983. “Aesthetic supervenience,” Southern Journal of Philosophy 22 Supplement (1983): 93-110.

_______1994. “Being realistic about aesthetic properties,” JAAC 52.3 Summer 1994 351-354.

_______2001. “Aesthetic properties, evaluative force, and differences of sensibility,” in E. Brady et al.

Mackinnon, John E. 2000. “Scruton, Sibley, and supervenience.” JAAC 58:4 Fall 2000 283-292.

_______2001. “Heroism and reversal: Sibley on aesthetic supervenience,” in E. Brady et al.

Pettit, Philip 1973. “The possibility of aesthetic realism,” in Shaper, Eva. ed., Pleasure, preference and value: studies in philosophical aesthetics. Cambridge : Cambridge University Press, 1983.

Scruton, Roger 1974. Art and imagination . London : Routledge and Kegan Paul, 1974.

Sibley, Frank 1959a. ‘Aesthetic Concepts,' The Philosophical Review 68. Reprinted with revisions in J. Margolis, Philosophy looks at the arts , 1962 and Sibley, Approach to Aesthetics 2001.

______ 1959b. ‘Aesthetics and the looks of things,' The Journal of Philosophy 66 (1959). Reprinted in Sibley 2001)

_______1973. “Particularity, art, and evaluation,” Proceedings of the Aristotelian Society, Supplementary volume 48 , 1973. Reprinted in Sibley 2001.

_______2001. Sibley, Frank, Approach to aesthetics: collected papers on philosophical aesthetics , Oxford : Oxford University Press, 2001. All page references to Sibley are from this volume..

Vaida, Iuliana 1998. “The quest for objectivity: secondary qualities and aesthetic qualities.” JAAC 56:3 Summer 1998 283-297.

Zemach, Eddy 1997. Real Beauty. University Park , Pa. : Pennsylvania State Press, 1997.

________2001. “What is an aesthetic property,” in Brady et al.


1. Notably by Noel Carroll 1999 and Eddy Zemach 1997.

2. The analysis by Eddy Zemach 1997 is an interesting new venture.

3. Cp. Sibley's distinction between ‘merit-constituting' and ‘merit-responsible” properties, i.e., aesthetic properties and their non-aesthetic base, respectively. Sibley 2001, 97.

4. Sibley, Levinson, Carroll and many, many others.

5. Beauty-constituting differs from causing beauty and giving reason to believe . It amounts to being that wherein the beauty lies . This requires that a B-constituting property be a beautiful property. Cp. Sibley's discussion of uses of beautiful and beautifully blue (Sibley 1959b2001, 30)

6. This limitation also excludes being mostly yellow, where that is a property of quantitative degree.

7. The intensity condition on beauty-constituting properties must not be misunderstood. It does not imply, pace Goldman's travesty of Sircello's view (Goldman 1995, 24), that only intensely vivid colors can be beautiful colorwise. The view is rather that beautiful vivacity must be high vivacity. Pastel colors are not beautifully vivid, but beautifully soft or delicate. The softness or delicacy is what is present in high degree. (Sircello 1975, 37)

8. Images of a windblown maiden and a forest scene are also given, in order of decreasing swiftness.

9. Many have acknowledged the role of resemblance without making it a criterion of aesthetic properties (Endnote 3).

10. An illuminating discussion of metaphor in this context is found in Berys Gaut 1997.

11. In discussion of metaphor a distinction is drawn between two and four-term formulations. The latter form takes A to be to B as C is to D. Cross-categorial resemblances have this form. Their peculiarity is that no full and coherent account can be given of the analogy of A to C except by the mediation of an impression. The specific way hot pink is like literal hotness is only that an impression of such likeness is created. The four-term form does not by itself impose that limitation. A being as strong for a human as Atlas is for a titan states a literal resemblance.

12. Granted, the fact of an objective resemblance does not by itself determine us to have an impression of that resemblance. But our selective receptivity to a given resemblance needs an explanation.

13..See Levinson 2001, p. 66n16 for variant ideas on the ontology of perceptual and aesthetic properties.

14. See endnote 3 for discussion of selected counterexamples.

15. Even if critics of supervenience, such as Goldman, prove right, still it separates elements of the causal chain, which in any case is implicated in the present conception.

16. This anatomizing occurs at two levels, in the base and in the phenomenological aspect. Experimenting with the base will often produce surprising results as to what actually affects the impression. But equally there are discoveries to be made by interrogating the impression with regard to the component resemblances.

17. I am far from supposing that it is easy to determine what is and what is not a genuine resemblance relative to a given aesthetic property. A major question is whether there is any disposition-independent criterion. After all, there is none such for key color properties, e.g., middle red, according to Clark . Clark 1993, 129.

18. See endnote 9 for a brief comment on this.

19. Zemach 1997, 2001 propounds a criterion of beauty that foregrounds cognitive facility. An additional factor in the beautiful is almost certainly the goodness of the properties constituting the resemblant-set in their own domains, the moral, practical, intellectual, etc. goodness. Sircello propounds as a condition on a beautiful PQD that it not be a property of defect, deficiency or lack, or the appearance of such. Similarly properties of defect (etc.) are ugliness-constituting. I believe some such condition operates in justifying aesthetic judgments. But I am not sure what role it plays.