For me, section 17 "Of the Ideal of Beauty," has always been one of the most difficult passages in the Critique of Judgment. (I am using the J. H. Barnard translation, Hafner Press, 1951). But I think I have found the key to its understanding. My current take on it is that it contains important intimations of what will happen later on in the book, particularly in the section on fine art, and can be seen as the entire CJ philosophy in a nutshell. I also see it as both taking into account cultural relativism but also seeking to transcend it by way of the notion of a rational ideal of beauty that goes beyond mere empirical considerations. It is consonant with his egalitarianism, and it is noteworthy that what would normally be taken to be the paradigm of a western-dominated conception of universal beauty in the sculpture by Polycleitus, which itself was taken as "the canon" in Ancient Greece, is rejected, or at least downplayed, as nothing more than something at the level of the empirical. One can also see here an anticipation of Hegel since the classical, as represented by Polycleitus, and humorously, perhaps, by Myron and his cow, is sublated in the last paragraph by the rational ideal which would be better represented by what Hegel would call "the Romantic," for example the self-portraits of Rembrandt, in which the inner spirit and inner beauty shows through, overcoming that which is merely average based on a mere mechanical combination of elements. When I discuss the concept of beauty in regards to humans the notion of inner beauty being more important than mere external beauty: this section exemplifies and elaborates on this perception, but also in terms of a creation of an ideal beauty within oneself, a kind of aesthetic self-molding that we elaborated later in Nietzsche, Foucault and Shusterman. Perhaps molding an inner beauty within oneself allows for perception of, and judgment of, inner beauty in others and in their representations.
The section (#17) begins by reiterating that there can be no objective rule of taste, and yet the sensation of satisfaction or dissatisfaction attending the object considered beautiful or ugly is universally communicable in the sense that we can expect, or even demand, that everyone experience it as being beautiful or as having the same attendant sensation. Kant suggests that it is hardly even probable that this accord will exist in all times and ages, but it ought to exist based on "deep-lying grounds of agreement in judging of the forms under which the objects are given" these grounds being the way that the faculties of the imagination and the understanding respond to forms that are designed or look designed and thus have a certain purposefulness of look. He then says that we consider "some products of taste as exemplary." Those who have already read the book may not that the word "exemplary" is also found in the the section on fine art. Also, as with the genius later, it turns out that taste cannot be acquired by imitating others, but must be "an original faculty" and that, although imitation of a model may show skill, taste requires ability to "judge the model itself." Again this reminds us of how genius will later be described as creating its own rules. He then says that "the highest model, the archetype of taste, is a mere idea which everyone must produce in himself and according to which he must judge every object of taste" etc. He is saying that the archetype or ideal is something that we create within ourselves, and is not, for example, simply a set of works that have stood the test of time, or even some paradigmatic works set up by a particular culture. The level of autonomy given to the good critic is stunning. Again, this reminds us of the ability of the genius to create his or her own rules or even his or her own world. Creating the internal model of beauty is interestingly like creating the original work of fine art, one that exhibits what Kant calls "aesthetic ideas." Then it turns out that this idea is a rational concept which is correlated with an ideal which is itself a "representation of an individual being" seen as implied by the idea. The ideal depends on the "indeterminate idea that reason has of a maximum," which is to say that it depends on the notion of "the best." So we create within ourselves an ideal of beauty which is connected to the ideal notion of the best. We are not in immediate possession of such an ideal, but we "strive to produce" this ideal of imagination in ourselves.
Then the question is whether we arrive at such an ideal, a priori or empirically. Kant's answer will be in a sense both. But the a priori method is much more important. Kant then surprisingly tells us that the beauty "for which an ideal is to be sought" must be "fixed by a concept of objective purposiveness." This is a paradox as it seems to go against everything he has previously told us about beauty. As a result, it is not aesthetic as ordinarily understood. Kant had already undercut or modified his original idea of the aesthetic as pure by introducing the notion of appendant beauty. Now he indicates that the ideal of beauty "cannot appertain to the object of a quite pure judgment of taste" as it is "in part intellectual." There has to be an "idea of reason in accordance with definite concepts" that lies at its basis. And this idea determines its internal possibility a priori. Hence there cannot be an ideal of beautiful flowers, furniture or views, nor even an ideal of appendant beauty as in a beautiful house, tree or garden, since these are nearly as vague as free beauty. It must be related to something the purpose of its existence is in itself, i.e. man, since man can "determine his purposes by reason" and this is why man is alone "susceptible of an ideal of beauty," given that "only humanity in his person, as an intelligence...is susceptible to the ideal of perfection."
This has always puzzled me: why turn to man as the ideal of beauty? Why introduce the notion of man as the ideal of beauty in this book or at this point? The reason seems to be that only man is connected in this way to the supersensible realm, the realm of soul, God and immortality, the realm that needs to be supposed but which cannot be proved to exist. Now, Kant continues, there are two "elements" in all of this, the one being the "normal idea" and "the rational idea" i.e. two ideals of beauty. He spends a lot of time on the normal idea, but what he is really interested in is the rational idea to which he devotes a short paragraph at the end. I take the normal idea to be a not-to-useful empirical approach to the ideal of beauty in man. It is the approach that most people take to the ideal of beauty, and it is superficial. The normal idea is "an individual intuition (of the imagination), representing the standard of our judgment (upon man) as a thing belonging to a particular animal species." We know Kant would be disinclined to take it seriously since it treats humans as merely animal.
The "rational idea" by contrast is one that focuses on the "purposes of humanity" which cannot be fully represented in sense but the phenomenal effect of which can be revealed. The normal idea is taken from experience. By contrast, when we are talking about the purposes of humanity or its purposiveness, a purpose to which "only the whole race [of humanity] and not any isolated individual is adequate" the rational idea of this "lies merely in the idea of the judging subject." It is this "aesthetical idea" which can be represented in a model, i.e. in a work of art, and presumably fine art, although he is not supposed to be talking about that in this part of his book. I take an example of such a model to be a work of art that represents humanity by way of representing an individual human, say a self-portrait by Rembrandt.
Where Kant may go wrong to some degree in this section, or perhaps he is merely being intentionally devious or playful, is in saying that we can make "how this comes to pass" intelligible through a "psychological explanation." Let us assume that he is just being tricky, for the next paragraph, the penultimate in the section, is almost a joke. In this paragraph he describes the imagination as reproducing "the image of the figure of the object" from a great number of objects of different kinds or even the same kind, and, by comparison, unconsciously letting "one image glide into another," thus coming up with an average which can serve as a common measure. For example, we see a thousand full-grown men (as all adults have seen at least), compare their sizes, and judge a "normal size" by way of letting the images of these "fall on one another." And so "this is the stature of a beautiful man," except that how can anyone take seriously the idea that a mere average size is what makes a man beautiful or have a beautiful stature? The idea of equating the beautiful with the average is absurd. Moreover, size has little to do with beauty anyway. Kant, in short, is messing with us. The normal idea is purely mechanical. It is as though we were dealing with the philosophy of Democritus, rather than Kant. The "various impressions of such figures on the organ of internal sense" through a "dynamical effect" produces the norm. As he admits, the result can be arrived at mechanically "by adding together all thousand magnitudes, heights, breadths, and thicknesses" and dividing by the sum of one thousand. The average man, based on all of this applied to all of his parts, is the "normal idea."
But then it turns out it is the normal idea "in the country where the comparison is instituted." The possibility of relativism in beauty raises its ugly head. The "normal idea" on this culturally relative interpretation or variation implies that the "Negro," "European," and "Chinaman," would all have a different normal idea of beauty, something which may be true empirically, given that what we see as physically beautiful in a human is what based on what we are used to looking at plus the fund of judgments of beauty we heard when growing up.
Kant then seems to transition from this culturally relative moment to "the whole race" by which he means humanity itself, but still talking in terms of the normal idea. The normal idea for the whole race "is the image for the whole race," and it is this idea "which nature takes as archetypes in her production of the same species" but which is not reached in any individual case. (Why Kant should intrude the intentions or plans of nature here is not clear.) But we know we have not gotten far here since Kant insists that this is "by no means the whole archetype of beauty in the race," and in fact is only a necessary condition for correctness in the mental presentation of humanity. The Doryphorus of Polycleitus and the Cow of Myron are examples he gives of such correct representations, but ones which do not reach beyond the normal idea. Kant emphasizes this by ending the paragraph with reiterating that the presentation of this sort "is merely correct." Moreover, the representation does not please by its beauty, he says, but simply by not violating the condition of correctness necessary for beauty. It is not, in short, sufficient for beauty. So we need to move beyond that which is merely considered racially or ethnically correct, or based on the dynamic psychology of the imagination in its associative mode, to the rational (and perhaps to imagination in its productive or creative mode).
The last paragraph contains the key to the section. As Kant previously observed, we can only expect the ideal in the human figure (as opposed to any other part of nature or in the human artifactual world which could be represented), and this ideal "consists in the expression of the moral" which provides the basis for the object itself pleasing universally, and not just as being merely correct. What we are looking for is "the visible expression of moral ideas that rule men inwardly." Although we can get this from experience, "the idea of highest purposiveness" in man found in such virtues as "goodness of heart, purity, strength, peace" is only visible in the body or portrait through a "union of pure ideas of reason with great imaginative power." This is precisely what we find later that the genius artist is able to accomplish with his or her aesthetic ideas. This further step away from the empirical is required not only in the object itself but in the perceiving subject and, Kant insists, in the creator of the work, i.e. "still more in him who wishes to present it." We know finally that this ideal of beauty is correct, or rather escapes the realm of the merely sensible, if the satisfaction is not infected by "sensible charm."
Now this judgment is not "purely aesthetical" since we take a "great interest" in it because of its connection with morality. But this is not a bad thing for Kant.
The great irony of this section is that it is immediately followed by the moral of the entire Third Moment, the "explanation of the beautiful derived from this Third Moment," which is that "Beauty is the form of the purposiveness of an object, so far as this is perceived in it without any representation of purpose." The irony is that this explanation has been transcended, bracketed, or perhaps even sublated, since the ideal of beauty does involve purposiveness, even more than dependent or appendent beauty, which was an odd enough exception earlier on. Here there is a representation of purpose, although not a logical one. Rather, it is the way that the moral qualities, or inner spirit, shines through some great portraits, for example. I find myself here thinking of Davide's Oath of the Horatii (1785) where the virtue of courage shines through action, but also, perhaps more appropriately, the portraits of Rembrandt, where an inner light of personal character, sometimes of tragic sadness, as well as deep humanity, seems to shine through. As said, previously, this is similar to what Hegel refers to as "The Romantic."