The Stranger discusses images explicitly beginning at 235b8ff. The relevant passages are worth quoting in full (235d-236c). In Fowler's translation, from Perseus:
Stranger
I see the likeness-making art as one part of imitation. This is met with, as a rule, whenever anyone produces the imitation by following the proportions of the original in length, breadth, and depth, and giving, besides, [235e] the appropriate colors to each part.Theaetetus
Yes, but do not all imitators try to do this?Stranger
Not those who produce some large work of sculpture or painting. For if they reproduced the true proportions of beautiful forms, the upper parts, you know, would seem smaller [236a] and the lower parts larger than they ought, because we see the former from a distance, the latter from near at hand.Theaetetus
Certainly.Stranger
So the artists abandon the truth and give their figures not the actual proportions but those which seem to be beautiful, do they not?Theaetetus
Certainly.Stranger
That, then, which is other, but like, we may fairly call a likeness, may we not? [236b] And the part of imitation which is concerned with such things, is to be called, as we called it before, likeness-making?Theaetetus
It is to be so called.Stranger
Now then, what shall we call that which appears, because it is seen from an unfavorable position, to be like the beautiful, but which would not even be likely to resemble that which it claims to be like, if a person were able to see such large works adequately? Shall we not call it, since it appears, but is not like, an appearance?Theaetetus
Certainly.Stranger
And this is very common in painting [236c] and in all imitation?Theaetetus
Of course.Stranger
And to the art which produces appearance, but not likeness, the most correct name we could give would be “fantastic art,” would it not?Theaetetus
By all means.Stranger
These, then, are the two forms of the image-making art that I meant, the likeness-making and the fantastic.
Here is part of the Greek:
Ξένος
...
μίαν μὲν τὴν εἰκαστικὴν ὁρῶν ἐν αὐτῇ τέχνην. ἔστι δ᾽ αὕτη μάλιστα ὁπόταν κατὰ τὰς τοῦ παραδείγματος συμμετρίας τις ἐν μήκει καὶ πλάτει καὶ βάθει, καὶ πρὸς [235ε] τούτοις ἔτι χρώματα ἀποδιδοὺς τὰ προσήκοντα ἑκάστοις, τὴν τοῦ μιμήματος γένεσιν ἀπεργάζηται.Ξένος
οὔκουν ὅσοι γε τῶν μεγάλων πού τι πλάττουσιν ἔργων ἢ γράφουσιν. εἰ γὰρ ἀποδιδοῖεν τὴν τῶν καλῶν ἀληθινὴν συμμετρίαν, οἶσθ᾽ ὅτι σμικρότερα μὲν τοῦ δέοντος [236α] τὰ ἄνω, μείζω δὲ τὰ κάτω φαίνοιτ᾽ ἂν διὰ τὸ τὰ μὲν πόρρωθεν, τὰ δ᾽ ἐγγύθεν ὑφ᾽ ἡμῶν ὁρᾶσθαι.
The distinction the Stranger makes here between likeness-making ("eikastic" imitation) and the making of appearances ("fantastic" imitation) has been the source of endless confusion. Among earlier interpreters, Philip (1961, p. 459, JSTOR), for instance, makes the strange suggestion that “[e]ikastike would appear to be a class without members, serving only a purpose of symmetry.” His argument relies too much on the Cratylus, where the distinction does not appear, though Socrates does discuss images. Cornford (1935, p. 198) makes the equally misguided suggestion, under the influence of arguments in the Republic (where, again, the distinction does not occur) that “[b]oth here and in the Republic the whole of fine art, considered as ‘imitative’, falls under the art of making ‘semblances’, not ‘likenesses’. Plato does not mean that there is a good and honest kind of art which makes ‘likenesses’ reproducing in all three dimensions and the natural colors of the original” – something which in fact he must mean.
More contemporary interpreters do better, though there is still no consensus regarding the meaning of the distinction. Blondell (2002, pp. 365ff.), for example, argues that εἰκόνα are “analytic images,” but she mars her argument by suggesting that only eikastic imitation is pedagogically beneficial. Like her, most interpreters of the distinction (so Notomi 1999, Palumbo in the Proceedings of the III Symposium Platonicum) operate under the assumption that Plato must obviously value eikastic imitation more than fantastic, or clearly identify philosophy with eikastic rather than fantastic imitation (Nightingale 2002 is an exception: she argues that Plato could have used both forms of imitation for pedagogical purposes, though her views are dismissed by Rutherford in the same volume). But the assumption is unwarranted; the Stranger says little about whether we are to consider eikastic imitation the sole province of the philosopher, or to value it always more than fantastic imitation.
What, then, is the Stranger trying to say with this distinction? I draw here on Benardete 's somewhat opaque, but useful interpretation of this passage (1984, pp. II.105-112). I take it that the Stranger is suggesting that there are two ways in which images can represent objects, and that both of them are flawed, though in different ways. Let's take, for example, the image of a straight road:
This is a "fantastic" or distorted image of the original: in order to represent the road so that it looks like a road, the picture had to distort the actual proportions of the road, primarily by making the parallel lines of the road appear to converge. The image's summetria or proportions are thus different from the summetria of the original, or more precisely, the image and the original do not have a common measure (a summetria). This was necessary because the road is "too large," as the Stranger notes, to be eikastically represented and still look like the road looks.
But what would an eikastic image of the road "look like"? Here is one (not, of course, of the same road):
Though this particular image is not eikastic in every respect described by the Stranger, it is eikastic in the sense that the image preserves the proportions of the original: the sides of the road really are parallel to each other, and they are depicted as parallel in the image. But of course, because from our normal point of view the road does not look like that, a fantastic image is required - a perspectivally distorted image - if we want the image of the road to look the way in which a road would normally look to us.
Fantastic images are distorted - and thus flawed as representations of their originals - precisely in order to account for our flawed perspective. Eikastic images are not distorted from the point of view of the original (that is, they preserve the original proportions of the object) but they appear flawed because they are not well adapted to our point of view. The relationship between an eikastic image and its original is like the relationship between the central circles in the Titchener circles illusion:
Both red circles are in fact identical in size, but they look different to us given the limitations of our perceptual apparatus.In the Greek context it seems likely to me that the Stranger is referring to a form of perspectival painting, where figures have to be systematically distorted in order to look right to us. The point of this distinction, however, is the subject of an upcoming post...
