As I mentioned in my previous post, the best evidence about Aristotle’s theoretical views about physiognomical inferences (if the passage is indeed by Aristotle) can be found at the very end of the Prior Analytics (in II 27, 70b7-38). For the application of these views, we have to turn to his History of Animals – a treatise in which Aristotle collects and organizes his data about the differences of animals in order to prepare for causal explanations at a later stage.
The collection starts by identifying the differences between animal parts (HA I 6, 491a14-16), and first among these the differences between the parts of humans, since those are best known to us (HA I 6, 491a19-23). The exposition starts at the top of the human body, and slowly works its way down. The first couple of chapters (HA I 8-11) discuss the human face and head, and this is where the physiognomical material comes in. I do not have space here to deal with the material extensively, so let me instead note four characteristics of Aristotle’s appeal to the science of physiognomy in these chapters, before ending my post in aporia.
First, Aristotle treats the physiognomical material as one kind of biological fact among many. The chapters list all kinds of information about the differentiations of parts on the head, using the human head as a guideline, including physiological descriptions of the part, the part’s technical name, its differentiations, and how those differentiations correlate with differences in the functionality of that part and with differences in the character of the being with that part. For instance, Aristotle states that, among the various eye-colors found in animals, greenish is ‘a sign of the best character and is strongest with regard to sharpness of vision’ (HA I 10, 492a3-4: ἐνίοις δὲ αἰγωπόν· τοῦτο ἤθους βελτίστου σημεῖον καὶ πρὸς ὀξύτητα ὄψεως κράτιστον). The physiognomical material thus does not receive any special treatment.
Second, most of the correlations that Aristotle notes in these chapters between facial differentiations and differences in character explicitly use the language of signs. For instance, differences in the nicks or corners of the eyes are signs: ‘when these are long, it is a sign of a bad character, when they are like combs and are fleshy, [it is a sign] of wickedness’ (HA I 9, 491b24-26: οἳ ἂν μὲν ὦσι μακροί, κακοηθείας σημεῖον, ἂν δ’ οἷον οἱ ἰκτῖνες κρεῶδες ἔχωσι τὸ πρὸς τῷ μυκτῆρι, πονηρίας). I only found one case where Aristotle does not use the language of signs (in HA I 8, 491b12-14: ‘this [i.e., the forehead] is big in some, and they are slower, small in others, and they are easily moved; and in some it is broad, and they are excitable, in others it is rounded, and they are spirited’). It also appears that Aristotle uses the language critically: about differences in ear-coverings, for instance, he says that ‘the ones in between’ being smooth and hairy ‘are the best with a view to hearing, but they indicate nothing of character’ (HA I 11, 492a32-34; 33-34: βέλτιστα δὲ τὰ μέσα πρὸς ἀκοήν, ἦθος δ’ οὐδὲν σημαίνει). Not every facial differentiation is a physiognomical sign!
Third, there is not one domain from which Aristotle draws his physiognomical signs. Some signs indicate ‘the best character’ in both humans and animals (such as having greenish eyes), suggesting that perhaps the physiognomical theory is of the zoologist kind, but other signs are unique to human beings and therefore cannot be drawn from other animals, suggesting that perhaps his physiognomy draws from the facial expressions of humans. Here his discussion of the differentiations in the position of the ear is interesting, as it identifies – in the language of APr II 27 – an affection that is distinctive of humans, such that the sign cannot be drawn from animals: ‘they [i.e., ears] are large or small or in between, or stick out a lot or stick out nothing or are in between, and the ones in between are a sign of the best character, while the large [ears] and the ones that stick out are [a sign] of a tendency for silly talk and garrulity’ (HA I 11, 492a34-b3: Καὶ ἢ μεγάλα ἢ μικρὰ ἢ μέσα, ἢ ἐπανεστηκότα σφόδρα ἢ οὐδὲν ἢ μέσον· τὰ δὲ μέσα βελτίστου ἤθους σημεῖον, τὰ δὲ μεγάλα καὶ ἐπανεστηκότα μωρολογίας καὶ ἀδολεσχίας).
Fourth, and perhaps most interestingly, Aristotle often describes the differentiations of parts on the head as forming a continuum between two extremes, with the one in the middle always being a sign of the ‘best character’, even it is not always the functional best. For instance, of eyes, Aristotle says that the ones that ‘are the most receding are the sharpest with regard to every animal, while the middle [i.e., the position in between protruding or receding] is a sign of the best character (HA I 10, 492a8-10; 9-10: τούτων οἱ ἐντὸς μάλιστα ὀξυωπέστατοι ἐπὶ παντὸς ζῴου, τὸ δὲ μέσον ἤθους βελτίστου σημεῖον). This language, of course, is strongly reminiscent of Aristotle’s ‘doctrine of the mean’ as presented in his ethical treatises, but what are we to make of this connection – if there is one? Does Aristotle think that having ‘intermediate’ facial features is a valid predictor of how successful one will be in developing virtues and hence in applying the ‘ethical mean’ in one’s feelings and actions? Is that why he tells future lawgivers to ‘look at’ the nations of the world – in order to diagnose (much in Pythagorean fashion) on the basis of their faces which peoples have the best (natural) character? Or is Aristotle just appropriating ‘reputable opinions’ in his own language here, and am I trying to make too much of it?
 I have found only two cases outside the History of Animals where Aristotle makes use of language that can perhaps be identified as physiognomical: see GA IV 5, 774a37-b2 and GA V 7, 786b34-787a2.
 Reading κτένες instead of ἰκτῖνες in HA I 9, 491b25.