In his recommendations to lawgivers about the material conditions of the ideal city in the Politics, Aristotle suggests that the question of what should be the natural qualities of its citizens can be answered fairly easily by reference to existing ethnographical opinions (Pol VII 7, 1327b20-23):
“One could almost grasp this by looking at the cities that are held in high esteem among the Greeks and, with respect to the whole inhabited world, at how it is divided into nations.” (σχεδὸν δὴ κατανοήσειεν ἄν τις τοῦτό γε, βλέψας ἐπί τε τὰς πόλεις τὰς εὐδοκιμούσας τῶν Ἑλλήνων καὶ πρὸς πᾶσαν τὴν οἰκουμένην, ὡς διείληπται τοῖς ἔθνεσιν.)
In order to be able to select his citizens well, the lawgiver does not need to have any specialized kind of knowledge about, for instance, the essences of human beings and their biological or psychological differentiations. Instead, he can simply rely on reputable opinions about which peoples are supposed to make good citizens, and observe what kinds of natural qualities they have. This reference to ethnographical endoxa in combination with an appeal to empirical evidence is significant, as it suggests the acceptance of characterological stereotypes, but also – and more surprisingly – of physiognomical theory in Aristotle’s thoughts.
Physiognomy is the ancient (but not quite extinct…) ‘science’ of ‘judging’ or ‘discriminating’ character that is traditionally traced back to Pythagoras, who allegedly applied it to his prospective students as a means to diagnose their character traits and intelligence before admitting them to his school (see Aulus Gellius, Noct. Att. I 9), although Galen prefers to credit ‘the divine Hippocrates’ as its originator, referring to the clear links between physiognomy and the humoral theories and sign-interpretations prevalent in ancient Hippocratic medicine. The ‘science’ is first treated systematically in a handbook from the third-century B.C., the Physiognomonics, consisting of two parts (each presumably written by a different Peripatetic author), which has been spuriously ascribed to Aristotle. The handbook describes physiognomy as a science that postulates a ‘sympathetic’ correspondence between a person’s body and soul, according to which affections of the one simultaneously change the other. This correspondence allows one to draw inferences from observations of certain bodily or facial features in a person – that is, from the so-called ‘physiognomical signs’ that are themselves identified on the basis of either similarities to features distinctive of animals, ethnicities, or moral states – to the presence of the corresponding psychological states or character traits in that person.
There is no doubt that Aristotle was familiar with the discipline and its methods. For instance, in the midst of explaining how occasionally monstrosities come about during the formation of the embryo, Aristotle refers to ‘a certain physiognomist’ who ‘reduced all faces to those of two or three animals’ (GA IV 3, 769b20-21: φυσιογνώμων δέ τις ἀνῆγε πάσας εἰς δύο ζῴων ἢ τριῶν ὄψεις). More significantly, Aristotle concludes an exposition of how the validity of non-deductive types of reasoning (such as inductions, inferences from examples, and inferences from signs, all of which are often used in rhetorical contexts) can be tested for through the three syllogistic figures, with a discussion of the validity of physiognomical inferences (APr II 27, 70b7-38). By using an example of a physiognomical inference that takes its signs from animals, Aristotle lays out the conditions under which ‘it is possible to judge natures’ (APr II 27, 70b7-38; see especially 70b7: τὸ δὲ φυσιογνωμονεῖν δυνατόν ἐστιν, εἴ τις δίδωσιν…; b11-14: εἰ δὴ … δυνησόμεθα φυσιογνωμονεῖν; b22-26: εἰ τοίνυν … δυνησόμεθα φυσιογνωμονεῖν).
What is less clear, however, is whether and to what extent Aristotle endorses this kind of science and relies on its contents and methods in his own works. The conditional language Aristotle uses in the Prior Analytics passage warrants caution (Aristotle may simply be discussing current inferential practices performed by others), and it has been suggested that the passage is not closely related to the rest of the treatise and is perhaps a later appendix. In addition, the Peripatetic author of the first treatise preserved in the Physiognomonics claims that his alternative, ‘philosophical’ method of physiognomy (according to which one infers a psychological state indirectly from the observation of the signs of a second psychological state with which the presence of the first is necessarily connected) is one that “nobody has ever tried” (Phgn 2, 807a4: οὐδεὶς μέντοι ἐπικεχείρηκεν). This suggests that at least this (more sophisticated) part of the Peripatetic physiognomical method is not to be traced back to Aristotle’s own views.
Perhaps in my next blog-post, I will be able to say more about the role of physiognomy – if there is one – in Aristotle’s works…
 See Smith (1989), 227: “This terminal section, on ‘recognizing natures’ (phusiognômonein), seems to have no very close connection with the Prior Analytics. It might have been included (or appended by an editor) because this skill also formed part of the orator’s bag of tricks. However, the discussion never refers to the other themes of the Prior Analytics. It seems more likely, therefore, that this passage found its place here only because it is concerned with certain physical states as signs of states of character.”