Erismann ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 2019 Christophe Erismann The Depicted Man: The Byzantine Afterlife of Aristotle’s Logical Doctrine of Homonyms Christophe Erismann NOTEWORTHY DEVELOPMENT in Iconophile discourse in the ninth century is the sudden engagement with Aristo- telian logical concepts.1 In order to strengthen their arguments in favour of image veneration, both Nicephorus, Patriarch of Constantinople, and Theodore the Stoudite make use of logical reasoning and concepts. In addition to syllogistic reasoning, two Aristotelian doctrinal elements are used. First, the doctrine of relatives (πρός τι): the image and the prototype, the latter considered as model, are relatives in the strict Aristo- telian sense. All the properties attributed by Aristotle to relatives, such as simultaneity in being2—if one exists the other has to exist as well—and the implication of knowledge—if one is known, the other is necessarily also known—are properly applicable to image and model. The second Aristotelian doctrine is the con- 1 See P. J. Alexander, The Patriarch Nicephorus of Constantinople: Ecclesiastical Policy and Image Worship in the Byzantine Empire (Oxford 1958) 189–213; T. Anagnostopoulos, “Aristotle and Byzantine Iconoclasm,” GRBS 53 (2013) 763–790; K. Parry, Depicting the Word: Byzantine Iconophile Thought in the Eighth and Ninth Centuries (Leiden 1996) 52–63, and “Aristotle and the Icon: The Use of the Categories by Byzantine Iconophile Writers,” in S. Ebbesen et al. (eds.), Aristotle’s Categories in the Byzantine, Arabic and Latin Traditions (Copenhagen 2013) 35–57. 2 On this aspect see C. Erismann, “Venerating Likeness: Byzantine Icono- phile Thinkers on Aristotelian Relatives and their Simultaneity,” BJHP 24 (2016) 405–425. A 312 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 cept of homonymy.3 The model and the image are homonyms4 as they share the same name (I can say “Paul” of the Apostle and of an icon which depicts him), but not the same definition (Paul is a human being, whereas the icon is a piece of wood that is painted and covered with gold). The latter doctrine has received significantly less attention than the former. It is nevertheless an excellent way to assess the philosophical and logical culture of ninth-century thinkers, the nature of their sources, and their contribution to the history of logic. It also offers an exceptional case study for understanding better the complex itinerary fol- lowed by logical ideas, from the original Aristotelian text to the ninth-century thinkers who make use of them in their polemical writings. Homonyms are discussed by Aristotle on several occasions, but the main passage for us and the one which proved most influential is the discussion which opens the Categories. There Aristotle describes homonyms as things which have a name in common, but which differ in their definition, or, to use his own expression, in their defining-statement of essence (λόγος τῆς οὐσίας).5 Homonyms are part of a bigger scheme, namely a clas- 3 This is not the first appearance of the concept of homonym in a theo- logical controversy. Gregory of Nazianzus in Oration 29.14 offers an interest- ing earlier example. He accuses Eunomius of using faulty logic and of a bad understanding of the concepts of homonyms and synonyms. He forms a very suggestive bridging with Aristotle’s famous illustration of homonyms: “Well, do you really mean that the Son is not ‘God’ in the proper sense of the word, in the same way that a picture of an animal is not an animal?” He is reacting to the claim attributed to Eunomius (for which we have no textual trace in his own writings) that Christ is named “God” homonymously with the Father, i.e. that the same name “God” is predicated of both, but that the essences of both are different. The term κύων is taken as an example for homonyms by Gregory’s opponents (ὁ δὲ ἧττον τοῦ ἑτέρου κυνός, οἷον ὁ θαλάττιος τοῦ χερσαίου). We will find this precise example again later, as it is used by John of Damascus. 4 I do not use the translation “equivocal,” as both Aristotle and the exe- getical tradition clearly speak of things and not of terms. 5 I follow here Harold Cooke’s Loeb translation (1938) of λόγος by “state- ment,” but I make it more precise (“defining-statement”), to render the idea CHRISTOPHE ERISMANN 313 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 sification of things, and this classification precedes and is in a way similar to Aristotle’s discussion of the famous ontological square. This is his classification of beings according to the cri- teria of inherence (“being in a subject”) and predication (“being said of a subject”). In the case of homonyms and synonyms, the two criteria are “having a name in common” and “having a statement of essence in common.” If you add polyonyms and heteronyms, as the exegetes will do, then you have a full square. According to the opening remarks of the Categories, two things are synonymous if both the same name and the same definition are applicable to them (as when “animal” is predicated of the lion and of the donkey). They are homonymous if they share only the name, the definitions being different in the two cases. Aristotle does not draw the entire scheme, but we can recon- struct it with the help of his late ancient commentators.6 Same definition Different definition Same name SYNONYMS (συνώνυµα) HOMONYMS (ὁµώνυµα) Different name [POLYONYMS] (πολυώνυµα) [HETERONYMS] (ἑτερώνυµα) The Aristotelian doctrine of homonyms turns out to be a very efficient tool for responding to one of the main Iconoclast claims. It may even appear as predestined to play such a role in the Byzantine dispute, because of Aristotle himself and the example that he chose to use to illustrate his theory. For to illustrate what of definition which is implied in this context. It is important to note that οὐσία here refers to secondary substance, to the species, as it is only a species which can be defined (see An.post. 83b5), for a genus is requested in the definition. On the Aristotelian expression see J. Anton, “The Meaning of λόγος τῆς οὐσίας in Aristotle’s Categories,” The Monist 52 (1968) 252–267. 6 See for example John Philoponus In Cat. (A. Busse, Comm. in Arist. Graeca XIII.1 14.11–15): τῶν οὖν πραγµάτων τὰ µὲν κοινωνεῖ ἀλλήλοις κατὰ τὸ ὄνοµα διαφέρει δὲ κατὰ τὸν ὁρισµὸν καὶ καλεῖται ὁµώνυµα, τὰ δὲ κοινωνεῖ µὲν κατὰ τὸν ὁρισµὸν διαφέρει δὲ κατὰ τὸ ὄνοµα καὶ καλεῖται πολυώνυµα, τὰ δὲ κατά τε τὸ ὄνοµα καὶ τὸν ὁρισµὸν κοινωνεῖ καὶ καλεῖται συνώνυµα, τὰ δὲ κατὰ τὸ ὄνοµα καὶ κατὰ τὸν ὁρισµὸν διαφέρει, ἃ καλεῖται ἑτερώνυµα. 314 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 he meant by homonyms, he gave the example of a man and a picture, both called by the same Greek word (ζῷον) without however sharing the same definition. The example given in the Categories may nevertheless be read in two ways. The less obvious reading will gradually be adopted by the exegetical tradition, rendering the adoption of the doctrine even more natural for Iconophile thinkers. I will proceed in five steps in this paper. This is my reconstruc- tion of the long history that resulted in the use of the Aristotelian doctrine of homonyms by the ninth-century Iconophile theo- logians. The first step consists of an analysis of Aristotle’s description of homonyms at the beginning of the Categories (1a16) and a presentation of the ambiguities of the example he gives. The second step is to discuss the interpretation of Aristotle’s example by two late ancient commentators, Porphyry and Sim- plicius, who consider it to be a case of intentional homonymy (ἀπὸ διανοίας) and of homonymy by resemblance (καθ᾽ ὁµοι- ότητα). This characterisation allows only one possible reading of Aristotle’s text. The third focuses on the reformulation of this passage by the vehicle chiefly responsible for transmitting Aristotelian logic from Alexandria to Constantinople, namely the tradition of logical compendia. These anonymous seventh-eighth century logical handbooks offer a tendentious rendering of Aristotle’s example which, very probably, led to its use by ninth-century Iconophile thinkers. It will also be shown here that John of Damascus incorporated this reading in his Dialectica. The fourth step addresses which argument or problem the doctrine of homonymy was intended to solve, or, to phrase it differently, against what argument did the Iconophile thinkers use this particular element of Aristotelianism? The fifth and last part is a philosophical discussion of the Iconophile understanding of homonyms, in order to establish their contribution to an aspect of the doctrine. Ninth-century Iconophile thinkers insist on the fact that homonyms are things, but this is not an innovation; what is a real contribution to the discussion is the elaboration of a relational characterization of CHRISTOPHE ERISMANN 315 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 homonyms. For them, homonyms are relatives in a strict Ari- stotelian sense.7 1. Aristotle’s homonyms Aristotle opens the Categories with the homonyms.8 Things9 or beings are homonyms when they share a name, but not the state- ment of essence (Cat. 1a1–6): ὁµώνυµα λέγεται ὧν ὄνοµα µόνον κοινόν, ὁ δὲ κατὰ τοὔνοµα λόγος τῆς οὐσίας ἕτερος, οἷον ζῷον ὅ τε ἄνθρωπος καὶ τὸ γεγραµ- µένον· τούτων γὰρ ὄνοµα µόνον κοινόν, ὁ δὲ κατὰ τοὔνοµα λόγος τῆς οὐσίας ἕτερος. ἐὰν γὰρ ἀποδιδῷ τις τί ἐστιν αὐτῶν ἑκατέρῳ τὸ ζῴῳ εἶναι, ἴδιον ἑκατέρου λόγον ἀποδώσει. When things have only a name in common and the defining- statement of essence which corresponds to the name is different, they are called homonyms. Thus, for example, both a man and a picture are animals. These have only a name in common and the defining-statement of essence which corresponds to the name is different; for if one is to say what being an animal is for each of them, one will give two distinct defining-statements. (transl. Ackrill, mod.) Depending on how one understands ζῷον and τὸ γεγραµµέ- νον, two different readings are possible. The traditional reading 7 This article will be followed by a study offering an assessment of Photius’ reading of the question. In Amphilochion 137, the Patriarch of Constantinople exposes the “ante-predicaments” (τὰ πρὸ τῶν κατηγοριῶν), i.e. the concepts treated at the beginning of Aristotle’s Categories, before the discussion of the ten categories themselves, homonyms included. Photius shows a clear knowl- edge of, at least, some of the Alexandrian commentaries on the Categories, but he retains also elements of the Christian logical tradition of the compendia and of previous Iconophile thinkers discussed in the present paper. 8 On the background of this doctrine see J. Anton, “The Aristotelian Doctrine of Homonyma in the Categories and its Platonic Antecedents,” JHPh 6 (1968) 315–326. 9 I follow here the interpretation suggested by J. L. Ackrill in the com- mentary to his translation of the Categories (Oxford 1963), according to which homonyms are things: “The terms ‘homonymous’ and ‘synonymous’, as de- fined by Aristotle in this chapter [i.e. Cat. ch. 1], apply not to words but to things” (71). 316 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 consists in understanding the example as follows: a man and a depicted man, who may even be the same man, or any kind of depicted beast, are both called “animal.” Adherents of this read- ing would even translate the phrase with “a man and a portrait.” The fact that the homonymy focuses on “animal” seems an ar- gument for claiming that the depicted entity is a non-human animal, as otherwise Aristotle would have said that the name in common is “man.” A different reading is also possible. The term ζῷον in Ari- stotle’s time had two meanings, the main one being “living creature,” but there was also the less frequent but nevertheless attested meaning, “a figure or image in painting, embroidery, sculpture,”10 in other words an image or a picture but not neces- sarily of an animal. An illustration of this meaning is found in Herodotus (4.88): ἀπ’ ὧν δὴ Μανδροκλέης ἀπαρχήν, ζῷα γραψάµενος πᾶσαν τὴν ζεῦξιν τοῦ Βοσπόρου καὶ βασιλέα τε Δαρεῖον ἐν προεδρίῃ κατή- µενον καὶ τὸν στρατὸν αὐτοῦ διαβαίνοντα. Mandrocles took the first-fruits of these and had a picture made with them, showing the whole bridge of the Bosporus, and Darius sitting aloft on his throne and his army crossing. If one adopts this sense of ζῷον, then the example reads as follows: the term ζῷον is used to designate two things—the animal and the image—which share a name, but not the state- ment of essence. Homonymy is then purely accidental, just a fact of language. This reading has the advantage of being in ac- cordance with other passages of the Corpus Aristotelicum in which Aristotle offers examples of homonyms. In both cases, the two unrelated things are homonyms because they share only a name. The first is in the Nicomachean Ethics, namely κλείς, which means both a particular bone and a key (1129a27–31): Now “justice” and “injustice” seem to be ambiguous, but because the homonymy is close, it escapes notice and is not obvious as it is, comparatively, when the meanings are far apart, e.g. (for here the difference in outward form is great) as the homonymy in the 10 Cf. LSJ s.v. A.II.: “in art, figure, image, not necessarily of animals.” CHRISTOPHE ERISMANN 317 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 use of kleis for the collar-bone of an animal and for that with which we lock a door (οἷον ὅτι καλεῖται κλεὶς ὁµωνύµως ἥ τε ὑπὸ τὸν αὐχένα τῶν ζῴων καὶ ᾗ τὰς θύρας κλείουσιν). The second appears in the Topics and involves the word for donkey. Here the concept of homonyms is not mentioned, and the λόγος τῆς οὐσίας of the Categories is reduced to simply λόγος (107a18–21): Look also at the genera of the objects denoted by the same name, and see if they are different without the one falling under the other, as (e.g.) onos is both the animal [i.e. the donkey] and the engine (οἷον ὄνος τό τε ζῷον καὶ τὸ σκεῦος). For their defining- statement that corresponds to the name is different; for the one will be declared to be an animal of a certain kind, and the other to be an engine of a certain kind (ἕτερος γὰρ ὁ κατὰ τοὔνοµα λόγος αὐτῶν· τὸ µὲν γὰρ ζῷον ποιόν τι ῥηθήσεται, τὸ δὲ σκεῦος ποιόν τι). According to this reading, ζῷον, κλείς, and ὄνος are three names which happen to be respectively predicated of different pairs of things which do not share the same definition, the same essential being. 2. The commentators and the introduction of the concept of homonymy by resemblance The ambiguity of the Aristotelian text will be solved by the systematicizing exegesis produced by the scholars of the Neo- platonic schools.11 Two crucial distinctions will be introduced. These distinctions are based on concepts that do not come from Aristotle’s logical writings, but from the Physics. The first is the opposition between what comes about by chance, ἀπὸ τύχης, and what is the result of thought, ἀπὸ διανοίας (Ph. 2.5, 197a2– 3). The second is a distinction described by Aristotle as follows: “some homonymies are far removed from one another, some have a certain likeness, and some are nearly related either 11 For an overview of the Neoplatonic exegesis of Aristotle’s homonyms see J. Anton, “Ancient Interpretations of Aristotle's Doctrine of Homonyma,” JHPh 7 (1969) 1–18. 318 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 generically or analogically” (7.4, 249a23–25). For our purposes, the important element introduced here is the point about “having a certain likeness” (ἔχουσαί τινα ὁµοιότητα). Porphyry offers the first presentation of homonyms that in- tegrates these two elements,12 while also offering a personal and vivid elucidation of Aristotle’s example:13 12 Porphyry introduces the notions of homonyms by chance and hom- onyms as a result of thought in his commentary on the Categories, but we already find a mention of the concepts a few decades before, in the eighth book of the Stromata of Clement of Alexandria, the so-called Liber Logicus (8.8.24 [95.16–26 Stählin): “Homonyms bear the same name, but do not have the same statement, e.g. ‘man’ is both the living being and the one on the picture. Of homonyms, some are homonymous by chance, like ‘Ajax’ from Locris and ‘Ajax’ from Salamis, others from thought. And of the latter group, some [are homonymous] according to similarity, like ‘man’ being both the living being and the one on the picture; some according to analogy, like the ‘feet’ of a mountain and our ‘feet’, because they are lower; some according to activity, like the ‘foot’ [i.e. rudder] of a ship, by which the ship is kept on course, and our ‘foot’, by which we move. Homonyms are called after the same thing and in view of the same thing, like a book and a scalpel are ‘medical’ after the medical man using them and in view of the same statement of the ‘medical’,” τὰ δὲ ὁµώνυµα τῷ αὐτῷ ὀνόµατι χρώµενα, λόγον δὲ οὐ τὸν αὐτὸν ἔχοντα, οἷον ἄνθρωπος τό τε ζῷον καὶ γεγραµµένος. τῶν δὲ ὁµωνύµων τὰ µὲν ἀπὸ τύχης ὁµωνυµεῖ, ὡς Αἴας ὁ Λοκρὸς καὶ ὁ Σαλαµίνιος, τὰ δὲ ἀπὸ διανοίας, καὶ τούτων τὰ µὲν καθ’ ὁµοιότητα, ὡς ἄνθρωπος τό τε ζῷον καὶ ὁ γεγραµµένος, τὰ δὲ κατὰ ἀναλογίαν, ὡς “πόδες Ἴδης” καὶ οἱ ἡµέτεροι πόδες διὰ τὸ κατωτέρω εἶναι, τὰ δὲ κατ’ ἐνέργειαν, ὡς ποὺς πλοίου, δι’ οὗ τὸ πλοῖον πλεῖ, καὶ ποὺς ὁ ἡµέτερος, δι’ οὗ κινούµεθα. λέγεται ὁµώνυµα ἀπὸ τοῦ αὐτοῦ καὶ πρὸς τὸ αὐτό, ὡς [ἀπὸ] τοῦ ἰατροῦ τὸ βιβλίον καὶ τὸ σµιλίον ἰατρικὰ ἀπό τε τοῦ χρωµένου ἰατροῦ καὶ πρὸς τὸν αὐτὸν λόγον τὸν ἰατρικόν (transl. Havrda p.115, mod.). Here as well it is the name “man” which is common to the two homonyms, i.e. to the living human being and to the depicted one. It is highly probable that the eighth book of the Stromata integrates elements of Galen’s logical writings, most likely of his lost treatise On Demonstration, see M. Havrda, The so-called Eighth Stromateus by Clement of Alexandria: Early Christian Reception of Greek Scientific Methodology (Leiden 2016) 34–50, with further refer- ences. 13 A. Busse, Comm. in Arist. Graeca IV.1 65.25–30; transl. S. K. Strange, Por- phyry: On Aristotle, Categories (Ithaca 1992) 45–46 (mod.). CHRISTOPHE ERISMANN 319 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 The others [homonyms] depend on thought (ἀπὸ διανοίας), as for example homonyms by likeness (καθ’ ὁµοίωσιν). Suppose I were to use the name “man” both of a mortal rational animal and of an image of a man (εἰκόνα ἀνθρώπου)—suppose I were to see it and say, “That is a man” (ἄνθρωπος τοῦτο). Clearly it is not just a matter of chance that I call that which is drawn in the image (τὸ ἐν τῇ εἰκόνι γράµµα) “man”: it is because it is an image of a living man. So it is due to thought (ἀπὸ διανοίας) that I call both the living man and the statue or image “man.” Simplicius, following Porphyry, advances a more systematic application of these elements to the homonyms and inscribes Aristotle’s example into this scheme. For him homonyms are divided between homonyms by chance (ἀπὸ τύχης) and hom- onyms by thought (ἀπὸ διανοίας), i.e. intentional homonyms. This latter group is divided into four: homonyms by likeness (τὰ καθ’ ὁµοιότητα), by analogy (κατὰ ἀναλογίαν), by derivation from one thing (ἀφ’ ἑνὸς), and by reference to a single goal (πρὸς ἕν). Simplicius interprets Aristotle's example as the result of a thought on the basis of an existing likeness:14 Having arrived at this point, the commentators are accustomed to enumerate the types of homonyms (τοὺς τρόπους τῶν ὁµωνύµων), and they say that with regard to the highest types, homonyms come about in two ways. Some of them come about by chance (ἀπὸ τύχης), as both Paris and the Macedonian are “Alexanders”; and some are by thought (ἀπὸ διανοίας). The latter occur when someone thinks the matter over, and for a specific reason imposes the same names [on different things]. Chance homonyms, being contingent and indefinite, admit of no di- visions. By thought homonyms, by contrast, are divided into four: Firstly, the [homonyms] by likeness (καθ’ ὁµοιότητα): this is the kind Aristotle used in his example of homonyms, when he said “both the man and the depicted [man] are animals.” These have this name in common, but their defining-statement is different (τὸν δὲ λόγον ἕτερον), since the man is an “animal” as it is an animate, sensitive substance (οὐσία ἔµψυχος αἰσθητική), while the image or statue of a man is an “animal” in the sense that it is 14 K. Kalbfleisch, Comm. in Arist. Graeca VIII 31.22–32; transl. M. Chase, Simplicius On Aristotle’s Categories (Ithaca 2003) 45–46 (mod.). 320 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 a semblance of an animate, sensitive substance (ἡ δὲ εἰκὼν τοῦ ἀνθρώπου ἢ ὁ ἀνδριὰς ζῷον οὕτως ὡς ὁµοίωµα οὐσίας ἐµψύχου αἰσθητικῆς). Porphyry’s and Simplicius’ contributions—widely endorsed by the Alexandrian commentators15—constitute an important step in the history of Aristotelian homonyms for several reasons. First, they offer a precise interpretation of Aristotle’s example by reformulating it in terms of “an image of the man,” ἡ δὲ εἰκὼν τοῦ ἀνθρώπου. Their reading seems to ignore the second sense of ζῷον. Then, they clearly state that this predication is not by chance, but that this homonymy results from thought and is based on a similarity or likeness (καθ’ ὁµοίωσιν, καθ’ ὁµοιότη- τα). Finally, the distinction of essence between the man and the image is clearly stated, as man is defined as animate and capable of sensation, whereas the icon is not.16 Porphyry (128.18–19) states as a general rule that “homonyms never belong to the same genus,” τῶν δὲ ὁµωνύµων κατὰ ταὐτὸν οὐδὲν δήπου γένος. Simplicius notably insists on the fact that the man and his image belong to two different categories: “Socrates and the painted Socrates, however, do not both participate in the essence of ‘animal’; but one participates in the substance [of ‘animal’], 15 For example in the very influential commentary on the Categories by Am- monius: A. Busse, Comm. in Arist. Graeca IV.4, at 21.15–22.11. 16 Aristotle himself gives an example of homonyms in the De anima, which functions exactly like the Neoplatonic reading of the man and the statue: “Suppose that the eye were an animal—sight would have been its soul, for sight is the essence of the eye which corresponds to the defining-statement, the eye being merely the matter of seeing; when seeing is removed the eye is no longer an eye, except in name—no more than the eye of a statue or of a painted figure,” εἰ γὰρ ἦν ὁ ὀφθαλµὸς ζῷον, ψυχὴ ἂν ἦν αὐτοῦ ἡ ὄψις· αὕτη γὰρ οὐσία ὀφθαλµοῦ ἡ κατὰ τὸν λόγον (ὁ δ’ ὀφθαλµὸς ὕλη ὄψεως), ἧς ἀπολειπούσης οὐκέτ’ ὀφθαλµός, πλὴν ὁµωνύµως, καθάπερ ὁ λίθινος καὶ ὁ γεγραµµένος (412b18–23; transl. J. A. Smith, mod.). The eye of a living being —i.e. an eye which sees—and the eye of a depicted or sculpted animal are homonyms because both have the name “eye” in common, but the statement of essence—in this case sight is described as the essence of the eye—is not the same (which is clear from the fact that a statue does not see). CHRISTOPHE ERISMANN 321 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 while the other participates in ‘colour’ or ‘surface figure’. Thus, they are not reduced under the same category, but Socrates comes under the category of essence, while the painted Socrates comes under the category of the quality.”17 Two main conditions necessary for the utilisation of the doctrine of the homonyms by the Iconophile theologians of the ninth century—the causal explanation of some homonyms by a likeness between two things and the necessary essential differ- ence between two homonyms18—are formulated here for the first time. 17 Simpl. 21.17–20: ὁ δὲ Σωκράτης καὶ ὁ γεγραµµένος ἅτε µὴ τῆς οὐσίας ἄµφω τοῦ ζῴου µετέχοντες, ἀλλ’ ὁ µὲν τῆς οὐσίας, ὁ δὲ χρώµατος ἢ σχήµατος ἐπιπολαίου, οὐχ ὑπὸ τὴν αὐτὴν ἀνάγονται κατηγορίαν, ἀλλ’ ὁ µὲν Σωκράτης ὑπὸ τὴν οὐσίαν, ὁ δὲ γεγραµµένος ὑπὸ τὸ ποιόν (transl. Chase 36, mod.). 18 The difference of essence between the man and the image is stated in very clear terms by the later commentator David, who says in his Introduction to Philosophy—a series of prolegomena to the study of logic preceding his commentary to Porphyry’s Isagoge (Prolegomena Philosophiae (A. Busse, Comm. in Arist. Graeca XVIII.2 35.9–20; transl. S. Gertz, Elias and David [London 2018] 118, mod.): “ ‘like’ is used in the case of image and original, as when we say that an image of Socrates is similar to Socrates. We should know that the archetype is called ‘model’ and that, as it were, the image was produced in relation to it. For example, we say that Socrates is the model of Socrates’ image, since it is in relation to him that the image was produced. We say that the philosopher is like god in this sense: just as we say that the image of Socrates is like Socrates, even though the image of Socrates and Socrates are different (for one is inanimate, but the other animate), in this sense we also say that the philosopher is like god, even though the essence of god and man are different,” πάλιν ὅµοιον λέγεται ὡς ἐπὶ τῆς εἰκόνος καὶ τοῦ παραδείγµα- τος, οἷον ὡς ὅταν τὴν εἰκόνα Σωκράτους ὁµοίαν εἶναι λέγωµεν τῷ Σωκράτει. ἰστέον δὲ ὅτι παράδειγµα λέγεται τὸ ἀρχέτυπον καὶ οἱονεὶ πρὸς ὃ ἐγένετο ἡ εἰκών· οἷον παράδειγµα τῆς εἰκόνος Σωκράτους λέγοµεν τὸν Σωκράτην· πρὸς γὰρ τοῦτον ἐγένετο ἡ εἰκὼν αὐτοῦ. κατὰ τοῦτο οὖν τὸ σηµαινόµενον λέγοµεν τὸν φιλόσοφον ὅµοιον τῷ θεῷ εἶναι· ὥσπερ γὰρ τὴν εἰκόνα Σωκράτους λέγοµεν ὁµοίαν εἶναι τῷ Σωκράτει, εἰ καὶ ἄλλο ἐστὶν ἡ εἰκὼν τοῦ Σωκράτους καὶ ἄλλος ὁ Σωκράτης (ἡ µὲν γὰρ ἄψυχός ἐστιν, ὁ δὲ ἔµψυχος), κατὰ τοῦτο τὸ σηµαινόµενον λέγοµεν τὸν φιλόσοφον ὅµοιον εἶναι τῷ θεῷ, εἰ καὶ ἄλλη ἐστὶν ἡ οὐσία τοῦ θεοῦ καὶ ἄλλη τοῦ ἀνθρώπου. 322 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 3. Logical handbooks and John of Damascus’ Dialectica Middle Byzantine compendia of logic, which usually offer short summaries of all or part of the content of Porphyry’s Isagoge, Aristotle’s Categories, and sometimes of the Peri Hermeneias or of some elementary syllogistic, are historically crucial. They constitute one of the most efficient transmitters of the ideas about logic worked out by the Neoplatonic exegetes of Aristotle in Alexandria to the early Byzantine thinkers in Constantinople. Homonyms were presented in various logical handbooks com- posed in the seventh and the eighth centuries. Interestingly enough, the kind of short formulations featured in these com- pendia reflect several changes in comparison to Aristotle’s phrasing. The same paragraph about homonyms is to be found in two different compendia of logical terminology. The first of these compendia is devoted to some terms and concepts related to the Categories (ὁµώνυµον, συνώνυµον, παρώνυµον, ἑτερώνυ- µον, but also κατηγορούµενα, πρότερον, ἅµα, ἄτοµον, κατάφα- σις, ἀπόφασις, ἀντίφασις). The second devotes more coverage to the concepts at stake in the Isagoge (γένος, εἶδος, διαφορά, ἴδιον, συµβεβηκός) and in the Categories (several definitions of οὐσία, a few lines about ὁµώνυµα καὶ συνώνυµα, and a para- graph for each of the nine remaining categories). Homonyms are defined as follows:19 ὁµώνυµόν ἐστιν, ὅταν δύο πράγµατα µόνῳ ὀνόµατι κοινωνοῦσιν, ὡς ἐπὶ εἰκόνος καὶ τοῦ Παύλου· τὰ γὰρ ἀµφότερα λέγεις ἄν- θρωπον, ἀλλὰ µόνῳ τῷ ὀνόµατι κοινωνοῦσιν, τῷ δὲ πράγµατι διαφέρουσι. A homonym is when two things have only a name in common, as with an image and Paul, for you say “man” of both of them, yet 19 M. Roueché, “Byzantine Philosophical Texts of the Seventh Century,” JÖB 23 (1974) 61–76, at 72.2–4. A new edition of this very short text has been published by B. Roosen and P. Van Deun, “Les collections de définitions philosophico-théologiques appartenant à la tradition de Maxime le Confes- seur: le recueil centré sur ὁµώνυµον, συνώνυµον, παρώνυµον, ἑτερώνυµον,” in M. Cacouros et al. (eds.), Philosophie et sciences à Byzance de 1204 à 1453. Les textes, les doctrines et leur transmission (Leuven 2006) 53–76 (there is no textual difference for this passage, at 70.1–4). CHRISTOPHE ERISMANN 323 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 they have in common only this name, while they differ with re- spect to the thing itself. This text represents an important step. The example is Christianised, the mentioned individual is no longer any given man or Socrates,20 but Paul. Further, “animal” is not used any more, but “man,” which is predicated of both Paul and an icon. It seems quite reasonable to read the text as referring to Paul and an icon representing the Apostle. John of Damascus compiled an extensive logical compen- dium, the Philosophical Chapters (Capita Philosophica), better known under the title of Dialectica.21 He considers the question of hom- onyms in two distinct paragraphs. Paragraph 32 focuses on the case of the two types of animals called κύων or “dog,” one being the familiar canine and the other a particular type of fish. Both share the same name,22 but not the same definition (ὁρισµός) be- cause they are of a different nature (ἐπειδὴ καὶ ἑτέρα φύσις καὶ ἑτέρα, 101.5 Kotter). The second paragraph deals with homon- ymy in the case of an image (86.8–13): Predication is homonymous when [the two subjects of the predi- cation] admit of the [same] name (τὸ µὲν ὄνοµα δέχεται), but not at all of the [same] definition (τὸν δὲ ὅρον οὐδαµῶς). For instance, the image of a man admits of the name of man, but it does not admit of the definition of man (οἷον ἡ εἰκὼν τοῦ ἀνθρώπου τὸ µὲν ὄνοµα τοῦ ἀνθρώπου δέχεται, τὸν δὲ ὅρον τοῦ ἀνθρώπου οὐ δέχε- ται). For the definition of man is an animal rational, mortal, and receptive of intellect and scientific understanding (ζῷον λογικόν, θνητόν, νοῦ καὶ ἐπιστήµης δεκτικόν). And the image, however, is not an animal; for it is not animate (ἔµψυχος), nor is it rational or receptive of intellect and scientific understanding. It is interesting to note the clear understanding of the Aristo- 20 Simplicius, on one occasion, rephrases the Aristotelian example as fol- lows (21.14–16): “Again, Socrates is an animal, and so, too, is an image of Socrates, even though the latter is a configuration of colours,” καὶ πάλιν ζῷον ὅ τε Σωκράτης καὶ ἡ τοῦ Σωκράτους εἰκὼν χρωµάτων οὖσα σχηµατισµός. 21 Ed. P. B. Kotter, Die Schriften des Johannes von Damaskos I (Berlin 1969). 22 The example of κύων as a homonymous term is found in Arist. Soph el. 166a16–17. 324 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 telian expression λόγος τῆς οὐσίας in both passages as definition. John of Damascus uses both terms ὅρος and ὁρισµός. This understanding is indeed perfectly justified from a philosophical point of view. John of Damascus also wrote three treatises defending the veneration of images. He does not use Aristotelian logical ter- minology in the argumentation developed there. The introduc- tion of such terminology is a Constantinopolitan phenomenon well documented in Iconophile writings composed during the years between the restoration of iconoclasm by the emperor Leo V in 814/5 and the death of Theodore the Stoudite in 826. Logical concepts like homonymy were introduced by Nicepho- rus and Theodore in attempts to answer Constantine V’s the- orising of iconoclasm, but were not yet in use among Iconophiles at the time of the Second Council of Nicaea (787), as we shall see. 4. Constantine V The Iconophile use of the theory of homonymy can partly be explained by the need to refute an argument put forth by the Iconoclast emperor Constantine V (741–775). Constantine is particularly important for the history of the problem of images because he transformed an old issue partly grounded in the exegesis of the Old Testament into a Christological one. In the time of his father and predecessor Leo III, the veneration of icons was combatted as idolatry;23 during Constantine’s reign it was fought as heresy. We know the thought of Constantine V thanks to 47 fragments of his polemic theological inquiries (Πεύσεις) about images.24 They are quoted by Nicephorus, who tries to answer them and to refute Constantine’s position.25 One 23 See S. Gerö, Byzantine Iconoclasm during the Reign of Leo III (Louvain 1973) 102–109. 24 S. Gerö, Byzantine Iconoclasm during the Reign of Constantine V (Louvain 1977) 37–52. 25 Constantine’s work was written as a succession of questions and answers, according to the traditional Christian literary model of the erotapokriseis. They CHRISTOPHE ERISMANN 325 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 of them, particularly relevant for our topic, is reported by Nicephorus as follows:26 εἰσάγει γοῦν εὐθύς· καὶ εἰ καλῶς, ὁµοούσιον αὐτὴν [i.e. τὴν εἰκόνα] εἶναι τοῦ εἰκονιζοµένου. That is why he [Constantine V] immediately adds: “If the icon is good, it is consubstantial with the one of whom it is an icon.” For Iconoclasts following Constantine V, Christ himself and a good icon of Christ are consubstantial and do not differ in essence.27 Testimony for this position is in the discussion of the Iconoclast Council of Hiereia (754) as recorded during the sixth session of the Iconophile Council of Nicaea II (787):28 τούτῳ δὲ τῷ τρόπῳ καὶ οὗτοι τὴν εἰκόνα τοῦ Χριστοῦ καὶ αὐτὸν τὸν Χριστὸν οὐδὲν κατ᾽ οὐσίαν λέγουσι διαφέρειν. In the same manner, they [the Iconoclasts] assert that the icon of Christ and Christ himself differ in nothing regarding their es- sence.29 have been collected and edited in H. Hennephof, Textus Byzantinos ad Icono- machiam pertinentes in usum academicum (Leiden 1969) 52–57. An earlier edition was included in G. Ostrogorsky, Studien zur Geschichte des byzantinischen Bilder- streites (Breslau 1929) 8–11. 26 Hennephof, Textus Byzantinos 52 no. 142 (= Nicephorus Antirrheticus 1, PG 100.225A). 27 See A. Giakalis, Images of the Divine: The Theology of Icons at the Seventh Ecumenical Council (Leiden 2005) 132: “The identifications made by the icono- clasts were numerous. First, all true images are taken to be ‘natural’ or consubstantial with their archetypes. From this derives the utter impossibility of distinguishing between ‘natural’ and ‘imitative’ icons. By further con- sequence, image and archetype or prototype are identified. Consequently, Christ and an icon of Christ must be identified in their essence, that is to say, icon and person represented (prototype) must always be consubstantial. Every icon not identical with the prototype in essence is an idol.” 28 E. Lamberz, Concilium Universale Nicaeum Secundum (Berlin/New York 2008) 668.20–21. 29 In the Acts of Nicaea II, this claim is followed by a comment by the Iconophile redactor: “if they recognized the difference [in essence between Christ and the image of Christ], they would not have uttered all this bizarre nonsense. For it is obvious to all that a prototype is one thing and an image 326 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 The condition for a good icon, for Iconoclasts, is that it is of the same essence as, and consubstantial with, its archetype. The nature of the model and of the copy has to be the same. If this condition is not fulfilled by an image, then venerating it is idolatry. This condition is indeed impossible to fulfil. The argu- ment is simple: rationality is part of the human essence, but it is impossible for an icon, as a piece of wood that is painted and covered with gold, to be rational.30 This criterion makes it, de facto, impossible for any material depiction to fulfil the condition required to be considered as a good (καλῶς) icon and a proper object of veneration. Even if we do not have it in its entirety, it is possible, on the basis of the fragment transmitted by Nicepho- rus and his refutation of this position, to reconstruct the argu- ment of Constantine V along the following lines: 1) A good image is consubstantial with its prototype. 2) Consubstantial entities share all the same essential properties [by definition]. 3) A prototype is by essence rational and animate. 4) A depiction being only material is without reason and inanimate. 5) Therefore the depiction does not share all the essential properties of the prototype. 6) Therefore the depiction and the prototype are not consubstantial. 7) Therefore the depiction is not a good image, it is an idol. 8) Therefore it is wrong to venerate it. The argument is sound and difficult to contradict, with the another; the former is animate, while the latter is inanimate,” εἰ τὴν διαφορὰν ἔγνωσαν, οὐκ ἂν ταύτας τὰς κενοφωνίας ἐτερατολόγησαν. ἀρίδηλον γὰρ πᾶσιν ὑπάρχει ὅτι ἄλλο ἐστὶν εἰκὼν καὶ ἄλλο πρωτότυπον· τοῦτο µὲν ἔµ- ψυχον, ἐκεῖνο δὲ ἄψυχον (668.21–24 Lamberz; transl. R. Price, The Acts of the Second Council of Nicaea [Liverpool 2018] 475, mod.). 30 The affirmation that a man and his image differ essentially is indeed not an Iconophile innovation. It is implicit in Aristotle, as homonyms do not share the same λόγος τῆς οὐσίας. Already Simplicius, as we have seen, has insisted on this by underlining the fact that they both belong to different categories (substance for the model, quality for the image). John of Damascus, in a passage discussed below, states clearly that the image, contrary to its model, it is not animate, nor rational. CHRISTOPHE ERISMANN 327 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 exception of the first premise (1). The understanding of con- substantiality as sharing the same essential properties (2) is a well-accepted thesis among both philosophers and theologians. The descriptions of the prototype (3) and the depiction (4) are based on common-sense experience. The conclusion that the prototype and the depiction do not share all essential properties (5) and the rejection of consubstantiality between them (6) are logical deductions. The conclusion that the depiction is an idol is based on a logical application of the first premise (1). The final conclusion about the wrongness of the veneration of an idol (8) is a traditional Christian theological statement. Any attempt to refute such a reasoning must focus on the first premise. Exactly this will be the argumentative strategy of the Iconophiles, first at the Second Council of Nicaea in 787, then in a more philosophical way and with the help of the concept of homonyms in the Antirrhetici of both Nicephorus and Theodore the Stoudite. The Second Council of Nicaea, an Iconophile council, adopts the principle that the prototype and the image share a common name but insists on their difference in essence. It is nevertheless interesting to note that the Aristotelian terminology of hom- onyms is never used. It is therefore clear that this terminology was introduced only later, by Nicephorus and Theodore the Stoudite. I will mention just two examples from the Acts which precisely state the commonality of name and clearly reject an identity of essence between the icon and the model; these two passages would have been the right place to introduce the Ari- stotelian terminology of homonyms. 658.16–20 Lamberz: Therefore, since Christ is depicted according to his human nature, it is obvious, as the truth has proved, that Christians con- fess that the icon which is seen has in common with the archetype only the name, and not the essence (κατὰ τὸ ὄνοµα µόνον ὁµο- λογοῦσιν οἱ Χριστιανοὶ κοινωνεῖν τὴν ὁρωµένην εἰκόνα τῷ ἀρχε- τύπῳ καὶ οὐ κατὰ τὴν οὐσίαν). However, these senseless men say that there is no difference between an icon and the prototype and they decide for the identity of essence in things which are different in essence (αὐτοὶ δὲ κεπφωθέντες ἀδιάφορον λέγουσιν εἶναι 328 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 εἰκόνα καὶ πρωτότυπον καὶ ἐν ἑτεροουσίοις τὸ ταὐτὸν τῆς οὐσίας κρίνουσι). 664.28–31: For the icon is one thing and the prototype another. No one among those who think rightly looks in any way to the icon for the qualities of the prototype. For the right reason discerns in the image nothing other than that it [i.e. the image] has the name in common with him of whom it is the image, and not the essence (ἄλλο γάρ ἐστιν εἰκὼν καὶ ἄλλο τὸ πρωτότυπον· καὶ τὰ ἰδίωµατα τοῦ πρωτοτύπου οὐδαµῶς τις τῶν εὖ φρονούντων ἐν τῇ εἰκόνι ἐπιζητεῖ. ἐν γὰρ τῇ εἰκόνι ἄλλο οὐδὲν ὁ ἀληθὴς λόγος γινώσκει ἢ κατὰ τὸ ὄνοµα κοινωνεῖν οὗτινός ἐστιν εἰκὼν καὶ οὐ κατὰ τὴν οὐσίαν). 5. The “Iconophile” understanding of homonyms: Nicephorus and Theodore Stoudites Iconophile thinkers saw that the answer provided by Nicaea II was not sufficient to prevent a second uprising of Iconoclasm. The superficial refutation of the definition of the Iconoclast Council of Hiereia that was included during Nicaea II was not enough to win the intellectual battle. More had to be done on a theoretical level against the stronger official theorisation of iconoclasm, that of Constantine V, which was still the doctrinal nucleus of Second Iconoclasm. This is exactly what Nicephorus tries to do in his Antirrhetici in offering counter-arguments, written between 818 and 820, to the statements made by Constantine V. Nicaea II focused on the justification of the veneration of images. In order to win the battle, new weapons had to be used, the main one being Aristotelian logic. Nicephorus of Constantinople Nicephorus adopts homonymy first and foremost for the case of the icon of Christ, which also receives the name “Christ.” He says that “the icon, taking the name common to both natures, is called homonymously with the archetype.”31 But he rejects the consubstantiality of the pattern and the image. For him, they are 31 Antirrh. 1.47 (PG 100.324A–B): ἡ εἰκὼν … οὕτως οἰκειωθεῖσα καὶ τῷ κοινῷ τῶν φύσεων ὀνόµατι προσχρωµένη, ὁµωνύµως τῷ ἀρχετύπῳ προσαγο- ρεύεται. CHRISTOPHE ERISMANN 329 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 not ὁµοούσιοι (Antirrh. 1.30 [PG 100.280B–C]): [The image] has not acquired identity according to essence [with the archetype], nor need everything that is predicated of the archetype qua archetype be predicable of the image of it (οὐ κατ᾽ οὐσίαν τὸ ταὐτὸν κεκτηµένη, οὐδὲ γὰρ ὅσα κατὰ τοῦ ἀρχετύπου ὡς ἀρχετύπου κατηγορεῖται, καὶ τῆς ἀπ᾽αὐτοῦ εἰκόνος κατηγορη- θήσεται πάντως). Indeed, the archetype may be animate, while the image is inanimate. The archetype may be rational and able to move, while the image is without reason and motionless. Con- sequently, these two are not identical, but they are similar to each other in their visible form and dissimilar from each other in es- sence (οὐκοῦν οὐ ταὐτὸν ἀµφότερα, ἀλλὰ πὴ µὲν ἔοικεν ἀλλήλοις τῷ εἴδει, πὴ δὲ ἀπέοικε τῇ οὐσίᾳ). This text clearly shows that Nicephorus has Constantine V’s argument in mind. He is taking care to mention both the con- substantiality—in terms of identity of essence—and the key definitional element of the consubstantiality, i.e. that what is essentially predicated of one of two consubstantial entities also has to be predicated of the second. He indeed concedes, like John of Damascus before him, that the model and the image do not share all essential properties (like rationality and motion in his example). The introduction of the concept of homonyms by Nicephorus is intrinsically linked to his relational understanding of images. For him, as for Theodore the Stoudite, the model and the image are relatives in the strict Aristotelian sense elaborated in the Categories. Here is one example of Nicephorus’ use of the concept of relatives (with the notable addition of the Neoplatonic concept of relation) (Antirrh. 1.30 [277C–D]): The image is related to the archetype and is the effect of a cause. Therefore, necessarily it belongs to, and is called, a relative (τῶν πρός τι). Relatives are said to be such as they are from their being of some other thing, and through their relation (σχέσει) they are mutual correlatives. A father, for instance, is called the son’s father … thus an archetype is called the archetype of an image and an image the image of an archetype, and nobody will call the image of an individual an unrelated image; for the one and the other are introduced and contemplated together. 330 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 The very interesting point is that Nicephorus elaborates on the two elements—a particular kind of πρός τι, i.e. two things linked by a relation of resemblance, and the homonyms—not sep- arately but together. One will be the condition of the other. On the basis of the distinction found in the Neoplatonic com- mentaries, Nicephorus introduces a fascinating claim about the grounding of the homonymy in the relation of resemblance (Antirrh. 1.30 [280B]): Moreover, the resemblance confers homonymy [on the icon and its archetype] (ἐκ περιουσίας δὲ καὶ τὴν ὁµωνυµίαν χαρίζεται ἡ ὁµοίωσις). The designation (προσηγορία) is one and the same for both [i.e. the icon and the archetype]. The icon of the king is called “the king,” and might well say: “the king and I are one,” despite the evident fact that they are different in essence (δῆλον δὲ ὅτι παρὰ τὸ τῆς οὐσίας διάφορον). We have said these things in order to demonstrate the way in which the image, being con- sidered together with the archetype (πρὸς τὸ ἀρχέτυπον θεωρου- µένη), is related to it (τὴν σχέσιν ἔχει). Homonymy is not accidental at all but supervenes on the exist- ing relation of resemblance. Iconophile thinkers are aware of the risk of a purely accidental homonymy, since in that case this logical tool would not be useful at all as an argument against the Iconoclasts. If there is merely an accidental commonality of name between Peter and the image of Peter, then there is no gain in calling them homonyms. Iconophile thinkers try to avoid this by grounding homonymy. This is exactly what Nicephorus does when he introduces a very interesting claim about ground- ing the homonymy in the relation of resemblance. This is only possible because of his endorsement of the Neoplatonic analysis of Aristotle’s example. This grounding is confirmed by the seventh of the ten syl- logisms written by Nicephorus to establish the priority of the image of Christ over the Cross (Antirrh. 3.35 [432B]): The name “Christ” is predicated homonymously of the image of Christ (τὸ Χριστὸς ὄνοµα ὁµωνύµως κατὰ τῆς εἰκόνος Χριστοῦ κατηγορεῖται). It is called “Christ” as the image of the emperor is called “Emperor.” But it is impossible to say this about the Cross, as no one among the people who are sound of mind would call CHRISTOPHE ERISMANN 331 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 the cross “Christ” in any possible way. That which has come to participate in the name itself because it has already shared in the form of the body (καθάπερ ἤδη καὶ τῷ τοῦ σώµατος τύπῳ κε- κοινώνηκεν) is more precious than that which participates in none of these. So the image is more precious than the cross. This is not the place to discuss this argument as such and its place in Nicephorus’ anti-Iconoclast argumentation about the superiority of the image of Christ over the Cross.32 The relevant aspect of the text for our purpose is the clear causal explanation of homonymy. There is homonymy because there was before some sharing of formal appearance, which is, phrased differ- ently, nothing else than the relation of resemblance described in Antirrheticus 1.30. Here as well, homonymy does not come at ran- dom, but expresses a previous likeness, a resemblance in reality. It is because two things resemble one another or share some features in their appearance that they can be homonyms. The homonymy is grounded in the relation of resemblance. Theodore the Stoudite Theodore follows a similar line of argumentation as Nicepho- rus.33 His claim about the fact that Christ and the icon of Christ share the same name is directly supplemented by the affirmation that there is no consubstantiality between the prototype (Christ) and the image (Antirrh. 1.8 [PG 99.337C]): It is not possible to distinguish one from the other [i.e. an image from its prototype] by the name which they have in common (τῇ ὁµωνυµίᾳ), but by their natures (τῇ φύσει). If the distinction is possible thanks to their natures, it is obviously because the two natures are different. This point is also stated in another passage: “His image … is called ‘Christ’ because of the signification of the name (τῇ σηµασίᾳ τῇς προσηγορίας), not be- cause of the nature of divinity and humanity” (οὐ τῇ φύσει τῆς 32 See C. Erismann, “Nicephorus I of Constantinople, Aristotelian Logic and the Cross,” in M. Knežević (ed.), Aristotle in Byzantium (Alhambra 2019) 193–206. 33 For some remarks on Theodore on homonymy see T. Tollefsen, St Theodore the Studite’s Defence of the Icons. Theology and Philosophy in Ninth-Century Byzantium (Oxford 2018) 42–49 and 121–123. 332 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 θεότητός τε καὶ ἀνθρωπότητος). The icon does not have the divine and human essence of Christ. For Theodore, sharing a name—and the veneration—is not the same as sharing an es- sence: “It shares the name of the prototype (τοὔνοµα µὲν κοι- νωνεῖ), as well as its honour and veneration, but it has no part in its nature (τῆς δὲ φύσεως ἠλλοτρίωται)” (Antirrh. 2.17 [361C]). It is interesting to note that Theodore uses the concept of homonymy also in his literary work, as in his poem On the Holy Icons (no. 30), where he articulates the identity of name (here κλῆσις) with the difference in essence or nature:34 The Icon which you see is [an icon] of Christ. Call it also “Christ,” but homonymously; For the identity is in appellation, but not in nature. (Χριστὸν δὲ καὐτὴν λέξον, ἀλλ’ ὁµωνύµως· κλήσει γάρ ἐστι ταυτότης, ἀλλ’ οὐ φύσει) For both, there is one undivided worship. Therefore, who venerates this [image], worships Christ; for who does not venerate it, is also completely his enemy, since who rages against him does not want his painted appearance in the flesh to be worshipped. Theodore also insists on the link between the πρός τι and the homonyms. This is clear in his Ep. 528, which according to its editor, G. Fatouros, was probably written between 821 and 826.35 This letter is an explanation sent to John the Grammarian about the terminology used about images in another letter (Ep. 428), notably about the terms σχέσις, ὁµωνυµική, ὑποστατική. This very rich text states (789.50–790.59 Fatouros): For relation (σχέσις), as they say, belongs to the relatives (τῶν πρός τι). For they both exist together with one another (ἅµα) and are predicated reciprocally one of the other (ἀντιστρέφει πρὸς ἄλ- ληλα), as archetype of image. For the one could not exist if the other were not present (οὐ γὰρ ἂν εἴη θάτερον µὴ θατέρου παρόν- τος), as has been philosophized also in the case of things that exist 34 P. Speck, Theodoros Studites, Jamben (Berlin 1968) 175.1–8. 35 G. Fatouros, Theodori Studitae Epistulae I (Berlin 1992) 470. See also V. Grummel, “Jean Grammaticos et saint Théodore Studite,” EchOr 36 (1937) 181–189. CHRISTOPHE ERISMANN 333 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 at the same time (ἅµα). There is added as well the word “homon- ymous” (ὁµωνυµική), and this word too is of the same meaning [i.e. it is relational]. For a name is a name of something that is named. Thus in this case too the account of homonyms belongs to the relatives, since according to the definition used in philoso- phy we are taught that homonyms are those things “which have only their name in common, while the defining-statement of es- sence that corresponds with the name is different,” such as Christ himself and Christ when he has been depicted (ὥστε κἀνταῦθα τῶν πρός τι ὁ λόγος, ἐπεὶ καὶ κατὰ φιλοσοφίας ὅρον ὁµώνυµά ἐστι διδασκόµεθα, ὧν ὄνοµα µόνον κοινόν, ὁ δὲ κατὰ τοὔνοµα λόγος τῆς οὐσίας ἕτερος, οἷον αὐτὸς Χριστὸς καὶ ὁ ἐγγεγραµµένος). The last sentence of the passage is a literal quotation of Ari- stotle’s Categories, who is not mentioned by name but referred to by the expression κατὰ φιλοσοφίας. However, Theodore inter- venes in the text to change the example used by Aristotle: the man is changed to Christ. Where Aristotle has ὅ τε ἄνθρωπος καὶ τὸ γεγραµµένον, Theodore writes οἷον αὐτὸς Χριστὸς καὶ ὁ ἐγγεγραµµένος. The iconophile interpretation and adaptation of the Aristotelian example are now complete. It is not even about the icon of Paul, but directly about the icon of Christ which was the absolute core of the debates in the ninth century. On a terminological level, noteworthy is the use of the un- common term—Theodore is known for his use of rare words and neologisms36—ὁµωνυµικός. A literal translation would be “re- lating to the identity of name.”37 For Theodore, the close link between homonyms and relatives is also due to the relational nature of homonyms. Homonyms are πρός τι. As there is no master without a slave, there is no homonym without another homonym. An entity is never a homonym alone, but always with something else. Being homon- ymous is a relational property. Once it is stated that homonyms 36 See E. Trapp, “Zum Wortschatz des Theodor Studites,” in M. Grünbart (ed.), Theatron. Rhetorische Kultur in Spätantike und Mittelalter (Berlin 2007) 449– 462. 37 Here I follow the translation die Gleichnamigkeit betreffend, as given in the Lexikon zur byzantinischen Gräzität V 1133, which cites precisely this example. 334 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 are relatives, Theodore can apply to homonyms the charac- teristics of relatives, such as simultaneity of being. Since a thing cannot be a homonym alone, it means that it only exists as a homonym when a second homonym exists. The link between the concepts of homonyms and relatives is so strong that in the same letter, when he quotes his letter 428, he dares a surprising glossing of the term “relational” by the term “homonymous”: “it is however relational, that is to say homonymous” (σχετικὴ δὲ ὅµως, ἤγουν ὁµωνυµική, 788.22–23). The fact that both Nicephorus and Theodore adopt relatives and homonyms in the same theoretical move has possibly influ- enced their understanding of both concepts. All the examples they discuss—i.e. a very specific kind of image—are both hom- onyms and relatives. Theodore seems to deduce from the two claims that Christ and his image are homonyms as they share the name “Christ,” and that Christ and his image, as archetype and image, are relatives (πρός τι), the understanding that hom- onyms are relatives. Such a conclusion would be corroborated by the fact that the homonymy is grounded in the relation of resemblance, which, precisely, is the explanation of the ontological fact that the model and the image are relatives. The two concepts function in a sim- ilar way, i.e. in both cases it requires two entities. One is neither a relative nor a homonym alone, but always of something or to something.38 Independently of the question of images, it is one of the great merits of the applied logic of Nicephorus and Theo- dore that they underline the relational nature of homonyms. 38 This point was already made by Ammonius in his commentary to the Categories (Comm. in Arist. Graeca IV.4 17.16–18): “See how precisely he [Ari- stotle] says not ‘homonym’ but ‘homonyms’, using the plural form of the word, since homonyms are thought of as many things, or at least two, but one never speaks this way in the case of one thing” (ἐπειδὴ ταῦτα θεωρεῖται ἐν πλείοσι πράγµασιν ἢ ἐν δύο τὸ ἐλάχιστον, ἐν ἑνὶ δὲ οὐδέποτε λέγεται). Philoponus (Comm. in Arist. Graeca XIII.1 15.34–16.1) has followed his master by claiming that one cannot find one homonym, as one does not say that one thing is homonymous to itself (οὐ γὰρ λέγεται αὐτὸ ἑαυτῷ τι εἶναι ὁµώνυµον). CHRISTOPHE ERISMANN 335 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 Later Iconophile writings The concept of homonyms is also present in later Iconophile literature; these texts very probably reflect the work done by Nicephorus and Theodore. Karl Hansmann has published nine homilies on the Gospels of John and Matthew that contain several remarks about images that defend an Iconophile point of view.39 These passages con- tain an application of logical terminology to the question of images which is extraordinarily close to Nicephorus’ and Theo- dore’s way of writing, notably in the endorsement of the re- lational understanding of images. Hansmann proposed a date of composition between 809 and 811. It has recently been sug- gested that these homilies could be the work of Metrophanes of Smyrna,40 the bitter adversary of Photius in the 860s. Their author offers an interesting remark about the homonymy in the case of images, which, for once, does not only concern religious icons (188.2–6 Hansmann): For, the image of the man, of the horse, and of the lion are named “man,” “horse,” “lion” homonymously but not synonymously. For it is not the case that one finds among those things admitted in things named homonymously complete identity of nature and form (ὁµωνύµως γὰρ ἡ τοῦ ἀνθρώπου καὶ ἵππου καὶ λέοντος εἰκὼν ὠνόµασται ἄνθρωπος καὶ ἵππος καὶ λέων ἀλλ’ οὐχὶ συνωνύµως. οὐκ ἔστι δὲ τῶν ἐνδεχοµένων ἐν τοῖς ὁµωνύµως ὠνοµασµένοις εὑρεῖν παντελῆ ταυτότητα φύσεως καὶ µορφῆς). In this text, in which the author shows his familiarity with the logical terminology by contrasting homonymy with synonymy, consubstantiality is clearly rejected. Two homonyms are not essentially identical: here the notion of essence is expressed by 39 K. Hansmann, Ein neuentdeckter Kommentar zum Johannesevangelium: Unter- suchungen und Text (Paderborn 1930). These homilies are transmitted in a single manuscript (British Library, Add. 39605), which dates to the early tenth cen- tury. The name of the author has unfortunately been erased. 40 P. Van Deun, “La chasse aux trésors: la découverte de plusieurs œuvres inconnues de Métrophane de Smyrne (IXe–Xe siècle),” Byzantion 78 (2008) 346–367. 336 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 the term φύσις, a traditional way in the Christian logical tra- dition to refer to species, also used by Theodore. One last text probably linked to the second period of Icono- clasm (814–843) has to be mentioned here, even if its authen- ticity is strongly contested. It is the so-called Letter of the Three Patriarchs (Epistula synodica ad Theophilum imperatorem, BHG 1386).41 The letter purports to have been written to Emperor Theophilos (829–842) by the Patriarchs of Alexandria (Christopher), Anti- och ( Job), and Jerusalem (Basil) during a synod held in Jerusalem in 836. The existence of this synod is not attested. Nevertheless, the ninth-century origin of the text is proved by the oldest testi- mony to the text, a ninth-century manuscript in majuscule, today in the Monastery of St. John on Patmos (MS. 48).42 The heterogeneous nature of the work and its complicated transmis- sion render very difficult a clear decision about which chapters are original and which are not. Scholars have given various judgements on the chapters reasserting the Iconophile convic- tions of the Patriarchs. It is not my purpose here to tackle this vexed issue. I will limit myself to mentioning a passage about homonyms. This passage is doctrinally interesting in any case, either as an original claim of the Patriarchs, or as part of a later ninth- century interpolation of the text. If the passage is part of the original letter, as claimed by Julian Chrysostomides, then it is interesting to note the relatively quick diffusion of the use of homonyms in the discussion about icons—once again, a very Constantinopolitan terminology—among Melkite theologians under Muslim rule. If the text is not authentic, then it constitutes one more example of the acceptance in the second half of the 41 Ed. and transl. J. A. Munitiz, J. Chrysostomides, E. Harvalia-Crook, and C. Dendrinos, The Letter of the Three Patriarchs to Emperor Theophilos and Related Texts (Camberley 1997). See also J. Signes Codoñer, The Emperor Theophilos and the East, 829–842: Court and Frontier in Byzantium during the Last Phase of Iconoclasm (Farnham 2013) 365–420. 42 A. D. Kominis, Πατμιακὴ βιβλιοθήκη ἤτοι Νέος κατάλογος τῶν χειρο- γράφων κωδίκων τῆς Ἱερᾶς Μονῆς Ἁγίου Ἰωάννου τοῦ Θεολόγου Πάτμου I (Athens 1988). CHRISTOPHE ERISMANN 337 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 ninth century of the lexical and conceptual innovations intro- duced by thinkers of the early ninth century (23.1–5 Munitiz): For the copies of the prototypes are [named] homonymously [with the prototypes], the copies display the particular features of the visible non-essential form, those by which it is recognized, e.g. the hookedness of the nose, greyness of the eyes or the whiteness and blackness of skin and the similar characteristics among qual- ities; for the image of a person is drawn not by nature, but by convention (καὶ γὰρ ὁµωνύµως τὰ παράγωγα τῶν πρωτοτύπων, ἐµφαίνουσι τὰ ἰδιώµατα τῆς ὁρωµένης ἐπουσιώδους µορφῆς, τὰ γνωρίσµατα, οἷον, τὸ γρυπόν, τὸ γλαυκόν, ἢ τὸ λευκόν, ἢ τὸ µέλαν, ἢ καὶ τὰ ὅµοια τῶν ποιοτήτων χαρακτηρίσµατα· οὐ γὰρ φύσει ἡ εἰκὼν ἐγγέγραπται τοῦ ἀνθρώπου, ἀλλὰ θέσει). According to this, homonymy is linked to the exemplification of common visible properties, of shared aspects of appearance. It is clear that for the author(s) of this passage there is no consubstan- tiality between prototype and image, as the shared properties are clearly described as non-essential (ἐπουσιώδης); yet consubstan- tiality consists in sharing all essential properties. The properties exemplified by both the prototype and the images are accidental qualities, that is to say, they do not belong to the essence of the two entities and are not subsumed in the category of οὐσία but in that of quality (ποιόν). Conclusion The main philosophical conviction of the Iconophile theo- logians of the first quarter of the ninth century in Constantinople consists of two tenets: the image is named “Christ” homon- ymously (κατὰ τὸ ὁµώνυµον) and it is Christ’s image relatively (κατὰ τὸ πρός τι). This is formulated thanks to Aristotelian logic, but, in the case of the homonyms as well as in the case of rela- tives,43 only after a long exegetical journey. As such, Aristotle’s example of a man and a picture in the first lines of the Categories was predestined to be used in a controversy about images, but not yet ready to be used. It is thanks to the various steps and layers of 43 The main proof is the use of the typically Neoplatonic concept of relation (σχέσις). 338 THE DEPICTED MAN ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 interpretation that his example became the perfect tool for Iconophile theologians. We have followed the evolution of the example over centuries, starting with Aristotle’s wording ὅ τε ἄνθρωπος καὶ τὸ γεγραµµένον. It became first ἡ δὲ εἰκὼν τοῦ ἀνθρώπου and ὁ δὲ Σωκράτης καὶ ὁ γεγραµµένος by the Neo- platonic commentators. It was then Christianised in middle Byzantine logical compendia as εἰκόνος καὶ τοῦ Παύλου, before reaching its final form of αὐτὸς Χριστὸς καὶ ὁ ἐγγεγραµµένος among Iconophiles, whose analysis was focused on the Christo- logical dimension of the iconoclast controversy. From a philosophical point of view, it is remarkable to observe how Iconophiles have so fruitfully developed the Neoplatonic classification of homonyms. Simplicius analysed the case of Aristotle’s example of a man and an image of a man as a kind of homonymy which is both intentional/by reason—i.e. which results from an act of thinking—and which is based on a similarity or likeness. Nicephorus and Theodore the Stoudite both adopted such a view, as for them the homonymous use of the name “Christ” for both Christ and an image of Christ is grounded in the relation of resemblance existing between Christ and his icon. It is not possible to decide if Nicephorus or Theodore did read specifically the commentary of Simplicius or that of another Neoplatonist, as this exegetical view on homonyms was common among the School and is mentioned in other expositions. It is far from impossible that the Iconophiles had direct recourse to Sim- plicius’ commentary, as this text was very probably accessible in the first half of the ninth century. A copy of Simplicius’ com- mentary on the Categories—today in Paris, Par.gr. 257544— was produced in connection with the so-called “Philosophical Col- lection,” a very important set of scientific manuscripts copied in Constantinople between 850 and 875.45 This implies the existence, at the time of copying, of an older codex, on the basis 44 The manuscript is a palimpsest and Simplicius’ commentary is the scriptio inferior. 45 A. C. Cataldi Palau, “Un nuovo codice della ‘Collezione filosofica’: il palinsesto Parisinus Graecus 2575,” Scriptorium 55 (2001) 249–274. CHRISTOPHE ERISMANN 339 ————— Greek, Roman, and Byzantine Studies 59 (2019) 311–339 of which the new manuscript was made. It remains to note that the question of homonyms offers an excellent illustration of the depth of the logical and philosophical culture of the preeminent theologians of the first half of the ninth century. It confirms also that the introduction of logical ideas into the dispute about image veneration was a Constantino- politan phenomenon. Detailed analysis of the question of hom- onyms allows us to claim that this introduction was posterior to the Second Council of Nicaea. If one accepts the attribution of Hansmann’s anonymous text to Metrophanes of Smyrna, and even with 836 as the date of composition of the Letter of the Three Patriarchs, then clearly Nicephorus and Theodore, mainly in their respective Antirrhetici, seem to be at the origin of the Icono- phile move. The study of the long history of homonyms offers also a clear proof that Aristotle was not read without his Neoplatonic exegetes. This history offers a remarkable case study of the modalities of the transmission of logical ideas. Above all, it shows that transmission of theses and concepts in the history of philoso- phy does not always follow the scheme of one medieval author reading one ancient text, but is often the result of a long suc- cession of engaged intermediaries with their own opinions, who silently but efficiently modify aspects of the ancient text.46 February, 2019 University of Vienna christophe.erismann@univie.ac.at 46 This paper was written under the auspices of the research project “Re- assessing Ninth Century Philosophy. A Synchronic Approach to the Logical Traditions” (9 SALT) generously granted by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 648298). Earlier versions of this article were presented at the conference “Aristotle and Christianity” (National and Kapodistrian University of Athens, November 2016) and at the conference “Philosophy in Byzantium: The Order of Nature and Order of Humankind” (University of Munich, October 2017).