¹ Avicenna, Kitāb al-Šifāʾ, Al-Ṭabīʿiyyāt, al-Nafs, ed. Rahman, Fazlur, Avicenna's De Anima, Being the Psychological Part of Kitāb al-Shifāʾ (Oxford 1959), p. 7, 8–10 Google Scholar. In the next passage Avicenna will argue that perfection is a better notion for the soul than “power”.

² “Form” replaces here “substance” due to the context.

³ Ibid ., p. 9, 1–7.

⁴ Both Arabic ǧawharī and English “substantial” are equivocal and can refer both to the fact that something is substance and that something is an integral ontological part of substance. In this article I will mean the latter whenever I use the word “substantial” or “substantiality”. To render the first idea I will use the expression “substancehood” or “being substance”. For the equivocity of Arabic “ǧawharī” see infra.

⁵ Avicenna, Nafs I, 1, p. 10, 3–6.

⁶ Ibid ., p. 10, 9–14.

⁷ Some research has already been done on Avicenna's theory of the soul (e.g. Hasse, Dag, Avicenna's De Anima in the Latin West: The Formation of a Peripatetic Philosophy of the Soul, 1160–1300 [London-Turin, 2000]Google Scholar, Druart, Thérèse-Anne, “The human soul's individuation and its survival after the body's death: Avicenna on the causal relation between body and soul”, Arabic Sciences and Philosophy, 10 [2000]: 259–73Google Scholar). His famous “Flying Man” has attracted particular attention (e.g. Marmura, Michael, “Avicenna's ‘Flying Man’ in context”, in Marmura, M. (ed.), Probing in Islamic Philosophy: Studies in the Philosophies of Ibn Sina, al-Ghazali and Other Major Thinkers (Bringhamton-New-York, 2005), pp. 181–96Google Scholar). The question about the ontological status of the soul and its being a separable substance was investigated in Wisnovsky, Robert, Avicenna's Metaphysics in Context (London, 2003), pp. 113–44Google Scholar).

⁸ I cannot agree with De Haas, Frans, John Philoponus’ New Definition of Prime Matter (Leiden-New-York-Köln, 1997), p. 198 Google Scholar that this definition should not be understood as a definition of accident. Indeed, using Avicenna's framework, it is the definition of accident as an ontological category but not the definition of “accidentality” (meaning “being extrinsic”).

⁹ Aristotle, Categoriae, ed. Minio-Paluello, Laurent, Aristotelis categoriae et liber de interpretatione (Oxford, 1949)Google Scholar, 3a29–32: “μὴ ταραττέτω δὲ ἡμᾶς τὰ μέρη τῶν οὐσιῶν ὡς ἐν ὑποκειμένοις ὄντα τοῖς ὅλοις, μή ποτε ἀναγκασθῶμεν οὐκ οὐσίας αὐτὰ φάσκειν εἶναι· οὐ γὰρ οὕτω τὰ ἐν ὑποκειμένῳ ἐλέγετο τὰ ὡς μέρη ὑπάρχοντα ἔν τινι.”

¹⁰ Most modern interpreters assume that Aristotle meant physical parts and not differentiae in this passage (see e.g. Oehler, Klaus, Aristoteles: Werke in deutscher Übersetzung, Bd. 1 (Oldenburg 2006), p. 265 Google Scholar and Rashed, Marwan, Essentialisme: Alexandre d'Aphrodise entre logique, physique et cosmologie (Berlin-New York 2007), p. 43, n. 133Google Scholar). However, the Late Antique commentators wouldn't have agreed with this interpretation, as will become clear shortly. For the view that it is hylomorphic compound here at stake see Rashed, Essentialisme, p. 44.

¹¹ Cf. Rashed, Essentialisme, pp. 50, 69–70, 75ff. This can be also related to comparing differentiae-genus relation to form-matter relation (cf. Aristotle, Met. Δ, 28, 1024b6-9 and Met. Z, 12, 1038a-9 and Porphyry, Isagoge, ed. Adolf Busse, Commentaria in Aristotelem Graeca 4.1 [Berlin, 1887], p. 11, 15–16).

¹² Cat. 3a20-22: “ὥστε οὐκ ἂν εἴη οὐσία τῶν ἐν ὑποκειμένῳ. —οὐκ ἴδιον δὲ οὐσίας τοῦτο, ἀλλὰ καὶ ἡ διαφορὰ τῶν μὴ ἐν ὑποκειμένῳ ἐστίν.”

¹³ On Lucius and the aporia handed down by Simplicius see Griffin, Michael, Aristotle's Categories in the Early Roman Empire (Oxford, 2015), pp. 108–11, 120–1Google Scholar and Moraux, Paul, Der Aristotelismus bei den Griechen: von Andronikos bis Alexander von Aphrodisias, Bd. 2 (Berlin-New York, 1984), pp. 528–63 esp. 537–8Google Scholar.

¹⁴ Simplicius, In Aristotelis categorias commentarium, ed. Kalbfleisch, Karl, Simplicii in Aristotelis categorias commentarium, Commentaria in Aristotelem Graeca 8 (Berlin, 1907), p. 48, 1–11 CrossRefGoogle Scholar.

¹⁵ P. Moraux rightly noticed that this aporia is rather strange against the background of what we find in Aristotle's Cat. It is Aristotle himself who firstly accepts that differentiae – that undoubtedly are to be numbered under Lucius’ “complements” of substance – are not-in-a-subject. Moreover, he agrees at Cat. 3a29–32 that the parts of a substance are substances themselves and correspondingly not-in-a-subject exactly because of his definition stated above in Cat. (Moraux, Aristotelismus, p. 538).

¹⁶ Despite Simplicius’ account it remains a question whether Porphyry really was the first to try to solve Lucius’ aporia. For instance, it is known that both Alexander of Aphrodisias and Plotinus agreed with a view identical with the first possibility given by Lucius. They could have anticipated Porphyry in this respect (Rashed, Essentialisme, p. 76ff on Alexander defining differentiae as substances and Plotinus, Enneades, ed. Henry, Paul and Schwyzer, Hanz-Rudolf, Plotini opera, vol. 2 [Berlin, 1959]: VII, 3, 5, 12–13 Google Scholar where Plotinus draws on substancehood on the ground of being a part of substance).

¹⁷ A possible source for this line of thought could have been Alexander's view that heat is substance as an intrinsic element of substances (Alexander apud Averroem, In Aristotelis metaphysica commentarium, in: Averroes, Tafsīr mā baʿd al-ṭabīʿa: Grand Commentaire de la Métaphysique, ed. Maurice Bouyges [Beirut, 1938–1948], p. 1519, 10–11). In Alexander, In Aristotelis topicorum libros octo commentaria, ed. Maximilian Wallies, Commentaria in Aristotelem Graeca 2.2 (Berlin, 1891), p. 50, 14–15 Alexander claims that heat is not an accident of fire though it is an accident in other things. Thereby he rejects the suggested definition of accident as that “which is accidental at least once”. Though this looks very similar to Avicenna's definition, there is a radical difference between them. Alexander's opponents apply this definition to the accident meaning “accidental” whereas Avicenna to “accident as ontological category”. Therefore Avicenna's position is rather closer to Alexander: that something can be both accidental and substantial in different cases. Cf. Ebert, Theodor, “Aristotelian accidents”, in Taylor, Christopher (ed.), Oxford Studies in Ancient Philosophy, vol. XVI (1998), pp. 133–59, 146–7Google Scholar and De Haas, Prime Matter, p. 203.

¹⁸ Porphyry applies Aristotle's second definition of accident in Top I, 5, 102b6–7, upgraded by Alexander, In Top., pp. 49, 25–50, 5 to accident defined as being-in-a-subject.

¹⁹ Simplicius, In Cat., p. 48, 11–33.

²⁰ Ibid ., p. 49, 1–9. Simplicius objects to Porphyry as follows. If “being not-in-a-subject” is predicated only of substances whereas “being in-a-substance” is predicated of accidents, then “complements” seem to belong to no category at all. Since Simplicius automatically equates being not-in-a-subject with substance he does not notice Porphyry's insistence that the substantial complements are precisely not-in-a-subject parallel to substances, so that being not-in-a-subject is characteristic of both.

²¹ Porphyry, In Aristotelis categorias commentarium, ed. Busse, p. 95, 17–20. In the same direction goes Porphyry's famous differentiation between genera and differentiae regarding the fact that the former are ἐν τῷ τί ἐστι κατηγορούμενον whereas the latter ἐν τῷ ποῖόν τί ἐστι κατηγορούμενον (id., Isagoge, pp. 2, 16; 11, 7–8).

²² Porphyry, In Cat., 95, 22–96, 2.

²³ Porphyry, , On Aristotle's Categories, tr. Strange, Steven K. (London, 1992), p. 87, n. 199Google Scholar.

²⁴ Indeed, the sentence “διὸ καὶ εἰς τὸν ὁρισμὸν τῆς οὐσίας παραλαμβάνεται ἡ διαφορὰ ὡς συμπληρωτικὴ οὖσα τῆς οὐσίας” can be interpreted in this way, especially because the assumption that complements are substances is connected to it by a “δὲ”. However one should avoid leaping to conclusions. For one can also suggest that Porphyry is just explaining here why differentiae are parts of the definition of substances next to genera.

²⁵ Simplicius, In Cat., p. 49, 1–9. Cf. n. 16.

²⁶ Ammonius, In Aristotelis categorias commentarius, ed. Busse, Adolf, Ammonius in Aristotelis categorias commentaries, Commentaria in Aristotelem Graeca 4.4 (Berlin, 1895), p. 46, 10–20 Google Scholar. Two points are not quite clear about this report. Firstly, Ammonius probably means Porphyry's account of differentiae when he speaks of a “middle” position. Elias supports this view (Eliae (olim Davidis) in Aristotelis categorias commentarium, ed. Busse, Adolf, Eliae in Porphyrii isagogen et Aristotelis categorias commentaria, Commentaria in Aristotelem Graeca 18.1 [Berlin, 1900], p. 173, 13ffGoogle Scholar). However, John Dillon remarks in his translation of Dexippus’ commentary on the Categories (transl. Dillon, p. 89, n. 50) that the “middle” position (that is also mentioned there) is rather to be ascribed to Plotinus. Secondly, it is not clear enough whether “related to substance” just means “substance”. The opposed case, when differentiae are “related to accidents” (πρὸς τοῖς συμβεβηκόσι), which is also mentioned by Ammonius, supports this suggestion. Ammonius gives an example of “white” in swan as a differentia “related to accident”. However, it is commonly known that “white” is just an accident of swan (although inseparable). Thus, one can conclude, even with this report, that for Porphyry some differentiae such as “rational” and “mortal” are substances.

²⁷ Ammonius, In Cat., pp. 46, 20–47, 5.

²⁸ Ibid ., pp. 45, 17–46, 9.

²⁹ Ibid ., p. 47, 9.

³⁰ John Philoponus, In Aristotelis categorias commentarium, ed. Adolf Busse, Philoponi (olim Ammonii) in Aristotelis categorias commentarium, Commentaria in Aristotelem Graeca 13.1 (Berlin, 1898), pp. 64, 9–69, 11 and Elias, In Cat., pp. 172, 31–174, 23.

³¹ This approach could have been helpful in the Commentators’ project of reconciling Plato and Aristotle. For secondary substances become more ontologically independent in this system. On this see infra in the conclusion of this article. Alexander obviously had different reasons for which see Rashed, Essentialisme, pp. 24–5.

³² G. Endress suggested that it could have been the need to defend his Trinitarian doctrine that pushed Ibn ʿAdī towards discussing the topic ( Endress, Gerhard, The Works of Yaḥyā Ibn ʿAdī: An Analytical Inventory [Wiesbaden, 1977], p. 49 Google Scholar). This suggestion does not seem to be quite convincing. On the one hand, Ibn ʿAdī’s try to prove that heat is not a substance in fire could be considered as a way to help him show that the Son in not a substance in God. However, since Ibn ʿAdī proves in On Fire and Heat that heat is in fire as an accident, it is highly implausible that he would parallel it to the Logos being in the Father.

³³ The first to attract attention to this treatise and point out that it is combined with the Late Antique discussions on the ontological status of the parts of substance was Endress in The Works, p. 49. Then, R. Wisnovsky found a manuscript of the treatise (Robert Wisnovsky, New Philosophical Texts of Yaḥyā ibn ʿAdī: A Supplement to Endress’ Analytical Inventory, in Opwis, Felicitas and Reisman, David [eds.], Islamic Philosophy, Science, Culture, and Religion: Studies in Honor of Dimitri Gutas [Leiden-Boston, 2012], p. 313 Google Scholar). I am grateful to R. Hansberger and R. Wisnovsky for providing me access to the draft of the edition of this MS. The following references will be to its folios and paragraphs as presented in the Hansberger-Wisnovsky forthcoming edition.

³⁴ Yaḥyā b. ʿAdī, On Fire and Heat, fol. 18^r, §5.

³⁵ tawahhum stays in Arabic for ἐπίνοια in Greek cf. Porphyry, Isagoge, p. 13, 1–2 and its translation in Manṭiq Arisṭū, ed. ʿAbd al-Raḥmān Badawī (Beirut, 1980), p. 1086. I render it with “imagining” in order to keep the idea of thought experiment but I do not mean any sense of φαντασία implied here.

³⁶ Ibid ., §10.

³⁷ Ibid ., §13.

³⁸ For instance something can be an accident in Porphyrian sense and still not-in-a-subject in Aristotelian sense – the possibility which Avicenna will cash out, see infra.

³⁹ Ibid ., §11. In another philosophical treatise, Ibn ʿAdī again points out that fire as an element is “a collection of predicables” (ǧumla muʾallafa min maqūlāt) but  adds that it can also be understood as a substance (ǧawhar) (see Yaḥyā b. ʿAdī, Maqālāt Yaḥyā b. ʿAdī al-falsafiyya, ed. Saḥbān Ḫalīfāt [Amman, 1988], p. 190). Unfortunately, he does not explain what these cases are. It is obvious that our case in On Fire and Heat deals with fire as with an element that is not a substance but rather a “collection”.

⁴⁰ Yaḥyā b. ʿAdī, On Fire and Heat, fol. 18^r, §14–16.

⁴¹ Comparing this to Ibn ʿAdī’s famous tripartite system of ontological levels ((1) essence qua universal, (2) essence qua essence, (3) essence qua particular, see Rashed, Marwan, “Ibn ʿAdī et Avicenne: sur les types d'existants”, in Chiaradonna, Riccardo [ed.], Aristotele e suoi esegeti neoplatonici: logica e ontolgia nelle interpretazioni greche e arabe [Bibliopolis, 2004], pp. 107–72Google Scholar) one can suggest that only (2) are substances properly speaking and other two are composites of substance and accidents.

⁴² This is obviously so in the first part of the treatise. However, Ibn ʿAdī comes back to this argument later in support of his second argument and there it seems to function already as an argument (see Ibid., fol. 19^r, §36).

⁴³ Ibn ʿAdī shows that this is his preferred argument when he mentions only it in his discussion in the presence of the vizier Abū al-Qasīm ʿĪsā Ibn ʿAlī Ibn ʿĪsā al-Ǧarrāḥ (see short treatise Yaḫyā b. ʿAdī, Nusḫat mā aṯbatahū Yaḥyā b. ʿAdī li-Abī Bakr al-Ādamī al-ʿAṭṭār, fol. 72^v in Hansberger-Wisnovsky edition). Ibn ʿAdī claims there that the genus of heat – if it were substance – would not include the heat in iron where it is an accident. This argument obviously functions in the same way as the argument to be presented.

⁴⁴ I am not aware of any earlier account of substance and accidents that would anticipate Ibn ʿAdī’s. The closest is the suggestion that Alexander refutes in In Top., p. 50, 12–13. In the modern scholarship, one finds a very similar interpretation of the Aristotelian account of accidents in Ebert, “Aristotelian accidents”, pp. 143–7.

⁴⁵ Yaḥyā b. ʿAdī, On Fire and Heat, fol. 18^v, §27.

⁴⁶ Ibid ., § 22.

⁴⁷ Ibid ., fol. 18^v, §26; or later: “muḥāl an yūǧada maʿnā wāḥid bi-ʿaynihi ǧawharan li-šayʾ mā wa-ʿaraḍan li-šayʾ āḫar” (Ibid., fol. 19^v, §52).

⁴⁸ Ibid ., fol. 18^v, §28.

⁴⁹ Ibn ʿAdī spends a lot of time proving the applicability of this rule to the substance/accident case. E.g. he has to prove that the suggestion that there is something like a universal (kullī) “heat” that can be once accident and once substance does not entail the possibility for heat to be once accident and once substance (Ibid., fol. 19^r, §32ff). Unfortunately, his argumentation is very subtle and long and we have to omit it for the sake of brevity.

⁵⁰ Ibid ., fol. 18^v, §23: “fa-laysa mumkin iḏān an yaṣduqa fī ḏāt wāḥida aʿnī mā yadullu ʿalayhi ḥadd wāḥid annahu fī mawḍūʿ wa-annahu lā fī mawḍūʿ”.

⁵¹ Notice the use of the same example the Porphyrian tradition used to illustrate when heat is not something substantial but accidental.

⁵² Ibid ., fol. 18^v, §24. Though Porphyry himself accepted substantial accidents because he had a middle way between substances and accidents. However later tradition led his analysis towards its logical end where no substantial accidents were possible any more (unless in different respects). “Substantial accidents” should not be confused with Aristotelian “essential accidents” (συμβεβηκότα καθ’αὑτά, see Met. Δ, 30, 1025a30–34.).

⁵³ Apart from Abū Bakr al-Ādamī al-ʿAṭṭar, who is mentioned in a short treatise on this topic (Ibn ʿAdī, Nusḫa), Ibn Suwār mentions also Abū Isḥāq Ibrāhīm Ibn Bakkūš as Ibn ʿAdī’s rival and Abū Sulaymān Muḥammad b. Ṭāhir as Ibn ʿAdī’s possible supporter on the topic of the “form of fire” (ṣūrat al-nār) (see Endress, Gerhard, “Ibn al-Ḫammār,” in Rudolph, Ulrich, Philosophie in der islamischen Welt, Bd 1, 8.–10. Jahrhundert [Basel, 2012], p. 33 Google Scholar).

⁵⁴ Georr, Khalil, Les Catégories d'Aristote dans leurs versions syro-arabes (Beirut, 1948), pp. 373–7Google Scholar (hereafter called Ibn Suwār, In Cat.). For the very convincing hypothesis which identifies this treatise with that mentioned in margine see Endress, The Works, p. 49.

⁵⁵ Ibn Suwār, In Cat., p. 373.

⁵⁶ Ibid ., p. 374.

⁵⁷ Ibn ʿAdī regards these kind of arguments as dialectic (al-ǧadaliyya) although he also agrees to play according to these rules and tries to show that Aristotle is on his side (Yaḥyā b. ʿAdī, On Fire and Heat, fol. 18^v, §17).

⁵⁸ Ibn Suwār, In Cat., p. 374.

⁵⁹ Ibid ., pp. 376–7.

⁶⁰ Ibid ., p. 377.

⁶¹ Ibid ., pp. 376–7. Ibn Suwār ascribes the same view to Ammonius. Moreover, he presents this linguistic interpretation as a quote from Ammonius. However, the source of this quote hasn't been identified. Anyway, it is hardly possible that Ammonius could have interpreted the Greek “ἐν ὑποκειμένῳ δὲ οὐδενί” in the same way as Ibn Suwār did the Arabic “fī mawḍūʿ mā”. For the Greek does not allow for two interpretations in the same way as the Arabic does.

⁶² Ibn ʿAdī’s correct understanding of the definition can be explained through his knowledge of Syriac. For the Syriac translation of “ἐν ὑποκειμένῳ δὲ οὐδενί” as “ba-medem dēn da-sīm” (Georr, Les Categories, p. 254, 6) does not leave any room for Ibn Suwār's interpretation either since medem always means “any” or “none” as in this context. Another Syriac translation puts it differently but agrees with Ibn ʿAdī’s interpretation (see King, Daniel, The Earliest Syriac Translation of Aristotle's Categories [Leiden-Boston, 2010], p. 96, 23–24 CrossRefGoogle Scholar).

⁶³ Abū al-Faraǧ b. al-Ṭayyib, Tafsīr Kitāb al-Maqūlāt, ed. Ferrari, Cleophea, Der Kategorienkommentar von Abū l-Faraǧ ʿAbdallāh Ibn aṭ-Ṭayyib (Leiden-Boston, 2006), p. 159, 21–23 Google Scholar.

⁶⁴ Avicenna, Kitāb al-Šifāʾ, al-Ilāhiyyāt, ed. Georges Anawatī and Saʿīd Zāyid (Cairo, 1960), II, 1, p. 48, 15.

⁶⁵ Avicenna, Kitāb al-Šifāʾ, al-Manṭiq, al-Maqūlāt, ed. Georges Anawatī, Maḥmūd al-Ḫuḍayrī, Aḥmad al-Ahwānī and Saʿīd Zāyid (Cairo 1959), I, 6, pp. 45, 16–46, 1.

⁶⁶ Ibid ., p. 46, 2–3.

⁶⁷ Ibid ., p. 46, 4–7.

⁶⁸ Ibid ., p. 48, 1–3.

⁶⁹ Avicenna, , The Metaphysics of Healing, tr. Marmura, Michael (Provo, 2005), p. 46, 20–28 Google Scholar.

⁷⁰ Maqūlāt I, 3. This issue deserves a separate study. To summarize: form is in matter not as an accident in-a-subject because matter needs form in order to exist. It was very much discussed by Alexander (Quaestio I, 8; I, 16, I, 26, on them see Rashed, Essentialisme, pp. 44ff).

⁷¹ Dexippus, In Aristotelis categorias commentarium, ed. Busse, Adolf, Dexippi in Aristotelis categorias commentarium, Commentaria in Aristotelem Graeca 4.2. (Berlin, 1888), p. 49 Google Scholar, 10ff – where the Porphyrian doctrine is reported – is the closest passage to this argument. Avicenna's reference to the fact that ‘differentiae are substances’ was  something known from the antique tradition (this is what the expression samiʿu (Maqūlāt I, 6, p. 46, 16) implies) could be a sign that he knew somebody from the Ammonian tradition that had claimed this.

⁷² Maqūlāt I, 6, p. 46, 8–13.

⁷³ Ibid ., p. 46, 16–19.

⁷⁴ Ibid ., p. 47, 8–13.

⁷⁵ Note the difference to Ibn ʿAdī’s linguistic analysis of the definition of substance. He thinks that “mā” in “mawḍūʿ mā” means “any” and on these grounds concludes that substance cannot be in any subject. On the contrary, Avicenna does not have the particle “mā” in his definition of substance and derives the same meaning of “never” as by Ibn ʿAdī from the expression “al-battata” that can be both interpreted as “never” and “not in any respect”.

⁷⁶ The expression substance in itself or accident in itself is used e.g. in Maqūlāt I, 6, p. 48, 6, 13 etc.

⁷⁷ Avicenna already provides this argument in Maqūlāt I, 4, p. 34, 5–8 and 17–18. His position there seems to relay once again on Ibn ʿAdī. For Avicenna agrees that one should differentiate between the subject of accident and the composite. Being in a composite is not like being in-a-subject. A composite (murakkab) consists of a substance and an accident. The accident exists in the composite as its part while being in the substance as an accident. This is indeed how Ibn ʿAdī’s theory works as was established above.

⁷⁸ This is a summary of the argument presented in Ibid., pp. 48, 5–49, 12. Though such parts of substances are substances in Avicenna one should not consider them as independent parts cf. Benevich, Fedor, “Die göttliche Existenz: Zum ontologischen Status der Essenz qua Essenz bei Avicenna”, Documenti e studi sulla tradizione filosofica medievale, XXVI (2015): 103–28, p. 121Google Scholar.

⁷⁹ This argument is further evidence for Avicenna's criticisms being directed at Ibn al-Ṭayyib. The latter wrote in his commentary on Categories that differentiae are in some respect accidents because they inhere in genus as in-a-subject. This is what Avicenna tries to escape here.

⁸⁰ Maqūlāt I, 6 p. 49, 13–15.

⁸¹ Ibid ., pp. 49, 20–50, 4.

⁸² Cf. Ibid., pp. 50, 12–51, 4–5: “al-šayʾ laysa ǧawharan bi-al-qiyās ilā šayʾ wa-in kāna ǧawhariyyan bi-al-qiyās ilā al-šayʾ allaḏī huwa fīhi.”

⁸³ Avicenna expresses the same idea already in Maqūlāt I, 3 p. 26, 5–8 although he hasn't proven it there yet. Previously he says that his opponents have confused being substantial and being substance while speaking about the notions of “maqūl ʿalā al-mawḍūʿ” and “maqūl fī al-mawḍūʿ” (Maqūlāt I, 3, p. 23, 4–6). According to Avicenna they claimed that only complements of substance (muqawwimāt) can be said “maqūl ʿalā al-mawḍūʿ” whereas accidents cannot. For accidents are all accidental and cannot be predicated of their subject as universals. To the contrary, Avicenna argues that accidents can also be called “maqūl ʿalā al-mawḍūʿ” when we speak about universal accidents (like universal “white”). However, it seems that Avicenna is rather unfair in this accusation. For e.g. Yaḥyā b. ʿAdī, who obviously recognizes the identity of accidental and accidents, also speaks of universal accidents in his On Fire and Heat, fol. 18^v, §22. What Avicenna probably implies is that if substantial (which is extensionally equal to “maqūl ʿalā al-mawḍū”) means substance, then the Porphyrians have problems with substantial attributes of accidents (like “color” for “white”).

⁸⁴ Accordingly Avicenna has three types of accidental: accidents par excellence, necessary concomitants (lawāzim) and essentially necessary concomitants (alias “essential accidents”, al-aʿrād al-ḏātīya), cf. Avicenna, Kitāb al-Šifāʾ, al-Manṭiq, al-Madḫal, ed. Georges Anawatī, Maḥmūd al-Ḫuḍayrī and Aḥmad al-Ahwānī (Cairo, 1952), p. 32.

⁸⁵ Maqūlāt I, 4, p. 37, 5–6.

⁸⁶ Ibid ., p. 32, 18.

⁸⁷ Avicenna, Madḫal I, 5, p. 32, 10. Example is Avicenna's.

⁸⁸ Ibid ., I, 6, pp. 34–5. By conceptual inseparability I mean here that conceiving of P is a necessary condition for conceiving of S. Otherwise, some immediate concomitants would be conceptually inseparable too as Avicenna states while denying that they are necessary conditions for conceiving of essences. For the synonymy of “essential” and “substantial” see infra.

⁸⁹ Maqūlāt I, 4, pp. 32, 17–33, 9.

⁹⁰ Cf. the idea of dependence of accidents on substances in Rashed, Marwan, Al-Ḥasan Ibn Mūsā al-Nawbaḫtī: Commentary on Aristotle “De Generatione et Corruptione” (Berlin, 2015), pp. 108–14CrossRefGoogle Scholar.

⁹¹ Going back to the starting point of this article – this is how Avicenna establishes that soul is a substance in Nafs I, 3. Note that both Avicenna and Ibn ʿAdī disregard the problems of induction. If a scientist decides that he knows all the cases and in all these cases something is not-in-a-subject he is able to conclude that it is a substance. It is always possible, though, that some particular case where it is in-a-subject escapes the scientist's attention.

⁹² Maqūlāt I, 6, p. 47, 8–13.

⁹³ This only applies to the differentiae that Avicenna calls “derivative” or “simple” like “rationality” (nuṭq) as opposed to “univocal” and “composite” differentiae like “rational” (nātiq). Whether the latter are substances or not is dependent on their bearers. On this see Vincenzo, Silvia Di, “Avicenna against Porphyry's definition of differentia specifica ,” Documenti e studi sulla tradizione filosofica medievale, XXVI (2015): 129–84, pp. 162–3Google Scholar.

⁹⁴ One should distinguish between two kinds of composites: C is composed of S and P where (1) S is matter (or like matter) and P is form (or like form). These composites are substances and their parts are substances (cf. Ilāhiyyāt V, 3, p. 214, 15); (2) S is substance and P is an accident like chair made of wood and a particular form or snubnose consisting of nose and concavity. Avicennas explicitly says that these are not like matter and form. Such composites are not substances (Maqūlāt I, 4, p. 34, 17–20).

⁹⁵ Ilāhiyyāt V, 8, p. 244–5. Here concavity and nose are mentioned again.

⁹⁶ Maqūlāt I, 6, p. 50, 15: fa-yakūnu al-ǧawharī makān al-ḏātī. The definition of “essential” as “integral part of essence” corresponds to Avicenna's definition of essential in Madḫal I, 5–6. However this is not the only understanding of “essential” in Avicenna. Apart from it there is also: Burhān definition of essential found in Avicenna, Kitāb al-Šifāʾ, al-Manṭiq, al-Burhān, ed. ʿAfīfī, Abū al-ʿAlā (Cairo, 1952), II, 2, p. 125–7Google Scholar; “primacy” definition of essential found in Burhān II, 2, p. 128 and e.g. in Avicenna, , Al-Išārāt wa-al-tanbīhāt maʿa šarḥ Naṣīr al-Dīn al-Ṭūsī, ed. Dunyā, Sulaymān, vol. 1 (Cairo, 1957), pp. 170, 5–173, 3Google Scholar.

⁹⁷ The very notion of “addition” implies the meaning of being dependent. This was apparently one of the reasons why the Porphyrian tradition started to define accident through its accidental character. The other reason should have been Aristotle himself (see Barnes, Porphyry, pp. 220–4).

⁹⁸ Avicenna, Kitāb al-Naǧāt min al-ġarq fī baḥr al-dalālāt, ed. Dānišpažūh, Muḥammad T. (Tehran, 1985), p. 12, 9–10 Google Scholar.

⁹⁹ Ilāhiyyāt III, 3, p. 106, 15–17.

¹⁰⁰ Avicenna explains what “white” means in Maqūlāt I, 15, p. 41, 15 and II, 1, p. 58, 3ff.

¹⁰¹ Ibid ., p. 50, 6–7.

¹⁰² Avicenna himself follows the Periptatetic tradition accepting that being is said of categories by analogy (bi-al-taškīk). On this see Alexander Treiger, “Avicenna's notion of transcendental modulation of existence (taškīk al-wuǧūd, analogia entis) and its Greek and Arabic sources”, in: Opwis and Reisman (eds.), Islamic Philosophy, pp. 327–65.

¹⁰³ Maqūlāt II, 1, p. 61, 2–14. In Išārāt, p. 155 he uses the same argument in order to establish that existence is not something essential for essence.

¹⁰⁴ By this he probably means proving the accidentality of numbers etc. see Stephen Menn, “Avicenna's metaphysics”, in: Adamson, Peter, Interpreting Avicenna: Critical Essays (Cambridge, 2013), pp. 143–69, p. 162CrossRefGoogle Scholar. Menn rightly demarcates between the Isagoge (predicable) and Categories (predicamental) sense of “accident” which is crucial for the present article.

¹⁰⁵ Maqūlāt II, 2, pp. 65, 12–66, 2. A good example could also be the parts of substance that are the topic of this paper.

¹⁰⁶ Ibid ., II, 2, p. 65, 13: “wa-ʿalā anna ḏātahu taqtaḍī hāḏihi al-nisba” and Ibid., 16–17: “ʿalā annahu fī ḏātihi bi-ḥayṯ lā budda lahu min mawḍūʿ”.

¹⁰⁷ Ibid ., II, 1, p. 61, 17–62, 1: “wa-in kāna qad yakūnu min al-ʿaraḍī mā ḥuṣūluhu laysa bi-ʿillat ḫariǧa ʿan al-māhiyya bal takūnu al-māhiyya mūǧiba lahu wa-muqtaḍiya ayyāhu.” This sense of essential is extensionally identical with the scientific essentiality of Burhān II, 2.

¹⁰⁸ Ibid ., II, 2, p. 66, 2–3: “fa-nisbat al-ʿaraḍ ilā hāḏih i nisbat al-mawǧūd ilā māhiyyāt min ḥayṯu laysa dāḫilan fī al-māhiyya.”

¹⁰⁹ It is worth noticing that this conclusion applies first of all to Avicenna's theory of substance as found in his Šifāʾ. On the contrary, we cannot find any traces of the discussion about fire and heat in his earlier work al-Muḫtaṣar al-awsaṭ (see Kalbarczyk, Alexander, “The Kitāb al-Maqūlāt of the Muḫtaṣar al-awsaṭ fī l-manṭiq: a hitherto unknown source for studying Ibn Sīnā’s reception of Aristotle's Categories ”, Oriens, 40 (2012): 305–54, pp. 329–32CrossRefGoogle Scholar), although it was obviously Avicenna's vorlage for several topics discussed in Maqūlāt.

¹¹⁰ Thus: If X is substance and Y is accidental for it, removing Y does not harm X (= accidentality of Y), whereas removing of X eliminates Y (= ontological dependence of Y). Following this logic one can easily imagine the same procedure for substantial parts so that the following is valid: if X is substance and Y is substantial for it, removing of Y eliminates X (= substantiality of Y), whereas removing of X does not harm Y (= ontological independence of Y). It follows that primary substances are the accidental instantiations of their substantial constituents and that these constituents are ontologically more independent that the substances. This is the conclusion that al-Fārābī derives from the Porphyrian approach to substantiality and accidentality ( Al-Fārābī, , Kitāb al-Ḥurūf, ed. Mahdi, Muhsin [Beirut, 1969], pp. 103, 12–104, 5Google Scholar).

¹¹¹ See e.g. Rashed, “Ibn ʿAdī et Avicenne” and Menn, “Avicenna's metaphysics”, pp. 154–7 on Ibn ʿAdī’s influence on Avicenna's theory of universals. Exactly as in our case Avicenna takes Ibn ʿAdī’s ideas as a starting point and develops them in what becomes one of the main new ideas of his metaphysics.