• 1 Prior Analytics, ii.27, 70b7-70b38, see especially Ross’s commentary, W.D. Ross, (ed.) (1949), Prio (...)
• 2 See Leslie A. Zebrowitz (1998), Reading Faces. Window to the Soul?, Boulder, CO, Westview Press, p. (...)
• 3 Prior Analytics, ii.27, 70b. The sign is discussed at 70a and defined as “a demonstrative premise w (...)
• 4 However, see Lisa Devriese (2021), “First Medieval Attestations of the Physiognomica,” in Lisa Devr (...)

1The huge extremities of a lion may be connected to his courage, Aristotle says in his Prior Analytics, and allow for generalization by inferring from bodily peculiarities to character traits connected or associated with them by way of a physiognomical syllogism.1 That is, having a huge forearm allows one to conclude that the forearm’s non-lion bearer is courageous. Required for this is choosing a certain affection that is typical for a group universally along with its correspondent sign, and then, based on the presence of this sign among members of another group that does not generally have it, to judge or, rather, infer the presence of the affection that is connected with this sign and always found among the members of the first group. This is done, as in Aristotle’s Prior Analytics and certain works of Avicenna, especially with animal species forming the first group and humans as forming the second.2 In Aristotle’s example, this includes courage for all lions, large extremities for all lions, and the inference from large extremities as found in an individual human to the presence of his or her courage.3 Note that physiognomy is thus, already in Aristotle, not limited to only facial expressions.4

• 5 See Martin Porter (2005), Windows of the Soul. The Art of Physiognomy in European Culture 1470-1780(...)
• 6 Also noted by Mariska Leunissen (2018), “Physiognomy,” in Paul T. Keyser & John Scarborough (eds.), (...)
• 7 70b7, εἴ τις δίδωσιν.
• 8 70b11-12, εἰ δὴ τοῦτό τε δοθείη καὶ ἓν...εἶναι.

2This kind of first-figure syllogism5 in its Aristotelian form has also influenced Avicenna. However, Aristotle could, at first sight, be understood as raising—or himself having—a few doubts about this syllogism, as he employs conditional phrases here.6 The first of them concerns the alteration of both body and soul under affections.7 He continues with another conditional on the singular or unique representation of an affection by a respective sign.8

• 9 Based on citations ascribed to Aristotle on physiognomy in al-Dimashqī that differ from the Greek t (...)
• 10 805a20-31, ps.-Aristotle (2007), Fisiognomica, Maria Fernanda Ferrini, (tr.), Milan, Bompiani, p. 1 (...)
• 11 This text especially influenced the Greek writer Adamantius and an anonymous Latin writer (ed. Jacq (...)
• 12 809b36-810a10. The panther is also found in the Arabic version of Polemon’s treatise, ed. Hoyland [ (...)
• 13 See the discussion in Anna Akasoy (2008), “Arabic Physiognomy as a Link between Astrology and Medic (...)
• 14 An edition of Barhebraeus’s Prior and Posterior Analytics is being prepared by the author.
• 15 See Johannes Thomann (1996), “Avicenna über die physiognomische Methode,” in Rüdiger Campe & Manfre (...)

3The peculiar signs in animals for certain character traits are also discussed in an independent ancient treatise on physiognomy that has long been ascribed to Aristotle.9 Besides also including a conditional sentence similar to Aristotle regarding the validity of physiognomical inferences (805a17-18), the realm of signs is not limited to animals only, but also includes ethnological and habitual signs.10 Another well-received physiognomical text, which includes a long list of these signs, was that by Polemon (2^nd ct. CE).11 Both texts were translated into Arabic, and at the very least it would seem that the pseudo-Aristotelian Physiognomics had a (direct or indirect) influence on Avicenna’s treatment of the physiognomical syllogism, which manifests in his using the example of the panther in addition to the lion.12 Polemon's text also found its way into Barhebraeus’s Syriac summa Cream of Wisdom, though in dealing with physiognomy proper in the ethical section, as already noted by Furlani,13 while Barhebraeus’s treatment of the physiognomical syllogism itself in his Prior Analytics follows Avicenna and Aristotle.14 Polemon’s text, in turn, influenced several Arabic philosophers. However, in this respect, an influence also by al-Fārābī on Avicenna, in addition to Aristotle, has also been suggested by Thomann.15

• 16 See, for example, Martin Porter (2005), Windows of the Soul.
• 17 As might be seen in Johann Caspar Lavater (2018), Physiognomische Fragmente. Zur Beförderung der Me (...)
• 18 See Anna Akasoy (2008), “Arabic Physiognomy” for other relevant works which have received attention(...)

4Physiognomy in general still played a role in later centuries, not only in the West,16 and even well after the Renaissance,17 but also in the Arabic tradition.18

• 19 Ii.27; Healing, Prior Analytics, p. 579-580.

6The very last part of Avicenna’s Prior Analytics, that is, al-Qiyās, in his summa The Healing (al-Shifāʾ) shares many parallels with Aristotle’s last sections of the second book of his Prior Analytics with respect to the sequence of topics dealt with. Among these similarities is the physiognomical syllogism (qiyās al-firāsa) that concludes both works.19

• 20 Salvation, vol. i, p. 76.

7A shorter discussion is also found in Avicenna’s shorter summa The Salvation (al-Najāh), where it is called al-qiyās al-firāsī with an adjectival form.20 It is likewise the very last part of the section corresponding to the Prior Analytics.

• 21 Johannes Thomann (1996), “Avicenna über die physiognomische Methode,” p. 53-61.
• 22 See Johannes Thomann (1996), “Avicenna über die physiognomische Methode,” p. 57, for a narrower usa (...)
• 23 Healing, Prior Analytics, p. 579.

8An initial investigation into and comparison of these two versions has already been provided by Thomann,21 who notes that, both summae offer a slightly different classification of the physiognomical syllogism. For while the Healing places it among syllogisms related to example or analogy (tamthīlī) and sign (ʿalāmī), the Salvation considers it as similar to both demonstrative sign (dalīl) and example (tamthīl), respectively, though even uses here dalīl.22 The Healing’s version has a broadening in this respect, as bodily affections (infiʿālāt) and dispositions (hayʾāt) are first said to be demonstrative signs (dalāʾil) of psychological affections, but, once a certain psychological affection of an animal is known and it is investigated which particular bodily disposition (ḥāl) is connected with it, the bodily disposition (hayʾa) is made a (general) sign (ʿalāma) for the psychical dispositon (hayʾa). The bodily sign (ʿalāma) is then exemplified by the lion’s large extremities.23 The example is amplified by the lion’s also having a magnanimity that is lacking in the panther, despite its also having large extremities. The panther, therefore, is only courageous.

• 24 Healing, Prior Analytics, p. 579.
• 25 Healing, Prior Analytics, p. 580. This might be influenced by Aristotle’s remark on conversion at t (...)

9The middle term, according to the Salvation’s version, has to be a bodily disposition that a human has. The Healing’s version explains that, after having come to know a certain animal’s psychical disposition, such as courage of the lion, one has to seek out a peculiar accompanying bodily state in order to make this bodily disposition (hayʾa) a sign (ʿalāma) for the psychological disposition.24 The Salvation’s version is the only one to list the terms of the syllogism, though without using formal variables, as Aristotle does at the very end of his discussion (70b). The version in the Healing, on the other hand, concludes with the statement that the syllogism is reducible to the previously mentioned types (وترجع إلى القياسات المذكورة).25

• 26 The following occurrences should be seen as an addition to Johannes Thomann (1996), “Avicenna über (...)
• 27 Middle Compendium, p. 242-243.
• 28 See Anna Akasoy (2008), “Arabic Physiognomy,” p. 125, for Galen’s influence on humoral mixtures or (...)
• 29 Middle Compendium, p. 243. Both here, in the Guidance to Logic, and in the Healing, it is called fi (...)
• 30 Middle Compendium, p. 243.
• 31 Healing, Prior Analytics, p. 580.
• 32 See also Dimitri Gutas (2014), Avicenna and the Aristotelian Tradition. Introduction to Reading Avi (...)

10Avicenna's treatment of the physiognomical syllogism, however, is not limited to these two works.26 In the recently edited Middle Compendium on Logic (al-Mukhtaṣar al-awsaṭ fī l- manṭiq), it appears at the very end of the part on the Prior Analytics, after example (tamthīl), objection (muqāwama), probability (raʾy), and sign (ʿalāma), though independently in its own paragraph and with the aforementioned explanation of finding a corresponding sign (ʿalāma).27 The treatment is quite Aristotelian, beginning with the usual conditional that when the fact that character traits follow natures and (humoral) temperaments,28 as bodily dispositions do, is conceded, then physiognomy will be granted. After the familiar lion example, he concludes with the often mentioned requirement that the middle term in a physiognomical syllogism29 be wider than the minor and equal with the major term, though without the statement on reducibility. He also includes the instruction to look into determining which two signs (ʿalāmatayn) follow which two meanings.30 This instruction also appears in the Healing,31 to which the whole passage is quite similar.32

• 33 This text is, for example, found neither in Gutas’s or Mehdevi’s lists of works, nor does the edito (...)
• 34 Guidance to Logic, vol. ii, p.282.
• 35 Also numūw, “growth,” p. 283, 11, should be corrected into nimr, “panther.”
• 36 See Johannes Thomann (1996), “Avicenna über die physiognomische Methode,” p. 57, and the discussion (...)

11However, the Guidance to Logic (al-Hidāya fī l-manṭiq), which has also only recently been edited, though its authenticity is not fully ascertained,33 includes it as well. It is treated together with the (demonstrative) sign (dalīl and ʿalāma) in the final section of the part on the Prior Analytics. The text is nearly identical to that of the Middle Compendium. Apart from the fact that he opens with the statement that the physiognomical syllogism is not among the set of syllogisms with a sign and an example (tamthīliyya ʿalāmiyya),34 where the copyist's or editor's negation may be an error,35 this is in line with terminological classification in the Healing and the Middle Compendium, though against the version in the Salvation.36 Another addition not in the Middle Compendium is the final sentence familiar from the Healing, namely, that the syllogism is reducible to other aforementioned types.

• 37 Guidance, p. 114, n° 42.

12The topic is, however, not dealt with in his shorter summa Guidance (al-Hidāya), even though he treats the example or analogy (tamthīl) very briefly at the end of the part corresponding to the Prior Analytics,37 giving the usual example of judging the heavens’ createdness via analogy from a house's being created.

• 38 Pointers, vol. i, p. 495.
• 39 Pointers, vol. i, p. 453.
• 40 Pointers, vol. i, p. 365.
• 41 Pointers, vol. i, p. 369 (ḍaʿīf). This has, for example, already been noted by al-Sayyed Aḥmad (198 (...)

13Even though Avicenna ends the section on logic in his Pointers and Admonitions (al-Ishārāt wa-l-tanbīhāt) with fallacies and non-syllogistical inferences,38 the respective logical part that is by context similar to the end of the Prior Analytics ends with the syllogism that proves by absurdity.39 He doesn’t explicitly consider the physiognomical syllogism there, rather, example and induction are placed before the discussion of the syllogistic figures;40 and the conclusion by analogy is considered “weak.”41

• 42 Or, rather, Elements, as suggested by Dimitri Gutas (2014), Avicenna and the Aristotelian Tradition(...)
• 43 Correspondingly, it is also not added by Fakhr al-Dīn al- Rāzī in his commentary on that work, ed. (...)
• 44 Logic of the Easterners, p. 74.

14Nor does Avicenna include it in his Sources42 of Wisdom (ʿUyūn al-ḥikma), although inference by example is dealt with at the end of the section of the Prior Analytics,43 nor in his technical Logic of the Easterners (Manṭiq al-Mashriqīn), where he voices concerns about example (tamthīl) as not yielding true definition (taʿrīf) and prone to grave errors.44

• 45 Book on Wisdom, tr. Achena & Massé, pt. i, p.104. Theologians and jurists are also mentioned as emp (...)
• 46 Book on Wisdom, tr. Zabeeh, p. 46-47.

15Likewise, it doesn’t occur in the Book on Wisdom (Dānesh-nāme), his earlier Persian summa, although he explicitly considers several non-syllogistic inferences within the section on logic, such as likely inferences from experience and observation—a context that would be well suited for a discussion of the physiognomical syllogism. He assigns syllogisms by analogy to the realms of politics and law.45 However, he ends the logical section with advice for “avoiding fallacies,”46 by which he diverges from the Aristotelian structure.

• 47 Sālim’s edition deals with the Poetics, while another edition by Ṣālih adds other parts, though not (...)
• 48 Philosophy for ʿArūḍī, ed. Sālim, p. 16.
• 49 Anna Akasoy (2008), “Arabic Physiognomy,” p. 137 (based on chapter headings). There’s, however, one (...)

16The extant manuscript of the Philosophy for ʿArūḍī (al-Ḥikma al-ʿArūḍiyya) is lacking, among others, the corresponding part on the Prior Analytics.47 Therefore, the question of whether or not it includes the physiognomical syllogism cannot be answered at this time. Avicenna does, however, mention courage as commonly ascribed to the lion in the extant part on the Poetics in that work.48 But he neither includes physiognomy in general nor the physiognomical syllogism in particular (as being for example, employed in this science) in his Canon of Medicine (al-Qānūn fī l-ṭibb).49

• 50 Divisions of Sciences, 110. See also Youssef Mourad (ed.) (1939), La Physiognomie arabe, p. 23, and (...)
• 51 See Henri Corbin (1960), Avicenna and the Visionary Recital, Princeton, Princeton University Press (...)

17However, in his Divisions of the Sciences (Fī ʾaqsām al-ʿulūm), Avicenna includes physiognomy as a branch science among the natural sciences, being concerned with inference (istidlāl) from physical constitution (khalq) to character (akhlāq).50 But he doesn’t go into further detail regarding the role of logic in this inference. Avicenna also mentions physiognomy in his earlier Ḥayy Ibn Yaqẓān (lit., Alive, Son of Awake), though more in a mystical respect and without mentioning the syllogism.51

• 52 See Dimitri Gutas (2014), Avicenna and the Aristotelian Tradition, p. 79, for Avicenna’s copying of (...)

18Roughly, then, there are two distinct versions that deal with the physiognomical syllogism, a longer one as is similarly found in the Healing, the Middle Compendium, and the Guidance in Logic, and a distinct, shorter one as found in the Salvation.52 Both versions feature the lion and the panther, which possibly goes back to the ps.-Aristotelian Physiognomics, in all occurrences. The majority of Avicenna’s extant logical works, however, include no mention of it at all.

• 53 See Dimitri Gutas (2014), Avicenna and the Aristotelian Tradition, p. 165, for a chronological over (...)

20A first hypothesis would be to assume a change in attitude that evolved toward a disfavoring of or disbelief in this syllogism, with Avicenna intentionally dropping it in later works. However, while this assumption would apply to the Pointers and the Book of Wisdom, which was, according to Gutas, written later than the Healing, it cannot be upheld as a general rule, since the Sources, which also omit physiognomical inferences, might, according to Gutas, predate the Healing,53 which includes it.

• 54 See the discussion by Anna Akasoy (2008), “Arabic Physiognomy,” p. 137, nt. 83, of two diverging in (...)

21Nonetheless, in all occurrences Avicenna includes the Aristotelian conditional sentence that might be understood as expressing doubt for this kind of syllogism. Moreover, he cautions against using the syllogism by analogy in the later Logic of the Easterners, also raising distrust in both analogy and induction in the Pointers. Therefore, a general personal disfavoring of physiognomics on his part might be possible.54

• 55 See Dimitri Gutas (2014), Avicenna and the Aristotelian Tradition, p. 323-334, for a shift toward m (...)

25Then again, nearness to Aristotle as well as length could both impact the possibility of inclusion. The omission might then also be an expression of a shift away from Aristotle. Such a general shift has been assumed by Gutas.55

• 56 Commentaire sur le livre Lambda, ed. Geoffroy/Janssens/Sebti.

26Avicenna might, therefore, have also dealt with it in the lost part on logic in the Fair Judgment (al-Inṣāf), as this work is, at least regarding the extant parts,56 closer to Aristotle’s text and of a certain length. Thus, it seems likely that he may have treated it there as well.

• 57 However, this might be far-fetched. Formal variables, however, are, for example, also employed for (...)

27The omission of formal variables in the discussion of the physiognomical syllogism, in contrast to Aristotle, could also indicate a perceived lack of formal conclusiveness, though a listing of the terms employed is found in the version in the Salvation.57

• 58 Ed. Mourad.
• 59 Foundation for Acquiring Knowledge, vol. i, p. 258, which also concludes the section on the Prior A (...)

28The omission from the Pointers, a text which attracted frequent commentary, is also mirrored in, for example, two important later commentators of this work, Naṣīr al-Dīn al-Ṭūsī and Fakhr al-Dīn al-Rāzī. The latter, however, dedicates a separate treatise to physiognomy,58 while Naṣīr al-Dīn al-Ṭūsī includes this syllogism in his own Persian logical treatise Foundation for Acquiring Knowledge (Asās al-Iqtibās),59 which was also translated into Arabic. Therefore, the non-inclusion of the physiognomical syllogism in their respective commentaries on the Pointers does not denote the commentators’ disinterest in the topic, but, rather, the fact that they followed the original text at that point.

• 60 Abridgment of Philosophy, vol. i, p. 552.
• 61 Well-Reflected Book, vol. i, p. 250.
• 62 On physiognomy in later Arabic encyclopedic works in the context of magic, see Anna Akasoy (2008), (...)

29However, despite Fakhr al-Dīn al-Rāzī’s treatment in an independent treatise, it’s not part of his summa the Eastern Investigations (al-Mabāḥith al-mashriqiyya), as there’s no full part covering the Prior Analytics. The same applies partially also to his influential summa the Abridgment of Philosophy and Logic (al-Mulakhkhaṣ fī l-ḥikma wa-l-manṭiq), even though he there mentions physiognomy as a science that makes an inference (istidlāl) from actions to character,60 which is similar to Avicenna’s definition in the Division. Similarly, the Well-Reflected Book on Philosophy (al-K. al-muʿtabar fī l-ḥikma) by Abū l-Barakāt al-Baghdādī mentions physiognomy briefly as a natural science concerned with animal attributes, though refuses to include further discussion on it in this book.61 The Renewal in Philosophy (al-Jadīd fī l-ḥikma) by Ibn Kammūna doesn’t include logic as found in the Aristotelian Analytics. However, it is also omitted from the summa Unveiling of the True Natures (Kashf al-ḥaqāʾiq) by Athīr al-Dīn al-Abharī, even though he dwells at length on the syllogism, that is, this omission is intentional rather than due to limitations of its content.62 Omissions such as these might be due to Avicenna’s own omission from the Pointers as well as his negative verdicts on the syllogism by analogy, or by other philosophers on the physiognomical inference itself.

• 63 See above.
• 64 This might be similar to what Martin Porter (2005), Windows of the Soul, p. 63, suggested as an att (...)

30This certain tendency to omit or exclude the physiognomical syllogism from philosophical works does not imply a direct general rejection of syllogisms using examples or analogies. However, Avicenna disfavors them (as explicitly stated in the Logic of the Easterners and the Pointers63) and perhaps doesn’t consider them sufficiently demonstrative (despite his distinguishing usage of demonstrative signs) to be included in the philosophical canon (as following from his statement in the Book of Wisdom). This would apply even more so to the physiognomical syllogism, as, besides concluding by (cross species) analogy, the attribution of dispositions as (general or even demonstrative) signs to character traits is also highly uncertain.64

• 65 See Martin Porter (2005), Windows of the Soul, p. 56, for a historical approach in the anonymous La (...)
• 66 An example might be Avicenna’s Aristotelian statement in his Animals, p. 64, that “all blooded walk (...)
• 67 This project has received funding from the European Research Council (ERC) under the European Union (...)

31However, as the physiognomical syllogism can be formulated by using general variables as Aristotle does, while Avicenna at least lists the respective terms of the syllogism in his Salvation, this is perhaps not due to a lack of formality in the concluding itself, but, rather, to the uncertainty in arriving at the content of these particular variables. For the determination of the corresponding bodily affections and psychological dispositions and their determination by analogy between members of different species is the primary contributor to uncertainty,65 while still being expressible in formal logical language. However, one wonders whether syllogisms with examples or analogy are perhaps implicitly employed—or at least present—for conclusions in his Book on Animals, rather than inductions only, which he also disfavors. This would at least apply to the descriptive parts of the book. These conclusions by induction or analogy wouldn’t have to be stated, but could, perhaps, be seen in stated generalizations across species, such as, for example, rules or generalizations for viviparous quadrupeds in general.66 They seem to be part of an Aristotelian lore.67