1 Throughout this paper, all Arabic texts will be cited initially by their Arabic titles. Since no universally recognized English translation exists for many of these works, an English translation for all the titles will be given in parentheses. In subsequent citations, a shortened, italicized version of the English translation will be used. Where medieval Latin versions of Arabic texts are available, I have in most cases provided references to these as well as to the Arabic original. References are also provided in initial citations to translations of the Arabic text into Western languages other than English, but subsequent citations are not provided to these translations.

2 I accept Rahman's contention that each of the internal senses is a “differentiation of Aristotle's phantasia.” See Rahman, Fazlur, Avicenna's Psychology (1952; rpt. Westport, CT: Hyperion, 1981), p. 83.Google Scholar

Recently, Portelli, J. P. has attempted to refute Rahman's claim and the related views of other scholars, in “The ‘Myth’ that Avicenna Reproduced Aristotle's ‘Concept of Imagination’ in De anima,” Scripta Mediterranean 3 (1982): 122–34, esp. p. 123Google Scholar. But in Al-Taʻlîqât ʻalâ hawâshin kitâb al-nafs li-Aristû (Marginal Glosses on Aristotle's De anima), edited by Badawi, A. R., in Aristû ʻinda al-ʻArab (Cairo: Maktabah al-Nahdah al-Misriyah, 1947), pp. 75–116Google Scholar, Avicenna himself explicitly identifies Aristotle's phantasia as a term covering all of the internal senses except the common sense: “And what [Aristotle] brings together here under the term ‘imagination’ (al-takhayyul) can be divided into a number of active powers, such as estimation and cogitation, and retentive powers, such as the formative faculty and memory” (p. 97, commenting on De anima 3.3.428b1 Iff.). On estimation in particular, cf. also Qanûn fî al-tibb (Canon of Medicine), edited by Zayʻur, Ali and Al-Qashsh, Idwar (Beirut: Muʻassasah Izz al-Din, 1987), p. 97Google Scholar: “And some people who speak loosely call this faculty [of estimation] ‘imagination’” (takhayyulan).

3 The canonical Avicennian presentations are found in Kitâb al-Najâh (Salvation), edited by Fakhry, Majid (Beirut: Dar al-‘Afaq al-Jadidah, 1985), pp. 200–2, 207–13Google Scholar; English translation by Fazlur Rahman, in Avicenna's Psychology, pp. 79–83; and Al-Shifâʼ: Al-Nafs (Healing: Psychology), edited by Rahman, F., Avicenna's De anima, Being the Psychological Part of Kitâb al-Shifâʼ’ (London: Oxford University Press, 1959), pp. 43–45; 58–61; 163–69Google Scholar. The medieval Latin translation of this work has been edited by Van Riet, S. in the Avicenna Latinus series, Liber de anima, seu sextus de naturalibus, 2 vols. (Louvain: Peeters; Leiden: Brill, 1968), Vol. 1, pp. 85–90, 114–20; Vol. 2, pp. 1–11Google Scholar. There is also another edition with French translation by Bakos, Jan, La psychologie d'lbn Sina (Avicenne), 2 vols. (Prague: Éditions de l'Académie Tchécoslovaque des Sciences, 1956).Google Scholar

4 The principal texts on the epistemic classification of premises are: Salvation, pp. 97–103; Al-Shifâʼ: Vol. 1, Al-Mantiq (Logic), part 5, Al-Burhân (Demonstration), edited by A. E. Affifi and I. Madkour (Cairo: Organisme General des Imprimeries Gouvernementales, 1956), pp. 63–67; Al-Ishârât wa-al-tanbîhât, edited by Forget, J. (Leiden: Brill, 1892), pp. 55–64Google Scholar; English translation by Inati, S. C., Remarks and Admonitions, Part One: Logic (Toronto: Pontifical Institute of Mediaeval Studies, 1984), pp. 118–28Google Scholar; French translation by Goichon, A.-M., Le livre de directives et remarques (Paris: Vrin, 1951)Google Scholar. Estimation's role in the division of the theoretical sciences is alluded to in Al-Shifâʼ: Vol. 1, part 1, Al-Madkhal (Isagoge), edited by G. Anawati, M. El-Khodeiri, F. Al-Ahwani and I. Madkour (Cairo: Al-Matbaʻah al-ʻAmiriyah, 1952), pp. 12–16; English translation with commentary by Marmura, M. E., “Avicenna on the Division of the Sciences in the Isagoge of his Shifâʼ,” Journal for the History of Arabic Science, 4 (1980): 239–50Google Scholar. The principal text decrying the estimation's influence on religious beliefs is Ithbât al-nubûwât (Proof of Prophecies), edited by Marmura, M. E. (Beirut: Dar al-Nahar, 1968)Google Scholar; English translation by Marmura, , “On the Proof of Prophecies and the Interpretation of the Prophets' Symbols and Metaphors,” in Lerner, R. and Mahdi, M., eds., Medieval Political Philosophy: A Sourcebook (Ithaca, NY: Cornell University Press, 1963), pp. 112–21Google Scholar. Estimation's role as the apperceptive faculty within the animal soul is outlined by Avicenna most explicitly in two late works, Al-Mubâhathât (Discussions), edited by A. R. Badawi, Aristû ʻinda al-ʻarab, pp. 118–239; and Al-Taʻlîqât (Notes), edited by Badawi, A. R. (Cairo: General Egyptian Book Organization, 1973)Google Scholar. Its place in the creation of fictional or “vain” ideas is treated in a short work, Al-Risâlah fî al-nafs (Letter on the Soul), edited with French translation by Michot, J. R., “‘LʻÉpitre sur la dispartition des formes intelligibles vaines après la mort’ d'Avicenne,” Bulletin de philosophie médiévale, 29 (1987): 153–70CrossRefGoogle Scholar; English translation by Michot, , “Avicenna's ‘Letter on the Disappearance of the Vain Intelligible Forms After Death’,” Bulletin de philosophie médiévale, 26–27 (1984–85): 94–103Google Scholar. The role of estimation in ethical precepts is also developed within the context of Avicenna's epistemic classification of premises, within the discussion of the premises known as “widely accepted” (al-mashhûrât).

5 For the Aristotelian background, cf. De anima 2.4.415a 14–22; cf. 2.5.417a14–20.

6 I.e., the proper and common sensibles of Aristotle, such as colour, sound, motion, magnitude, etc.

7 For the distinction between form and intention as the basis for a distinction of faculties, see, for example, Salvation, p. 200; Rahman transl., p. 32.

8 Cf. Healing: Psychology, p. 167; Van Riet, Vol. 2, p. 8: “It has been customary to call the thing apprehended by the sense a form, and the thing apprehended by the estimation an intention.”

I have opted throughout to retain the traditional translations of “estimation” for wahm and “intention” for maʻnâ, both of which are derived from the medieval Latin renditions of the terms, and both of which are somewhat controversial as translations. I believe that these terms, while not literal renditions of the Arabic, do in fact capture their technical philosophical connotations for the Islamic philosophers, although for the sake of brevity I do not intend to argue that point here. Rather, the traditional terms will be used simply for the sake of convenience.

9 Avicenna gives these examples in almost all of his initial enumerations of the internal senses. See, for example, Healing: Psychology, pp. 165–67; Van Riet, Vol. 2, pp. 6–7; Salvation, pp. 200–2; Rahman transl., pp. 30–31; Remarks and Admonitions, p. 125; Maqâlah fî al-nafs (Treatise on the Soul), edited with German translation by Landauer, S., “Die Psychologie des Ibn Sina,” Zeitschrift der deutschen morgenländischen Gesellschaft, 29 (1876): 339–72, esp. pp. 359–60Google Scholar; Sources of Wisdom, p. 38; Al-Mabdâ' wa-al-maʻâd (Genesis and Return), edited by Nurani, A. (Montreal: McGill University Press; Tehran: Tehran University Press, 1984), p. 95Google Scholar; Canon of Medicine, p. 96; Fî al-quwâ al-insânî wa-idrâkât-hâ (On Human Faculties and their Perceptions), in Tisʼ Rasâʻil (Nine Treatises) (Constantinople: Matbaʻah al-Jawaʼib, 1880), pp. 42–48, esp. p. 43.

10 See Salvation, p. 209; Rahman transl., pp. 39–40; Healing: Psychology, pp. 60–61; Van Riet, Vol. I, pp. 118–19; Sources of Wisdom, p. 42; Genesis and Return, p. 103; On Human Faculties, p. 44.

11 For the sake of convience, both Ghazâlî's critique of Avicenna in the Tahâfut alfalâsifah and Averroes's reply will be cited from Averroes's Tahâfut al-tahâfut (Incoherence of The Incoherence), edited by Bouyges, M. (Beirut: Imprimerie Catholique, 1930), pp. 543–53Google Scholar; English translation by Van Den Bergh, Simon, Averroes' Tahafut at-Tahafut, 2 vols. (Oxford: Oxford University Press, 1954), pp. 333–41Google Scholar; medieval Latin translation edited by Zedler, Beatrice, Averroes's Destructio Destructionum Philosophiae Algazelis in the Latin Version of Calo Calonymos (Milwaukee: Marquette University Press, 1961), pp. 423–32Google Scholar.

12 Commenting on his overview of Avicenna's division of the soul's powers, Ghazâlî says: “Nothing of what we havex mentioned need be denied on religious grounds, for all these things are observable facts whose habitual course has been provided by God. We only want now to refute their claim that the soul being an essence subsistent by itself can be known by demonstrative rational proofs … ” (Incoherence, p. 548; Van Den Bergh transl., p. 336).

13 All of Avicenna's presentations of the internal senses include their assignment to a particular organ or ventricle of the brain. In the standard account, the estimative sense is assigned to the posterior part of the middle ventricle of the brain, giving it ready access to the imagination and memory. See, for example, Salvation, p. 202; Rahman transl., p. 31. In some works, however, Avicenna suggests that the estimative faculty can be said to have the whole brain as its organ generally, and this ventricle as its special seat. See, for example, Treatise on the Soul, p. 360; Remarks and Admonitions, p. 125.

14 Incoherence, p. 549; Van Den Bergh transl., pp. 337–38. On this point, cf. Druart, T.-A., “Imagination and the Soul-Body Problem in Arabic Philosophy,” Analecta Husserliana, 16 (1983): 327–42, esp. p. 337Google Scholar.

15 Ghazâlî, of course, is well aware of Avicenna's use of estimation in logical contexts, and Ghazâlî himself freely employs estimative premises in his own logical works, the Miʻyâr al-ʻilm (Standard for Knowledge) (Beirut: Dar al-Andalus, 1964), pp. 151–52; and the Maqâsid al-falâsifah (Intentions of the Philosophers), edited by Dunya, S. (Cairo: Dar al-Maʻarif bi-Misr, 1961), pp. 104–5Google Scholar; medieval Latin version of the latter work edited by Lohr, C. H., “Logica Algazelis: Introduction and Critical Text,” Traditio, 21 (1974): 223–90Google Scholar, esp. p. 276 (the term ‘estimatives’ is rendered into Latin as opinabiles).

16 It might be objected that most of the properties which Avicenna cites as examples of intentions are relational properties, and hence need not inhere in the extramental thing: wolves are not inherently hostile, but only so in relation to sheep. But Avicenna's view of relational properties requires that they have some external ground in a particular accident of a particular substance, even if some aspect of the notion of a relation is a purely conceptual construct. On Avicenna's views on relations, see Marmura, M. E., “Avicenna's Chapter ‘On the Relative,’ in the Metaphysics of the Shifâʼ,” in Hourani, G. F., ed., Essays on Islamic Philosophy and Science (Albany: State University of New York Press, 1977), pp. 83–99.Google Scholar

17 In the Salvation, for example, Avicenna initially appeals only to the fact that the intentions are not perceived by both internal and external sense faculties (p. 201; Rahman transl., p. 30); only when he comes to establish a hierarchy of abstraction among the soul's apprehending powers does he explicitly characterize estimative intentions as non-material (p. 209; Rahman transl., pp. 39–40). For other discussions of estimation's place in the hierarchy of abstractive powers, see Healing: Psychology, pp. 60–61; Van Riet, Vol. 1, pp. 118–19; Sources of Wisdom, p. 42; On Human Faculties, p. 44; Genesis and Return, p. 103.

18 Cf. Ushida, N., Étude comparative de la psychologie d'Aristote, d'Avicenne, et de St-Thomas (Tokyo: Keio Institute of Cultural and Linguistic Study, 1968), p. 159Google Scholar: “Ainsi les significations des objets ne sont elle-mêmes ni purement sensibles, ni purement intelligibles.… Par conséquent, Avicenne les détermine commes entités immatérielles en elles-mêmes, mais engagées par accident dans le matière.”

19 Salvation, p. 209; Healing: Psychology, p. 60; Van Riet, Vol. 1, p. 119. Avicenna's proposition is thus of the form, “Not necessarily p ” (∼Lp), whereas Ghazâlî appears to be assuming a proposition of the form, “Necessarily not-p” (L∼p). The only exception to this careful way of describing intentions is in the early Genesis and Return, p. 103, where intentions are referred to as intelligibles: “As for the estimation, it abstracts form from matter the most, because it understands non-sensible, intelligible intentions, although it does not understand them as intelligible universals, but rather, as bound to sensible intentions.”

20 However, in Ghazâlî's initial presentation of Avicenna's account of the soul's powers, he does represent Avicenna's claims accurately. See Incoherence, p. 544; Van Den Bergh transl., p. 334, where he says that the meaning of an intention is “that which does not require a body for its existence, although it can happen that it occurs in a body.”

21 The Avicennian quiddity or common nature is discussed in Healing: lsagoge, p. 15; the passage is translated and discussed in Marmura, “Avicenna on the Division of the Sciences,” pp. 247–50; see also Shifâʼ: Ilâhîyât (Healing: Metaphysics), edited by Madkour, I., Musa, M., Dunya, S. and Zayid, S., 2 vols. (Cairo: General Egyptian Book Organization, 1960), Vol. 1, pp. 31–32Google Scholar; 5.1–2; Vol. 1, pp. 195–212; medieval Latin version edited by Van Riet, S., Liber de philosophia prima, 3 vols. (Leiden: Brill; Louvain: Peeters, 1977–83), Vol. 1, pp. 34–36Google Scholar; Vol. 2, pp. 227–45; French translation by Anawati, G. C., La métaphysique du Shifâʼ, 2 vols. (Paris: Vrin, 1978, 1985)Google Scholar. For a discussion of this doctrine and its influence on the Latin West, see Owens, Joseph, “Common Nature: A Point of Comparison Between Thomistic and Scotistic Metaphysics,” Mediaeval Studies, 19 (1957): 1–14.CrossRefGoogle Scholar

22 I am not suggesting that an estimative intention is itself a common nature; my point is simply that the notion of a property undetermined to any specific mode of existence is not peculiar to Avicenna's doctrine of estimation, and is in fact central to his entire metaphysical outlook.

23 De anima 3.7.431a8–14. On the difference between Avicenna and Averroes with regard to this point, cf. Rahman, Avicenna's Psychology, p. 82.

In the Summa theologiae (la.87.4), Thomas Aquinas supplies an implicit defence of Avicenna on this point, arguing that there is a difference between those pleasures and pains attendant upon perception that involve a sort of aesthetic harmony or discord between the object and the percipient—e.g., when a colour is pleasing to the eye, or a sound cacophonous to the ear—and the perception of pleasure and pain under the guise of harmfulness or utility to the percipient's well-being, which Avicenna's examples of animal estimation imply.

24 See Averroes, Incoherence, p. 547; Van Den Bergh transl., p. 336.

This rejection of estimation is already implicit in a work as early as Averroes's Epitome of the Parva naturalia. In this text, Averroes mentions estimation only once, saying that it is an ability in animals which has no name, but which Avicenna has called “estimation.” See Talkhîs kitâb al-hiss wa-al-mahsûs, edited by Blumberg, H. (Cambridge, MA: Mediaeval Academy, 1972), p. 39Google Scholar. In the rest of the text, Averroes refers only to the discriminative faculty (al-mumayyizah), which is a human faculty equivalent to cogitation, to which Averroes assigns the role of analyzing and synthesizing images. Blumberg's translation of this text is misleading in this regard, since he renders al-qûwah al-mumayyizah as “estimation” or “estimative faculty” throughout. See Epitome of the Parva naturalia, translated by Blumberg, H. (Cambridge, MA: Mediaeval Academy, 1961)Google Scholar.

25 Averroes's acceptance of cogitation in humans and his rejection of animal estimation suggest that this is part of his own position (cf. the preceding note). Any real grasp of something like an intention is attributable to humans by virtue of their access to the intellect; in animals, the reactions to which Avicenna alludes are purely instinctual, and presumably involve no authentic perception of the intention itself.

26 My contention here is that it is principally the phenomenon of certain peculiarly human types of cognitive judgments that best illustrates why imagination alone cannot, on Avicenna's view, fulfil the functions that he assigns to the estimative faculty, and why those same activities cannot be assigned to the intellect. I do not intend to deny the presence of estimation in animals, a point on which Avicenna is insistent. Rather, I am arguing that human cognition best exemplifies Avicenna's underlying rationale, even though the activities in question can be shown to have their counterparts in nonhuman, non-rational animals as well.

27 For incidental perception in Aristotle, see De anima 2.6.418a20–25; 3.1.425a 14–b4. For estimation and incidental perception in Avicenna, see especially Healing: Psychology, p. 166; Van Riet, Vol. 2, pp. 6–7: “Then we may make judgments concerning the sensibles by means of intentions which we do not sense, and which are either not sensibles in their natures at all, or which are sensibles, but we do not sense them at the time of the judgment. … As for those which are sensibles, we see, for example, something yellow, and we judge that it is honey and sweet. For the sense does not convey this to us at this moment, although it is of the genus of the sensible, even if the judgment itself does not take place through anything sensible at all. …”

28 Cf. De anima 2.6.418a20–24; 3.2.425a24–27.

29 For this formula, see, for example, Salvation, p. 201: “And the distinction between the perception of form and the perception of intention is that the form is something which both the internal soul and the external sense perceive together, but the external sense perceives it first and conveys it (wa-yuʼaddî-hi) to the soul;… whereas the intention is something which the soul perceives from the sensibles (min al-mahsûs), without the external sense perceiving it first”; and Remarks and Admonitions, p. 124: “For rational and irrational animals perceive in the particular sensibles particular non-sensible intentions which are not conveyed by way of the senses” (wa-lâ mutaʼaddiyah min tarîq al-hawâss).

30 Even Aristotle, from whom the principle of faculty differentiation according to a difference in objects derives, locates the actualization of perception in the patient or percipient, and hence argues that the object of perception is in some way defined in terms of its effect on a percipient. See, for example, De anima 2.7.419a 17–18; and especially 3.2.426a15–26. This principle, too, would allow the same object to count as a form at one time and an intention at another, depending on which perceptual faculty it was affecting, in much the same way as the same object is visible if it affects sight, but tactile if it affects touch.

31 Healing: Psychology, pp. 184–85; Van Riet, Vol. 2, p. 39. Rahman emphasizes this facet of estimation in his commentary on the Salvation. See Rahman, Avicenna's Psychology, p. 81.

32 In Avicenna's scheme of internal senses, the formative faculty acts as a kind of sensememory where the sensible forms or images are stored, whereas memory proper is the storehouse of estimative intentions.

33 Healing: Psychology, pp. 165; Van Riet, Vol. 2, pp. 49–50.

34 I am using ‘icon’ here after Aristotle's use of the Greek eikôn at De memoria 1.450a22–451a14, to indicate a depiction that is taken in its special function of portraying a particular individual, e. g., Coriscus, rather than simply as a picture of a man who happens to have this particular physical form.

35 Two passages in Avicenna's Notes reinforce my claim that estimation is the agent of incidental perception because of its ability to grasp concrete individuals in their individuality. In the first passage, Avicenna addresses the question of how we perceive that the sensible forms which we perceive have an external existence (wujûdan min khârijin), that is, how we distinguish real percepts from those in dreams, since the senses themselves have only discrete sensations—for example, touch, only feeling heat or weight, not a hot or heavy body. Avicenna answers that this is done by either the estimation or the intellect (pp. 68–69). In a later passage, he likens the awareness that this collection of sensible qualities belongs to this individual, Zayd, to the incidental perception that allows us to be aware of the flavour of honey when we perceive its colour (p. 147).

36 Healing: Psychology, p. 182; Van Riet, Vol. 2, pp. 34–35; Canon of Medicine, p. 97; Treatise on the Soul, pp. 359, 363; in the latter passage, Avicenna provides an elaborate analogy between the division of labour amongst the servants of a prince and the functions of the soul's faculties. Here estimation is likened to the vizier in the animal soul. In Remarks and Admonitions (p. 124), Avicenna likens the estimative sense's judgment of intentions to the external sense's judgment of what they observe. On estimation as a judge, cf. Michot, La destinée de l'homme, pp. 148–49.

37 See Canon of Medicine, p. 96, where Avicenna calls estimation “the power which is in reality the internal percipient in the animal” (qûwah hiya bi-al-haqîqah al-mudrikah al-bâtinah fî al-hayawân).

38 See the Treatise on the Soul, p. 359, where Avicenna calls the estimative sense “opinionative” (al-zânnah).

39 Treatise on the Soul, pp. 359, 360. Here Avicenna also refers to imagination being controlled by whatever estimation “sanctions and assents to” (istaswaba-hu… wa saddaqa-hu).

40 See Healing: Psychology, pp. 164–65; Van Riet, Vol. 2, p. 4, for the interplay of estimation and common sense in dreams; the language of authority, control and obedience recurs in relation to estimation in Avicenna's consideration of sleep and dreams at p. 171; Van Riet, Vol. 2, pp. 14–16; and in his discussion of estimation and memory at pp. 184–85; Van Riet, Vol. 2, pp. 38–40. On this point one should also recall the text from the Notes cited in n. 35 above, in which estimation is said to determine whether or not perceptual objects have an external existence.

41 See especially Sources of Wisdom, p. 39: “And the peculiar characteristic of the imaginative faculty is its continuity of motion, so long as it is not controlled. And its motion is to imitate things through their likenesses and contraries.” The random and unconscious nature of compositive imagination is also emphasized by Nasr al-Dîn al-Tûsî (d. 1274), a commentator on Avicenna's Remarks and Admonitions. Replying to the charge made by an earlier commentator, Fakhr al-Dîn al-Râzî (d. 1209), that compositive imagination is superfluous because estimation alone is able to account for the ability to compose and divide images and intentions, Tûsî argues that compositive imagination is a faculty “which has free disposal over two things whose presence it does not perceive.” Tûsî's commentary, along with that of Qutb al-Dîn al-Râzî (d. 1374), can be found in the anonymous edition of the Ishârât wa-tanbîhât, 3 vols., 2nd ed. ([Tehran]: Daftar Nashr Kitab, 1983), Vol. 2, p. 45. Fakhr al-Dîn al-Râzî's criticisms are repeated in several works: Sharh ʻuyûn al-hikmah (Commentary on Sources of Wisdom), edited by Al-Saqa, A. H. A., 3 vols. (Cairo: Maktabah al-Anjlu al-Misriyah, [1986]), Vol. 2, p. 252Google Scholar; Lubâb al-lshârât (Core of the Remarks and Admonitions), edited by Shahaby, M., in Al-tanbihat val-Esharat (Tehran: Tehran University Press, [1960]), p. 237Google Scholar; and Sharh al-lshârât (Commentary on Remarks and Admonitions), edited by Khalaf, A. J. (Cairo: N.p., 1907), pp. 151–52Google Scholar. (Râzî's commentary is printed in the margins of this edition.)

42 This does not mean that humans do not possess this power, but that they possess it in virtue of having an animal soul, not in virtue of being rational.

43 In Remarks and Admonitions, Avicenna even suggests that the intellect itself can only use the compositive imagination with the help of estimation: “And it is as if [the compositive imagination] were a faculty belonging to estimation, and, through the mediation of estimation, to the intellect” (p. 125).

44 In this regard, it is interesting to note that Averroes, in his early Talkhîs kitâb al-nafs (Epitome of the De anima), edited by Al-Ahwani, A. F. (Cairo: Maktabah al-Nahdah al-Misriyah, 1950)Google Scholar, seems to run into problems precisely in his attempt to assign the controlling and combining powers to a single imaginative faculty. Thus we find Averroes claiming in his chapter on the imagination in this work both that the imagination is necessary and involuntary (p. 64) and that it is under our control (pp. 60, 64).

45 For the hint that the estimative power has a role to play in the creation of fictional images, see the Letter on the Soul, p. 159: “The second premise is that it is not possible for these forms, which are contrary and opposed to the real, to arise in the soul unless through the mediation of imagination, sensation, and estimation. For estimation indeed also has some influence upon this” (my translation; cf. Michot transl., p. 101). Avicenna argues that it is false to hold that such fictions as the phoenix are only images, on the grounds that any idea which can be entertained in such as way as to involve commonality is universal and intelligible. But one is able to imagine many individual phoenixes sharing many traits, so there must be a universal idea of phoenix as well as images of particular phoenixes.

Similar activities of estimation are described in chap. 13 of Ahwâl al-nafs (States of the Soul), edited by Al-Ahwani, A. F. (Cairo: Issa El Baby el-Halaby, 1952)Google Scholar, where Avicenna speaks of how estimation, with the help of imagination, governs one's activities when one is overcome by emotions like fear, causing one to perceive “eerie things” (ʻumûran mûhishatan; pp. 119, 121); French translation by Michot, J. R., “Prophétie et divination selon Avicenne,” Revue philosophique de Louvain, 83 (1985): 507–35CrossRefGoogle Scholar. Also related to estimation's link with fictive creations is Avicenna's suggestion of the possibility of estimative felicity and misery in the afterlife for those who pursue material pleasures in this life. Such people may, Avicenna suggests, use other bodies, perhaps the celestial ones, to continue an imaginary life amongst the forms and intentions in which their estimative senses once delighted. On this point, see the Genesis and Return, chap. 15, pp. 114–15. It is important to note that Avicenna explicitly reports these views as those of serious philosophers, although he does not endorse them openly in this text. Avicenna's notion of an imaginal or estimative afterlife is the principal subject of Michot's La destinée de l'homme. See especially chap. 1, pp. 22–56, for the context of Avicenna's consideration of this topic.

46 For further discussion of the role of estimation in poetic creativity, see my Logic and Aristotle's Rhetoric and Poetics in Medieval Arabic Philosophy (Leiden: Brill, 1990), pp. 204–7.

47 Healing: Psychology, p. 166; Van Riet, Vol. 2, p. 8.

48 Ibid., p. 166; Van Riet, Vol. 2, p. 8.

49 Ibid., pp. 182–83; Van Riet, Vol. 2, p. 35. The metaphor of honey and bile is a favourite one, especially in Avicenna's discussions of poetic premises. It may be an allusion to Aristotle's remarks on incidental perception at De anima 3.1.425a30-b4. For discussion of the honey-bile metaphor, see my “The ‘Imaginative Syllogism’ in Medieval Arabic Philosophy,” Mediaeval Studies, 51 (1989): 243–67; and my Logic and Aristotle's Rhetoric and Poetics, pp. 231–38, esp. p. 237, n. 75.

50 See n. 4 above for the principal texts.

51 See Remarks and Admonitions, p. 59; Inati transl., p. 123; Healing: Demonstration, p. 64; Salvation, p. 98.

52 In the Sources of Wisdom, p. 12, Avicenna explicitly makes the link between some estimative premises and sophistics; the same is true of the Persian work, Dânesh-nâmeh ‘Aiâ’î (Book of Science), French translation by Achena, M. and Massé, H., Le livre de science, 2 vols. (Paris: Belles Lettres, 1955–58), Vol. 1, p. 75Google Scholar. The Remarks and Admonitions likens the estimative premises to “resembling” or ambiguous propositions, which are assigned to sophistics, although Avicenna admits that estimatives would be widely accepted (mashhûrât), and thus dialectical or rhetorical, were it not for philosophy and religious law (p. 60; Inati transl., pp. 123–24). In the Demonstration of the Healing, Avicenna makes no positive assignment of estimative premises to any logical art, but he does explicitly exclude estimative necessity as a ground for demonstrative certitude (p. 65).

53 Healing: Demonstration, p. 64; cf. Salvation, p. 98.

54 Remarks and Admonitions, p. 59; Sources of Wisdom, p. 12. Inati's claim (p. 124, n. 24 of Remarks and Admonitions) that “pure” here refers to the purity of the objects of estimation (i.e., the immateriality of intentions) is probably incorrect. Avicenna's comments in his various discussions seem to indicate that the purity in question is the estimative sense's independence-that is, its failure to be guided by the intellect.

55 Sources of Wisdom, p. 12.

56 Healing: Demonstration, p. 64. Cf. Notes, pp. 22–23, where Avicenna suggests that estimation's co-operation with the intellect produces epistemic confidence, whereas its obstruction of the intellect produces perplexity and impedes the attainment of certitude.

57 For these examples, see Salvation, p. 98; Healing: Demonstration, p. 64; Remarks and Admonitions, p. 60; Inati transl., p. 124.

58 See Proof of Prophecies, sec. 38–39, pp. 56–59; Marmura transl.. pp. 119–20. Here Avicenna argues that since the estimative sense is prone to deny the reality of immaterial beings, particularly in people whose intellects are weak, it is the source for materialist views of God and the afterlife: “Indeed, all who have perished have suffered thus because they have conformed with the estimative faculty, which is the animal faculty that gives false judgments regarding abstracted form (al-sûrah al-mujarradah) when the senses are dormant” (p. 56; Marmura transl., p. 119, slightly modified).

59 Cf. Healing: Demonstration, p. 67; Remarks and Admonitions, p. 60; Inati transl., pp. 123–24; Salvation, p. 99. The first two texts emphasize that estimative propositions are even stronger in their influence than widely accepted propositions (i.e., those with widespread social sanction).

60 Remarks and Admonitions, p. 59; Inati transl., p. 123.

61 Of course, animals can fall prey to errors in incidental perception, as when a bird mistakes the scarecrow for a real person (my example), but Avicenna makes no claim that incidental perception in general partakes of estimative necessity. On the issue of estimation and sense error, cf. the discussion of Ushida, Étude comparative, pp. 163–68.

62 On the soul's unity, cf. the discussion in the Salvation, chap. 15, pp. 228–30; Rahman transl., pp. 65–68.

63 Remarks and Admonitions, pp. 59–60 (my translations); Sources of Wisdom, p. 12; Salvation, p. 99. On Avicenna's dualism, see Druart, “Imagination and the Soul-Body Problem,” pp. 330–35; and “The Soul-Body Problem: Avicenna and Descartes,” in Druart, , ed., Arabic Philosophy and the West: Continuity and Interaction (Washington, DC: Center for Contemporary Arab Studies, Georgetown University, 1988), pp. 27–49; esp. pp. 41–42Google Scholar.

64 Salvation, p. 98; cf. Remarks and Admonitions, p. 59: “Because estimation is consequent upon sensation, and what does not agree with the sensible the estimation does not accept” (my translation).

65 Remarks and Admonitions, pp. 59–60 (my translation); cf. Healing: Demonstration, p. 64: “[The intellect] draws the conclusion that the sensibles have principles which are different from the sensibles [themselves].” On the estimation's inability to represent itself, cf. Book of Science, Achena-Masse transl., Vol. 1, p. 70.

Along the same lines, Avicenna emphasizes in his Notes that the estimative faculty is tied to the temporal realm, so that time cannot be removed from its conceptions. Moreover, because time itself is not in time, estimation has no cognizance of the nature of time itself, just as it has no ability to imagine its own nature or the nature of its other principles: “Estimation establishes a ‘when’ for everything; and it is impossible for time itself to have a ‘when’ “(p. 142). See Notes, pp. 138–39 and 141–42 for further consideration of estimation and temporality.

66 Cf. Saivation, p. 218; Rahman transl., p. 52, on the lack of self-awareness in the sensible soul. For the Greek background, see Rahman's commentary, Avicenna's Psychology, pp. 103–4; and Pines, S., “La conception de la conscience de soi chez Avicenne et chez Abu'l-Barakat al-Baghdadi,” Archives d'histoire doctrinale et littéraire du moyen age, 29 (1954): 21–98Google Scholar, esp. pp. 36–43. Cf. also Remarks and Admonitions, p. 179. Avicenna's views on estimation's role in self-awareness and apperception in general are discussed in section 8 below.

67 Salvation, p. 99; cf. Remarks and Admonitions: “But if [the intellect and estimation] reach a conclusion together, the estimative power will recoil and impede the acceptance of that whose necessity was granted” (p. 60; my translation).

68 Healing: Demonstration, p. 65; Salvation, p. 99. This list is divided between the essential non-sensibles, such as the Creator and the Intellect, and things “more general” than the sensible—i.e., the principles of sensible things which are themselves prior to what is sensible.

69 One should not forget in this regard that the function of the sensible soul in the formation of all rational judgments is for Avicenna totally preparatory and ancillary. See, e.g., Salvation, chap. 11, pp. 220–22; Rahman transl., pp. 54–56.

70 There are some exceptions in the list of the Salvation, which includes the [Agent] Intellect and God as Creator. But these are the only principles which Avicenna openly identifies as essentially non-sensible, as indicated in n. 68 above.

71 Ultimately, then, Avicenna needs estimation to account for these judgments, because attributing them to the intellect would represent a violation of the principle of noncontradiction: the intellect would have to affirm simultaneously both “S is P” and “S is ∼P.” There are two specific features of estimative propositions that implicitly lead Avicenna to hold that attributing them to the intellect would produce this result: (1) the fact that these propositions have the psychological force of primary propositions or first principles; and (2) the fact that most people continue to feel the pull of these materialist views, despite their intellectual convictions to the contrary. Together, these two phenomena make it impossible for Avicenna to claim that false beliefs of this sort are secondary intelligibles, arrived at as conclusions of faulty processes of reasoning. These propositions have the feel of intuitive first principles; because they are false, they cannot truly be first principles of the intellect; the only remaining alternative is to make them principles, but on some pre-intellectual level.

72 Concerns similar to these may be behind the criticisms of Fakhr al-Dîn al-Râzî, which revolve around the alleged “equality in force” between the true necessity of first principles and the psychological compellingness of estimative premises. See Commentary on Sources of Wisdom, Vol. 1, pp. 198–99; 201–2; Core of Remarks and Admonitions. pp. 197–98.

73 Healing: Demonstration, p. 64.

74 Salvation, p. 99. In the Notes, Avicenna also applies the term fitrah to estimation: “A human being is constituted by nature (futira) to acquire knowledge and to apprehend things naturally (tabʻan) by way of the senses, and then by way of the estimation, which is a copy (nuskhat-hâ) of them” (p. 22).

75 Salvation, p. 99.

76 On Avicenna's dualism, cf. n. 63 above.

77 It is not entirely clear to me whether Avicenna's use of the Qurʻanic term fitrah, rather than the Aristotelian term tabʻ, is meant to overcome the obvious appearance that Avicenna has indeed allowed nature in the Aristotelian sense to labour in vain. In the passage from the Notes cited in n. 74 above, tabʻ seems to be confined to the sensible soul, since Avicenna goes on to refer to the process of acquiring intelligibles as a “non-natural acquisition (iktisâb Iâ tabʻan) (p. 22). Still, the use of fitrah to refer to estimation as well as intellection is not, on the whole, especially comforting, given its place of honour in other contexts, in particular Avicenna's doctrine of intuition (hads). According to Gutas, Dimitri, Avicenna and the Aristotelian Tradition (Leiden: Brill, 1988), p. 170Google Scholar, Avicenna uses fitrah as “a concept of natural intelligence providing innate, a priori knowledge.” It is therefore the theological equivalent to hads, an intellectual intuition which is prophetic in its highest form. That fitrah could be used both for such an infallible noetic activity and for the erroneous judgments of estimation certainly creates ambiguity, and raises serious epistemological problems as to how innateness and connaturality can be used as guarantors of knowledge and truth. For if estimative fitrah and intuitive fitrah differ so radically in their reliability, then any appeal to something's being in accordance with fitrah must of necessity be epistemologically irrelevant: it can tell us nothing in itself about whether or not we can trust our cognitive instincts.

78 The practical analogue of this is, of course, the Socratic problem of weakness of will: how can the intellect be overcome by the passions and fail to follow its own dictates? See Aristotle's formulation of the problem in Nicomachean Ethics 7.3.1145b21–31.

79 Salvation, pp. 98–99.

80 Cf. n. 69 above.

81 Salvation, p. 99.

82 See, e.g., Remarks and Admonitions, pp. 56–58; Inati transl., pp. 120–21.

83 Healing: Isagoge, pp. 12–13; translated by Marmura in “Avicenna on the Division of the Sciences,” p. 243. For a discussion of the passage, and the role of estimation in it, see Marmura, pp. 243–44.

84 Healing: Isagoge, p. 13; Marmura transl., p. 245.

85 Healing: Metaphysics, Vol. 1, pp. 23–24; Van Riet, Vol. 1, pp. 26–27. There are several other places in the Metaphysics where Avicenna alludes to the nature of estimative apprehension in order to establish similar points. Following is a summary of the more important references: (1) Matter cannot be divided in the estimation (Vol. 1, p. 75; Van Riet, Vol. 2, p. 87). (2) Surface can be separated from body in the estimation, to the extent that the estimation does not attend to its relation to a body even though it recognizes that there cannot be a surface without a body (Vol. 1, pp. 111–15; Van Riet, Vol. 1, pp. 122–27). (3) Estimation is called upon to refute the denial of the existence of the circle among the atomists: “And all of this is something about whose destruction there is no doubt intuitively (ʻaid al-badîhah), nor does estimation, which is the rule (al-qânûn) concerning sensible things and what is attached to them, as you know, conceptualize it (yatasawwaru-hu)” (Vol. 1, p. 148; Van Riet, Vol. 1, pp. 167–68). (4) Mathematical objects (al-ʻumûr al-ta ʻlimîyah) can be separated from matter in the estimation, although such objects are necessarily considered in the estimation according to the division and configuration (al-tashakkul) they have in matter (Vol. 2, pp. 11–13; Van Riet, Vol. 2, p. 347).

86 Seep. 231 above.

87 On the “Flying Man” see Druart, “The Soul-Body Problem,” pp. 31–38; Marmura, M. E., “Avicenna's ‘Flying Man’ in Context,” Monist, 69 (1986): 383–95CrossRefGoogle Scholar; Pines, “La conception de la conscience de soi,” pp. 21–56.

88 Cf. Ushida, Étude comparative, pp. 170–72, on the estimative sense as the apperceptive faculty within the sensible soul, with reference to Healing: Psychology, p. 67; Van Riet, Vol. 1, p. 130: “As to perceiving that it perceives, this does not belong to the sense, for the perception [in question] is not a colour which is seen, nor a sound which is heard. Rather, this is only perceived through intellectual activity or the estimation. …”

89 Discussions, sec. 305, p. 184.

90 Notes, p. 161: “This is not peculiar to human beings, but rather, all animals are aware of themselves in some way.”

91 Ibid., p. 82.

92 Discussions, sec. 421, pp. 220–21, attributes awareness that I am seeing or hearing to the estimative faculty or the rational soul; cf. Notes, p. 147: “The perception that one is perceiving is either through the intellect or through the estimation. For other animals perceive that they are perceiving through estimation.”

93 Thus here, as in the discussions of incidental perception, aspects of animal awareness that are usually associated with Aristotle's common sense come to be assigned to estimation, although Avicenna still views the common sense as a necessary, though insufficient, condition for both sensible apperception and incidental perception. The reason for this shift is obvious: once intentions have been identified as distinct perceptual objects from sensible forms, sensible awareness cannot be completed by any faculty that is limited to the perception of forms and unable to grasp intentions, like the Avicennian common sense.

For a discussion of the relations between common sense, estimation, and incidental perception in Avicenna and Averroes, see Blaustein, Michael, Averroes on the Imagi nation and the Intellect (Ph.D. dissertation, Harvard University, 1984), pp. 80–81, esp. p. 81, n. 15Google Scholar. For a consideration of the common sense as the faculty of consciousness for Aristotle, see Modrak, D. K., “An Aristotelian Theory of Consciousness?” Ancient Philosophy, 1 (1981): 160–70CrossRefGoogle Scholar; and Aristotle: The Power of Perception (Chicago and London: University of Chicago Press, 1987), pp. 133–54Google Scholar.

94 Notes, pp. 147–48; cf. n. 35 above.

95 Discussions, sec. 365, p. 204. The topic of self-awareness is considered throughout sec. 365.

96 On the indirectness of estimative awareness, cf. Discussions, sec. 423, p. 221, which raises the question of whether the estimative faculty could be aware of the intellective faculty, just as the intellect is aware of estimation, since in both cases the awareness would be through something else, not through the primary subject of the awareness itself.

97 Discussions, sec. 371, p. 208: “He was asked, ‘By what faculty do we perceive our particular essences (dhawât-nâ al-juz'îyah)? For the soul perceives intentions (al-nafs idrâk-hâ li-al-maʻânî) either through the intellective faculty-but the awareness of the particular essence is not an object of the intellect; or through the estimative faculty—but the estimative faculty perceives intentions conjoined to images, whereas it has been shown that I can perceive my essence even if I do not perceive it through my limbs nor imagine my body.’ “

98 Discussions, sec. 371, p. 208. This passage and those that follow it are translated and discussed in Pines, “La conception de la conscience de soi,” pp. 46–53. Avicenna's views on the relationship between the soul and body in relation to individuation are developed in Healing: Psychology, pp. 227–34; Van Riet, Vol. 2, pp. 113–26; Salvation, pp. 223–27; Rahman transl., pp. 50–63 (chap. 13). Briefly, while Avicenna maintains that the generation of a body provides the occasion for the initial individuation of the soul, it is not a cause of that individuation, except accidentally and by way of preparation. Rather, the separated intelligences create an immaterial, individual soul whenever a body which is suited to be used as that soul's instrument comes about.

99 See Notes, pp. 30; 79–80; 147–48. The latter two passages emphasize that the acquisition of the awareness that one is aware of oneself does not imply that one is “aware of oneself twice,” because one act of awareness is innate, the other acquired. Avicenna also suggests that this secondary awareness must always be acquired, to the extent that the primary awareness involves a pure identity between subject and object, whereas the secondary awareness seems to imply some element of distance or otherness. On this point, cf. Marmura, “Avicenna's ‘Flying Man’ in Context,” pp. 386–87.

100 For the importance of the notion of tanbîh in the Flying Man, see Marmura, “Avicenna's ‘Flying Man’ in Context,” pp. 386–88.

101 See Healing: Psychology, p. 16; Van Riet, Vol. 1, pp. 36–37; Remarks and Admonitions, p. 119. In the shorter version of the argument found in Book 5, chap. 7 of the Psychology of the Healing (p. 255; Van Riet, Vol. 2, p. 162), Avicenna does not use verbs alluding to estimation. A translation of all three versions of the argument is provided in Marmura's “Avicenna's ‘Flying Man’ in Context.”

102 Various forms of the verb associated with compositive imagination—takhayyalnâ, natakhayyalu— are also employed in the two versions of the argument in the Healing.

103 This ability also seems to be at work in the role of estimation in the processes of physical and mathematical abstraction discussed in section 7. For like a thought-experiment, such acts of abstraction draw upon the estimative sense's ability to direct the attention of the sensitive soul to particular features of an object to the exclusion of its other features.

104 Cf. the passage cited at n. 75 above.

105 Healing: Psychology, p. 167; Van Riet, Vol. 2, p. 8.

106 Ibid., p. 183; Van Riet, Vol. 2, pp. 35–36.

107 Ibid., p. 187; Van Riet, Vol. 2, pp. 43–44. On the centrality of estimation to animal appetition, see Rahman, Avicenna's Psychology, p. 82.

108 This parallel is also noted by Pines, “La conception de la conscience de soi,” p. 27.

109 See Remarks and Admonitions, pp. 58–59; Inati transl., p. 122; Salvation, pp. 99–100 (here the term ‘widespread’ [al-dhâʻiʻâtb;is used, but it is then defined as “widelyaccepted, esteemed beliefs” [ârâ’ mashûrah mahmûdah]); Healing: Demonstration, pp. 65–66; Sources of Wisdom, p. 12; Book of Science, Achena-Massé transl., Vol. 1, pp. 70–71. All of these discussions contain some version of the thought-experiment, and most also allude to estimation's role in it.

The link with endoxa is evident in the standard division of widely accepted premises according to whether all people, the majority or the wise accept them. Cf. Topics, l.2.100b21–23.

110 Remarks and Admonitions, pp. 58–59; Inati transl., p. 122.

111 This point is evidenced, for example, in Avicenna's reference in the quoted passage to the possibility of inductive establishment of such premises. In the Healing: Demonstration, Avicenna speaks of the need for some argumentation (hujjah) in order to make true propositions of this sort certain (p. 66); in the Sources of Wisdom, he alludes to the possibility of establishing some widely accepted propositions by demonstration (burhân), distinguishing these from widely accepted premises established through custom alone (p. 12).

In this regard, it is also important to note that Avicenna uses the term ‘widely accepted’ in broad and narrow senses. In the broad sense, any proposition accepted on the basis of its popular appeal is widely accepted, and on this construal, all primary premises such as ‘The whole is greater than the part’ are widely accepted as well, as are many sensible, estimative and empirical premises, since unreflecting people often believe them because of popular consensus. On the narrow meaning, the term ‘widely accepted’ is reserved for the “esteemed” or “praiseworthy” (al-mahmûdât) premises (i.e., for ethical dicta), and excludes the primary propositions and others which are necessary of acceptance. It is this narrower meaning that concerns us here: in the Remarks and Admonitions, Avicenna says that on this construal, “the widely-known propositions are concerned with either obligations, reformative education, and those things on which divine laws agree, character and sentiments or inductive conclusions” (p. 59; Inati transl., p. 123).

For a discussion of ethical premises in Avicenna and Ghazâlî, see Marmura, M. E., “Ghazâlî on Ethical Premises,” Philosophical Forum, 1 (1969): 393–403Google Scholar; this article contains a translation of Ghazâlî's discussion of widely accepted and related premises in his Standard for Knowledge.

112 Salvation, p. 100. In this context, Avicenna notes that on the broad construal of “widespread,” not only all primary propositions, but also all estimative ones, are also widespread. Cf. Remarks and Admonitions, p. 60; Inati transl., pp. 123–24. For the contrast with fitrah, cf. also Healing: Demonstration, p. 66.

113 Remarks and Admonitions, pp. 58–59; Inati transl., p. 122.

114 Kitâb al-radd ‘aiâ al-mantiqîyîn, edited by al-Kutubi, A. S. S. (Bombay: Qayyimah Press, 1949), pp. 428–29Google Scholar. My attention was first drawn to the passage by Marmura, “Ghazâlî on Ethical Premises,” pp. 397–98, n. 9.

115 Remarks and Admonitions, Eighth Namat, pp. 190–91; quoted by Ibn Taymîyah, Refutation of the Logicians, pp. 431–32.

116 Refutation of the Logicians, p. 432.

117 The terms qabîh (ugly), and jamîl or husn (beautiful or noble) are used for both aesthetic and ethical judgments in Arabic.

118 Refutation of the Logicians, p. 428.

119 Healing: Demonstration, p. 66; Salvation, p. 100.

120 This difference is akin in some ways to the difference between natural law and positive law in Thomistic ethics.