CROSS-REFERENCE

Résumé. Cet article démontre que la théorie de l’expérience ( taǧriba ) d’Ibn Sīnā né-cessite une référence croisée entre la logique et la psychologie. Suivant la tradition lin-guistique basrane, il paraphrase les noms dérivés ( ism muštaqq ) dans la formule li-x y (« y appartient à x ») : par exemple, ʿālim («savoir») est paraphrasé en lahu ʿilm («un acte de savoir lui appartient»). Sa théorie de l’expérience emploie cette formule pour arranger les phénomènes observés dans une certaine forme de syllogisme et décrire les fonctions des sens internes du cerveau. Ibn Sīnā arrange les phénomènes observés dans la formule li-x y ou la proposition dont un prédicat est un nom dérivé de y . En même temps, il soutient que le sens du souvenir est impliqué dans le processus de l’expérience. Le souvenir est un sens interne du cerveau, qui préserve maʿnā perçue par le sens de l’estimation. L’estimation consiste à percevoir maʿnā ( y ) qui est inhérente à un sujet ( x ), et à porter un jugement sur x en fonction de y . Ainsi, dans ce processus, l’analyse des noms dérivés range les phénomènes observés dans un ordre rationnel, et le sens interne établit la causalité entre l’appréhension et l’aﬀirmation logique. L’article se termine par une discussion sur la relation de cette


INTRODUCTION
This article examines Ibn Sīnā's (d. 1037) theory of experience (taǧriba) from a logical and psychological perspective. The following argument of this article is that his ideas regarding causality require the cross-reference between the linguistic doctrines of the Basran Muʿtazilite school of the science of kalām (ʿilm al-kalām, hereafter written as kalām) and the functions of the inner senses of the brain. 1 Although previous studies have described the impact of kalām on Ibn Sīnā's metaphysical and psychological concepts, there is still room for further study on the relationship between Ibn Sīnā's logical and methodological perspectives and the Basran linguistic tradition. 2 This article focuses on the key linguistic concept, derived name (ism muštaqq -i. e., paronym), that is, participles such as fāʿil ("acting / agent") and ʿālim ("knowing / knower"). The grammatical form of a derived name stems from its maṣdar (pl. maṣādir), or prototypic verbal noun; for example, the maṣdar of ʿālim is ʿilm ("an act of knowing / knowledge"). Thus, a derived name can be paraphrased into one sentence that contains its maṣdar; ʿālim is paraphrased into a nominal proposition lahu ʿilm ("an act of knowing belongs to him"). 3 For Ibn Sīnā, the paraphrased form li-x y (or, "y belongs to x") is not a mere grammatical rule, but a logical structure corresponding to reality; thus, the li-x y formula should have some counterparts in other fields of philosophy. 4 In this light, this article argues that Ibn Sīnā's theory of experience applies this formula for arranging observed phenomena into a form of a syllogism and describes how our brains' inner senses apprehend reality. Whereas previous studies have mentioned this formula in connection with Ibn Sīnā's Categories (Al-maqūlāt) and Metaphysics (Alilāhiyyāt), 5 this article approaches it with relevance to classical Arabic grammar and Basran kalām.
Although many scholars have studied Ibn Sīnā's theory of experience and indicated the involvement of the inner senses in this process, they are not clear on 1) the specific form of an Arabic proposition by which observed phenomena are expressed, and 2) how the brain's inner senses are involved in this process. 6 This article clarifies these mechanisms and contributes to a new perspective on Ibn Sīnā's epistemology, which Jon McGinnis characterized as "naturalized epistemology." 7 Although this author's standpoint is similar to McGinnis's view that Ibn Sīnā considered epistemological processes as embedded in psychological processes, this article argues that Ibn Sīnā elaborated on the nested structure in which psychology also presupposes logic in a certain way and that his 4 Concerning Ibn Sīnā's view on the correspondence between language and reality, see Ibn Sīnā, Al-šifāʾ. Al-ʿibāra, ed. Maḥmūd al-Ḫuḍayrī (al-Qāhira, 1970(al-Qāhira, [repr. 1984), 1-6. 5 Several studies have addressed this formula with relevance to the relative (muḍāf ) -one of the ten predicables. See Michael E. Marmura, "Avicenna's chapter on the relative in the metaphysics of the Shifāʾ," in G. F. Hourani (ed.), Essays on Islamic philosophy and science (Albany, 1975), p. 83-99; Olga Lizzini, "Causality as relation: Avicenna (and al-Ġazālī)," Quaestio, vol. 13 (2013), p. 165-195. Alexander Kalbarczyk has argued that the li-x y form, which is called the having mode of attribution in his study, was important in reforming the traditional scheme of Aristotle's Categories (Kalbarczyk, Predication, p. 74-115). 6 Only Jules Janssens mentions the inner senses' involvement in the process of experience. However, even he does not delve into the mechanisms behind it. Jules L. Janssens, "Experience (taǧriba) in classical Arabic philosophy (al-Fārābī -Avicenna)," Quaestio, vol. 4 (2004), p. 45-62, and p. 53-54 in particular. Dimitri Gutas and Jon McGinnis have also discussed the epistemological roles of the inner senses but did not discuss them in relation to experience. See Dimitri Gutas, "The empiricism of Avicenna," Oriens, vol. theory of experience is an outcome of the Arabicization of Aristotelian logic.

DERIVED NAMES IN CLASSICAL ARABIC GRAMMAR AND KALĀM
The fundamental ideas of Arabic grammar were expressed in The Book (Al-kitāb), written by Sībawayh (d. 796). One key idea is the name of the agent (ism al-fāʿil), which is paradigmatically presented as fāʿil ("acting / agent"). This is the starting point of the genealogy of derived names. It signifies an act -either perfected or not -insofar as it is attributed to a subject. 8 Sībawayh suggested the concept of maṣdar, which is connoted in the name of the agent and signifies a particular act that is distinguished from its subject. Taking ʿālim ("knowing / knower") as an example, this participle contains the maṣdar, namely, ʿilm ("an act of knowing"), which is both distinguished from and attributed to its subject. 9 Based on this analysis, Al-Mubarrad (d. 898) divided nominal propositions into two types. In the first type, the predicate is "identical to the subject in terms of meaning;" for example, "Zayd is your brother" (Zayd aḫūka). The second type includes a predicate that is "different from the first term;" e. g., "Zayd is standing" (Zayd qāʾim). 10 The second proposition, which contains the name of the agent, is distinguished from the first because the second's predicate contains a reference to something other than the subject (i. e., the act of standing signified by its maṣdar, qiyām) and the first's subject is identical to its predicate in meaning.
This classification of nominal propositions has had many repercussions throughout the history of Arabic thought. Basran mutakallimūn (kalām scholars) developed their own form of language analysis based on these arguments and then applied their analysis to ontological and epistemological issues. Both Basran Muʿtazilite and Ashʿarite scholars began their ontological investigations with linguistic study; they did not study the nature of things, but how things are expressed in the Arabic language. They examined a variety of ontological assertions (ithbāt), that is, judgments on which entities (šayʾ) are required to exist if a cer-tain speech act is true that are made by classifying the predication of names (ism) and adjectives (ṣifa). 11 These scholars also classified existing entities into three categories: God, atoms (ǧawhar), and entitative accidents (ʿaraḍ). Each of these are particular entities and can be subjects of the predicate "existent" (mawǧūd) in its proper meaning. 12 Ibn Sīnā does not explicitly cite mutakallimūn as his influence. Some scholars have discussed the influence of Abū Hāšim al-Ǧubbāʾī's (d. 933) mode (ḥāl) theory on Ibn Sīnā's ontology. 13 However, this article pays attention to his father, Abū ʿAlī al-Ǧubbāʾī's (d. 915) language analysis. He equated ṣifa (attribute) with an adjective or predicate as a part of speech. According to Frank, al-Ǧubbāʾī divided attributes into six classes based on their ontological assertions. 14 The relevant class for us is the attribute of the entitative ground (ṣifat al-maʿnā). Predicating this attribute implies that an entitative accident, which is different from something signified by a subject, really exists. Conversely, the subject is validly described by a given expression in virtue of the accident's existence. Hence, al-Ǧubbāʾī called this entitative accident "ground" (maʿnā) or "cause" (ʿilla). 15 Al-Ǧubbāʾī examines this ontological assertion by taking apart the derived name into the li-x y formula. He paraphrases the attribute ʿālim ("knowing / knower") into one sentence, lahu ʿilm ("an act of knowing belongs to him"). This paraphrased sentence itself is further predicated on the subject of the original sentence. Thus, the sentence Zayd ʿālim ("Zayd is knowing / a knower") is paraphrased into a sentence with two subjects: Zayd lahu ʿilm ("Zayd, an act of knowing belongs to him"). 16 11 Concerning the overview of the linguistic background of Basran kalām, see Frank, "Attribute." For more on the concept of entity, see Al-Ashʿarī, Kitāb maqālāt alislāmiyīn wa-iḫtilāf al-muṣallīn, ed. Hellmut Ritter (Wiesbaden, 1980), p. 122-123; Frank, Beings, p. 14. 12 Al-Ǧubbāʾī did not recognize the existence of universals. Regarding his nominalism, see Frank, "Attribute," p. 263. 13 For more on Abū Hāšim's impact on Ibn Sīnā, see Rashed, "Chose," p. 167-185; Damien Janos, Avicenna on the ontology of pure quiddity (Berlin, 2020), p. 366-375, 591-612. Early Basran Muʿtazilite and classical Ashʿarite methods of analyzing derived names are basically the same as al-Ǧubbāʾī's; they all paraphrase it into li-x y (Frank, Beings, p. 11-14). Abū Hāšim's mode theory does not examine ontological assertions by taking apart predicates but interprets the whole sentence's connotations by paraphrasing huwa ʿālim ("he is knowing") into kawnuhu ʿāliman ("that he is knowing"). See Frank, Beings, p. 19-27; Frank, "Attribute," p. 267-269. 14 Frank, "Attribute," p. 261-262. 15 Frank, Beings, p. 16; Frank, "Attribute," p. 262. 16 Frank, Beings, p. 16; Frank, "Attribute," p. 262.
The second sentence obtains clearer references to the entitative accident and its relationship with the subject; here, the act of knowing (the entitative accident), signified by ʿilm, belongs to the thing signified by the pronoun hu, which is identical to the subject Zayd.
Al-Ǧubbāʾī differentiated the attribute of the entitative ground from the attribute predicated by virtue of the subject itself (li-nafsihi). The latter attribute (hereafter called the attribute of essence) only describes the subject's essence. 17 Whereas the attribute of essence is used to state the tautological proposition x = y, the attribute of the entitative ground yields the relative proposition li-x y.

Ibn Sīnā's concept of a derived name
Before Ibn Sīnā's time, falāsifa (i. e., Arabic Aristotelians) had been discussing the grammatical structure of derived names in connection with Arabic Aristotelian logic. 18 In this context, Ibn Sīnā adopted a form of language analysis similar to the Basran tradition. The following section describes Ibn Sīnā's method of language analysis and how it differs from his predecessors.
Whereas al-Ǧubbāʾī's kalām is constituted by only particular entities, Ibn Sīnā's falsafa employs a protean concept, maʿnā (pl. maʿānī), which can indicate an existing thing that is concrete / material, mental, universal, or particular. 19 Although maʿnā can be translated as "notion," when it comes to its use in logic, the word has other meanings in other contexts (as shown below). Hence, this article does not attempt a direct English translation but instead provides a running commentary on the word's various uses and meanings.
Even though the works of Ibn Sīnā and al-Ǧubbāʾī contain different metaphysical content, they adopt formally similar methods. Ibn Sīnā analyzes derived names as follows: 17 19 The following passage demonstrates the multifaceted nature of maʿnā: "What goes out through a voice indicates what is in a soul, which is called impression (āthār). Things that are in the soul indicate things, which are called maʿānī, that is, intentions that belong to the soul (maqāṣid li-l-nafs). In comparison with words, impressions are maʿānī as well." Ibn Sīnā, Al-ʿibāra, p. 2-3.
[Predicates are divided into] primary names (ism mawḍūʿ), derived names (ism muštaqq), and verbs (kalima). […]. Derived names signify an unidentified subject in which a thing (amr), from which the name is derived, exists. Hence, it signifies a maʿnā or thing, its unidentified subject, and the relationship between them. You say, for example, "walking / walker" (māšin) because it signifies an act of walking (mašy), its unidentified subject, and that the act of walking belongs to him (anna al-mašy lahu). 20 In this passage, derived names signify three things: the maʿnā (signified by mašy), the unidentified subject (signified by the pronoun hu), and the relationship between them (signified by the particle li). The subject to which the maʿnā belongs is unidentified here because the sentence al-mašy lahu originally contains one derived name (māšin) and this name has not yet been predicated on a certain subject. Here, we can see that Ibn Sīnā's analysis of the derived name is formally parallel to al-Ǧubbāʾī's attribute of the entitative ground. The derived name is analyzed via the li-x y formula, and the maʿnā (y), which is signified by the maṣdar, belongs to the subject (x). 21 However, Ibn Sīnā does not regard the maʿnā as a ground of a certain proposition in the above passage. He describes the logical structure of the derived name and the significance of each term without exploring the cause of predication. Although he does not follow kalām's technical use of maʿnā as an entitative ground here, his use of maʿnā reflects grammatical tradition in that the participle contains the distinguished significance of the maʿnā, which is simpler than the participle, from which the participle is derived. 22 Ibn Sīnā employs two other types of paraphrased formulae, fī x y (y inheres in x) and ḏū y (an owner of y). For instance, abyaḍ ("white / a white thing") can be paraphrased as fīhi bayāḍ ("whiteness inheres in it") or ḏū bayāḍ ("an owner of whiteness"). To the best of the author's knowledge, falāsifa preferred ḏū y, possibly because it is more useful 20 Ibn Sīnā, Al-ʿibāra, p. 18. I translate ism mawḍūʿ, which can be directly translated as "established name," as "primary name" to emphasize its difference from derived name. I consider ism mawḍūʿ to be the ism ("name") that signifies a maʿnā by correspondence (bi-tawāṭuʾ) without temporal connotations, which Ibn Sīnā explains in the beginning of Chapter 2 of Al-ʿibāra. Ibn Sīnā, Al-ʿibāra, p. 7. See also n. 27 p. 222 below. 21 For more on the structure of derived names, see also Ibn Sīnā, Al-šifāʾ. Al-maqūlāt, for arranging a syllogism than either li-x y or fī x y, which inevitably yield sentences with two subjects. However, this formula was not used in classical Basran kalām. According to Ibn Sīnā, ḏū implies a particle of ascription, and thus another term can be attached to it. 23 Ḏū might be a comprehensive word that can be used to signify any type of ascription. 24 Concerning fī and li, Ibn Sīnā differentiates between them in the following passage.
The derived name has a certain ascription (nisba) -any ascription of any maʿnā, no matter whether it exists in [a subject] (mawǧūd fīhi) like eloquence, it belongs to [a subject] (lahu) like wealth, or it is attached to an action (mawḍūʿan li-ʿamal min aʿmālihi) like "sharp" (ḥadīd). 25 It seems that fī means the spatial relationship of inherence and li means the ascription itself without entailing spatial connotations. Therefore, the fī-ascription can also be described by the li-ascription; for example, Ibn Sīnā describes the relationship of whiteness with its subject through two phrases "this (whiteness) is in it (the subject)" (fīhi ḏālika) and "[whiteness] is in the possession of the subject" (mawǧūd li-l-mawḍūʿ). 26 In any case, the predication of any derived name implies the maʿnā, its subject, and the ascription of the maʿnā to its subject.
Ibn Sīnā also elaborates on the concept of maṣdar. He distinguishes between two maṣādir. The first signifies "an original meaning of a simple noun" (ḥaqīqat ḥāl al-ism al-muṭlaq), i. e., an act or state where a name is "named as a primary name" (mawḍūʿan waḍʿan awwalan), such as ḍarb ("hitting"). 27 The second signifies a relationship in which an 23 Ibn Sīnā, Al-maqūlāt, p. 144. According to this account, li-x y and fī x y more minutely express the logical structure of derived names. 24 Ibn Sīnā uses ḏū in various cases: "an owner of a disposition" (ḏū al-malaka) -a fī-ascription, "a thing owning a relationship" (šayʾ ḏū iḍāfa) -a li-ascription, and "a thing owning reason" (šayʾ ḏū nuṭq) -the differential ascription. Ibn Sīnā, Almaqūlāt, p. 148, 159, 102. 25 Ibn Sīnā, Al-maqūlāt, p. 16. The third kind of relationship does not ascribe an act or state to a subject, but rather ascribes a feature to an act. This article does not deal with this kind of ascription. 26 Ibn Sīnā, Al-maqūlāt, p. 20. 27 Mawḍūʿan waḍʿan awwalan can be translated as "established as primary establishment" verbatim. This phrase is associated with the linguistic view of the establishment of language (waḍʿ al-luġa). According to this view, language emerged out of a primordial establishment of the vocables (alfāẓ) for their meanings. Opinion is divided over whether the establisher of the linguistic code (wādiʿ) is God or primordial human society. Ibn Sīnā enumerates five possible sources for the origin of language: divine inspiration, human natural disposition to call a certain maʿnā by a certain word, human agreement (iṣṭilāḥ), gradual development, and any combination of these. However, he is not very interested in the origin of language; hence, he act or state arises out of (or is actualized by) its subject, as in ibyidād ("becoming white") and tabyīd ("making white"). 28 He also introduces the expanded concept of maṣdariyya, which encompasses both of these maṣādir. 29 He further distinguishes between two types of predications: that of a univocal name (ḥaml muwāṭaʾa) and a derived name (ḥaml ištiqāq). The former is "essentially and univocally" (bi-l-ḥaqīqa wa-l-muwāṭaʾa) predicated to its subject. For example, in the sentence "Zayd is a human" (Zayd insān), humanness is predicated to Zayd insofar as "the subject is identical [to the predicate]" (mawḍūʿ huwa); in short, x = y. 30 Conversely, the second type of predication is where whiteness is predicated to a human because whiteness is inherent in a human: for example, "the human is white or an owner of whiteness" (al-insān abyaḍ aw ḏū bayāḍ). These predicates are used in cases where we cannot say, "it is whiteness." 31 Ibn Sīnā's classification of predications is reminiscent of al-Mubarrad's classification of nominal propositions and Basran kalām's classification of attributes. His predication of the univocal name is parallel to al-Mubarrad's tautological proposition and al-Ǧubbāʾī's attribute of essence (the x = y proposition), and his predication of the derived name corresponds to al-Mubarrad's proposition whose predicate is different from the first term and al-Ǧubbāʾī's attribute of the entitative ground (the li-x y proposition). Kalbarczyk suggests that by instituting the division between the two types of attribution, Ibn Sīnā reformed the traditional scheme of Aristotle's Categories. 32 This article surmises that the correspondence between these ideas suggests that Ibn Sīnā's logic is influenced by both Greek and Arabic peripatetic tradition as well as the classical Basran linguistic tradition, and thus his work represents an attempted Arabicization of Aristotelian logic.
does not give an answer. When he mentions the establishment, he seems to mean the first connection of a simple maʿnā with a simple noun, from which a derived name is obtained, without asking the establisher. B. G. Weiss, "Al-waḍʿ al-lugha," in P.

The li-x y formula and related topics
Here, it is necessary to list the concepts by which the li-x y formula is analyzed. The terms x and y, and the relationship between them expressed by the particle li, can be examined in terms of the ten predicables. Three of these predicables -substance (ǧawhar), potentiality (quwwa; a subcategory of quality, or kayfiyya), and relative (muḍāf )are considered below since they are relevant to the discussion of Ibn Sīnā's theory of experience.

Substance
First, we should differentiate between substance and accident in essence. Ibn Sīnā defines substance as a thing of which essence is "existing not in a subject" (mawǧūd lā fī mawḍūʿ). 33 Thus, he equally recognizes the substantiality of immaterial substances (i. e., soul and intellect), matter, form, and form-matter composites. 34 Conversely, he defines accident as "existing in (fī) another thing." 35 Ibn Sīnā focuses on the ontological status of being in or not in a subject. Here, we should be careful not to confuse forms with accidents because, although forms are categorized as substance, we can say that they inhere in matter in a certain meaning. Thus, both form-in-matter and accident-in-subject relationships can be put into the fī x y formula; however, their meanings are completely different. An accident requires a substance in which the accident's essence is established; in other words, a substance's essence is established before an accident is attached to it. Hence, the substance is the subject of the accident. Meanwhile, form is an extrinsic attribute that gives matter existence without being a part of it, and matter has no actual existence in itself. Thus, matter is a place (maḥall) in which form exists, but not its subject. 36

Potentiality
Ibn Sīnā defines quality as "any constitution (hayʾa) that resides in a thing described by it and does not necessitate or require supposing it (= a thing described), and what can be conceived in itself without the necessity of paying attention to a relationship with things other than this constitution." 37 According to him, quality should be conceptualized 33  as such without considering relationships with other things, including the substance in which the quality resides. He then divides quality as follows: Quality is either related to the existence of a soul or not. What is not related is either related to quantity or not. What is not related [to quantity] is either a thing of which essence is a preparedness (istiʿdād) or a thing of which essence -in which the preparedness has arisen -is actuality. 38 Although he conceptualizes qualities as such, he also classifies these qualities into several subcategories according to the things to which they are accidentally related. The relevant subcategory for this discussion is preparedness. In another passage, he identifies preparedness as potentiality, and suggests that potentiality is not related to the soul or some quantity but rather ascribed to a body insofar as it is natural. 39 Ibn Sīnā defines potentiality as "physically perfected preparedness that is directed to an outer thing in a certain direction;" in other words, a physical predisposition directed to one of two opposites. For example, every person can be sick; however, a fragile person has the potential (or predisposition) to be sick rather than healthy. 40 It should be noted that Ibn Sīnā maintains another, disparate concept of potentiality. Using a wrestler as an example, he argues that the potentiality (as a subcategory of quality) for wrestling attributed to a wrestler is not the "primary psychological active potentiality" (alquwwa al-ūlā al-muḥarrik al-nafsāniyya). He interprets the latter as a substance, which is presumed to be the animal soul. Rather, the potentiality as quality is the physical perfection of the wrestler's organs. 41

Relative
The category of relative is essentially related to the particle li. Ibn Sīnā defines it as "a thing of which quiddity is said with regard to (bi-lqiyās) another thing." 42 Taking potentiality as an example, we shall see how it functions in this category.
Ibn Sīnā defined quality as such without paying attention to its relationship with other things, including its subject. However, potentiality necessarily involves a relationship, as any potentiality is the potential for something. Thus, the quiddity of potentiality might mistakenly be conceived as relating to something else and identified as the relative.
To solve this problem, Ibn Sīnā distinguishes between per se and accidental relationships. Whereas potentiality has a per se relationship with its object, it has an accidental relationship with its subject -signified by the particle li. He states that "potentiality and faculty are potentiality and faculty ascribed to a thing (li-šayʾ) for a thing (ʿalā šayʾ)." 43 Whereas "for a thing" is potentiality's essential relationship with an object, "ascribed to a thing" is its accidental relationship with a subject, which renders the potentiality relative. When potentiality is conceptualized in the form of lahu quwwa ("potentiality belongs to him"), "its conceptualization requires the conceptualization of a thing external to it." 44 This account is consistent with the significance of derived names. Qawiyy (potential) -the derived name of quwwa -contains three significations: potentiality, its subject, and the ascription of the former to the latter. Hence, it does not refer to the essence of the potentiality as such. To sum up, a thing of which essence is conceived as such can be rendered relative by attaching the particle of ascription li. 45 Ibn Sīnā also differentiates between the unilateral relationshipi. e., ascription (nisba) -and the bilateral relationship -i. e., correlation (iḍāfa). In the former, the particle li attaches to one of the two terms (for example, knowledge is ascribed to a knower, but not vice versa); and in the latter, it attaches to both terms (the relationship between two brothers). Although he does not employ these terms consistently, the current article adopts this classification. 46 Note that the interpretation of the term y of li-x y is not always unambiguous; y can mean both an act and potentiality, and the latter can be interpreted as a substance as well as an accident. 47 47 If we interpret the term y as an act, we might be inclined to make li-x y another category, "to act" (an yafʿala). However, Ibn Sīnā defines the category as "ascription of a substance to a thing (i. e., an act) that emanates from it and is not essentially stable, but repeats its renewal and passing." In this definition, he renders the subject relative to its act; in other words, the definition is put into li-y x but not li-x y. Thus, li-x y, which renders the act relative to its subject, cannot be subsumed under this substituted for y, whether this mašy is interpreted as an act of walking or as the two types of potentiality is determined through context.

BRAIN, LOGIC, AND EXPERIENCE
This section explores how Ibn Sīnā's doctrine of experience combines the analysis of derived names and the inner senses of the brain. Previous works on Ibn Sīnā's doctrine of experience have tended to focus on accounts in Al-burhān and Al-išārāt wa-l-tanbīhāt; however, this article focuses on texts in Al-naǧāt and Al-muḫtaṣar al-awsaṭ fī al-manṭiq because the latter explicitly mention the involvement of the inner senses. I will briefly examine all four accounts here.
In Al-burhān, according to Ibn Sīnā, the process of our mental judgement (ḥukm al-ḏihn) that a repeated observation induces within us produces a conviction of causality as follows. When the phenomenon of scammony purging yellow bile repeats, we perceive this phenomenon as taking place not by chance, but due to a certain cause. This cause is neither a volitional selection -which does not belong to the scammony -nor the scammony's body, insofar as it is a mere body. Accordingly, we assume that the cause of the phenomenon lies in scammony's nature or its concomitants. Thus, Ibn Sīnā identifies the faculty for purging attributed to scammony's nature as the middle term of the syllogism, although he admits that identifying the real cause is not a proper subject of logic. 48 In Al-išārāt, he briefly summarizes his idea as follows: experience can produce within us a conviction that there is some causal relationship between observed phenomena, but that as far as logic is concerned, we cannot determine what this causal relationship is. 49 In both works, Ibn Sīnā outlines the process of mental judgement; however, he does not specify the mental faculties involved in this process. The account in Al-naǧāt fills this gap: Experienced things (muǧarrabāt) are things by which the perception (alḥiss) in cooperation with syllogism brings about a logical assent (taṣdīq). When [the phenomenon] that the existence of a thing belongs to another thing (wuǧūd šayʾ li-šayʾ) -for example, that the purgation belongs to the category even though we can interpret y as an act. See Ibn Sīnā, Al-naǧāt: min alġaraq fī baḥr al-ḍalālāt, ed. Muḥammad Taqī Dānishpazhūh (Tehran, 1986) Y. HOKI scammony (al-ishāl li-l-suqammūniyā), or that the observed movement belongs to the celestial body -is repeated in our perception, this is repeated in our recollection (ḏikr). Then, when this repeats in our recollection, by means of the syllogism connected with the recollection, experience occurs. 50 Here, he describes experienced things with reference to the li-x y formula and suggests that the syllogism cooperates with our faculty of recollection to produce a logical assent. In his view, recollection is one of the brain's five internal senses. It functions by preserving maʿānī that are extracted from objects of perception by another inner sense, estimation. In Al-muḫtaṣar, Ibn Sīnā explicitly mentions the faculty of estimation regarding experience: When perception is repeated over perceived objects, one maʿnā, out of which memory (ḥifẓ) emerges as a faculty of thought, is brought to estimation from them. When this memory is repeated several times, experience emerges out of it. In short, memory emerges out of repetition of perception, and experience emerges out of repetition of memory. 51 Ibn Sīnā suggests that estimation, which extracts the maʿnā from objects of outer senses, contributes to the emergence of experience. However, the explanations in Al-naǧāt and Al-muḫtaṣar do not provide a full picture of the involvement of this psychological faculty in experience. Therefore, we should examine the concept of estimation below. 52 According to Ibn Sīnā, estimation takes place at the rear of the middle part of the brain. He insists that human beings share the sense of estimation with other animals, and therefore it is not a rational faculty. 53 He recognizes the following two functions of estimation. The first is to perceive a maʿnā -which cannot be perceived by exterior senses -that resides in a perceived object: for example, a sheep perceives a maʿnā of muḍādd ("an opposed thing") in a wolf. 54 This maʿnā is itself immaterial, but nevertheless embedded in a particular body. In other words, it is not a subjective interpretation of sense data, but it is "existent in particular objects of perception" (al-mawǧūda fī al-maḥsūsāt al-ǧuzʾiyya). 55 The second is to "make a decisive judgment on a thing that it is so-and-so or not so-and-so" 56 calling for reaction; for example, the sheep's judgment that the wolf is a threat and the action of fleeing from it. 57 Ibn Sīnā detects causation between the first and second functions. He insists that the maʿnā grasped by estimation "necessitates" the sheep's fear of the wolf and reaction to run from it. 58 In short, our sense of estimation perceives the maʿnā (y) inherent in something (x) and makes a judgment about x based on y.
These functions of estimation are reminiscent of al-Ǧubbāʾī's analysis of the attribute of the entitative ground: the maʿnā, in both, functions by causing judgments about the thing in which the maʿnā resides. Therefore, we might expect some parallels between the functions of estimation and Ibn Sīnā's analysis of derived names. However, he does not expound on estimation in a strictly logical way because he does not regard it as a rational faculty. He does not describe judgments from estimations as propositions that contain predicates derived from the corresponding maʿānī: for example, in the sheep's judgement that "al-ḏiʾb ʿadūw" (the wolf is an enemy). 59 Besides, some maʿānī do not come in the form of maṣdar -e. g., in the sheep's perception of "an opposed thing" (muḍādd). In short, he recognizes causation between the first and second functions of estimation, but does not arrange them in a rational order.
Nevertheless, Ibn Sīnā admits that human estimation can become rational through "adjacency of reason (or articulated speech)" (muǧāwarat al-nuṭq). 60 He does not clarify how this adjacency functions; however, by comparison with animal estimation, it is reasonable to see this nuṭq (reason / speech) as a faculty that divides a perceived object in accordance with linguistic order. Thus, we can infer what this joining of estimation and reason / speech brings to his doctrine of experience. In short, he introduces the inner senses to establish a causal relationship between the apprehension of reality and the logical assertion. Whereas the maʿnā 1i. e., the maʿnā as the object of logic -obtained through analysis of a derived name is not the cause of predication, but the connoted signification 55  of a derived name, our estimation can extract the cause that requires a certain judgment. Conversely, reason can provide an estimation with logical structure by describing the object in the li-x y formula because the estimation perceives the maʿnā 2 (y) -the maʿnā as the object of the estimation -as inherent in the other thing (x). Moreover, this structure grants estimation the potential to compose a proposition where the subject is x and the predicate is the derived name of y.
This cooperation requires describing a repeated phenomenon in a specific Arabic formula. In Al-naǧāt, Ibn Sīnā uses the formula wuǧūd šayʾ li-šayʾ ("an existence of a thing belongs to another thing"): for example, purgation belongs to scammony. Conversely in Al-burhān, he describes the same phenomenon as "the scammony is purging yellow bile" (al-suqammūniyā mushil li-l-ṣafrāʾ) and its paraphrase, "the scammony, the purgation of yellow bile occurs to it" (al-suqammūniyā yaʿriḍu lahu ishāl al-ṣafrāʾ). 61 These sentences indicate a crossover between the analysis of a derived name and the functions of estimation.
From the viewpoint of logic, "the scammony is purging yellow bile" is a singular proposition that includes the derived name, mushil ("purging"). This predicate can be paraphrased into the sentence lahu ishāl ("the purgation belongs to it"). Hence, the original sentence can be paraphrased into a sentence with two subjects, such as "the scammony, the purgation of yellow bile belongs to it (or occurs to it)." This ishāl is a maṣdariyya, which can be written as ishāl 1 corresponding to maʿnā 1 . It should be noted that the latter proposition does not express a mere simultaneous occurrence of the two things, but the unilateral relationship (i. e., ascription of purgation to the scammony). This analysis can arrange the observed phenomenon into a rational order but cannot determine its cause.
Conversely, when this phenomenon is repeated via recollection, maʿnā 2 must be repeatedly perceived by the estimation. Using the analysis of mushil as an analogy, we can posit maʿnā 2 as ishāl 2 . To be exact, the object of the estimation should be described as lahu ishāl 2 because it always perceives the maʿnā 2 as an object embedded in a particular body. Then, because the maʿnā 2 is the basis of the judgment, the perception of lahu ishāl 2 can produce the proposition "the scammony purges" (al-suqammūniyā mushil) with the assistance of logic. This combination of logic and psychology yields the following syllogism: (minor premise) The scammony is the owner of ishāl 2 61 Ibn Sīnā, Al-burhān, p. 95.
(major premise) The owner of ishāl 2 purges yellow bile when digested.
Here, by employing both the analysis of the derived name and the functions of the inner senses, we can form the general framework of a syllogism. Indeed, this syllogism is not scientific knowledge, because that would require the abstraction (taǧrīd) of the essence from particulars. 62 However, it seems here that logic and psychology have already started to advance in harmony toward finding the true cause.
The above argument contradicts Ibn Sīnā's assertion that the maʿnā 2 is immaterial and not perceived by the external senses if ishāl 2 is interpreted as a physical phenomenon. The external senses perceive the phenomenon described as lahu ishāl 1 as the attribution of an act or preparedness (i. e., a physical disposition); however, these are not objects of estimation. According to the above quotation from Al-naǧāt, 63 the phenomenon perceived by recollection can also be written as lahu ishāl 1 . We should interpret it as an immaterial phenomenon. The following section addresses how this problem can be resolved by consulting the natural sciences.
When equating the cause of purgation with the scammony's nature or its concomitants in Al-burhān, Ibn Sīnā suggested that the scammony's effect might be identified as its proximate faculty, specific property (ḫāṣṣa), or other factors related to them. 64 Of these, specific property is presumed to be the most probable answer. Specific property is discussed in medicine as well as logic. In medicine, the term refers to any medicinal effect that cannot be reduced to the four primary qualities of the elements and their mixture (e. g., the magnetic attraction and purgative effect for a specific kind of humor). In Al-qānūn fī al-ṭibb, Ibn Sīnā categorizes the magnetic attraction and the nature of all animal and plant species as "agents by virtue of the substance" (fāʿil biǧawhar). 65  hence, it comes from the substance itself and is not explained in terms of the qualities. At the concrete level, a substance is a form-matter composite; however, form also refers to the substance itself or the immaterial factor that determines what the thing is. Thus, Ibn Sīnā identifies the cause of a specific property as "divine emanation" (fayḍ ilāhī), 66 that is, a specific form (ṣūrat nawʿ) granted by active intellect. 67 In logic, specific property is a concomitant of a subject's quiddity. It is not a per se attribute that constitutes the essence of the subject, but instead is accidental. Nevertheless, specific property is not separable from its subject. 68 According to Al-išārāt, properties that comprise a species, are confined to it, and are not separable from it are most properly called specific properties. 69 These properties are attached to their subject by virtue of its quiddity; for instance, triangles have the specific property that their three angles add up to 180 degrees by virtue of their triangularity. 70 Here, Ibn Sīnā recognizes the necessary dependence of the specific property on the subject's quiddity. To summarize, in both medicine and logic, specific property is actualized by its subject's quiddity. 71 Considering these arguments, ishāl 2 cannot be equated with a real act and preparedness, which are perceived by the external senses. It is instead related to the scammony's specific form that determines its quiddity. However, human estimation cannot comprehend abstracted forms as such; it can only perceive an insufficiently abstracted connotation, which dwells in the scammony. As ishāl 2 is not the scammony's form, but its accidental property, its relationship to the scammony is not a form-in-matter relationship, but an accident-in-substance relationship. 66 Ibn Sīnā defines specific properties as "nature, which is existent in a body composed of elements by a specific mixture that acquires a specific preparedness that occurs to the body, through the divine emanation." Ibn Sīnā However, this accidental property is brought about by the scammony's form; therefore, it suggests the real cause. As mentioned above, ishāl 1 is the maṣdariyya, which itself has a subsystem in which the subject actualizes the act of purgation. Hence, ishāl 2 can be further analyzed and abstracted to obtain the true cause of the subject's actualization. In short, to perceive the maʿnā 2 is to realize that the subject actualizes the act through a certain immaterial cause. Although this perception alone cannot provide scientific knowledge, it can capture a sense of causality that leads to a rough syllogistic framework.
The preceding discussion shows that Ibn Sīnā's theory of experience necessitates the cross-reference between logic and psychology; logic provides psychology with rational structure, and psychology establishes the causal relationship between perception and proposition. By introducing estimation and recollection, Ibn Sīnā is attempting to obtain pieces of causality before examining the nature of objects. The upshot of this approach is that the fundamental structure of causality is instilled in the human brain. Thus, humans can innately perceive the maʿnā 2 (y) as a cause inherent in something else (x) and compose a proposition concerning x based on y. Although human estimation can mistakenly attribute causality to the wrong maʿnā 2 , it still can produce a formally correct syllogistic framework in every case.

REPERCUSSIONS IN MEDICINE
Given the relationship between logic and medicine discussed above and the way in which Ibn Sīnā gave an empirical basis for obtaining new knowledge through experience, next we should ask what impacts his theory of experience had on medical science. The answer is that it had some contradictory effects in that it restricted his approach to medicine.
In the medieval Arabic medical system, contrary to Ibn Sīnā's assertion, it was not obvious whether a certain medicine's effect on a human body was attributable solely to the nature of the medicine. This is because the qualities and effects of a medicine can only be estimated by observing its interaction with the human body. Ibn Sīnā acknowledges this principle of medicine in the Al-qānūn. 72  Medical essentialism poses problems regarding the prescription of medical doses. For instance, Hippocrates describes a case where prescribing a large amount of purgative medicine purges other kinds of humors that are not being targeted. 74 ʿAlī b. al-ʿAbbās al-Maǧūsī (d. 994) recounts a similar case. 75 Both insisted that the cause of this phenomenon is the "surplus faculty" that remains in a medicine. 76 However, their explanations do not answer why the surplus faculty purges other kinds of humors, since this faculty was originally conceived to be effective only for a specific kind of humor. Thus, physicians before and after Ibn Sīnā's time had to contend with the unsolved problem of how to gauge a medicine's quality, quantity, and its interaction with the human body when prescribing doses.
Ibn Sīnā simply underestimated this problem. For him, the proposition "scammony purges yellow bile by its nature" was necessary and sufficient. He recognized the existence of anomalies in this empirical proposition; however, he considered them accidental. 77 Rather, he criticized people who were not satisfied with his explanation and forged diverse factors that operated between the two objects: the effect of heat, the correspondence of the nature of the two objects, and so on. 78 Ibn Sīnā concludes the following: When a questioner asks why the fire burns, the only answer is that it is hot. The meaning of this answer is that the fire has a burning faculty by virtue of its nature. Likewise, when a questioner asks why the magnet attracts iron, the only answer is that it has an attracting faculty by virtue of its nature. 79 This statement is a natural consequence of his doctrine of experience; causes of effects are attributed solely to their subjects in accordance with the li-x y formula. A more meticulous investigation leads to methodological error.
Ibn Rušd refuted Ibn Sīnā's essentialism by asserting that if this theory was correct, a great deal of a purgative medicine would not purge humors other than the ones targeted. 80 Ibn Rušd insisted that the cause of the scammony's effect was not its natural faculty, but the "proportion and the correspondence" (nisba wa-muwāfaqa) of the mixture of elements between the scammony and the humors. In other words, the scammony's effect on yellow bile is actualized by a certain proportion of elements within the medicine and the humor, and therefore different humors can be purged by different proportions of the elements. Thus, this effect is an accident of the medicine, not inherent in it. 81 To summarize, Ibn Sīnā's essentialism approach is consistent in diverse fields including logic, psychology, and medicine; however, this consistency precludes an examination of the correlation between doses of medicine and the human body. Conversely, Ibn Rušd investigated how each medicine has varied effects on the human body according to its dosage. This is a proper approach to medicine; however, it reduces the nature of medicine to a relative phenomenon. In short, the former approach viewed the medical system as based on ascription, whereas the latter is based on correlation.

CONCLUSION
This study found that Ibn Sīnā's theory of experience presupposes four conditions: a singular proposition containing a derived name, its paraphrase into li-x y, our estimation's perception of y embedded in x, and our ability to determine y as the cause of the proposition concerning x. In his mind, the logical process arranges observed phenomenon into this rational order but cannot determine causality on its own, and our estimation detects the cause that necessitates a judgment but lacks the rational order. However, the logical structure of derived names itself does not presuppose psychology, and the functions of the inner senses 80 Ibn Rušd, Al-kulliyyāt, p. 376-378. Concerning Ibn Rušd's theory of the specific property, see Yu Hoki, "Ibn Rushd's criticism of the theory of the inherence of the specific property (khāṣṣa)," Orient, vol. 57 (2014), p. 33-48 (in Japanese with an English summary). 81 Ibn Rušd, Al-kulliyyāt, p. 377.
are not logical in themselves. To create a full epistemological picture, by "adjacency" of both, Ibn Sīnā created a nested structure in which these two factors, which are irrelevant to each other, presuppose each other.
We should also note that the Basran linguistic tradition played a decisive role in this epistemological process. Ibn Sīnā has two methodological features in common with Baṣran mutakallimūn; however, he deals with each in different fields. The first is the method of analyzing derived names in logical writings; here, he does not see the maʿnā as the cause of a predication. The second is that in psychology he detects the maʿnā, which is inherent to its subject, as the cause of a judgement about the subject. Thus, Ibn Sīnā's doctrine of experience might be best understood as an outcome of the Arabicization of Aristotelian logic. This study addressed a limited part of his philosophy. However, this study also suggests the possibility that, in other fields, Ibn Sīnā employed a holistic approach in which Arabic grammar, Aristotelian logic, psychology, and the nature of objects share a parallel structure.