It has been recognized for some time that the view that the subject matter of logic is “second intentions”, widely held in late-medieval and early-modern Latin scholasticism, can be traced back to the writings of Arabic logicians. As shown by Professor A.I. Sabra in a classic article from 1980, the concept of “second intentions” is already in evidence in the works of Fārābī, but the view that such second intentions form the special subject matter of logic seems first to have been explicitly advanced by Avicenna.Footnote ¹ This position was later adopted in Latin scholastic circles in the course of the thirteenth and fourteenth centuries via translations of the relevant writings of Avicenna (particularly the Eisagoge and Metaphysics of al-Shifāʾ). Curiously, a number of prominent Arabic logicians abandoned Avicenna's position in the very same centuries during which it was being assimilated by Latin scholastics. As has been noted recently by Tony Street, a rival and influential view on the subject matter of logic emerged in the writings of Afḍal al-Dīn al-Khūnajī (d. 1248) and gained widespread acceptance.Footnote ² The present paper will explore the vigorous debate by thirteenth and fourteenth century Arabic logicians on the respective merits of the views of Avicenna and Khūnajī.

I

The Persian-born Cairo-based scholar and judge Afḍal al-Dīn al-Khūnajī is a largely forgotten figure today, but he was considered to have been a major figure in the history of Arabic logic by the well-known North African historian Ibn Khaldūn (d. 1406).Footnote ³ Ibn Khaldūn noted that Khūnajī's works on logic were “relied upon” by the eastern logicians of his day. North African scholars were obviously also beginning to show interest in these works by the fourteenth century, for Ibn Khaldūn had himself memorized Khūnajī's short epitome of logic al-Jumal, and his teacher Muḥammad al-Sharīf al-Tilimsānī (d. 1370) had written a commentary on it. It can be shown that on a range of issues, Khūnajī's innovative works on logic, and especially his lengthy summa Kashf al-asrār, had a major impact on the later logical tradition. Khūnajī appears, for example, to have been the first to offer an extensive discussion of the immediate implications of conditional and disjunctive propositions, recognizing along the way De Morgan's laws: an affirmative disjunction (P or Q) is logically equivalent to a negated conjunction (Not both not-P & not-Q), and a negative disjunction (Not: P or Q) is logically equivalent to a conjunction (Both not-P & not-Q). Khūnajī also appears to have been the first Arabic logician to suggest that possibility propositions do not convert and that first-figure syllogisms with possibility minors cannot be shown to be productive, one of the most distinctive features of later Arabic modal logic. On these points, and a number of others, Khūnajī's innovative suggestions were taken up by the later Arabic tradition and enshrined in the Arabic handbooks on logic that continued to be studied in colleges until the modern age.

On the question of the subject matter of logic, Khūnajī also departed from received Avicennan doctrine. In his Kashf al-asrār, the subject matter of logic is presented as being “the objects of conception and assent” (al-maʿlūmāt al-taṣawwuriyya wa-al-taṣdīqiyya).Footnote ⁴ Khūnajī did not engage in polemics on this point, nor did he explicitly state that he was departing from received views on this issue. However, in the works of Najm al-Dīn al-Kātibī (d. 1277) and Khūnajī's student Ibn Wāṣil al-Ḥamawī (d. 1298), the view that the subject matter of logic is “the objects of conception and assent” is attributed to Khūnajī and explicitly presented as a superior alternative to the Avicennan view that it is “second intentions” (al-maʿqūlāt al-thāniya).Footnote ⁵ Ibn Wāṣil wrote in his Nukhbat al-fikar, which he dedicated to the Hohenstaufen ruler of Sicily Manfred (r. 1258–1266), that the received view is that the subject matter of logic is second intentions (al-maʿqūlāt al-thāniya). He then added:

Ibn Wāṣil did not, however, expand on why this view is superior to the older view that the subject matter of logic is second intentions. By comparison, a relatively detailed and clear account of the motivation behind this departure from Avicenna is given by Kātibī in his monumental commentary on Khūnajī's Kashf al-asrār and in his own summa of logic Jāmiʿ al-daqāʾiq.

Kātibī's point of departure was the generally accepted view that the subject matter of a science is that whose “intrinsic accidents” (ʿawāriḍ dhātiyya or aʿrāḍ dhātiyya) are studied in that science. The notion of “intrinsic accidents” is worth an independent study. It is central to Avicenna's conception of science and elicited considerable discussion and controversy in the later Avicennan tradition.Footnote ⁷ In the present context, I confine myself to Kātibī's own brief explication of the notion. In his Jāmiʿ al-daqāʾiq, he explained that aʿrāḍ dhātiyya are accidents that accrue to a subject because of (i) the subject's essence (talḥaquhu li-dhātihi), such as the ability to wonder which is true of humans because of their essence, viz. rational animal; or (ii) because of something extrinsic that is co-extensional with the subject (li-mā yusāwīhi), such as the ability to laugh which is true of humans because of the ability to wonder, or (iii) because of something that is of more general extension than the subject but is part of its essence (li-amrin aʿamma dākhilin fīhi), such as moving at will which is true of humans because of their being animals.Footnote ⁸

Kātibī went on to give a succinct overview of the dispute on the subject matter of logic. He mentioned that the older view of Avicenna and Fakhr al-Dīn al-Razī (d. 1210) was that the subject matter of logic is second intentions. The notion of second intentions he explained thus:

On this account, the realities of things constitute the primary objects or “first intentions” of the mind. The mind then makes delimitative and predicative judgments about these first intentions. The delimitative and predicative judgments to which the first intentions are subjected are called “second intentions” since they are in the second order of intellection. The standard view is that logic investigates these second intentions, not in themselves but insofar as it is possible by their means to derive what was previously unknown. The fact that second intentions lead to new knowledge is not part of their definition (and in this sense it is an “accidental” fact about them) but nevertheless accrues to them by virtue of what they are (and in this sense it is “intrinsic” to them). For example, the conception of a genus and a differentia (li-anna taṣawwura al-jinsi wa-al-faṣli al-qarībayn) leads intrinsically to the conception of the thing defined (yūṣilu li-dhātihi ilā taṣawwuri al-maḥdūd), and an assent to two premises sharing a middle-term (li-anna ḥuṣūla taṣdīqayni mushtarikayni fī ḥaddin huwa al-awsaṭu fī al-ʿaql) and arranged in a productive mood of the syllogism leads intrinsically to a further assent involving the two extremes (yūṣilu ilā al-ʿilmi bi-intisābi aḥadi al-ḥaddayni al-bāqiyayni ilā al-ākhar).Footnote ¹⁰ Since the subject matter of a science is that whose “intrinsic accidents” are investigated in that science, it follows that second intentions are the subject matter of logic.

Kātibī went on to add that “some” have made the subject matter of logic the objects of conceptions and assents, rather than second-intentions. (In his commentary on Khūnajī's Kashf al-asrār, Kātibī made it clear that this was the position of Khūnajī.) On this account, the subject matter of logic is “more general” (aʿamm) than second intentions since it would include second intentions and also conceptions that are not delimitative combinations. This view is preferable, he added, since:

Kātibī's point may be summarized as follows: The received view is that the subject matter of logic is second intentions, i.e. delimited conceptions and predicative judgments, insofar as these lead to new conceptions and assents. The view that delimited conceptions and judgments are the province of logicians is attested already in Fārābī,Footnote ¹² though what is not attested in Fārābī's extant works is the identification of such delimited conceptions and predicative judgments with second intentions. The objection mentioned and endorsed by Kātibī is that such second intentions are not exhaustive of the subject matter of logic. Rather, logic also investigates predicates that accrue to (yalḥaqu) simple notions (maʿānī mufrada). What he had in mind becomes clearer from his commentary on Khūnajī's Kashf al-asrār. He there wrote that it is an intrinsic accident of the first intention “rational” that it is a differentia of human, and it is an intrinsic accident of the first intention “animal” that it is a genus. The logician “investigates” (yabḥathu ʿan) concepts such as “differentia” and “genus” and, crucially, these are intrinsic accidents of first intentions. It follows that the subject matter of logic includes first intentions as well as second intentions.

It is important to note that Kātibī, as well as other Arabic logicians who debated the subject matter of logic in the thirteenth and fourteenth centuries, drew a distinction between what logicians “investigate” (yabḥathu ʿan) or “look into” (yanẓuru fī) and what is the subject matter (mawḍūʿ) of their discipline. What is “investigated” are the intrinsic accidents of the subject matter, and not the subject matter itself. Unless one is alive to this distinction the thrust of Kātibī's argument might well appear baffling, for he argues against the view that the subject matter of logic is second intentions by pointing out that logic “investigates” concepts such as “genus” and “differentia” which are textbook examples of second intentions!

Another noteworthy aspect of Kātibī's discussion is his explanation of second intentions. As mentioned above, a second intention in his view is either a delimited conception like “rational animal” or a categorical judgment like “The rational is an animal”. These are called “second intentions” because they involve a bringing together of simple, first-order concepts such as “rational” and “animal”. This view seems to be derived from al-Mulakhkhaṣ of Fakhr al-Dīn al-Rāzī which explicates “second intentions” in very similar words.Footnote ¹³ One might compare Rāzī's and Kātibī's explication of “second intention” with that given in Avicenna's Eisagoge of al-Shifāʾ: Quiddities have accidents (ʿawāriḍ) that accrue to them on account of their existing extra-mentally, and accidents that accrue to them on account of their existing in the mind, such as being essential or accidental, universal or particular, subject or predicate. According to the explanation given by Avicenna, a “second intention” is what accrues to first intentions due to the latter's existence in the mind. For example, the second intention “universal” accrues to animal due to the latter's existence in the mind, whereas the first intention “moving” accrues to it due to its existence in the external world. The subject matter of logic, Avicenna added, is the accidents that accrue to quiddities by virtue of these quiddities being in the mind, from the perspective (min jihati) of these accidents being useful for the task of deriving new knowledge from what we already know.Footnote ¹⁴ Avicenna went on to specify that it is not sufficient to say in general that logic is occupied with mental existents (wujūd al-ashyāʾ fī al-dhihn). For mental existents are of two types: (i) those derived from extra-mental entities, and (ii) those derived from other mental entities insofar as they are in the mind and have no correlate in the extra-mental world. It is the second type of mental existents that is the province of the logician.Footnote ¹⁵ A summary of this view of second intentions is given by Khūnajī's student Ibn Wāṣil in his Nukhbat al-fikar:

Although the explication offered by Avicenna and Ibn Wāṣil seems very different from the one offered by Rāzī and Kātibī, it is not clear that the explications were intended to be incompatible or involved any disagreement about what is and is not a second intention. It is revealing that Kātibī, in his commentary on Kashf al-asrār, first cited Avicenna's explanation of second intentions from al-Shifāʾ and then cited Rāzī's explanation from al-Mulakhkhaṣ without any indication that they are competing or irreconcilable. On the contrary, he claimed that Rāzī's account of “second intentions” filled in the details (faṣṣala) of the general account of Avicenna. He wrote:

An account that bridges the apparent gap between Avicenna's and Rāzī's explications of second intentions is arguably to be found in al-Baṣāʾir al-Naṣīriyya by ʿUmar b. Sahlān al-Sāwī (fl. 1145).Footnote ¹⁸ According to Sāwī, the subject matter of logic is the notions (maʿānī) of which the “explicatory phrase” (qawl shāriḥ) and argument (ḥujja) are formed. These maʿānī are second intentions in the following sense: the mind intellects the forms (ṣuwar) and quiddities (māhiyyāt) of extra-mental things and then manipulates these forms/quiddities in various ways: for example combining them with other extrinsic matters (yulḥiqu bi-baʿḍihā umūran laysat minhā), or predicating some of others (yaḥkumu bi-baʿḍihā ʿalā baʿḍ), or abstracting from them their non-essential attributes (yujarridu baʿḍahā ʿan ʿawāriḍin khārijatin ʿan ḥaqīqatihā). Such abstractions and combinations are states that only accrue to the forms/quiddities as they are in the mind (al-aḥwāl al-ʿāriḍa lahā baʿda ḥuṣūlihā fī al-dhihn). The forms/quiddities are first intentions, whereas the states that accrue to these first intentions in the mind are second intentions. The subject matter of logic is these second intentions insofar as they lead to acquiring new knowledge.Footnote ¹⁹

Sāwī's explication is clearly in line with both that of Avicenna and that of Rāzī. His explanation of “second intentions” is close to the one given by Avicenna but like Rāzī he gave propositions and delimited conceptions as paradigm examples of second intentions. There is therefore good reason to believe that the explication of “second intentions” offered by Rāzī (who was familiar with Sāwī's work) was not intended to be incompatible with the explication of Avicenna and that it did not result in diverging views about what is and is not a “second intention”. One might speculate that Rāzī's alternative explication was designed in part to avoid any explicit appeal to “mental existence” (wujūd dhihnī) which was a topic of contention between theologians and philosophers, the former typically denying mental existence altogether.Footnote ²⁰ Simply identifying “second intentions” with delimited conceptions and predicative judgments on the grounds that these presuppose a primary intellection of simple realities might have appeared like a neat way of sidestepping a host of contentious metaphysical claims.

II

Khūnajī's view that the subject matter of logic is the objects of conception and assent was accepted by a number of prominent thirteenth-century Arabic logicians. Kātibī and Ibn Wāṣil have already been mentioned, but the same is true of Kātibī's teacher Athīr al-Dīn al-Abharī (d. 1265) and Khūnajī's slightly younger associate Sirāj al-Dīn al-Urmawī (d. 1283). More importantly, the view came to be enshrined in the classic handbooks on logic that continued to be studied in Muslim colleges until the modern period. These include: al-Risāla al-Shamsiyya by Kātibī, Maṭāliʿ al-anwār by Urmawī, and Tahdhīb al-manṭiq by the Timurid scholar Saʿd al-Dīn al-Taftāzānī (d. 1390).Footnote ²¹ Even much later Arabic handbooks on logic continued to expound this view, for example Sullam al-ʿulūm by the Mughal scholar Muḥibbullāh Bihārī (d. 1707), al-Burhān fī ʿilm al-mīzān by the Turkish Ottoman scholar Ismāʿīl Gelenbevī (d. 1791), and al-Laʾālī al-muntaẓama by the Qajar scholar Mullā Hādī Sabzevārī (d. 1878).Footnote ²² However, despite the near-consensus of authors of handbooks, the question of the subject matter of logic continued to be discussed from the thirteenth century to the nineteenth, and this was often noted in the numerous commentaries and glosses that these much-studied handbooks elicited.

The brilliant polymath and philosopher Naṣīr al-Dīn al-Ṭūsī (d. 1274) was an important figure in what might broadly be described as the resurgence of more orthodox Avicennism after the critiques of Avicenna's philosophy and logic by the likes of Rāzī, Khūnajī, Khūnajī's associates Urmawī, Ibn Wāṣil and Abharī, and Abharī's student Kātibī. It is not that Rāzī and Khūnajī, for example, were consistently anti-Avicennan (as is often falsely believed). They took their point of departure from Avicenna's writings, and accepted the main innovations of Avicenna in logic, and in these respects they may be contrasted to those who criticized Avicenna from the perspective of a more orthodox Aristotelianism such as Averroes (d. 1198) or ʿAbd al-Laṭīf al-Baghdādī (d. 1231). However, they did not approach the writings of Avicenna with anything like the veneration given to the writings of Aristotle by the earlier Baghdad Aristotelians or Averroes. They delighted in raising and discussing doubts (shukūk) and showed little patience with the practice of doing logic by means of textual exegesis. Ṭūsī was certainly no mere uncritical follower of Avicenna, but on a whole range of issues he did believe that Rāzī and those who followed in his wake had been too quick to abandon Avicenna's own positions. The question of the subject matter of logic was one of these issues. In his Taʿdīl al-miʿyār, a systematic critique of Abharī's Tanzīl al-afkār, Ṭūsī demurred from Abharī's view that the subject matter of logic is “conceptions and assents” (al-taṣawwurāt wa-al-taṣdīqāt). He argued that:

Alternatively, argued Ṭūsī, we might mean by “conception and assent” what these terms are true of insofar as they are conceptions and assents (madlūlātahumā min ḥaythu hiya taṣawwurātun wa-taṣdīqāt), i.e. the two concepts signified by the terms “conception” and “assent” respectively. However, these two concepts cannot be the subject matter of logic either, for a number of reasons:

1. (i) Conception insofar as it is conception (min ḥaythu huwa-taṣawwurun) cannot be anything else, such as universal or particular or essential and accidental; and assent, insofar as it is assent (min ḥaythu huwa taṣdīqun), cannot be anything else such as categorical or hypothetical.

2. (ii) Logic investigates that which leads to conception and assent (bāḥithun ʿan al-mūṣili ilā al-taṣawwuri wa-al-taṣdīqi), and conception insofar as it is conception, and assent insofar as it is assent, cannot lead to conception or assent for this would mean that they lead to themselves.

3. (iii) Conception insofar as it is conception cannot be, for example, definition, and assent insofar as it is assent cannot be syllogism, for obviously “conception” is a distinct quiddity from “definition”, and “assent” is a distinct quiddity from “syllogism”. Definition and syllogism are not even accidents of conception insofar as it is conception, or assent insofar as it is assent, by which Ṭūsī seems to mean that the particular concepts “conception” and “assent” considered as such (i.e. without regard to what they are true of) are simply distinct from the particular concepts “definition” and “syllogism” considered as such. Conception and assent can of course be true of (yaṣduqu ʿalā) the parts of definition or of the parts of syllogism, but not by virtue of just being conception and assent respectively (for conception insofar as it is conception, and assent insofar as it is assent, cannot be anything else) but rather by virtue of being general accidents (aʿrāḍ ʿāmma) that can be predicated of their instances (ʿalā maʿrūḍātihimā).

Ṭūsī's objection may be summarized as follows: By “conception and assent” Abharī can either mean “everything of which these terms are true”, or he can mean the concepts “conception” and “assent” themselves. If the former is meant, then the subject matter of logic becomes all knowledge. If the latter, then this would amount to saying that the subject matter of logic is the two concepts “conception” and “assent” which is plainly implausible. Having thus rejected the view that the subject matter of logic is conception and assent, Ṭūsī concluded his discussion by reiterating the standard Avicennan view that it is second intentions insofar as they lead from knowns to unknowns.Footnote ²⁴

A generation after Ṭūsī, another influential criticism of Khūnajī's views on the subject matter of logic was raised by the prominent logician and Māturīdī theologian Shams al-Dīn al-Samarqandī (d. 1303).Footnote ²⁵ In his summa of logic Qisṭās al-afkār and his own commentary on it, Samarqandī maintained that the correct position is that the subject matter of logic is second intentions.Footnote ²⁶ For logic investigates (yabḥathu ʿan) the universal and particular, essential and accidental, subject and predicate (al-kullī wa-al-juzʾī wa-al-dhātī wa-al-ʿaraḍī wa-al-mawḍūʿi wa-al-maḥmūl), and these are second intentions. It does not investigate second intentions as such, but only insofar as they are a means to reach new knowledge. Samarqandī explained second intentions in terms that are reminiscent of Rāzī and Kātibī: If we simply conceive of quiddities and realities without taking into account any judgment (bi-dūni iʿtibāri ḥukmin ʿalayhā) then these are first intentions. The mind can then make delimitative or predicative judgments about these quiddities, and these judgments are second intentions (fa-kawnuhā kadhālika maʿqūlātun thāniyatun), such as this first intention being “essential” (dhātī) or “accidental” (ʿaraḍī) or a “universal” (kullī) or particular (juzʾī). If we then make further delimitations or predicative judgments of second intentions then these third-order delimitations or judgments are third intentions, and so on. Logic investigates the third-order intrinsic accidents that accrue to second intentions (wa-baḥthu al-manṭiqi waqaʿa fī al-darajati al-thālithati wa-mā baʿduhā), for the logician investigates second intentions being for example a genus or differentia or proprium or general accident or definition or description or proposition or syllogism or analogy (li-annahu yabḥathu ʿan kawni al-maʿqūlāti al-thāniyati jinsan wa-faṣlan wa-khāṣṣatan wa-ʿaraḍan ʿāmman wa-ḥaddan wa-rasman wa-kawnihā qaḍiyyatan wa-qiyāsan wa-tamthīlan wa-ghayri dhālika). Samarqandī's reference to third- and higher-order intentions is interesting. In the above-mentioned article from 1980, Sabra had noted that by the eighteenth century Arabic logicians were recognizing third- and higher-order intentions.Footnote ²⁷ This step had already explicitly been taken by Samarqandī in the late thirteenth century. It led him to deviate somewhat from earlier accounts. His examples of second intentions are “universal”, “particular”, “essential” and “accidental”. “Genus” and “differentia” are on his account not second but third intentions, i.e. intelligibles that come about by mental manipulation of second intentions.

Samarqandī continued by noting that Khūnajī and those who follow him (ṣāḥib al-Kashf wa-qawmun mimman tabiʿahu) have an alternative view of the subject matter of logic. They claim, according to Samarqandī, that the logician may “investigate” the universal and the particular, the essential and the accidental, the subject and the predicate (qad yabḥathu ʿan al-kullī wa-al-juzʾī wa-al-dhātī wa-al-ʿaraḍī wa-al-mawḍūʿi wa-al-maḥmūl), i.e. that the logician may “investigate” second intentions that accrue to first intentions, thus making the latter part of the subject matter of logic. Therefore, the subject matter of logic should be taken to be broader than second-intentions and to include the objects of conceptions and assents in general insofar as these are a means to acquiring new knowledge. On this account, logic would investigate second intentions as well as the third- or higher-order intentions recognized by older logicians (li-yandarija al-maʿqūlatu al-thāniyatu wa-mā dhakartum min al-maʿqūlāti al-thālithati wa-mā baʿduhā taḥta al-manṭiq). In other words, it would not only investigate third-and higher-order intentions that accrue to second intentions (as earlier logicians have held) but also second intentions that accrue to first intentions.

It may be recalled that Samarqandī had started his discussion of this whole issue by claiming that the subject matter of logic is second intentions “since logic investigates the universal and the particular, the essential and the accidental, the subject and the predicate” (limā kāna al-manṭiqu yabḥathu ʿan al-kullī wa-al-juzʾī wa-al-dhātī wa-al-ʿaraḍī wa-al-mawḍūʿi wa-al-maḥmūl). Disquietingly, at this stage in his discussion he noted that Khūnajī and others rejected this position because logic “may investigate the universal and the particular, the essential and the accidental, the subject and the predicate” (qad yabḥathu ʿan al-kullī wa-al-juzʾī wa-al-dhātī wa-al-ʿaraḍī wa-al-mawḍūʿi wa-al-maḥmūl). How can the fact that logic “investigates the universal and the particular, the essential and the accidental, the subject and the predicate” both be a ground for accepting the thesis that the subject matter of logic is second intentions and then be a ground for rejecting that very thesis? I assume that Samarqandī's initial formulation was careless and that his considered position was, as he also wrote, that logic “investigates second intentions being for example a genus or a differentia or a proprium or general accident or definition or description or proposition or syllogism or analogy” (yabḥathu ʿan kawni al-maʿqūlāti al-thāniyati jinsan wa-faṣlan wa-khāṣṣatan wa-ʿaraḍan ʿāmman wa-ḥaddan wa-rasman wa-kawnihā qaḍiyyatan wa-qiyāsan wa-tamthīlan wa-ghayri dhālika). On this interpretation, there would be an important difference between “investigating second intentions being this or that” (al-baḥth ʿan kawn al-maʿqūlāt al-thāniya jinsan wa-faṣlan…) and “investigating second intentions” (al-baḥth ʿan al-maʿqūlāt al-thāniya). This would be in line with the distinction made above, in connection with the discussion of Kātibī, between the subject matter of a discipline and what is “investigated” in that discipline (namely the intrinsic accidents of the subject matter). This would also be in line with the later defense of the Avicennan view by Quṭb al-Dīn al-Rāzī (d. 1365) – which will be discussed below – in which a distinction is made between logic investigating the states (aḥwāl) of second intentions, which is held to be the correct position, and investigating second intentions themselves, which is held to be the basis of Khūnajī's alternative position.

Samarqandī went on to claim that this novel position of Khūnajī and his followers is wrong. Logic investigates second intentions (al-baḥth ʿan al-maʿqūlāti al-thāniya) as part of the principles (mabādiʾ) of the science, not as part of the issues (masāʾil) of that science. The distinction between the “issues” and “principles” of a discipline was thoroughly conventional. It is to be found, for example, in Avicenna's Ishārāt. The principles of a science are there explained as being either the premises (muqaddimāt) that are treated as axiomatic in that science (immā wājibatu al-qabūli aw musallamatun), or the definitions of (i) the subject matter of the discipline (mawḍūʿ al-ṣināʿa) or (ii) its parts or instances (ajzāʾ wa-juzʾiyyāt) or (iii) its intrinsic accidents (aʿrāḍ dhātiyya). Ṭūsī in his commentary gave the following examples of the three kinds of basic definition in the discipline of physics: (i) “The body is … def”; (ii) “The simple body is … def”; and (iii) “Movement is … def”.Footnote ²⁸ Samarqandī's claim seems to be that the cases of second intentions themselves being “investigated” in logic fall into one of these kinds. They would accordingly be either propositions accepted in logic in which second intentions are in predicate position, or basic definitions of the second intentions that form the subject matter of the discipline (or the parts or intrinsic accidents of these second intentions). An example of the first kind would presumably be statements like “Animal is a universal” and “Zayd is a particular”. In such cases Samarqandī surely has a point, for such propositions are hardly more than conventionally accepted examples in the discipline of logic, as opposed to issues that would be the subject of ratiocination or debate by logicians. Indeed, to have such propositions belong to the “principles” of logic is arguably to concede too much and, as will be seen below, Quṭb al-Dīn al-Rāzī, who was aware of and echoed Samarqandī's discussion, maintained that such propositions do not have anything to do with logic. The second kind of “principles”, i.e. the basic definitions of the various second intentions, arguably poses a more serious problem for the Avicennan claim that the subject matter of logic is second intentions. The sources do not explicitly spell out exactly why such definitions are a problem, but the challenge is presumably this: The predicate in a definition of a second intention must also be a second intention, and as such must “accrue to” first intentions. If, in other words, a second intention “accrues” to first intentions, then any definition of a second intention must also accrue to first intentions. For example, “universal” is uncontroversially a second intention. As such, it must – equally uncontroversially – accrue to (yalḥaqu bi-) or be accidentally true of (yaʿriḍu li-) first intentions such as “human” or “animal”. However, any definition of “universal”, since it merely explicates the definiendum, would also accrue to, or be accidentally true of, these very same first intentions. If we, for example, suppose the definition of “universal” to be “that whose conception does not prevent it being true of many,” then this definition becomes true of the first-order instances of “universal” such as “human” and “animal”. To this extent it would be true to say that in such definitions it is second intentions themselves that are “investigated”, and not second intentions being this or that third intention. It is possibly this kind of definition that Kātibī had had in mind when he argued – as mentioned above – that the subject matter of logic includes first intentions since the logician “investigates the concept (yabḥathu ʿan mafhūm) of “essential”, “accidental”, “species”, “genus”, “differentia”, “subject” and “predicate” and other things that accrue to single notions that we intellect”.

It has been seen that two counter-arguments to Khūnajī's thesis on the subject matter of logic were made in the second half of the thirteenth century, by Ṭūsī and Samarqandī respectively. Ṭūsī's rejection was based on drawing a distinction between understanding “conception” and “assent” to indicate everything falling under those terms or alternatively to indicate the two concepts themselves. He then argued that neither reading produced a defensible position regarding the subject matter of logic. Samarqandī rejected Khūnajī's position by drawing a distinction between the “principles” and “issues” of logic and then claiming that the adduced examples of logic supposedly “investigating” second intentions themselves (as opposed to their intrinsic accidents) all belonged to the principles and not the issues of the discipline. Both arguments against Khūnajī would be echoed by subsequent discussions, as will become clear in the following section.

III

Quṭb al-Dīn al-Rāzī al-Taḥtānī (d. 1365), who was active in Ilkhanid Persia and later moved to Damascus, was arguably the most prominent and influential Arabic logician from the fourteenth century. He wrote an esteemed commentary on Kātibī's al-Risāla al-Shamsiyya that was still being studied in the nineteenth century in Egypt, Ottoman Turkey, Iran and India (as attested by the fact that it was repeatedly printed or lithographed in these places in the nineteenth and twentieth centuries in editions obviously intended for madrasa students).Footnote ²⁹ His most extensive treatment of logic, however, is his commentary on Urmawī's Maṭāliʿ al-anwār, approximately twice as long as his commentary on the Shamsiyya.Footnote ³⁰ Nicholas Rescher, in his The Development of Arabic Logic (1964), suggested that Quṭb al-Dīn was a central figure in the reconciliation and synthesis of the “western” anti-Avicennan and the “eastern” pro-Avicennan trends in Arabic logic.Footnote ³¹ However, this suggestion needs to be qualified significantly. Recent research by Tony Street has undermined the assumption that there were opposed “eastern” and “western” traditions of Arabic logic in the twelfth and thirteenth centuries.Footnote ³² In any case, Quṭb al-Dīn was far from being a reconciler. It is true that in his widely-studied commentary on Kātibī's Shamsiyya he rarely expressed serious disagreement with Kātibī, and in this sense he may have contributed to the prominence of the logic of what Tony Street has called “revisionist Avicennism”.Footnote ³³ However, the commentary on the Maṭāliʿ reveals Quṭb al-Dīn to have been an avid supporter of Avicenna who regularly (and sometimes angrily) rejected the departures from Avicenna suggested by Khūnajī and his followers (including Urmawī and Kātibī). The topic of the subject matter of logic is a case in point. Urmawī had written that the subject matter of logic is “conceptions and assents” (al-taṣawwurāt wa-al-taṣdīqāt), for logic investigates (yabḥathu ʿan) their intrinsic accidents (ʿawāriḍihā al-lāḥiqati li-mā hiya hiya) which is that they lead to further conceptions or assents.Footnote ³⁴ Quṭb al-Dīn, on the other hand, started his treatment of the issue by stating that the position of “the people of verification” (ahl al-taḥqīq) is that the subject matter of logic is second intentions. He explicated second intentions as follows: Things (al-ashyāʾ) can have accidents (ʿawāriḍ) that accrue to them in the extra-mental world (such as color or length or movement), and they can have accidents that accrue to them insofar as they are represented in the mind (min ḥaythu hiya mutamaththilatun fī al-ʿaql), and specifically accidents that have no correlate in the extra-mental world, such as “universal” or “particular”. The latter accidents are called “second intentions”. These are the subject matter of logic since logic investigates the states (aḥwāl) of second intentions like essential and accidental, species, genus and differentia, definition and description, categorical and hypothetical proposition, or syllogism and induction and analogy, all insofar as they lead to new knowledge. Logic thus investigates the third-intentions that accrue to second intentions (or the fourth intentions that accrue to third intentions, etc.). Quṭb al-Dīn, who was familiar with Samarqandī's Qisṭās, thus accepted the view that there are third- and higher-order intentions and that it is these that the logician “investigates” (yabḥathu ʿan), whereas second intentions are the subject matter (mawḍūʿ) to which these higher-order intentions accrue.

Quṭb al-Dīn went on to present the alternative view of Khūnajī and his followers. He wrote:

Quṭb al-Dīn rejected this new-fangled position. His first objection is obviously related to that of Samarqandī: The later logicians support their position by pointing out that in logic second intentions themselves are “investigated” and not merely the third- and higher-order intentions that accrue to second intentions. Quṭb al-Dīn replied that if what is meant by “investigating” certain second intentions is that the logician “clarifies their conceptions” (yubayyinu taṣawwurātihā) then this is not part of the issues (masāʾil) of logic but of the principles (mabādiʾ) of logic. If what is meant is the determination of the extension of second intentions (in arādū al-taṣdīqa bihā li-al-ashyāʾi), i.e. determine that X is a genus or Y is a differentia, then it is clear that this is not part of logic at all. In other words, logical propositions in which second intentions are themselves “investigated” are of two types: (i) definitions of basic second intentions such as “universal” or “essential”; or (ii) propositions in which the second intentions are predicated of first intentions, as in “Animal is a universal” or “Zayd is a particular”. The first case is handled by resorting, like Samarqandī, to the distinction between the “issues” and “principles” of a discipline. The second case is dismissed by Quṭb al-Dīn as not being part of logic at all.

Quṭb al-Dīn also reiterated the objection that had been raised by Ṭūsī: If by “conceptions” and “assents” is meant what these terms are true of (mā ṣadaqatā ʿalayhi min al-afrād) then the subject matter of logic would be all knowledge. Alternatively, by the terms “conception” and “assent” might be meant the two concepts “conception” and “assent” (mafhūmahumā). However, it seems clear that logic does not investigate the intrinsic accidents of the concepts “conception” and “assent”. For it is not by virtue of being a conception (maʿlūm taṣawwurī) that a definition leads to knowledge of the essence of the definiendum, but rather by virtue of being something more specific, namely a conception of the genus and differentia. Similarly, it is not by virtue of being an assent (maʿlūm taṣdīqī) that something converts to a universal-negative proposition, but by virtue of something more specific, namely being a universal-negative proposition.Footnote ³⁶

Quṭb al-Dīn went on to raise an anticipated response: Could this objection not equally be raised against the view that the subject matter of logic is second intentions? In other words, could one not object that by “second intentions” we can either mean (i) the concept itself, or (ii) that of which the concept is true? In the former case, it is clear that logic does not investigate the essential attributes of the concept “second intentions”. In the latter case, the subject matter of logic would include all second intentions. Quṭb al-Dīn's answer to this was to select the second option and say: We do not concede that it follows that the subject matter of logic is all second intentions.Footnote ³⁷ Rather, the subject matter is only the subset of second intentions that is (i) relevant to the question of leading to new knowledge and (ii) relevant in such a way that what is true of them “transmits” (yasrī) to first intentions. For example, one may consider the following example:

The subject of the proposition is a second intention. However, what makes it a proposition properly belonging to logic is not only this fact but the additional facts that (i) the proposition is relevant (albeit indirectly) to the task of acquiring new knowledge, and (ii) the predication “transmits” to the first intentions that fall under the second intentions, so that we for example know that “animal” must be one of the five mentioned divisions. One might compare this with a proposition such as:

This too has a second intention as subject. The definition may even be said to lead to the acquisition of the concept “Noun”. Why then is it not a proposition that belongs to logic? The answer is that the definition can lead to acquiring new knowledge (in this case a new concept) but not by stating a general principle (qānūn) that is relevant to this task. On the basis of the definition, we know that “Animal” is a noun, but not by way of giving a general rule for the acquisition of new knowledge.

The prominent and influential Timurid scholar al-Sayyid al-Sharīf al-Jurjānī (d. 1413), in his esteemed gloss on Quṭb al-Dīn's commentary, anticipated that someone might object that it is equally possible to use this strategy to defend the view of Khūnajī and Urmawī. Thus, one would say that what is meant by “objects of conception” and “objects of assent” are what these terms are true of, and then – to disarm the argument of Ṭūsī that this would make the subject matter of logic equivalent to all knowledge – specify that the province of logic is only objects of conception and assent that are appropriate in the proper manner to the obtaining of new knowledge.Footnote ³⁸ In fact, the older contemporary and rival of Jurjānī, Saʿd al-Dīn al-Taftāzānī (d. 1390) had replied along these lines to the objection of Ṭūsī. Taftāzānī, in his commentary on Kātibī's Shamsiyya, wrote that it had been objected to Kātibī's view about the subject matter of logic that if we mean by “objects of conception” and the “objects of assent” the two concepts themselves (mafhūmahumā) then it is clear that the predicates that are relevant to the acquisition of new knowledge are not essential attributes of these two concepts. If, on the other hand, we mean what the terms “objects of conception” and “objects of assent” are true of (mā ṣadaqat hiya ʿalayhi), then all objects of conception and assent, i.e. the sum of human knowledge, would be the subject matter of logic. He replied to the objection thus:

Taftāzānī's answer is, in other words, to say that what is meant by “objects of conception” and “objects of assent” are what these terms are true of, but to make it clear that these are the subject matter of logic only insofar as they lead to conceptions and assents in general, i.e. to indefinite conceptions and assents (ilā taṣawwurin mā aw taṣdīqin mā) as opposed to this specific conception or that specific assent (lā ilā taṣawwurin aw taṣdīqin makhṣūṣ). The conceptions and assents of the various sciences do lead to new conceptions and assents, yet this has nothing to do with the particular contents of these conceptions and assents but with their being definitions or arguments in general. It is these general features of the definitions and arguments used in the various sciences that are the subject matter of logic.

Jurjānī may have had Taftāzānī's argument in mind when he wrote his glosses on Quṭb al-Dīn's commentary on Maṭāliʿ al-anwār. He responded that an attempt to offer further specifications of the “objects of conception and assent” that constitute the subject matter of logic would in effect be to concede that the departure from the Avicennan position was unwarranted:

Jurjānī's rejection might appear dogmatic. How could he in the present context help himself to the premise that logic only investigates the states of second intentions that are applicable to first intentions? It would seem that he accepted Samarqandī's independent argument (endorsed earlier by Quṭb al-Dīn) that apparent counter-examples adduced by Khūnajī and his followers are part of the “principles” and not the “issues” of logic. Taking this as settled, one might object, as Jurjānī did, that further specifications of the kind of objects of conception and assent that are relevant for logic would simply lead back to second intentions. However, the merits of Samarqandī's argument were not universally accepted by later logicians. The debate on the subject matter of logic seems to have continued unabated until the modern period, and it is clear from a sampling of this later literature that Samarqandī's point did not remain unchallenged.Footnote ⁴¹ However, pursuing the debate beyond the fourteenth century is beyond the scope of this article.

IV

From the early thirteenth century, the Avicennan view that the subject matter of logic is second intentions ceased to be the default position of Arabic logicians. Instead, the standard madrasa handbooks on logic typically reiterated Khūnajī's view concerning the subject matter of the discipline. On Avicenna's view, logic has a distinct subject matter, namely the properties that accrue to quiddities by virtue of being in the mind. Later supporters of Avicenna sometimes qualified the view by specifying that it is only the second intentions that are relevant in a general way to the acquisition of new knowledge that constitute the province of logic. On Khūnajī's alternative view, the subject matter of logic is more generally the objects of conception and assent. Defenders of Avicenna were quick to object that this would make the subject matter of logic coterminous with all human knowledge and all sciences. Supporters of Khūnajī's view seem to have accepted this. In other words, they accepted that logic has no peculiar subject matter but that its province is all human ratiocination from the known to unknown insofar as it exhibits certain general, content-neutral features. In support of Khūnajī's view they pressed the point that in logic second intentions themselves, and not merely their intrinsic accidents, may be investigated. Defenders of the Avicennan view replied that this investigation was only part of the “principles” and not the “issues” of logic. There is no indication that this debate was resolved or died out in later centuries. The issue was clearly still being discussed in the nineteenth and twentieth centuries.

It may be natural to ask if the intense and often abstruse debates concerning the subject matter of logic that went on for half a millennium were related to more general philosophical or theological differences. It may be tempting, for example, to see the theologians' aversion to “mental existence” lurking in the background of these debates. However, matters are clearly not that simple. There is no evidence that Khūnajī, Abharī, Urmawī, Kātibī and Ibn Wāṣil self-identified as “theologians” (mutakallimūn), and in any case they did not deny that second intentions form part of the subject matter of logic. Furthermore, the theologians' opposition to “mental existence” seems in fact to have waned in later centuries. Fakhr al-Dīn al-Rāzī accepted “mental existence” and later standard handbooks on theology such as al-Mawāqif by Ījī (d. 1355) and al-Maqāṣid by Taftāzānī did so as well.Footnote ⁴² It cannot be ruled out, of course, that the debate concerning the subject matter of logic can be shown by future research to have been a reflection of some grander difference in philosophical or theological commitments. However, it is also possible that the debate was self-contained and pursued for its own sake. Surely, logicians and philosophers of that age, as well as ours, could discuss intricate issues at a high level of abstraction, in the process drawing further distinctions or refining arguments to meet real or anticipated objections, without there necessarily being a point or motive that lies outside the explicit aims of the discussion itself.Footnote ⁴³