Khaled El-Rouayheb's latest monograph presents an account of the history of Arabic logic over a period of 1,000 years, from 900 to 1900, and over a vast geographical region, from the Balkans to India and from Egypt to Morocco. His examination is no small accomplishment because it requires familiarity with unpublished manuscripts in numerous libraries in various countries. His purpose is to dislodge the thesis of Nicholas Rescher, who argues that the Arabic logical tradition was in “stagnation and decline” after 1300 (pp. 2–3). El-Rouayheb claims, and rightly so, that the thesis has become “enshrined in the only extensive modern survey of the history of Arabic logic, Nicholas Rescher's The Development of Arabic Logic” (p. 2). Rescher says specifically: “By around 1500 this process of the disintegration of logical study to a purely text-oriented comment-mongering was complete. Logic was dead in Islam as a branch of inquiry” (The Development of Arabic Logic [Cleveland, Ohio: John T. Zubal, 2009—a reprint of the book first published in 1964 by the University of Pittsburgh Press], 81). Rescher argues that the best logicians of the Arabic tradition were the Baghdadian Aristotelians, including al-Farabi, and the Andalusians, including Ibn Bajja and Ibn Rushd, and that their influence lasted about 100 years after the death of Ibn Rushd (d. 1198), at which point the logical tradition deteriorated. El-Rouayheb's primary evidence in criticism of Rescher's thesis is that there were developments in logic after 1300 amongst several key logicians in the Middle East and after 1600 amongst the Ottoman logicians; in one chapter El-Rouayheb focuses on the logical handbook of Ismaʿil Gelenbevi (d. 1791) as a pre-eminent Ottoman example of such improvements.
El-Rouayheb's refutation focuses on one type of syllogism, the “relational syllogism,” in order to show that its development in the 18th century by Ottoman logicians led to the abandonment of the classical Aristotelian definition of the syllogism. El-Rouayheb writes: “As I will show in this study, logicians within the Arabic tradition long struggled with the anomaly of arguments involving relations. In the 18th century, several prominent Ottoman logicians decisively abandoned the classic Aristotelian assumption that a valid categorical syllogism must consist of two premises with one term in common and began to investigate what they called ‘unfamiliar syllogisms’ with relational premises” (p. 2). According to El-Rouayheb, this shift away from Aristotle confirms a contention of Augustus De Morgan (d. 1871) that Aristotelian logic could not validate such simple inferences as “every man is an animal” to “every head of a man is the head of an animal” (p. 1) because there are more than three terms in the two premises. Thus, El-Rouayheb's intention is to show that Aristotelian logic is inadequate, although his proof is not De Morgan's arguments but several Arabic logicians of the 13th to 17th centuries and especially certain Ottoman logicians of the 18th and 19th centuries whose writings predate De Morgan's work. That this Arabic logical development emerged in the time and place that it did challenges two a priori assumptions embedded in Rescher's thesis: “that Islamic civilization declined in general after the thirteenth century, [and] that commentaries (shurūḥ) and glosses (ḥawāshī) can only involve pedantic explication of received opinions” (p. 3; see also pp. 39–40 and 255–59). El-Rouayheb criticizes both.
El-Rouayheb distinguishes between an Aristotelian and a relational syllogism in the following manner. The Aristotelian definition of a syllogism, as defined by al-Farabi, Ibn Sina, and Ibn Rushd, amongst others, is “a statement in which things more than one are posited; if these are composed together then something else is implied by them necessarily, by themselves and not by accident” (El-Rouayheb's translation of al-Farabi's definition, p. 14). One of El-Rouayheb's examples is “every human is a mammal” and “every mammal is an animal,” and therefore, “every human is an animal” (p. 1). Relational syllogisms not only possess the three terms of a perfect syllogism, the middle term and the two extremes, but in addition to these terms they express relations, such as equality or possession. The result is that in the syllogism of relation the syllogism has more than three terms. The first example that El-Rouayheb cites from De Morgan is the one we have already mentioned: “every man is an animal” and “every animal has a head,” and therefore “every head of a man is the head of an animal” (p.1). The confusion is created by the genitive relations introduced between the subject and predicate of the conclusion. It is argued that the standard Aristotelian definition of a syllogism is unable “to handle such arguments” (p. 1).
El-Rouayheb claims that relational syllogisms can be productive of knowledge, but they do not fit this Aristotelian definition of syllogism. Relational syllogisms were known in Latin Europe, and they were called “oblique syllogisms” to indicate that the syllogism has an oblique term, that is, a term in a grammatical case other than nominative. In Aristotelian syllogistic, the terms would have to be reduced or the premises would not be productive of knowledge.
The scholarly literature pertaining to El-Rouayheb's research raises several considerations for further study. First, Paul Thom has remarked, “It was Aristotle himself who saw the way around this difficulty [of relational syllogisms]. For the purpose of reckoning syllogistic figure and mood we may ignore the relational expression in such a syllogism and treat it as a sort of modified Celarent with terms [in the example] ‘things that laugh,’ ‘man’ and ‘ass.’ Aristotle stated this brilliant intuition in [Prior Analytics] A 36, but did not work out the details: that was done by the medievals” (“Termini Obliqui and the Logic of Relations,” Archiv für Geschichte der Philosophie, 59 [1977], 144). In short, Aristotle says that the relational syllogism can be reformulated as a typical syllogism with only three terms without any loss of content. Thom's observation indicates the resourcefulness of Aristotelian syllogistic in providing an answer to El-Rouayheb's criticism.
Second, there are several studies on logic of the classical Arabic period showing that al-Farabi's and Ibn Rushd's accounts of Aristotle may not simply be, as El-Rouayheb suggests, “standard Aristotelian fare” (p. 16; see also p. 34). Rescher himself seems puzzled by his own observations that al-Farabi's Short Commentary on the Prior Analytics “is devoted solely and exclusively to: (i) the theory of conditional syllogisms with respect to which the Anal. Pr. is virtually silent, apart from the ‘promissory note’ of 50a40; and (ii) the logic of arguments which can be related to the syllogism in various ways, dealt with in the final portion of book II of Anal. Pr. To this second item, some 6 percent of Anal. Pr., some 60 percent of al-Fārābī's Short Commentary is devoted!” (Al-Fārābī's Short Commentary on Aristotle's Prior Analytics, trans. and notes by Nicholas Rescher [Pittsburgh, Pa.: Pittsburgh University Press, 1963], 39). Why this highly selective and qualified use of Aristotle by al-Farabi? Similar procedures have been shown to exist in Ibn Rushd's commentaries on Aristotle. Charles E. Butterworth's readings of both Aristotle's text and Ibn Rushd's “commentary” reveal that Ibn Rushd is selective and he may even improve on Aristotle (“Comment Averroès lit les Topiques d'Aristote,” Penser avec Aristote, Etudes réunies sous la direction de M. A. Sinaceur. Ouvrage publié avec le concours de l'UNESCO [Toulouse: Edition Erès, 1991], 701–24; see esp. 722–23). Given these presentations of Aristotle, it remains uncertain that al-Farabi and Ibn Rushd are “standard Aristotelian fare.” The logicians of the Classical period need further clarification.
El-Rouayheb's book is essential reading for students of philosophy, and his comprehensive reading of unpublished Arabic manuscripts is impressive. His argument corrects insufficiently established judgments of Rescher, and his book establishes a dialectical study of Rescher's criticisms of Arabic logic after 1300. We look forward to further conversation on this subject. In this 1,000-year period of largely uncharted intellectual history, El-Rouayheb's book is one to which we will often return.
