Introduction

Abū Yūsuf Yaʿqūb b. Isḥāq al-Kindī (d. ca. 259/873) ranks among the greatest philosophers in the Arabo-Islamic tradition, along with his successors al-Fārābī (d. 338/950), Ibn Sīnā (d. 429/1037), and Ibn Rushd (d. 595/1198).¹ Because only a small fraction of al-Kindī’s books has been preserved, modern investigators have a woefully incomplete picture of his contributions to philosophy and other fields. The bibliography of al-Kindī in the Fihrist of Ibn al-Nadīm (d. 380/990), which lists over two hundred individual titles, thus provides one of the main avenues of access to his works and thought. The first two titles in the section of al-Kindī’s bibliography in the Fihrist devoted to his philosophical works, al-Falsafa al-ūlā … (First Philosophy …) and al-Falsafa al-dākhila … (Internal Philosophy …), pose problems of interpretation on which modern investigators have commented occasionally but not resolved. The following remarks examine these titles in detail, considering them in light of the extant portion of al-Falsafa al-ūlā, references and quotations in later works, and al-Kindī’s On the Quantity of Aristotle’s Books. I argue the following: 1) that al-Kindī gave the two titles parallel structures and intended them to form a pair; 2) that al-Kindī envisaged the relationship between the two works as that of foundation to detailed exposition; 3) that the title al-dākhila is related to the Greek term “esoteric” and invokes a division of Aristotle’s works into esoteric and exoteric that was common in Late Antique understandings of Aristotle’s corpus; 4) that the second half of the title al-Falsafa al-dākhila…, which lists several categories of content, was corrupted already in the earliest extant manuscript of Ibn al-Nadīm’s Fihrist; and 5) that al-Falsafa al-dākhila covered the main categories of Aristotle’s philosophical works as conceived by al-Kindī and presented in his work On the Quantity of Aristotle’s Books, something which facilitates reconstruction of the original title, at least tentatively. These arguments demonstrate 6) that al-Kindī intended the two works to provide, in combination, a substantial survey of Aristotle’s philosophical works.

The Bibliography of Al-Kindī in the Fihrist of Ibn Al-Nadīm (d. 380/990)

The Fihrist’s extensive bibliography of al-Kindī contains 249 titles in all, divided into sixteen topical sections.² Moreover, in the earliest manuscript of the Fihrist, Şehid Ali Paşa MS 1934, the copyist – and presumably Ibn al-Nadīm in the original – left nearly half a page blank at the end of the bibliography, suggesting that, in Ibn al-Nadīm’s estimation, it was still incomplete and lacked as many as several dozen additional titles.³ Later bibliographies of al-Kindī by al-Qifṭī (d. 646/1248) and Ibn Abī Uṣaybiʿa (d. 668/1270) repeat the list from Ibn al-Nadīm’s entry verbatim, with some augmentations, al-Qifṭī adding four titles and Ibn Abī Uṣaybiʿa adding fifty-three titles.⁴ In 1857, Gustav Flügel published the Arabic text of Ibn al-Nadīm’s bibliography of al-Kindī, along with a German translation, in his work al-Kindî genannt “der Philosoph der Araber”: Ein Vorbild seiner Zeit und seines Volkes (Al-Kindī, Called “The Philosopher of the Arabs”: A Paragon of His Time and His People).⁵ An English translation of Ibn al-Nadīm’s bibliography of al-Kindī is available in Bayard Dodge’s translation of the Fihrist, and Peter Adamson and Peter E. Pormann have translated the bibliography into English as well in the introduction to their collection of al-Kindī’s philosophical works.⁶ The bibliography includes twenty-one titles on philosophical topics, as well as ten on logic and twelve on politics and ethics. Comparison with al-Kindī’s extant works suggests that many of the titles listed in the bibliography were probably short treatises and reveals that many were dedicated to his student Aḥmad b. al-Muʿtaṣim, the son of the Abbasid Caliph al-Muʿtaṣim (r. 218–227/833–842).⁷

Ibn al-Nadīm must have copied al-Kindī’s bibliography from a pre-existing individual bibliography devoted to al-Kindī’s works, for its length and structure show the characteristics of a carefully authored text rather than a mere working list. It is possible that al-Kindī himself originally compiled the bibliography. Perhaps the most famous auto-bibliography, that of Galen, was known by al-Kindī’s contemporaries and so could have served as a model.⁸ In another case, Ibn al-Nadīm copied the auto-bibliography of the famous physician Abū Bakr al-Rāzī (d. 313/925).⁹ It is also possible that the bibliography was compiled by one of al-Kindī’s students such as Aḥmad b. al-Ṭayyib al-Sarakhsī (d. 286/899).¹⁰ In any case, Ibn al-Nadīm probably gained access to al-Kindī’s bibliography through Yaḥyā b. ʿAdī (d. 363/974), Ibn al-Nadīm’s friend and fellow book-seller, who was one of his main consultants on the history of philosophy. Thus, for example, Yaḥyā b. ʿAdī was Ibn al-Nadīm’s source for the bibliography of Aristotle.¹¹

The two titles addressed here focused on the works of Aristotle. This is not surprising, given the tremendous role played by Aristotle’s works in the curricula of many centers of learning in the Near East during Late Antiquity and the central role they continued to play in the Islamic period.¹² Al-Kindī explicated and commented on Aristotle’s oeuvre in many of his works, and the fact that he was engaged in paraphrasing, adapting, presenting, and commenting on Aristotle’s corpus throws light on many titles in his bibliography. Fortunately, al-Kindī provides a relatively detailed overview of his understanding of Aristotle’s corpus in his extant treatise Risāla fī kammiyyat kutub Arisṭūṭālīs wa-mā yuḥtāju ilayhi fī taḥṣīl al-falsafa (Treatise on the Quantity of Aristotle’s Books and on What is Required for the Acquisition of Philosophy). This text was first published in 1940 by Michelangelo Guidi and Richard Walzer, along with an Italian translation.¹³ Muḥammad ʿAbd al-Hādī Abū Rīda published a second edition of the work in 1950, prefaced by a short discussion.¹⁴ Nicholas Rescher analyzed al-Kindī’s presentation of Aristotle’s Organon in the treatise.¹⁵ Jean Jolivet published a study of the work in 2004,¹⁶ and Jules Janssens wrote on al-Kindī’s use of Qurʾānic quotations therein in 2007.¹⁷ Peter Adamson made extensive use of the treatise in his monograph on al-Kindī’s thought, and he and Peter Pormann included an English translation in their collection of translations of al-Kindī’s philosophical works.¹⁸ As Jolivet points out, this work is probably identical with the title that appears in the Fihrist of Ibn al-Nadīm as Kitāb Tartīb kutub Arisṭāṭālīs (Book on the Order of Aristotle’s Books). In addition to the fact that the general topic seems to overlap, the fact that the text of the treatise makes frequent use of cognates of tartīb such as marātib, “degrees, stages,” justifies the identification.¹⁹ This text, which grants modern investigators access to al-Kindī’s understanding of the Aristotelian corpus, facilitates interpretation of the titles presented by Ibn al-Nadīm.

Paired Philosophical Works

كتاب الفلسفة الأولى في ما دون الطبيعيّات والتوحيد

كتاب الفلسفة الداخلة والمسائل المنطقيّة والمعتاصة وما فوق الطبيعيّات

The first two titles in the bibliography, which fall under the first section rubric kutub falsafiyya, “Philosophical Books,” are evidently intended to form a complementary pair. The titles both begin with Kitāb al-Falsafa, “The Book of Philosophy.” In both, “Philosophy” is modified by a following adjective. They both include the term al-ṭabīʿiyyāt, “Physics” or “Physical Things.” The first title refers to mā dūna l-ṭabīʿiyyāt, “What is below Physics,” while the second title refers to mā fawqa l-ṭabīʿiyyāt, “What is above Physics.” These elements of parallelism are accentuated by the fact that the two titles occur together, and also by the fact that they appear at the very beginning of the list of philosophical titles, itself the first section in the bibliography. It seems clear that these are not simply two separate works among al-Kindī’s philosophical works; they are two related works that form an intentional pair.

However, the two titles raise several questions, given that it is difficult to explain exactly how they relate to each other. As Ibn al-Nadīm’s text presents it, the first title is Kitāb al-Falsafa al-ūlā fī mā dūna l-ṭabīʿiyyāt wa-l-tawḥid, which one might translate literally as follows: “The Book of First Philosophy, on What Is below Physics and (on) Oneness.” The second title is Kitāb al-Falsafa al-dākhila wa-l-masāʾil al-manṭiqiyya wa-l-muʿtāṣa wa-mā fawqa l-ṭabīʿiyyāt,²⁰ which one might render literally as “The Book of Internal Philosophy, and Logical and Difficult Questions, and What Is above Physics.” Scholars to date have translated these titles in various ways. Flügel gives the translation of the first as Das Buch der ersten Philosophie über das was nicht über die (Vorkenntnisse der) physischen Dinge und die Lehre von der Einheit (Gottes) hinausgeht (The Book of First Philosophy, on That Which Does Not Go beyond the [Preliminary Knowledge of] Physical Things and the Teaching of the Unity [of God]), and the second as Das Buch der tiefer eingehenden Philosophie, der logischen und verwickelten Lehrsätze und dessen was über die physischen Dinge (als gewöhnliche Erscheinungen) hinausgeht (The Book of More Profound Philosophy, Logical and Complex Propositions, and That Which Goes beyond Physical Things [as Ordinary Phenomena]).²¹ George N. Atiyeh translates the first as “On First Philosophy” and the second as “On Introduced Philosophy and Logical and Abstruse Problems and Metaphysics.”²² Dodge translates the first as “Elementary (First) Philosophy, Introductory to Natural Phenomena and Unity,” and the second as “Intrinsic (Inner) Philosophy, Logical and Difficult Questions, and Metaphysics (the Supernatural).”²³ Adamson and Pormann translate the first title as “On First Philosophy, Regarding What Is below Natural Things, and Oneness (tawḥīd),” and the second as “On Internal Philosophy, Logical and Recondite Questions, and What Is above Natural Things.”²⁴ Jean Jolivet and Roshdi Rashed translate the first title as Livre de la philosophie première, sur ce qui est au-delà de la nature et sur le tawḥīd (Book of First Philosophy, on That Which is beyond Nature, and on tawḥīd).²⁵ In general, these translations do not clarify how the two titles complement each other, in some cases suggesting that they overlap instead. What has proved especially puzzling has been the relationship between the two phrases mā dūna l-ṭabīʿiyyāt “what is below physics, or natural things,” and mā fawqa l-ṭabīʿiyyāt “what is above physics, or natural things.”

Al-Kindī’s Work on Metaphysics, al-Falsafa al-ūlā

The first work, al-Falsafa al-ūlā, is well known and partially extant. Around 1930, Hellmut Ritter discovered a copy of a substantial section of the work in a single manuscript in Istanbul, MS Aya Sofya 4832, fols. 34r–53r.²⁶ Three editions of the text have been published: the first by Aḥmad Fuʾād al-Ahwānī in 1948, the second by Muḥammad ʿAbd al-Hādī Abū Rīda in 1950, and the third by Jean Jolivet and Roshdi Rashed in 1998.²⁷ Alfred L. Ivry translated the work into English in 1974, Rafael Ramón Guerrero and Emilio Tornero Poveda published a Spanish translation in 1986,²⁸ and Jean Jolivet and Roshdi Rashed published a French translation in 1998.²⁹ Ivry and other scholars of philosophy in the Arabic tradition have shown that al-Falsafa al-ūlā is to a large extent a presentation of material related to Aristotle’s Metaphysics.³⁰

I argue that the title Kitāb al-Falsafa al-ūlā fī mā dūna l-ṭabīʿiyyāt wa-l-tawḥid (The Book of First Philosophy, on What Is below Physics and [on] Oneness) is correct as presented in Ibn al-Nadīm’s bibliography of al-Kindī. It is clear from the extant portion of al-Kindī’s First Philosophy that “first philosophy” designates metaphysics. He derived this terminology from Aristotle himself, who uses “first philosophy” to refer to metaphysics in his Metaphysics, and al-Kindī clearly modeled his work after that of Aristotle.³¹ It is well known from studies to date on al-Kindī that, in this work, he assimilated Aristotle’s metaphysics to Islamic theology. The reference to al-tawḥīd, “the Oneness (of God),” fits both with this trend in his thought in general and with what is evident from the extant portion of the work.³² It is corroborated in general by a shorter treatise of al-Kindī, titled Risāla fī waḥdāniyyat allāh wa-tanāhī jirm al-ʿālam (Treatise on the Oneness of God and the Finitude of the Body of the Earth).³³

The title given in the manuscript is Kitāb al-Kindī ilā l-Muʿtaṣim bi-llāh fī l-Falsafa al-ūlā (Epistle to al-Muʿtaṣim on First Philosophy).³⁴ Earlier investigators have raised questions concerning the relationship of this extant work to Ibn al-Nadīm’s title Kitāb al-Falsafa al-ūlā fī mā dūna l-ṭabīʿiyyāt wa-l-tawḥid: the epistle to al- Muʿtaṣim may be identical with the book on first philosophy, but it may also be a distinct, shorter work devoted to the same topic.³⁵ My own assessment is that they are in fact the same, the main evidence being that Ibn al-Nadīm’s bibliography of al-Kindī, despite its length and detail, contains no mention of another work also devoted to an exposition of first philosophy. The title “Epistle to al-Muʿtaṣim” may not have been the original title of the extant work but rather came to be applied to it later, drawing on indications in the text itself that the work had been dedicated to the Abbasid caliph al-Muʿtaṣim.³⁶ It must have been a comprehensive work that adapted and presented Aristotle’s Metaphysics in Arabic for an Arabic readership, and it is unlikely that al-Kindī wrote two such comprehensive works covering the same ground.

The original text of First Philosophy was probably at least twice the length of the extant text.³⁷ The surviving manuscript of al-Kindī’s al-Falsafa al-ūlā contains four chapters, each termed fann, which made up the entire first section (juzʾ) of the work. The references to chapters appear in the text of Abū Rīda’s edition as follows:

al-fann al-thānī wa-huwa l-juzʾ al-awwal fī l-falsafa al-ūlā, “The second chapter, and it is the first section in First Philosophy”³⁸

fa-l-nukmil al-āna hādhā l-fann al-thānī, “So let us now complete this second chapter”³⁹

al-fann al-thālith min al-juzʾ al-awwal, “The third chapter of the first section”⁴⁰

al-fann al-rābiʿ wa-huwa l-juzʾ al-awwal, “The fourth chapter, and it is the first section”⁴¹

tamma l-juzʾ al-awwal min kitāb Yaʿqūb b. Isḥāq al-Kindī, “Thus has been completed the first section of the book of Yaʿqūb b. Isḥāq al-Kindī”⁴²

The chapter divisions are evidently original and constituted sub-divisions of a larger “section” (juzʾ) of the work. The first juzʾ of the work ends with the end of the fourth chapter; the manuscript copy thus apparently preserves the entire first juzʾ intact. The original work must have covered most of the topics found in Aristotle’s Metaphysics, and this provides an estimation that the complete work was originally two to three times as long as the extant text. It is not known with certainty how many parts (ajzāʾ) the work had, but it appears likely that it had two sections.

Al-Ahwānī suggests that the original work contained two sections and that each section corresponded to part of the title: the first to al-tawḥīd, God’s unicity, and the second to mā dūna l-ṭabīʿiyyāt, which he interprets as metaphysics.⁴³ He claims that the original work resembled two separate treatises, because, in his view, the end of the fourth chapter in the extant manuscript reads like a conclusion.⁴⁴ This argument is not convincing. It is more likely, as Ivry has argued, that the work followed the outline of Aristotle’s Metaphysics throughout and therefore did not resemble two separate treatises.

Three references to al-Kindī’s work al-Falsafa al-ūlā in later texts throw additional light on the original work’s form and content, though they somewhat blur the picture with regard to its title. Coincidentally, they were all recorded in al-Andalus, between the late third/ninth century and the mid-fifth/eleventh century, by Ibn ʿAbd Rabbih (d. 328/940), Ibn Ḥazm (d. 456/1064), and Ṣāʿid al-Andalusī (d. 462/1070). In his voluminous literary anthology, al-ʿIqd al-farīd, Ibn ʿAbd Rabbih quotes a passage of what appears to be al-Falsafa al-ūlā in a discussion of divine decree or predestination. He introduces the quotation with the statement, qāla l-Kindī fī l-fann al-tāsiʿ min al-tawḥīd, “Al-Kindī stated in the ninth chapter of Divine Unicity.”⁴⁵ This quotation suggests that al-Kindī’s work reached al-Andalus quickly and was well known there already by the early fourth/tenth century. It also indicates that Ibn ʿAbd Rabbih – or perhaps an intermediate source – had access to a more complete copy than that which has survived, for it included the ninth chapter (fann) of the work. Ibn ʿAbd Rabbih’s reference to the ninth chapter of course reveals that the original work contained at least nine chapters. The ninth chapter likely belonged to the second section (juzʾ) of the work, and the chapters in the second section presumably continued the numbering of those in the first section, beginning with the fifth fann. If the numbering had resumed in each section, one would expect Ibn ʿAbd Rabbih to have specified to which juzʾ the fann belonged. His reference gives the title of the work as al-Tawḥīd.

Over a century after Ibn ʿAbd Rabbih quoted the ninth chapter of al-Falsafa al-ūlā, Ibn Ḥazm penned a substantial critique of al-Kindī’s views on metaphysics in a treatise that has been published under the title Risālat al-Radd ʿalā l-Kindī al-faylasūf, “Treatise of Refutation of al-Kindī, the Philosopher.” Ibn Ḥazm’s original title is not known; the text appears without a title in the manuscript.⁴⁶ As Hans Daiber has argued, Ibn Ḥazm’s treatise responds primarily to al-Falsafa al-ūlā.⁴⁷ In the course of the discussion, Ibn Ḥazm refers to al-Kindī’s work as Kitābuhu fī l-tawḥīd, “His book on God’s unicity.”⁴⁸ He may not have intended by this to report the exact title of the work, instead merely indicating the topic discussed therein. Alternatively, he could have understood that the work’s title was actually Fī l-tawḥīd, (On [God’s] Unicity), or that this was the second half of the title, which normally indicates the content. The last possibility is most in keeping with the form of the title that Ibn ʿAbd Rabbih reports, al-Tawḥīd, as well as the title as Ibn al-Nadīm reports it: Kitāb al-Falsafa al-ūlā fī mā dūna l-ṭabīʿiyyāt wa-l-tawḥīd.

The oddest of the three references is that of Ibn Ḥazm’s contemporary Ṣāʿid al-Andalusī, who applies to al-Kindī’s work a title that occurs nowhere else in extant sources and is in addition difficult to explain. In his work on the history of the sciences, Ṭabaqāt al-umam, al-Andalusī provides a relatively extensive notice on al-Kindī, stressing his Arab tribal ancestry.⁴⁹ He clearly did not have access to al-Kindī’s full bibliography, for he reports that al-Kindī authored more than fifty works, a far cry from the over two hundred works in Ibn al-Nadīm’s Fihrist. The first work al-Andalusī lists in his short bibliography of al-Kindī is apparently al-Falsafa al-ūlā:

fa-min kutubihi l-mashhūra Kitāb al-Tawḥīd al-maʿrūf bi-Fam al-dhahab, dhahaba bihi ilā madhhab Aflāṭūn min al-qawl bi-ḥudūth al-ʿālam fī ghayr zamān wa-naṣara hādhā l-madhhab bi-ḥujaj ghayr ṣaḥīḥa baʿḍuhā sūfisṭāʾiyya wa-baʿḍuhā khiṭābiyya

Among his famous books is The Book of Divine Unicity, which is known as “The Mouth of Gold.” In it, he adopted the view of Plato, the opinion that the creation of the world did not occur in time. He supported this view with unsound arguments, some of which were sophistical, and some of which were rhetorical.⁵⁰

The title Fam al-dhahab (The Mouth of Gold) appears quite odd. It does not appear in any of the other references to al-Kindī’s al-Falsafa al-ūlā, and, furthermore, it appears as a surprising designation for a work of any kind. The phrase Fam al-dhahab is known in Arabic, however, as the translation of the epithet of the Christian writer John Chrysostom (d. 407 CE).⁵¹ Such a title is not attributed to al-Kindī elsewhere, nor is such a book title generally known. The surviving text of al-Falsafa al-ūlā gives no indication that would allow one to connect the content with the meaning of fam al-dhahab, whereas the title Kitāb al-Tawḥīd is perfectly comprehensible and recalls the title al-Tawḥīd cited by Ibn ʿAbd Rabbih and the phrase fī l-tawḥīd mentioned by Ibn Ḥazm. It seems possible that a copyist’s error occurred in Ṣāʿid al-Andalusī’s text or an earlier source on which he drew, and that al-Kindī’s work al-Falsafa al-ūlā was known in al-Andalus primarily as Kitāb al-Tawḥīd.

Modern attempts to translate Ibn al-Nadīm’s full title of al-Falsafa al-ūlā are in general agreement on one aspect: that the word al-tawḥīd in the title is parallel to mā dūna l-ṭabīʿiyyāt and understood to be governed by the preposition fī and not parallel to al-ṭabīʿiyyāt and understood to follow dūna. In other words, “First Philosophy” addresses two topics: A, that which is below Physics, and B, “God’s Oneness.” They all agree in excluding the possible sense “First Philosophy, on that which is below Physics and below God’s Oneness.”

The real problem regarding the interpretation of the title is the phrase mā dūna l-ṭabīʿiyyāt “what is below physics, or natural things.” It does not appear in the surviving manuscript of al-Kindī’s work or in any of the later references to the work. It must refer to metaphysics, because the contents of the work clearly belong to that topic. However, it seems to reflect an intended contrast with mā fawqa l-ṭabīʿiyyāt, which appears not only in the second philosophical title in the bibliography but also many times in al-Falsafa al-ūlā itself, referring to metaphysics. This presents an apparent contradiction. Al-Ahwānī, for example, argues that it is simply a reference to metaphysics, though he interprets mā fawqa l-ṭabīʿiyyāt as referring to metaphysics also.⁵² Adamson and Porman translate this phrase as “what is below natural things,” adding, in a note,

As they stand the first and second titles seem to deal with contrasting topics: what is “below (dūna)” natural things, and what is “above (fawqa)” them. But it is hard to see what could be “below natural things” – perhaps matter? And we may be dealing simply with two alternate titles for our On First Philosophy.⁵³

They sense that the titles are meant to contrast in some fashion, and bring out the problem of understanding what is “below” natural things, but then suggest again that the two titles might refer to the same work.

Mā dūna l-ṭabīʿiyyāt, if interpreted as “that which is below physics,” does not seem to correspond to metaphysics, but the problem may be resolved by the realization that dūna has other meanings. While dūna can mean “below, beneath, inferior, less than, nearer, without,” it can also mean, as Edward W. Lane puts it, “other than, beside, besides, exclusively of, or not.”⁵⁴ The phrase mā dūna l-ṭabīʿiyyāt could therefore mean “that which is not physics” or “that which is other than physics,” as opposed to “that which is below physics.” Interpreting the phrase in this manner has the advantage of sidestepping the contradiction described above. This interpretation finds some support in the second chapter of al-Falsafa al-ūlā, in which al-Kindī delimits clearly between what is physical and what is not physical with regard to the methods required to investigate them.

Whoever examines things that are above nature (al-ashyāʾ allatī fawqa l-ṭabīʿa), that is, those things that have no matter and are not joined to matter, will not find any representation of them in the soul, but rather may only perceive them through intellectual inquiries (al-abḥāth al-ʿaqliyya)…. For this reason, many inquirers regarding things that are above nature (al-ashyāʾ allatī fawqa l-ṭabīʿa) have been perplexed, since they have attempted to use representations of such things in the soul in order to investigate them, following their customary practice regarding the senses, as does a child…. The same holds with regard to physical things (al-ashyāʾ al-ṭabīʿiyya), when they attempt to use mathematical investigation [to examine them], for this should only be applied with regard to what has no matter. For matter is disposed to being acted upon, and therefore has motion, and nature (al-ṭabīʿa) is the primary cause of everything that may be in motion or at rest. Therefore, every physical thing (ṭabīʿī) is material, and hence it is not possible for mathematical investigation to be applied in the investigation of physical things, on the grounds that it pertains only to that which has no matter. Then, since mathematics is such that it may be used only with regard to investigation of what is not physical (mā laysa bi-ṭabīʿī), whoever uses it in the investigation of physical things (al-ṭabīʿiyyāt) will miss the mark⁵⁵ and will not attain the truth….

The physical (al-ṭabīʿī) is everything that moves, and hence the science of physics (ʿilm al-ṭabīʿiyyāt) is the science of everything that moves. And therefore, that which is above physics (mā fawqa l-ṭabīʿiyyāt) does not move, because it is not possible that something should be the cause of its own generation…. Therefore, what is above physics (mā fawqa l-ṭabīʿiyyāt) does not have motion. Hence, it has been clarified that the science of what is above physics (mā fawqa l-ṭabīʿiyyāt) is the science of what does not move.⁵⁶

Here, al-Kindī argues that one may investigate physical things using sense perception, but that this method does not work for non-physical things. Conversely, one must use intellectual methods (al-abḥāth al-ʿaqliyya) or mathematical investigation (al-faḥṣ al-riyāḍī) for non-physical things, and ordinary sense perception will not work for these topics. In this passage, he uses the phrases al-ashyāʾ allatī fawqa l-ṭabīʿa, “things that are above nature,” twice and the phrase mā fawqa l-ṭabīʿiyyāt, “what is above physics,” three times. In the course of the argument, however, al-Kindī stresses the distinction between what is physical and what is not. It is my contention that the phrase mā dūna l-ṭabīʿiyyāt is connected directly with this distinction and is used in a manner similar to a phrase that occurs at a critical point in this passage: mā laysa bi-ṭabīʿī, “what is not physical.” This explains how mā dūna l-ṭabīʿiyyāt and mā fawqa l-ṭabīʿiyyāt plausibly both refer to metaphysics, so that the contradiction no longer obtains.

Al-Falsafa al-Dākhila: The Evidence of Al-Kindī’s Treatise on Ptolemy

Ibn al-Nadīm presents the second title as Kitāb al-Falsafa al-dākhila wa-l-masāʾil al-manṭiqiyya wa-l-muʿtāṣa wa-mā fawqa l-ṭabīʿiyyāt,⁵⁷ which one might render literally as “The Book of Internal Philosophy, and Logical and Difficult Questions, and What Is above Physics.” Flügel translates the title as Das Buch der tiefer eingehenden Philosophie, der logischen und verwickelten Lehrsätze und dessen was über die physischen Dinge (als gewöhnliche Erscheinungen) hinausgeht (The Book of More Deeply Penetrating Philosophy, of Logical and Complex Propositions, and of That Which Goes beyond Physical Things (as Ordinary Phenomena)).⁵⁸ George N. Atiyeh translates it as “On Introduced Philosophy and Logical and Abstruse Problems and Metaphysics.”⁵⁹ Dodge translates the title as “Intrinsic (Inner) Philosophy, Logical and Difficult Questions, and Metaphysics (the Supernatural).”⁶⁰ Adamson and Pormann translate the title as “On Internal Philosophy, Logical and Recondite Questions, and What Is above Natural Things.”⁶¹

The meaning of the term al-dākhila and the relationship of al-Falsafa al-dākhila to al-Falsafa al-ūlā has puzzled modern scholars. Emma Gannagé recognizes this, omitting the troublesome term and translating the title as “Book of […] (?) Philosophy and of the Logical and Difficult Questions and What Is above Physics.”⁶² While metaphysics is certainly the topic of the first work, several of the translations mentioned earlier include metaphysics in the second title as well, suggesting that the second book is redundant or covers some of the same ground as the first. The translations of the second title vary a great deal because it is problematic and difficult to interpret not only on account of the adjective dākhila, “internal, inner, entering,” that modifies “philosophy” but also because the term al-muʿtāṣa disturbs the parallelism and does not have a clear referent. Most scholars have passed over this matter in silence. Adamson has suggested that the first two titles might both refer to a single work on “first philosophy” or metaphysics.⁶³ Discussing al-Falsafa al-ūlā, Adamson writes, “The Fihrist has two distinct titles that might refer to the work, which are the first two entries; in both the phrase ‘On First Philosophy’ is only the beginning of a longer title.”⁶⁴ In other words, he interprets al-Falsafa al-dākhila as conveying a meaning similar to that of al-Falsafa al-ūlā, “first philosophy.” It is my contention that this confusion may be cleared up, at least partially. The two works may be recognized as properly complementary, and not redundant.

One might be tempted to interpret the phrase al-Falsafa al-dākhila as meaning “Introduction to Philosophy.” The adjective dākhila is cognate with madkhal, one of the most common terms for introduction in Ibn al-Nadīm’s Fihrist and in Arabic works in the Greek sciences in particular. The Fihrist records seventeen titles that begin with the term al-Madkhal, “Introduction,”⁶⁵ not to mention other titles that include the term but do not begin with it. The section of al-Kindī’s bibliography that lists his logical works in the Fihrist includes two with the title Risālatuhu fī l-madkhal al-manṭiqī … “His Treatise on the Logical Introduction,” one extensive and one abridged, both of which may refer to Porphyry’s famous introduction to Aristotelian logic, the Isagoge.⁶⁶ However, to interpret the title al-Falsafa al-dākhila as equivalent to al-Madkhal ilā l-falsafa appears unjustified. The form dākhila, “entering,” rarely occurs in titles, and it is difficult to construe the active participle modifying the ostensible topic as equivalent in meaning to the noun madkhal, and meaning introduction to the topic.

An alternative suggestion is that the word has been corrupted in the manuscript tradition and that al-dākhila is simply a copyist’s error for something else. This supposition is precluded, however, by several early texts that use the term al-falsafa al-dākhila to refer unambiguously to al-Kindī’s philosophy. One reference occurs in a treatise by al-Kindī himself on Ptolemy’s Almagest, while the other occurs in the famous debate between the philosopher Abū Bishr Mattā b. Yūnus (d. 328/940) and the grammarian Abū Saʿīd al-Sīrāfī (d. 368/979) as recorded in Kitāb al-Imtāʿ wa-l-muʾānasa by Abū Ḥayyān al-Tawḥīdī (d. 414/1023).

The first reference occurs in a work al-Kindī wrote on astronomy that is titled Kitāb fī l-ṣināʿa al-ʿuẓmā (Book on the Greatest Craft) in the unique surviving manuscript. Franz Rosenthal analyzed the manuscript in a 1956 article and edited and published it much later, in 1987.⁶⁷ As Rosenthal points out, this work is probably identical to the book titled by Ibn al-Nadīm Risāla fī ṣināʿat Baṭlamyūs al-falakiyya (Treatise on Ptolemy’s Astronomical Craft). The work, as it has been preserved, treats the first eight chapters of the first book of Ptolemy’s Almagest. Early on in the text, al-Kindī quotes one of his own works regarding the idea that the divine is separate from all sensory data.⁶⁸ The Arabic text reads as follows:

ammā l-ilāhī fa-huwa ghayr mudrak bi-ʿilm yuḥīṭu bihi li-dhālika wa-li-annahu lā yaẓharu bi-shayʾ min al-ḥawāss wa-lā yuqārinu l-maḥsūsa bal mufāriq lahā abadan kamā bayyannā fī kitābinā fī l-falsafa al-ūlā al-dākhila wa-innamā l-mawjūd afʿāluhu llatī min ajlihā lazimat maʿrifatuhu jalla thanāʾuhu wa-khaḍaʿat al-ʿuqūl al-insiyya li-l-iqrār bihi.⁶⁹

As for the category of divine knowledge, it cannot be grasped by a science that would comprehend it. Therefore – and because [this category of divine knowledge] is not manifest to any of the senses and is not joined to sensible things but rather is always separate from them, as we have shown in our book on al-falsafa al-ūlā al-dākhila – only those actions of His are perceptible on account of which one is required to know Him – may His praise be manifest – and [on account of which] human intellects must submit to a confession of His existence.⁷⁰

The topic discussed evidently fits into the rubric presented in al-Kindī’s al-Falsafa al-ūlā, and the title is given as Kitāb fī l-falsafa al-ūlā al-dākhila (either: Book on First, Internal Philosophy; or: Book on Internal First Philosophy). Rosenthal realizes that this title is problematic, because it conflates the two titles al-Falsafa al-ūlā … and al-Falsafa al-dākhila…, yet it is clear in Ibn al-Nadīm’s list that the second is distinct from that of the first.⁷¹ Rosenthal suggests the possibility that the title as reported in this passage may be corrupt:

It seems possible that in the title quoted, the particle wa- was omitted before ad-dāḫilah, that al-Kindī actually referred to a combined edition of his Falsafah al-ūlā and Falsafah ad-dāḫilah, and that the passage referred to must be sought in some section of the Falsafa ad-dāḫilah. However, the assumption of an inexact reference cannot be dismissed.⁷²

In Rosenthal’s assessment, the relevant passage alluded to in this quotation does not appear in the text of al-Falsafa al-ūlā that has been preserved. He thought it more likely that it occurred in the other text, al-Falsafa al-dākhila. However, since the extant text of al-Falsafa al-ūlā is obviously incomplete, the passage may have occurred in one of the missing chapters of that work instead. To whichever of the two books one assigns the quotation, and however one resolves the form of the title of the work quoted, this indicates that the adjective al-dākhila that appears in the title al-Falsafa al-dākhila in Ibn al-Nadīm’s list is correct. It also corroborates the supposition that the two books were closely related and perhaps even formed a combined work.

Al-Falsafa al-Dākhila: The Evidence of the Debate between Abū Bishr Mattā and Abū Saʿīd al-Sīrāfī

The term al-falsafa al-dākhila also occurs in association with al-Kindī in the course of the famous debate between Abū Saʿīd Ḥasan b. ʿAbdallāh al-Sīrāfī (d. 368/969) and Abū Bishr Mattā b. Yūnus al-Qunnāʾī (d. 328/940) in front of the vizier Abū l-Fatḥ al-Faḍl b. Jaʿfar Ibn al-Furāt (d. 327/939). The debate has generally been viewed as emblematic of the tension between the Greek sciences and Arabo-Islamic learning during the formative period of Islamic intellectual history. It has been studied by a number of scholars, who agree on the gist of the exchange: as reported, the debate shows that al-Sīrāfī ably defended the thesis that Arabic grammar is superior to logic as a tool for distinguishing truth from falsehood. It is generally agreed as well that the report is biased against the philosophers. Al-Tawḥīdī’s (d. after 400/1009) work Kitāb al-Imtāʿ wa-l-muʾānasa is the only independent source of the debate, but modern scholars first became aware of it through the text of Irshād al-arīb by Yāqūt al-Ḥamawī (d. 626/1229), who quoted the text in full from al-Imtāʿ.⁷³ Al-Tawḥīdī reports that he recorded the text of the debate, which occurred in the year 326/937–8, from the dictation of ʿAlī b. ʿĪsā al-Rummānī (d. 384/994).⁷⁴

A number of scholars have translated and discussed the debate. D. S. Margoliouth translated the text into English, Taha Abderrahmane and Abdelali Elamrani-Jamal provided French translations, and Gerhard Endress translated the text into German. Muhsin Mahdi paraphrased the text of the debate but actually edited out all mention of al-Kindī, perhaps because he wanted to avoid raising an additional set of questions. A number of passages in the text have proved difficult to render partly because of the use of odd technical terms, but partly because the general flow of the argument is difficult to follow. As will be seen, the meaning of the term al-falsafa al-dākhila proved difficult for the translators to interpret, and they rendered it in several disparate fashions.

The reference to al-falasafa al-dākhila occurs in one passage of the debate in which al-Sīrāfī chides Abū Bishr by ridiculing al-Kindī as leader of the philosophers. He reports that, on one occasion, al-Kindī was posed a number of contrived philosophical questions by interlocutors who presented them as matters belonging to al-falsafa al-dākhila. Al-Kindī, who did not realize that they were setting him up to look like a fool, blithely answered them.⁷⁵ This section of the debate occurs as part of a subsidiary argument against Abū Bishr Mattā. If al-Kindī was one of the leading philosophers, and if al-Kindī can be shown to have made a number of ridiculous or nonsensical statements, then this reflects rather badly on the entire discipline that Abū Bishr Mattā is representing. The nature of this argument introduces an additional facet into the quest for an understanding of the term al-falsafa al-dākhila. It is used here in a parodic fashion, and one might therefore need to distinguish between the original or straight meaning intended by al-Kindī in his works and the distorted or parodic meaning intended by al-Sīrāfī in the debate. In the following excerpts, I have italicized the phrase used to render al-falsafa al-dākhila.

In his 1905 article, Margoliouth gives the following rendering of the passage in question:

… so that some persons made up questions of this style, and deluded him with them, making him suppose they belonged to the foreign philosophy; he did not perceive that they were inventions, and thought he must be deranged or diseased or indisposed or confused.⁷⁶

Margoliouth interprets al-falsafa al-dākhila here as “the foreign philosophy.”⁷⁷ In other words, he understands al-dākhila to convey the meaning of al-dakhīla, “intrusive, foreign.” He is influenced in this by the association of philosophy and logic with the Greeks. Abderrahmane adopts a similar interpretation. He published a French translation of the debate, without any discussion, in 1978, in which the passage devoted to al-Kindī is rendered as follows:

… au point qu’on lui a posé maintes questions qui l’induisirent en erreur en lui faisant croire qu’elles relevaient de la philosophie adoptée. Comme il ne pouvait pas voir qu’elles étaient de pures inventions, il prétextait qu’il devait avoir l’esprit fatigué ou le tempérament affecté ou le caractère perturbé ou le coeur indisposé.⁷⁸

Abderrahmane translates al-falsafa al-dākhila here as la philosophie adoptée “adopted philosophy.” This is close to Margoliouth’s translation, for by using this phrase he apparently intends to convey the idea that the philosophy that al-Kindī and his peers practice has been borrowed from the Greeks.

Elamrani-Jamal’s interpretation differs significantly from those of Margoliouth and Abderrahmane. He also translated the entire debate into French, in 1983, rendering the passage in question as follows:

On lui soumit par la suite d’autres questions sous une forme identique pour l’induire en erreur. On lui montra enfin qu’elles relevaient d’une pseudo-philosophie admise de sa part simplement par ignorance du sujet. Il était malade d’esprit, de tempérament corrompu, d’instinct dégénéré et de coeur trouble.⁷⁹

El-Amrani-Jamal points out in a note that the text contains a report of a discussion by al-Kindī on ontology that has been badly transmitted by al-Tawḥīdī or his informants. He states that he has translated it literally, without attempting to render the proper original sense.⁸⁰ He interprets al-falsafa al-dākhila here as pseudo-philosophie, “pseudo-philosophy.” I consider this a better rendering than those of Margoliouth and Taha Abderrahmane. It captures the parodic use of the term by al-Sīrāfī, and it fits the context well. However, it does not represent directly the original meaning intended by al-Kindī, who obviously intended it in a positive sense.

Gerhard Endress discussed this text in a 1986 publication, pointing out that it is difficult especially because it is parodic in nature. He writes, “Since the questions are expressly introduced as malicious pseudo-problems, we do not attempt to grant them a sensible meaning through the systematic breakdown of the text.” He adds, “The transmission of the text is hence quite uncertain.”⁸¹ Endress renders this section of the passage in question as follows:

… und dann legten sie ihm weitere Fragen dieser Art vor und verstrickten ihn damit in Fehler und gaben vor, es seien Themen der innersten Philosophie; er aber merkte nicht, was mit ihm gespielt wurde, und glaubte darob, sein Verstand sei krank, seine Gesundheit hinfällig, seine Natur verdorben und sein Herz verstört.⁸²

In his translation, Endress renders al-falsafa al-dākhila as innersten Philosophie, “innermost philosophy.” This rendition conveys the idea that the term is emphatic, something that goes along with the parody and the sarcastic tone of the passage. It suggests that the philosophy in question is recondite, arcane, or inaccessible to most observers. I believe, therefore, that Endress’s translation gets very close to the parodic meaning intended by al-Sīrāfī in the debate, while also conveying some of the original sense of the term.

This passage in the debate throws some light on the intellectual legacy of al-Kindī, despite the fact that it comes from a partisan and unsympathetic source. A commentary on al-Kindī’s oeuvre from just one or two generations after his death, it shows that he was looked upon as a leading figure in the study of philosophy in Baghdad and as a model for later generations of philosophers. However, the portrayal of al-Kindī shows that not long after his death, his work was known to be difficult to follow, in part because he used idiosyncratic technical terms, and in part because his discussions were convoluted. In the case of scholars who had a low opinion of philosophy, to this was added the idea that this difficulty was not backed up by actual erudition; in their estimation, al-Kindī was, in fact, as bewildered by his own discussions as anyone else. This view on the part of the Baghdadi audience is corroborated by a book title attributed in the Fihrist to al-Jāḥiẓ (d. 255/868): Risāla fī farṭ jahl Yaʿqūb b. Isḥāq al-Kindī (Treatise on the Extreme Ignorance of Yaʿqūb b. Isḥāq al-Kindī).⁸³ The reference to al-falsafa al-dākhila in the debate of al-Sīrāfī confirms, again, that this term was part of al-Kindī’s usage and thus can be assumed to belong properly to the title reported in the Fihrist. Moreover, it suggests that al-Falsafa al-dākhila was one of his most famous books, one that was emblematic of his entire philosophical oeuvre.

The translations of the debate render al-falsafa al-dākhila in as many ways as there are translators, though one may see some similarity among several of the renditions: “foreign philosophy,” “adopted philosophy,” “pseudo-philosophy,” and “innermost philosophy.” The text of the debate between Abū Saʿīd al-Sīrāfī and Abū Bishr Mattā shows that al-falsafa al-dākhila was a technical term associated with the writings of al-Kindī in particular, but the problem of its interpretation remains, exacerbated especially by the fact that it did not become a standard term in later philosophical discourse in Arabic. The term al-falsafa al-dākhila is understood by al-Sīrāfī to be odd and unnatural in addition to being obscure in meaning to him and to his contemporaries already in the early fourth/tenth century. Al-Kindī seems to have intended by al-falsafa al-dākhila the detailed, advanced treatment of the questions of philosophy, as opposed to its broad concepts and outlines. This usage led al-Sīrāfī to use the term to mean the confused or overly complex and convoluted discussion of philosophical questions.

This interpretation finds significant support in one passage of al-Falsafa al-ūlā that discusses the relationship of first philosophy to the rest of philosophy. In the first chapter of the work, al-Kindī explains that “knowledge of the First Cause [= God]” has been called first philosophy “since all of the rest of philosophy is contained in the knowledge of it” (ʿilm al-ʿilla al-ūlā … idh jamīʿ bāqī l-falsafa munṭawin fī ʿilmihā).⁸⁴ In other words, first philosophy contains, encompasses, or subsumes all the rest of philosophy. This statement hints at al-Kindī’s understanding of al-Falsafa al-dākhila: it is the rest of philosophy, that which first philosophy contains. It is called internal because it is internal to first philosophy, being comprised, contained, or subsumed within it. This interpretation fits with al-Kindī’s evident concern to present philosophy in its entirety as a unified system. Both works therefore address metaphysics, but one must understand the relationship between al-Falsafa al-ūlā and al-Falsafa al-dākhila as that of the foundation or outline to the subsidiary topics and details.

This sense of the term al-dākhila, as well as its suggestion of matters that are arcane and difficult, is likely related to the use of the term “esoteric,” a label scholars in the Aristotelean philosophical tradition used to refer to a category of Aristotle’s works. As Jonathan Barnes writes, “The distinction in Aristotle’s works between the esoteric and the exoteric is an ancient one.”⁸⁵ Aristotle’s books were divided into categories in many ways, but one of the most prevalent was a division into “exoteric” and “acroatic” or “acroastic” works. These divisions have been interpreted in two main ways, focusing either on their form or on their intended audience. Scholars who interpret the terms as referring to the form argue that the exoteric works are those which were edited and polished with care and shaped into final copies, whereas the acroatic works were in the form of unedited lectures not yet edited into their final form. Scholars who interpret the terms as referring to the audience argue that the exoteric works are intended for a wide, public audience, and so are more general and less technically complex, while the acroatic works are intended for a narrower audience of scholars and students of philosophy, originally those who attended the Lycaeum, and so focus on complex and technical matters. In the end, the two interpretations overlap, but most scholars lean toward the latter definition.⁸⁶ It is thus possible that al-Kindī’s use of al-dākhila originated as a translation of the Greek esoterike, “esoteric, internal.”

Al-Kindī’s adoption of the term al-dākhila as a calque on a Greek expression would be one of many cases in the philosophical tradition when a Greek term has been translated literally into Arabic, creating subsequent confusion. A well-known example of this is the translation of Aristotle’s Physics as Samʿ al-kiyān, a rather hopeless and misleading rendition of the Greek, probably through Syriac. The original title in Greek meant “the lecture on natural things,” but the Arabic rendition, taken literally, would mean “the hearing of being.” Scholars have also remarked on the oddity of al-Kindī’s Arabic, including his technical terms.⁸⁷ An obvious instance of overly literal translation from al-Kindī’s work is the title of the treatise investigated above, on the “quantity” of Aristotle’s books. Idiomatic Arabic would use ʿadad, “number” and not kammiyya, “quantity.” Similarly, for example, al-Kindī uses iqnāʿ, literally, “convincing” to mean mathematical proof, something that is not idiomatic from the point of view of later Arabic.⁸⁸ The use of the term “esoteric” could have been transmitted to al-Kindī’s circle through the works of Christian philosophers of Late Antiquity, who adopted the terms used in the Aristotelian tradition and mapped them onto a distinction between ordinary or “exoteric” philosophy, which had to do with secular topics, and special “esoteric” philosophy, which had to do with theological doctrine and with Christian doctrine in particular.⁸⁹

Corruption of the Title al-Falsafa al-dākhila

كتاب الفلسفة الأولى في ما دون الطبيعيّات والتوحيد

كتاب الفلسفة الداخلة والمسائل المنطقيّة والمعتاصة وما فوق الطبيعيّات

As we have seen above, Ibn al-Nadīm presents the full title of al-Falsafa al-dākhila as Kitāb al-Falsafa al-dākhila wa-l-masāʾil al-manṭiqiyya wa-l-muʿtāṣa wa-mā fawqa l-ṭabīʿiyyāt,⁹⁰ which one might render literally as “The Book of Internal Philosophy, Logical and Difficult Questions, and What is above Physics.” This title presents several problems of interpretation. It is evidently parallel to that of al-Falsafa al-ūlā, but if the intention indeed was to form a pair of these works and to present them as such in al-Kindī’s bibliography, certain issues come to the fore.

A first problem has to do with the usual form adopted by medieval Arabic book titles: X fī Y. They commonly consist of two parts, the first giving a general or descriptive title and the second, beginning with the preposition fī, “on, about, regarding,” announcing the topic discussed in the work. In this case, al-Falsafa al-ūla has this structure, and one might expect al-Falsafa al-dākhila, its counterpart, to follow the same pattern. However, in the second title, the conjunction wa-, “and,” appears where the preposition fī, “on,” should occur: Kitāb al-Falsafa al-dākhila wa-l-masāʾil al-manṭiqiyya….

The second problem has to do with the descriptions of the contents of the two works, which show a curious lacuna. The first title reports that the work is concerned with what is below Physics, while the second states that it is concerned with what is above Physics. Whatever happened to Physics? One would expect physics to appear somewhere in the two titles, unless there were originally a third title that has for some reason dropped out of Ibn al-Nadīm’s list.

The third problem has to do with the word al-muʿtāṣa, “difficult, abstruse.” This particular form is rare and represents odd diction in this context. The much more common word for “difficult” is mushkil, which in fact occurs in myriad book titles; somewhat less common is the synonym muʿḍil, which also occurs in book titles occasionally. Much more common than muʿtāṣ is its cognate ʿawīṣ, “difficult to comprehend, recondite,” which appears in one passage of the extant portion of al-Falsafa al-ūlā.⁹¹ An interpretation of the title would have to address the choice of this relatively rare term.

The fourth problem has to do with the parallel structure of the phrase wa-l-masāʾil al-manṭiqiyya wa-l-muʿtāṣa wa-mā fawqa l-ṭabīʿiyyāt. This part of the title has the structure A and B and C. If one adopts the reasonable assumption that al-ṭabīʿiyyāt is shorthand for physical questions, al-masāʾil al-ṭabīʿiyya, then A and C are properly parallel, but B, al-muʿtāṣa, meaning “difficult,” ruins the parallelism. The terms al-manṭiqiyya and wa-mā fawqa l-ṭabīʿiyyāt designate areas of content addressed in the work, but al-muʿtāṣa does not, describing a quality of the questions to which it refers instead. One would either want it to be an adjective attached to one of the other descriptions of the content, without the conjunction wa-, or else to be an adjective designating content, like its counterparts in the list.

These four problems suggest the possibility that the second title has been corrupted in transmission. Recourse to the oldest extant manuscript of the Fihrist, Şehid Ali Paşa MS 1934, preserved in Istanbul, does not resolve the issue. Dodge already consulted the manuscript for his translation of the Fihrist, as did Ayman Fuʾād Sayyid in his edition of the Arabic text. Close examination of the titles show that the editors have read them correctly, as they are recorded, except that they have provided dots for some letters that are not dotted in the manuscript. The corruption could have occurred between Ibn al-Nadīm’s original and the Şehid Ali Paşa MS, in Ibn al-Nadīm’s original, or already in the source on which he drew, but it is not the fault of the editors or translators of the Fihrist.

The first problem may be resolved by a simple emendation. In my view, the preposition fī, “on,” was necessarily part of the original title, which must have read Kitāb al-Falsafa al-dākhila fī l-masāʾil al-manṭiqiyya…. This would be more in keeping with the usual form of book titles, and also with the parallelism intended between this title and that of al-Falsafa al-ūlā.

Al-Kindī’s On the Quantity of Aristotle’s Books

Help in resolving the second, third, and fourth issues regarding designation of the content of al-Falsafa al-dākhila may be sought in al-Kindī’s treatise On the Quantity of Aristotle’s Books. In that treatise, al-Kindī divides Aristotle’s books into four categories, after propaedeutics or preliminary exercises (riyāḍiyyāt). He takes up this taxonomy three times in the treatise, each time following the same order. First, he enumerates the works; secondly, he presents succinct summaries of the works; and thirdly, he treats Aristotle’s aims or intentions in authoring the works. In the first presentation of the taxonomy, al-Kindī states that the first category is logic, on which Aristotle wrote eight books: The Categories, On Interpretation, (Prior) Analytics, Second (Posterior) Analytics or Apodeictics, Topics, Sophistics, Rhetoric, and Poetics. To refer to this category of books al-Kindī uses the term al-manṭiqiyyāt or al-manṭiqiyya.⁹² The second category consists of Aristotle’s physical books and includes seven works: Lecture on Physics, On the Heavens, On Generation and Corruption, On Atmospheric and Terrestrial Phenomena (also known as Things on High or Treatise on Extremes), On Minerals, On Plants, and On Animals. Al-Kindi uses the terms al-ṭabīʿiyyāt or al-ṭabīʿiyya to refer to these books.⁹³ The third category consists of the books On the Soul, On Sense-Perception and Sense-Objects, On Sleep and Waking, and On the Length and Brevity of Life. Al-Kindī does not use a particular label for this category when he introduces them, but describes them as books that are about “things that have no need for bodies for their persistence and subsistence, but may exist together with bodies.”⁹⁴ Jolivet interprets this third category as a reference to Aristotle’s “psychological” works.⁹⁵ In the later section of the treatise that is devoted to explaining the aims (aghrāḍ) of Aristotle’s works, al-Kindī uses the term kutubuhu l-nafsāniyya, “his psychological books,” to refer to the works of this category.⁹⁶ The fourth category consists of one book, the Metaphysics (mā baʿda l-ṭabīʿiyyāt), “on things that have no need for bodies and are unconnected to bodies.”⁹⁷ Al-Kindī also uses the phrase mā fawqa l-ṭabīʿiyyāt to refer to this category, as will be seen below.

These four categories, I argue, map onto the second title in Ibn al-Nadīm’s list of al-Kindī’s philosophical works. The title as it stands reads: Kitāb al-Falsafa al-dākhila wa-l-masāʾil al-manṭiqiyya wa-l-muʿtāṣa wa-mā fawqa l-ṭabīʿiyyāt. It is clear that al-masāʾil al-manṭiqiyya, “logical questions,” refers to logic, the first of the four categories. Similarly, one may connect mā fawqa l-ṭabīʿiyyāt, “what is above physics,” with the fourth category, that of metaphysical works.⁹⁸ The likelihood of this interpretation is confirmed by a paragraph in the treatise in which al-Kindī summarizes the curricular stages of the philosopher’s education:

Fa-qad yanbaghī li-man arāda ʿilm al-falsafa an yuqaddima stiʿmāl kutub al-riyāḍiyyāt⁹⁹ ʿalā marātibihā llatī ḥaddadnā, wa-l-manṭiqiyyāt ʿalā marātibihā llatī ḥaddadnāhā ayḍan, thumma l-kutub ʿalā l-ashyāʾ al-ṭabīʿiyya ʿalā l-qawl alladhī ḥaddadnā ayḍan, thumma mā fawqa l-ṭabīʿiyyāt, thumma kutub al-akhlāq wa-siyāsat al-nafs bi-l-akhlāq al-maḥmūda …

Whoever wants to acquire the science of philosophy must first make use of the books of mathematics according to their proper order that we have specified, and the logical [books] according to their proper order that we have specified as well, then the books regarding physical things, according to what we have said about them as well, then that which is above physics, then the books of ethics and the control of one’s soul through praiseworthy morals.¹⁰⁰

What is especially important here is that the term used is mā fawqa l-ṭabīʿiyyāt, the same term that appears in the second title. Especially in combination with reference to al-manṭiqiyyāt, this suggests that the title al-Falsafa al-dākhila is related to a taxonomy of Aristotle’s works similar to that which appears in al-Kindī’s presentation of them in this work.

The reference to logical questions (al-masāʾil al-manṭiqiyya) and the use of the term mā fawqa l-ṭabīʿiyyāt both suggest a connection between the second title and al-Kindī’s taxonomy of Aristotle’s works, covering two of the four categories, but then this suggests that the two middle categories, those of physics and psychology, are missing. And in the place where they would have occurred in the title stands the single, puzzling term wa-l-muʿtāṣa. This corroborates the suspicion that the term wa-l-muʿtāṣa is problematic and that the title is corrupt at this point.

Physics in al-Falsafa al-Dākhila?

It is very likely that al-Falsafa al-dākhila included a discussion of physics. Two pieces of evidence, beyond the simple matter that a reading of the two titles suggests that mention of “physics” is oddly lacking, indicate that this was the case. The first piece of evidence is provided by the questions al-Kindī’s opponents pose to him, as recorded in the debate between al-Sīrāfī and Abū Bishr Mattā. Al-Sīrāfī mentions several example problems that were put to al-Kindī, including the following:

qālū lahu ayḍan: mā nisbat al-ḥarakāt al-ṭabīʿiyya ilā l-ṣuwar al-hayūlāʾiyya? wa-hal hiya mulābisa li-l-kiyān fī ḥudūd al-naẓar wa-l-bayān, aw muzāyila lahu muzāyalatan ʿalā ghāyat al-iḥkām?

They also asked him, “What is the relationship of physical motions to material forms? Are they coextensive with these forms within limits that can be determined by speculation and demonstration? Or are they completely separate from them, in the strictest fashion?”¹⁰¹

Even though the questions posed to al-Kindī are malicious pseudo-problems, as Gerhard Endress points out, this question evidently refers to motion in particular, which, al-Kindī stresses, is the proper province of physics. This question also refers explicitly to physics in the term al-ḥarakāt al-ṭabīʿiyya, “physical motions.” Since all of the questions posed to al-Kindī are presented as being related to his “internal philosophy,” this suggests that the work al-Falsafa al-dākhila included a discussion of physics.

The second piece of evidence is a passage in al-Falsafa al-ūlā in which al-Kindī refers to a discussion of his own on physics. The passage reads as follows:

Therefore, it is incumbent upon anyone inquiring into any science to inquire first about the cause of the things that fall under that science. If we ask what causes natural dispositions (al-ṭibāʿ), which are the cause of physical things (al-ashyāʾ al-ṭabīʿiyya), we find, as we said in the early parts of The Physics, that it is the cause of all motion. Therefore, the physical (al-ṭabīʿī) is everything which moves, and hence the science of physics (ʿilm al-ṭabīʿiyyāt) is the science of everything that moves.¹⁰²

The key phrase reads kamā qulnā fī awāʾil al-ṭabīʿa, which would be translated as “as we said in the early parts of The Physics.” In my view, this refers to a particular book, and the title of the book invoked is al-ṭabīʿa, “The Physics,” though this could be a type of shorthand reference. Scholars to date have interpreted this phrase somewhat differently. Ivry also believes that al-Kindī is referring to a book, but he gives the title as “The Principles of Physics.”¹⁰³ In other words, he interprets the title as being made up of the words Awāʾil al-ṭabīʿa, and he translates awāʾil as “principles,” rather than meaning “the first parts, early parts.” In a note, he proposes two possible identifications of the work invoked with titles recorded in the Fihrist: “Book on the Action and Affection of First Natures” (Kitāb fī l-fāʿila wa-l-munfaʿila min al-ṭabīʿiyyāt al-ūlā) or the “Book on the Principles of Sensible Things” (Kitāb fī awāʾil al-ashyāʾ al-maḥsūsa).¹⁰⁴ Jolivet and Rashed translate the phrase as “comme nous l’avons dit à propos des principes de la nature,” – i.e., “as we have said with regard to the principles of nature” – which suggests that they do not interpret it as a reference to a particular title.¹⁰⁵ In my view, al-Kindī’s statement ostensibly refers to the early parts or questions in a particular work devoted to physics. Since it is not known that al-Kindī wrote a work with the title al-Ṭabīʿa, it appears likely that this reference alludes to part of al-Falsafa al-dākhila that was devoted to physics. The most likely alternative is that this statement refers to an independent work al-Kindī wrote on physics, the title of which has not been recorded.

If al-Falsafa al-dākhila contained a substantial discussion of physics, then it stands to reason that that topic would have appeared in the original title. The title may have read, therefore, al-Falsafa al-dākhila fī al-masāʾil al-manṭiqiyya wa-l-ṭabīʿiyya … wa-mā fawqa l-ṭabīʿiyyāt, “Internal Philosophy, on Logical and Physical Questions … and on What is above Physics.” This title may have been corrupted by a type of haplology. Since the term al-ṭabīʿiyyāt already occurs in the title, a copyist may have felt that it was repeated unnecessarily or had been introduced by mistake, or he may have simply omitted it out of inadvertence.

Psychology in al-Falsafa al-Dākhila?

A remaining question is how to explain the term al-muʿtāṣa in al-Kindī’s title. The title has the form X, on A and B and C. This adjective, B, follows the adjective al-manṭiqiyya, A; if the two were meant to be paired, a more satisfying parallelism would have been created by an adjective that ends in the nisba form, -iyya, and that refers to one category among the contents of the work. The word al-muʿtāṣa possibly arose as a corruption or misreading of an original word ending in -iyya, so that the end of the word, -ṣa actually was originally -Ciyya (C here representing an unknown consonant). A likely candidate is al-nafsāniyya, “psychological” (a.l-n.f.s.a.n.y.h), which may have been corrupted to al-muʿtāṣa (a.l-m.ʿ.t.a.ṣ.h). The definite article al- is the same in both cases, as is the final tāʾ marbūṭa (h). The loop of the letter fāʾ could have been mistaken for that of an ʿayn; the alif is in the same position in the two words; and the combination n-y could have been confused for one larger letter, ṣ. There is quite a bit of difference between the written forms of wa-l-nafsāniyya and wa-l-muʿtāṣa, but it nevertheless is possible to imagine the corruption, especially if the diacritical points were lacking in the original.

النفسانيّة >> المعتاصة

The term al-nafsāniyya fits well from the perspective of meaning. As we have seen above, al-Kindī uses it to refer to Aristotle’s psychological works. In his taxonomy of Aristotle’s works the category of psychological works comes after the categories of logical and physical works and before the category of metaphysical works.

The Reconstructed Form of the Title of al-Falsafa al-Dākhila

My proposal is therefore that the original form of the second title was as follows: Kitāb al-Falsafa al-dākhila fī l-masāʾil al-manṭiqiyya wa-l-ṭabīʿiyya wa-l-nafsāniyya wa-mā fawqa l-ṭabīʿiyyāt, “The Book of Esoteric Philosophy, on Logical, Physical, and Psychological Questions, and What Is above Physics.”

<< كتاب الفلسفة الداخلة والمسائل المنطقيّة والمعتاصة وما فوق الطبيعيّات

كتاب الفلسفة الداخلة في المسائل المنطقيّة والطبيعيّة والنفسانيّة وما فوق الطبيعيّات

Even if one has doubts about the specific emendation intended to restore the original al-nafsāniyya for the corruption al-muʿtāṣa, it is highly probable that the original, intact title of al-Falsafa al-dākhila referred to the topics of physics and psychology in some fashion. The title clearly invokes the taxonomy of Aristotle’s works according to al-Kindī because it mentions the categories of logic and metaphysics, and physics and psychology are the two categories that fit in between the two categories of logic and metaphysics in that taxonomy. The work itself presumably covered those four categories in the same order in which al-Kindī ordinarily presented them.

The division of contents of al-Falsafa al-dākhila envisaged here is corroborated not only by reference to al-Kindī’s work On the Quantity of Aristotle’s Books and by general knowledge of the Aristotelian curriculum, but also by reference to the work of al-Fārābī on the philosophy of Aristotle. Al-Fārābī discusses the sciences according to Aristotle as follows:

According to him [i.e., Aristotle], therefore, there emerge three sciences: the science of logic, natural science, and voluntary science. He let logic take the lead in the latter two sciences and gave it the authority to judge them and examine whatever takes place in them. Since the beings covered by these two sciences – that is, natural science and voluntary science – are one in genus, and since the primary intention of the science of logic is to give an account of the above-mentioned things with respect to the beings covered by natural science and voluntary science, he came to the view that the materials and subjects of the three sciences are subjects that are one in genus. And since the science of logic should precede the other two sciences, he began to enumerate at the outset the beings that are the materials and subjects of the three sciences and that comprise what exists by nature and what exists by will. Those existing in nature are the subject of natural science, those existing by will alone are the subject of voluntary science, and those that are common – that is, can be produced by either nature or will – are the subject of both sciences. The art of logic gives one part of what he has to know about the subjects of these two sciences. Hence the science of logic shares with these sciences their primary subjects and materials.¹⁰⁶

This passage shows clearly a tripartite taxonomy of al-Kindī’s works that maps onto the division suggested above with regard to al-Kindī’s work al-Falsafa al-dākhila. The main apparent difference is caused by Muhsin Mahdi’s translation of al-nafsāniyya here as “voluntary” rather than “psychological.” The rest of al-Fārābī’s treatise adopts a structure based on this division, treating first Aristotle’s logical works, secondly his works on the natural sciences, and thirdly his works on the soul or psychology.¹⁰⁷ This is followed by a section on Aristotle’s metaphysics.¹⁰⁸ Al-Fārābī may have arrived at this taxonomy independently, drawing on his knowledge of Late Antique sources, and may not have been influenced by al-Kindī in this regard, but his understanding nevertheless suggests that this taxonomy, with the same categories presented in the same order, was influential.

Conclusion: A Two-Volume Guide to Aristotle’s Philosophy

The first two titles presented in Ibn al-Nadīm’s list of al-Kindī’s philosophical works formed a pair and represented a fundamental attempt on the part of al-Kindī to create a compact epitome of the philosophical works of the Aristotelian corpus. This is suggested by the prominent and adjacent placement of the two titles, the evident parallelism in form between them, and a consideration of the contents that they cover, taking into account al-Kindī’s understanding of the taxonomy of Aristotle’s works. This confirms Rosenthal’s suspicion, mentioned above, which was based on al-Kindī’s reference in his work on Ptolemy’s Almagest: the fact that he seems to refer both to al-Falsafa al-ūlā and to al-Falsafa al-dākhila occurring in what looks like one title suggests that al-Kindī was referring to a combined edition of the two works. The first work, in al-Kindī’s view, presented the grand scheme into which the rest of philosophy fits, and the second work presented the detailed discussions that fit into that scheme. It is probably on this account that he uses the term al-Falsafa al-dākhila: it is what is contained in or subsumed by al-Falsafa al-ūlā.

The title of the first work as presented by Ibn al-Nadīm, Kitāb al-Falsafa al-ūlā fī mā dūna l-ṭabīʿiyyāt wa-l-tawḥīd is correct in its present form and means “Book of First Philosophy, on What is Other than Physics and on God’s Unicity.” The title of the second work is given by Ibn al-Nadīm in the form Kitāb al-Falsafa al-dākhila wa-l-masāʾil al-manṭiqiyya wa-l-muʿtāṣa wa-mā fawqa l-ṭabīʿiyyāt, which appears to be corrupt for several reasons. I have suggested that it should be emended to Kitāb al-Falsafa al-dākhila fī l-masāʾil al-manṭiqiyya wa-l-ṭabīʿiyya wa-l-nafsāniyya wa-mā fawqa l-ṭabīʿiyyāt, meaning “Book of Internal Philosophy, on Logical, Physical, and Psychological Questions and Metaphysics.” This involves three particular changes. First is the replacement of the first conjunction wa-, “and,” with fī, “on,” which makes the title adopt the form X fī Y that is typical of medieval Arabic book titles and also match the first title, Kitāb al-Falsafa al-ūlā fī mā dūna l-ṭabīʿiyyāt wa-l-tawḥīd. Second is the addition of al-ṭabīʿiyya, referring to physical questions, which may have dropped out of the original title on account of haplology. Third is the emendation of al-muʿtāṣa to al-nafsāniyya, referring to psychological questions. Overall, these emendations serve to make the titles more parallel and are based on the supposition that al-Falsafa al-dākhila covered the four main categories of Aristotle’s philosophical works.

While al-Falsafa al-dākhila is not known to be extant, al-Kindī’s successors likely drew on the work for their understanding of Aristotle, and some material from the work may have been preserved in later works or in manuscripts that have yet to be investigated. Together with al-Falsafa al-ūlā, it formed part of a two-volume compendium that provided an overview of all of philosophy, confirming the centrality of Aristotle’s corpus in al-Kindī’s thought. They suggest that al-Kindi was engaged in a grand philosophical project which involved translating and adapting Aristotle’s oeuvre as completely as possible and presenting it to an Arabic-speaking audience. This general interpretation does not depend on acceptance of the emendations proposed here. Those emendations, however, suggest possible ways in which the full title of al-Falsafa al-dākhila might be shown to match that of al-Falsafa al-ūlā more closely in form, while reconciling its contents more fully with al-Kindī’s taxonomy of Aristotle’s works.