SOURCE: Galston, Miriam. “Alfarabi's Method of Writing.” In Politics and Excellence: The Political Philosophy of Alfarabi, pp. 22-54. Princeton, N.J.: Princeton University Press, 1990.

[In the following essay, Galston analyzes al-Fārābī's writing style—a multilevel method which both conceals and reveals—and his purpose in adopting this method.]

Those who wish to succeed in arriving at answers will find it useful to go over the perplexing points well. For answers successfully arrived at are solutions to the perplexing points that have previously been raised. A person cannot untie a knot if he is not aware of it.

—Aristotle Metaphysics III. 1¹

Philosophic discourse has been the object of philosophic inquiry since the time of Plato. According to Socrates, as presented by Plato in the Phaedrus,² concern with the relative merits of oral and written communication can be traced back to the ancient Egyptian king Thamus, who expressed the fear that the invention of writing would eventually dull people's memories and breed a class of men laden with information, but lacking genuine wisdom (Phaedrus 274-275). Plato's Socrates asserts the superiority of oral instruction to its written counterparts and, as a corollary, advances the view that those who really know will only put pen to paper playfully, to amuse themselves, as others indulge in drinking parties and similar diversions (Phaedrus 276). This view did not prevail, nor did the Greek philosophical world arrive at a consensus. Socrates left no writings. And of those philosophers who wrote, some wrote conversations, some wrote treatises, and some wrote commentaries or supercommentaries on the writings of others.

The problem of the method or methods of philosophic communication is regularly addressed by students of medieval philosophy, especially students of medieval Islamic and Jewish philosophy, for a number of reasons. The vocabulary, the style, and very often the themes that preoccupy medieval authors are unfamiliar and thus obscure to modern readers. Further, several medieval philosophers themselves advocated or attributed to others a special manner of constructing books that proceed on several levels of argument simultaneously.³ Others, such as the twelfth-century philosopher Maimonides, laboring under a religious prohibition against communicating certain kinds of wisdom to the community at large, claimed to have composed books in such a way as to conceal from some readers what they revealed to others.⁴ Bearing a superficial resemblance to these two types of multilevel writing is a religious doctrine that was current in medieval times, namely, the doctrine that revealed texts consist in large measure of hidden truths expressed in metaphorical language, symbols, and parables.⁵ Even the importance of esotericism for Islamic mysticism may have helped sustain interest in this subject among Islamicists well after its urgency was lost for historians of philosophy in general.⁶

Because doctrines of Islamic esotericism are well established and have been widely discussed and because of a tendency on the part of some to equate the various modes of philosophic multilevel writing with their religious counterparts, the nature of philosophic writing is frequently addressed by those who study Islamic philosophy, even though scholars still disagree about the presence or type of multilevel writing in the works of individual Muslim philosophers. Thus, George Hourani argues for “an impressive tradition of secrecy among philosophers preceding Averroes,”⁷ while E. I. J. Rosenthal “cannot see any convincing reason [for interpreting Averroes esoterically] any more than in the case of Maimonides.”⁸ Salomon Munk, Shlomo Pines, and Louis Gardet all accept multilevel writing of some kind as common, if not universal, among Muslim philosophers.⁹ According to Leo Strauss certain kinds of textual difficulties and irregularities in the works of great philosophers impose an obligation on the reader to reflect on the possibility of multilevel writing but do not point unequivocally to its presence. However, when in addition persecution by religious or political authorities is known to have been a real danger for those who voiced heterodox views, he believes that there should be a presumption that the author in question had recourse to multilevel writing.¹⁰ Muhsin Mahdi, who prefaces his study of Ibn Khaldūn with a general discussion of the question of philosophic discourse, singles out an author's apparently disinterested discussion of multilevel writing in other authors' works or in Scripture as a signal that the author himself writes in this manner.¹¹ He attributes multilevel writing to Alfarabi, in particular, both because of Alfarabi's discussions and apparent approval of the esoteric methods of writing of Plato and Aristotle and because exoteric writing is a necessary consequence of the nature of the relationship between the philosopher and the nonphilosophers.¹² Several recent studies of Averroes discuss in some detail the special features of medieval philosophic commentaries, which can be viewed as a distinctive subset of multilevel writing.¹³

Alfarabi's method of writing has received somewhat less attention than that of Averroes or Avicenna. Strauss, who was among the first to analyze in detail Alfarabi's method of writing, takes the position that all of Alfarabi's writings are “exoteric,” that is, not to be taken literally, although he singles out Falsafat Aflāṭun as the least exoteric of the philosopher's writings.¹⁴ Likewise, Mahdi characterizes Alfarabi's writings generally as exoteric, while observing gradations among the works and even within individual works.¹⁵ Richard Walzer attributes the difficulties in Alfarabi's style of writing to the fact that his treatises summarize the conclusions of his philosophic investigations, but leave to the reader the discovery of their application.¹⁶ While eschewing doctrines of deliberate secrecy and levels of writing,¹⁷ E. I. J. Rosenthal also takes note of the obscurity of Alfarabi's prose and traces the problematic character of his writing to its being “diffuse, repetitive and lacking in clarity and precision.”¹⁸

There is, then, a consensus as to the impenetrability of Alfarabi's works, although scholars disagree about the underlying reason for the difficulty. It is thus necessary to begin a study of Alfarabi's thought with an examination of the problem of his writing. For one's view of the origin of the obscurity of his style will influence the manner in which one reads his works and, as a consequence, the interpretation of his teachings one eventually reaches.

In his “Translator's Introduction” to Maimonides' Guide of the Perplexed, Pines issues a word of caution against presuming an identity between the doctrines or the methods of Alfarabi and those of Maimonides: “none of Alfarabi's writings is as carefully designed as is the Guide to throw the unqualified readers and many qualified ones off the right track.”¹⁹ According to Pines, not only the scope of the Guide of the Perplexed but “its composition and its purpose are quite different from those of any of al-Fārābī's works.”²⁰ Despite this assurance as to the relative openness of Alfarabi, as compared with Maimonides, the obstacles to grasping his method of writing are still considerable.

Given the currency in the medieval world of assorted theories of multidimensional writing, an analysis of the genus of multilevel writing will serve as a useful starting point for this inquiry. Of course, even when the full range of types of multilevel writing has been clarified, it will be difficult to demonstrate the philosopher's commitment to the genre, much less to any specific type of multilevel writing. For the presence of stylistic devices associated with this type of writing in a particular author's works does not definitively resolve the question of the author's intention in thus composing them. Some of the hallmarks of multilevel writing—such as inconsistencies, contradictions, digressions, and silences where the reader expects a lengthy discussion—can be deliberate or inadvertent; and even when deliberate, they may be the result of a variety of circumstances surrounding a work's construction. For example, Averroes explains the presence of contradictions within a single book of Aristotle in terms of a teacher's need to make use of “imprecise, provisional assumptions” in the early stages of instruction, even when he intends to offer a more thorough and consistent discussion subsequently. “For, it is easier to lead the student from what is commonly accepted to what is true than to have him attack the heart of the matter at once.”²¹ Likewise Alfarabi's discussions of multilevel writing as he understood it to be practiced by some of his predecessors, especially Plato and Aristotle, must be examined, because to some extent they reveal his understanding of the character and purpose of philosophic discourse, the advantages of various modes of communication, and the contexts in which each is most fitting or effective. At the same time, it is not safe to assume Alfarabi adopted for his own use every practice he praises on the part of others, even other philosophers. That assumption would entail further, intuitively questionable assumptions, such as that there is a single proper method of philosophic discourse effective as well as desirable, whatever the epoch or the habits and experiences of an audience. The assumption of a single best mode is belied—at the very least, for Alfarabi—by the fact that Plato and Aristotle chose diametrically opposed styles of writing, each of which earned Alfarabi's praise.²² Finally, it is even riskier to extrapolate from an author's approval of multilevel writing in the religious sphere to his own procedure. For, on its face, it would seem that a philosopher could have believed that Scripture admits of levels of interpretation, and viewed this as part of Scripture's excellence, without adopting its method himself. This is especially but not exclusively the case if the writer sees the ultimate end of Scripture as different from the ultimate purpose of a philosopher who writes.

The present chapter will examine Alfarabi's method of writing by first clarifying the concept of multilevel writing. Since the theories of multilevel writing adopted by Muslim and Jewish philosophers were an outgrowth, if not a continuation, of the Greek philosophical tradition, and since the main varieties of multilevel writing are represented in that tradition, it will be pedagogically useful to elaborate the concept through an analysis of two aspects of that tradition: the idea of exoteric writing as it appears in Aristotle's works and the development of the idea in the classical world in response to Aristotle.²³ A historical analysis is also recommended by the fact that important parts of Alfarabi's philosophy are presented through commentaries on Plato and Aristotle, and Alfarabi's understanding of the two earlier philosophers is premised, in part, on the belief that each practiced a species of multilevel writing. This belief of Alfarabi's has occasioned forceful and, at times, empassioned controversy among academics and others.²⁴ Although the analysis that follows will not definitively resolve the question whether Alfarabi's belief about Plato and Aristotle was justified or whether he projected back onto classical philosophic texts a method of writing and interpretation more suited to the medieval philosophic tradition, the results are very suggestive.

The historical analysis of the concept of exoteric writing as it appears in Aristotle and his successors occupies the first section of this chapter. The second section discusses Alfarabi's understanding of the writing of Plato and Aristotle, concentrating on what he sees as their motive for choosing the methods they employed and the general defense he offers for some of their practices. His analysis of multilevel philosophic writing will then be contrasted with his theory of multilevel religious writing. The final section of the chapter will deal with the different approaches to reading and interpreting medieval philosophical texts suggested by the different kinds of multilevel writing.

A. ARISTOTLE'S EXōTERIKOI LOGOI AND THE POST-ARISTOTELIAN TRADITIONS

The distinction between exoteric and esoteric works is not made by Aristotle.²⁵ We can conjecture that the distinction came to be attributed to him because he does speak of exōterikoi logoi (literally: exoteric speeches, arguments, or works); because some of his teachings are difficult to reconcile with others of his teachings; and, in the case of particular versions of the exoteric-esoteric tradition, because of mystical theories prevalent in the centuries after Aristotle's death. Several of the references²⁶ to exōterikoi logoi in Aristotle's extant works follow a pattern: Aristotle is about to embark on a discussion that builds on a distinction or classification that will not itself be examined.²⁷ For example, the definition of virtue presupposes the division of the soul into rational and irrational (Nicomachean Ethics I. 13); the distinction between art and practical wisdom presupposes the distinction between making and doing (Nicomachean Ethics VI. 4); the nature of the best regime depends on the nature of the best life (Politics VII. 1); the number and character of regimes (politeiai) corresponds to the number and character of the possible types of rule (Politics III. 4); and the definition of virtue depends on the division of good things into goods of the soul and goods external to the soul (Eudemian Ethics II. 1).²⁸ In each instance Aristotle alludes to the discussion of an issue contained in the exōterikoi logoi and either explicitly or implicitly finds the earlier discussion adequate to serve as one of the starting points for the inquiry at hand.

Modern interpreters of the allusions to exōterikoi logoi in Aristotle's works can be divided into those who maintain that Aristotle has in mind other works or discussions of his own and those who see some or all of the passages as referring the reader to books or discussions outside the Lyceum.²⁹ In the latter case, exoteric could mean books or discussions of non-Peripatetic philosophers, conversations among educated but not philosophic men, or anything in between.³⁰ Among those who connect the exoterikoi logoi with Aristotle's own writings, opinion appears to be divided between those who see the reference as specifying books with a distinctive style and character (usually popular or nonscientific) and those for whom exoteric refers in a general way to discussions “elsewhere”—i.e., in other places. For example, Thomas Aquinas explains the exōterikoi logoi referred to in Nicomachean Ethics I. 13 as the discussion of the rational and irrational aspects of the soul presented in the De Anima.³¹ In contrast, many modern interpreters view the exōterikoi logoi as referring to Aristotle's popular works in general or the lost dialogues in particular.³² According to Aquinas's interpretation, in contrast, by “exoteric” Aristotle could mean his philosophic writings, and not simply or primarily his popular works.

Those who identify the exoteric discussions with a particular group of Aristotle's works that have not survived—either the dialogues well known in antiquity or the dialogues together with other works written early in Aristotle's career—themselves disagree as to the character of the lost works. Beginning from the observation that the teachings of the lost dialogues and early treatises seem, from the fragments in our possession and reports of their contents, to be at odds with the doctrines of the books in our possession, these commentators disagree as to the reason for the discrepancies. At bottom, the dispute is between those who attribute the discrepancies to a difference in style between the exoteric works and the surviving books and those who discern in the discrepancies a difference in substance. Those who attribute the differences to style argue that the exoteric writings represent a watered-down version of the doctrines or arguments contained in the philosophic works. The prose of the exoteric works, according to this view, was more accessible, less technical, and more polished—if not elegant³³—than the prose in the scientific books we possess, because Aristotle designed them to be read by educated laymen familiar with Platonic philosophy, probably in a popularized version, or by beginners in philosophy.³⁴ In either case the relation of the exoteric works to the nonexoteric ones can be compared to the relation between elementary and advanced courses of study. Those who attribute the differences to substance, on the other hand, usually associate the exoteric works with the early, Platonic period of Aristotle's thought.³⁵

To a large extent the controversy about the meaning of “exoteric” may be due to the vagueness of Aristotle's references to exōterikoi logoi noted above. In the most famous of the references, however, the antecedent is perfectly clear. This occurs in the section of the Physics where Aristotle is beginning to discuss the nature and attributes of time (Physics IV. 10 217b29ff.). “First we would do well [kalōs echei] to go over the perplexing points, using the exōterikoi logoi.” Aristotle proceeds to the kind of dialectical discussion for which he is famous: reviewing the pros and cons of a variety of ways of approaching certain questions presupposed by a basic understanding of time, such as whether time is a being and whether time can be divided into parts.

The purpose of such dialectical discussions in Aristotle's works and their role in his philosophy as a whole has long been a subject of controversy. In the nineteenth century, George Grote connected Aristotle's procedure in Physics IV. 10 with his remarks about dialectic in the beginning of the Topics (I. 2 100b21, 101a25, 34-36, b2) and with the third book of the Metaphysics (III. 1 995a28ff.), arguing that “exoteric” in the Physics refers to the dialectical introduction to philosophy that Aristotle often extols.³⁶ Grote concluded that exoteric means extraneous to philosophy understood as “the didactic or demonstrative march” to truth.³⁷ Although outside philosophy proper and having recourse to popular opinion, such prephilosophic inquiries serve philosophy in that “numerous points are canvassed and few settled; the express purpose being to bring into full daylight the perplexing aspects of each.”³⁸ Thus, according to Grote, the exōterikoi logoi refer not to a substantive doctrine, but to a method—a method based on popular and other opinions, but without itself being popular, because it aims at learning from opinion the defects of opinion.³⁹ Grote's understanding is especially attractive because it is consistent with the circumstance that most of the passages where Aristotle uses the expression exōterikoi logoi refer to questions of definition and classification, i.e., to ideas inherently incapable of demonstration. A few more recent commentators have emphasized Aristotle's dialectical discussions of first principles and basic concepts as an integral part of the philosophic attempt to verify primary things, which, by their nature, are not susceptible of demonstrative proof.⁴⁰

Many of the modern interpretations of Aristotle's exōterikoi logoi can be traced to ancient sources or have counterparts in those sources. Cicero (first century B.C.) identifies Aristotle's exoteric works with his writings in a popular style, and contrasts them with those that are “more carefully wrought” (Rackham) and in the form of notes or comments.⁴¹ This, Cicero acknowledges, sometimes gives Aristotle⁴² the appearance of inconsistency, which is in fact illusory (De Finibus V. v. 12). Since the term “esoteric” is missing from the passage in Cicero, although the term “exoteric” appears and is written in Greek, it seems that “esoteric” was not at that time used to describe either Aristotle's works or the character of his discussions.

The earliest clear evidence of the view that Aristotle's works contain secret doctrines occurs in Plutarch (first and second centuries A.D.).⁴³ The context is an account of the decision by Philip of Macedonia to entrust Alexander's education to Aristotle. Three sections of the account are relevant to the question at hand: Plutarch's general remarks on the nature of Alexander's education; the text of a letter that Alexander is supposed to have sent Aristotle, rebuking him for committing his philosophy to writing; and a paraphrase of Aristotle's reply (Alexander VII. 3-5).⁴⁴ The passage describing Alexander's education is as follows: “It would appear, moreover, that Alexander not only received from his master his ethical and political doctrines, but also participated in those secret and more profound teachings [tōn apporrētōn kai bathuterōn didaskaliōn] which philosophers designate by the special terms ‘acroamatic’ and ‘epoptic,’ and do not impart to many” (Perrin) (Alexander VII. 3).

Although the terms “exoteric” and “esoteric” are missing, the implication of the contrast between the ethical and political doctrines, on the one hand, and the acroamatic teachings, on the other, is that certain doctrines, especially those of practical philosophy, are imparted to people in general, whereas other philosophic teachings are reserved for the few. In the same period as Plutarch, Gellius explicitly contrasts Aristotle's exoteric and acroamatic works—a distinction repeated by several later writers. As was the case with Plutarch, the use of “exoteric” by Gellius seems to be associated with particular subject matters: the exoteric works are devoted to rhetorical exercises, argumentative ability,⁴⁵ and politics; in the acroamatic, “a more profound and recondite [remotior subtiliorque] philosophy was discussed, which related to the contemplation of nature and dialectic discussions” (Rolfe) (Attic Nights XX. v. 2-3).

On the basis of Plutarch and Gellius we can say that by the second century A.D. the distinction between exoteric and nonexoteric was connected with the division of philosophy into theoretical and practical, with specific logical topics treated in one or the other manner. Practical philosophy and certain areas of logic were considered exoteric, or at least it was thought that the doctrines associated with them should be communicated in a popular way. Theoretical philosophy and other aspects of logic, it was believed, should not be expressed with an eye to the general public. Like Plutarch and Gellius, Cicero linked practical philosophy to the exoteric-nonexoteric distinction, although with a different effect. In De Finibus V. v. 12 the study of the highest human good is presented as having been communicated by the Peripatetics in two ways, one popular or exoteric and the other “more carefully wrought.” The implication is that the other subdivisions of philosophy and logic were always communicated in the latter, less popular fashion. In the case of Plutarch and Gellius, particular subject matters invite distinct modes of exposition. Their view makes sense if practical philosophy is intrinsically less philosophic than theoretical philosophy or if the practical end—action with a view to happiness or living well—dictates that the mode of exposition be popular. The Ciceronian account of the practice of the Peripatetics may suggest that in their view neither of the two aspects of practical philosophy could be simply subordinated to the other, i.e., that theory and practice were both necessary to achieve practical philosophy's larger purpose.

The major difference between Cicero's account and that of the two later authors is that for the later authors the doctrines of the acroamatic works are clearly depicted as secret or hidden. Plutarch's account of the correspondence between Alexander and Aristotle reinforces the view that his reference to acroamatic teachings is to secret doctrines. According to Plutarch, in Alexander's letter to Aristotle, the Greek conqueror reproaches the philosopher for having made his acroamatic teachings (tous akroamatikous tōn logōn) public, on the ground that once they become common property (koinoi), Alexander will no longer be superior to other men. The conqueror adds that as far as he is concerned, superiority means superiority in the highest things (ta arista) and not merely superior power. Plutarch depicts the philosopher as replying that Alexander need not worry: “Know that [the akroamatikoi logoi] have been both made public and not made public [kai ekdedomenous kai mē ekdedomenous], since they are only comprehensible to those who have heard me.”⁴⁶ There follows the comment that “in truth his treatise on metaphysics is of no use for those who would either teach or learn the science, but is written as a memorandum [hupodeigma] for those already trained therein” (Perrin) (Alexander VII. 4-5).

The usual interpretation of works “at once made public and not made public” has been that they contain a hidden as well as a literal meaning. George Boas and Ingemar Düring argue against the usual interpretation on the grounds that Plutarch's final comment about Aristotle's Metaphysics as a memorandum constitutes an alternative and better explanation.⁴⁷ For a memorandum written for initiates, if both technical and elliptical, might appear both to convey and to suppress information. It would provide key words and propositions to people well acquainted with the subject and be meaningless or appear garbled to all others.⁴⁸ In either case, Plutarch represents a departure from his predecessors in that “exoteric” no longer refers to a distinct group of writings. Rather, in Plutarch can be seen the first expression of a notion that became axiomatic in medieval times, namely, that one and the same work is both exoteric and esoteric, depending on how it is read and understood.

In a passage in the Vitarum auctio of the second-century satirist Lucian, the doctrine of a secret teaching is associated with Aristotle in an unambiguous way. The setting is a parody of a slave market where philosophies are for sale in the form of handsome young men. A Pythagorean, a Cynic, a Heraclitean, a Socratic, and a Stoic have so far commanded good prices. The owner has a Peripatetic brought out. The hawker cries to the buyers gathered round, “Come and buy the one with the most understanding, who knows absolutely everything!” A potential buyer inquires, “What's he like?” “Moderate, decent [epieikēs], and adaptable to life,” comes the reply. “Moreover,” boasts the hawker, “he's double.” “What do you mean?” asks the perplexed buyer. “He appears to be one thing on the outside and another on the inside. So if you buy him, remember to call the one ‘exoteric’ and the other ‘esoteric’” (Vitarum auctio XXVI). This seems to be the earliest use of “esoteric” to designate the nonexoteric Aristotle. Yet because the work is a satire and because throughout Lucian chooses only the most obvious doctrines of each school to ridicule, it would appear that Aristotle was already well known as a purveyor of secret doctrines. Despite Boas's and Düring's arguments against counting the authors mentioned thus far as witnesses for the prevalence of this view, the reference in Lucian suggests that the doctrine of a secret teaching was established by the second century.

Among the late Greek commentators, Themistius, Olympiodorus, Elias, and Simplicius all raise the question of a secret teaching in connection with Aristotle. Of these, Themistius believes that the esoteric level is not only obscure, but deliberately so.⁴⁹ According to Olympiodorus and Elias, Alexander of Aphrodisias claimed that the esoteric teaching is not merely different from the exoteric because it is more advanced; rather, the exoteric teaching is simply false and the esoteric teaching is the simple truth.⁵⁰ Ammonius mentions this interpretation of the two levels in Aristotle's works without naming Alexander as the source.⁵¹ All three authors make clear their disagreement with Alexander's interpretation of Aristotle and put forth the view that exoteric works differ not in doctrine, but in their simplified exposition and more eloquent style. Düring conjectures that the view these commentators attribute to Alexander stems from a misunderstanding of something Alexander is likely to have said, namely, that the teaching in Aristotle's exoteric works cannot be equated with Aristotle's own views. The hypothetical statement, Düring argues, could have been prompted by Alexander's belief in the elementary-advanced version of the exoteric-esoteric distinction as easily as by the secret teaching version. Because we do not possess the disputed statement of Alexander, because no other passages in Alexander repeat this point of view, and because he holds that the attribution of a secret doctrine to Aristotle is nonsense, Düring concludes that these late Greek testimonials to Alexander's understanding of Aristotle's method of writing should be discounted.⁵²

B. ALFARABI AND PHILOSOPHIC MULTILEVEL WRITING

In Alfarabi's works the division made in antiquity between a philosopher's popular and philosophic treatises is superseded by the theory, foreshadowed in Plutarch of single works admitting of levels of interpretation. There is, however, a unique reference in Alfarabi's corpus to separate works that Aristotle apparently wrote with the sole purpose of popularizing philosophical ideas. The contents of these books Alfarabi labels “exoteric philosophy” (al-falsafah al-khārijah) (Jadal 37:6-7/203r19-v2). It appears from the passage that Alfarabi does not know these books themselves; he knows of them from references to them in the works of Aristotle or others that he did possess. However, no mention is made of special books devoted to exoteric philosophy in Falsafat Arisṭūṭālīs, which presents itself as a description of the totality of Aristotle's philosophy (Arisṭūṭālīs 59:2-3). This omission can be explained by the fact that Alfarabi considered exoteric philosophy to be primarily a species of religion rather than a species of philosophy (see Sa‘ādah 90:10-21/40:9-19, 94:7-10/44:6-9).⁵³ Like other kinds of religion, it instills in people beliefs about subjects for which wisdom or science is available. Exoteric philosophy differs from other species of religion in that it provides true opinions, whereas most religions contain a mixture of images and true opinions or even images and generally accepted but false opinions. To achieve their ends, however, both employ persuasive and imaginative methods of instruction (Sa‘ādah 90:10-14/40:9-13, 90:19-21/40:17-19). Mention of separate books devoted to exoteric philosophy is also missing in the list of Aristotle's works in Qabl Ta‘allum al-Falsafah (Ta‘allum 50:16-52:15). This omission cannot be explained in the same way as the omission in Falsafat Arisṭūṭālīs, however, since the subject of the passage is Aristotle's books, not the parts of his philosophy, and it appears that Al-farabi intended the list to be exhaustive.

The dominant understanding of multilevel writing for Alfarabi is thus of works at once made public and not made public. Within this genre he distinguishes the method of Plato from the method of Aristotle. He conveys the contrast between them most clearly in the Jam‘, a work devoted to harmonizing the apparent conflicts between Platonic and Aristotelian teachings.⁵⁴ Plato, we are told, eschewed committing what he knew to writing on the ground that the proper place for knowledge is not in books, but in “pure hearts and pleasing minds.” Then, as Plato's knowledge increased, he became afraid of forgetting some of his discoveries, which would be increasingly difficult to recover as his wisdom grew.⁵⁵ So he decided to put what he knew in writing in the form of riddles and enigmas, in order that only worthy people would be able to detect his meaning and only as a result of study and effort (Jam‘ 5:23-6:5). Aristotle, apparently for the same reason as Plato, chose a deceptively straightforward style that served to conceal the subtlety and controversial character of his thought (Jam‘ 6:5-12). After an enumeration of some of the devices Aristotle used to achieve this end, Alfarabi concludes by quoting a letter Aristotle allegedly wrote in response to Plato's accusation that Aristotle had made the sciences public. “Although I committed these sciences and the wisdoms contained therein to writing, I arranged them in such a way that only people with training will understand them, and I expressed them in such a way that only experts will grasp them” (Jam‘ 7:6-8).⁵⁶ Alfarabi's partial catalogue of Aristotelian devices includes the following:

For example, in many of the syllogisms he advances for natural, divine, and ethical [subjects], his arguments omit the necessary premise. The commentators have pointed out where these occur. Alternatively, he omits [the name of] many of the authorities [to whom he refers]; or he omits one member of a pair, and limits himself to one member; … Alternatively, he mentions the two premises of some syllogism and follows them with the conclusion from another, or mentions the two premises of a syllogism and follows them with the conclusion from the necessary concomitants of these premises. … Alternatively, he enumerates the individual instances of something obvious at great length to display his unstinting, strenuous effort to be thorough, while he passes over something obscure without discussing it at length or giving it its due. Alternatively, he arranges, orders, and organizes the contents of his scientific books in a way that makes one suppose that this is part of his immutable nature. But if one contemplates his letters, one will find the discussions there constructed and arranged according to different [types of] format and order than are in those [scientific] books.

(Jam‘ 6:10-7:3)

In the introduction to his commentary on Plato's Laws, Talkhīs Nawāmīs Aflāṭun, Alfarabi adds to the above description of Plato's writing the observation that Plato frequently combined unmistakable riddles and enigmas with simple, straightforward declarations, knowing the latter would be buried among and thus not easily distinguishable from the overtly cryptic remarks (Nawāmīs 4:10-16).⁵⁷ In short, according to Alfarabi Plato's writing is for the most part overtly ambiguous, although at times unexpectedly clear, whereas Aristotle's works display an untroubled surface that masks a complex interior. The works of both, then, conceal as they reveal, but the presence of concealment is revealed by Plato and concealed by Aristotle. As far as Aristotle's surviving works are concerned, therefore, Alfarabi saw as provisional the obvious and often conventional doctrines elaborated and apparently endorsed by the author.

Alfarabi refers to a number of reasons why the two Greek philosophers wrote works to be interpreted in different ways on different levels. His commentary on Plato's Laws begins with a reference to the philosopher's concern for physical safety. Alfarabi makes this point by telling the story of a pious ascetic who wanted to escape from a tyrannical ruler seeking to arrest him. The ascetic was able to flee by disguising himself as a drunken vagabond and appearing at the city's gate singing and playing a musical instrument. In response to the gatekeeper's query, he said in a joking manner, “I am so and so, the ascetic.” Feeling sure that nothing could be further from the truth, the gatekeeper let him pass (Nawāmīs 4:1-9).

Alfarabi introduces his commentary to Plato's Laws with this tale, he informs the reader, because the ascetic's behavior provides an analogy for Plato's method of writing. Fearing that science would “fall into the hands of those who do not deserve it and be deformed, or fall into the hands of someone ignorant of its worth or who might use it improperly,” Plato decided to write in the indirect manner previously described (Nawāmīs 4:10-21). The story of the ascetic is an obvious instance of persecution. However, Plato's fears for his writing, as reported by Alfarabi, are not necessarily connected with a concern for his own safety. Without the tale of the ascetic we might interpret Plato's twofold concern as directed toward the integrity of science or philosophy and the well-being of “insufficiently gifted or uneducated” readers who might suffer from an exposure to philosophy. With the tale as an introduction, the implication is that Plato feared that, as a result of their failure to understand philosophy's purpose and worth, readers reared with nonphilosophic beliefs, or the authorities charged with protecting such beliefs, would be likely to persecute philosophers.⁵⁸

Alfarabi elaborates this point in his commentary on Aristotle's Topics, where he claims that a philosopher protects himself from the multitude (al-jamhūr) by disseminating some of the discoveries of philosophy in popular form, since people tend to despise what they find strange (Jadal 37:9-11/203v4-6, 37:18-38:2/203v17-19). At first the impulse to instruct the multitude is presented as part of the philosopher's natural affection for his fellow men; but Alfarabi soon indicates that popular philosophy (“fourth philosophy”) is part of a trade of a portion of the good (al-khayr) for a portion of the goods (al-khayrāt) (Jadal 36:16-37:2/203r8-14). Although Alfarabi sees an increase in the philosophers' safety as one consequence of thus educating their fellow men, the philosophers' turn toward popular philosophy cannot simply be reduced to a strategy for ensuring the philosophers' preservation. For elsewhere Alfarabi suggests there is a spontaneous antipathy between philosophers and the adherents of religion, which is likely to be accentuated by the former meddling in the affairs of the latter (Hurūf No. 149, 155:1-18), and he warns that a truly virtuous person increases the risk of being harmed by attempting to reform the way of life of those around him (Siyāsah 101:14-16). Thus, the reasons Alfarabi advances for the philosophers' commitment to enlightening nonphilosophers are not entirely consistent.

Protecting oneself from possible or actual hostility on the part of those whose opinions or ways of life are threatened by the activity of philosophy is one motivation for constructing the surface of multilevel texts so as to disseminate certain kinds of myths among nonphilosophic readers. Another version of the identification of exoteric teachings with politically salutary myths stresses the moral and, possibly, the intellectual development of nonphilosophic readers. According to this view the conclusions of philosophic inquiry can harm laymen because they may undermine the layman's current beliefs without furnishing a replacement. The immediate effect of publicizing philosophic discoveries can be destructive because, in the absence of the reasoning on which philosophic insights are based and without the commitment to a life of inquiry, a person could realize the falseness of his existing beliefs, yet fail to appreciate the significance of philosophic truths. To take a nonrandom example, it is far easier to grasp that philosophic doctrines about the soul prove conventional notions about heaven and hell false than to understand how a good life, or virtuous activity, constitutes the soul's reward and their opposites its punishment.

This concern with the well-being of nonphilosophers is to be found in Alfarabi's works as well. According to Al-Madīnah al-Fāḍilah a person confused after realizing the untenability of his fundamental beliefs may, as a result, seek and discover the truth, succumb to moral hedonism, or retreat into skepticism (Madīnah 280:13-284:12/70:16-71:21). Intellectual capacity and character together appear to determine the direction the life of such a person will take. On the basis of this passage, then, it is likely that Alfarabi had the possible intellectual and moral corruption of readers in mind when he offered as one reason for Aristotle's decision to write on several levels “clearing up any doubts that the student's nature is fit for instruction” (Ta‘allum 54:5-6). Similarly, in the passage in the Jam‘ on the writing of Plato and Aristotle, Plato decides to commit philosophy to writing for his own edification; yet he favors the particular mode he chooses in order to hide his real thoughts from people at large and reveal them to the industrious and deserving (Jam‘ 6:1-5). There is thus a gap between Alfarabi's Plato's avowed purpose and the strategy he eventually adopts. For Plato could have reinforced his flagging memory without laboring to construct such carefully crafted works; he could have jogged his memory by making suitable abbreviated notes, without fear of enlightening anyone else should the notes fall into another's hands. Alfarabi's account of Plato's fears explains the form his writing took once the decision was made to construct works which would benefit a few, while harming no one. It does not explain the decision to enlighten the deserving.

In Qabl Ta‘allum al-Falsafah Alfarabi gives as one of Aristotle's reasons for writing in a concealed fashion the need to train people's minds for the rigors of inquiry (li-yarūḍ al-fikr bi'l-ta‘ab fī al-ṭalab) (Ta‘allum 54:7). Alfarabi expands on this theme in two places, the Kitāb al-Jadal and his summary of Aristotle's philosophy, Falsafat Arisṭūṭālīs. According to both works, the logical art that Aristotle had in mind for the mental training of the sort needed by an investigator is dialectic, that is, a method of argument usually but not necessarily employed in debates between two people that reasons deductively and with syllogistic necessity from generally accepted premises or inductively to generally accepted universal statements (Jadal 13:5-8/187v7-11, 14:9-10/188r12-14, 25:13-15/195v7-10, 97:4/241v15-16).⁵⁹ The art of dialectic so understood

trains a person and prepares his mind for the certain sciences. It does this by habituating a person to investigation; by making known to him how an investigation takes place and how things should be ordered and statements arranged in an investigation so that he attacks the problem; by enabling his mind to grasp the middle term [of syllogisms] quickly; by making him capable of grasping quickly the syllogisms for any problem posed; by imparting to him the ability to oppose every opinion he hears or hears about and to grasp quickly the points that can be opposed in every statement advanced.⁶⁰ Thus, [the art of dialectic] habituates a person not to be persuaded by unexamined opinion, the dictates of the first thoughts that cross one's mind, first impressions, and a quick inspection, without close study and careful examination.

(Jadal 29:18-30:3/198r17-v6)

So understood, dialectical training is not merely useful for philosophy; without training of this kind and the habit of investigation it creates, human beings cannot attain what is real (al-ḥaqq) and philosophy (Jadal 31:1-3/199r11-13, see 30:12-13/198v17-18). This is the reason Aristotle begins the treatment of every subject with dialectical arguments and with a dialectical investigation (Jadal 31:3-6/199r13-17).⁶¹ And this is why Plato has Socrates advise Parmenides to train himself in dialectic as well as why Socrates proceeds to discourse dialectically with him (Jadal 31:6-12/199r17-v6). Training people's minds for inquiry thus means giving them experience in dialectical investigation so that they can then carry out dialectical investigations on their own.

Alfarabi explains at length in Kitāb al-Jadal the reason dialectical investigation must precede scientific or demonstrative investigation, that is, the reason dialectical investigation is part of philosophy proper. All objects of investigation, except the subjects of the mathematical sciences,⁶² have a tendency to lead those who study them astray, because their material component causes them to exist with a wide range of frequently contradictory attributes (Jadal 32:11-34:4/200r9-201r16). One of the tasks of the philosopher is to probe such things characterized by contraries to the point of distinguishing their essential from their accidental properties. Because it is the character of dialectic to reason to either or both of a pair of contradictories, dialectic can deal with the contrariness inherent in nature in a way that demonstration cannot. With the rules of demonstration a philosopher can evaluate dialectical statements and proofs, thereby eliminating some of them and exposing their falseness. But the demonstrative rules cannot be applied until the contrariety inherent in the objects of investigation is itself exposed (Jadal 34:4-17/201r16-v14, see 31:14-32:10/199v12-200r9). Thus, Alfarabi also refers to dialectical investigation as the method of creating doubts (ṭarīq al-tashkīk) (Jadal 31:11/199v5, see 34:4-7/201r16-19, 34:17-19/201v14-16).

In Falsafat Arisṭūṭālīs Alfarabi develops the case for dialectical inquiry as a partner of demonstration in philosophic investigation. The theory of demonstration, as expounded in the Posterior Analytics, describes the requirements of scientific reasoning, but it fails to explain how one arrives at a demonstrative syllogism in the first place. Figuring out what demonstration is needed for a specific problem is “extremely difficult,” because it involves hitting upon the appropriate middle term (Arisṭūṭālīs 78:12-15). Dialectic can facilitate this process by making a person adept at constructing dialectical syllogisms, which can then be evaluated in light of the rules governing demonstrations, as set forth in the Posterior Analytics. What survives this test, whether propositions or arguments, can be reclassified from dialectical to scientific and incorporated into the certain sciences (Arisṭūṭālīs 78:6-12). In the case of deductive reasoning, then, dialectic's utility comes from providing an indirect access to demonstrative truths. The dialectician turned investigator will have a large pool of arguments from which to extract demonstrations or which can be transformed into demonstrations through revision and refinement.

Much of the analysis of dialectic's use for philosophy in Kitāb al-Jadal deals with the special problem of verifying self-evident or primary propositions—premises which by their very nature can never be demonstrated or proved syllogistically, but must be grasped through insight (baṣīrah) (Jadal 30:16-19/199r3-7). Difficulties arise because many, if not most, of the beliefs people hold as a result of childhood rearing appear to be of this sort. In point of fact, what usually happens is that people mistake generalities for universals and partial truths for pure ones (Jadal 21:18-22:6/193r11-19). Primary truths must, therefore, be screened to verify their status as primary truths. And dialectic is the only logical method equipped to differentiate truth from falsehood at the level of primary propositions (Jadal 30:19-31:3/199r7-13, 34:4-17/201r16-v14). It is thus indispensable for establishing the foundations of all thought as well as the axioms of particular sciences (see Jadal 35:14-21/202r17-v7).

The theme, then, of the passage in Kitāb al-Jadal just discussed is mainly the elusiveness of certainty about primary truths rather than the elusiveness of the truths themselves. Primary truths are first grasped as generally accepted opinions or received on the authority of others (Jadal 19:6-13/191r17-v7, 28:9-11/197r19-v3, 31:14-16/199v8-10). Alfarabi's references to insight suggest that recognition of these truths occurs a second time through a more rigorous, yet still nondiscursive, process. Their precise origin, he concedes in his commentary on Aristotle's Posterior Analytics, is one of the enduring controversies of philosophy and science (Burhān 23:1-25:9/138v12-140r4). However, to use primary truths with confidence, we do not need to understand their origin, since we can speak with some clarity about the method of verifying such insights, once grasped. Dialectic, in short, is indispensable because of the critical faculties it develops; and those critical faculties, in turn, are necessary for testing and thereby verifying primary truths not amenable to proof.⁶³

To acquire the art of dialectic, one must engage in dialectical question and answer. Only by continued practice can a person acquire the mental agility philosophical investigation requires (Jadal 39:11-12/204r12-13). One must practice by staging debates, even with partners who prize victory over learning, since the effect of competitiveness can be improved argumentative skills appropriate for serious inquiry (Jadal 39:3-40:8/204r1-v4). It is true that dialectical investigation can be carried out in isolation, but the enterprise is easier when people pool their resources (Jadal 45:6-10/207v9-14). Perhaps Alfarabi's final teaching is that expertise in dialectical debate with others is not only the best training for philosophy; properly understood, it exists for the sake of private philosophic inquiry (Arisṭūṭālīs 78:15-79:2). In any event, dialectic so conceived requires two participants with approximately the same talents and experience; through parity, the strengths of each will contribute to the improvement of the other (Jadal 40:8-15/204v4-13). When live confrontations are not possible, however, books can supply the requisite training in their stead (Jadal 25:13-16/195v7-11).

There are, then, nondefensive, philosophic reasons for choosing to write dialectically.⁶⁴ In light of Alfarabi's remarks about the nonphilosophic character of received opinions, however true, and his insistence on personal insight as a condition of philosophic certainty, an indirect mode of communication is dictated by the philosopher's responsibility to his readers. Were textbooks in geometry to serve as the model for philosophical treatises, they would create belief where they should promote understanding. The problem for a philosopher, given Alfarabi's views on education, is to construct books that provide the reader with practice in the dialectic presupposed by success in philosophical investigation. In the best possible case, a book will be so constructed that it can serve as a partner in inquiry for readers with a wide range of backgrounds or at different stages in their pursuit of specific subjects. Multilevel writing which “conceals as it reveals and reveals as it conceals” performs this service. In the last analysis, it originates less in the desire to exclude than from concern for the intellectual development of those who are inclined toward the truth.

C. ALFARABI AND RELIGIOUS MULTILEVEL WRITING

It appears that the idea of exoteric doctrines with esoteric meanings first arose in the Islamic world in the religious sphere, both among those engaged in Qur'anic exegesis and among Islamic sects with mystical tendencies. Although the view of the Qur'an as having one meaning when taken literally and other meanings when subjected to interpretation was in no way universal, it was widely endorsed by philosophers and certain schools of theology and jurisprudence. Of course interpretation meant different things to each of these groups; and their efforts to penetrate to the deepest level of understanding were, as a result, made in accordance with different canons of interpretation.

To Alfarabi can be traced one of the major philosophic versions of the view that the Qur'an is an exoteric work, namely, the doctrine that religion is an imitation of philosophy.

Both comprise the same subjects and both give an account of the ultimate principles of the beings. For both supply knowledge about the first principle and cause of the beings, and both give an account of the ultimate end for the sake of which man is made—that is, supreme happiness—and the ultimate end of every one of the other beings. In everything of which philosophy gives an account based on intellectual perception or conception, religion gives an account based on imagination. In everything demonstrated by philosophy, religion employs persuasion.

(Mahdi) (Sa‘ādah 90:15-21/40:13-19)⁶⁵

Religion is an imitation of philosophy, in other words, because both seek to make known the same things and because the account of these things in philosophy is superior to the parallel account in religion. Because philosophic insights are often beyond the grasp of the layman, religion conveys those insights using vocabulary and concepts easy to understand or analogies with objects and events taken from ordinary men's everyday experiences. Religious doctrines, then, are imitations of philosophic truths designed to meet the needs while conforming to the intellectual abilities of people in general. Religion is popular in this sense.

The doctrine that religion is an imitation of philosophy entails viewing religious texts as exoteric statements because it assumes that philosophic truths provided the original insights which the prophet or founder of a religion then recast or else that awareness of philosophical truths is the ultimate goal of religious teachings. This raises the thorny problem of the extent to which and the manner in which imitations can be said to promote a grasp of what, in principle, they imitate. The notion of imitation implies a limit on the pool of images available for conveying a particular truth. It precludes the possibility that images—whether religious teachings or the literal meaning of multilevel philosophic texts—are simply arbitrary. There is a range of possibilities along a continuum, some “closer” and some “further” from what is being imitated (Siyāsah 85:14-18). The closer the imitation, the greater its resemblance to the original. The reflection of a person in a pool of water is in this sense closer to the living person than is the reflection in water of a statue of the person (Siyāsah 85:6-11). At the same time, the notion of imitation entails that no image—whether in poetic or prose form—is true in the strict sense that the original is true. Statements about the first cause of the universe illustrate this limitation of imitations well. We can say that the first cause is one or a unity in order better to understand something of its nature, although the first cause is not one in the same sense as a physical object is one. The attribute “one” in connection with incorporeal entities is an image, a reflection of the truth, without itself being true.

The thesis that imitations can be evaluated on the basis of their ability to lead people to the original ideas which they represent is also open to the objection that in some works Alfarabi presents proximity to truth as the goal of imitation, while elsewhere proximity to truth is only one, and not the decisive, variable determining the suitability of images. According to the passage in Taḥṣīl al-Sa‘ādah which advances the thesis that religion is an imitation of philosophy, religion “attempts to bring the similitudes of [the things it imitates] as close as possible to their essences” (Mahdi) (Sa‘ādah 91:11/41:10-11). The images thus “represent the theoretical things that have been demonstrated in the theoretical sciences” (Sa‘ādah 94:3-4/44:2-3, Millah 47:6-7). In Fusūl Muntaza‘ah Alfarabi lists three classes of praiseworthy imaginative impressions—those that promote the well-being of the rational, the irascible, and the appetitive faculties of the soul (Fuṣūl No. 56, 64:5-65:5). According to this classification, then, images can contribute to the development of men's minds as well as to their character; and they seem to be able to affect the rational faculty directly, not merely as a byproduct of improving men's morals. Finally, Alfarabi's theory of prophecy as some kind of interaction between the agent intellect and imagination may be seen as supporting a close connection between the original theoretical discoveries and the subsequent imaginative recasting of them. Something of this sort would seem dictated by the crucial distinction between a powerful imagination independent of rational control and a powerful imagination under the direction of reason.

The teaching of Al-Siyāsah al-Madaniyyah, on the other hand, is that the perfection of an imaginative representation is distinct from its truth content or proximity to the original (Siyāsah 86:11-12).

Now if imitations are made equally well or are equal in having few or hidden controversial points, one can use all or any one of them indifferently. But if some are better than others, one should choose the imitations that have been made most perfectly and that have no controversial points at all or else have few or hidden controversial points. After that, one should choose the imitations closest to the truth [al-ḥaqīqah] and discard the rest.

(Siyāsah 86:17-87:4)

The implication is that the primary index of excellence in imitations is whether they “work,” i.e., whether they generate an image that grips the audience; and this, in turn, depends largely on how familiar the material used to create the images is to a particular audience (Siyāsah 85:17-86:4). When one must choose, therefore, between an image that is truly fitting and one that is effective, the former consideration must bow to the latter.

Examined in light of Al-Siyāsah al-Madaniyyah, the doctrine that religion is an imitation of philosophy turns out to be more ambiguous than first appeared. Even Taḥṣīl al-Sa‘ādah, where the doctrine is prominent and the integrity of religious imagery is most clearly proclaimed, offers by way of illustration images that are exceedingly remote from the things they are designed to imitate. Intelligibles are imitated by sensibles, metaphysical principles are represented by political counterparts, the ontological hierarchy among the beings is depicted by means of temporal and spatial sequences, and the classes of supreme happiness are portrayed by means of generally accepted and often illusory goods (Sa‘ādah 90:22-91:11/41:1-10). The Farabian dictum, with which the passage ends, that philosophy is prior to religion in time (Sa‘ādah 91:13/41:12), is thus meant to suggest philosophy's priority in nature. In short, the gulf separating the corporeal from the incorporeal epitomizes the limits of imaginative fidelity. To say that illusory goods are an image of the true human end, as demonstrated in theoretical science, is tantamount to admitting the antithetical character of religion and philosophy. Alfarabi comes closest to making this opposition explicit in Kitāb al-Hurūf, in a section devoted to the relationship between religion and philosophy. The section is especially important because it occurs in a work that defends the thesis that religion is an imitation of philosophy and describes the events that will occur when the religion in question is actually based on “perfect philosophy” (Hurūf No. 149, 155:1).⁶⁶ According to Alfarabi, were such a religion to be transferred from the nation in which it arose to another nation, without the second nation being informed that its religion is an imitation of philosophy, two situations are likely to occur. The adherents of that religion will be hostile to true philosophy should the latter ever reach their nation, and the followers of true philosophy will be hostile to the adherents of that religion (Hurūf 155:1-10). If the relationship between that religion and philosophy were then made public (apparently by the founders of the religion in the first nation), the followers of philosophy would cease being hostile to the adherents of religion, whereas the adherents of religion would continue to oppose true philosophy (Hurūf 155:10-11). The conclusion is inescapable that because of the nature of imitation and imagery, religious doctrines, even in the best case, will be so different from philosophic teachings that they will be perceived by the philosophers themselves as antagonistic. To apply the principles for the construction of images suggested in Taḥṣīl al-Sa‘ādah, we can conjecture that this genetic tale in Kitāb al-Hurūf was intended to portray the natural opposition between religion and philosophy in graphic terms. If so, the gulf separating philosophic insights and the best imaginative reworking of them necessitates as one of its practical consequences an adversary relationship between religion and philosophy.

Why, then, does Alfarabi assert and develop the view that religion is an imitation of philosophy? The remainder of the passage in Kitāb al-Hurūf just summarized provides one explanation. When philosophy is perceived as an enemy, the adherents of religion are likely to persecute philosophers. When forced to defend themselves, Alfarabi advises, philosophers should avoid attacking religion itself; they should confine their attack to the specific religious doctrine that philosophy is the enemy of religion. The view that religion is an imitation of philosophy, in other words, is designed to counter religious attacks on philosophy and philosophers without undermining religious teachings in general. As far as Kitāb al-Hurūf is concerned, philosophic claims about the affinity between religion and philosophy appear to be practical and defensive in origin.

The relationship between philosophy and religion is somewhat different in the case of one particular species of religion, namely, the one Alfarabi calls exoteric philosophy. As was noted above, exoteric philosophy is the religion of nonphilosophers who adhere to philosophic doctrines as a result of rhetorical persuasion.⁶⁷ Those who adhere to exoteric philosophy thus believe true opinions, in contrast to the members of other religions, who believe in images of philosophic teachings or a mixture of opinions and images. Alfarabi's purpose in referring to exoteric philosophy as a religion appears to be his desire to emphasize that from the perspective of philosophy, true opinion is opinion first and foremost and truth only secondarily and accidentally (see Jadal 28:14/197v6-7). This position is a consequence of his understanding that truth must be discovered to be possessed, whether through reasoning or personal insight (baṣīrah nafsih) (Jadal 30:16-17/199r3-4, Millah 46:18-19). This is also the teaching of Al-Madīnah al-Fāḍilah, where Alfarabi first distinguishes three modes of grasping the core beliefs shared by all members of a political community of excellence: recognizing them as conclusions of demonstrations or through one's own insights, recognizing them as they really are but on the authority of others, and recognizing them by means of similitudes instead of as they really are (Madīnah 278:8-14/69:19-70:3). However, in the next paragraph he refers to these three modes as “two forms of knowledge” (ma‘rifatān)—that of the wise man and that of the believer in similitudes (Madīnah 278:14-280:1/70:3-6). True opinions are, then, true, but those who possess such opinions must in the last analysis be considered believers and not wise men.

D. CONCLUSION

The idea of multilevel writing itself admits of levels of interpretation, depending on an author's ultimate purpose and the means employed to achieve the author's ends. Religious multilevel writing has as its primary purpose instilling in ordinary readers a set of beliefs that will shape their view of the world and man's place in it. It effects this through a variety of rhetorical and poetical devices, such as relying on analogies, whether appropriate or not; bypassing argument altogether when images are just as effective; and simplifying issues, even at the risk of distortion. The author of religious multilevel writing may intend to manipulate the believers he creates to further his own selfish ends; but he is just as likely to intend his handiwork to improve the way of life of his fellow men. Whatever the ultimate purpose, the hallmark of the surface of this kind of writing is the creation of a feeling of satisfaction or confidence in people (sukūn al-nafs) about their opinions and beliefs.

As Alfarabi depicts them, philosophers who write have several goals. They seek to secure their own place in the community, improve if not reform the way of life there, and, above all, contribute actively to the intellectual development of those who are especially gifted. The first two activities are carried out by persuasive methods, often in much the same way as is done by the practitioners of religious multilevel writing. The last of these three activities, on the other hand, is possible only by means of the method of dialectic. The hallmark of the surface of this aspect of philosophic writing is reasoning on the basis of generally accepted opinions to both of a pair of contradictories. Such reasoning has a twofold effect. As noted above, it provides access to a grasp of the essences of things—in some cases, the only access, given the special obstacles to knowledge posed by the contrariness inherent in nature and the indemonstrableness of primary premises. Further, in consequence of furnishing equally persuasive accounts of opposing points of view, the art of dialectic destroys people's confidence in their own beliefs by heightening their awareness of the plausibility of alternatives. The second effect of philosophic multilevel writing, then, is creating perplexity (Jadal 34:14-17/201v10-14, 22:15-16/193v12-14, see 31:11/199v5, 34:5/201r17, 34:19/201v15-16).

To ascertain the character of multilevel texts, it is sometimes necessary to do more than identify the character of isolated arguments. For on occasion the art of dialectic will employ the same devices as are useful to the art of rhetoric, namely, in those cases where specific arguments or modes of argument are at once dialectical and rhetorical. Since the immediate objective of each art is to persuade, their distinctiveness will in those instances derive from the use to which persuasion is put. Religious multilevel writing persuades in order to soothe; dialectical philosophic writing persuades in order to unsettle. Using persuasive arguments, religious multilevel writing tries to present the reader with an apparent demonstration; in contrast, through carefully juxtaposed persuasive arguments, philosophic multilevel writing seeks to guide readers to their own discovery of the appropriate proofs. In short, the ultimate goal of philosophic multilevel writing is to provide those who wish to know with the passion, the tools, and the awareness necessary to engage in the arduous search for truth. For, in the best case, perplexity founded on a growing appreciation of the complexity of the phenomena will give way to insight on the same basis.

These differences in intention and method are responsible for a fundamental difference in the relationship between the surface and deepest meanings in religious multilevel texts and in that aspect of philosophic multilevel texts without a parallel in the religious variety, i.e., the dialectical inquiry directed toward the reader's intellectual development. To be sure, by virtue of being multilevel texts, the deepest meaning of each is different from its surface meaning, but this surface agreement masks a deeper disagreement. The preoccupation in religious multilevel writing with inculcating salutary opinions in ordinary readers leads the author of such texts to indicate the content—and sometimes even the existence—of the true teaching in ways that pose the least risk of disrupting the surface teaching. The consequences of subordinating the instructional function of such texts to their political purpose can best be seen by contrasting the religious variety of multilevel writing with one of the philosophic modes recognized in antiquity,⁶⁸ which intends the surface teaching to be an elementary version of the highly technical, philosophic doctrine beneath. The latter model, for which instruction is the primary and perhaps the sole purpose, implicitly presupposes an affinity between the outermost and innermost levels. Where the exoteric teaching is seen as primarily political, in contrast, the distinction between exoteric and esoteric is easily transformed from the distinction between popular and philosophic into the distinction between spurious and serious. The political, religious, or rhetorical mode, according to which the exoteric teaching is a “noble lie,” need not, yet tends to, go hand in hand with a theory of multilevel writing that stresses the polarity between the literal and hidden meanings.

Alfarabi's dialectical method of multilevel writing cannot be equated with the elementary-advanced philosophic model any more than it can with the religious model. On the basis of what we know from ancient testimony, the exoteric stratum of philosophic multilevel writing in the elementary-advanced mode merely furnished a rough replica of the underlying philosophic teachings. The exoteric layer does not appear to have embodied the impetus for and the means of progressing from the elementary to the advanced levels. As far as we know, in other words, the exoteric level was fundamentally static. The surface of Alfarabi's dialectical writing, on the other hand, is dynamic. It is the surface in Alfarabi's works that leads the reader to what lies beneath.

Another explanation of the differences between religious multilevel writing and Alfarabi's dialectical mode is possible: they may originate less in a disagreement about the ultimate purpose of multilevel writing than in a difference in judgment about the means to realize that end. According to this hypothesis, both modes seek first and foremost to enlighten those capable of attaining truth and only secondarily to promote the moral well-being of other people or to avoid being harmed by them. However, the religious mode presupposes that instruction can be realized successfully through the largely concealed methods associated with religious multilevel writing. Hence, the religious approach is able to proceed on the level of rhetorical arguments and still accommodate its ultimate purpose to a large extent. Alfarabi's dialectic, according to this hypothesis, is predicated on the belief that the inherent difficulties of instruction and inquiry force the writer, whatever the writer's fears for the ordinary reader, to labor toward the primary purpose relatively openly—as openly as the requirements of personal discovery permit. Both methods, in short, effect instruction indirectly; the rhetorical mode, however, proceeds indirectly because of the need to conceal, whereas the dialectical mode conceals because of the necessity to proceed indirectly.

Whatever the ultimate reason for the differences between the two modes of multilevel writing—whether a disagreement about ends or about means—a conflict between the two modes remains. An argumentative work dominated by investigative dialectic is at bottom incompatible with a purely persuasive work limited to rhetorical and poetic devices coupled with occasional allusions to true teachings. This conflict is most clearly visible when one addresses the problem of interpreting multilevel texts. There are, to be sure, some procedures common to the interpretation of both modes. First, the overall structure of a book, the order in which its chapters or other divisions are arranged, the juxtaposition of specific passages within a chapter, and even the order in which the premises of an argument are presented should be construed as implicit statements about the subject under discussion. The implicit statements should be identified and compared with the explicit discussion (or discussions) of the same subject. Repetitions should be examined for changes that alter the meaning of the initial formulation in decisive ways. The absence of particular words, assertions, or arguments where a reader versed in the subject would expect them should be taken as indicating an author's reservations about opinions apparently adopted in other treatises or in other parts of the same treatise.⁶⁹

At the same time, belief in the polarity, affinity, or dialectical relationship between the levels of a multilevel work will affect the way a reader applies these general guidelines to specific textual problems. In particular, which of the views one holds will to a large extent determine the weight one assigns to conflicting pieces of evidence, the significance one attaches to omissions, the purpose one attributes to structural features and, hence, the inferences one makes on the basis of them. In other words, the rules of interpretation do not in and of themselves inform the reader when an omission is a “silent rejection” and when it is dictated by the focus of a particular passage or argument and is made good in another place.⁷⁰ Some repetitions supersede previous formulations, others clarify, and still others furnish subdivisions subsumed under the initial, more general formulation.

Further, each of the several views of the nature of multilevel writing has distinctive principles of interpretation. For example, the reader who assumes a polarity between the surface and deepest teachings is likely to argue that pronouncements made “most frequently or more conspicuously” do not represent an author's real beliefs, or that “a hint … deserves to be taken more seriously than the most emphatic and frequently stated doctrines of [an author's] more exoteric works.”⁷¹ The reader who posits a dialectical relationship between levels is likely to be more concerned with the presence and validity of arguments than with frequency of assertions and to evaluate the various doctrines endorsed by an author on the basis of the rational defense provided for each (whether the defense takes the form of an explicit justification for a doctrine or the form of relevant doctrines and arguments advanced elsewhere in the author's works which can be brought to bear on the subject at hand). Again, the belief that one can uncover the deepest teaching of a work merely by discovering which of a pair of contradictory statements should be discarded or the belief that in the last analysis it will be proper to accept one of the author's statements in toto and reject the others⁷² can be traced to an underlying conviction of a polarity between the levels of meaning. In contrast, the reader who assumes a dialectical method will begin by assigning each explicit doctrine, or each formulation of individual doctrines, approximately the same weight and then play them and their consequences off against one another, without the expectation that any one doctrine or formulation will prove entirely true or false. Again, the reader who believes in a polarity between literal text and inner meaning may well endorse the method of interpretation that has been dubbed “reading between the lines.” Although Strauss, who popularized the expression, intended it as a metaphor,⁷³ the expression is unfortunate because it can lead to neglect of or even contempt for the literal text.⁷⁴

The interpretation of Alfarabi's writing as fundamentally dialectical rather than rhetorical, as animated primarily by the desire to enlighten and only secondarily by fear of persecution, and as relying on mutually dependent rather than polar levels of meaning finds an analogue in Maimonides' account in the Guide of the Perplexed of the Solomonic as against the Rabbinic approach to interpreting Biblical parables. Biblical parables, like religious texts in general and multilevel philosophic works, admit of literal and hidden meanings. According to the Rabbinic method of interpretation, the act of discerning the real meaning of a text of the Torah is analogous to recovering a pearl dropped in a dark and cluttered room. To find the precious gem, one must light a candle, itself worth nothing. Similarly, the exoteric meaning of a parable is worth nothing; yet it enables a person to grasp the precious inner meaning (Guide 6:25-7:15/16:4-17:5). Maimonides follows this description of the interpretive method of the Rabbis with a saying of Solomon: A word fitly spoken is like apples of gold in settings of silver (Proverbs 25:11). According to Maimonides, this dictum refers to parables, and it means that a well-constructed parable has two meanings, which can be compared to a golden apple and the silver filigree casing that surrounds it. Because the holes in the filigree are extremely small, the silver casing obscures the existence of the golden apple inside to distant or inattentive people at the same time that it directs the attention of careful observers to the golden apple's existence (Guide 7:15-8:3/17:6-21). Solomon's saying, as interpreted by Maimonides, thus differs from the Rabbis' Midrash in its evaluation of the worth of the exoteric meaning of parables. Although inferior to gold, silver is itself valuable: weighed in its own right—not to mention when compared to a candle of wax—it is worth a great deal.⁷⁵ Further, after lighting a candle, one looks away from the flame. Accordingly, the exoteric meaning is presented as ultimately worthless. In contrast, after perceiving a glimmer of gold beneath the filigree, one must look through the filigree to see the object beneath. In the latter case, then, the exoteric meaning itself leads the careful observer to the inner meaning and always remains a reflection of that meaning.

The question then becomes, does the thesis that the exoteric meaning of parables—and by implication the surface meaning of philosophic texts written on several levels—consists in politically salutary beliefs or simplified true opinions do justice to the Solomonic correction of the Rabbinic method of interpretation? The silver filigree imagery would seem to indicate some kind of cognitive role for the exoteric meaning, whatever its practical function. Maimonides' exact words are that the exoteric meanings constitute wisdom useful in many respects, among which (min jumlatihā) are the well-being of communities of men (Guide 7:29-8:1/17:19-20). In other words, in this passage of the Guide of the Perplexed Maimonides suggests that political well-being is not the exclusive beneficiary of the effect of exoteric formulations. Further, the hypothesis that the exoteric meaning acts as some kind of cognitive forerunner of ultimate theoretical insights gains in force from the fact that it is compatible with Maimonides' characterization of the exoteric meaning in terms of silver, his depiction of the exoteric meaning as access to the inner meaning, and his portrayal of the exoteric meaning as exhibiting roughly the same form as the inner meaning.

There is little doubt that some of those who write in the religious multilevel mode, whether philosophers or not, intend to “reveal the truth to those able to understand by themselves.”⁷⁶ In contrast, those who choose the dialectical mode appear to be less sanguine about the prospects for gifted people if left to their own resources. The possibility should also be considered that the explicit claims about secrecy made by some philosophers who appear to engage in the religious mode of multilevel writing are exaggerated, that is, that these claims are part of the rhetoric of the works. This possibility should be considered, not out of perversity, but as a consequence of taking these works on their own terms. For if all the literal teachings are suspect, then an author's professions on the subject of concealment should be considered provisional as well. One might, then, discern as separate species of philosophic writing in the religious mode the method of Avicenna, who advocates keeping the existence of concealment secret, and the method of Maimonides, whose discourse on the secret character of the Guide of the Perplexed still rings loud and clear.

The chapters which follow examine three of the most fundamental questions addressed by Alfarabi in his political works: the nature of happiness and perfection, especially the role of theory and action as elements of human excellence in the highest case; the qualifications of rulers of excellence, in particular the contribution of both theoretical and practical wisdom to the formation of practical judgments; and the kind or kinds of political orders that make possible a political community of excellence. Throughout the assumption is that both rhetorical and dialectical modes of exposition are present in Alfarabi's treatises and that the former are subordinated to the latter. As a consequence of this assumption, the initial procedure will be to examine all the Farabian doctrines advanced and developed in connection with the above themes with equal care, including both members of obvious contradictions. This will entail pulling together Alfarabi's various treatments of each theme, so as to demonstrate at the outset the seriousness with which Alfarabi views doctrines that may be dismissed as exoteric by one or another school. The bulk of each chapter will be devoted to analyzing and evaluating the positions thus established. This will be followed by an effort to resolve the contradictions explored or, when appropriate, an indication of conflicts without resolution. The final chapter treats the most intractable problem of Farabian exegesis, that of the character and purpose of his parallel works—one of the most conspicuous examples of repetition and reformulation to be found in medieval literature.

Notes

1. Translating aporia as “perplexing point,” aporeō as “raising the perplexing points,” and diaporeō as “going over the perplexing points” in the sense of “examining.” Although this translation is wooden, it reveals the centrality of the concept of aporia in the passage. Euporia and euporein are from the same Greek root, but they have been translated as “reaching answers” instead of, for example, “resolving the perplexing points,” which would reveal the root shared by the two sets of terms. For a discussion of the meanings of these expressions, see Aubenque (1960).

2. For Alfarabi's understanding of this section of the Phaedrus, see Falsafat Aflāṭun 16:4-10.

3. See the references in Hourani (1961), p. 106, n. 142, and Section B below.

4. For an alternative understanding of Maimonides' project, see Section D below.

5. For the theory of the Qur'an as an exoteric text, see Keddie (1963); Averroes Faṣl al-Maqāl 12:11ff.; Hourani (1961), esp. pp. 22-25. For the comparable theory in connection with sacred Jewish texts, see Maimonides Guide of the Perplexed 9:6-20:18/2:6-9:25 (Introduction).

6. For the esotericism of Islamic mystical sects, see Blochet (1902) and Corbin (1960), pp. 28-35. According to Blochet (1902), pp. 490-491, the Muslim philosophers had no tradition of esotericism: once the reader masters the technical vocabulary and ideas, it will become obvious that Islamic philosophy is a continuation of the early Greek and Neoplatonic traditions. According to Corbin (1960), p. 21, the Muslim philosophers were influenced by the esotericism of the mystics.

7. Hourani (1961), p. 106, n. 142.

8. Rosenthal (1968), p. 432. Similarly Guttmann (1950), p. 206, and (1964), p. 434, n. 125.

9. Munk (1859), p. 332; Pines (1937), p. 71; Gardet (1951), p. 676.

10. Strauss (1952), pp. 30-32.

11. Mahdi (1957), p. 118, n. 1. On this kind of writing in general, see Mahdi (1957), pp. 71-72, 113-125, and (1986), the latter of which is entirely devoted to the theme of Alfarabi's method of writing. The texts upon which the latter article is based are discussed in Section B below.

12. Mahdi (1961B), p. 10.

13. Butterworth (1977), pp. 7-10, 12-13, 22-28; Lerner (1974), pp. xvff.; Kurland (1958), pp. xv-xvi; Davidson (1969), pp. xiii-xv; Strauss (1945), p. 375. Lerner notes the ways that the commentary form enables the Muslim philosophers to conceal the extent of their disagreement with earlier philosophers (Lerner 1974, p. xv). Both Davidson and Butterworth observe that Averroes' short or small commentaries constitute largely independent reworkings of Aristotle's texts and that the differences are of substance as well as form.

14. Strauss (1945), p. 375. The exoteric character of Alfarabi's writings, according to Strauss, derives from the doctrine of the Phaedrus that written works as such are exoteric.

15. See note 11 above, Mahdi (1969A), pp. 3-4, 9, and (1975A), p. 50, n. 4.

16. Walzer (1965), pp. 780-781.

17. Rosenthal (1968), p. 432; contrast Madkour (1963), p. 453.

18. Rosenthal (1955), p. 158.

19. Pines (1963), p. 1xxxvi.

20. Pines (1963), p. 1xxix.

21. Davidson (1969), pp. xx, 51.

22. See Section B below.

23. There was also a tradition of secret doctrines associated with Platonic philosophy that Alfarabi knew (see Section B below). For the case against the existence of secret Platonic doctrines, see Boas (1953), pp. 85ff. For the argument that such doctrines did exist, see Strauss (1946), pp. 326-347.

24. See, for example, Burnyeat (1985).

25. See Eudemian Ethics I. viii. 6 1217b22, where Aristotle distinguishes in passing between treatments en tois exōterikois logois and those en tois kata philosophian (sc. logois), i.e., discussions or arguments carried out philosophically. Aristotle never refers to esoteric speeches or arguments.

26. The list is in Ross (1953), vol. 2, pp. 408-409 (commenting on Metaphysics 1076a28).

27. Note, however, that in some instances Aristotle refines the distinction or classification with which he begins. See, for example, Nicomachean Ethics I. 13, esp. 1102a32, 1102b13, 1103a1-3. For the authorities that view exōterikoi logoi as referring to a distinction made elsewhere, see Dirlmeier (1969), p. 53.

28. In Nicomachean Ethics I. 8, the division of goods is tripartite.

29. See the discussions of this question in Gauthier & Jolif (1970), vol. 1, pp. 63-67; Dirlmeier (1969); Ross (1953), vol. 2, pp. 409-410; Moraux (1951), p. 167, n. 79; Zeller (1897), vol. 1, pp. 110-123; Grant (1885), vol. 1, pp. 398-409; Grote (1880), pp. 45-53. Simplicius Physics 695:28 (commenting on Aristotle Physics 217b31) equates the exōterikoi logoi with common and generally accepted opinions.

30. For criticism of this view, see Dirlmeier (1969), pp. 53-55, and Zeller (1897), vol. 1, p. 115, n. 4, p. 121, n. 2.

31. Aquinas Commentary on Aristotle's Nicomachean Ethics I. L. 19, C. 229. Aquinas says that Aristotle calls the discussion of the soul in the De Anima “exoteric” because “he wrote the book as an epistle to persons living at a considerable distance” (Litzinger), as contrasted with the lecture notes used in connection with the oral instruction of students. Aquinas says that alternatively “exoteric” may refer to discussions outside the scope of the inquiry in question.

32. Bernays (1863), pp. 36-40; Zeller (1897), vol. 1, pp. 115-120; Ostwald (1962), p. 9, n. 17 (on ta enkyklia, which Ostwald connects with exōterikoi logoi). Cf. Grant (1885), vol. 1, pp. 401-404, and Grote (1880), p. 52, who believe Aristotle's popular writings in general are meant and not just his dialogues.

33. According to Ross (1949), p. 15, Cicero's description of Aristotle's prose as a “golden stream” refers to the dialogues. Likewise Kerferd (1967), p. 152. Zeller (1897), vol. 1, pp. 106-108, connects Cicero's statement with Aristotle's early works, which were addressed to popular audiences. These may have included the dialogues, but were not limited to them.

34. According to Stahr (1967), vol. 1, p. 322, Aristotle wanted “to come to an understanding with the public” (emphasis in original) because educated men were followers of Plato, and Aristotle wanted to “break ground for his newer philosophy by enlightening the public on certain practical points.”

35. See especially Jaeger (1923). Gauthier & Jolif (1970), vol. 1, pp. 64-67, say that the difference between the exoteric and other works was one of both substance and style—style, because the exoteric works were addressed to the public, which, although informed, still required a “vulgarisation”; and substance, because the exoteric works reflect Aristotle's early views.

36. Grote (1880), pp. 46-48.

37. Grote (1880), p. 52.

38. Grote (1880), p. 53.

39. Grote (1880), pp. 49-53. Similarly Dirlmeier (1969), pp. 52-53.

40. See Wieland (1975); Weil (1975); and Aubenque (1961) on the centrality of Aristotle's dialectical method for his philosophy as a whole. Barnes (1975), pp. 77-87, argues that demonstration was intended to be a mode of exposition, not a method of investigation. For the traditional view, namely, that Aristotle considered dialectic a precursor of and significantly inferior to the activity of philosophy proper, see Solmsen (1968), p. 55, and Huby (1962), p. 72, n. 1, pp. 76, 79-80.

41. Commentarii, according to Zeller (1897), vol. 1, p. 111, n. 5, refers to the strictly scientific (acroamatic) works, which were “continuous expositions.” See Grote (1880), p. 44, n. a (for Cicero commentarii refers to the “general heads—plain, unadorned statements of fact or reasoning, which the orator or historian is to employ his genius in setting forth and decorating so that it may be heard or read with pleasure and admiration by a general audience”). The earliest known catalogue of Aristotle's writings, which also dates from approximately this period, groups all the dialogues together, regardless of subject matter, instead of placing each dialogue near the other works devoted to the same questions. Moraux (1951), pp. 169-170, concludes from this that there is a high degree of probability that the tradition of viewing the dialogues as a separate category was well established in the first century B.C., although he notes that the author of the catalogue saw the fundamental division in Aristotle's writings as being between the exoteric and the more philosophic works, not between the dialogues and the more philosophic works. Cicero says that his source is Antiochus (of Ascalon) (De Finibus V. iii. 8). Both Moraux and Düring consider Cicero's account of his source plausible, and Moraux defends Antiochus as competent to discourse on Peripatetic philosophy as it was then generally understood. What remains, he warns, is to establish the fidelity of the author of the catalogue to Aristotle. Cicero refers to Aristotle's exōterikoi logoi in one other place, his Epistolae ad Atticum, where he says that Aristotle began each exoteric work with an introduction (prohoemium) (Epistolae ad Atticum IV. 16).

42. And Theophrastus.

43. For the secret teaching interpretation of Aristotle, see Düring (1957), pp. 432-436; Boas (1953), pp. 79-85; Moraux (1951), pp. 169-170; Zeller (1897), vol. 1, pp. 120-123.

44. Both letters are summarized and then quoted by Gellius (second century A.D.), who cites Andronicus as his source. These passages are printed in Düring (1957), pp. 431-432. According to Düring the letters are a “more or less verbal quotation” from Andronicus.

45. Translating facultas argutiarum as “argumentative ability.” Rolfe has “logical subtlety”; Düring (1957), p. 432, has “good literary style.”

46. On the contrast between ekdedomenoi and mē ekdedomenoi in Aristotle, see Zeller (1897), vol. 1, p. 108, n. 3 (the contrast is between published or made public and not published or made public, and not merely between already published and not yet published, or between works published by Aristotle and works published by others).

47. Düring (1957), p. 429; Boas (1953). Boas also argues that the so-called evidence of a secret teaching tradition has been misinterpreted. In contrast, Düring finds that a secret teaching tradition did in fact exist, but argues that the tradition misunderstood Aristotle. Likewise Moraux (1951), pp. 169-170; Zeller (1897), vol. 1, pp. 120-121; and Grant (1885), vol. 1, pp. 399-400, deny the plausibility of the secret teaching doctrine, but not its existence.

48. See the discussion of Alfarabi's account of philosophic writing in Section B below.

49. 26 Oratio 319D (= 385-386 Dindorf). This passage and its context (319-320) are quoted in Düring (1957), p. 435. The passage is noted by Boas (1953), p. 84, n. 6, as evidence against his thesis.

50. Olympiodorus Prolegomena 7:6-21; Elias Categories 114:18-115:13. The passage in Olympiodorus is quoted in Düring (1957), p. 438.

51. Ammonius Categories 4:18-27.

52. Düring (1957), pp. 435-436.

53. In Taḥṣīl al-Sa‘ādah the expression for exoteric philosophy is al-falsafah al-barrāniyyah, although the Hyderabad edition has al-falsafah al-b*t*rā'iyyah (Sa‘ādah 90:14/40:13). In Kitāb al-Jadal 37:5-6/203r19, Alfarabi uses the expression bi'l-falsafah al-khārijah wa'l-barrāniyyah. In translating this passage, Mahdi (1986), pp. 112-113, uses the expression “diffused” for al-khārijah and “public” for al-barrāniyyah. Contrast the use of al-falsafah al-khārijah in Kitāb al-Burhān 82:4-5/177r10-11 (“philosophy external to what human beings can do,” i.e., philosophy not about human things).

54. The full title is Kitāb al-Jam‘Bayn Ra'yay al-Hakīmayn Aflāṭun al-Ilāhī wa-Arisṭūṭālīs (“The Harmonization of the Opinions of the Two Wise Men: Plato, the Divine, and Aristotle”). On this work in general, see Fakhry (1965). For an extended discussion of the section of this work devoted to the difference between Plato's and Aristotle's methods of writing, see Mahdi (1986), pp. 104-109.

55. Contrast Plato Seventh Letter 344e (“It cannot be that [a person] has written to assist his memory; there is no danger of a man forgetting the truth, once his soul has grasped it, since it lies within a very small compass” [Hamilton]). Alfarabi mentions Plato's Letters in his catalogue of Plato's works in Falsafat Aflāṭun, but he does not refer to this letter by name or describe its contents.

56. On the practice in late antiquity of making up a correspondence between great men, see Düring (1957), pp. 433-434.

57. On this passage in Alfarabi's Nawāmīs, see Strauss (1959), pp. 134-138, and Mahdi (1986), pp. 109-112. The implication of the account appears to be that philosophic truths can in principle be conveyed by being stated directly. For a different view, see below, this section. See also Mahdi (1986), p. 104.

58. Strauss (1952), p. 17, calls fear of persecution on the part of the philosophers “the most obvious and crudest reason” why they employ exoteric-esoteric writing.

59. For dialectical training by oneself, see Falsafat Arisṭūṭālīs 78:15-79:2.

60. Reading fī with MS Teheran Malik 1583 instead of wa-fī with MS Bratislava No. 231, TE 41, and MS Hamidiyyah 812.

61. Alfarabi here goes further than Aristotle in the claims he makes on behalf of dialectical training. According to Topics I. 2 (the chapter on which Alfarabi is commenting in the above quotation), training in dialectic makes it easier (rhaion) to grasp truth and falsehood (101a34-36). See note 40 above and note 63 below for dialectic in connection with grasping premises.

62. And among the mathematical sciences, those that study physical objects have the same tendency (Jadal 33:20-21/201r7-8).

63. Aristotle seems to agree with Alfarabi about this use of dialectic. Although Aristotle at first says that dialectic is useful (khrēsimos) with respect to the premises of philosophy (Topics I. 2 101a25-26, 36-37), subsequently he says it is necessary (anankē) “to deal with [the ultimate bases of each science] through the generally accepted opinions on each point. This process belongs peculiarly, or most appropriately, to dialectic; for being of the nature of an investigation, it lies along the path to the principles of all methods of inquiry” (Forster) (Topics 101b1-4).

64. For a comparable theory in relation to Maimonides, see Berman (1974), p. 164 and n. 33.

65. Compare Kitāb al-Millah 46:22-47:17, where religion is subsumed under philosophy.

66. Alfarabi also stipulates that the religion based on perfect philosophy uses images throughout or for the most part. This religion would not, then, seem to be the equivalent of what is elsewhere called exoteric philosophy.

67. Above, p. 35.

68. See Section A above (the doctrine that exoteric works are elementary or simplified versions of esoteric or technical teachings).

69. Strauss (1946), p. 352.

70. Strauss (1945), p. 358. See Butterworth (1975), pp. 120-121, (1977), pp. 25-26, for a more developed theory of how to interpret silence.

71. Strauss (1945), p. 375; Najjar (1964), p. 20 (Arabic Introduction). Compare Strauss (1945), p. 392, n. 99, where he infers the importance of certain terms from their frequency in the text.

72. Berman (1965), p. 410; Strauss (1945), p. 369 (1952), p. 59. This view appears to be implied by Twersky's use of the expression as a “split-level composition” (Twersky 1966, p. 556).

73. Strauss (1952), p. 24. He also cautions that “only such reading between the lines as starts from an exact consideration of the explicit statements of the author is legitimate.” Elsewhere he observes about one of Alfarabi's books that it is “doubtful … whether it would be wise of us to attach great importance to its explicit argument” (Strauss 1945, p. 359).

74. Strauss (1952), p. 52, speaks of the “art of revealing by not revealing, and of not revealing by revealing.” Najjar (1964), p. 20 (Arabic Introduction), says that for the most part “the philosopher means what he does not say and says what he does not mean.” The understanding of Alfarabi as a dialectical writer first and foremost accepts the first half of such statements while rejecting the second half.

75. Against this interpretation of the passage is the fact that Maimonides cites “promoting the well-being of communities of men” as one instance of the worth of the exoteric meaning of parables. He thus seems to agree with those who deny the philosophic content of exoteric statements. If so, then the formulation of exoteric teachings in terms of silver would not elevate them from the political to the philosophic realm; at best, the Solomonic formulation would differ from the Rabbinic mainly in the dignity it ascribes to what is politically useful. And given the peculiar subject that occasions Maimonides' reflections (speeches susceptible of literal and hidden meanings), the reader would have to consider the possibility that Maimonides' correction of the Rabbis is a criticism of the rhetoric and not the substance of their remarks. If people's awareness of the exoteric-esoteric distinction is potentially disruptive to society, then one should indulge in whatever noble lies are necessary to defend the integrity of exoteric teachings. According to this line of argument, no significance should be attached to the identification of popular teachings with silver instead of wax.

76. Strauss (1952), p. 94. This phrase is meaningful in the context of the dialectical interpretation of multilevel writing also. In the latter context, however, it refers not to the absence of guidance but to personal discovery—in other words, to the ability to reach certain insights on one's own, as a result of instruction or guidance by others.

Bibliography

Pierre Aubenque (1961). “Sur la notion aristotélicienne d’aporie.” In Aristote et les problèmes de méthode. Symposium Aristotelicum 2 (1960). Louvain and Paris: 1961.

J. Barnes (1975). “Aristotle’s Theory of Demonstration.” In Barnes, Schofield, and Sorabji (1975), 65-87. An earlier version appeared in Phronesis 14 (1969), 123-152.

J. Barnes, M. Schofield, and R. Sorabji, eds. (1975). Articles on Aristotle. Vol. 1: Science. London: 1975.

Lawrence V. Berman (1974). “Maimonides, the Disciple of Alfārābī.” Israel Oriental Studies 4 (1974), 154-178.

——— (1965). Review of S. Pines, Moses Maimonides. The Guide of the Perplexed. Journal of the American Oriental Society 85 (1965), 410-413.

Jakob Bernays (1863). Die Dialoge des Aristoteles in ihrem Verhältnis zu seiner übrigen Werken. Berlin: 1863.

M. E. Blochet (1902). “Études sur l’ésotérisme musulman.” Journal Asiatique (Ninth Series) 19 (1902), 49-111.

George Boas (1953). “Ancient Testimony to Secret Doctrines.” Philosophical Review 62 (1953), 79-92.

M. F. Burnyeat (1985). “Sphinx without a Secret” (review of Leo Strauss, Studies in Platonic Political Philosophy). New York Review of Books, 30 May 1985, 30ff.

Charles E. Butterworth (1977). Averroes’ Three Short Commentaries on Aristotle’s Topics, Rhetoric, and Poetics. Albany, N.Y.: 1977.

——— (1975). “New Light on the Political Philosophy of Averroës.” In Essays on Islamic Philosophy and Science, ed. George F. Hourani. Albany, N.Y.: 1975.

Henri Corbin (1960). Avicenna and the Visionary Recital. Tr. Willard Trask. New York: 1960. Originally published as Avicenne et le récit visionnaire. Tehran and Paris: 1954. Reprinted, Paris: 1979.

Herbert A. Davidson (1969). Averroes. Middle Commentary on Porphyry’s Isagoge and on Aristotle’s Categories. Cambridge, Mass.: 1969.

Franz Dirlmeier (1969). “Physik IV 10.” In Naturphilosophie bei Aristoteles und Theophrast, ed. I. Düring. Heidelberg: 1969.

Ingemar Düring (1957). Aristotle in the Ancient Biographical Tradition. Göteborg: 1957.

Majid Fakhry (1965). “Al-Farabi and the Reconciliation of Plato and Aristotle.” Journal of the History of Ideas 26 (1965), 469-478.

Louis Gardet (1951). “Philosophie et loi religieuse en Islām et dans le Judaïsme médiéval.” Revue Thomiste 51 (1951), 671-684.

R. A. Gauthier & J. Y. Jolif (1970). Aristote. L’Éthique à Nicomaque. Second edition. 2 vols. Louvain and Paris: 1970.

Alexander Grant (1885). The Ethics of Aristotle. Fourth edition, revised. 2 vols. London: 1885.

George Grote (1880). Aristotle. Ed. Alexander Bain and G. Croom Robertson. Second edition, with additions. London: 1880. Reprinted, New York: 1973.

Julius Guttmann (1964). Philosophies of Judaism. Tr. David W. Silverman. New York: 1964. Originally published as Die philosophie des judentums. Munich: 1933.

——— (1950). Philosophia Judaica: Maimonides. London: 1950.

George F. Hourani (1961). Averroes. On the Harmony of Religion and Philosophy (Kitāb Faṣl al-Maqāl). London: 1961.

Pamela Huby (1962). “The Date of Aristotle’s Topics and Its Treatment of the Theory of Ideas.” Classical Quarterly N.S. 12 (1962), 72-80.

Werner Jaeger (1923). Aristoteles: Grundlegung einer Geschichte seiner Entwicklung. Berlin: 1923.

Nikki R. Keddie (1963). “Symbol and Sincerity in Islam.” Studia Islamica 19 (1963), 27-63.

G. B. Kerford (1967). “Aristotle.” In The Encyclopedia of Philosophy. Vol. 1. New York and London: 1967.

Samuel Kurland (1958). Averroes on Aristotle’s De Generatione et Corruptione. Middle Commentary and Epitome. Translation from Arabic, Hebrew, and Latin with an introduction. Cambridge, Mass.: 1958.

Ralph Lerner (1974). Averroes on Plato’s Republic. Translation with an introduction and notes. Ithaca, N.Y., and London: 1974.

Ibrahim Madkour (1963). “Al-Fārābi.” In A History of Muslim Philosophy, ed. M. M. Sharif. 2 vols. Wiesbaden: 1963.

Muhsin Mahdi (1986). “Man and His Universe in Medieval Arabic Philosophy.” In L’homme et son univers au moyen âge, ed. Christian Wenin. Louvain-la-Neuve: 1986.

——— (1975A). “Remarks on Alfarabi’s Attainment of Happiness.” In Essays on Islamic Philosophy and Science, ed. George F. Hourani. Albany, N.Y.: 1975.

——— (1969A). Alfarabi’s Philosophy of Plato and Aristotle. Translated with an introduction. Revised edition. Ithaca, N.Y.: 1969.

——— (1961B). “The Editio Princeps of Fārābī’s Compendium Legum Platonis.” Journal of Near Eastern Studies 20 (1961), 1-24.

——— (1957). Ibn Khaldûn’s Philosophy of History. London: 1957.

Paul Moraux (1951). Les listes anciennes des ouvrages d’Aristote. Louvain: 1951.

Salomon Munk (1859). Mélanges de philosophie juive et arabe. Paris: 1859. Reprinted, Paris: 1955.

Fauzi Najjar (1964). Al-Fārābī’s The Political Regime (Al-Siyasa al-Madaniyya Also Known as the Treatise on the Principles of Beings). Arabic text, edited with an introduction and notes. Beirut: 1964.

Martin Ostwald (1962). Aristotle. Nicomachean Ethics. Translation, with an introduction and notes. Indianapolis: 1962.

G. E. L. Owen, ed. (1968). Aristotle on Dialectic. Oxford: 1968.

Shlomo Pines (1963). “Translator’s Introduction.” In Moses Maimonides. The Guide of the Perplexed. Translation with an introduction and notes by Shlomo Pines, with an introductory essay by Leo Strauss. Chicago: 1963.

——— (1937). “Some Problems of Islamic Philosophy.” Islamic Culture 11 (1937), 66-80.

E. I. J. Rosenthal (1968). “Political Philosophy in Islam and Judaism.” Judaism 17 (1968), 430-440.

——— (1955). “The Place of Politics in the Philosophy of Al-Fārābī.” Islamic Culture 29 (1955), 157-178.

W. David Ross (1953). Aristotle. Metaphysics. Edited with commentary. 2 vols. Revised edition. Oxford: 1953.

——— (1949). Aristotle. Fifth edition, revised. London: 1949.

Friedrich Solmsen (1968). “Dialectic without the Forms.” In Owen (1968), 49-68.

Adolph Stahr (1967). “Aristoteles.” In A Dictionary of Greek and Roman Biography and Mythology, ed. W. Smith. 3 vols. New York: 1967.

Leo Strauss (1959). “How Fārābī Read Plato’s Laws.” In What Is Political Philosophy? New York: 1959.

——— (1952). Persecution and the Art of Writing. Glencoe, Ill.: 1952.

——— (1946). “On a New Interpretation of Plato’s Political Philosophy.” Social Research 13 (1946), 326-367.

——— (1945). “Farabi’s Plato.” In Louis Ginzberg Jubilee Volume. 2 vols. New York: 1945.

Isadore Twersky (1966). Review of Moses Maimonides. The Guide of the Perplexed. Speculum 41 (1966), 555-558.

Richard Walzer (1965). “Al-Fārābī.” Encyclopedia of Islam. New edition. Leiden and London: 1965.

E. Weil (1975). “The Place of Logic in Aristotle’s Thought.” In Barnes, Schofield, and Sorabji (1975), 88-112. Originally published as “La place de la logique dans la pensée aristotélicienne.” Revue de métaphysique et de morale 56 (1951), 283-315.

W. Weiland (1975). “Aristotle’s Physics and the Problem of Inquiry into Principles.” In Barnes, Schofield, and Sorabji (1975), 127-140. Originally published as “Das Problem des Prinzipienforschung und die aristotelische Physik.” Kant-Studien 52 (1960-1961), 206-219.

Eduard Zeller (1897). Aristotle and the Earlier Peripatetics. Tr. B. F. C. Costelloe and J. H. Muirhead. 2 vols. London and New York: 1897. Reprinted, New York: 1962. Originally published as Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung. Third edition. Vol. 2, Part 2. Leipzig: 1879.