George Grote

In the fourth and fifth chapters of my work on ‘Plato and the Other Companions of Sokrates,’ I investigated the question of the Platonic Canon, and attempted to determine, upon the best grounds open to us, the question, What are the real works of Plato? I now propose to discuss the like question respecting Aristotle.

But the premisses for such a discussion are much less simple in regard to Aristotle than in regard to Plato. As far as the testimony of antiquity goes, we learn that the Canon of Thrasyllus, dating at least from the time of the Byzantine Aristophanes, and probably from an earlier time, was believed by all readers to contain the authentic works of Plato and none others; an assemblage of dialogues, some unfinished, but each undivided and unbroken. The only exception to unanimity in regard to the Platonic Canon, applies to ten dialogues, which were received by some (we do not know by how many, or by whom) as Platonic, but which, as Diogenes informs us, were rejected by agreement of the most known and competent critics. This is as near to unanimity as can be expected. The doubts, now so multiplied, respecting the authenticity of various dialogues included in the Canon of Thrasyllus, have all originated with modern scholars since the beginning of the present century, or at least since the earlier compositions of Wyttenbach. It was my task to appreciate the value of those doubts; and, in declining to be guided by them, I was at least able to consider myself as adhering to the views of all known ancient critics.

Very different is the case when we attempt to frame an Aristotelian Canon, comprising all the works of Aristotle and none others. We find the problem far more complicated, and the matters of evidence at once more defective, more uncertain, and more contradictory.

The different works now remaining, and published in the Berlin edition of Aristotle, are forty-six in number. But, among these, several were disallowed or suspected even by some ancient critics, while modern critics have extended the like judgment yet farther. Of several others again, the component sections (either the books, in our present phraseology, or portions thereof) appear to have existed once as detached rolls, to have become disjointed or even to have parted company, and to have been re-arranged or put together into aggregates, according to the judgment of critics and librarians. Examples of such doubtful aggregates, or doubtful arrangements, will appear when we review the separate Aristotelian compositions (the Metaphysica, Politica, &c.). It is, however, by one or more of these forty-six titles that Aristotle is known to modern students, and was known to mediÆval students.

But the case was very different with ancient literati, such as Eratosthenes, Polybius, Cicero, Strabo, Plutarch, &c., down to the time of Alexander of Aphrodisias, AthenÆus, Diogenes Laertius, &c., towards the close of the second century after the Christian era. It is certain that these ancients perused many works of Aristotle, or generally recognized as his, which we do not now possess; and among those which we do now possess, there are many which it is not certain that they perused, or even knew.

Diogenes Laertius, after affirming generally that Aristotle had composed a prodigious number of books (p?p?e?sta ???a), proceeds to say, that, in consequence of the excellence of the author in every variety of composition, he thinks it proper to indicate them briefly.1 He then enumerates one hundred and forty-six distinct titles of works, with the number of books or sections contained in each work. The subjects are exceedingly heterogeneous, and the form of composition likewise very different; those which come first in the list being Dialogues,2 while those which come last are Epistles, Hexameters, and Elegies. At the close of the list we read: “All of them together are 445,270 lines, and this is the number of books (works) composed by Aristotle.”3 A little farther on, Diogenes adds, as an evidence of the extraordinary diligence and inventive force of Aristotle, that the books (works) enumerated in the preceding list were nearly four hundred in number, and that these were not contested by any one; but that there were many other writings, and dicta besides, ascribed to Aristotle — ascribed (we must understand him to mean) erroneously, or at least so as to leave much doubt.4

1 Diog. La. v. 21. S?????a?e d? p?p?e?sta ???a, ?pe? ????????? ???s??? ?p?????a?, d?? t?? pe?? p??ta? ?????? t??d??? ??et??.

2 Bernays has pointed out (in his valuable treatise, Die Dialoge des Aristoteles, p. 133) that the first in order, nineteen in number, among the titles enumerated by Diogenes, designate Dialogues. The longest of them, those which included more than one book or section, are enumerated first of all. Some of the dialogues appear to have coincided, either in title or in subject, with some of the Platonic:— ?e?? ???a??s????, in four books (comparable with Plato’s Republic); ????t????, in two books; S?f?st??, ?e???e???, S?p?s???, each in one book; all similar in title to works of Plato; perhaps also another, ?e?? ??t?????? ? G??????, the analogue of Plato’s Gorgias.

3 Diog. La. v. 27. ??????ta? a? p?sa? ????de? st???? t?tta?e? ?a? tetta?????ta p??? t??? pe?ta??s??????? ?a? d?a??s???? ?d?????ta. ?a? t?sa?ta ?? a?t? pep?a??te?ta? ???a.

4 Diog. La. v. 34. Heitz (Die Verlorenen Schriften des Aristoteles, p. 17) notices, as a fact invalidating the trustworthiness of the catalogue given by Diogenes, that Diogenes, in other places, alludes to Aristotelian compositions which are not mentioned in his own catalogue. For example, though Diogenes, in the catalogue, allows only five books to the Ethica, yet he himself alludes (v. 21) to the seventh book of the Ethica. But this example can hardly be relied upon, because ?? t? ?d?? t?? ?????? is only a conjecture of H. Stephens or MÉnage. The only case which Heitz really finds to sustain his remark, is the passage of the Prooemium (i. 8), where Diogenes cites Aristotle ?? t? ?a????, that work not being named in his catalogue. But there is another case (not noticed by Heitz) which appears to me still stronger. Diogenes cites at length the Hymn or PÆan composed by Aristotle in honour of Hermeias. Now there is no general head of his catalogue under which this hymn could fall. Here Anonymus (to be presently mentioned) has a superiority over Diogenes; for he introduces, towards the close of his catalogue, one general head — ?????a ? ?????, which is not to be found in Diogenes.

We have another distinct enumeration of the titles of Aristotle’s works, prepared by an anonymous biographer cited in the notes of MÉnage to Diogenes Laertius.5 This anonymous list contains only one hundred and twenty-seven titles, being nineteen less than the list in Diogenes. The greater number of titles are the same in both; but Anonymus has eight titles which are not found in Diogenes, while Diogenes has twenty-seven titles which are not given by Anonymus. There are therefore thirty-five titles which rest on the evidence of one alone out of the two lists. Anonymus does not specify any total number of lines; nevertheless he gives the total number of books composed by Aristotle as being nearly four hundred — the same as Diogenes. This total number cannot be elicited out of the items enumerated by Anonymus; but it may be made to coincide pretty nearly with the items in Diogenes,6 provided we understand by books, sections or subdivisions of one and the same title or work.

5 MÉnage ad Diog. tom. ii. p. 201. See the very instructive treatise of Professor Heitz, Die Verlorenen Schriften des Aristoteles, p. 15 (Leipzig, 1865).

6 Heitz, Die Verl. Schrift. des Aristot. p. 51. Such coincidence assumes that we reckon the ????te?a? and the Epistles each as one book.

I think it unnecessary to transcribe these catalogues of the titles of works mostly lost. The reader will find them clearly printed in the learned work of Val. Rose, Aristoteles Pseudepigraphus, pp. 12-20.

The two catalogues just mentioned, agreeing as they do in the total number of books and in the greater part of the items, may probably be considered not as original and copy, but as inaccurate transcripts from the same original authority. Yet neither of the two transcribers tells us what that original authority was. We may, however, be certain that each of them considered his catalogue to comprehend all that Aristotle could be affirmed on good authority to have published; Diogenes plainly signifies thus much, when he gives not only the total number of books, but the total number of lines. Such being the case, we expect to find in it, of course, the titles of the forty-six works composing the Berlin edition of Aristotle now before us. But this expectation is disappointed. The far greater number of the Aristotelian works which we now peruse are not specified either in the list of Diogenes, or in that of Anonymus.7 Moreover, the lists also fail to specify the titles of various works which are not now extant, but which we know from Aristotle himself that he really composed.8

7 Heitz, Verl. Schr. Aristot. p. 18, remarks that “In diesem Verzeichnisse (that of Diogenes) die bei weitem grÖsste Zahl derjenigen Schriften fehlt, welche wir heute noch besitzen, und die wir als den eigentlichen Kern der aristotelischen Lehre enthaltend zu betrachten gewohnt sind.” Cf. p. 32. Brandis expresses himself substantially to the same effect (Aristoteles, Berlin, 1853, pp. 77, 78, 96); and Zeller also (Gesch. der Phil. 2nd ed. Aristot. Schriften, p. 43).

8 Heitz, Verl. Schr. des Aristoteles, p. 56, seq.

The last-mentioned fact is in itself sufficiently strange and difficult to explain, and our difficulty becomes aggravated when we combine it with another fact hardly less surprising. Both Cicero, and other writers of the century subsequent to him (Dionysius Hal., Quintilian, &c.), make reference to Aristotle, and especially to his dialogues, of which none have been preserved, though the titles of several are given in the two catalogues mentioned above. These writers bestow much encomium on the style of Aristotle; but what is remarkable is, that they ascribe to it attributes which even his warmest admirers will hardly find in the Aristotelian works now remaining. Cicero extols the sweetness, the abundance, the variety, the rhetorical force which he discovered in Aristotle’s writings: he even goes so far as to employ the phrase “flumen orationis aureum” (a golden stream of speech), in characterizing the Aristotelian style.9 Such predicates may have been correct, indeed were doubtless correct, in regard to the dialogues, and perhaps other lost works of Aristotle; but they describe exactly the opposite10 of what we find in all the works preserved. With most of these (except the History of Animals) Cicero manifests no acquaintance; and some of the best modern critics declare him to have been ignorant of them.11 Nor do other ancient authors, Plutarch, AthenÆus, Diogenes Laertius, &c., give evidence of having been acquainted with the principal works of Aristotle known to us. They make reference only to works enumerated in the Catalogue of Diogenes Laertius.12

9 Cicero, Acad. Prior. ii. 38, 119: “Quum enim tuus iste Stoicus sapiens syllabatim tibi ista dixerit, veniet flumen orationis aureum fundens Aristoteles, qui illum desipere dicat.” Also Topica, i. 3. “Quibus (i.e. those who were ignorant of Aristotle) eo minus ignoscendum est, quod non modo rebus iis, quÆ ab illo dictÆ et inventÆ sunt, adlici debuerunt, sed dicendi quoque incredibili quÂdam quum copiÂ, tum suavitate.” Also De Oratore, i. 11, 49; Brutus, 31, 121; De Nat. Deor. ii. 37; De Inventione, ii. 2; De Finibus, i. 5, 14; Epistol. ad Atticum, ii. 1, where he speaks of the “Aristotelia pigmenta,” along with the ????????? of Isokrates. Dionysius Hal. recommends the style of Aristotle in equal terms of admiration: pa?a??pt??? d? ?a? ???st?t??? e?? ??s?? t?? te pe?? t?? ????e?a? de???t?t?? ?a? t?? saf??e?a?, ?a? t?? ?d??? ?a? p???a???? (De Veter. Script. CensurÂ, p. 430, R.; De Verb. CopiÂ, p. 187). Quintilian extols the “eloquendi suavitas” among Aristotle’s excellences (Inst. Or. X. i. p. 510). Demetrius Phalereus (or the author who bears that title), De EloquentiÂ, s. 128, commends a? ???st?t????? ????te?. David the Armenian, who speaks of him (having reference to the dialogue) as ?f??d?t?? ?????? ???? (the correction of Bernays, Dial. des Arist. p. 137) ?a? ?a??t?? ???est??, probably copies the judgment of predecessors (Scholia ad Categor. p. 26, b. 36, Brandis).

Bernays (Die Dialoge des Aristoteles, pp. 3-5) points out how little justice has been done by modern critics to the literary merits, exhibited in the dialogues and other works now lost, of one whom we know only as a “dornichten und wortkargen Systematiker.”

10 This opinion is insisted on by Ravaisson, Essai sur la MÉtaphysique d’Aristote, pp. 210, 211.

11 Valentine Rose, Aristoteles Pseudepigraphus, p. 23: “Cicero philosophicis certe ipsius Aristotelis libris nunquam usus est.” Heitz, Die Verlor. Schrift. des Aristot. pp. 31, 158, 187: “Cicero, dessen Unbekanntschaft mit beinahe sÄmmtlichen heute vorhandenen Werken des Aristoteles eine unstreitige Thatsache bildet, deren Bedeutung man sich umsonst bemÜht hat abzuschwÄchen.” Madvig, Excursus VII. ad Ciceron. De Finibus, p. 855: “Non dubito profiteri, Ciceronem mihi videri dialogos Aristotelis populariter scriptos, et Rhetorica (quibus hic Topica adnumero) tum p???te?a? legisse; difficiliora vero, quibus omnis interior philosophia continebatur, aut omnino non attigisse, aut si aliquando attigerit, non longe progressum esse, ut ipse de subtilioribus Aristotelis sententiis aliquid habere possit explorati.” The language here used by Madvig is more precise than that of the other two; for Cicero must be allowed to have known, and even to have had in his library, the Topica of Aristotle.

12 See this point enforced by Heitz, pp. 29-31. AthenÆus (xiv. 656) refers to a passage of Philochorus, in which Philochorus alludes to Aristotle, that is, as critics have hitherto supposed, to Aristot. Meteorol. iv. 3, 21. Bussemaker (in his PrÆfat. ad Aristot. Didot, vol. iv. p. xix.) has shewn that this supposition is unfounded, and that the passage more probably refers to one of the Problemata Inedita (iii. 43) which Bussemaker has first published in Didot’s edition of Aristotle.

Here, then, we find several embarrassing facts in regard to the Aristotelian Canon. Most of the works now accepted and known as belonging to Aristotle, are neither included in the full Aristotelian Catalogue given by Diogenes, nor were they known to Cicero; who, moreover, ascribes to Aristotle attributes of style not only different, but opposite, to those which our Aristotle presents. Besides, more than twenty of the compositions entered in the Catalogue are dialogues, of which form our Aristotle affords not a single specimen: while others relate to matters of ancient exploit or personal history; collected proverbs; accounts of the actual constitution of many Hellenic cities; lists of the Pythian victors and of the scenic representations; erotic discourses; legendary narratives, embodied in a miscellaneous work called ‘Peplus’ — a title perhaps borrowed from the Peplus or robe of AthÊnÊ at the Panathenaic festival, embroidered with various figures by Athenian women; a symposion or banquet-colloquy; and remarks on intoxication. All these subjects are foreign in character to those which our Aristotle treats.13

13 Brandis and Zeller, moreover, remark, that among the allusions made by Aristotle in the works which we possess to other works of his own, the majority relate to other works actually extant, and very few to any of the lost works enumerated in the Catalogue (Brand. Aristoteles, pp. 97-101; Zeller, Phil. der Griech. ii. 2, p. 79, ed. 2nd). This however is not always the case: we find (e.g.) in Aristotle’s notice of the Pythagorean tenets (Metaphys. A. p. 986, a. 12) the remark, d????sta? d? pe?? t??t?? ?? ?t????? ??? ?????ste???; where he probably means to indicate his special treatises, ?e?? t?? ???a???e??? and ???? t??? ???a???e????, enumerated by Diog. L. v. 25, and mentioned by Alexander, Porphyry, and Simplikius. See Alexander, Schol. ad Metaphys. p. 542, b. 5, 560, b. 25, Br.; and the note of Schwegler on Metaphys. i. 5, p. 47.

The difficulty of harmonizing our Aristotle with the Aristotle of the Catalogue is thus considerable. It has been so strongly felt in recent years, that one of the ablest modern critics altogether dissevers the two, and pronounces the works enumerated in the Catalogue not to belong to our Aristotle. I allude to Valentine Rose, who in his very learned and instructive volume, ‘Aristoteles Pseudepigraphus,’ has collected and illustrated the fragments which remain of these works. He considers them all pseudo-Aristotelian, composed by various unknown members of the Peripatetic school, during the century or two immediately succeeding the death of Aristotle, and inscribed with the illustrious name of the master, partly through fraud of the sellers, partly through carelessness of purchasers and librarians.14 Emil Heitz, on the other hand, has argued more recently, that upon the external evidence as it stands, a more correct conclusion to draw would be (the opposite of that drawn by Rose, viz.): That the works enumerated in the Catalogue are the true and genuine; and that those which we possess, or most of them, are not really composed by Aristotle.15 Heitz thinks this conclusion better sustained than that of Rose, though he himself takes a different view, which I shall presently mention.

14 Valent. Rose, Aristoteles Pseudepigr. pp. 4-10. The same opinion is declared also in the earlier work of the same author, De Aristotelis Librorum Ordine et Auctoritate.

15 Heitz, Die Verlor. Schrift. des Ar. pp. 29, 30.

It will be seen from the foregoing observations how much more difficult it is to settle a genuine Canon for Aristotle than for Plato. I do not assent to either of the two conclusions just indicated; but I contend that, if we applied to this question the same principles of judgment as those which modern Platonic critics often apply, when they allow or disallow dialogues of Plato, we should be obliged to embrace one or other of them, or at least something nearly approaching thereto. If a critic, after attentively studying the principal compositions now extant of our Aristotle, thinks himself entitled, on the faith of his acquired “Aristotelisches GefÜhl,” to declare that no works differing materially from them (either in subject handled, or in manner of handling, or in degree of excellence), can have been composed by Aristotle — he will assuredly be forced to include in such rejection a large proportion of those indicated in the Catalogue of Diogenes. Especially he will be forced to reject the Dialogues — the very compositions by which Aristotle was best known to Cicero and his contemporaries. For the difference between them and the known compositions of Aristotle, not merely in form but in style (the style being known from the epithets applied to them by Cicero), must have been more marked and decisive than that between the Alkibiades, Hippias, Theages, ErastÆ, Leges, &c. — which most Platonic critics now set aside as spurious — and the Republic, Protagoras, Gorgias, PhilÊbus, &c., which they treat as indisputably genuine.16

16 Thus (for example) in Bernays, who has displayed great acuteness and learning in investigating the Aristotelian Canon, and in collecting what can be known respecting the lost dialogues of Aristotle, we read the following observations:— “In der That mangelt es auch nicht an den bestimmtesten Nachrichten Über die vormalige Existenz einer grossen aristotelischen Schriftenreihe, die von der jetzt erhaltenen durch die tiefste formale Verschiedenheit getrennt war. Das Verzeichniss aristotelischer Werke fÜhrt an seiner Spitze sieben und zwanzig BÄnde jetzt verlorener Schriften auf, die alle in der kÜnstlerischen GesprÄchsform abgefasst waren,” &c. (Bernays, Die Dialoge des Aristoteles, p. 2; compare ibid. p. 30).

If, as Bernays justly contends, we are to admit these various writings, notwithstanding “the profound difference of form,” as having emanated from the same philosopher Aristotle, how are we to trust the Platonic critics when they reject about one-third of the preserved dialogues of Plato, though there is no difference of form to proceed upon, but only a difference of style, merit, and, to a certain extent, doctrine?

Zeller (Die Phil. der Griechen, ii. 2, pp. 45, 46, 2nd ed.) remarks that the dialogues composed by Aristotle are probably to be ascribed to the earlier part of his literary life, when he was still (or had recently been) Plato’s scholar.

In discussing the Platonic Canon, I have already declared that I consider these grounds of rejection to be unsafe and misleading. Such judgment is farther confirmed, when we observe the consequences to which they would conduct in regard to the Aristotelian Canon. In fact, we must learn to admit among genuine works, both of Plato and Aristotle, great diversity in subject, in style, and in excellence.I see no ground for distrusting the Catalogue given by Diogenes, as being in general an enumeration of works really composed by Aristotle. These works must have been lodged in some great library — probably the Alexandrine — where they were seen and counted, and the titles of them enrolled by some one or more among the literati, with a specification of the sum total obtained on adding together the lines contained in each.17 I do not deny the probability, that, in regard to some, the librarians may have been imposed upon, and that pseudo-Aristotelian works may have been admitted; but whether such was partially the fact or not, the general goodness of the Catalogue seems to me unimpeachable. As to the author of it, the most admissible conjecture seems that of Brandis and others, recently adopted and advocated by Heitz: that the Catalogue owes its origin to one of the Alexandrine literati; probably to Hermippus of Smyrna, a lettered man and a pupil of Kallimachus at Alexandria, between 240-210 B.C.. Diogenes does not indeed tell us from whom he borrowed the Catalogue; but in his life of Aristotle, he more than once cites Hermippus, as having treated of Aristotle and his biography in a work of some extent; and we know from other sources that Hermippus had devoted much attention to Aristotle as well as to other philosophers. If Hermippus be the author of this Catalogue, it must have been drawn up about the same time that the Byzantine Aristophanes arranged the dialogues of Plato. Probably, indeed, Kallimachus the chief librarian, had prepared the way for both of them. We know that he had drawn up comprehensive tables, including, not only the principal orators and dramatists, with an enumeration of their discourses and dramas, but also various miscellaneous authors, with the titles of their works. We know, farther, that he noticed Demokritus and Eudoxus, and we may feel assured that, in a scheme thus large, he would not omit Plato or Aristotle, the two great founders of the first philosophical schools, nor the specification of the works of each contained in the Alexandrine library.18 Heitz supposes that Hermippus was the author of most of the catalogues (not merely of Aristotle, but also of other philosophers) given by Diogenes;19 yet that nevertheless Diogenes himself had no direct acquaintance with the works of Hermippus, but copied these catalogues at second-hand from some later author, probably Favorinus. This last supposition is noway made out.

17 Stahr, who in the first volume of his work Aristotelia (p. 194), had expressed an opinion that the Catalogue given by Diogenes is the Catalogue “der eigenen Schritten des Stageiriten, wie sie sich in seinem Nachlasse befanden,” retracts that opinion in the second volume of the same work (pp. 68-70), and declares the Catalogue to be an enumeration of the Aristotelian works in the library of Alexandria. Trendelenburg concurs in this later opinion (Prooemium ad Commentar. in Aristot. De AnimÂ, p. 123).

18 ???pp?? ? ?a?????e??? ?? t? p??t? pe?? ???st?t?????, is cited by AthenÆus, xv. 696; also v. 213.

Among the Tables prepared by Kallimachus, one was ?a?t?d?p?? S????a?t?? ???a?; and in it were included the ??a????t?p????? s?????ata ???????, ?a? ???s?pp??, ?a? ??t?????, ?t? d? Fa?t?? (AthenÆus, xiv. 644). If Kallimachus carried down his catalogue of the contents of the library to works so unimportant as these, we may surely believe that he would not omit to catalogue such works of Aristotle as were in it. He appears to have made a list of the works of Demokritus (i.e. such as were in the library) with a glossary. See Brandis (Aristoteles, Berlin, 1853, p. 74); also Suidas v. ?a???a???, Diogen. Laert. viii. 86; Dionys. Hal. De Dinarcho, pp. 630, 652 R.; AthenÆus, viii. 336, xv. 669.

19 Heitz, Die Verl. Schr. des Aristot. pp. 45-48.

Patricius, in his Discuss. Peripatetic. (t. i. pp. 13-18), had previously considered Hermippus as having prepared a Catalogue of the works of Aristotle, partly on the authority of the Scholion annexed to the conclusion of the Metaphysica of Theophrastus. Hermippus recited the testament of Aristotle (AthenÆus, xiii. 589).

Both Valentine Rose and Bernays regard Andronikus as author of the Catalogue of Aristotle in Diogenes. But I think that very sufficient reasons to refute this supposition have been shown by Heitz, pp. 49-52.

The opinion given by Christ, respecting the Catalogue which we find in Diogenes Laertius — “illum catalogum non AlexandrinÆ bibliothecÆ, sed exemplarium Aristotelis ab Apelliconte Athenas translatorum fuisse equidem censeo” — is in substance the same as that of Rose and Bernays. I do not concur in it. (Christ, Studia in Aristotelis Libros Metaphysicos, Berlin, 1853, p. 105).

It seems thus probable that the Catalogue given by Diogenes derives its origin from Hermippus or Kallimachus, enumerating the titles of such works of Aristotle as were contained in the Alexandrine library. But the aggregate of works composing our Aristotle is noway in harmony with that Catalogue. It proceeds from a source independent and totally different, viz., the edition and classification first published by the Rhodian Andronikus, in the generation between the death of Cicero and the Christian era. To explain the existence of these two distinct and independent sources and channels, we must have recourse to the remarkable narrative (already noticed in my chapter on the Platonic Canon), delivered mainly by Strabo and less fully by Plutarch, respecting the fate of the Aristotelian library after Aristotle’s death.

At the decease of Aristotle, his library and MSS. came to Theophrastus, who continued chief of the Peripatetic school at Athens for thirty-five years, until his death in 287 B.C. Both Aristotle and Theophrastus not only composed many works of their own, but also laid out much money in purchasing or copying the works of others;20 especially we are told that Aristotle, after the death of Speusippus, expended three talents in purchasing his books. The entire library of Theophrastus, thus enriched from two sources, was bequeathed by his testament to a philosophical friend and pupil, Neleus;21 who left Athens, and carried away the library with him to his residence at the town of SkÊpsis, in the Asiatic region known as Æolis, near Troad. At SkÊpsis the library remained for the greater part of two centuries, in possession of the descendants of Neleus, men of no accomplishments and no taste for philosophy. It was about thirty or forty years after the death of Theophrastus that the kings of Pergamus began to occupy themselves in collecting their royal library, which presently reached a magnitude second only to that of Alexandria. Now SkÊpsis was under their dominion, and it would seem that the kings seized the books belonging to their subjects for the use of the royal library; for we are told that the heirs of Neleus were forced to conceal their literary treasures in a cellar, subject to great injury, partly from damp, partly from worms. In this ruinous hiding-place the manuscripts remained for nearly a century and a half — “blattarum ac tinearum epulÆ,” — until the Attalid dynasty at Pergamus became extinct. The last of these kings, Attalus, died in 133 B.C., bequeathing his kingdom to the Romans. All fear of requisitions for the royal library being thus at end, the manuscripts were in course of time withdrawn by their proprietors from concealment, and sold for a large sum to Apellikon, a native of Teos, a very rich resident at Athens, and attached to the Peripatetic sect. Probably this wealthy Peripatetic already possessed a library of his own, with some Aristotelian works; but the new acquisitions from SkÊpsis, though not his whole stock, formed the most rare and precious ingredients in it. Here, then, the manuscripts and library both of Aristotle and Theophrastus became, for the first time since 287 B.C., open to the inspection of the Athenian Peripatetics of the time (about 100 B.C.), as well as of other learned men. Among the stock were contained many compositions which the Scholarchs, successors of Theophrastus at Athens, had neither possessed nor known.22 But the manuscripts were found imperfect, seriously damaged, and in a state of disorder. Apellikon did his best to remedy that mischief, by causing new copies to be taken, correcting what had become worm-eaten, and supplying what was defective or illegible. He appears to have been an erudite man, and had published a biography of Aristotle, refuting various calumnies advanced by other biographers; but being (in the words of Strabo) a lover of books rather than a philosopher, he performed the work of correction so unskilfully, that the copies which he published were found full of errors.23 In the year 86 B.C., Sylla besieged Athens, and captured it by storm; not long after which he took to himself as a perquisite the library of Apellikon, and transported it to Rome.24 It was there preserved under custody of a librarian, and various literary Greeks resident at Rome obtained access to it, especially Tyrannion, the friend of Cicero and a warm admirer of Aristotle, who took peculiar pains to gain the favour of the librarian.25 It was there also that the Rhodian Andronikus obtained access to the Aristotelian works.26 He classified them to a great degree anew, putting in juxtaposition the treatises most analogous in subject;27 moreover, he corrected the text, and published a new edition of the manuscripts, with a tabulated list. This was all the more necessary, because some booksellers at Rome, aiming only at sale and profit, had employed bad writers, and circulated inaccurate copies, not collated with the originals.28 These originals, however, were so damaged, and the restitutions made by Apellikon were so injudicious, that the more careful critics who now studied them were often driven to proceed on mere probable evidence.

20 Diog. L. iv. 5; Aulus Gellius, N. A. iii. 17.

21 From a passage of Lucian (De Parasito, c. xxxv.) we learn that Aristoxenus spoke of himself as friend and guest of Neleus: ?a? t?? pe?? t??t?? ???e?; ?????? ?? ?a? ?????, ???st??e??? d? ? ??s????, p????? ????? ????? ?a? a?t?? d? pa??s?t?? ???e?? ??.

22 Strabo, xiii. 608, 609; AthenÆus, v. 214. The narrative of Strabo has been often misunderstood and impugned, as if he had asserted that none of the main works of Aristotle had ever been published until they were thus exhumed by Apellikon. This is the supposed allegation which Stahr, Zeller, and others have taken so much pains to refute. But in reality Strabo says no such thing. His words affirm or imply the direct contrary, viz., that many works of Aristotle, not merely the exoteric works but others besides, had been published earlier than the purchase made by Apellikon. What Strabo says is, that few of these works were in possession of the Peripatetic Scholarchs at Athens before the time of that purchase; and he explains thus how it was that these Scholarchs, during the century intervening, had paid little attention to the profound and abstruse speculations of Aristotle; how it was that they had confined themselves to dialectic and rhetorical debate on special problems. I see no ground for calling in question the fact affirmed by Strabo — the poverty of the Peripatetic school-library at Athens; though he may perhaps have assigned a greater importance to that fact than it deserves, as a means of explaining the intellectual working of the Peripatetic Scholarchs from Lykon to Kritolaus. The philosophical impulse of that intervening century seems to have turned chiefly towards ethics and the Summum Bonum, with the conflicting theories of Platonists, Peripatetics, Stoics, and Epikureans thereupon.

23 Strabo, xiii. 609. ?? d? ? ?pe?????? f??????? ????? ? f???s?f??, d?? ?a? ??t?? ?pa?????s?? t?? d?a???t??, e?? ??t???afa ?a??? et??e??e t?? ??af?? ??ap????? ??? e?, ?a? ???d??e? ?a?t?d?? p???? t? ???a.

24 Strabo, xiii. 609; Plutarch, Sylla, c. xxvi.

25 Strabo, xiii. 609. ???a?????, ? ??aat???? d?e?e???sat? f??a??st?t???? ??, ?e?ape?sa? t?? ?p? t?? ????????. Tyrannion had been the preceptor of Strabo (xii. 548); and BoÊthus, who studied Aristotle along with Strabo, was a disciple of the Rhodian Andronikus. See Ammonius ad Categorias, f. 8; and Ravaisson, Essai sur la MÉtaphysique d’Aristote, Introduction, p. 10.

26 Plutarch, Sylla, c. xxvi.

27 The testimony of Porphyry in respect to Andronikus, and to the real service performed by Andronikus, is highly valuable. Porphyry was the devoted disciple and friend, as well as the literary executor, of Plotinus; whose writings were left in an incorrect and disorderly condition. Porphyry undertook to put them in order and publish them; and he tells us that, in fulfilling this promise, he followed the example of what Andronikus had done for the works of Aristotle and Theophrastus. ?pe? d? a?t?? (Plotinus) t?? d?????s?? ?a? t?? d??ta??? t?? ????? p??e?s?a? ??? ?p?t?e?e?, ??? d? ??e??? ???t? ?pes???? ?a? t??? ?????? ?ta????? ?p???e????? p???sa? t??t?, p??t?? ?? t? ???a ?? ?at? ??????? ??sa? f??d?? ??ded???a ?d??a??sa, ??s?e??? d’ ?p????d???? t?? ????a??? ?a? ??d??????? t?? ?e??pat?t????, ?? ? ?? ?p??a??? t?? ???d?????f?? e?? d??a t???? f???? s????a?e?, ? d? t? ???st?t????? ?a? Te?f??st?? e?? p?a?ate?a? d?e??e, t?? ???e?a? ?p???se?? e?? ta?t?? s??a?a???, ??t? d? ?a? ??? pe?t????ta t?ssa?a??ta ???? t? t?? ???t???? ???a d?e???? ?? e?? ?? ???e?da?, t? te?e??t?t? t?? ?? ?????? ?a? ta?? ???e?s?? ?s???? ?p?t????, ???st? d? ???e?d? t? ???e?a f???? s??ef???sa, d??? ?a? t???? p??t?? t??? ??af??t????? p????as??. (Porphyry, Vita Plotini, p. 117, Didot.) Porphyry here distinctly affirms that Andronikus rendered this valuable service not merely to the works of Aristotle, but also to those of Theophrastus. This is important, as connecting him with the library conveyed by Sylla to Rome; which library we know to have contained the manuscripts of both these philosophers. And in the Scholion appended to the Metaphysica of Theophrastus (p. 323, Brandis) we are told that Andronikus and Hermippus had made a catalogue of the works of Theophrastus, in which the Metaphysics was not included.

28 Strabo, xiii. 609: ????p??a? t??e? ??afe?s? fa????? ???e??? ?a? ??? ??t??????te?, &c.

This interesting narrative — delivered by Strabo, the junior contemporary of Andronikus, and probably derived by him either from Tyrannion his preceptor or from the Sidonian BoÊthus29 and other philosophical companions jointly, with whom he had prosecuted the study of Aristotle — appears fully worthy of trust. The proceedings both of Apellikon and of Sylla prove, what indeed we might have presumed without proof, that the recovery of these long-lost original manuscripts of Aristotle and Theophrastus excited great sensation in the philosophical world of Athens and of Rome. With such newly-acquired materials, a new epoch began for the study of these authors. The more abstruse philosophical works of Aristotle now came into the foreground under the auspices of a new Scholarch; whereas Aristotle had hitherto been chiefly known by his more popular and readable compositions. Of these last, probably, copies may have been acquired to a certain extent by the previous Peripatetic Scholarchs or School at Athens; but the School had been irreparably impoverished, so far as regarded the deeper speculations of philosophy, by the loss of those original manuscripts which had been transported from Athens to SkÊpsis. What Aristotelian Scholarchs, prior to Andronikus, chiefly possessed and studied, of the productions of their illustrious founder, were chiefly the exoteric or extra-philosophical and comparatively popular:— such as the dialogues; the legendary and historical collections; the facts respecting constitutional history of various Hellenic cities; the variety of miscellaneous problems respecting Homer and a number of diverse matters; the treatises on animals and on anatomy, &c.30 In the Alexandrine library (as we see by the Catalogue of Diogenes) there existed all these and several philosophical works also; but that library was not easily available for the use of the Scholarchs at Athens, who worked upon their own stock, confining themselves mainly to smooth and elegant discourses on particular questions, and especially to discussions, with the Platonists, Stoics, and Epikureans, on the principia of Ethics, without any attempt either to follow up or to elucidate the more profound speculations (logical, physical, metaphysical, cosmical) of Aristotle himself. A material change took place when the library of Apellikon came to be laid open and studied, not merely by lecturers in the professorial chair at Athens, but also by critics like Tyrannion and Andronikus at Rome. These critics found therein the most profound and difficult philosophical works of Aristotle in the handwriting of the philosopher himself; some probably, of which copies may have already existed in the Alexandrine library, but some also as yet unpublished. The purpose of Andronikus, who is described as Peripatetic Scholarch, eleventh in succession from Aristotle, was not simply to make a Catalogue (as Hermippus had made at Alexandria), but to render a much greater service, which no critic could render without having access to original MSS., namely, to obtain a correct text of the books actually before him, to arrange these books in proper order, and then to publish and explain them,31 but to take no account of other Aristotelian works in the Alexandrine library or elsewhere. The Aristotelian philosophy thus passed into a new phase. Our editions of Aristotle may be considered as taking their date from this critical effort of Andronikus, with or without subsequent modifications by others, as the case may be.

29 Strabo, xvi. 757. Stahr, in his minor work, Aristoteles unter den RÖmern, p. 32, considers that this circumstance lessens the credibility of Strabo. I think the contrary. No one was so likely to have studied the previous history of the MSS. as the editors of a new edition.

30 Strabo, xiii. 609: s???? d? t??? ?? t?? pe??p?t?? t??? ?? p??a? t??? et? Te?f?ast??, ???? ??? ????s? t? ???a p??? ?????? ?a? ???sta t?? ???te?????, ?d?? ??e?? f???s?fe?? p?a?at????, ???? ??se?? ???????e??· t??? d’ ?ste???, ?f’ ?? t? ???a ta?ta p?????e?, ?e???? ?? ??e???? f???s?fe?? ?a? ???st?te???e??, ??a????es?a? ??t?? t? p???? e???ta ???e?? d?? t? p????? t?? ?a?t???. Also Plutarch, Sylla, c. xxvi.

The passage of Strabo is so perspicuous and detailed, that it has all the air of having been derived from the best critics who frequented the library at Rome, where Strabo was when he wrote (?a? ???ade ?a? ?? ??e?a?d?e??, xiii. 609). The Peripatetic Andronikus, whom he names among the celebrated Rhodians (xiv. 655), may have been among his informants. His statements about the bad state of the manuscripts; the unskilful emendations of Apellikon; the contrast between the vein of Peripatetic study, as it had stood before the revelation of the manuscripts, and as it came to stand afterwards; the uncertain evidences upon which careful students, even with the manuscripts before them, were compelled to proceed; the tone of depreciation in which he speaks of the carelessness of booksellers who sought only for profit, — all these points of information appear to me to indicate that Strabo’s informants were acute and diligent critics, familiar with the library, and anxious both for the real understanding of these documents, and for philosophy as an end.

31 Plutarch, Sylla, c. xxvi. Spengel (“Ueber die Reihenfolge der naturwissenschaftlichen Schriften des Aristoteles,” MÜnchen. philol. Abhandl. 1848,) remarks justly that the critical arrangement of Aristotle’s writings, for collective publication, begins from the library of Apellikon at Rome, not from that of Alexandria. See p. 146: “Mehr als zweihundert Jahre lang fehlt uns alle nÄhere Kunde Über die peripatetische Schule. Erst mit der viel besprochenen Auffindung der Bibliothek des Aristoteles in Athen und deren WegfÜhrung nach Rom durch Sulla wird ein regeres Studium fÜr die Schriften des Philosophen bemerkbar — und zwar jetzt eigentlich der Schriften, weniger der Lehre und Philosophie im Allgemeinen, welche frÜher allein beachtet worden ist. Wir mÖchten sagen, von jetzt an beginne das philologische Studium mit den Werken des Aristoteles, die kritische und exegetische Behandlung dieser durch Tyrannion, Andronikus, Adrastus und viele andre nachlfolgende,” &c.

The explanation just given, coinciding on many points with Brandis and Heitz, affords the most probable elucidation of that obscurity which arises about the Aristotelian Canon, when we compare our Aristotle with the Catalogue of Diogenes — the partial likeness, but still greater discrepancy, between the two. It is certain that neither Cicero32 nor the great Alexandrine literati, anterior to and contemporary with him, knew Aristotle from most of the works which we now possess. They knew him chiefly from the dialogues, the matters of history and legend, some zoological books, and the problems; the dialogues, and the historical collections respecting the constitutions of Hellenic cities,33 being more popular and better known than any other works. While the Republic of Plato is familiar to them, they exhibit no knowledge of our Aristotelian Politica, in which treatise the criticism upon the Platonic Republic is among the most interesting parts. When we look through the contents of our editions of Aristotle the style and manner of handling is indeed pretty much the same throughout, but the subjects will appear extremely diverse and multifarious; and the encyclopedical character of the author, as to science and its applications, will strike us forcibly. The entire and real Aristotle, however, was not only more encyclopedical as to subjects handled, but also more variable as to style and manner of handling; passing from the smooth, sweet, and flowing style — which Cicero extols as characterizing the Aristotelian dialogues — to the elliptical brevity and obscurity which we now find so puzzling in the De Anim and the Metaphysica.34

I shall assume this variety, both of subject and of handling, as a feature to be admitted and allowed for in Aristotle, when I come to discuss the objections of some critics against the authenticity of certain treatises among the forty-six which now pass under his name. But in canvassing the Aristotelian Canon I am unable to take the same ground as I took in my former work, when reviewing the Platonic Canon. In regard to Plato, I pointed out a strong antecedent presumption in favour of the Canon of Thrasyllus — a canon derived originally from the Alexandrine librarians, and sustained by the unanimous adhesion of antiquity. In regard to Aristotle, there are no similar grounds of presumption to stand upon. We have good reason for believing that the works both of Plato and Aristotle — if not all the works, at least many of them, and those the most generally interesting — were copied and transmitted early to the Alexandrine library. Now our Plato represents that which was possessed and accredited as Platonic by the Byzantine Aristophanes and the other Alexandrine librarians; but our Aristotle does not, in my judgment, represent what these librarians possessed and accredited as Aristotelian. That which they thus accredited stands recorded in the Catalogue given by Diogenes, probably the work of Hermippus, as I have already stated; while our Aristotle is traceable to the collection at Athens, including that of Apellikon, with that which he bought from the heirs of Neleus, and to the sifting, correction, and classification, applied thereto by able critics of the first century B.C. and subsequently; among whom Andronikus is best known. We may easily believe that the library of Apellikon contained various compositions of Aristotle, which had never been copied for the Alexandrine library — perhaps never prepared for publication at all, so that the task of arranging detached sections or morsels into a whole, with one separate title, still remained to be performed. This was most likely to be the case with abstruser speculations, like the component books of the Metaphysica, which Theophrastus may not have been forward to tender, and which the library might not be very eager to acquire, having already near four hundred other volumes by the same author. These reserved works would therefore remain in the library of Theophrastus, not copied and circulated (or at least circulated only to a few private philosophical brethren, such as EudÊmus), so that they never became fully published until the days of Apellikon.35

But though the edition published by Andronikus would thus contain many genuine works of Aristotle not previously known or edited, we cannot be sure that it would not also include some which were spurious. Reflect what the library of Apellikon, transported to Rome by Sylla, really was. There was in it the entire library of Theophrastus; probably, also, that of Neleus, who must have had some books of his own, besides what he inherited from Theophrastus. It included all the numerous manuscript works composed by Aristotle and Theophrastus, and many other manuscript works purchased or acquired by them, but composed by others — the whole in very bad order and condition; and, moreover, the books which Apellikon possessed before, doubtless as many Aristotelian books as he could purchase. To distinguish, among this heterogeneous mass of manuscripts, which of them were the manuscripts composed by Aristotle; to separate these from the writings of Theophrastus, EudÊmus, or other authors, who composed various works of their own upon the same subjects and with the same titles as those of Aristotle — required extreme critical discernment and caution; the rather, since there was no living companion of Aristotle or Theophrastus to guide or advise, more than a century and a half having elapsed since the death of Theophrastus, and two centuries since that of Aristotle. Such were the difficulties amidst which Apellikon, Tyrannion, and Andronikus had to decide, when they singled out the manuscripts of Aristotle to be published. I will not say that they decided wrongly; yet neither can I contend (as I argued in the case of the Platonic dialogues) that the presumption is very powerful in favour of that Canon which their decision made legal. The case is much more open to argument, if any grounds against the decision can be urged.

Andronikus put in, arranged, and published the treatises of Aristotle (or those which he regarded as composed by Aristotle) included in the library conveyed by Sylla to Rome. I have already observed, that among these treatises there were some, of which copies existed in the Alexandrine library (as represented by the Catalogue of Diogenes), but a still greater number which cannot be identified with the titles remaining of works there preserved. As to the works common to both libraries, we must remember that Andronikus introduced a classification of his own, analogous to the Enneads applied by Porphyry to the works of Plotinus, and to the Tetralogies adopted by Thrasyllus in regard to the Dialogues of Plato; so that even these works might not be distributed in the same partitions under each of the two arrangements. And this is what we actually see when we compare the Catalogue of Diogenes with our Aristotle. Rhetoric, Ethics, Physics, Problems, &c., appear in both as titles or subjects, but distributed into a different number of books or sections in one and in the other; perhaps, indeed, the compositions are not always the same.

Before I proceed to deal with the preserved works of Aristotle — those by which alone he is known to us, and was known to mediÆval readers, I shall say a few words respecting the import of a distinction which has been much canvassed, conveyed in the word exoteric and its opposite. This term, used on various occasions by Aristotle himself, has been also employed by many ancient critics, from Cicero downwards; while by mediÆval and modern critics, it has not merely been employed, but also analysed and elucidated. According to Cicero (the earliest writer subsequent to Aristotle in whom we find the term), it designates one among two classes of works composed by Aristotle: exoteric works were those composed in a popular style and intended for a large, indiscriminate circle of readers: being contrasted with other works of elaborated philosophical reasoning, which were not prepared for the public taste, but left in the condition of memorials for the instruction of a more select class of studious men. Two points are to be observed respecting Cicero’s declaration. First, he applies it to the writings not of Aristotle exclusively, but also to those of Theophrastus, and even of succeeding Peripatetics; secondly, he applies it directly to such of their writings only as related to the discussion of the Summum Bonum.36 Furthermore, Cicero describes the works which Aristotle called exoteric, as having proems or introductory prefaces.37

In the main, the distinction here drawn by Cicero, understood in a very general sense, has been accepted by most following critics as intended by the term exoteric: something addressed to a wide, indiscriminate circle of general readers or hearers, and intelligible or interesting to them without any special study or training — as contrasted with that which is reserved for a smaller circle of students assumed to be specially qualified. But among those who agree in this general admission, many differences have prevailed. Some have thought that the term was not used by Aristotle to designate any writings either of his own or of others, but only in allusion to informal oral dialogues or debates. Others again, feeling assured that Aristotle intended by the term to signify some writings of his own, have searched among the works preserved, as well as among the titles of the works lost, to discriminate such as the author considered to be exoteric: though this search has certainly not ended in unanimity; nor do I think it has been successful. Again, there have not been wanting critics (among them, Thomas Aquinas and Sepulveda), who assign to the term a meaning still more vague and undefined; contending that when Aristotle alludes to “exoteric discourses,” he indicates simply some other treatise of his own, distinct from that in which the allusion occurs, without meaning to imply anything respecting its character.38

To me it appears that this last explanation is untenable, and that the term exoteric designates matter of a certain character, assignable to some extent by positive marks, but still more by negative; matter, in part, analogous to that defined by Cicero and other critics. But to conceive clearly or fully what its character is, we must turn to Aristotle himself, who is of course the final authority, wherever he can be found to speak in a decisive manner. His preserved works afford altogether eight passages (two of them indeed in the Eudemian Ethics, which, for the present at least, I shall assume to be his work), wherein the phrase “exoteric discourses” (???te????? ?????) occurs. Out of these eight passages, there are seven which present the phrase as designating some unknown matter, not farther specified, but distinct from the work in which the phrase occurs: “Enough has been said (or is said, Aristotle intimates) about this subject, even in the exoteric discourses.” To what it is that he here alludes — whether to other writings of his own or oral discussions of his own, or writing and speech of a particular sort by others — we are left to interpret as we best may, by probable reason or conjecture. But there is one among the eight passages, in which Aristotle uses the term exoteric as describing, not what is to be looked for elsewhere, but what he is himself about to give in the treatise in hand. In the fourth book of the Physica, he discusses the three high abstractions, Place, Vacuum, Time. After making an end of the first two, he enters upon the third, beginning with the following words:— “It follows naturally on what has been said, that we should treat respecting Time. But first it is convenient to advert to the difficulties involved in it, by exoteric discourse also — whether Time be included among entities or among non-entities; then afterwards, what is its nature. Now a man might suspect, from the following reasons, that Time either absolutely does not exist, or exists scarcely and dimly,” &c. Aristotle then gives a string of dialectic reasons, lasting through one of the columns of the Berlin edition, for doubting whether Time really exists. He afterwards proceeds thus, through two farther columns:— “Let these be enumerated as the difficulties accompanying the attributes of Time. What Time is, and what is its nature, is obscure, as well from what has been handed down to us by others, as from what we ourselves have just gone through;”39 and this question also he first discusses dialectically, and then brings to a solution.

Now what is it that Aristotle here means by “exoteric discourse?” We may discover by reading the matter comprised between the two foregoing citations. We find a string of perplexing difficulties connected with the supposition that Time exists: such as, “That all Time is either past or future, of which the former no longer exists, and the latter does not yet exist; that the Now is no part of Time, for every Whole is composed of its Parts, and Time is not composed of Nows,” &c. I do not go farther here into these subtle suggestions, because my present purpose is only to illustrate what Aristotle calls “exoteric discourse,” by exhibiting what he himself announces to be a specimen thereof. It is the process of noticing and tracing out all the doubts and difficulties (?p???a?) which beset the enquiry in hand, along with the different opinions entertained about it either by the vulgar, or by individual philosophers, and the various reasons whereby such opinions may be sustained or impugned. It is in fact the same process as that which, when performed (as it was habitually and actively in his age) between two disputants, he calls dialectic debate; and which he seeks to encourage as well as to regulate in his treatise entitled Topica. He contrasts it with philosophy, or with the strictly didactic and demonstrative procedure: wherein the teacher lays down principles which he requires the learner to admit, and then deduces from them, by syllogisms constructed in regular form, consequences indisputably binding on all who have admitted the principles. But though Aristotle thus distinguishes Dialectic from Philosophy, he at the same time declares it to be valuable as an auxiliary towards the purpose of philosophy, and as an introductory exercise before the didactic stage begins. The philosopher ought to show his competence as a dialectician, by indicating and handling those various difficulties and controversies bearing on his subject, which have already been made known, either in writings or in oral debate.40

We thus learn, from the example furnished by Aristotle himself, what he means by “exoteric discourses.” The epithet means literally, extraneous to, lying on the outside of; in the present case, on the outside of philosophy, considered in its special didactic and demonstrative march.41 Yet what thus lies outside philosophy, is nevertheless useful as an accompaniment and preparation for philosophy. We shall find Aristotle insisting upon this in his Topica and Analytica; and we shall also find him introducing the exoteric treatment into his most abstruse philosophical treatises (the Physica is one of the most abstruse) as an accompaniment and auxiliary — a dialectic survey of opinions, puzzles, and controverted points, before he begins to lay down and follow out affirmative principles of his own. He does this not only throughout the Physica (in several other passages besides that which I have just cited),42 but also in the Metaphysica, the treatises De AnimÂ, De Generatione et Corruptione, &c.

Having thus learnt to understand, from one distinct passage of Aristotle himself, what he means by “exoteric discourses,” we must interpret by the light of this analogy the other indistinct passages in which the phrase occurs. We see clearly that in using the phrase, he does not of necessity intend to refer to any other writings of his own — nor even to any other writings at all. He may possibly mean this; but we cannot be sure of it. He means by the phrase, a dialectic process of turning over and criticizing diverse opinions and probabilities: whether in his own writings, or in those of others, or in no writings at all, but simply in those oral debates which his treatise called Topica presupposes — this is a point which the phrase itself does not determine. He may mean to allude, in some cases where he uses the phrase, to his own lost dialogues; but he may also allude to Platonic and other dialogues, or to colloquies carried on orally by himself with his pupils, or to oral debates on intellectual topics between other active-minded men. When Bernays refers “exoteric discourse” to the lost Aristotelian Dialogues; when Madvig, Zeller, Torstrick, Forchhammer, and others, refer it to the contemporary oral dialectic43 — I think that neither of these explanations is in itself inadmissible. The context of each particular passage must decide which of the two is the more probable. We cannot go farther, in explaining the seven doubtful passages where Aristotle alludes to the “exoteric discourses,” than to understand the general character and scope of the reasonings which he thus designates. Extra-philosophical, double-sided, dialectic, is in general (he holds) insufficient by itself, and valuable only as a preparation and auxiliary to the didactic process. But there are some particular points on which such dialectic leaves a result sufficient and satisfactory, which can be safely accepted as the basis of future deduction. These points he indicates in the passages above cited; without informing us more particularly whether the dialectic was written or spoken, and whether by himself or by others.44

From the time of Cicero downward, a distinction has been drawn between some books of Aristotle which were exoteric, and others that were not so; these last being occasionally designated as akroamatic. Some modern critics have farther tried to point out which, among the preserved works of Aristotle, belonged to each of these heads. Now there existed, doubtless, in the days of Cicero, Strabo, Plutarch, and Gellius, books of Aristotle properly called exoteric, i.e. consisting almost entirely of exoteric discourse and debate; though whether Aristotle himself would have spoken of an exoteric book, I have some doubt. Of such a character were his Dialogues. But all the works designated as akroamatic (or non-exoteric) must probably have contained a certain admixture of “exoteric discourse”; as the Physica (F?s??? ????as??) and the Metaphysica are seen to contain now. The distinction indicated by Cicero would thus be really between one class of works, wherein “exoteric discourse” was exclusive or paramount, — and another, in which it was partially introduced, subordinate to some specified didactic purpose.45 To this last class belong all the works of Aristotle that we possess at present. Cicero would have found none of them corresponding to his notion of an exoteric book.

To understand fully the extent comprehended by the word exoteric, we must recollect that its direct and immediate meaning is negative — extraneous to philosophy, and suitable to an audience not specially taught or prepared for philosophy. Now this negative characteristic belongs not merely to dialectic (as we see it in the example above cited from the Aristotelian Physica), but also to rhetoric or rhetorical argument. We know that, in Aristotle’s mind, the rhetorical handling and the dialectical handling, are placed both of them under the same head, as dealing with opinions rather than with truth.46 Both the one and the other are parted off from the didactic or demonstrative march which leads to philosophical truth; though dialectic has a distant affinity with that march, and is indeed available as an auxiliary skirmisher. The term exoteric will thus comprehend both rhetorical argument and dialectical argument.47 Of the latter, we have just seen a specimen extracted from the Physica; of the former, I know no specimen remaining, but there probably were many of them in the Aristotelian dialogues now lost — that which was called ‘Eudemus,’ and others. With these dialogues Cicero was probably more familiar than with any other composition of Aristotle. I think it highly probable that Aristotle alludes to the dialogues in some of the passages where he refers to “exoteric discourses.” To that extent I agree with Bernays; but I see no reason to believe (as he does) that the case is the same with all the passages, or that the epithet is to be understood always as implying one of these lost Aristotelian dialogues.48

46 See the first two chapters of Aristotle’s Rhetorica, especially pp. 1355 a. 24-35, 1358 a. 5, 11, 25, also p. 1404 a. 1.: ???? ??s?? p??? d??a? t?? p?a?ate?a? t?? pe?? t?? ??t??????, which is exactly what he says also about Dialectic, in the commencement of the Topica.