Chapter XI: Appendix: The Sources of the Fragments (1)

The value of a quotation depends on two things, (1) the habit of accuracy in the person who quotes it, and (2) whether it is quoted from the original or from some intermediate source. Consequently the careful student of the early Greek philosophers, who depends wholly on quotations for his direct knowledge of these thinkers, cannot neglect the consideration of these two questions. Closely connected with the accuracy of quotations is the question as to the accuracy of later writers in the opinions which they have attributed to these thinkers. These topics I propose to consider very briefly, that the student may have at least some clue to guide him in his studies.

I.

§ 1. We find in Plato[96] scarcely any quotations, since the literary character of the dialogue excludes anything that might seem pedantic. There are allusions to certain phrases of Herakleitos which had already become all but proverbs:—the Herakleitean sun, the harmony of opposites, ‘all in motion’ with the example of the river; and the comparison ‘god:man::man:ape’ is also given as the teaching of Herakleitos.[97] Similarly phrases of Anaxagoras are brought into the dialogues—‘all things were together,’ ‘νοῦς disposed all things,’[98] but they hardly deserve the name of quotations. Other allusions to his theory do not even suggest a quotation. The only real quotations are from Parmenides,[99] and in two of these passages the text as read by Simplicius was corrupt and unmetrical. Simplicius quotes the same passage at one time from Plato, at another time apparently from the original,[100] so that he enables us to correct the form of the quotation which he (or the writer from whom he drew) read in his MS. of Plato. Plato’s writings betray no particular interest in any of the pre-Sokratic thinkers except Parmenides and the Pythagorean school, nor do they convey any hint as to the value of the work of the other early thinkers. So it need not surprise us that he alludes to popular phrases and seems rather to avoid exact quotation.

§ 2. Beyond these allusions we get comparatively little light from Plato as to the teachings of his predecessors. Xenophanes is once spoken of as the founder of the Eleatic school and of its doctrine of unity. Parmenides is a far more interesting character to Plato, and the highest regard is expressed for him.[101] When his position as to the unity of being and the non-existence of not-being is discussed, there is no reason to think that his opinions are not correctly given; but when Parmenides is introduced as a speaker, we are not to believe that he states the opinions of the real Parmenides any more than the Platonic Sokrates states the positions of the real Sokrates. Of Zeno we learn that he was skilled in the art of dialectic.[102] Zeno’s statement of the occasion and purpose of his book[103] is of course Plato’s deduction from the book itself. The speculations of Anaxagoras are several times mentioned.[104] The statement that he regarded the heavenly bodies as ‘λίθοι’ is a welcome addition to our knowledge of his doctrines; and Plato’s criticism of Anaxagoras’ use of his fundamental principle is most important. Of Empedokles we hear but little; the statement of his doctrine of sense-perception is a happy exception to the rule. The accuracy of Plato’s statements where they can be tested gives an added importance to what he says about the Pythagoreans.[105] In a word all the data which we have from Plato are valuable, but these data are much fewer than we might expect.

§ 3. Both the citations from earlier philosophers and the statement of their opinions are much more frequent in the writings of Aristotle. Two of his references to the sayings of Herakleitos are not new to the reader of Plato; indeed Fr. 41 _ap._ Meta. 1010 a 13 is cited with direct reference to the passage where it is cited in Plato. Fr. 37, if we may accept the conjecture of Patin,[106] is a sarcastic phrase of Herakleitos which Aristotle has introduced seriously into a theory of sense-perception. Fr. 46 and 57 are summary phrases stating the fundamental positions of Herakleitos; Fr. 51 and 55 proverbial sayings attributed to him; Fr. 59 alone has the form of a genuine quotation.[107] It is evident that summary phrases give the philosopher’s impression, just as proverbial sayings may come through the medium of popular thought, so that neither have quite the value of direct quotation.

From Xenophanes Aristotle gives two _mots_, which were attributed naturally enough to the poet-skeptic. There is no proof that Xenophanes was the original author of either of them.

From Parmenides four passages are quoted; strangely enough three of them are passages that had been quoted by Plato. Lines 52-53 in our texts of Aristotle repeat the same error that appears in our texts of Plato; ll. 103-105 are not so near to what seems to be the original (judged by the quotation in Simplicius) as is the Platonic version. Unless our MSS. are greatly at fault, two of the four passages were very carelessly reproduced, and we have reason to believe that they were drawn from Plato. The fourth passage, given by Aristotle and Theophrastos, has the appearance of careful quotation, though one verb has an unmetrical form in our Aristotle (where Theophrastos gives a correct form). Aristotle does not quote directly from either Zeno or Melissos.

Coming now to Empedokles, we find two extended passages which can only be regarded as genuine quotations, namely ll. 287-311 and 316-325. On the other hand several phrases (ll. 208, 326, 443) give only a general idea of the language of Empedokles. Most of the quotations consist of from one to four lines preserving their metrical form, so that they deserve the name of quotations; but their accuracy is doubtful in matters of detail. This is most clearly seen by an examination of the ten cases where the same passage is quoted twice by Aristotle, namely: lines 36-39, 104-107, 146-148, 167, 208, 244, 270-271, 330-332, 333-335. In only three of these instances (38-39, 270-271, 333-335) is the quotation identical; in the other cases there is some slight difference in the text, although commonly both versions scan correctly. An examination of the lines quoted only once in Aristotle shows very frequent deviation from the same lines as quoted by others. In two instances a line is omitted from the context (37 and 99); a case is changed, a connecting particle changed or omitted entirely, a common word is substituted for a rarer one (236-237) or an Aristotelian word for the word required by the full context (e.g. _Meta._ 1015 a 1), or finally only the substance of the line is given (e.g. lines 91, 92). These variations are so numerous as to justify the conclusion that the text furnished by Simplicius or by Sextus Empiricus deserves quite as much weight as that furnished by Aristotle, since the latter cares only for the thought and not at all for the exact language in which the thought had been clothed.

§ 4. In addition to these quotations we find in the writings of Aristotle a comparatively full statement of the opinions of the pre-Sokratic philosophers. Aristotle was interested in the work of his predecessors, since he rightly regarded his own system as the crowning result of partial views that had been set forth before. All that is valuable in their work he would give its place in his own philosophy, and their false or partial opinions he would controvert. Accordingly his ordinary method is to commence the discussion of a theme by stating the opinions of his predecessors and criticising them; and it is natural that the early thinkers who first set forth characteristic views with force and vigour should receive the fullest consideration, for indeed this position is still due to them in the history of philosophy.

Inasmuch as Aristotle set the fashion for later philosophic writers in collecting and criticising the opinions of earlier thinkers, it is important to form a clear conception of both the excellence and the defects of his method.

On a first examination of his statements of these opinions the student is struck by their fullness and comparative accuracy. Emminger[108] has collected and discussed these data, and arrives at the conclusion in every instance that Aristotle’s statement is based on a use of the best materials at his command, and that it reproduces correctly the view of the philosopher in question. It is true that Emminger takes the position of an apologist. There is no doubt, however, that Aristotle was very familiar with the poems of Empedokles, the arguments of Zeno, the system of the Pythagoreans; when he cannot verify his opinions, as in the case of Thales, they are commonly introduced with a λέγεται of caution; and where the views of earlier thinkers seem to be distorted, it is generally due to one of several simple causes which we can estimate with considerable accuracy.

My own conclusion is that the data given by Aristotle are of the greatest value for the study of his predecessors, though they are to be used with caution.

Turning to the defects of the Aristotelian method, I would point out that there is apparently no little difference in the care with which Aristotle had studied the writings of his predecessors. His general attitude towards the Eleatic school is well known, and there is no evidence that he was really familiar with the works of Xenophanes or Parmenides or Melissos. The fact that three of the four quotations from Parmenides were at least suggested by Plato’s writings should not receive undue weight, yet it is certainly suggestive. Several _sayings_ are quoted from Herakleitos, and his logic is severely criticised; we do not, however, obtain from Aristotle any conception of the real importance of Herakleitos. In fact, Aristotle does not seem at all to have understood the meaning of Herakleitos’ work, whether we are to attribute it to his inability to put himself in sympathy with so different a thinker, or to his failure to study his writings. If we had only the data from Aristotle, we should really know more of the significant work of Anaximandros than of Herakleitos.

The conception of the earlier Greek thinkers which we obtain from Aristotle’s writings is distorted along four lines.

1. Whether or not it was due to his failure to study certain of these thinkers, Aristotle’s comparative estimate of them is not one with which we can agree. As for Herakleitos, we can say that Aristotle assigns him a very important place in early thought, even though he gives us but little clue to what his work really was. Perhaps he overestimates the work of Anaximandros and Anaximenes because he finds in them so clear an anticipation of his own thought. Certainly he does not give due weight to the Eleatic school as a whole, and in particular to Melissos. Melissos was not a great original thinker along entirely new lines, but his work in systematising Eleatic thought was very important. Perhaps because he resembled Aristotle in what he sought to do, although from so very different premisses, he is handled with the greater disdain.

2. We may get from Aristotle a slightly distorted view of the earlier thinkers because he stated their views in the terms of his own philosophic system. The commonest philosophical terms, such as ἄπειρον, ἕν, φύσις, κενόν, τὰ ὄντα, στοιχεῖον, σῶμα, οὐσία, πάθη, slightly changed their meanings as they gradually took their place in a definite philosophical terminology. ἄρχη is regularly used by Aristotle to denote the original principle of all things which the early thinkers sought, εἶδος is used in the statement of Herakleitos’ position[109] and of the Pythagorean philosophy[110]: the latter a word introduced into philosophy by Plato, the former probably not used in this sense before Aristotle himself.

3. This tendency, however, is not limited to the use of philosophical terms. Aristotle states the general position of earlier thinkers from the standpoint of his own developed system. The arguments of Zeno and Melissos are thrown into logical form that he may the better criticise them. Herakleitean teachings also are stated in Aristotelian logic, and thereby lose the truth they might have had. Aristotle finds his own theory of indeterminate potential matter in Anaximandros, and it is no easy task to discern what is due to Aristotle and what to Anaximandros in the Aristotelian account. Again in the case of Parmenides we may well question the statement[111] that his two principles were heat = fire = _being_, and cold = earth = _not-being_.

4. Finally Aristotle may be said to give a false impression of his predecessors when he assigns the probable causes for their opinions. Cf. _Meta._ 983 b 18, supra p. 2; _Phys._ 204 b 26, supra p. 10 ‘in order that other things may not be blotted out by the infinite;’ _de anima_ 405 a 25, supra p. 58.

The mere statement of these lines, along which Aristotle may be said slightly to distort the views of his predecessors, is sufficient to put the reader on his guard; and it is comparatively easy to make allowance for them.

§ 5. The fragments of Theophrastos that remain are sufficient only to show that he studied the work of the pre-Sokratic thinkers even more carefully than Aristotle; to make any exact inferences as to his method of making quotations, however, is impossible on the basis of these fragments. Four of his quotations are also cited by Aristotle,[112] and it is interesting to notice that in the second and the fourth of this list Theophrastos gives a text that is probably more correct than that found in our MSS. of Aristotle. The remaining quotations found in Theophrastos[113] show a familiarity only with Empedokles. Only one of these scans correctly, and that by the change of one word, which probably was erroneously copied. Ll. 191-192 have lost some words, and ll. 423-424 are quite rewritten in prose. Apparently Theophrastos was even more careless of the form of his quotations than Aristotle, though he knows the early thinkers at first hand and can correct Aristotle’s quotations. The statement of the _opinions_ of these thinkers by Theophrastos will be considered later in connection with the doxographic tradition.

§ 6. From the time of Aristotle to Plutarch we know comparatively little of the works of the early philosophers, or of the habit of quoting from them. There is abundant evidence, however, that they were studied; the positions and sayings of Herakleitos especially seem to have attracted much attention. The works extant under the name of Hippokrates are attributed by some writers to a period even before Aristotle. In these works there are allusions to the positions of Empedokles and Anaxagoras, and Book I of the treatise περὶ διαιτῆς contains much Herakleitean material. There is scarcely one direct quotation (cf. Fr. 60), and Bernays cannot be said to be successful in reconstructing phrases of Herakleitos from this source. The book, however, is a comparatively early witness to the work of Herakleitos, and doubly important because it is independent of that Stoic study to which is due most of our knowledge of him.

§ 7. More than the other schools that succeeded Aristotle the Stoics devoted themselves to the history of philosophy, and they were interested in Herakleitos for the same reason that Aristotle had been interested in Anaximandros, because they regarded him as a precursor in their own line of thought. Herakleitean phrases occur already in the hymn of Kleanthes to Zeus, thus showing that they had already been adopted into the Stoic phraseology.[114] Philodemos (vii. 81) quotes Chrysippos also as giving a quotation from Herakleitos.

It is only from later writers, however, that we can ascertain how much Herakleitos was studied in this period. Apparently collections were made of his sayings, which soon displaced the more complete form of his writings. Indeed, it is hard to prove that his book existed at all in later times, although Sextus Empiricus quotes a passage of some length which is considered to be the beginning of the work. Further, the works of at least some Stoic writers must have abounded in quotations from Herakleitos. In the writings of Philo there are numerous allusions to sayings of Herakleitos; and the Stoic context, the connection with Stoic ethics, as well as Philo’s general interest in the Stoic school, make it probable that he finds his Herakleitos in his Stoic sources. But while Philo is thus an important witness to the study of Herakleitos among the Stoics, he is of little value in reconstructing the text of the Ephesian philosopher. The carelessness of his method of quotation is shown by the form in which he gives three lines of Empedokles (48-49, 386). To seven fragments of Herakleitos (1, 22, 24, 46, 56, 64, 70) Philo makes a mere allusion; in another series of instances (10, 67, 69, 79, 80, 82) a phrase, often a single word, of Herakleitos is worked into the context. Fr. 68 and 85 are quoted very carelessly, and 76 and 89 have assumed a form very different from that which they originally had. Commonly the name of the author (Herakleitos) is not given.

Cicero quotes Herakleitos 113 in Greek without the author’s name, and translates 114 carefully; Bywater, p. x, suggests that he found the latter in somebody’s _de exilio commentatio_. Returning to the Stoic school, we find in Seneca an accurate translation of Herakleitos 77 and 81, so that we are inclined to trust his version of 120. What seems to be Herakleitos 113, however, is assigned to Demokritos in an expanded form. The epistles attributed to Herakleitos belong to approximately this period, and are interesting only as additional evidence to the study of Herakleitos by Stoic philosophers. Stobaeos quotes several Herakleitean phrases from Musonius. Fr. 20 and 69 are given only in substance, a phrase from 114 is worked into the context, and 75 is quoted in a later form. Fr. 75 as well as 27 and 67 is found in the second and third books of Clement’s _Paedagogos_, books which draw largely from Musonius. The use of Herakleitean material by Lucian, especially in his _Vitarum auctio_, ch. xiv., is doubtless based on a Stoic source, as is indicated by the work ἐκπύρωσις. We may conclude this survey of Stoic writers with Marcus Aurelius. In his writings we find bare allusion to Herakleitos 2, 5, 20, 73, and perhaps to 97; a word or two of 34, 84, and 98 are worked into the text; while 25, 69, 90, 93, 94 are half quoted in the text. Apparently all are allusions to, or abbreviated citations of, sentences with which the reader was supposed to be familiar. It is wholly improbable that citations made in this manner were drawn from the book itself; rather they seem to point to a collection of ‘sayings’ of Herakleitos which must have been quite generally known. Unless such a collection is assumed, they must be regarded as phrases which were familiar to all because they were so often quoted. The former hypothesis seems to me the more tenable.

§ 8. We find in Plutarch one of the principal sources of our fragments. Nearly fifty fragments of Herakleitos are quoted more or less fully in his writings. Many of these quotations consist of a single phrase containing perhaps only a word or two of the original writer, so that they are not of much value for purposes of reconstruction. Sometimes the citation is given in Plutarch’s own words;[115] sometimes there is only a careless allusion, as to Fr. 41, 43, and 120. Even when we seem to have a real quotation, it may be expanded, as in the case of Fr. 108 ap. _Moral._ 143 D compared with _Moral._ 644 F, or Fr. 31 ap. _Moral._ 98 D as compared with _Moral._ 957 A. So I am inclined to regard Fr. 11, 22, and 44 as having been expanded by Plutarch. We cannot therefore place much reliance on the form of Plutarch’s quotations from Herakleitos. As to the source of these quotations we should notice that two of them (Fr. 41 and 45) had been mentioned by Plato, and others (38, 41, 43, and 105) by Aristotle; it is probable that Plutarch quotes these because they were familiar to the readers of Plato and Aristotle. Fr. 20, 22, 24, 25, 34, 44, 75, and 85 occur in Stoic writers, and Plutarch himself refers 91 to the Stoics. Fr. 45-56 are made Stoic in Plutarch by the addition of the word κόσμου (defining ἁρμονίη) which does not appear e.g. in Plato; and Fr. 19, 20, 74, 75, and 87 have a decided Stoic colouring. Thus we may suspect that about half the quotations from Herakleitos were drawn from Stoic sources. On the other hand 78 with its context seems to be based on a considerable passage of Herakleitos, and 11, 12, and 127 have the appearance of careful quotation.

Plutarch’s method in handling quotations from philosophers who wrote in poetry is more satisfactory. It is only rarely that the thought is put in his own words,[116] or that the quotation consists of less than a full line. Sometimes lines are grouped which do not belong together, as ap. _Moral._ 607 C and 618 B. In some instances the text itself seems to be at fault.[117] In general, however, the poetic form protected such quotations from change, and the poetic form was naturally retained in quotations for the purpose of embellishment. I may add that Plutarch rarely neglects to give the name of the author from whom he quotes. As to the source of these poetic quotations, we cannot doubt that Plutarch sometimes quotes Empedokles from the original. A literary man could hardly fail to be acquainted with his poems, and it is by no means likely that the quotations _Moral._ 607 C, 1111 F, 1113 are taken from an intermediate source. Five of the quotations from Parmenides, on the other hand, were not new to the readers of Plato and Aristotle, and the two remaining ones, together with some of the lines from Empedokles, as I have tried to show elsewhere,[118] were probably drawn from a collection of passages on the moon. There is no evidence that Plutarch knew Parmenides at first hand. Many passages of Empedokles also had become common property in the time of Plutarch, and in some instances Plutarch no doubt found collections of quotations suitable for his purpose, so that we cannot attribute all the single lines quoted from Empedokles to Plutarch’s own study of his poems.

§ 9. Judged by the Herakleitos fragments which they yield, the works of Clement and Hippolytos are hardly second in importance to Plutarch for the student of early Greek philosophy. In the _Protreptikos_ of Clement there is an interesting series of passages from Herakleitos on popular worship; in the _Paedagogos_ and the first and fourth books of the _Stromata_ there are scattered quotations most of which bear clear marks of their secondary origin; book II contains several quotations from the introduction to Herakleitos’ works; while the third and fifth books of the _Stromata_ contain a much larger collection of passages from Herakleitos, Xenophanes, Parmenides, and Empedokles. A casual glance at the whole series of quotations shows that Clement’s method was by no means uniform, and that he was often contented with a secondary source for his quotations, not taking the trouble to look them up in the original. In the first book of the _Stromata_ the first quotation from Herakleitos is a proverb familiar in Greek literature, the second passage a bare allusion to a sentence quoted by Plutarch, and the two remaining ones refer to two quotations also given by Diogenes. That Clement used the βίοι which were the basis of the work of Diogenes Laertios is probable from his quotation of Parmenides 28-30 and Empedokles 26-28, 383-384. It is also highly probable that Clement found much of his material in Stoic sources. It is generally agreed that in _Paedagogos_ ii. and iii. he freely used Musonius. Hera. 122 _ap._ Clement 188 ‘what men do not expect at death’ is interpreted by Clement as referring to Stoic fire, and Clement 649 (Hera. 123) also attributes to Herakleitos and the Stoics an idea belonging to the latter only. Hera. 77 is alluded to by Seneca as familiar to his Stoic readers, and other fragments cited by Clement were apparently found by Philo in his Stoic sources. Hera. 69 _ap._ Clement 718 looks like another form of Hera. 19 which Plutarch quotes from a Stoic source, and perhaps we may regard 20 also as from the Stoic source from which Plutarch drew. Hera. 31 _ap._ Clement 87 includes an added phrase (as to the stars) which appears also in _one_ of the two passages in Plutarch where it is quoted. One of the lines of Parmenides and six of the single lines of Empedokles given by Clement are also found in Plutarch. Consequently I regard it as not improbable that Clement drew quotations from Plutarch, and as all but certain that he drew from the Stoic sources of Plutarch. The wrong interpretation of Hera. 116 (_ap._ Clement 699), 122 (_ap._ 18), 67 (_ap._ 261), 79 (_ap._ 111), and perhaps 27 (_ap._ 229) is additional proof that Clement was entirely unfamiliar with the context in which these passages originally stood, and therefore probably did not draw from the original. While we are quite unable to trust Clement’s interpretation of his quotations, it should be remarked that he is exceedingly careful to give the correct form (e.g. Hera. 101 _ap._ Clement 586 as compared with the same fragment in Hippolytos; in this quotation he gives the dialect forms with his usual fidelity).

It remains to consider several series of passages, and to ask whether these were quoted at firsthand. In the _Protreptikos_ we find Herakleitos fragments 122, 124, 125 together, and a little farther on 126-127 (cf. 122 _ap._ Clement 680, and 123 _ap._ 649) on the topic of popular worship. These are clearly quoted from a connected passage, and not phrases that have been passed on as proverbs. Moreover 124-127 are somewhat closely connected with each other (perhaps 122 belongs with them). It is evident that Clement (or possibly the immediate source of Clement) drew them from a somewhat extended passage in the original. Another series of passages from Herakleitos and Empedokles (_ap._ Clement 516 and 520) are quoted as illustrating the misery of human life. They occur together in a long series of quotations on this topic, and at least one line, Empedokles 404, is not quite pertinent; its lack of fitness in this connection may mean that Clement is adapting a collection of passages made (wholly or in part) by another hand for a slightly different purpose. Again, a considerable number of fragments, especially in books ii. and v. of the _Stromata_, are pithy proverbial statements of the fundamental attitude of Herakleitos toward other men (cf. Herakl. 5-8, 104, 2-3, 49, 111 b with its addition from Demosthenes _de corona_ p. 324). These are all marked by their proverbial form, and are many of them quoted by other writers. It is most natural to think that they were drawn from a collection of Herakleitean sayings such as is presupposed by the allusions of Marcus Aurelius and perhaps by the parody of Lucian.

As to the poetic citations in the fifth book of the _Stromata_ it seems to me wholly likely that the verses of Xenophanes, and Parmenides 133-139, are quoted from the original poems. Empedokles lines 74 and 165 are repeated as proverbs; lines 33, 74, 104 (quoted with Herakleitos 68) are often-quoted verses on the favourite topic of the elements; lines 342-343 are quoted with Herakleitos 49, lines 16-17 with Parmenides 28-30 and Herakleitos 111, and it is quite probable that Clement found the topical groups of quotations ready to his hand. Empedokles 26 f., 55 f., 81, 130 f., are all _introductory_ lines, and these too may have been collected by some earlier writer. We may conclude, then, that many of the citations in Clement were not taken from the original works, but that some may have been; the most important fact is that Clement transcribes his quotations with great faithfulness.

§ 10. The citations given in the works of Sextus Empiricus are important because they are in a measure independent of the Stoic line of tradition; we may even say with confidence that some of them are cited from the original works. For Herakleitos there is only one important series of fragments, namely that found in _adv. Math._ vii. §§ 126-134. Fragments 52 and 54 of Herakleitos are indeed mentioned in a series of epigrams with no name attached to them (_Pyrrh._ i. 55), and a little later (_Pyrrh._ iii. 115 and 230) there is an allusion to the well-known Fr. 42 and a statement of Herakleitos’ opinion as to life and death (cf. Fr. 78). The discussion _adv. Math._ vii. §§ 126-134 is a statement of the doctrine of sense-perception which Sextus attributes to Herakleitos. Diels has given good reasons (_Dox._ 209-211) for believing that this passage is based on Aenesidemus, a skeptic philosopher with strong Herakleitean leanings of the first century B.C. In it are contained the full form of Fr. 2 (cited in part by other writers) and Fr. 4 and 92 (with comment based on a longer passage); there is also a phrase reminding the reader of Fr. 77 in § 130. This is the fullest extant material for reconstructing the introduction to Herakleitos’ book, and was evidently based on the text of Herakleitos. While it is cited quite accurately, it is probable that Sextus took the citation from the same source as the rest of the discussion; still, when we remember Sextus’ fondness for citing proœmiums, we cannot say definitely that he did not take it himself from the work of Herakleitos.

Xenophanes is cited in passages varying in length from one to four lines. Most of these passages are not known from other writers or known only from late Homeric commentators. Where the same passage is cited twice, there is no variation except in the arrangement of the lines. Fr. vii. is given in part twice—once lines 3-4, and again lines 1, 2, and 4 (see _supra_ p. 66).—From Parmenides (in addition to the line 132 given by Plato and Aristotle) Sextus gives the proœmium of his work. Although earlier editors have extensively rearranged this passage, I believe it is substantially correct in Sextus, and I see no reason to doubt that it was taken from the work itself. The citation of other lines before 53 by Plato and by Simplicius confirms the suspicion, however, that Sextus had omitted something at this point. From Empedokles’ main philosophical work Sextus gives a portion of the proœmium (lines 2-23), as well as four lines from the introduction to the καθάρματα. It is reasonable to believe that these lines with 428-435 were cited from the original poem; the only errors are copyists’ blunders. Sextus also cites Empedokles 33-35 and 78-80. These are much copied lines, and the form in Sextus includes some obvious errors, e.g. ἀήρ for αἰθήρ (l. 78) and φιλία for φιλότης (l. 80), (cf. ἤπιον l. 79)—errors which very likely were found in the source from which Sextus drew the lines.[119] We may conclude that Sextus cited sometimes from the original, sometimes at second hand; and that his citations reproduce his source accurately except that he sometimes omits verses from their connection.

§ 11. The quotations in the _Refutatio omnium haeresium_, which is now attributed to Hippolytos, include some that are very accurate and others of which the text is hopeless, an anomaly that is very difficult to explain. In the fifth book one phrase reminds the reader of Herakleitos 71, while Herakleitos 68 a is quoted with the author’s name, and 101 without it. In the sixth book there is an allusion to two forms of fire (Hera. 21), and Herakleitos 29 combined with 95 is quoted under the name of Pythagoras. Most of the quotations from Herakleitos, however, are closely grouped in ix. ch. 9-10. Some of these are phrases familiar in earlier writers (e.g. Hera. 3, 47, and 69); 2, 44, 45, and 35 are passages of some length which Hippolytos gives in accurate form; 24 is accompanied by a Stoic explanation, and probably the phraseology of 28 and 36 is Stoic; in most of the citations in this group the text is very carefully given, even to the connecting particles, but besides the fragments in Stoic form just mentioned, the text of 123 is corrupted beyond possibility of restoration, and 58 is almost as bad. These fragments are consistently interpreted as anticipating the views of a Christian sect, and it is possible that the κρινέει of 26 is due to this influence rather than to the Stoics. Bywater (p. ix) suggests that Hippolytos drew his quotations directly from the work of Herakleitos; but it is not easy to regard the difference in accuracy as wholly a difference in the accuracy of one man’s copying.

The quotations from Empedokles, as indeed from other poets, show that Hippolytos was often very careless. The omission of a word (e.g. lines 334, 335 _ap._ Hipp. 165, l. 84 _ap._ 246) is too common to be attributed wholly to the carelessness of copyists, nor would the rest of the text of Hippolytos justify this supposition. Lines 33-35 are quoted twice (p. 246 and p. 318), and the last line differs in the two cases; such a change as from τέγγει to σπόνδε (p. 313) is not one that a copyist would be very likely to make. On the other hand, it is hardly conceivable that the errors in ll. 110 f. _ap._ p. 247, 222 f. _ap._ p. 251, 338 f. _ap._ p. 254 existed in any text that Hippolytos copied. The only possible explanation for this phenomenon is that sometimes Hippolytos quoted from memory, paying no attention either to metre or to phraseology, and sometimes (as in his quotations from Herakleitos generally) from either the original or a source that was very close to the original. Since so many of the Empedoklean passages are not cited by any other writer, we may suppose that Hippolytos drew them from the original.

§ 12. Of the quotations in Diogenes Laertios from Herakleitos, Bywater says (p. x): ‘Laertium ... libro pervetusto usum esse nemo jam adfirmaverit.’ We do find four sentences of some length from Herakleitos, the genuineness of which is not questioned (Fr. 16, 17, 112, 114); it is noticeable that these fragments, together with the allusions to Fr. 33 and 119, all refer to particular men, and so possessed a special interest for the biographical writers, who were Diogenes’ main source. Three other fragments of more than two words are given by Diogenes (71, 100, and 103), and these are not found in any other Greek writer. The remaining fragments consist of only one or two words (22, 48, 62, 69, 80, 113), or are now regarded as spurious (131, 132). There is no reason to think that the fragments of Herakleitos contained in this work are not copied with reasonable accuracy; on the other hand, we may assume from what we know of Diogenes’ method of work that they were not drawn directly from the writings of Herakleitos.

Diogenes quotes Xenophanes xiv. 1-2, and Empedokles l. 6, in a series of passages on skepticism, Xen. xviii. in a series on Pythagoreanism; Fr. xxiv., the only one not found elsewhere, relates to the life of Xenophanes. From Parmenides are quoted lines 28-30 and 54-56. The last passage does not really illustrate the point for which it is quoted (the senses inexact), and our text of Diogenes contains two blunders from some copyist. Portions of the proœmium of Empedokles’ main work on philosophy (1, 24-32, _ap._ viii. 60 and 59) are mentioned in connection with the name of Satyros. It is pretty clear (_ap._ viii. 62) that a ‘Herakleitos’ is the source from which lines 352-363 are taken; if so, the statement viii. 54 that this is the beginning of the καθάρματα comes from the same writer. Lines 384-385 are quoted much in the form in which they appear in Athenaeos, though with one copyist’s error; from the same work of Empedokles we have also lines 355, 415, 417 in passages where Diogenes had just mentioned Timaeos. The familiar lines 35 and 67-68 are found here—line 35 in a very confused form. In general these lines from poetic writers show numerous small errors, which may be due to the state of our manuscripts. Both the fragments from Herakleitos and those in poetic form are of great value, though we are in the dark as to their immediate source.

§ 13. The works of neo-Platonic writers frequently mention the earlier philosophers, but yield few fragments of value. Plotinos refers to ten fragments of Herakleitos. Four of these (80, 82, 83, 85) have the form of quotations, and in two instances the name of Herakleitos is mentioned; they are, however, very short, and give no clue to their source. Sometimes Plotinos plays on words that were evidently known as Herakleitean, e.g. Fr. (47?), 54, 69, 80; or again an Herakleitean idea is stated in his own words, Fr. 32, 83, 99, 130. The manner in which these quotations and allusions are made shows that the phrases were very familiar, either in earlier writers or possibly in some collection of sayings. Line 81 b of Parmenides is quoted with no name; line 40 b is quoted with the author’s name, and is followed by an account of the context which shows that it was drawn from a passage of some length. From Empedokles we find only two phrases, taken from lines 381 and 382, that are worked into the text of Plotinos.

Porphyry quotes from Herakleitos only familiar phrases, and these in the briefest form (74 ap. _de antr. nym._ xi. and 72 ap. _de antr. nym._ x.). The phrases were so familiar that it was only necessary to suggest the idea (e.g. 56 ap. _de antr. nym._ xxix.) without mentioning the name of the philosopher. Parmenides is not so well known; Greeks and Egyptians, we read, say that he mentioned the two gates in his _Physika_ (_de antr. nym._ xxiii.). Only the καθάρματα of Empedokles is quoted, but here Porphyry knows the subjects treated in the work (_de abst._ II. xxi.), and sometimes the full context of the passage he quotes (e.g. _de antr. nym._ viii.). In the case of lines 415-420 we are not sure that Porphyry was right in applying the verses to Parmenides; still, the quotations would seem to be taken directly from the καθάρματα and copied with fair accuracy.

Iamblichos draws a few quotations from his predecessors in the neo-Platonic school (Empedokles, lines 415-420 from Porphyry; and Herakleitos, Fr. 69, 82, 83 from Plotinos, if Stobaeos is correct in attributing this group of fragments to Iamblichos). Most of the allusions to fragments of Herakleitos, however, cannot be traced to this source. The combination of Herakleitos 29 and 95, which Hippolytos had attributed to Pythagoras, Iamblichos also attributes to the same thinker; his language, however, differs in detail from that used by Hippolytos. Two words of Herakleitos 114 (which had been cited by the Stoics and by Diogenes) are given, with the additional statement that Herakleitos gave laws to the Ephesians. Bywater’s number 128 is an allusion probably including a single word from Herakleitos, as does 129 also. Two words each from Fr. 11 and 12 (both found in Plutarch) are worked into the text of Iamblichos—in the former instance with the name of Herakleitos. Finally 105, which also appears in Plutarch, is given here in more accurate form. These references to Herakleitos, like those of the earlier neo-Platonists, are all made to fragments assumed to be familiar because they had been quoted often by earlier writers.

The writings of his predecessors in this same school are frequently mentioned by Proklos, but his quotations from pre-Sokratic thinkers seem not to be derived from them. In the commentary on Parmenides several scattered lines are quoted from the works of the original Parmenides. The quotations are very brief; they include in all only parts of six or seven lines, and sometimes these are cited more than once. It is therefore quite unlikely that Proklos drew them directly from the poem of Parmenides. In his commentary on the _Timaeos_ Proklos uses the form of quotation from Herakleitos six times (alluding to Fr. 16, 32, 44, 68, 79, 80), but only 32 and 44 can be called quotations, while even these are very brief. On p. 106 E we find part of what Diogenes gives in connection with Fragment 80, but no part of 80 itself; 79 was cited by the early Christian writers, and Proklos interpreted it in the same manner that they had done; 68 also had been paraphrased in the source from which Proklos drew it. So far as Herakleitos is concerned, we see how far from their origin the tradition of the fragments had gone, but we get no new light on their original form.

A few lines of Parmenides we know only from Proklos. Verses 29-30 had been given by Diogenes and Clement, but some of the verses 33-40 are new. In these instances, as is usually the case with the quotations in Proklos, the text of the quotations is in a condition almost hopeless. Indeed, at p. 160 D a line and a half of Parmenides are filled out with half a line from Empedokles under the name of the former writer. From Empedokles only single lines (once two lines together) are given, and they aid but little in the reconstruction of the text. Proklos, like Plutarch, is very careful to cite the name of his authorities; but the text of the quotations is so carelessly reproduced that they are of little value.

§ 14. The commentators on Aristotle early began to illustrate his statements about earlier thinkers by passages copied from their works. Alexander of Aphrodisias and Joh. Philoponos seldom add fragments not contained in the works of Aristotle himself; but Simplicius copies long extracts, so that, except for Herakleitos, his commentaries are the most important source for our knowledge of the writings of the pre-Sokratic philosophers. There can be no doubt that most of these quotations—at least in his commentary on the _Physics_ of Aristotle—were drawn from the original works. The most careful scrutiny of the passages from Zeno, Melissos, and Anaxagoras fails to reveal any reason for questioning their character as genuine quotations, except in the case of some of the fragments of Melissos. Pabst (and independently Burnet) has shown that the so-called Fragments 1-5 of Melissos, though given in the form of quotations, are in reality an epitome covering more briefly the same ground that is covered by the following fragments, and adding almost nothing to our knowledge of Melissos. It is wholly unlikely that Simplicius made this epitome himself, for that would be at variance with his ordinary method of work, and with his custom later in dealing with Melissos. So we are driven to assume either that he drew them from some epitome of Melissos to which he had access, or, what seems to me more probable, that he copied them from an earlier commentator, whose habit it was to condense his quotations rather than to copy them at full length. If now we examine the quotations in Simplicius’ commentary on the _de caelo_ (Melissos Fr. 17 and numerous lines from Parmenides and Empedokles), it is noticeable that a considerable number of them occur also in the scholia to Aristotle. It is possible that as they appear in our scholia they all come from Simplicius. One long quotation (Melissos Fr. 17) is, however, taken by Eusebios from Aristokles, a much earlier commentator on Aristotle. This fact of course confirms the belief that earlier commentators on Aristotle accessible to Simplicius already contained quotations from the philosophers in question;[120] and the presence in our scholia of so many fragments quoted by Simplicius on the _de caelo_ would at least suggest an investigation of the question whether our scholia drew them from an earlier source than Simplicius—in other words, whether Simplicius did not in all probability take them from the commentaries of his predecessors. So when we find Parmenides line 78 _ap._ Simplicius, _Physica_ 29, 18 in the form that Plato had quoted it,[121] when we find line 60 _ap._ 120, 23 quoted from an indirect source (cf. p. 145, 4, where it is quoted in context), we may conclude that Simplicius took those quotations from Parmenides at second hand, and not improbably from earlier commentators on Aristotle. The quotations from Herakleitos are all of them in a late form, and show that Simplicius was not familiar with any work under the name of Herakleitos.[122] Nor did Simplicius know Xenophanes at first hand. The two quotations from his poem occur in the discussion of a passage from Theophrastos, and are probably taken from him. The quotations show, however, that Simplicius knew at first hand the works of Zeno, Melissos, Anaxagoras, Parmenides, and Empedokles, and it remains to examine the numerous quotations from the last two thinkers in order to form some idea as to the probable accuracy of Simplicius’ method of quotation.

Stein in his attempt to restore the text of Parmenides finds numerous misarrangements of the lines and breaks where one or more lines have dropped out. Certainly there is evidence that Simplicius omitted four or more lines between 89 and 94, nor does he indicate the break in any way. Several times a phrase of his own is inserted in the middle of a line (e.g. _Phys._ 39, 28; 143, 22), and once a line is filled out metrically, according to our manuscripts, by a phrase which is generally regarded as a comment from Simplicius (_Phys._ 145, 16). The text itself of these fragments is often very dubious in our manuscripts (e.g. lines 96, 98, 100), but Simplicius may not be responsible for this. In our manuscripts also we read sometimes ωὐτός, sometimes αὐτός, and when either ὤν or ἐών (ὄντα or ἐόντα) is metrically possible, the shorter is usual; here again we cannot with any confidence hold that Simplicius is responsible.

The quotations from Empedokles shed more light on the method of Simplicius. Not infrequently lines are omitted in sequence, as two lines between 68 and 70 (_Phys._ 158, 1 f.), and again in the same quotation one line between 90 and 92, and two lines between 93 and 94. According to Bergk the line between 174 and 176 should be omitted (it is identical with 184); and Schneidewin inserts here line 175 (of Stein) from Stobaeos; the passage occurs twice in the same form in Simplicius, however (and once in the scholia to Aristotle), so that this error probably existed in the text from which Simplicius copied. On p. 33, 19 of the _Physica_ two passages from different parts of the poem of Empedokles are joined without break, and the end of line 95 (Stein 115) is modified to make the connection between the two passages. In two instances I believe that Simplicius (or some copyist) has repeated in a quotation some lines from the last previous quotation. On p. 159 of the _Physica_ the end of the first quotation is repeated as the end of the second, except that a summary phrase is substituted for the last half-line; again on p. 160 (lines 6-8) we find three lines which had occurred in the last previous quotation, and which are inserted here with the change of a connecting word. Sometimes we can point out an error that probably existed in the text from which Simplicius copied, as in the case of line 175 mentioned above. Thus ἐδεῖτο in line 99, κῆρυξ in 93, βεβλάστηκε at 105, and probably ἤερος in 78 appear in repetitions of the same quotation at different points, and so may be assigned to the source of Simplicius. In other instances we may say that Simplicius copied carelessly, as in the case of line 89, which is corrected in the prose paraphrase, and possibly 138, where the curious text in the _Physica_ may be corrected from the _de caelo_. The state of our manuscripts of Simplicius, however, is probably responsible for most of the numerous errors in the forms of words.

From this survey of the sources I have omitted the names of many writers who furnish some little addition to our knowledge of the fragments, for their method of quotation is relatively unimportant, nor have I thought it necessary to consider later writers who throw light only on the later history of the fragments. Accordingly I have not spoken of Eusebios, who repeats quotations from Plutarch and from Clement, or of Theodoret, who drew from Clement, or of Julian, who drew from Plutarch. Again, I have not spoken of Stobaeos, or Eustathios, or the scholia generally, as sources, for we are not at present able to determine the line of tradition for these fragments. I have, however, examined the more important sources of fragments, in order that the student may be able to estimate the relative value of the sources, both as to text and as to directness of transmission, in his own study of them.

II.

§ 15. Turning now to the doxographic tradition, we may state the problem as follows:—In the _Placita philosophorum_ attributed to Plutarch, in the _Eclogae physicae_ of Stobaeos, in fragments from Arius Didymos, in Hippolytos, and in other writers, we find copious statements as to the _opinions_ of the early philosophers. These opinions shed light on many points not mentioned in the fragments of their writings now remaining, and so they have great importance for the student of their systems. At the same time they are often confused and unreliable. The problem is to determine the relation of these writers to each other, as well as to the source of the whole series, in order that we may estimate their relative value. This work has been most successfully accomplished in the Prolegomena, to Diels’ _Doxographi Graeci_, a work that is absolutely indispensable to the student of this subject. There is no occasion to reopen here a question that Diels has so successfully solved, but I propose to state briefly a few of the conclusions which the reader will find substantiated in the work of Diels.

The most obvious fact to one who takes up the study of the doxographic writers is that the _Placita_ attributed to Plutarch, and the _Eclogae physicae_, which was originally a part of the _Florilegium_ of Stobaeos, are intimately related; and when the two are printed side by side, as the reader finds them in the text of Diels, the likeness of the two is most striking. At the same time the two books are not identical, and each gives much material that the other omits. Stobaeos cannot have copied from the work attributed to Plutarch, for even in passages that occur in the _Placita_ Stobaeos not infrequently gives the fuller form; nor can the writer of the _Placita_ have copied from Stobaeos, for his work can be traced back nearly three centuries before the time of Stobaeos. It was used by Athenagoras in his defence of the Christians 177 A.D. (_Dox._ p. 4); it was mentioned by Theodoret (_Dox._ p. 47); and important corrections of the text are made by Diels on the authority of Eusebios, Cyril, and the pseudo-Galen, all of whom had used it. Theodoret (_Therap._ IV. 31, _Dox._ 47) mentions the epitome by Plutarch, but only after he has mentioned the _Placita_ of Aetios, Ἀετίου τὴν περὶ ἀρεσκόντων συναγωγήν, and it is this work of Aetios which Diels vindicates as the source both of Plutarch and of Stobaeos, while Theodoret also quotes from it occasionally. A careful study of these three writers and their methods enables Diels to reconstruct a large part of the work of Aetios; and it is the sections of this work bearing on the earlier philosophers which I have translated (see III. English Index under ‘Aetios’). Of Aetios himself almost nothing is known; the work assigned to him must have been written between the age of Augustus and the age of the Antonines (_Dox._ 100). It was in four books, divided into chapters by topics, and in each chapter the opinions of the philosophers were given not by schools but by affinity of their opinions.