Anaximander, son of Praxiades, was also a citizen of Miletos, and Theophrastos described him as an “associate” of Thales.45 We have seen how that expression is to be understood (§ XIV).

According to Apollodoros, Anaximander was sixty-four years old in Ol. LVIII.2 (547/6 B.C.); and this is confirmed by Hippolytos, who says he was born in Ol. XLII. 3 (610/9 B.C.), and by Pliny, who assigns his great discovery of the obliquity of the zodiac to Ol. LVIII.46 We seem to have something more here than a combination of the ordinary type; for, according to all the rules, Anaximander should have “flourished” in 565 B.C., half-way between Thales and Anaximenes, and this would make him sixty, not sixty-four, in 546. Now Apollodoros appears to have said that he had met with the work of Anaximander; and the only reason he can have had for mentioning this must be that he found in it some indication which enabled him to fix its date. Now 547/6 is just the year before the fall of Sardeis, and we may perhaps conjecture that Anaximander mentioned what his age had been at the time of that event. We know from Xenophanes that the question, “How old were you when the Mede appeared?” was considered an interesting one in those days.47 At all events, Anaximander was apparently a generation younger than Thales.48

Like his predecessor, he distinguished himself by certain practical inventions. Some writers credited him with that of the gnomon; but that can hardly be correct. Herodotos tells us this instrument came from Babylon, and Thales must have used it to determine the solstices and equinoxes.49 Anaximander was also the first to construct a map, and Eratosthenes said this was the map elaborated by Hekataios. No doubt it was intended to be of service to Milesian enterprise in the Black Sea. Anaximander himself conducted a colony to Apollonia,50 and his fellow-citizens erected a statue to him.51

Nearly all we know of Anaximander's system is derived in the last resort from Theophrastos, who certainly knew his book.52 He seems once at least to have quoted Anaximander's own words, and he criticised his style. Here are the remains of what he said of him in the First Book:

Anaximander of Miletos, son of Praxiades, a fellow-citizen and associate of Thales,53 said that the material cause and first element of things was the Infinite, he being the first to introduce this name of the material cause. He says it is neither water nor any other of the so-called54 elements, but a substance different from them which is infinite; from which arise all the heavens and the worlds within them.—Phys. Op. fr. 2 (Dox. p. 476; R. P. 16).

He says that this is “eternal and ageless,” and that it “encompasses all the worlds.”—Hipp. Ref. i. 6 (R. P. 17 a).

And into that from which things take their rise they pass away once more, “as is meet; for they make reparation and satisfaction to one another for their injustice according to the ordering of time,” as he says55 in these somewhat poetical terms.—Phys. Op. fr. 2 (R. P. 16).

And besides this, there was an eternal motion, in which was brought about the origin of the worlds.—Hipp. Ref. i. 6 . (R. P. 17 a).

He did not ascribe the origin of things to any alteration in matter, but said that the oppositions in the substratum, which was a boundless body, were separated out —Simpl. Phys. p. 150, 20 (R. P. 18).

Anaximander taught, then, that there was an eternal, indestructible something out of which everything arises, and into which everything returns; a boundless stock from which the waste of existence is continually made good. That is only the natural development of the thought we have ascribed to Thales, and there can be no doubt that Anaximander at least formulated it distinctly. Indeed, we can still follow to some extent the reasoning which led him to do so. Thales had regarded water as the most likely thing to be that of which all others are forms; Anaximander appears to have asked how the primary substance could be one of these particular things. His argument seems to be preserved by Aristotle, who has the following passage in his discussion of the Infinite:

Further, there cannot be a single, simple body which is infinite, either, as some hold, one distinct from the elements, which they then derive from it, or without this qualification. For there are some who make this (i.e. a body distinct from the elements) the infinite, and not air or water, in order that the other things may not be destroyed by their infinity. They are in opposition one to another—air is cold, water moist, and fire hot—and therefore, if any one of them were infinite, the rest would have ceased to be by this time. Accordingly they say that what is infinite is something other than the elements, and from it the elements arise.—Arist. Phys. Γ. 204 b 22 (R. P. 16 b).

It is clear that Anaximander is here contrasted with Thales and with Anaximenes. Nor is there any reason to doubt that the account given of his reasoning is substantially correct, though the form is Aristotle's own, and in particular the “elements” are an anachronism.56 Anaximander started, it would seem, from the strife between the opposites which go to make up the world; the warm was opposed to the cold, the dry to the wet. These were at war, and any predominance of one over the other was an “injustice” for which they must make reparation to one another at the appointed time.57 If Thales had been right in saying that water was the fundamental reality, it would not be easy to see how anything else could ever have existed. One side of the opposition, the cold and moist, would have had its way unchecked, and the warm and dry would have been driven from the field long ago. We must, then, have something not itself one of the warring opposites, something more primitive, out of which they arise, and into which they once more pass away. That Anaximander called this something by the name of φύσις is the natural interpretation of what Theophrastos says; the current statement that the term ἀρχή was introduced by him appears to be due to a misunderstanding.58 We have seen that, when Aristotle used the term in discussing Thales, he meant what is called the “material cause,”59 and it is hard to believe that it means anything else here.

It was natural for Aristotle to regard this theory as an anticipation or presentiment of his own doctrine of “indeterminate matter,”60 and that he should sometimes express the views of Anaximander in terms of the later theory of “elements.” He knew that the Boundless was a body,61 though in his own system there was no room for anything corporeal prior to the elements; so he had to speak of it as a boundless body “alongside of” or “distinct from” the elements (παρὰ τὰ στοιχεῖα). So far as I know no one has doubted that, when he uses this phrase, he is referring to Anaximander.

In a number of other places Aristotle speaks of some one who held the primary substance to be something “intermediate between” the elements or between two of them .62 Nearly all the Greek commentators referred this to Anaximander also, but most modern writers refuse to follow them. It is, no doubt, easy to show that Anaximander himself cannot have said anything of the sort, but that is no real objection. Aristotle puts things in his own way regardless of historical considerations, and it is difficult to see that it is more of an anachronism to call the Boundless “intermediate between the elements” than to say that it is “distinct from the elements.” Indeed, if once we introduce the elements at all, the former description is the more adequate of the two. At any rate, if we refuse to understand these passages as referring to Anaximander, we shall have to say that Aristotle paid a great deal of attention to some one whose very name has been lost, and who not only agreed with some of Anaximander's views, but also used some of his most characteristic expressions.63 We may add that in one or two places Aristotle certainly seems to identify the “intermediate” with the something “distinct from” the elements.64

There is even one passage in which he speaks of Anaximander's Boundless as a “mixture,” though his words may perhaps admit of another interpretation.65 But this is of no consequence for our interpretation of Anaximander. It is certain that he cannot have said anything about “elements,” which no one thought of before Empedokles, and no one could think of before Parmenides. The question has only been mentioned because it has given rise to a lengthy controversy, and because it throws light on the historical value of Aristotle's statements. From the point of view of his own system, these may be justified; but we shall have to remember in other cases that, when he seems to attribute an idea to some earlier thinker, we are not bound to take what he says in an historical sense.66

Anaximander's reason for conceiving the primary substance as boundless was, no doubt, as indicated by Aristotle, “that becoming might not fail.”67 It is not clear, however, that these words are his own, though the doxographers speak as if they were. It is enough for us that Theophrastos, who had seen his book, attributed the thought to him. And certainly his view of the world would bring home to him the need of a boundless stock of matter. The “opposites” are, we have seen, at war with one another, and their strife is marked by “unjust” encroachments on either side. The warm commits “injustice” in summer, the cold in winter, and this would lead in the long run to the destruction of everything but the Boundless itself, if there were not an inexhaustible supply of it from which opposites might continually be separated out afresh. We must picture, then, an endless mass, which is not any one of the opposites we know, stretching out without limit on every side of the world we live in.68 This mass is a body, out of which our world once emerged, and into which it will one day be absorbed again.

We are told that Anaximander believed there were “innumerable worlds in the Boundless,”69 and we have to decide between the interpretation that, though all the worlds are perishable, there are an unlimited number of them in existence at the same time, and Zeller's view that a new world never comes into existence till the old one has passed away, so that there is never more than one world at a time. As this point is of fundamental importance, it will be necessary to examine the evidence carefully.

In the first place, the doxographical tradition proves that Theophrastos discussed the views of all the early philosophers as to whether there was one world or an infinite number, and there can be no doubt that, when he ascribed “innumerable worlds” to the Atomists, he meant coexistent and not successive worlds. Now, if he had classed two such different views under one head, he would have been careful to point out in what respect they differed, and there is no trace of any such distinction. On the contrary, Anaximander, Anaximenes, Archelaos, Xenophanes, Diogenes, Leukippos, Demokritos, and Epicurus are all mentioned together as holding the doctrine of “innumerable worlds” on every side of this one,70 and the only distinction is that, while Epicurus made the distances between these worlds unequal, Anaximander said all the worlds were equidistant.71 Zeller rejected this evidence72 on the ground that we can have no confidence in a writer who attributes “innumerable worlds” to Anaximenes, Archelaos, and Xenophanes. With regard to the first two, I hope to show that the statement is correct, and that it is at least intelligible in the case of the last.73 In any case, the passage comes from Aetios,74 and there is no reason for doubting that it is derived from Theophrastos, though the name of Epicurus has been added later. This is confirmed by what Simplicius says:

Those who assumed innumerable worlds, e.g. Anaximander, Leukippos, Demokritos, and, at a later date, Epicurus, held that they came into being and passed away ad infinitum, some always coming into being and others passing away.75

It is practically certain that this too comes from Theophrastos through Alexander.

We come next to a very important statement which Cicero has copied from Philodemos, the author of the Epicurean treatise on Religion found at Herculaneum, or perhaps from the immediate source of that work. “Anaximander's opinion was,” he makes Velleius say, “that there were gods who came into being, rising and passing away at long intervals, and that these were the innumerable worlds”;76 and this must clearly be taken along with the statement of Aetios that, according to Anaximander, the “innumerable heavens” were gods.77 Now it is much more natural to understand the “long intervals” as intervals of space than as intervals of time;78 and, if that is right, we have a perfect agreement among our authorities.

It may be added that it is very unnatural to understand the statement that the Boundless “encompasses all the worlds” of worlds succeeding one another in time; for on this view there is at a given time only one world to “encompass.” Moreover, the argument mentioned by Aristotle that, if what is outside the heavens is infinite, body must be infinite, and there must be innumerable worlds, can only be understood in one sense, and is certainly intended to represent the reasoning of the Milesians ; for they were the only cosmologists who held there was a boundless body outside the heavens.79 Lastly, we happen to know that Petron, one of the earliest Pythagoreans, held there were just one hundred and eighty-three worlds arranged in a triangle,80 which shows at least that the doctrine of a plurality of worlds was much older than the Atomists.

The doxographers say it was the “eternal motion” that brought into being “all the heavens and all the worlds within them.” We have seen (§ VIII.) that this is probably only the Aristotelian way of putting the thing, and that we must not identify the primordial motion of the Boundless with any purely mundane movement such as the diurnal revolution. That would be quite inconsistent, moreover, with the doctrine of innumerable worlds, each of which has, presumably, its own centre and its own diurnal revolution. As to the true nature of this motion, we have no definite statement, but the term “separating off” (ἀπόκρισις) rather suggests some process of shaking and sifting as in a riddle or sieve. That is given in Plato's Timaeus as the Pythagorean doctrine,81 and the Pythagoreans followed Anaximander pretty closely in their cosmology (§ 54). The school of Abdera, as will be shown (§ 179), attributed a motion of the same kind to their atoms, and they too were mainly dependent on the Milesians for the details of their system. This, however, must remain a conjecture in the absence of express testimony.

When, however, we come to the motion of the world once it has been “separated off,” we are on safer ground. It is certain that one of the chief features of early cosmology is the part. played in it by the analogy of an eddy in water or in wind, a δίνη (or δῖνος),82 and there seems to be little doubt that we are entitled to regard this as the doctrine of Anaximander and Anaximenes.83 It would arise very naturally in the minds of thinkers who started with water as the primary substance and ended with “air,” and it would account admirably for the position of earth and water in the centre and fire at the circumference, with “air” between them. Heavy things tend to the centre of a vortex and light things are forced out to the periphery. It is to be observed that there is no question of a sphere in revolution at this date; what we have to picture is rotary motion in a plane or planes more or less inclined to the earth's surface.84 It is in favour of the conjecture given above as to the nature of the primordial motion that it provides a satisfactory dynamical explanation of the formation of the δίνη, and we shall find once more (§180) that the Atomists held precisely this view of its origin.

The doxographers also give us some indications of the process by which the different parts of the world arose from the Boundless. The following statement comes ultimately from Theophrastos:

He says that something capable of begetting hot and cold out of the eternal was separated off at the origin of this world. From this arose a sphere of flame which fitted close round the air surrounding the earth as the bark round a tree. When this had been torn off and shut up in certain rings, the sun, moon and stars came into existence.—Ps.-Plut. Strom. fr. 2 (R. P. 19).85

We see from this that, when a portion of the Boundless was separated off from the rest to form a world, it first differentiated itself into the two opposites, hot and cold. The hot appears as flame surrounding the cold; the cold, as earth with air surrounding it. We are not told here how the cold was differentiated into earth, water and air, but there is a passage in Aristotle's Meteorology which throws some light on the question. After discussing the views of the “theologians” regarding the sea, he says:

But those who are wiser in the wisdom of men give an origin for the sea. At first, they say, all the terrestrial region was moist; and, as it was dried up by the sun, the portion of it that evaporated produced the winds and the turnings back of the sun and moon,86 while the portion left behind was the sea. So they think the sea is becoming smaller by being dried up, and that at last it will all be dry. Meteor, B, 1. 353 b 5.

And the same absurdity arises for those who say the earth too was at first moist, and that, when the region of the world about the earth was heated by the sun, air was produced and the whole heavens were increased, and that it (the air) produced winds and caused its (the sun's) turnings back.87—Ib. 2. 355 a 21 (R. P. 20 a).

In his commentary on the passage, Alexander says this was the view of Anaximander and Diogenes, and cites Theophrastos as his authority for the statement. This is confirmed by Anaximander's theory of the sea as given by the doxographers (§ 20). We conclude, then, that after the first separation of the hot and the cold by the δίνη, the heat of the flame turned part of the moist, cold interior of the world into air or vapour—it is all one at this date—and that the expansion of this mist broke up the flame itself into rings. We shall come back to these rings presently, but we must look first at what we are told of the earth.

The origin of earth and sea from the moist, cold matter which was “separated off” in the beginning is thus described:

The sea is what is left of the original moisture. The fire has dried up most of it and turned the rest salt by scorching it. - Aet. iii. 16, 1 (R. P. 20 a).

He says that the earth is cylindrical in form, and that its depth is as a third part of its breadth.—Ps.-Plut. Strom fr. 2 (R. P. ib.).

The earth swings free, held in its place by nothing. It stays where it is because of its equal distance from everything. Its shape is hollow and round, and like a stone pillar. We are on one of the surfaces, and the other is on the opposite side.88—Hipp. Ref. i. 6 (R. P. 20).

Adopting for a moment the popular theory of “elements,” we see that Anaximander put fire on one side as the hot and dry, and all the rest on the other as the cold, which is also moist. This may explain how Aristotle came to speak of the Boundless as intermediate between fire and water. And we have seen also that the moist element was partly turned into “air” or vapour by the fire, which explains how Aristotle could say the Boundless was something between fire and air, or between air and water.89

The moist, cold interior of the world is not, in fact, water. It is always called “the moist” or “the moist state.” That is because it has to be still further differentiated under the influence of heat into earth, water, and vapour. The gradual drying up of the water by the fire is a good example of what Anaximander meant by “injustice.”

Thales had said that the earth floated on the water, but Anaximander realised that it was freely suspended in space (μετέωρος) and did not require any support. Aristotle has preserved the argument he used. The earth is equally distant from the circumference of the vortex in every direction, and there is no reason for it to move up or down or sideways.90 The doctrine of innumerable worlds was inconsistent with the existence of an absolute up and down in the universe, so the argument is quite sound. The central position of the earth is due to the δίνη; for the greater masses tend to the centre of an eddy.91 There is good evidence that Anaximander made the earth share in the rotary movement.92 It is not, however, a sphere, so we must not speak of an axial revolution. The shape given to the earth by Anaximander is easily explained if we adopt the view that the world is a system of rotating rings. It is just a solid ring in the middle of the vortex.

We have seen that the flame which had been forced to the circumference of the vortex was broken up into rings by the pressure of expanding vapour produced by its own heat. I give the statements of Hippolytos and Aetios as to the formation of the heavenly bodies from these rings.

The heavenly bodies are a wheel of fire, separated off from the fire of the world, and surrounded by air. And there are breathing-holes, certain pipe-like passages, at which the heavenly bodies show themselves. That is why, when the breathing-holes are stopped, eclipses take place. And the moon appears now to wax and now to wane because of the stopping and opening of the passages. The wheel of the sun is 27 times the size of (the earth, while that of) the moon is 18 times as large.93 The sun is the highest of all, and lowest are the wheels of the stars. —Hipp. Ref. i. 6 (R. P. 20).

The heavenly bodies were hoop-like compressions of air, full of fire, breathing out flames at a certain point through orifices.Aet. ii. 13, 7 (R. P. 19 a).

The sun was a wheel 28 times the size of the earth, like a chariot-wheel with the felloe hollow, full of fire, showing the fire at a certain point through an orifice, as through the nozzle of a pair of bellows.—Aet. ii. 20, i (R. P. 19 a).

The sun was equal to the earth, but the wheel from which it breathes out and by which it is carried round was 27 times the size of the earth.—Aet. ii. 21, 1.

The sun was eclipsed when the orifice of the fire's breathing-hole was stopped.—Aet. ii. 24., 2.

The moon was a wheel 19 times the size of the earth, like a chariot-wheel with its felloe hollow and full of fire like that of the sun, lying oblique also like it, with one breathing-hole like the nozzle of a pair of bellows. [It is eclipsed because of the turnings of the wheel.]94 —Aet. ii. 25, 1.

The moon was eclipsed when the orifice of the wheel was stopped.—Aet. ii. 29, 1.

(Thunder and lightning, etc.) were all caused by the blast of the wind. When it is shut up in a thick cloud and bursts forth with violence, then the tearing of the cloud makes the noise, and the rift gives the appearance of a flash in contrast with the blackness of the cloud.—Aet. iii. 3, 1.

Wind was a current of air (i.e. vapour), which arose when its finest and moistest particles were stirred or melted by the sun.—Aet. iii. 7, 1.

There is a curious variation in the figures given for the size of the wheels of the heavenly bodies, and it seems most likely that 18 and 27 refer to their inner, while 19 and 28 refer to their outer circumference. We may, perhaps, infer that the wheels of the “stars” were nine times the size of the earth; for the numbers 9, 18, 27 play a considerable part in primitive cosmogonies.95 We do not see the wheels of fire as complete circles; for the vapour or mist which formed them encloses the fire, and forms an outer ring except at one point of their circumference, through which the fire escapes, and that is the heavenly body we actually see.96 It is possible that the theory of “wheels” was suggested by the Milky Way. If we ask how it is that the wheels of air can make the fire invisible to us without becoming visible themselves, the answer is that such is the property of what the Greeks at this date called “air.” For instance, when a Homeric hero is made invisible by being clothed in “air,” we can see right through both the “air” and the hero.97 It should be added that lightning is explained in much the same way as the heavenly bodies. It, too, was fire breaking through condensed air, in this case storm clouds. It seems probable that this was really the origin of the theory, and that Anaximander explained the heavenly bodies on the analogy of lightning, not vice versa. It must be remembered that meteorology and astronomy were still undifferentiated,98 and that the theory of “wheels” or rings is a natural inference from the idea of the vortex.

So far we seem to be justified, by the authority of Theophrastos, in going; and, if that is so, certain further inferences seem to be inevitable. In the first place, Anaximander had shaken himself free of the old idea that the heavens are a solid vault. There is nothing to prevent us from seeing right out into the Boundless, and it is hard to think that Anaximander did not believe he did. The traditional cosmos has given place to a much grander scheme, that of innumerable vortices in a boundless mass, which is neither water nor air. In that case, it is difficult to resist the belief that what we call the fixed stars were identified with the “innumerable worlds” which were also “gods.” It would follow that the diurnal revolution is only apparent; for the stars are at unequal distances from us, and can have no rotation in common. It must, then, be due to the rotation of the cylindrical earth in twenty-four hours. We have seen that the earth certainly shared in the rotation of the δίνη. That gets rid of one difficulty, the wheel of the “stars,” which is between the earth and the moon; for the fixed stars could not be explained by a “wheel” at all; a sphere would be required. What, then, are the “stars” which are accounted for by this inner wheel? I venture to suggest that they are the morning and the evening stars, which, we have seen (p. 23, n. 1), were not recognised yet as a single luminary. In other words, I believe that Anaximander regarded the fixed stars as stationary, each rotating in its own vortex. No doubt this involves us in a difficulty regarding the rotation of the sun and the moon. It follows from the nature of the vortex that they must rotate in the same direction as the earth, and, on the assumption just made, that must be from west to east, and it must be a slower rotation than that of the earth, which is inconsistent with the fact that the circumference of a vortex rotates more rapidly than the centre. That, however, is a difficulty which all the Ionian cosmologists down to Demokritos had to face. Holding, as they did, that the whole rotation was in the same direction, they had to say that what we call the greatest velocities were the least. The moon, for instance, did not rotate so rapidly as the sun, since the sun more nearly keeps up with the fixed stars.99 That Anaximander failed to observe this difficulty is not surprising, if we remember that he was the first to attack the problem. It is not immediately obvious that the centre of the vortex must have a slower motion than the circumference. This serves to explain the origin of the theory that the heavenly bodies have a rotation of their own in the opposite direction to the diurnal revolution which we shall see reason for attributing to Pythagoras (§ 54).

We have, in any case, seen enough to show us that the speculations of Anaximander about the world were of an extremely daring character. We come now to the crowning audacity of all, his theory of the origin of living creatures. The Theophrastean account of this has been well preserved by the doxographers:

Living creatures arose from the moist element as it was evaporated by the sun. Man was like another animal, namely, a fish, in the beginning.—Hipp. Ref. i. 6 (R. P. 22 a).

The first animals were produced in the moisture, each enclosed in a prickly bark. As they advanced in age, they came out upon the drier part. When the bark broke off,100 they survived for a short time.101—Aet. v. 19, 4 (R. P. 22).

Further, he says that originally man was born from animals of another species. His reason is that while other animals quickly find food by themselves, man alone requires a lengthy period of suckling. Hence, had he been originally as he is now, he would never have survived.—Ps.-Plut. Strom. fr. 2 (R. P. ib.).

He declares that at first human beings arose in the inside of fishes, and after having been reared like sharks,102 and become capable of protecting themselves, they were finally cast ashore and took to land.—Plut. Symp. Quaest. 730 f (R. P. ib.).

The importance of these statements has sometimes been overrated and still more often underestimated. Anaximander has been called a precursor of Darwin by some, while others have treated the whole thing as a mythological survival. It is therefore important to notice that this is one of the rare cases where we have not merely a placitum, but an indication of the observations on which it was based. It is clear from this that Anaximander had an idea of what is meant by adaptation to environment and survival of the fittest, and that he saw the higher mammals could not represent the original type of animal. For this he looked to the sea, and he naturally fixed upon those fishes which present the closest analogy to the mammalia. The statements of Aristotle about the galeus levis were shown by Johannes Müller to be more accurate than those of later naturalists, and we now see that these observations were already made by Anaximander. The way in which the shark nourishes its young furnished him with the very thing he required to explain the survival of the earliest animals.103

45. R. P. 15 d. That the words πολίτης καὶ ἑταῖρος, given by Simplicius, De caelo, p. 615, 13, are from Theophrastos is shown by the agreement of Cic. Acad. ii. 118, popularis et sodalis. The two passages represent independent branches of the tradition. See Note on Sources, §§ 7, 12.

46. Diog. ii. 2 (R. P. 15); Hipp. Ref. i. 6 (Dox. p. 560); Plin. N.H. ii. 31.

47. Xenophanes, fr. 22 (= fr. 17 Karsten; R. P. 95 a).

48. The statement that he “died soon after” (Diog. ii. 2; R. P. 15) seems to mean that Apollodoros made him die in the year of Sardeis (546/5), one of his regular epochs.

49. For the gnomon, see Introd. p. 26, n. 1; and cf. Diog. ii. 1 (R. P. 15); Herod. ii. 109 (R. P. 15 a). Pliny, on the other hand, ascribes the invention of the gnomon to Anaximenes (N.H. ii. 187).

50. Aelian, V.H. iii. 17. Presumably Apollonia on the Pontos is meant.

51. The lower part of a contemporary statue has been discovered at Miletos (Wiegand, Milet, ii. 88), with the inscription ΑΝ]ΑΞΙΜΑΝΔΡΟ. It was not, we may be sure, for his theories of the Boundless that Anaximander received this honour; he was a statesman and an inventor, like Thales and Hekataios.

52. In this and other cases, where the words of the original have been preserved by Simplicius, I have given them alone. On the various writers quoted, see Note on Sources, §§ 9 sqq.

53. Simplicius says “successor and disciple” (διάδοχος καὶ μαθητής) in his Commentary on the Physics; but see above, p. 50, n. 4.

54. For the expression τὰ καλούμενα στοιχεῖα, see Diels, Elementum, p. 25, n. 4.

55. Diels (Vors. 2, 9) begins the actual quotation with the words ἐξ ὧν δὲ ἡ γένεσις . . . The Greek practice of blending quotations with the text tells against this. Further, it is safer not to ascribe the terms γένεσις and φθορά in their technical Platonic sense to Anaximander, and it is not likely that Anaximander said anything about τὰ ὄντα.

56. See p. 12, n. 2.

57. The important word ἀλλήλοις is in all the MSS. of Simplicius, though omitted in the Aldine. This omission made the sentence appear to mean that the existence of individual things (ὄντα) was somehow a wrong (ἀδικία) for which they must be punished. With ἀλλήλοις restored, this fanciful interpretation disappears. It is to one another that whatever the subject of the verb may be make reparation and give satisfaction, and therefore the injustice must be a wrong which they commit against one another. Now, as δίκη is regularly used of the observance of an equal balance between the opposites hot and cold, dry and wet, the ἀδικία here referred to must be the undue encroachment of one opposite on another, such as we see, for example, in the alternation of day and night, winter and summer, which have to be made good by an equal encroachment of the other. I stated this view in my first edition (1892), pp. 60-62, and am glad to find it confirmed by Professor Heidel (Class. Phil. vii., 1912, p. 233 sq.).

58. The words of Theophrastos, as given by Simplicius (Phys. p. 24, 15: R. P. 16), are ἀρχήν τε καὶ στοιχεῖον εἴρηκε τῶν ὄντων τὸ ἄπειρον, πρῶτος τοῦτο τοὔνομα κομίσας τῆς ἀρχῆς, the natural meaning of which is “he being the first to introduce this name (τὸ ἄπειρον) of the material cause.” Hippolytos, however, says (Ref. i. 6, 2) πρῶτος τοὔνομα καλέσας τῆς ἀρχῆς, and this has led most writers to take the words in the sense that Anaximander introduced the term ἀρχή. Hippolytos, however, is not an independent authority (see Note on Sources, § 13), and the only question is what Theophrastos wrote. Now Simplicius quotes Theophrastos from Alexander, who used the original, while Hippolytos represents a much more indirect tradition. Obviously, καλέσας is a corruption of the characteristically Peripatetic κομίσας, and the omission of τοῦτο is much more likely than its interpolation by Alexander or Simplicius. But, if τοῦτο is genuine, the ὄνομα referred to must be τὸ ἄπειρον, and this interpretation is confirmed by Simpl. De caelo 615, 15, ἄπειρον δὲ πρῶτος ὑπέθετο. In another place (p. 150, 23) Simplicius says πρῶτος αὐτὸς ἀρχὴν ὀνομάσας τὸ ὑποκείμενον, which must mean, as the context shows, “being the first to name the substratum of the opposites as the material cause,” which is another point altogether. Theophrastos is always interested in noting who it was that “first” introduced a concept, and both ἄπειρον and ὑποκείμενον were important enough to be noted. Of course he does not mean that Anaximander used the word ὑποκείμενον. He only infers that he had the idea from the doctrine that the opposites which are “in” the ἄπειρον are “separated out.” Lastly, the whole book from which these extracts were taken was Περὶ τῶν ἀρχῶν, and the thing to note was who first applied various predicates to the ἀρχή or ἀρχαί.

59. See p. 47 n. 6 and Introd. p. 11 n. 3.

60. Arist. Met. Λ, 2. 1069 b 18 (R. P. 16 c).

61. This is taken for granted in Phys. Γ, 4. 203 a 16; 204 b 22 (R. P. 16 b), and stated in Γ, 8. 208 a 8 (R. P. 16 a). Cf. Simpl. Phys. p. 150, 20 (R. P. 18).

62. Aristotle speaks four times of something intermediate between Fire and Air (Gen. Corr. B, 1. 328 b 35; ib. 5. 332 a 21; Phys. A, 4. 187 a 14; Met. A, 7. 988 a 30). In five places we have something intermediate between Water and Air (Met. A, 7. 988 a 13; Gen. Corr. B, 5. 332 a 21; Phys. Γ, 4. 203 a 18; ib. 5. 205 a 27; De caelo, Γ, 5. 303 b 12). Once (Phys. A, 6. 189 b 1) we hear of something between Water and Fire. This variation shows at once that he is not speaking historically. If any one ever held the doctrine of τὸ μεταξύ, he must have known which “elements” he meant.

63. Arist. De caelo, Γ, 5. 303 b 12, ὕδατος μὲν λεπτότερον, ἀέρος δὲ πυκνότερον, ὃ περιέχειν φασὶ πάντας τοὺς οὐρανοὺς ἄπειρον ὄν.

64. cf. Phys. Γ, 5. 204 b 22 (R. P. 16 b), where Zeller rightly refers τὸ παρὰ τὰ στοιχεῖα to Anaximander. Now, at the end (205 a 25) the whole passage is summarised thus: καὶ διὰ τοῦτ' οὐθεὶς τὸ ἓν καὶ ἄπειρον πῦρ ἐποίησεν οὐδὲ γῆν τῶν φυσιολόγων, ἀλλ' ἢ ὕδωρ ἢ ἀέρα ἢ τὸ μέσον αὐτῶν. In Gen. Corr. B, 1. 328 b 35 we have first τι μεταξὺ τούτων σῶμά τε ὂν καὶ χωριστόν, and a little further on (329 a 9) μίαν ὕλην παρὰ τὰ εἰρημένα. In B, 5. 332 a 20 we have οὐ μὴν οὐδ' ἄλλο τί γε παρὰ ταῦτα, οἷον μέσον τι ἀέρος καὶ ὕδατος ἢ ἀέρος καὶ πυρός.

65. Met. Λ, 2. 1069 b 18 (R. P. 16 c). Zeller (p. 205, n. 1) assumes an “easy zeugma.”

66. For the literature of this controversy, see R. P. 15. Professor Heidel has shown in his “Qualitative Change in Pre-Socratic Philosophy” (Arch., xix. p. 333) that Aristotle misunderstood the Milesians because he could only think of their doctrine in terms of his own theory of ἀλλοίωσις. That is quite true, but it is equally true that they had no definite theory of their own with regard to the transformations of substance. The theory of an original “mixture” is quite as unhistorical as that of ἀλλοίωσις. Qualities were not yet distinguished from “things,” and Thales doubtless said that water turned into vapour or ice without dreaming of any further questions. They all believed that in the long run there was only one “thing,” and at last they came to the conclusion that all apparent differences were due to rarefaction and condensation. Theophrastos (ap. Simpl. Phys. 150, 22) says ἐνούσας γὰρ τὰς ἐναντιότας ἐν τῷ ὑποκειμένῳ . . . ἐκκρίνεσθαι. I do not believe these words are even a paraphrase of anything Anaximander said. They are merely an attempt to “accommodate” his views to Peripatetic ideas, and ἐνούσας is as unhistorical as the ὑποκείμενον.

67. Phys. Γ, 8. 208 a 8 (R. P. 16 a). Cf. Aet. i. 3, 3 (R. P. 16 a). The same argument is given in Phys. Γ, 4. 203 b 18, a passage where Anaximander has just been named, τῷ οὕτως ἂν μόνον μὴ ὑπολείπειν γένεσιν καὶ φθοράν, εἰ ἄπειρον εἴη ὅθεν ἀφαιρεῖται τὸ γιγνόμενον. I cannot, however, believe that the arguments at the beginning of this chapter (203 b 7; R. P. 17) are Anaximander's. They bear the stamp of the Eleatic dialectic, and are, in fact, those of Melissos.

68. I have assumed that the word ἄπειρον means spatially infinite, not qualitatively indeterminate, as maintained by Teichmüller and Tannery. The decisive reasons for holding that the sense of the word is “boundless in extent” are as follows: (1) Theophrastos said the primary substance of Anaximander was ἄπειρον and contained all the worlds, and the word περιέχειν everywhere means “to encompass,” not, as has been suggested, “to contain potentially.” (2) Aristotle says (Phys. Γ, 4. 203 b 23) διὰ γὰρ τὸ ἐν τῇ νοήσει μὴ ὑπολείπειν καὶ ὁ ἀριθμὸς δοκεῖ ἄπειρος εἶναι καὶ τὰ μαθηματικὰ μεγέθη καὶ τὰ ἔξω τοῦ οὐρανοῦ· ἀπείρου δ' ὄντος τοῦ ἔξω, καὶ σῶμα ἄπειρον εἶναι δοκεῖ καὶ κόσμοι. The mention of σῶμα shows that this does not refer to the Atomists. (3) Anaximander's theory of the ἄπειρον was adopted by Anaximenes, and he identified it with Air, which is not qualitatively indeterminate.

69. Cf. [Plut.] Strom. fr. 2 (R. P. 21 b).

70. Aet. ii. 1, 3 (Dox. p. 327). Zeller seems to be wrong in understanding κατὰ πᾶσαν περιαγωγήν here of revolution. It must mean “in every direction we turn,” as is shown by the alternative phrase κατὰ πᾶσαν περίστασιν. The six περιστάσεις are πρόσω, ὀπίσω, ἄνω, κάτω, δεξιά, ἀριστερά (Nicom. Introd. p. 85, 11, Hoche).

71. Aet. ii. 1, 8 (Dox. p. 329), τῶν ἀπείρους ἀποφηναμένων τοὺς κόσμους Ἀναξίμανδρος τὸ ἴσον αὐτοὺς ἀπέχειν ἀλλήλων, Ἐπίκουρος ἄνισον εἶναι τὸ μεταξὺ τῶν κόσμων διάστημα.

72. He supposed it to be only that of Stobaios. The filiation of the sources had not been traced when he wrote.

73. For Anaximenes see § 30; Xenophanes, § 59; Archelaos, § 192.

74. This is proved by the fact that the list of names is given also by Theodoret. See Note on Sources, § 10.

75. Simpl. Phys. p. 1121, 5 (R. P. 21 b). Cf. Simpl. De caelo, p. 202, 14, οἱ δὲ καὶ τῷ πλήθει ἀπείρους κόσμους, ὡς Ἀναξίμανδρος . . . ἄπειρον τῳ μεγέθει τὴν ἀρχὴν θέμενος ἀπείρους ἐξ αὐτοῦ τῷ πλήθει κόσμους ποιεῖν δοκεῖ.

76. Cicero, De nat. d. i. 25 (R. P. 21).

77. Aet. i. 7, 12 (R. P. 21 a). The reading of Stob., ἀπείρους οὐρανούς, is guaranteed by the ἀπείρους κόσμους of Cyril, and the ἀπείρους νοῦς (i.e. ουνους) of the pseudo-Galen. See Dox. p. 11.

78. It is natural to suppose that Cicero found διαστήμασιν in his Epicurean source, and that is a technical term for the intermundia.

79. Arist. Phys. Γ, 4. 203 b 25, ἀπείρου δ' ὄντος τοῦ ἔξω (sc. τοῦ οὐρανοῦ), καὶ σῶμα ἄπειρον εἶναι δοκεῖ καὶ κόσμοι (ἄπειροι). The next words—τί γὰρ μᾶλλον τοῦ κενοῦ ἐνταῦθα ἢ ἐνταῦθα—show that this refers to the Atomists as well; but the ἄπειρον σῶμα will not apply to them. The meaning is that both those who made the Boundless a body and those who made it a κενόν held the doctrine of ἄπειροι κόσμοι in the same sense.

80. See below, § 53. Cf. Diels, Elementum, pp. 63 sqq.

81. Plato, Tim. 52 e. There the elemental figures (which have taken the place of the “opposites”) “being thus stirred (by the irregular motion of the τιθήνη), are carried in different directions and separated, just as by sieves and instruments for winnowing corn the grain is shaken and sifted; and the dense and heavy parts go one way, while the rare and light are carried to a different place and settle there.

82. Aristophanes, referring to the Ionian cosmology, says (Clouds, 828) Δῖνος βασιλεύει τὸν Δι' ἐξεληλακώς, which is nearer the truth than the modern theory of its religious origin.

83. I gratefully accept the view propounded by Prof. W. A. Heidel (“The δίνη in Anaximenes and Anaximander,” Class. Phil. i. 279), so far as the cosmical motion goes, though I cannot identify that with the “eternal motion.” I had already done what I could to show that the “spheres” of Eudoxos and Aristotle must not be imported into Pythagoreanism, and it strengthens the position considerably if we ascribe a rotary motion in a plane to Anaximander's world.

84. This is the plain meaning of Aet. ii. 2, 4, οἱ δὲ τροχοῦ δίκην περιδινεῖσθαι τὸν κόσμον, which is referred to Anaximander by Diels (Dox. p. 46). Zeller's objections to the ascription of the δίνη to Anaximander are mainly based on an inadmissible rendering of the word τροπαί (p. 63 n. 2). Of course, the rotations are not all in the same plane; the ecliptic, for instance, is inclined to the equator, and the Milky Way to both.

85. This passage has been discussed by Heidel (Proceedings of the American Academy, xlviii. 686). I agree that ἀπὸ τοῦ ἀπείρου must be supplied with ἀποκριθῆναι, and I formerly thought that ἐκ τοῦ αἰδίου might be equivalent to that, and might have been displaced if the order of words was too harsh. I cannot believe that it means “from eternity,” as Heidel thinks. On the other hand, he is clearly right in his interpretation of περιφυῆναι and ἀπορραγείσης. He also points out correctly that “the sphere of flame” is an inaccuracy. The comparison to the bark of a tree distinctly suggests something annular.

86. Zeller (p. 223, n. 5) asks what can be meant by τροπαὶ τῆς σελήνης, but his difficulty is an imaginary one. The moon has certainly a movement in declination and therefore τροπαί. In other words, the moon does not always rise at the same point of the horizon any more than the sun. This is admitted by Sir T. L. Heath (Aristarchus, p. 33, n. 3), though he has unfortunately followed Zeller in supposing that τροπαί here means “revolutions.” This seems to me impossible; for τρέπεσθαι means “to turn back” or “to turn aside,” never “to turn round,” which is στρέφεσθαι. It is conceivable, indeed, that τροπαὶ ἠελίοιο in Od. xv. 404 means the place where the sun sets and turns back from west to east, though it is not very likely, as Hesiod already uses τροπαὶ ἠελίοιο of the winter and summer solstices (O.D. 479, 564, 663). Zeller's statement (repeated by Heath) that Aristotle speaks of τροπαί of the fixed stars in De caelo, B, 14. 296 b 4, is erroneous. What Aristotle does say is that, if the earth is in motion, there ought to be πάροδοι (movements in 1atitude) and τροπαί of the fixed stars, which there are not. The passage is correctly rendered by Sir T. L. Heath himself in a subsequent chapter (p. 241). For the other passages referred to, see p. 64, n. 1, and p. 76, n. 3.

87. From the whole context it is plain that τὰς τροπὰς αὐτοῦ means τὰς τοῦ ἡλίου τροπάς, and not τὰς τοῦ οὐρανοῦ, as Zeller and Heath say. The “air” in this passage answers to “the portion that evaporated” (τὸ διατμίσαν) in that previously quoted, and τοῦτον must therefore refer to it. Cf. the paraphrase of Alexander (p. 67, 3 from Theophrastos, Dox. p. 494). τὸ μέν τι τῆς ὑγρότητος ὑπὸ τοῦ ἡλίου ἐξατμίζεσθαι καὶ γίνεσθαι πνεύματά τε ἐξ αὐτοῦ καὶ τροπὰς ἡλίου τε καὶ σελήνης (see last note). In this chapter of the Meteorology, Aristotle is discussing the doctrine that the sun is “fed” by moisture and the relation of that doctrine to its τροπαί at the solstices, and we must interpret accordingly.

88. The MSS. of Hippolytos have ὑγρὸν στρογγύλον, and so has Cedrenus, a writer of the eleventh century who made extracts from him. Roeper read γυρὸν [στρογγύλον], supposing the second word to be a gloss on the first. Diels (Dox. p. 218) holds that the first applies to the surface of the earth; while the second refers to its circuit. Professor A. E. Taylor has pointed out to me, however, the great improbability of the view that γυρόν means convex. The Ionians down to Archelaos (§ 192) and Demokritos (Aet. iii. 10, 5, κοίλην τῷ μέσῳ) regularly regarded the surface of the earth as concave, and γυρός can just as well mean that. The next words are also of doubtful meaning. The MSS. of Hippolytos have χίονι λίθῳ, while Aetios (iii. 10, 2) has λίθῳ κίονι. Diels doubtfully conjectures λίθῳ κίονι, which he suggests might represent an original λιθέῃ κίονι (Dox. p. 219). In any case the pillar seems genuine, and the general sense is guaranteed by the Plutarchean Stromateis (loc. cit.), ὑπάρχειν . . . τῷ μὲν σχήματι τὴν γῆν κυλινδροειδῆ.

89. See above, p. 55, n. 4.

90. Arist. De caelo, B, 13. 295 b 10 εἰσὶ δέ τινες οἳ διὰ τὴν ὁμοιότητά φασιν αὐτὴν (τὴν γῆν) μένειν, ὥσπερ τῶν ἀρχαίων Ἀναξίμανδρος· μᾶλλον μὲν γὰρ οὐθὲν ἄνω ἢ κάτω ἢ εἰς τὰ πλάγια φέρεσθαι προσήκειν τὸ ἐπὶ τοῦ μέσου ἱδρυμένον καὶ ὁμοίως πρὸς τὰ ἔσχατα ἔχον. One point of the δίνη is no more “down” than another. Apparently, the Pythagoreans adopted this reasoning; for Plato makes Sokrates in the Phaedo say (108 e) [Perseus 109a] ἰσόρροπον γὰρ πρᾶγμα ὁμοίου τινὸς ἐν μέσῳ τεθὲν οὐχ ἕξει μᾶλλον οὐδὲ ἧττον οὐδαμόσε κλιθῆναι. From this it appears that ὁμοιότης means something like “indifference.” There is nothing to differentiate one radius of a circle from another.

91. Arist. De caelo, B, 13. 295 a 9 (ἡ γῆ) συνῆλθεν ἐπὶ τὸ μέσον φερομένη διὰ τὴν δίνησιν· ταύτην γὰρ τὴν αἰτίαν πάντες λέγουσιν ἐκ τῶν ἐν τοῖς ὑγροῖς καὶ περὶ τὸν ἀέρα συμβαινόντων· ἐν τούτοις γὰρ ἀεὶ φέρεται τὰ μείζω καὶ τὰ βαρύτερα πρὸς τὸ μέσον τῆς δίνης. διὸ δὴ καὶ τὴν γῆν πάντες ὅσοι τὸν οὐρανὸν γεννῶσιν ἐπὶ τὸ μέσον συνελθεῖν φασιν.

92. This was expressly stated by Eudemos (ap. Theon. Smyrn. p. 198, Ἀναξίμανδρος δὲ ὅτι ἐστὶν ἡ γῆ μετέωρος καὶ κινεῖται περὶ τὸ μέσον. Anaxagoras held the same view (§ 133).

93. I assume with Diels (Dox. p. 56o) that something has fallen out of the text, but I have made the moon's circle 18 and not 19 times as large, as agreeing better with the other figure, 27. See p. 68, n. 1.

94. There is clearly some confusion here, as Anaximander's real account of lunar eclipses is given in the next extract. There is also some doubt about the reading. Both Plutarch and Eusebios (P.E. xv. 26, 1) have ἐπιστροφάς, so the τροπάς of Stob. may be neglected, especially as the codex Sambuci had στροφάς. It looks as if this were a stray reference to the theory of Herakleitos that eclipses were due to a στροφή or ἐπιστροφή of the σκάφη (§ 71). In any case, the passage cannot be relied on in support of the meaning given to τροπαί by Zeller and Heath (p. 63, n. 2).

95. See Tannery, Science hellène, p. 91; Diels, “Ueber Anaximanders Kosmos” (Arch. x. pp. 231 sqq.).

96. The true meaning of this doctrine was first explained by Diels (Dox. pp. 25 sqq.). The flames issue per magni circum spiracula mundi, as Lucretius has it (vi. 493). The πρηστῆρος αὐλός, to which these are compared, is simply the mouthpiece of the smith's bellows, a sense the word πρηστήρ has in Apollonios of Rhodes (iv. 776), and has nothing to do with the meteorological phenomenon of the same name (see Chap: III. § 71), except that the Greek sailors very likely named the fiery waterspout after the familiar instrument. It is not necessary now to discuss the earlier interpretations of the phrase.

97. This is not so strange a view as might appear. An island or a rock in the offing may disappear completely when shrouded in mist (ἀήρ), and we seem to see the sky beyond it.

98. See above, p. 27.

99. Lucretius, v. 619 sqq.

100. This is to be understood in the light of what we are told about γαλεοί below. Cf. Arist. Hist. An. Z, l0. 565 a 25, τοῖς μὲν οὖν σκυλίοις, οὓς καλοῦσί τινες νεβρίας γαλεούς, ὅταν περιρραγῇ καὶ ἐκπέσῃ τὸ ὄστρακον, γίνονται οἱ νεοττοί.

101. The true reading is ἐπ' ὀλίγον χρόνον μεταβιῶναι, the omission of χρόνον by Diels in Vors.1 and Vors.2 being apparently a slip. In the Index to Dox., Diels s.v. μεταβιοῦν says “mutare vitam [cf. μεταδιαιτᾶν],” and I followed him in my first edition. Heidel well compares Archelaos, ap. Hipp. Ref. i. 9, 5 (of the first animals) ἦν δὲ ὀλιγοχρόνια.

102. Reading ὥσπερ οἱ γαλεοί for ὥσπερ οἱ παλαιοί with Doehner, who compares Plut. De soll. anim. 982 a, where the φιλόστοργον of the shark is described.

103. On Aristotle and the galeus levis, see Johannes Müller, “Ueber den glatten Hai des Aristoteles” (K. Preuss. Akad., 1842), to which my attention was directed by my colleague, Professor D'Arcy Thompson. The precise point of the words τρεφόμενοι ὥσπερ οἱ γαλεοί appears from Arist. Hist. An. Z, l0. 565 b 1, οἱ δὲ καλούμενοι λεῖοι τῶν γαλεῶν τὰ μὲν ᾠὰ ἴσχουσι μεταξὺ τῶν ὑστερῶν ὁμοίως τοῖς σκυλίοις, περιστάντα δὲ ταῦτα εἰς ἑκατέραν τὴν δικρόαν τῆς ὑστέρας καταβαίνει, καὶ τὰ ζῷα γίνεται τὸν ὀμφαλὸν ἔχοντα πρὸς τῇ ὑστέρᾳ, ὥστε ἀναλισκομένων τῶν ᾠ ῶν ὁμοίως δοκεῖν ἔχειν τὸ ἔμβρυον τοῖς τετράποσιν. It is not necessary to suppose that Anaximander referred to the further phenomenon described by Aristotle, who more than once says that all the γαλεοί except the ἀκανθίας “send out their young and take them back again” (ἐξαφιᾶσι καὶ δέχονται εἰς ἑαυτοὺς τοὺς νεοττούς, ib. 565 b 23), for which compare also Ael. i. 17; Plut. De amore prolis 494 c ; De soll. anim. 982 a. The placenta and umbilical cord described by Johannes Müller will account sufficiently for all he says.