The welding of incongruous elements is very apparent in the Semitic Version. For the statement that man will be created in order that the gods may have worshippers is at once followed by the announcement that the gods themselves must be punished and their "ways" changed. In the Sumerian Version the gods are united and all are naturally regarded as worthy of man's worship. The Sumerian Creator makes no distinctions; he refers to "our houses", or temples, that shall be established. But in the later version divine conflict has been introduced, and the future head of the pantheon has conquered and humiliated the revolting deities. Their "ways" must therefore be altered before they are fit to receive the worship which was accorded them by right in the simpler Sumerian tradition. In spite of the epitomized character of the Sumerian Version, a comparison of these passages suggests very forcibly that the Semitic-Babylonian myth of Creation is based upon a simpler Sumerian story, which has been elaborated to reconcile it with the Dragon myth.

The Semitic poem itself also supplies evidence of the independent existence of the Dragon myth apart from the process of Creation, for the story of Ea and Apsû, which it incorporates, is merely the local Dragon myth of Eridu. Its inclusion in the story is again simply a tribute to Marduk; for though Ea, now become Marduk's father, could conquer Apsû, he was afraid of Tiamat, "and turned back".(1) The original Eridu myth no doubt represented Enki as conquering the watery Abyss, which became his home; but there is nothing to connect this tradition with his early creative activities. We have long possessed part of another local version of the Dragon myth, which describes the conquest of a dragon by some deity other than Marduk; and the fight is there described as taking place, not before Creation, but at a time when men existed and cities had been built.(2) Men and gods were equally terrified at the monster's appearance, and it was to deliver the land from his clutches that one of the gods went out and slew him. Tradition delighted to dwell on the dragon's enormous size and terrible appearance. In this version he is described as fifty bêru(3) in length and one in height; his mouth measured six cubits and the circuit of his ears twelve; he dragged himself along in the water, which he lashed with his tail; and, when slain, his blood flowed for three years, three months, a day and a night. From this description we can see he was given the body of an enormous serpent.(4)

A further version of the Dragon myth has now been identified on one of the tablets recovered during the recent excavations at Ashur,(1) and in it the dragon is not entirely of serpent form, but is a true dragon with legs. Like the one just described, he is a male monster. The description occurs as part of a myth, of which the text is so badly preserved that only the contents of one column can be made out with any certainty. In it a god, whose name is wanting, announces the presence of the dragon: "In the water he lies and I (. . .)!" Thereupon a second god cries successively to Aruru, the mother-goddess, and to Pallil, another deity, for help in his predicament. And then follows the description of the dragon:

The text here breaks off, at the moment when Pallil, whose help against the dragon had been invoked, begins to speak. Let us hope we shall recover the continuation of the narrative and learn what became of this carnivorous monster.

There are ample grounds, then, for assuming the independent existence of the Babylonian Dragon-myth, and though both the versions recovered have come to us in Semitic form, there is no doubt that the myth itself existed among the Sumerians. The dragon motif is constantly recurring in descriptions of Sumerian temple-decoration, and the twin dragons of Ningishzida on Gudea's libation-vase, carved in green steatite and inlaid with shell, are a notable product of Sumerian art.(1) The very names borne by Tiamat's brood of monsters in the "Seven Tablets" are stamped in most cases with their Sumerian descent, and Kingu, whom she appointed as her champion in place of Apsû, is equally Sumerian. It would be strange indeed if the Sumerians had not evolved a Dragon myth,(2) for the Dragon combat is the most obvious of nature myths and is found in most mythologies of Europe and the Near East. The trailing storm-clouds suggest his serpent form, his fiery tongue is seen in the forked lightning, and, though he may darken the world for a time, the Sun-god will always be victorious. In Egypt the myth of "the Overthrowing of Apep, the enemy of Ra" presents a close parallel to that of Tiamat;(3) but of all Eastern mythologies that of the Chinese has inspired in art the most beautiful treatment of the Dragon, who, however, under his varied forms was for them essentially beneficent. Doubtless the Semites of Babylonia had their own versions of the Dragon combat, both before and after their arrival on the Euphrates, but the particular version which the priests of Babylon wove into their epic is not one of them.

We have thus traced four out of the five strands which form the Semitic-Babylonian poem of Creation to a Sumerian ancestry. And we now come back to the first of the strands, the Birth of the Gods, from which our discussion started. For if this too should prove to be Sumerian, it would help to fill in the gap in our Sumerian Creation myth, and might furnish us with some idea of the Sumerian view of "beginnings", which preceded the acts of creation by the great gods. It will be remembered that the poem opens with the description of a time when heaven and earth did not exist, no field or marsh even had been created, and the universe consisted only of the primaeval water-gods, Apsû, Mummu, and Tiamat, whose waters were mingled together. Then follows the successive generation of two pairs of deities, Lakhmu and Lakhamu, and Anshar and Kishar, long ages separating the two generations from each other and from the birth of the great gods which subsequently takes place. In the summary of the myth which is given by Damascius(1) the names of the various deities accurately correspond to those in the opening lines of the poem; but he makes some notable additions, as will be seen from the following table:

In the passage of the poem which describes the birth of the great gods after the last pair of primaeval deities, mention is duly made of Anu and Nudimmud (the latter a title of Ea), corresponding to the {'Anos} and {'Aos} of Damascius; and there appears to be no reference to Enlil, the original of {'Illinos}. It is just possible that his name occurred at the end of one of the broken lines, and, if so, we should have a complete parallel to Damascius. But the traces are not in favour of the restoration;(1) and the omission of Enlil's name from this part of the poem may be readily explained as a further tribute to Marduk, who definitely usurps his place throughout the subsequent narrative. Anu and Ea had both to be mentioned because of the parts they play in the Epic, but Enlil's only recorded appearance is in the final assembly of the gods, where he bestows his own name "the Lord of the World"(2) upon Marduk. The evidence of Damascius suggests that Enlil's name was here retained, between those of Anu and Ea, in other versions of the poem. But the occurrence of the name in any version is in itself evidence of the antiquity of this strand of the narrative. It is a legitimate inference that the myth of the Birth of the Gods goes back to a time at least before the rise of Babylon, and is presumably of Sumerian origin.

Further evidence of this may be seen in the fact that Anu, Enlil, and Ea (i.e. Enki), who are here created together, are the three great gods of the Sumerian Version of Creation; it is they who create mankind with the help of the goddess Ninkharsagga, and in the fuller version of that myth we should naturally expect to find some account of their own origin. The reference in Damascius to Marduk ({Belos}) as the son of Ea and Damkina ({Dauke}) is also of interest in this connexion, as it exhibits a goddess in close connexion with one of the three great gods, much as we find Ninkharsagga associated with them in the Sumerian Version.(1) Before leaving the names, it may be added that, of the primaeval deities, Anshar and Kishar are obviously Sumerian in form.

It may be noted that the character of Apsû and Tiamat in this portion of the poem(1) is quite at variance with their later actions. Their revolt at the ordered "way" of the gods was a necessary preliminary to the incorporation of the Dragon myths, in which Ea and Marduk are the heroes. Here they appear as entirely beneficent gods of the primaeval water, undisturbed by storms, in whose quiet depths the equally beneficent deities Lakhmu and Lakhamu, Anshar and Kishar, were generated.(2) This interpretation, by the way, suggests a more satisfactory restoration for the close of the ninth line of the poem than any that has yet been proposed. That line is usually taken to imply that the gods were created "in the midst of (heaven)", but I think the following rendering, in connexion with ll. 1-5, gives better sense:

If the ninth line of the poem be restored as suggested, its account of the Birth of the Gods will be found to correspond accurately with the summary from Berossus, who, in explaining the myth, refers to the Babylonian belief that the universe consisted at first of moisture in which living creatures, such as he had already described, were generated.(1) The primaeval waters are originally the source of life, not of destruction, and it is in them that the gods are born, as in Egyptian mythology; there Nu, the primaeval water-god from whom Ra was self-created, never ceased to be the Sun-god's supporter. The change in the Babylonian conception was obviously introduced by the combination of the Dragon myth with that of Creation, a combination that in Egypt would never have been justified by the gentle Nile. From a study of some aspects of the names at the beginning of the Babylonian poem we have already seen reason to suspect that its version of the Birth of the Gods goes back to Sumerian times, and it is pertinent to ask whether we have any further evidence that in Sumerian belief water was the origin of all things.

For many years we have possessed a Sumerian myth of Creation, which has come to us on a late Babylonian tablet as the introductory section of an incantation. It is provided with a Semitic translation, and to judge from its record of the building of Babylon and Egasila, Marduk's temple, and its identification of Marduk himself with the Creator, it has clearly undergone some editing at the hands of the Babylonian priests. Moreover, the occurrence of various episodes out of their logical order, and the fact that the text records twice over the creation of swamps and marshes, reeds and trees or forests, animals and cities, indicate that two Sumerian myths have been combined. Thus we have no guarantee that the other cities referred to by name in the text, Nippur, Erech, and Eridu, are mentioned in any significant connexion with each other.(1) Of the actual cause of Creation the text appears to give two versions also, one in its present form impersonal, and the other carried out by a god. But these two accounts are quite unlike the authorized version of Babylon, and we may confidently regard them as representing genuine Sumerian myths. The text resembles other early accounts of Creation by introducing its narrative with a series of negative statements, which serve to indicate the preceding non-existence of the world, as will be seen from the following extract:(2)

Here we have the definite statement that before Creation all the world was sea. And it is important to note that the primaeval water is not personified; the ordinary Sumerian word for "sea" is employed, which the Semitic translator has faithfully rendered in his version of the text.(4) The reference to a channel in the sea, as the cause of Creation, seems at first sight a little obscure; but the word implies a "drain" or "water-channel", not a current of the sea itself, and the reference may be explained as suggested by the drainage of a flood-area. No doubt the phrase was elaborated in the original myth, and it is possible that what appears to be a second version of Creation later on in the text is really part of the more detailed narrative of the first myth. There the Creator himself is named. He is the Sumerian god Gilimma, and in the Semitic translation Marduk's name is substituted. To the following couplet, which describes Gilimma's method of creation, is appended a further extract from a later portion of the text, there evidently displaced, giving additional details of the Creator's work:

Here the Sumerian Creator is pictured as forming dry land from the primaeval water in much the same way as the early cultivator in the Euphrates Valley procured the rich fields for his crops. The existence of the earth is here not really presupposed. All the world was sea until the god created land out of the waters by the only practical method that was possible in Mesopotamia.

In another Sumerian myth, which has been recovered on one of the early tablets from Nippur, we have a rather different picture of beginnings. For there, though water is the source of life, the existence of the land is presupposed. But it is bare and desolate, as in the Mesopotamian season of "low water". The underlying idea is suggestive of a period when some progress in systematic irrigation had already been made, and the filling of the dry canals and subsequent irrigation of the parched ground by the rising flood of Enki was not dreaded but eagerly desired. The myth is only one of several that have been combined to form the introductory sections of an incantation; but in all of them Enki, the god of the deep water, plays the leading part, though associated with different consorts.(1) The incantation is directed against various diseases, and the recitation of the closing mythical section was evidently intended to enlist the aid of special gods in combating them. The creation of these deities is recited under set formulae in a sort of refrain, and the divine name assigned to each bears a magical connexion with the sickness he or she is intended to dispel.(2)

We have already noted examples of a similar use of myth in magic, which was common to both Egypt and Babylonia; and to illustrate its employment against disease, as in the Nippur document, it will suffice to cite a well-known magical cure for the toothache which was adopted in Babylon.(1) There toothache was believed to be caused by the gnawing of a worm in the gum, and a myth was used in the incantation to relieve it. The worm's origin is traced from Anu, the god of heaven, through a descending scale of creation; Anu, the heavens, the earth, rivers, canals and marshes are represented as each giving rise to the next in order, until finally the marshes produce the worm. The myth then relates how the worm, on being offered tempting food by Ea in answer to her prayer, asked to be allowed to drink the blood of the teeth, and the incantation closes by invoking the curse of Ea because of the worm's misguided choice. It is clear that power over the worm was obtained by a recital of her creation and of her subsequent ingratitude, which led to her present occupation and the curse under which she laboured. When the myth and invocation had been recited three times over the proper mixture of beer, a plant, and oil, and the mixture had been applied to the offending tooth, the worm would fall under the spell of the curse and the patient would at once gain relief. The example is instructive, as the connexion of ideas is quite clear. In the Nippur document the recital of the creation of the eight deities evidently ensured their presence, and a demonstration of the mystic bond between their names and the corresponding diseases rendered the working of their powers effective. Our knowledge of a good many other myths is due solely to their magical employment.

Perhaps the most interesting section of the new text is one in which divine instructions are given in the use of plants, the fruit or roots of which may be eaten. Here Usmû, a messenger from Enki, God of the Deep, names eight such plants by Enki's orders, thereby determining the character of each. As Professor Jastrow has pointed out, the passage forcibly recalls the story from Berossus, concerning the mythical creature Oannes, who came up from the Erythraean Sea, where it borders upon Babylonia, to instruct mankind in all things, including "seeds and the gathering of fruits".(1) But the only part of the text that concerns us here is the introductory section, where the life-giving flood, by which the dry fields are irrigated, is pictured as following the union of the water-deities, Enki and Ninella.(2) Professor Jastrow is right in emphasizing the complete absence of any conflict in this Sumerian myth of beginnings; but, as with the other Sumerian Versions we have examined, it seems to me there is no need to seek its origin elsewhere than in the Euphrates Valley.

Even in later periods, when the Sumerian myths of Creation had been superseded by that of Babylon, the Euphrates never ceased to be regarded as the source of life and the creator of all things. And this is well brought out in the following introductory lines of a Semitic incantation, of which we possess two Neo-Babylonian copies:(1)

Here the river as creator is sharply distinguished from the Flood; and we may conclude that the water of the Euphrates Valley impressed the early Sumerians, as later the Semites, with its creative as well as with its destructive power. The reappearance of the fertile soil, after the receding inundation, doubtless suggested the idea of creation out of water, and the stream's slow but automatic fall would furnish a model for the age-long evolution of primaeval deities. When a god's active and artificial creation of the earth must be portrayed, it would have been natural for the primitive Sumerian to picture the Creator working as he himself would work when he reclaimed a field from flood. We are thus shown the old Sumerian god Gilimma piling reed-bundles in the water and heaping up soil beside them, till the ground within his dikes dries off and produces luxuriant vegetation. But here there is a hint of struggle in the process, and we perceive in it the myth-redactor's opportunity to weave in the Dragon motif. No such excuse is afforded by the other Sumerian myth, which pictures the life-producing inundation as the gift of the two deities of the Deep and the product of their union.

But in their other aspect the rivers of Mesopotamia could be terrible; and the Dragon motif itself, on the Tigris and Euphrates, drew its imagery as much from flood as from storm. When therefore a single deity must be made to appear, not only as Creator, but also as the champion of his divine allies and the conqueror of other gods, it was inevitable that the myths attaching to the waters under their two aspects should be combined. This may already have taken place at Nippur, when Enlil became the head of the pantheon; but the existence of his myth is conjectural.(1) In a later age we can trace the process in the light of history and of existing texts. There Marduk, identified wholly as the Sun-god, conquers the once featureless primaeval water, which in the process of redaction has now become the Dragon of flood and storm.

Thus the dualism which is so characteristic a feature of the Semitic-Babylonian system, though absent from the earliest Sumerian ideas of Creation, was inherent in the nature of the local rivers, whose varied aspects gave rise to or coloured separate myths. Its presence in the later mythology may be traced as a reflection of political development, at first probably among the warring cities of Sumer, but certainly later in the Semitic triumph at Babylon. It was but to be expected that the conqueror, whether Sumerian or Semite, should represent his own god's victory as the establishment of order out of chaos. But this would be particularly in harmony with the character of the Semitic Babylonians of the First Dynasty, whose genius for method and organization produced alike Hammurabi's Code of Laws and the straight streets of the capital.

We have thus been able to trace the various strands of the Semitic-Babylonian poem of Creation to Sumerian origins; and in the second lecture we arrived at a very similar conclusion with regard to the Semitic-Babylonian Version of the Deluge preserved in the Epic of Gilgamesh. We there saw that the literary structure of the Sumerian Version, in which Creation and Deluge are combined, must have survived under some form into the Neo-Babylonian period, since it was reproduced by Berossus. And we noted the fact that the same arrangement in Genesis did not therefore prove that the Hebrew accounts go back directly to early Sumerian originals. In fact, the structural resemblance presented by Genesis can only be regarded as an additional proof that the Sumerian originals continued to be studied and translated by the Semitic priesthood, although they had long been superseded officially by their later descendants, the Semitic epics. A detailed comparison of the Creation and Deluge narratives in the various versions at once discloses the fact that the connexion between those of the Semitic Babylonians and the Hebrews is far closer and more striking than that which can be traced when the latter are placed beside the Sumerian originals. We may therefore regard it as certain that the Hebrews derived their knowledge of Sumerian tradition, not directly from the Sumerians themselves, but through Semitic channels from Babylon.

It will be unnecessary here to go in detail through the points of resemblance that are admitted to exist between the Hebrew account of Creation in the first chapter of Genesis and that preserved in the "Seven Tablets".(1) It will suffice to emphasize two of them, which gain in significance through our newly acquired knowledge of early Sumerian beliefs. It must be admitted that, on first reading the poem, one is struck more by the differences than by the parallels; but that is due to the polytheistic basis of the poem, which attracts attention when compared with the severe and dignified monotheism of the Hebrew writer. And if allowance be made for the change in theological standpoint, the material points of resemblance are seen to be very marked. The outline or general course of events is the same. In both we have an abyss of waters at the beginning denoted by almost the same Semitic word, the Hebrew tehôm, translated "the deep" in Gen. i. 2, being the equivalent of the Semitic-Babylonian Tiamat, the monster of storm and flood who presents so striking a contrast to the Sumerian primaeval water.(2) The second act of Creation in the Hebrew narrative is that of a "firmament", which divided the waters under it from those above.(3) But this, as we have seen, has no parallel in the early Sumerian conception until it was combined with the Dragon combat in the form in which we find it in the Babylonian poem. There the body of Tiamat is divided by Marduk, and from one half of her he constructs a covering or dome for heaven, that is to say a "firmament", to keep her upper waters in place. These will suffice as text passages, since they serve to point out quite clearly the Semitic source to which all the other detailed points of Hebrew resemblance may be traced.

In the case of the Deluge traditions, so conclusive a demonstration is not possible, since we have no similar criterion to apply. And on one point, as we saw, the Hebrew Versions preserve an original Sumerian strand of the narrative that was not woven into the Gilgamesh Epic, where there is no parallel to the piety of Noah. But from the detailed description that was given in the second lecture, it will have been noted that the Sumerian account is on the whole far simpler and more primitive than the other versions. It is only in the Babylonian Epic, for example, that the later Hebrew writer finds material from which to construct the ark, while the sweet savour of Ut-napishtim's sacrifice, and possibly his sending forth of the birds, though reproduced in the earlier Hebrew Version, find no parallels in the Sumerian account.(1) As to the general character of the Flood, there is no direct reference to rain in the Sumerian Version, though its presence is probably implied in the storm. The heavy rain of the Babylonian Epic has been increased to forty days of rain in the earlier Hebrew Version, which would be suitable to a country where local rain was the sole cause of flood. But the later Hebrew writer's addition of "the fountains of the deep" to "the windows of heaven" certainly suggests a more intimate knowledge of Mesopotamia, where some contributary cause other than local rain must be sought for the sudden and overwhelming catastrophes of which the rivers are capable.