---

§ 7. The Sources of Genesis.

The Book of Genesis has always been the strategic position of Pentateuchal literary criticism. It was the examination of this book that led Astruc, in 1753,¹ to the important discovery which was the first positive achievement in this department of research. Having noticed the significant alternation of the divine names in different sections of the book, and having convinced himself that the phenomenon could not be explained otherwise than as due to the literary habit of two writers, Astruc proceeded to divide the bulk of Genesis into two documents, one distinguished by the use of the name אֱלֹהִים, and the other by the use of יְהֹוָה; while a series of fragmentary passages where this criterion failed him brought the total number of his mémoires up to twelve. Subsequent investigations served to emphasise the magnitude of this discovery, which Eichhorn² speedily put on a broader basis by a characterisation of the style, contents, and spirit of the two documents. Neither Astruc nor Eichhorn carried the analysis further than Exodus 2,—partly because they were influenced by the traditional opinion (afterwards abandoned by Eichhorn) of Mosaic authorship, and did not expect to find traces of composition in the history contemporaneous with Moses. We shall see presently that there is a deeper reason why this particular clue to the analysis could not at first be traced beyond the early chapters of Exodus.

While the earlier attempts to discredit Astruc’s discovery took the direction of showing that the use of the two divine names is determined by a difference of meaning which made the one or the other more suitable in a particular connexion, the more recent opposition entrenches itself mostly behind the uncertainties of the text, and maintains that the Versions (especially LXX) show the Massoretic Text to be so unreliable that no analysis of documents can be based on its data: see Klostermann, Der Pentateuch (1893), page 20 ff.; Dahse, Archiv für Religionswissenschaft., vi. (1903), 305 ff.; Redpath, American Journal of Theology, viii. (1904), 286 ff.; Eerdmans, Die Komposition der Genesis (1908), 34 ff.; Wiener, Bibliotheca Sacra and Theological Review (1909), 119 ff.—It cannot be denied that the facts adduced by these writers import an element of uncertainty into the analysis, so far as it depends on the criterion of the divine names; but the significance of the facts is greatly overrated, and the alternative theories propounded to account for the textual phenomena are improbable in the extreme. (1) So far as I have observed, no attention is paid to what is surely a very important factor of the problem, the proportion of divergences to agreements as between LXX and Massoretic Text. In Genesis the divine name occurs in one or other form about 340 times (in Massoretic Text, יהוה 143 times + אלהים 177 times + י׳ א׳ 20 times). The total deviations registered by Redpath (296 ff.) number 50; according to Eerdmans (34 f.) they are 49; i.e. little more than one-seventh of the whole. Is it so certain that that degree of divergence invalidates a documentary analysis founded on so much larger a field of undisputed readings? (2) In spite of the confident assertions of Dahse (309) and Wiener (131 f.) there is not a single instance in which LXX is ‘demonstrably’ right against Massoretic Text. It is readily conceded that it is probably right in a few cases; but there are two general presumptions in favour of the superior fidelity of the Massoretic tradition. Not only (a) is the chance of purely clerical confusion between κ̅ς and θ̅ς greater than between יהוה and אלהים, or even between י׳ and א׳, and (b) a change of divine names more apt to occur in translation than in transcription, but (c) the distinction between a proper name יהוה and a generic אלהים is much less likely to have been overlooked in copying than that between two appellatives κύριος and θεός. An instructive example is 4²⁶, where LXX κύριος ὁ θεός is ‘demonstrably’ wrong. (3) In the present state of textual criticism it is impossible to determine in particular cases what is the original reading. We can only proceed by the imperfect method of averages. Now it is significant that while in Genesis LXX substitutes θεός for יהוה 21 times, and κύριος ὁ θεός 19 times (40 in all), there are only 4 cases of κύριος and 6 of κύριος ὁ θεός for אלהים (10 in all: the proportions being very much the same for the whole Pentateuch). LXX thus reveals a decided (and very natural) preference for the ordinary Greek θεός over the less familiar κύριος. Dahse urges (page 308) that Massoretic Text betrays an equally marked preference for יהוה, and has frequently substituted it for אלהים; but that is much less intelligible. For although the pronunciation of יהוה as אֲדֹנָי might have removed the fear of the Tetragrammaton,—and that would be a very good reason for leaving יהוה where it was,—it suggests no motive at all for inserting it where it was not. There is force, however, in Gray’s remark on a particular case (A Critical and Exegetical Commentary on Numbers page 311), that “wherever [ὁ] κ̅ς appears in LXX it deserves attention as a possible indication of the original text.” (4) The documentary theory furnishes a better explanation of the alternation of the names than any other that has been propounded. Redpath’s hypothesis of a double recension of the Pentateuch, one mainly Yahwistic and the other wholly (?) Elohistic, of which one was used only where the other was illegible, would explain anything, and therefore explains nothing; least of all does it explain the frequent coincidence of hypothetical illegibility with actual changes of style, phraseology, and standpoint. Dahse (following out a hint of Klostermann) accounts for the phenomena of Massoretic Text (and The Samaritan Recension of the Pentateuch) by the desire to preserve uniformity within the limits of each several pericope of the Synagogue lectionary; but why some pericopes should be Yahwistic and others Elohistic, it is not easy to conceive. He admits that his view cannot be carried through in detail; yet it is just of the kind which, if true, ought to be verifiable in detail. One has but to read consecutively the first three chapters of Genesis, and observe how the sudden change in the divine name coincides with a new vocabulary, representation, and spiritual atmosphere, in order to feel how paltry all such artificial explanations are in comparison with the hypothesis that the names are distinctive of different documents. The experience repeats itself, not perhaps quite so convincingly, again and again throughout the book; and though there are cases where the change of manner is not obvious, still the theory is vindicated in a sufficient number of instances to be worth carrying through, even at the expense of a somewhat complicated analysis, and a very few demands (see page xlviii f.) on the services of a redactor to resolve isolated problems. (5) It was frankly admitted by Kuenen long ago (see Historisch-critisch Onderzoek naar het ontstaan en de verzameling van de boeken des Ouden Verbonds i. pages 59, 62) that the test of the divine names is not by itself a sufficient criterion of source or authorship, and that critics might sometimes err through a too exclusive reliance on this one phenomenon.¹ Nevertheless the opinion can be maintained that the Massoretic Text is far superior to the Versions, and that its use of the names is a valuable clue to the separation of documents. Truth is sometimes stranger than fiction; and, however surprising it may appear to some, we can reconcile our minds to the belief that the Massoretic Text does reproduce with substantial accuracy the characteristics of the original autographs. At present that assumption can only be tested by the success or failure of the analysis based on it. It is idle to speculate on what would have happened if Astruc and his successors had been compelled to operate with LXX instead of Massoretic Text; but it is a rational surmise that in that case criticism would still have arrived, by a more laborious route, at very much the positions it occupies to-day.

The next great step towards the modern documentary theory of the Pentateuch was Hupfeld’s¹ demonstration that אלהים is not peculiar to one document, but to two; so that under the name Elohist two different writers had previously been confused. It is obvious, of course, that in this inquiry the divine names afford no guidance; yet by observing finer marks of style, and the connexion of the narrative, Hupfeld succeeded in proving to the ultimate satisfaction of all critics that there was a second Elohistic source (now called Elohist), closely parallel and akin to the Yahwistic (Yahwist), and that both Yahwist and Elohist had once been independent consecutive narratives. An important part of the work was a more accurate delimitation of the first Elohist (now called the Priestly Code: Priestly-Code), whose outlines were then first drawn with a clearness to which later investigation has had little to add.²

Though Hupfeld’s work was confined to Genesis, it had results of the utmost consequence for the criticism of the Pentateuch as a whole. In particular, it brought to light a fact which at once explains why Genesis presents a simpler problem to analysis than the rest of the Pentateuch, and furnishes a final proof that the avoidance of יהוה by two of the sources was not accidental, but arose from a theory of religious development held and expressed by both writers. For both Priestly-Code (Exodus 6^2 ff.) and Elohist (Exodus 3^13 ff.) connect the revelation of the Tetragrammaton with the mission of Moses; while the former states emphatically that God was not known by that name to the patriarchs.¹ Consistency demanded that these writers should use the generic name for Deity up to this point; while Yahwist, who was bound by no such theory, could use יהוה from the first.² From Exodus 6 onwards Priestly-Code regularly uses יהוה; Elohist’s usage fluctuates between א׳ and י׳ (perhaps a sign of different strata within the document), so that the criterion no longer yields a sure clue to the analysis.

It does not lie within the scope of this Introduction to trace the extension of these lines of cleavage through the other books of the Hexateuch; and of the reflex results of the criticism of the later books on that of Genesis only two can here be mentioned. One is the recognition of the unique position and character of Deuteronomy in the Pentateuch, and the dating of its promulgation in the eighteenth year of Josiah.¹ Although this has hardly any direct influence on the criticism of Genesis, it is an important landmark in the Pentateuch problem, as furnishing a fixed date by reference to which the age of the other documents can partly be determined. The other point is the question of the date of Priestly-Code. The preconception in favour of the antiquity of this document (based for the most part on the fact that it really forms the framework of the Pentateuch) was nearly universal among scholars down to the publication of Wellhausen’s Geschichte Israels, i., in 1878; but it had already been shown to be groundless by Graff² and Kuenen in 1866–69.

This revolutionary change was brought about by a comparison of the layers of legislation in the later Pentateuch books with one another, and with the stages of Israel’s religious history as revealed in the earlier historical books; from which it appeared that the laws belonging to Priestly-Code were later than Deuteronomy, and that their codification took place during and after, and their promulgation after, the Exile. There was hesitation at first in extending this conclusion to the narratives of Priestly-Code, especially those of them in Genesis and Exodus 1–11. But when the problem was fairly faced, it was perceived, not only that Priestly-Code in Genesis presented no obstacle to the theory, but that in many respects its narrative was more intelligible as the latest than as the oldest stratum of the book.

The chief positions at which literary criticism has arrived with regard to Genesis are, therefore, briefly these: (1) The oldest sources are Yahwist and Elohist, closely parallel documents, both dating from the best period of Hebrew literature, but distinguished from each other by their use of the divine name, by slight idiosyncrasies of style, and by quite perceptible differences of representation. (2) These sources were combined into a composite narrative (Jehovist) by a redactor (Redactor^Jehovist), whose hand can be detected in several patches of a literary complexion differing from either of his authorities. He has done his work so deftly that it is frequently difficult, and sometimes impossible, to sunder the documents. It is generally held that this redaction took place before the composition of Deuteronomy, so that a third stage in the history of the Pentateuch would be represented by the symbols Jehovist + Deuteronomy. (3) The remaining source Priestly-Code is a product of the Exilic or post-Exilic age, though it embodies older material. Originally an independent work, its formal and schematic character fitted it to be the framework of the Pentateuchal narrative; and this has determined the procedure of the final redactor (Redactor^Pentateuch), by whom excerpts from Jehovist have been used to fill up the skeleton outline which Priestly-Code gave of the primitive and patriarchal history.

The above statement will, it is hoped, suffice to put the reader in possession of the main points of the critical position occupied in the Commentary. The evidence by which they are supported will partly be given in the next four §§; but, for a full discussion of the numerous questions involved, we must here refer to works specially devoted to the subject.¹

Some idea of the extent to which conservative opinion has been modified by criticism, may be gathered from the concessions made by Professor Orr, whose book, The Problem of the Old Testament, deservedly ranks as the ablest assault on the critical theory of the Pentateuch that has recently appeared in English. Dr. Orr admits (a) that Astruc was right in dividing a considerable part of Genesis into Elohistic and Yahwistic sections; (b) that Eichhorn’s characterisation of the style of the two documents has, in the main, ‘stood the test of time’; (c) that Hupfeld’s observation of a difference in the Elohistic sections of Genesis ‘in substance corresponds with facts’; and (d) that even Graf and Wellhausen ‘mark an advance,’ in making Priestly-Code a relatively later stratum of Genesis than Jehovist (pages 196–201). When we see so many defences evacuated one after another, we begin to wonder what is left to fight about, and how a theory which was cradled in infidelity, and has the vice of its origin clinging to all its subsequent developments (Orr, 195 f.), is going to be prevented from doing its deadly work of spreading havoc over the ‘believing view’ of the Old Testament. Dr. Orr thinks to stem the torrent by adopting two relatively conservative positions from Klostermann. (1) The first is the denial of the distinction between Yahwist and Elohist (216 ff.). As soon as Hupfeld had effected the separation of Elohist from Priestly-Code, it ought to have been perceived, he seems to suggest, that the sections thus disentangled are really parts of Yahwist (217). And yet, even to Dr. Orr, the matter is not quite so simple as this, and he makes another concession. The distinction in the divine names remains; and so he is driven to admit that Yahwist and Elohist were, not indeed independent works, but different literary recensions of one and the same old work (229). What is meant by two versions in circulation alongside of each other, which never had currency as separate documents, is a point on which Dr. Orr owes his readers some explanation; if there were two recensions they certainly existed separately; and he cannot possibly know how far their agreement extended. The issue between him and his critical opponents is, nevertheless, perfectly clear: they hold that Yahwist and Elohist are independent recensions of a common body of tradition, while he maintains that they were recensions of a single document, differing in nothing but the use of יהוה or אלהים. What reasons, then, hinder us from deserting the critical view, and coming over to the side of Dr. Orr? In the first place, the difference between Yahwist and Elohist is not confined to the divine names. The linguistic evidence is very much clearer than Dr. Orr represents; and differences of conception, though slight, are real. It is all very well to quote from candid and truth-loving opponents admissions of the close resemblance of the narratives, and the difficulty and uncertainty of the analysis, in particular instances, and to suggest that these admissions amount to a throwing up of the case; but no man with an independent grasp of the subject will be imposed on by so cheap a device. In the second place, Yahwist and Elohist consist largely of duplicate narratives of the same event. It is true, this argument is lost on Dr. Orr, who has no difficulty in conceiving that Abraham twice told the same lie about his wife, and that his son Isaac followed his example, with very similar results in the three cases. But he will hardly affect to be surprised that other men take a more natural view,¹ and regard the stories as traditional variations of the same theme.—(2) The second position is that Priestly-Code was never a distinct or self-subsisting document, but only a “framework” enclosing the contents of Jehovist (341–377). Again we have to ask what Dr. Orr means by a ‘framework,’ which, in his own words, “has also, at certain points, its original, and, in parts, considerable contributions to bring to the history” (272); and how he can possibly tell that these original and considerable contributions did not come from an independent work. The facts that it is now closely interwoven with Jehovist, and that there are gaps in its narrative (even if these gaps were more considerable than there is any reason to suppose), prove nothing except that it has passed through the hands of a redactor. That its history presupposes a knowledge of Jehovist, and is too meagre to be intelligible apart from it, is amply explained by the critical view that the author wished to concentrate attention on the great religious turning-points in the history (the Creation, the Flood, the Covenant with Abraham, the Blessing of Jacob by Isaac, the origin of the name Israel, the Settlement in Egypt, etc.), and dismissed the rest with a bare chronological epitome. When we add that on all these points, as well as others, the ‘original and considerable contributions’ are (Dr. Orr’s protestations notwithstanding) radically divergent from the older tradition, we have every proof that could be desired that Priestly-Code was an independent document, and not a mere supplementary expansion of an earlier compilation (see, further, page lvii ff. below). But now, supposing Dr. Orr to have made good his contentions, what advantage has he gained? So far as we can see, none whatever! He does indeed go on to assert a preference for the term ‘collaboration’ as expressing the ‘kind and manner of the activity which brought the Pentateuchal books into their present shape’ (375).² But that preference might just as easily have been exercised on the full literary results of the critical theory. And Dr. Orr deceives himself if he imagines that that flimsy hypothesis will either neutralise the force of the arguments that have carried criticism past the barren eccentricities of Klostermann, or save what he chooses to consider the ‘essential Mosaicity’ of the Pentateuch.

Professor Eerdmans of Leiden, in a series of recent publications, has announced his secession from the Graf-Wellhausen school, and commenced to lay down the programme of a new era in Old Testament criticism (Hibbert Journal vii. [1909], 813 ff.). His Komposition der Genesis (1908) gives a foretaste of his literary method; and certainly the procedure is drastic enough. The divine names are absolutely misleading as a criterion of authorship; and the distinction between Priestly-Code and Jehovist goes overboard along with that between Yahwist and Elohist. Criticism is thus thrown back into its original chaos, out of which Eerdmans proceeds to evoke a new kosmos. His one positive principle is the recognition of a polytheistic background behind the traditions, which has been obscured in various degrees by the later monotheistic interpretation. By the help of this principle, he distinguishes four stages in the development of the tradition. (1) The first is represented by remnants of the original undiluted polytheism, where Yahwe does not appear at all; e.g. 35^1–7; the Israel-recension of the Joseph-stories; the groundwork of chapters 1. 20. 28^1–9 6⁹–9¹⁷. (2) Legends which recognise Yahwe as one among many gods; 4. 9^18–27 22. 27. 28^11–22 29. 30. 31. 39. (3) In the third stage, polytheistic legends are transferred to Yahwe as the only God: 2. 3. 6^1–8 7^1–5 8^20–22 11^1–9 16. 18. 19. 24. 25^19–34 26. (4) Late additions of purely monotheistic complexion: 15^1–6 17. 35^9–15 48^3–6. Now, we are quite prepared to find traces of all these stages of religion in the Genesis-narratives, if they can be proved; and, indeed, all of them except the second are recognised by recent critics. But while any serious attempt to determine the age of the legends from their contents rather than from their literary features is to be welcomed, it is difficult to perceive the distinctions on which Eerdmans’s classification is based, or to admit that, for example, chapter 17 is one whit more monotheistic than 20 or 27, or 24. In any case, on Eerdmans’s own showing, the classification affords no clue to the composition and history of the book. In order to get a start, he has to fall back on the acknowledged literary distinction between a Jacob-recension and an Israel-recension of the Joseph-narratives (on this see page 439 below). Since the former begins אלה תלדות יעקב, it is considered to have formed part of a comprehensive history of the patriarchs, commencing with Adam (5¹), set in a framework of Tôlĕdôth. This is the groundwork of Genesis. It is destitute of monotheistic colouring (it contains, however, legends of all the first three classes!), Yahwe being to the compiler simply one of the gods; and must therefore have originated before the Exile: a lower limit is 700 B.C. This collection was soon enlarged by the addition of legends not less ancient than its own; and by the insertion of the Israel-recension, which is as polytheistic in character as the Tôlĕdôth-collection! The monotheistic manipulation of the work set in after Deuteronomy; but how many editions it went through we cannot tell for certain. The last thorough-going reviser was the author of chapter 17; but additions were made even later than that, etc. etc. A more bewildering hypothesis it has never been our lot to examine; and we cannot pretend to believe that it contains the rudiments of a successful analysis. There is much to be learned from Eerdmans’s work, which is full of acute observations and sound reasoning in detail; but as a theory of the composition of Genesis it seems to us utterly at fault. What with Winckler and Jerome, and Cheyne, and now Eerdmans, Old Testament scholars have a good many new eras dawning on them just now. Whether any of them will shine unto the perfect day, time will show.