ii. 33, it is doubtful if the grammatical argument can be

sustained; but if it had any force it ought certainly to lead to the excision of the second member rather than of the first (Kuenen _Theologisch Tijdschrift_, 1884, 136; van Doorninck, _ib._, 1905, 225 f.; Eerdmans, _ib._ 494 ff.). A more important point is the absence of את‎ before the definite object. The writer’s use of this participle is very discriminating; and its omission suggests that ⁹ᵇ is really a nominal clause, as rendered above. If we were to indulge in analysis of sources, we might put ⁹ᵇ (in whole or in part) after ⁸ᵃ, and assign it to that secondary stratum of narrative which undoubtedly spoke of a tree of life (3²²).

* * * * *

=9.= _all sorts of trees ... food_] The primitive vegetation is conceived as consisting solely of trees, on whose fruit man was to subsist; the appearance of herbs is a result of the curse pronounced on the ground (3¹⁷ ᶠᐧ).――_and the tree of life_ (was) _in the midst_] On Budde’s strictures on the form of the sentence, _v.i._ The intricate question of the two trees must be reserved for separate discussion (pages 52 f., 94); for the present form of the story both are indispensable. The tree of life, whose fruit confers immortality (3²²; compare Proverbs 3¹⁸ 11³⁰ 13¹² 15⁴; further, Ezekiel 47¹², Revelation 22²), is a widely diffused idea (see Dillmann 49; Wünsche, _Die Sagen vom Lebensbaum und Lebenswasser_). The tree of knowledge is a more refined conception; its property of communicating knowledge of good and evil is, however, magical, like that of the other; a connexion with oracular trees (Lenormant, _Les Origines de l’histoire_ i. 85 f.; Baudissin, _Studien zur semitischen Religionsgeschichte_ ii. 227) is not so probable. As to what is meant by ‘knowing good and evil,’ see page 95 ff.

The primitive Semitic tree of life is plausibly supposed by Barton (_A Sketch of Semitic Origins_¹, 92 f.) to have been the date-palm; and this corresponds to the sacred palm in the sanctuary of Ea at Eridu (IV R. 15*), and also to the conventionalised sacred tree of the seals and palace-reliefs, which is considered to be a palm combined with some species of conifer. Compare also the sacred cedar in the cedar forest of _Gilgamesh_, _Tablets_ IV. V. For these and other Babylonian parallels, see _Das Alte Testament im Lichte des alten Orients_², 195 ff.

=10.= _a river issued_ (or _issues_) _from Eden_] The language does not necessarily imply that the fountain-head was outside the garden (Driver, Bennett); the verb יָצָא‎ is used of the rise of a stream at its source (Exodus 17⁶, Numbers 20¹¹, Judges 15¹⁹, Ezekiel 47¹, Zechariah 14⁸, Joel 4¹⁸). Whether the participle expresses past or present time cannot be determined.――_from thence it divides itself_] The river issues from the garden as a single stream, then divides into four branches, which are the four great rivers of the world. The site of Paradise, therefore, is at the common source of the four rivers in question (page 62‒66 below). That is the plain meaning of the verse, however inconsistent it may be with physical geography.――=11.= _Pîšôn_] The name occurs (along with Tigris, Euphrates, Jordan, and Gihon) in Sirach 24²⁵, but nowhere else in Old Testament. That it was not a familiar name to the Hebrews is shown by the topographical description which follows. On the various speculative identifications, see Delitzsch and Dillmann, and page 64 f. below.――_the whole land of Ḥăvîlāh_] The phraseology indicates that the name is used with some vagueness, and considerable latitude. In 10⁷ᐧ ²⁹ 25¹⁸ etc., Ḥavilah seems to be a district of Arabia (see page 202); but we cannot be sure that it bears the same meaning in the mythically coloured geography of this passage.――=12.= Two other products of the region are specified; but neither helps to an identification of the locality.――_bĕdōlaḥ_] a substance well known to the Israelites (Numbers 11⁷), is undoubtedly the fragrant but bitter gum called by the Greeks βδέλλιον or βδέλλα. Pliny (_Naturalis Historia_, xii. 35 f.) says the best kind grew in Bactriana, but adds that it was found also in Arabia, India, Media, and Babylonia.――_the šōham stone_] A highly esteemed gem (Job 28¹⁶), suitable for engraving (Exodus 28⁹ etc.), one of the precious stones of Eden (Ezekiel 28¹³), and apparently used in architecture (1 Chronicles 29²). From the Greek equivalents it is generally supposed to be either the onyx or the beryl (_v.i._). According to Pliny, the latter was obtained from India, the former from India and Arabia (_Naturalis Historia_, xxxvii. 76, 86).――=13.= _Gîḥôn_] The name of a well on the East of Jerusalem (the Virgin’s spring: 1 Kings 1³³ etc.), which Abraham Ibn Ezra strangely takes to be meant here. In Jewish and Christian tradition it was persistently identified with the Nile (Sirach 24²⁷; LXX of Jeremiah 2¹⁸ [where שִׁחוֹר‎ is translated Γηών]; Josephus _Antiquities of the Jews_ i. 39, and the Fathers generally). The great difficulty of that view is that the Nile was as well known to the Hebrews as the Euphrates, and no reason appears either for the mysterious designation, or the vague description appended to the name.――_land of Kûš_] Usually Ethiopia; but see on 10⁶.――=14.= _Ḥiddeḳel_] is certainly the Tigris, though the name occurs only once again (Daniel 10⁴).――_in front of ’Aššûr_] Either between it and the spectator, or to the east of it: the latter view is adopted by nearly all commentaries; but the parallels are indecisive, and the point is not absolutely settled. Geographically the former would be more correct, since the centre of the Assyrian Empire lay East of the Tigris. The second view can be maintained only if אַשּׁוּר‎ be the city which was the ancient capital of the Empire, now _Ḳal‛at Šerḳāt_ on the West bank of the river. But that city was replaced as capital by Kalḫi as early as 1300 B.C., and is never mentioned in Old Testament. It is at least premature to find in this circumstance a conclusive proof that the Paradise legend had wandered to Palestine before 1300 B.C. (Gressmann, _Archiv für Religionswissenschaft._, x. 347).――_Euphrates_] The name (פְּרָת‎) needed no explanation to a Hebrew reader: it is the נָהָר‎ _par excellence_ of the Old Testament (Isaiah 8⁷ and often).

_The site of Eden._――If the explanation given above of verse ¹⁰ be correct,――and it is the only sense which the words will naturally bear,――it is obvious that a real locality answering to the description of Eden exists and has existed nowhere on the face of the earth. The Euphrates and Tigris are not and never were branches of a single stream; and the idea that two other great rivers sprang from the same source places the whole representation outside the sphere of real geographical knowledge. In ¹⁰⁻¹⁴, in short, we have to do with a semi-mythical geography, which the Hebrews no doubt believed to correspond with fact, but which is based neither on accurate knowledge of the region in question, nor on authentic tradition handed down from the ancestors of the human race. Nevertheless, the question where the Hebrew imagination located Paradise is one of great interest; and many of the proposed solutions are of value, not only for the light they have thrown on the details of ¹⁰⁻¹⁴, but also for the questions they raise as to the origin and character of the Paradise-myth. This is true both of those which deny, and of those which admit, the presence of a mythical element in the geography of ¹⁰⁻¹⁴.

1. Several recent theories seek an exact determination of the locality of Paradise, and of all the data of ¹⁰⁻¹⁴, at the cost of a somewhat unnatural exegesis of verse ¹⁰. That of Friedrich Delitzsch (_Wo lag das Paradies?_, 1881) is based partly on the fact that North of Babylon (in the vicinity of Bagdad) the Euphrates and Tigris approach within some twenty miles of each other, the Euphrates from its higher level discharging water through canals into the Tigris, which might thus be regarded as an offshoot of it. The _land_ of Eden is the plain (_edinu_) between the two rivers from Tekrit (on the Tigris: nearly a hundred miles North of Bagdad) and ‛Ana (on the Euphrates) to the Persian Gulf; the _garden_ being one specially favoured region from the so-called ‘isthmus’ to a little South of Babylon. The _river_ of verse ¹⁰ is the Euphrates; Pishon is the Pallakopas canal, branching off from the Euphrates on the right a little above Babylon and running nearly parallel with it to the Persian Gulf; Giḥon is the _Shaṭṭ en-Nil_, another canal running East of the Euphrates from near Babylon and rejoining the parent river opposite Ur; Ḥiddeḳel and Euphrates are, of course, the lower courses of the Tigris and Euphrates respectively, the former regarded as replenished through the canal system from the latter. Ḥavilah is part of the great Syrian desert lying West and South of the Euphrates; and Kush is a name for northern and middle Babylonia, derived from the Kaššite dynasty that once ruled there. In spite of the learning and ingenuity with which this theory has been worked out, it cannot clear itself of an air of artificiality at variance with the simplicity of the passage it seeks to explain. That the Euphrates should be at once the undivided Paradise-stream and one of the ‘heads’ into which it breaks up is a glaring anomaly; while verse ¹⁴ shows that the narrator had distinctly before his mind the upper course of the Tigris opposite Assur, and is therefore not likely to have spoken of it as an effluent of the Euphrates. The objection that the theory confuses rivers and canals is fairly met by the argument that the Babylonian equivalent of נָהָר‎ is used of canals, and also by the consideration that both the canals mentioned were probably ancient river-beds; but the _order_ in which the rivers are named tells heavily against the identifications. Moreover, the expression ‘the _whole_ land of Ḥavilah’ seems to imply a much larger tract of the earth’s surface than the small section of desert enclosed by the Pallakopas; and to speak of the _whole_ of northern Babylonia as ‘surrounded’ by the _Shaṭṭ en-Nil_ is an abuse of language.――According to Sayce _The Higher Criticism and the Verdict of the Monuments_, 95 ff.; _A Dictionary of the Bible_, i. 643 f.), the garden of Eden is the sacred garden of Ea at Eridu; and the river which waters it is the Persian Gulf, on the shore of which Eridu formerly stood. The four branches are, in addition to Euphrates and Tigris (which in ancient times entered the Gulf separately), the Pallakopas and the Choaspes (now the _Kerkha_), the sacred river of the Persians, from whose waters alone their kings were allowed to drink (Herodotus i. 188). Besides the difficulty of supposing that the writer of verse ¹⁰ meant to trace the streams _upwards_ towards their source above the garden, the theory does not account for the order in which the rivers are given; for the Pallakopas is West of Euphrates, while the Choaspes is East of the Tigris.¹ Further, although the description of the Persian Gulf as a ‘river’ is fully justified by its Babylonian designation as _Nâr Marratum_ (‘Bitter River’), it has yet to be made probable that either Babylonians or Israelites would have thought of a garden as watered by ‘bitter’ (_i.e._ salt) water.――These objections apply with equal force to the theory of Hommel (_Aufsätze und Abhandlungen arabistisch-semitologischen Inhalts_, iii. 1, page 281 ff., etc., _The Ancient Hebrew Tradition as illustrated by the Monuments_, 314 ff.), who agrees with Sayce in placing Paradise at Eridu, in making the single stream the Persian Gulf, and one of the four branches the Euphrates. But the three other branches, Pishon, Giḥon, and Ḥiddeḳel, he identifies with three North Arabian wādīs,――Wadi Dawāsir, Wadi Rummā, and Wadi Sirhān (the last the ‘wādī of Diḳlah’ = _ḫad-deḳel_ [see on verse ¹⁴ above], the name having been afterwards _transferred_ to the Tigris).

¹ This objection is avoided by the modified theory of Dawson, who identifies Pishon with the Karun, still further East than the Kerkha. But that removes it from all connexion with Ḥavilah, which is one of the recommendations of Sayce’s view.

2. Since none of the above theories furnishes a satisfactory solution of the problem, we may as well go back to what appears the natural interpretation of verse ¹⁰, and take along with it the utopian conception of four great rivers issuing from a single source. The site of Paradise is then determined by the imaginary common source of the two known rivers, Euphrates and Tigris. As a matter of fact, the western arm of the Euphrates and the eastern arm of the Tigris do rise sufficiently near each other to make the supposition of a common source possible to ancient cosmography; and there is no difficulty in believing that the passage locates the garden in the unexplored mountains of Armenia. The difficulty is to find the Pishon and the Giḥon. To seek them amongst the smaller rivers of Armenia and Trans-Caucasia is a hopeless quest; for a knowledge of these rivers would imply a knowledge of the country, which must have dispelled the notion of a common source. Van Doorninck has suggested the Leontes and Orontes (_Theologisch Tijdschrift_,