Vetus

Testamentum
Vetus Testamentum IOSOT (2013) 32-35 brill.com/vt

The “Shield” of Abraham?

Martin Kessler
Brandeis University, Waltham

Genesis xv, which contains an account of Yahweh’s covenant with Abraham,

opens with the Heilsorakel: ʾal tira⁠ʾ ʾabram ʾanoki magen lak śekarka harbeh
meʾod, which the Revised Standard Version translates: “Fear not, Abram, I am
your shield; your reward shall be very great”.
The term magen (ʾabram) with which we are here primarily concerned, may
be interpreted in a number of ways. Alt suggested in 1929 in his article “Der
Gott der Väter”1 that it might be one of the designations of a patriarchal deity,
along with ʾabir yaʿaqob2 and paḥad yiṣḥaq.3 Actually, he was quite hesitant
about the possibility of postulating magen ʾabram as the name of Abraham’s
personal deity.4
Several objections may be raised against the plausibility of this interpreta-
tion. Whereas Jacob swears by the God of his father, paḥad yiṣḥaq (while the
designation ʾabir yaʿaqob occurs in connection with the blessing on Joseph),
Abraham is addressed directly in Gn. xv 1. Further, purely formal consider-
ations tend to argue against this being the divine disclosure of a new name,
as in Gn xvii 1 and xxxv 11: ʾani ʾel šadday. Use of the phrase ʾelohey ʾabraham
ʾabika weʾlohey yiṣḥaq (Gn. xxviii 13) does not prove one way or the other as to
whether there was a special designation for Abraham’s personal deity.
If magen is not likely to be a designation for Abraham’s God, let us consider
the traditional interpretation, viz. that God declares that He is “a shield” to

1) Re-published in Kleine Schriften zur Geschichte des Volkes Israëls I, München 1953, pp. 1-78.
2) Mentioned in the blessing of Jacob, Gn. xlix 24; also in Ps. cxxxii 2, 5; Is. lx 16, while ʾabir yiśra⁠ʾel
occurs in Is. i 24.
3) By whom Jacob is reported to have sworn when he made a covenant with Laban, Gn. xxxi 42, 53.
4) He suggests that in Gn. xv 1 we may have a Kultstiftungssage for the God of Abraham, and
that the divine address contains a reference to the lost personal name of Abraham’s God: magen
ʾabraham. (Kleine Schriften, op. cit., p. 67, note 4).
© Koninklijke Brill NV, Leiden, 1964 DOI: 10.1163/15685330-99000007
 M. Kessler / Vetus Testamentum IOSOT (2013) 32-35 33

Abraham.5 Von Rad remarks that this expression “derives from the language
of the cult”.6 Indeed, many psalms refer to Israel’s God as a “shield”, that is, a
weapon of warfare furnishing protection.7 This is how the MT seems to have
understood magen. The context then would be that Abraham was admon-
ished not to fear (a typical introductory phrase for a Heilsorakel ), for God is his
“shield”, who had kept him from danger as in the raid described in Gn. xiv.
A third interpretation, which holds that mgn in xv 1 is a verbal form (as in
xiv 20) to be vocalized mogen has not received the attention it deserves.8 This
view derives support from a consideration of the keywords which occur in
xiv 20-xv 3.9 Firstly, we note that the verbs natan and laqaḥ occur several times
(four and three times, respectively).10
What is still more important is that the root mgn, used in Melchizedek’s
blessing: baruk ʾel ʾelyon ʾašer-miggen ṣareyka beyadka, xiv 20, appears again
in xv 1: also in poetry. In xiv 20 it clearly means “to deliver up, to give”.11 Might
it not also mean this in xv 1? Then the oracle in xv 1 may be seen as a spiri-
tual “heightening” of xiv 20: it is important that God should “give” Abraham’s
enemies into his hand, but far superior is the “very great reward” which God
is about to give Abraham by virtue of the covenant. This would also bring the

5) This interpretation is accepted by Skinner, A Critical and Exegetical Commentary on Genesis,

Edinburgh 19562, p. 278; von Rad, Genesis, Philadelphia 1961, p. 178; Gunkel, Genesis, Göttingen
19646, p. 179, believes that this oracle presupposes a specific condition (Sitz im Leben) in which
Abraham needed a shield for protection.
 6) Op. cit., p. 178.
  7) See e.g. Ps. iii 4; xviii 3; xxviii 7, et al. Kaiser who affirms that most of the Psalms in which
magen occurs deal with Yahweh’s help to the king, raises the question of whether the designation
of Yahweh as a shield could be found already in pre-royal times, in the tradition of the holy war.
(“Traditionsgeschichtliche Untersuchung von Genesis 15”, ZAW 70, 1958, p. 113).
 8) It was suggested by Winckler, Forschungen III, p. 411, quoted by Gunkel, op. cit., p. 179; also
by Jeremias, Das AT im Lichte des Alten Orients Leipzig 19062, p. 351, and Ehrlich, Randglossen
zur Hebräischen Bibel, Leipzig 1908, pp. 57 f.
 9) The importance of keywords (which he terms Leitworte), as an aid to exegesis, has been
emphasized especially by Buber, in Buber & Rosenzweig, Die Schrift und Ihre Verdeutschung,
Berlin 1936, pp. 211 ff.
10) Abraham gave Melchizedek a tithe (xiv 20). The king of Sodom said: Give me the people, but
take the goods for yourself (21). Abraham had sworn not to take anything from the king of Sodom
(24). Let Aner, Eshkol and Mamre take their share (24). Abraham said to Yahweh: What wilt thou
give me? (xv 2). Thou hast not given me offspring (xv 3).—That these keywords are present in
both chapters is perhaps the more remarkable in view of the evident “editorial joining” in xv 1 in
the phrase ʾaḥar haddebarim ha⁠ʾelleh.
11)  Mgn is a poetic motif-word for “to give” ( Jeremias, op. cit., p. 351). In Hos. xi 8 the verbs ntn
and mgn occur in parallel stichs.
34 M. Kessler / Vetus Testamentum IOSOT (2013) 32-35

two parts of the oracle in closer harmony, as follows: “Fear not, Abram, I am
about to give you your very great reward”. In turn, it would lend better sense to
Abraham’s retort: “O Lord God; what wilt thou give me . . .” (v. 2) and be wholly
in line with the prevalence of the idea of “giving” in this section.
The vocalization of the MT (magen) seems to stem from a time when the
verbal force of mgn (Qal active participle) was no longer understood. On
the other hand, Yahweh’s designation as magen (shield) is well attested in
the Psalms.
Another objection, pointed out by Gunkel,12 is that where mgn occurs it is
always in the Piel elsewhere. But verbs which occur predominantly in the Piel,
sometimes appear in Qal particularly as active participles, as e.g. in the case
of dbr.13
Ehrlich14 has some illuminating comments to offer in his discussion of this
phrase. He begins by saying that ʾanoki is used here because it is more emphatic
than ʾani. ʾAnoki here, as elsewhere, emphasizes a verbal form containing a
personal pronoun; for the emphasis of a verbal suffix expressing an object,
however, ʾani must be used. Ehrlich further makes the important point that
in light of the context of Gn. xiv Abraham did not need any assurance of the
protection of Yahweh. On the other hand, Abraham, after his first generous
impulse of refusing any of the booty, might have had some second thoughts
on the matter. Thus, Ehrlich proposes to read mogen (poetic equivalent of
natan) and understands śekarka as an accusative of material, while lak repre-
sents an accusative of person.15
Considering the context, Ehrlich’s comment is well taken when he points
out that God’s offer of military protection is irrelevant after Abraham’s success-
ful campaign narrated in the previous chapter; on the other hand, Abraham
returns (voluntarily) with empty hands from the battle!
The use of ʾanoki (emphatic: containing three syllables) rather than ʾani is
significant here. The fact is, that it is too emphatic to stand before magen, the
predicate, which naturally attracts emphasis because it is the new element in
the clause. We may compare the divine self-revelations: ʾani ʾel šadday, etc.,
where likewise the predicate is the new factor, needing emphasis, so that the

12) Op. cit., p. 179. Gunkel would read memaggen (Piel participle) instead of Winckler’s miggen.
13) The places where dbr occurs in Qal are almost exclusively active participles (31 times). Like-
wise, kzb occurs most frequently in Piel, but once in Qal, again an active participle, as pointed out
by Ehrlich, op. cit., p. 58.
14) Op. cit., pp. 57 f.
15) The same construction is found in Gn. xvii 19: śarah išteka yoledet leka ben (emphatic subject,
followed by an active participle, and two objects, one personal and one material).
 M. Kessler / Vetus Testamentum IOSOT (2013) 32-35 35

pronoun, which receives ample attention by its position at the beginning of the
phrase, must be (less emphatic) ʾani. But in the MT both ʾanoki and magen are
emphatic: a highly unlikely situation.
On the other hand, the proposed reading ʾanoki mogen has scores of syn-
tactic parallels. To mention but a few examples: ʾanoki holek (Gn. xv 2); ʾanoki
boraḥat (xvi 2); ʾanoki mebiʾ . . . ʾanoki dober ( Jr. xxxii 42), etc. etc. Because we
have here a participial form without a suffix, the subject needs to be expressed
separately, and this is virtually always done by the emphatic pronoun ʾanoki
rather than ʾani.16
If mogen was indeed the original reading, the change to magen must have
taken place at some time prior to the Massoretes, undoubtedly under the
strong influence of magen in cultic language. The LXX rendering of mgn as
a verb (ʿuperaspἰzo=to cover with a shield) may reflect an intermediate stage
between mogen and magen.

16) Stated formally, the proposed reading would change the “incomplete synonymous parallelism
with compensation” to “synthetic or formal parallelism”. See N. K. Gottwald, “Poetry, Hebrew”,
Interpreter’s Dictionary of the Bible, K-Q, New York - Nashville 1962, pp. 829-838.