1. The Priority of the Promise (3:15–18)

15 My brothers, let me give you an illustration. Even in ordinary life, when a man’s will and testament has been duly executed, no one else can set it aside or add a codicil. 16 Now the promises were pronounced to Abraham and to his “issue.” It does not say “issues” in the plural, but in the singular, “and to your issue”; and the “issue” intended is Christ. 17 What I am saying is this: a testament, or covenant, had already been validated by God; it cannot be invalidated, and its promises rendered ineffective, by a law made four hundred and thirty years later. 18 If the inheritance is by legal right, then it is not by promise; but it was by promise that God bestowed it as a free gift on Abraham.

15 Paul begins his argument for the priority (and permanence) of the promise to Abraham with a human analogy (literally “I speak according to man,” cf. AV, RV). V. 15b begins in the Greek with homōs, which is usually understood to carry the meaning “nevertheless” (cf. AV, RV “yet”) and to be shifted forward from its natural position before “no one” by “trajection” (cf. 1 Cor. 14:7). 1  Less probably, the word could, if accented differently, be translated “also” or “likewise” (i.e., as with God’s covenant; cf. NIV “Just as … , so it is in this case”). 2 

“Will and testament” represents Gk. diathēkē, which is the LXX translation of Heb. berı̂ṯ, “covenant.” In support of the rendering “covenant” in our verse (AV, RV, NASB, NIV) it has been said that “the whole force of the argument demands that the diathēkē of verse 15 be of the same type as that in verse 17.” 3  But this seems to overstate the case, since “the point of comparison is simply that of inviolability, unalterability and therefore absolute validity.” 4  Although some scholars prefer not to tie down the meaning here to one or the other, 5  the many legal terms in the passage clearly suggest that the word is being used in the sense of Hellenistic law and point to “will, testament” as the more correct interpretation. 6  It is possible, however, to take diathēkē here in its original meaning of “an arrangement, a settlement drawn up and legally in force.” 7  This option seems to have an advantage over the other two, since it simplifies the divine-human comparison without affecting the point of the analogy. 8 

The NEB rendering of v. 15b (“when a man’s will and testament … or add a codicil”) seems to presuppose the practice of Roman law, in which “the testator … could add a codicil at any time that he chose, but after his death (or before it, for that matter) nobody else might do so” 9  —the validitation of the will coming only after the testator’s death (cf. Heb. 9:16f.). If (as has been maintained) the verse has Greek procedure in view, then Paul might be referring to the validation of a will by duly registering it and depositing it in the public record office of a city, after which “not even the testator was permitted to alter it, unless such permission had been expressly written into it”; 10  in this case it would be preferable to speak of the will being “confirmed” (AV, RV), “established” (NIV), or “ratified” (RSV, NASB) rather than “executed.” It has also been suggested that the reference is to the Jewish institution of mattenaṯ bārı̂ʾ, which had to do with the irrevocable gift of a healthy person. 11  Still, perhaps taking diathēkē in its original sense of a disposition or settlement will most readily permit a simple understanding of Paul’s words without need of modification: “When a deed of settlement is properly signed, sealed and delivered and the property legally conveyed, not even the original owner can revoke it or alter its terms.” 12  Whatever the precise background of the human analogy, Paul’s point is plain: even a human legal settlement is irrevocable in nature.

16 Next, Paul draws attention to a fact of biblical history, namely, that “the promises were pronounced to Abraham and to his `issue.’ ” The plural “promises” refers to God’s promise to Abraham as this was repeated on several occasions and couched in different expressions; 13  but in view of the association of the “inheritance” with the promise in v. 18, Paul may have in mind particularly that clause in the promise which concerned the giving of the land to Abraham and his descendants as an everlasting possession (Gen. 13:15; 17:8; cf. 12:7; 15:7). 14  Paul doubtless understands this in a spiritual sense, 15  although he does not pause to make this explicit.

Paul interprets the “issue” as a reference to Christ. He is well aware of the collective sense of sperma (Greek) or zera` (Hebrew) in the Genesis passages; 16  his identification of the “issue” spoken of in the promise as the Christ of history is not derived from a direct exegesis of the OT texts, but rather from an interpretation of them in the light of the Christ-event. His argument seems to be this: in the promise to Abraham, “it” (the subject of “does not say” is probably Scripture, as in v. 8, or possibly God) does not use the plural “issues,” referring to many, but “issue,” which, being a noun in the singular number, can refer to a single person—in fact, it does refer primarily to one individual, Christ. 17 

This piece of Pauline interpretation 18  is charged with being “allegorical in the sense that it no longer takes into account the original meaning of the words and has overstepped the limits set to allegory in rabbinical hermeneutics.” 19  Over against this, however, it should be remembered that (a) the word “issue,” both in Hebrew and in Greek, can be used to designate one definite descendant (e.g. Gen. 4:25; 1 Sam. 1:11); 20  (b) parallel arguments based on the singular or plural of Heb. zera` are not wanting in rabbinic literature; 21  and (c) Paul is speaking from the standpoint of fulfilled prophecy in the conviction that the “issue” of the original promise can, in the event, refer only to Christ. 22  According to Paul’s reading of history, then, Christ “is the true Heir of the promise, of the universal inheritance, and He determines the fellow-heirs” 23  —as vv. 26–29 will show.

17 With the words “What I am saying is this” Paul proceeds to apply the human analogy to the divine dealings in history: “a covenant” (RV, RSV, NASB) 24  had been validated by God 25  and therefore in force long before the giving of the law, which came only 430 years later. 26  From this minor premise, and the major premise already laid down in v. 15b, the conclusion follows that the law cannot possibly invalidate the original settlement so as to render the promise ineffective. 27  It is thus clear that Paul regards the promise to Abraham as a divinely ratified settlement or covenant and argues from its considerable priority to the law that its provisions cannot be made null and void by the later introduction of the law. 28 

18 Since the “promises” of v. 16a are spoken of as “covenant” in v. 17b, whose principal content is in turn referred to as “the promise” in v. 17c (the Gk. has epangelia, “promise,” in the singular; cf. AV, RV, RSV, NASB, NIV), we might speak of the original “covenant of promise.” 29  Its nature is underlined yet again when Paul points out that it was “by a promise” (v. 18b, RSV) that God bestowed the inheritance as a free gift on Abraham. 30 

As H. Schlier observes, three things are emphasized in this brief statement: (a) it is God who has made the decision, which therefore stands and is not to be broken; (b) the decision was for promise, not law; and (c) it was a gracious decision which forestalled all merit and excluded every meritorious achievement. 31  Since Abraham is a “conclusive test case” 32  and the OT event “still retains its (exemplary) meaning,” 33  the inference becomes inevitable (v. 18a): the inheritance is not “based on law” (NASB), more specifically the Mosaic law—if it were, the inheritance would belong to “those who rely on obedience to the law” (v. 10)—but is rather based on a promise and hence on faith. For just as Abraham was justified by faith in God’s promise (v. 6), so “the men of faith” are the true children of Abraham (v. 7) and will inherit Abraham’s blessing (v. 9). 34 



 1 E.g., Thayer s.v.; MHT 3: 337.

 2 E.g., BDF §450.2, followed by BAGD s.v. In this case the second object of comparison would have to be regarded as following (vv. 17f.; cf. 1 Cor. 14:9) rather than preceding ὅμως.

 3 Morris, Preaching, 91. Cf. Hughes, “Hebrews ix 15ff. and Galatians iii 15ff.”

 4 J. Behm, TDNT II: 129.

 5 E.g., Clarke, NT Problems, 154, n. 1; Whiteley, Theology, 85f.

 6 Cf., e.g., Behm, loc. cit.; F. Hauck, TDNT V: 760; G. Stählin, TDNT IX: 459, n. 172.

 7 Ridderbos 130.

 8 It would render unnecessary, e.g., the explanation that “no regard is paid to the fact that in the case of God’s testament the presuppositions of this validity … are very different from that of a human will, i.e., the death of the testator” (J. Behm, TDNT II: 129; cf. III: 1099).

 9 Bruce 171.

 10 Bruce 170f., describing the view of W. M. Ramsay.

 11 Betz 155f. (following E. Bammel).

 12 Bruce 170.

 13 Cf. Burton 184. Note the alternation between the plural, here and in v. 21, and the singular, vv. 17, 22.

 14 See Lightfoot 142 for two considerations which would seem to limit the references to Gen. 13:15 and 17:8.

 15 Cf. Rom. 4:13, where the promise to Abraham and his offspring appears in the form “that the world should be his inheritance”—a promise which is fulfilled in the penetration of the gospel of justification by faith into the world of the Gentiles.

 16 Cf. Rom. 4:18, where Abraham’s offspring of Gen. 15:5 is identified with the many nations of Gen. 17:5, interpreted as Gentile believers.

 17 On Paul’s argument here as an example of one mode of rabbinic direct or explicit exegesis (fastening on “the strict sense of a term”), cf. Cohn-Sherbok, “Exegesis,” 120f. For the suggestion that the clue to Paul’s interpretation of the “issue” as Christ lies “in the link-word ξύλον (`ṣ), found in both Deut 21:23 (cf. Gal 3:13) and Gen 22:6a, 7b, 9,” see Wilcox, “Upon the Tree,” especially 97f.

 18 Burton 509 reckons with the possibility that v. 16b (“It does not say … Christ”) is an early editorial comment which became incorporated in the text; this is definitely maintained by O’Neill 50f. It is true that Paul’s interpretation of “issue” here is at variance with that given in v. 29, but “Pauline dialectic leads not uncommonly to formal logical contradictions” (A. Oepke, TDNT IV: 619, cf. n. 82). It is also true that the omission of v. 16b would leave “a consistent connection” (Burton, loc. cit.); in this case the “issue” in v. 19c would find its antecedent in the “issue” of v. 16a. But, on the other hand, the reference to the arrival of the “issue” finds a more natural explanation in v. 16b than in v. 16a (cf. vv. 23–25, where the “coming” of faith is connected with the coming of Christ).

 19 Schoeps, Paul, 181, n. 3 (cf. 42, n. 2). But here this reminder is apt: “The 7 rules of Hillel and the 13 of Ishmael … show what can be gleaned from the literal wording” (F. Büchsel, TDNT I: 263, n. 24).

 20 Schoeps is therefore not accurate in saying that Paul’s deduction from the singular of Gen. 22:18 is “both in contradiction to the meaning of the word and to Biblical linguistic usage” (Paul, 181). Cf. Wilcox, “Promise,” who observes that the application of the promises (of the “seed” to Abraham and to David) to the Messiah “seems to have begun already with Jewish thought and exegesis” (16).

 21 Cf. Bruce 172f.; Longenecker, Paul, 59, n. 204.

 22 Cf. Newman, “Seed.” The matter is well expressed by Collins, “Exegesis,” 150: “The apostle bases his conclusion on two grounds: the use of the singular noun and the conviction that Christ is par excellence the seed of Abraham. In the first one can detect the scribe, but in the second the Christian is more in evidence.” Ultimately, however, “the argument is not a grammatical one but a theological one” (C. Brown, NIDNTT III: 524).

 23 J. Schniewind and G. Friedrich, TDNT II: 583. Cf. S. Schulz, TDNT VII: 545.

 24 It is interesting to note that NEB translates the single word διαθήκη as “a testament, or covenant”—thus correcting itself, as it were, or at least providing an alternative. To speak of God ratifying a will (e.g., Hooker, “Paul,” 50f.) in this instance is to complicate the thought unnecessarily, as may be demonstrated by the following comment: “The image and the thought are here very contradictory, for whereas a human will comes into effect only with the death of the testator, the will and testament of God (v. 15 …) … is put into effect as soon as it is drawn up, and from that point on it is exclusively and incontrovertibly valid” (J. Behm, TDNT III: 1100); cf. n. 8 above. If it be insisted that the same sense be given to διαθήκη in vv. 15 and 17, the best word would seem to be, not “covenant” or “will,” but “legal settlement” (see above on v. 15).

 25 On the variant reading with εἰς Χριστόν see Metzger 594.

 26 It is with reference to this law that the covenant is said to have been “confirmed before of God” (AV) or “previously ratified by God” (RSV, NASB). The figure 430 appears also in the MT of Ex. 12:40 as the period of Israel’s sojourning in Egypt, and in the LXX (and Samaritan text) of the same verse as covering the time of Israel’s sojourning in Egypt “and in the land of Canaan.” Gen. 15:13 (cf. Acts 7:6) gives the round figure of 400 years as the predicted period of Israel’s oppression in Egypt. The precise solution of the chronological difficulties—which in any case probably would not be of any particular interest to Paul (cf. Lightfoot 144; Betz 158)—seems to have eluded the commentators, but fortunately this does not in the least affect Paul’s argument here, which emphasizes only the considerable length of time separating the law from the promise.

 27 “Paul’s point is that what God has said is already in force by the mere fact that He has said it” (W. Foerster, TDNT III: 784, n. 36). This is borne out by the perfect participle προκεκυρωμένην (“validated”) as well as by the emphatic ὑπὸ τοῦ θεοῦ (“by God”).

 28 This temporal argument appealing from the law back to the Abrahamic promise “has an exact parallel in Jesus’ reference in the question of divorce to the ordinance of creation as revealing a deeper and more original form of the will of God” than the provision of the Mosaic law (Schoeps, Paul, 180). For other examples of this temporal argument as an accepted rabbinic method see Longenecker, Paul, 123, n. 63. The “law” here is identified by J. Guhrt, NIDNTT I: 370, with the elaborations of late Judaism; but on this see the editorial comment by C. Brown (371f.).

 29 This phrase in fact occurs in Barrett, First Adam, 60. Hooker (“Paul,” 51), however, thinks that Paul does not speak of God’s original promises to Abraham in terms of a covenant, the term διαθήκη being used (in her opinion) in the sense of “will” (cf. n. 24 above).

 30 Ridderbos 136, n. 13 says that χαρίζομαι without the object (as here) is a technical term for the law of inheritance, meaning: make a grant, deed something by will. But it is more natural to supply the object from the context; cf. BAGD s.v. 1.

 31 Schlier 149f.

 32 Guthrie 103 (he notes the emphatic position of “to Abraham”). Cf. E. Hoffmann, NIDNTT III: 72.

 33 BDF §342.5. Cf. Moule 14f., who calls this “the Perfect of Allegory.”

 34 “In laying the supreme emphasis on [the promise to Abraham] Paul parts company with Judaism. For Judaism the great thing was the deliverance from Egypt together with the giving of the law… . Paul reverses the emphasis. Judaism would read the patriarchal narratives in the light of Sinai. Paul insists on looking at Sinai from the standpoint of the promise to Abraham” (Manson, Paul and John, 45). Cf. Barrett, “Allegory,” 15.