Covenant (in the New Testament)

See also Covenant

COVENANT (in the NEW TESTAMENT) (διαθήκη, G1347, testament). A legal disposition, esp. of God for man’s redemption.




The Gr. noun διαθήκη, G1347, occurs thirty-three times in the NT, with over half of these in Hebrews; G. Vos designated this book as “The Epistle of the Diatheke,” PTR 13 (1915), 587. In the KJV the noun is tr. thirteen times as “testament” and twenty times as “covenant,” though six times with “testament” in the mg. The general RSV rendering is “covenant,” though on three occasions (see below, II) it renders διαθήκη, G1347, as “will,” “testament”. At two points the KJV also utilizes the verb “to covenant” and on one other occasion the adjective “covenant breaker.”


At the same time, there existed in classical Gr., from the time of Aristophanes (427 b.c.; Birds, 440ff.) a secondary and limited usage by which διαθήκη, G1347, also signified an ordinance or even a dipleuric, treaty-like “convention or arrangement between two parties, a covenant,” to which definition the 8th ed. of Liddell and Scott added, “and so in later writers” (p. 346; cf. TWNT, II:127, 128). More recent lexicographers have, however, insisted that by Hel. times the term’s signification was “exclusively last will and testament” (AG, p. 182), which was the “ordinary and invariable contemporary [1st Christian cent.] meaning,” employed with “absolute unanimity” in the papyri and inscrs. (MM, pp. 148, 149).

The thirty-three NT occurrences of diathēkē break down into three general groupings, determined by the subject or initiator of the “covenant” concerned: whether the subject be man, variously involved (three times); Christ, acting on the behalf of His Church (fourteen times); or Jehovah, accomplishing redemption for Israel (sixteen times, including Gal 4:24 which makes reference to the Church as well). The last mentioned group consists primarily of quotations from the OT.

Covenants between men.

Standing in contrast with the OT’s frequent descriptions of human agreements and treaties, the NT does not deal with man-made covenants as specific historical phenomena. Three times, however, it does allude to a diathēkē between men, for the purpose of illustrating Christ’s redemptive activity for His own, namely in Galatians 3:15; Hebrews 9:16, and 17. The first adduces the subject of “a man’s will,” for the sake of emphasizing its feature that “no one annuls...or adds to it, once it has been ratified” (Gal 3:15 RSV). The KJV rendering at this point of diathēkē as “covenant” appears unwise (cf. J. Murray, The Covenant of Grace, p. 30), not simply because of the term’s normal testamentary connotations and because of the contextual stress upon an “inheritance” (v. 18), but primarily because covenants are not, under most circumstances, incapable of modifications. Bequests, on the other hand, remain fixed, particularly after the death of the testator, and even prior to this point. While the latter characteristic may not have been rigidly established in Rom. law, the provisions of a last will in Syro-Grecian law (which applied in Galatia) were not permitted to become subject to modification, once they had received public sanction and had led to such adoption proceedings as may have been involved (cf. W. M. Ramsay, EXP [1899], 57ff.; G. Vos, Biblical Theology, p. 34). The other two vv. speak to the necessary dependence of a diathēkē upon the death of the party who has set it up: “Where a will is involved, the death of the one who made it must be established”; and, “A will takes effect only at death” (Heb 9:16, 17 RSV). Clearly, in its diathēkēs between men, the NT signifies what one means by “testament.”

Christ’s covenant with God the Father.


The problem of such relatively slight utilization of the covenant concept, outside of Hebrews, to describe the work of Jesus Christ, esp. when one compares the dominant position that it occupies in the OT, has led to no small discussion. Because of an assumed deterioration within the covenant idea during later OT days, G. E. Mendenhall proposes that “the covenant patterns were not really useful as a means of communication” (IDB, I:723). His assertion appears to be based upon negative critical assumptions about the externalized character of the Pentateuchal law and about its initial identification with the covenant only in the time of Ezra. This then leads to his disparaging conclusion that “the NT experience of Christ was one which could not be contained within the framework of a quasi-legal terminology or pattern of thought and action” (ibid.). This same writer later states, with greater plausibility, that it was an overemphasis upon law among the intertestamental Qumran sectarians and the NT Pharisees that worked in combination with the imperial government’s opposition to covenanting (anti-Rom.) secret societies, so as to “make it nearly impossible for early Christianity to use the term meaningfully” (ibid., I:722).

G. Vos, on the other hand, refuses to disparage OT thought in the former way and suggests rather that the NT’s emphasis upon the Person of Jesus Christ caused it to turn primarily to those portions of the OT that were descriptive of the coming Messiah, but which were so often separated from God’s revelations of His covenant (PTR, 13 [1915], 588). Vos proceeds to note that the additional and fresh emphases of Christ and the apostles upon concepts such as “the kingdom of God” or “the church” tended to restrict their use of covenant terminology to those passages that consciously sought to compare the older and the newer Testaments, as, for example, at the Lord’s Supper or in the Epistle to the Hebrews, where almost the whole of God’s redemptive plan is summed up in the doctrine of the two covenants (ibid., pp. 587-590). At these points the newer Christian development may then be compared either with valid Mosaic truths (as in 2 Cor 3) or with the perversions that are attributable to NT Pharisaism (as in Gal 4) and which hardly deserve the designation “covenant” in the first place. Early patristic writers exhibited no hesitancy in employing covenant terminology and concepts, e.g., Epistle of Barnabas, 13, 14, or Irenaeus, Against Heresies, III, 11, 8. Even Mendenhall grants that “for a time at least, the early Christians did regard themselves as a community bound together by covenant” (op. cit., I:722), though he confuses the covenant as a human, organizational compact with its basic character as a divinely redemptive testament (see below, IV-C; and COVENANT, THE NEW, II)


As suggested by the above quotations, Christ’s diathēkē is, moreover, definitely testamentary in character. The implications of the Lord’s words, “This is my blood of the diathēkē, which is poured out for many for the forgiveness of sins” (Matt 26:28), have been analyzed by A. R. Fausset as follows: “These requisites of a testament occur—1. A testator; 2. heir; 3. goods; 4. the testator’s death; 5. the fact of the death brought forward; 6. witnesses; 7. a seal, the sacrament of the Lord’s Supper, the sign of His blood....The heir is ordinarily the successor of him who dies and so ceases to have possession. But Christ comes to life again, and is Himself (including all that He had), in the power of His now endless life, His people’s inheritance” (A Commentary, Critical, Experimental and Practical on the Old and New Testaments, VI: 556). Shortly thereafter the Lord affirms, “I διατίθεμαι, G1415, bequeath unto you a kingdom” (Luke 22:29), though this literal meaning cannot always be insisted upon since His next words require a broader, non-testamentary usage: “as my Father appointed, διέθετο, unto me” (cf. Vos, PTR, 13 [1915], 608). Vos notes a trend, “moving away from the rendering ‘covenant’ to the other tr.” (ibid., p. 593); and, as he later states, “There are passages...for instance, those recording the institution of the Lord’s Supper, where a further return to ‘testament’ may seem advisable” (Biblical Theology, p. 35). The emphasis of the communion supper upon Christ’s sacrificial death, not simply upon a mealtime fellowship, with other characteristics of Christ’s diathēkē, are developed further under The New Covenant (q.v.); but this much does appear, that, just as in the case of the diathēkēs betw een men, so in the case of Christ’s diathēkē with God the Father, “covenant” in the NT warrants more precise definition as one’s last will and testament.

God’s covenant with Israel.


Finally, Romans 11:26, 27 quotes Isaiah 59:20, 21, in its anticipation of God’s still future covenant with Israel: “My covenant with them when I take away their sins.” The point of the reference seems to go beyond that of Jeremiah’s new testament (Jer 31:31-34), which was fulfilled in Christ’s sacrificial first coming (see THE NEW COVENANT, IV), and to imply Ezekiel’s testament of peace (Ezek 37:25-28), which will some day transform not only the Jewish nation, but also all the nations of the world, at His victorious Second Coming. The progressive revelations of the diathēkē will thus have achieved that same goal of final salvation which the prophets anticipated. In fact, the Apostle John closes out human history in the first half of his Apocalypse with an OT covenant-object, “In heaven...there was seen in his temple the ark of his covenant” (Rev 11:19).



God’s NT revelation provides His own normative explanation about the nature of the old diathēkē: it not only confirms the validity of the OT’s affirmations; it also establishes the ultimate character of those aspects of divine redemption that had, perforce, to remain unclarified prior to the incarnation of Jesus Christ (see COVENANT IN THE OT). Within the context of the NT the most crucial passage is Hebrews 9:15-22, esp. vv. 16 and 17. The former reads, “For where a diathēkē is, there must also of necessity be the death of the testator” (KJV). These words indicate a last will. Marcus Dodds, B. F. Westcott, and a few others have argued that the “death” signifies only a self-imprecation of dismemberment for non-fulfillment (cf. Jer 34:18-20), which was “brought out” by the ratifying ceremonies of a covenant. Such an explanation appears unlikely, for it is not the threat of death but the death itself that is brought out or adduced. Furthermore, the threat of death does not seem always to have been a necessary element in covenantal, as opposed to testamental, thought; cf. David and Jonathan. Verse 17 is decisive: “For a diathēkē is of force after men are dead: otherwise it is of no strength at all while the testator liveth” (KJV). As Dean Alford stated: “It is quite vain to deny the testamentary sense of diathēkē in this v....I believe it will be found that we must at all hazards accept the meaning of testament, as being the only one which will in any way meet the plain requirement of the verse” (The Greek Testament, IV:173, 174; cf. the renderings of ASV, RSV).

From this point it follows naturally to conclude with E. Riggenbach (Der Begriff Διαθήκη im Hebräerbrief) that in Hebrews as a whole the term likewise connotes “testament” (contrast Vos, PTR, 3 [1915], 617, 618). The same conclusion may, however, be applied to all the NT’s references to the older testament. Vos favors a true testamentalism in Acts 3:25 (following Deissmann, op. cit., p. 175; cf. MM, p. 148), where Luke states that the diathēkē that God made with the Israelite fathers has been realized in Jesus Christ; Paul’s stress rests upon legal terminology and the fact of inheritance (Gal 4:24-30). The testamentary significance of the NT term diathēkē when used for Christ’s newer covenant with the Father and for covenants between men has already been noted (see above, II, iii). Concerning the latter, Vos contends: “That Paul (Gal 3:15)...gives this specific turn to the idea...cannot, of course, give plausibility to the assumption that the LXX associated God with the idea of a ‘last will’” (op. cit., pp. 601, 602; vs. Deissmann, loc. cit.). Furthermore, the fact that the NT uses diathēkē for God’s covenant with Israel and that it specifies the meaning “testament” does not prove what the LXX trs. may have had in mind either. It does show that the divine author of the OT intended the saving bequest of His Son Jesus Christ from the very inception of His revelations. Patristic writers even went so far as to say that, since Israel had from the beginning forfeited the covenant by their idolatry, there really was no “new covenant” at all, but just the one testament in the death of Jesus, for Christians only (Barnabas, 4:8-18; 13-14:5).


B. F. Westcott, The Epistle to the Hebrews (1889), esp. 298-302; G. B. Stevens, The Theology of the NT (1906), 490-514; E. Riggenbach, Der Begriff Διαθήκη im Hebräerbrief (1908); A. Carr, “Covenant or Testament? A note on Heb 9:16f.,” EXP, VII: 7 (1909), 347ff.; L. G. da Fonseca, “Διαθήκη—foedus an testamentum?” Biblica, 8 (1927), 9 (1928); MM, 148, 149; J. Behm, “διατίθημι, διαθήκη,” TWNT (1935), II: 105-137; G. Vos, Biblical Theology (1948), esp. 32-36; R. Bultmann, Theology of the NT (1951), esp. I: 97, 98, 110, 340; R. Campbell, Israel and the New Covenant (1954); J. Murray, The Covenant of Grace (1954), esp. 25-30; J. B. Payne, Theology of the Older Testament (1962), 82-85.

International Standard Bible Encyclopedia (1915)

Diatheke, was the word chosen by the Septuagint translators to render the Hebrew berith, and it appears thus nearly 300 times in the Greek Old Testament in the sense of covenant, while suntheke and entolai are each used once only. The choice of this word seems to have been occasioned by a recognition that the covenant which God makes with men is not fully mutual as would be implied in suntheke, the Greek word commonly used for covenant (although not a New Testament word), while at the same time the rarity of wills among the Jews made the common sense of diatheke relatively unfamiliar. The Apocryphal writers also frequently use the same word in the same sense and no other.

In the New Testament diatheke is used some thirty times in a way which makes it plain that its translation must be "covenant." In Ga 3:15 and Heb 9:15-17 it is held by many that the sense of covenant must be set aside in favor of will or testament. But in the former passage it can be taken in the sense of a disposition of affairs or arrangement made by God, a conception in substantial harmony with its regular New Testament use and with the sense of berith. In the passage in Hebrews the interpretation is more difficult, but as it is acknowledged on all hands that the passage loses all argumentative force if the meaning testament is accepted, it seems best to retain the meaning covenant if possible. To do this it is only necessary to hold that the death spoken of is the death of the animal sometimes, if not, indeed, commonly slain in connection with the making of a covenant, and that in the mind of the author this death symbolized the death of the contracting parties so far at least as to pledge them that thereafter in the matter involved they would no more change their minds than can the dead. If this view is taken, this passage falls in line with the otherwise invariable use of the word diatheke by Jewish Hellenists.

See Testament.


Lightfoot, Commentary on Gal; Ramsay, Commentary on Gal; Westcott, Commentary on Hebrews; article on Heb 9:15-17, Baptist Review and The Expositor., July, 1904.