(thus also Luke 11:4) things that over time "a fixed form of the doxology"17
S. Matt 6:12 dipiepev K1 f13 311 syc? co? Cl.
In this scheme (2) and (3) are only dialectal variants.
Is the present tense original, and is tli£ Didache ver sion older than the Matthean at this point?14 Or does the aorist d<pr)Kapev reflect an Aramaic perfect—and if so, in what sense?15 with some characteristic deviations from the latter.9 It is hard to suppose that the Didache quotes directly from the text of Matthew's Gospel.10 The agreements would rest on a common liturgical tradition.
Excursus:TheText of the Lord's Prayer It is important that the Didache contains a formu lated doxology at the end of the prayer. Clearly, the form of the Our Father handed on by Q did not yet have such a doxology; it is lacking in the Lukan text and also in the original text of Matthew. Adolf
Schlatter and Joachim Jeremias have instructed us, however,16 that the Our Father was one of the prayers
"with the 'seal,' i.e., with a freely formulated conclu sion." This means that, from the beginning, the one who prayed such prayers formulated a conclusion to the last petition ex tempore. It is in the nature of things that over time "a fixed form of the doxology"17
We will trace the differences between the Didache and the Matthean text.11 The singular ev tw ovpava ("in heaven") in place of the plural in Matthew is essen tially meaningless.12 In the "fifth petition" we find
TTjjc dtpeiXpv ("debt") against Matthew's tol opeiXppara ("debts").13 The Didache's dipiepev ("we for give") is interesting. Cf.:
1. Matt 6:12 aippuapev S* B Z f1 pc vgst syP f; Gregory of NyssaP12. Matt 6:12 dipLopev D (L) W A 0 565 pc syc? co?
(thus also Luke 11:4) Bahr, "The Use of the Lord's Prayer in the Primitive Church," JBL 84 (1965) 154. Nevertheless, the agreements between Matthew and the Didache show that the paradigm of the Lord's Prayer was already more or less fixed in this sphere of tradi tion.
9 Cf. Ernst Lohmeyer, Das Vaterunser (5th ed.; Gottingen: Vandenhoeck & Ruprecht, 1962) 7 (ET:
The Lord's Prayer [New York: Harper & Row, 1965] 15-16). Cf. also the synopsis in table 9. Schollgen,
"Church Order," 66: Did. 8.2 is tradition.
10 Thus Funk, Patres apostolici, 1.19; Hemmer, Doctrine, xcv; Schaff, Church Manual, 188; Massaux, Influence,
3.154-55. To the contrary: Koester, Synoptische Uberlieferung, 203-9; Audet, Didache, 171-73; Lohmeyer,
Vaterunser, 7 (ET: Lord's Prayer, 16); Giet, L'Enigme, 200; Rordorf and Tuilier, Doctrine, 86. Against Koester, most recently J. B. Bauer, "Aspekte des
Kanonproblems," Meqor Hajjim: Festschrift fur Georg Molin zum 75 Geburtstag (Graz: Akademische Druckund Verlagsanstalt, 1983) 30-31. The arguments presented by Koester, says Bauer, are not conclusive.
The Didache's aipiepev was also found in the Koine text of Matthew's Gospel and in the version in the Apostolic Constitutions-, the latter also has ei> rolq ovpavoiq and b<peiXr)paT<x (the last again, in my opinion, not persuasive, because it is reasonable to suppose that the Constitutor used his text of Matthew to make corrections). Bauer asserts:
"However the problem now stands, whether the Didachist knew the Gospel of Matthew or not, that he records the Matthean form can also be under stood in this way: He knew not only this longer ver sion but also the shorter text of the Our Father as it is found in Luke, and he wished, so to speak, to inculcate the longer version as canonical" (p. 31; with reference to J. B. Bligh). This can scarcely be the case. Wengst (Didache, 26-27) has now also asserted that this passage is a citation from Matthew's Gospel. The liturgy or a written source are not genuine alternatives (p. 26), and the devia tions from the traditional text of Matthew "do not weigh heavily" (p. 27).
11 These differences are noted in table 9 by the use of underlining in the text of the Didache.
12 Koester (Synoptische Uberlieferung 206) suggested that the singular in the Didache might even be origi nal. Gerhard Schneider, "Das Vaterunser des
Matthaus," in A Cause de TEvangile: Etudes sur les Synoptiques et les Actes offertes au P. Jacques Dupont,
O.S.B., a Voccasion de son 70e anniversaire (LD 123; Paris: Cerf, Publications de Saint-Andre, 1985) 63 n.
23: the singular is secondary. The Apostolic Constitutions has the plural ev tolq ovpavolq (proba bly corrected according to the familiar text of Matthew).
13 Koester (Synoptische Uberlieferung, 206) considered whether the didachistic or Matthean form is the older. Schneider ("Vaterunser," 63 n. 23) thinks this is probably a secondary accommodation. The Apostolic Constitutions again accommodated to Matthew with rd oipeiX-ppara.
14 Cf. the suggestions in Koester, Synoptische Uberliefe rung, 206-7. Differently Joachim Jeremias, Theology,
196 n. 2: dipiopev or d<pLepev are softenings ("so far as there is no question of Lucan influence").
15 Jeremias, Theology, 201: 'ld(pr)Kapev goes back to the Aramaic iebaqnan, which is meant as a perfectum coincidentiae and is therefore to be translated: 'as herewith we forgive our debtors.'"
16 Adolf Schlatter, Der Evangelist Matthaus (6th ed.; Stuttgart: Calwer, 1963) 217; Jeremias, Theology,
203. Cf. also Joachim Jeremias, "Abba," in idem, Abba, 170-71 (ET: The Prayers of Jesus [SBT 2.6;
8.2-3 was adopted. Such a fixed form for the doxology appears for the first time here in the Didache; it later emerges (although expanded; see below) in the "majority text" of Matt 6:13 (end) (L W 0 0233 f13 2Jt f q boP').18 We do not know when and where these doxologies emerged. The passage in 1 Chr 29:11-12 that is often adduced as the model for them is too far removed in its wording. Rabbinic traditions also offer only analogies. Cf. the doxology: "praised be the name of his glorious kingdom forever and ever." This is how the priest and people respond to the great confession of sins in the temple cult (Str-B 1.423).
The same benediction was spoken after the begin ning of the Shema (Str-B, ibid.); compare also the Alenu prayer, whose conclusion is introduced by "for thine is the kingdom" (Str-B, ibid.). The object of the benediction is always the kingdom, as is the begin ning of the doxology to the Matthean Our Father as later attested by the "majority text." Here we should note that it is precisely this kingdom benediction, with which the later doxology of the Lord's Prayer in Matt 6:13 in the majority text begins, that is lacking in the Didache. The Didache doxology is accordingly not in three parts (kingdom, power, glory), but only two (power and glory). In addition we note that the mention of the kingdom is also lacking in the Sahidic version of the doxology to the Lord's Prayer (Matt
6:13 [end]), and the same is true of one manuscript of the Fayyumic. Thus the two-part textual form of the doxology in the Didache agrees with the Coptic witnesses just listed (with the sole exception of the "Amen," which is lacking in the Didache):19
Sahidic:20 Didache: se tiuk ts tsom on oov eanv T) dvpafug mm neooy nai r) So£a uix meNS221 eig rovg aiavag.
2XMHN.
This should be compared with the text of a fragment in the Fayyumic(?) dialect: xe tiok ts ts^m hn rie^y oixeiiG2 2XHHH.22 Characteristic of all these examples is the absence of the first member, r\ (daouAeLa KaL
("the kingdom and").23 The agreement of the Didache with this form of the text is remarkable. Nevertheless, it would be difficult to draw conclusions from this regarding the origins of the Didache,24 London: SCM, 1967] 106-7). The quotation is from edited by M. Maspero) contains Matt 5:46—6:18 with lacunae. David locates the fragments "in the central Egyptian dialect that has customarily been called Fayyumic" (p. 80). The text in question is on p. 84 of the article. The Middle Egyptian version of Codex Scheide (ed. Hans-Martin Schenke [TU 127;
Berlin: Akademie-Verlag, 1981] 64) has no doxology with the Lord's Prayer. The Bohairic version is split.
Parts of it evidence the (three-part) doxology, and others have the text of the Lord's Prayer without a doxology. Cf. George Horner, The Coptic Version of the New Testament in the Northern Dialect Otherwise
Called Memphitic and Bohairic (3 vols.; Oxford:
Clarendon, 1898-1905; repr. Osnabriick: Zeller, Theology, 203.
17 Jeremias, Theology, 203.
18 For the variants in 2148 pc g1 k syc p sa, see the information in Nestle-Aland. In particular, for the absence of the kingdom benediction in sa (and fay), and in the forms in k and syc, see immediately below.
19 It is true that the kingdom benediction is also lack ing in k, but the text of k shows only a one-part form overall. See n. 23 below.
20 Text from George Horner, The Coptic Version of the New Testament in the Southern Dialect Otherwise Called
Sahidic and Thebaic (7 vols.; Oxford: Clarendon, 1911-24; repr. Osnabriick: Zeller, 1969) 1.46.
Identical in its text is the version of ms. sa 9 (New York, Pierpont Morgan Library, M 569), not yet appearing in Horner's edition. Cf. Franz-Jurgen
Schmitz and Gerd Mink, eds., Lisle der koptischen Handschriften des Neuen Testaments, I: Die sahidischen
Handschriften der Evangelien, 1. Teil (ANTF 8; Berlin and New York: de Gruyter, 1986) 19-24. The text of this passage was very kindly provided for me by Franz-Jurgen Schmitz of the Institut fur neutestamentliche Textforschung.
21 Some mss. read rtaaiX2Te instead of neooy; f1 (cf.
Horner, Coptic Version, 47) reads SN62 FieNe2 instead of NieNS2 (cf. also the text of 2148 pc k).
22 Text from J. David, "Fragments de l'Evangile selon Saint Matthieu en dialecte moyen-egyptien," RB, n.s.
7 (1910) 80-92. The text consists of three fragments of Matthew's Gospel, the first of which (originally
1969) 1.38-39.
23 In Codex Bobbiensis (k), not only is the correspond ing term to rj /SomAeio: kccl lacking, but also that for Kai r) <56£a. The text contains only a one-part doxology. At the same time (with 2148 pc sa™5), the additional to)p aiapoip is presupposed: quoniam est tibi virtus in saecula saeculorum (Itala 1.31). In the Cureton Syriac the corresponding term for r) SvpaHLq kcx'l is absent. See the other variants in NestleAland.
24 Harnack (Lehre, 26; and "Prolegomena," 168-70) originally saw this agreement with the Sahidic tradi tion as an indication of the Didache's origins in Upper Egypt. Should we consider the possibility of a subsequent adaptation of the Didache text to the Sahidic version, so that we would have to conjecture the original Didache text? Still, the three-part text is also attested in the Apostololic Constitutions. But the
The presentation of the Lord's Prayer is followed in More important than the (purely accidental?) agreement of our Didache doxology with the Coptic textual witnesses just mentioned is another matter.
The specific form of the doxology in Did. 8.2 is found again in the doxology of the meal prayers in 10.5, and with only slight variation also in 9.4. In general, the doxology of the Lord's Prayer and those of the meal celebration in chaps. 9-10 belong together, as the following synopsis illustrates:
(a) otl aov eaTiv r\ bvvapLq Kai p Soija eiq rovq v. 3 by a rubrical note: the readers or hearers are called upon to speak this prayer three times a day.56 This corre sponds to the Jewish custom of having a fixed time for prayer "three times each day" (Dan 6:9, 11, 12 LXX: rpiq rrjq ppepaq).27 The provision in the Didache (or that of the liturgist) maintains the Jewish Tplq rrjq ppepaq but requires the replacement of the Jewish prayer with something different: namely, the Lord's
Prayer just cited. In this way also the Christians show themselves different from the invoKpLTaL. This passage in Did. 8.3 is of immense importance for the history of Christian worship. It may be noted that the hours for the threefold daily prayer are not yet fixed.28 aiuvaq (8.2) (b) aoi. p bo^a eiq roiiq aiuvaq (9.2) (b) aoi p bo^a eiq Tovq aiavaq (9.3) (a) otl aov eaTiv p bo^a Kai p bvvapLq bed Ipaov
XpLorov eiq rovq aiuvaq (9.4) (b) aoi 17 bo^a eiq rovq aiuvaq (10.2) (b) aoi p <5o|a eiq Tovq aiavaq (10.4) (a) otl aov eaTiv p bvvapLq Kai p bo^a eiq rovq
In summary, the whole section Did. 8.1-3 reveals a Jewish-Christian community in the process of setting aiuivaq (10.5).
The constancy of the formula forces us to view the doxology of the Lord's Prayer that appears in the Didache alongside those of the meal celebration. The one-part short formula (b) is contrasted with the long formula (a), which has two parts.25 In each of these formulae the mention of the (iaaiXeia is lacking. In conclusion, the Didache tradition is acquainted with a doxology existing in a short and a long form, both of which have found their way into the meal prayers and itself apart in polemical fashion from its surrounding
Jewish environment. This distinction is brought into action in this passage by means of the rejection of specif ic ritual customs of Judaism and their replacement by rit ual customs proper to Christians. the Lord's Prayer. We cannot exclude the possibility that their original Sitz im Leben was in the meal prayers and that they were transferred from there to the Lord's Prayer. already of the hours of "terce, sext, and none."
Tertullian De orat. 25 (CChrSL 1.272-73) attests prayer at the third, sixth, and ninth hours (with bib lical and trinitarian reasoning), but distinguishes these hours of prayer from the legitimis orationibus, quae sine ulla admonitione debentur ingressu lucis et noctis ("the legitimate prayers, which, without any command, ought to be offered at first light and at dusk," 25.5 [272-73]). The hours of terce, sext, and none are also found in Tertullian De ieiun. 10.3
(CChrSL 2.1267). Clement of Alexandria Strom. 7.7, 40.3 (GCS2 3.30) mentions the third, sixth, and ninth hours as times for prayer, "but the Gnostic prays throughout life" (aAA' ovv ye 6 yvwaTLKoq
Ttapa oXov evxerat tov (3Lov). See Origen De orat.
Apostololic Constitutions would have been accommo dated here to the text of Matthew, which was famil iar to its author, and the two-part form of the doxology (to be discussed immediately below), which has its parallels in the doxologies of the meal prayers, speaks against a conjecture in the Didache text. See below.
25 For the one-part formula cf. Rom 11:36b; for the two-part formula see the similar expressions in 1 Pet
4:11; Rev 1:6.
26 Of course, this refers to private prayer: Harnack, Lehre, 27; Klauser, Doctrina, 22; Mattioli, Didache,
95ff.; denied by Clerici, Einsammlung, 95; and Wrege, Bergpredigt, 103: "public, liturgical praxis of the communities." Our passage may be quoted in Lib. grad. 7.20 (Kmosko, 186).
27 For the usage in the tannaitic period, see Str-B 2.697-702.
28 Funk differs [Patres apostolici, 1.20) and thinks 12.1 (GCS 2.325); Cyprian De Dominica Oratione 34
(CSEL 3.1.292).