Church Order or Canons of the Apostles (Can. 4.1—13.4 or out adding or subtracting anything"] is reminiscent of
enough, however, the quotation from the Two Ways trac tate extends at first only as far as 13.4 (cf. Did. 4.8), ttoow paAAov ev roiq ^vr\Tolq. The rest of the "way of life" common tradition. In that case, Doctrina apostolorum appears to be an independent writing, or better, the Latin translation (made in antiquity) of a Greek original containing nothing but the Two Ways tractate.17 The Greek original has been lost.18 Whether the Greek origi nal was already entitled 5l8qlxV airooToAon', or whether that title was first given to the Latin translation
(as [de] doctrina apostolorum)19 is a point deserving fur ther consideration.20
(with the exception of Did. 4.13b, for which see immedi ately below), and the whole "way of death" have not been incorporated into the canon. (Something from the epilogue to the tractate, which is absent in the Didache, could be concealed in 14.3a: see below. The expression in 14.3b, pvAa^ecq a TrapeAafieq juqre Trpoodelq ppre vipaipwv ["You shall keep what you have received, with out adding or subtracting anything"] is reminiscent of Did. 4.13.) It is tempting to see the text as dependent on 3. The Two Ways tractate in the so-called Apostolic
Church Order or Canons of the Apostles (Can. 4.1—13.4 or 14)21 would be related to the text of the Didache in some what similar fashion, if it were not that the tractate is here attributed to the words of individual apostles (cer tainly a later stylization of the material).22 Remarkably the Didache,23 the more so as Can. 12.3 includes the expression t)%lu>0€P not dodrjvou TrvevpaTiKRV TpcHppv
K<xi ttotop kocl focx)v a'iWVLOV ("He graced you with spiri17 Those (e.g., Muilenburg) who derive the Two Ways tractate from Barnabas, and consider Doctrina apos tolorum the Latin translation of the Didache, typi cally get into serious difficulties. They must explain not only the (in this case) "epitomatic" character of Doctrina apostolorum, but also its occasional resem blance to Barnabas (against the Didache). But how?
"The Latin version of the Teaching of the Twelve Apostles shows clear traces of familiarity with the Epistle of Barnabas" (Muilenburg, Literary Relations,
27; cf. 45). This, and the homiletic context of the tradition, "supports the belief that the Latin version is in actuality a homiletical extract and not a true copy or translation of an original Two Ways
Teaching" (ibid., 42). Cf. 45: "The homiletic charac ter of the Latin version. . . ." Cf. earlier Robinson,
Didache, 73-83; and thereafter Vokes, Riddle, 17-21, Can. 12 (Thomas) = Did. 4.1-2. (This canon again inserts a short excursus.) Can. 13 (Cephas [!]) = Did. 4.3-8.
On Can. 14 (Bartholomew), see below.
The rest of the text is also attributed to apostles, but the material is no longer taken from the Two
Ways tractate. On the list of apostles, see below.
23 Thus Bryennios, Aibaxp, and o' (dependent on Barnabas, Didache, Apostolic Constitutions, and oth ers); Schaff, Church Manual, 19, 127-32 (dependent on Didache); Vokes, Riddle, 72, and passim (depen dent on Didache and Barnabas); Barnard, "Later
Flistory," 103-4 (Duae viae, Barnabas, and Didache); for Harnack and Wengst, see below.
24 pplp be exotpioa TrvevpaTiKpv Tpoippp kocl ttotop kcu fapp aiuPLOP bia too Troabog oov. Did. 10.3c and Canons, would show that its author also knows the second part of the Didache, and this would sup port the idea of literary dependence in the first part. Cf. Harnack, "Prolegomena," 210-11 n. 34.
(Harnack supported, on pp. 210-11, the opinion that the Canons of the Apostles was dependent on the Didache and Barnabas.) Of course, Harnack wished and passim (the Latin version is simply a homiletic extract and corrected the Didache by using
Barnabas).
18 For the type of Latin translation, see the general opinion of Giet, L'Enigme, 118. But his supposition that there were two stages of translation remains arbitrary.
19 Doctrina apostolorum g: de doctrina apostolorum f. The title in g may be the more original.
20 For more on this, see below.
21 On this writing, see above.
22 Can. 4 (John) = Did. 1.1-1,2b.
Can. 5 (Matthew) = Did. 1.2c; 3a.
Can. 6 (Peter) = Did. 2.2-7.
Can. 7 (Andrew) = Did. 3.1-2.
Can. 8 (Philip) = Did. 3.3a. (This canon inserts a rather long excursus on the key word emBopiot.)
Can. 9 (Simon) = Did. 3.3b.
Can. 10 (James) = Did. 3.4.
Can. 11 (Nathanael) = Did. 3.5-10. to establish dependence not only on Did. 10.3 but also on 13.1-2. In Apostellehre, 31-32, he attempted to understand the whole complex in light of his hypothesis of a basic document (see below), and this led him to a singular variant on this hypothesis.
Against the thesis that Can. 12.3 rests on Did. 10.3 + 13.1-2, see Ehrhard, Altchristliche Litteratur, 57.
25 Kraft (Didache, 10) complicates the matter: "Most witnesses also include brief verbal parallels to Didache 10.3b and 13.1/2 (?) in the expansion which follows the admonitions of Didache 4.2, but this material is not extensive enough to encourage the belief that our full Didache was used for the Two
Ways of CO [= Apostolic Church Order]. Rather, CO
Introduction Can. 12.2
Did. 4.2
'EK$r)Tr)oeiq be Kad' ppepav Ta npoouira rav dylav tual food and drink and eternal life"), which is reminis cent of Did. 10.3.24 One must admit that this is a relative
Epit. 9
'EK^rpoecq be ainov 'EK$r]Tr)06iq be to npoouarov avrov Kad' rpj.epai' ly strong argument. Still, while noting that this is a liturgical expression that the Canons did not necessarily derive from the Didache, 1 wish to maintain the hypothe sis that the text also received the tractate not directly from the Didache but from a tradition common to it and to the author of the Didache.25 The same will apply, con sequently, to the Epitome, which stands in a literary rela tionship to the Canons 26 Here again, the material of the tractate is attributed to apostles (eleven in number),27 and this time the whole "way of life" is written out (but the "way of death" is again absent).28 In making a deci sion about dependence on an independent Two Ways tractate, one should not be led astray by the expression p^Lwoe oe 8C avrov bo'dpvac TrvevpartKpv rpoippv nod fapv aiuviov (Epit. 9), which appears again here, some what shortened.29
Kai rovq Xocirovq Kai rovq XoLiroiiq dyiovq ayLovq Wengst thinks that the deviations between Epitome and the Canons derive from an additional use of the Didache or its Two Ways tractate. The text of the Canons is said to be a combination of the "Moral
Teaching of the Twelve Apostles" and the Didache.
But Barn. 19.10 offers something similar:
Kai €nt;r)Tr)06Lq nad' enaoTiqv ppepav to. irpoouTra rcov ayLuv In my opinion, it is not the text of the Canons but that of Epitome that requires explanation. It rests simply on an abbreviation of the source of its version of the Two Ways tractate. Wengst supposes that there is a similar phenomenon a little later:
In his investigation of the relations existing here, Wengst reached a different result.30 He discusses
(Didache, 10) the deviations between Epitome 9 and Canons 12, referring first to the passage c/cfrjr. be . . . ktA. (I add the Didache parallels): tains the whole "way of life" (without the piece found in Did. 4.5, 7, 10-11), while Canons breaks off at Did. 4.8 and then follows immediately with the epilogue of the tractate (as I suspect; see above);
(4) that the two writings (Canons and Epitome) pres ent only the way of life from the Two Ways tractate
(and in Canons it is already abbreviated) is connect ed with this naive stylization in terms of utterances of the eleven: the text breaks off when the list has been exhausted (cf. Giet, "Enseignement," 235).
Wengst's opinion is somewhat different (Didache, seems to have added excerpts from at least
Barnabas and possibly the Didache to the Dctr-like form of the Two Ways on which it is based."
26 On this document, see below.
27 Epit. 1 (John) = Did. 1.1-2b.
Epit. 2 (Matthew) = Did. 1.2c.
Epit. 3 (Peter) = Did. 2.2-7.
Epit. 4 (Andrew) = Did. 3.1-2.
Epit. 5 (Philip) = Did. 3.3a (without the excursus found at this point in the Canons).
Epit. 6 (Simon) = Did. 3.3b.
Epit. 7 (James) = Did. 3.4.
Epit. 8 (Nathanael) = Did. 3.5-10.
Epit. 9 (Thomas) = Did. 4.1-2.
Epit. 10 (Cephas [!]) = Did. 4.3-4, 6, 8.
Epit. 11 (Bartholomew) = Did. 4.9, 14a, 13a, 14b, 12, 10): "The way of death is also omitted, probably as a result of the division of the material among the eleven apostles, who preach the way of life, not the way of death." There is still another proposal in Rordorf and Tuilier, Doctrine, 119 n. 1.
28 That in Epitome the stylization of the material of Did. 3.1-6 in "teknon sayings" is lacking is not some
13b, 14c [sic].
For this list of eleven apostles, cf. Giet, L'Enigme, thing that I consider a sign of greater originality with respect to the Canons (where it is present). I think that it was available to the author of the Epitome, but that it was removed in the course of the development of the fiction.
29 Schermann attributed excessive significance to what he called the "X recension" of the two ways, and turned the genealogy of the tradition on its head
(Elfapostelmoral, 18-23, 80-88).
30 Wengst, Didache, 7-14, esp. 10-11.
122-23; and idem, "La Didache: Enseignement des douze Apotres?" Melto 3 (1967) 223-36. It is clear
(1) that the distribution of the material among the apostles is a subsequent stylization; it is also clear
(2) that this stylization must go back to a revision of the Two Ways tractate, on which both Canons and Epitome are dependent; the origins of the list of eleven aposdes (despite Giet's investigation) requires a new analysis; (3) that the text attributed to Bartholomew is different, because Epitome con
agree with Barnabas against the Didache. (These will be discussed below.)33 In my opinion, this phenome non suggests a different genealogical relationship among these writings. In sum, despite Wengst's objec tions, I wish to maintain the hypothesis that the Canons was not influenced by the Didache.
(4) In the Arabic Life of Shenoute of Atripe34 the intro duction to the speech of Visa (in Amelineau, 291, 11. 6ff.)
Did. 10.3 (end) Can. 12.3
Epit. 9 (end) EL yap 6 Kvpiog O yap KvpLoq ijfiuae ae bdavrov 8l ' avTov r]fcccoev aoi bodrjvai i)plv be eyapLoa bodf/va l ■KvevpaTLKpv ■nvevpaTiKrjv TTVeVpaT IKT)V Tpo<pr)v Kal KOTOV Kai far)v aluvcov Tpo<pi)v rpoippv Kal KOTOV Kai farjv aluviov Kal farjv alwviov is followed by a didactic section having the character of a set of rules: in content, it is nothing but a (fairly late) version of the Two Ways tractate. We are no longer sur prised to discover that the sectio evangelica is missing.
Otherwise, the tractate reappears with numerous alter ations, including the 65og tou davarov (which, however,
This is the passage already mentioned above.
According to Wengst, the added Kal itotov in Canon 12 should suggest that the Canons is additionally dependent on the Didache (Didache, 11). But if the Didache is already supposed to be in use (which I, nota bene, do not believe), it is much more inviting to suppose that both versions (Canons and Epitome) used the Didache, or (better) that their common source used the Didache,31 and that Kal kotov in the Epitome has (for unknown reasons) been omitted. The two conditions for supposing independence of the Canons from the Didache, as proposed by Wengst (Didache, is treated very cursorily).35 It is interesting that the shap ing of the material into teknon sayings does not begin with the section corresponding to Did. 3.1-6, but already in that corresponding to Did. 2.6 (and extends past Did. 3.6). There can probably be no question of the author's knowing the whole Didache.36
(5) Finally, I should mention two texts with a mutual literary relationship: Ps.-Athanasius Syntagma doctrinae
(ed. Batiffol), and Fides patrum (ed. Batiffol; also PG 28.1637-43).37 As regards the Fides patrum, the quota tions or allusions to the Two Ways tractate begin only in the second part of the writing, from chap. 3 (Batiffol), equivalent to PG 28.1639C. The literary relationship of the two writings to one another seems to be best under stood if we suppose that the Fides patrum (in its second part) draws either on the Syntagma doctrinae38 or on a source common to both.39 This whole question (includ11) are that the Didachist "reproduced his source word for word" and the Canons used "exactly the same source as the Didachist." In my opinion, these should be distinguished. The Didachist was fairly faithful in copying his source, but I do not believe that he used exactly the same source as did the Canons. What he had as a source was similar to that used by the Canons, because the relationship between the Didache and the Canons is not so exclusive as Wengst's reasoning implies.32 This is evident espe cially in the passages where the Canons (and Epitome)
36 H. Benigni attempted to derive the material from the Didache, but from a recension of the writing that would have been available to the Canons as well
("Didache copdca: 'Duarum viarum' recensio coptica monastica, Shenudii homiliis attributa, per arabicam versionem superstes," Bessarione 3 [1898]
326). According to Vokes (Riddle, 72) the use of the Two Ways tractate stems from the Didache, with the author assimilating Did. 1.2 to Barnabas and the Gospels. Barnard ("Later History," 104-5) writes that Didache and the Canons may have been used in addition to the Two Ways tractate. "The 'Two Ways' known to Schnudi was probably the Greek Original of the Doctrina" (p. 105).
37 On these two writings, see above.
38 Cf. Connolly, "Use of the Didache," 156; Vokes, Riddle, 72; Barnard, "Later History," 105.
39 Thus Harnack, Apostellehre, 28. Pierre Batiffol writes 31 It is true that Wengst himself (Didache, 11) raises the objection that this could also be a tradition "from which the compiler acquired knowledge by other means as well." But for him this objection is not conclusive in light of the considerations that follow.
32 Wengst himself (Didache, 10 n. 24) mentions "four exceptions," where the Canons and Epitome agree against the Didache.
33 Wengst supposes (Didache, 10 n. 23) that the Canons used Barnabas, but does not wish to explain Can.
12.1 (dyaKpoeLc; aq Kopr]v 6<pi)aAp.ov oou; cf. Epit. 9 and Barn. 19.9) by dependence on Barnabas. In stead, he explains it as showing dependence on the Two Ways tractate, because according to Wengst the Canons is dependent on Barnabas, but Epitome is not.
34 See above.
35 Striking here is the absence of the material con tained in Did. 4.9-14.
Introduction 12.4. For Synt. doctr. 6.6 (cf. Fides patr. 10 [1642D]) as well, dependence on Did. 13.3 cannot be demonstrated with certainty.43 The result is that these two writings have a secure relationship only to the Two Ways sec tion.44 The corresponding items probably do not go back to the Didache,45 but to the independent Two Ways tractate in one of its forms.46 ing the relationship of the Coptic version of the Fides patrum to the other documents) requires a new investiga tion. Quotations from the Two Ways tractate are distrib uted throughout both documents:
Synt. doctr. 1.4 = Fides patr. 3 (1639C): cf. Did. 1.2.
Synt. doctr. 1.5 = Fides patr. 3 (1639C): cf. Did. 2.2.
Synt. doctr. 1.9 = Fides patr. 4 (1639D): cf. Did. 2.4; 3.5, 6.
Synt. doctr. 2.4 = Fides patr. 5 (1640A): cf. Did. 2.2.
Synt. doctr. 2.5 = Fides patr. 5 (1640A): cf. Did. 3.4.
Synt. doctr. 3.1 = Fides patr. 7 (1641B): cf. Did. 3.5.
Synt. doctr. 4.1 = Fides patr. 8 (1641D):40 cf. Did. 3.8.
Add to these Fides patr. 3 (1639C): cf. Did. 6.1 (lacking a parallel in Syntagma doctrinae). All these passages have their parallels only in the Two Ways tractate of the Didache,41 and quotations from the sectio evangelica are
(of course) absent.42 All other parallels, which would indicate knowledge of later passages in the Didache (and thus knowledge of the Didache itself) are questionable:
Synt. doctr. 2.10 (cf. Fides patr. 6 [1640B]) knows of fast ing on Wednesdays and Fridays, but it cannot be proved that the passage is borrowed from Did. 8.1. Synt. doctr.
3.2, eav rexvotq M ^XwaiV duayKa^ovrat npayptaTeveofioti (cf. Fides patr. 7 [164 IB]), can, but need not, stem from Did. 42.4. Synt. doctr. 3.10 (cf. Fides patr. 7 [1641C]) can scarcely come from Did. 13.3, no more than Synt. doctr 6.4 (cf. Fides patr. 10 [1642D]) can come from Did. b. Further Details on the Hypothesis of an Independent
"Ways" Tractate In the preceding section, we have discussed the relation ship between the Two Ways portion of the Didache and similar passages in Barnabas, Doctrina apostolorum, the Canons of the Apostles, Epitome, the Life of Shenoute,
Syntagma doctrinae, and Fides patrum. We could not emerge from this without positing the idea of an inde pendent tractate de duabus viis, a tractate handed on in different versions and used as a source in various ways.
This model of interpretation (sometimes called the "basic document hypothesis") appeared early in the his tory of Didache research, was presented in the widest variety of versions, and dominates research today
(although the more precise genealogical relationships are differently described).47 I cannot enter into detail on these kinds of differences and nuances here. I will restrict myself to a description of what, in light of the history of research, appears to me most probable.
"Die Grundschrift der Didache und ihre Recensionen," ZNW 2 (1901) 59 n. 3, and passim; Muilenburg, Literary Relations, 39; Vokes, Riddle, 72-73
(with complicated demonstrations); Dix, "Didache,"
246-47; Barnard, "Later History," 105, suggested both dependence on the Duae viae and on the Didache and Canons.
44 Goodspeed, "Doctrina," 233.
45 Giet (L'Enigme, 146-47, 150) differs, and gives a complicated and quite arbitrary hypothesis.
46 Interesting is Synt. doctr. 1.6, a reminiscence of the James clauses; cf. Fides patr. 3 (1639C).
47 This model of interpretation stems from the very earliest period of research. Ehrhard (Altchristliche
Litteratur, 53) formulated it this way: "The first part, that the two versions of the Fides patrum (Greek and Coptic) go back, independently of one another, to the same source, from which the Syntagma doctrinae also derives ("Le Syntagma Doctrinae dit de Saint
Athanase," in idem, ed., Studia Patristica [Etudes d'ancienne litterature chretienne, fasc. 2; Paris:
Leroux, 1890] 135).
40 Synt. doctr. 4.1 and Fides patr. 8 (1641D) read (with Didache and against Barnabas, Canons, Epitome',
Doctrina apostolorum: omnia verba) dca ttarrog. One would strike this expression in Didache if it were not found in Syntagma doctrinae and Fides patrum. Const.
2.1.5 probably reads thus (it is uncertain whether the source, the Didascalia apostolorum, also had this reading); in 7.8.3 (from the Didache), 8loc -navroq is absent.
41 In fact, in Syntagma doctrinae only from the begin ning of the Two Ways tractate (analogous to Didache the description of the two ways, is Jewish in origin and may have been formulated even before Christ as a catechism for proselytes." (The description of the source as a catechism for proselytes is not, how ever, shared by all defenders of this hypothesis.) On p. 54, Ehrhard also names the earliest representa tives of this opinion. Its principal founder (after some precursors) was Taylor, "Early Jewish Manual
1-3).
42 Differently Dix ("Didache," 247), who sees Did. 1.5a quoted in Synt. doctr. 3.8-9.
43 Dependence of Syntagma doctrinae and Fides patrum on the Didache is advocated by Edgar Hennecke,