Didache

One Codex, Signed and Dated

In 1873 Philotheos Bryennios, then head of the Patriarchal Great School of the Nation in Constantinople and later Metropolitan of Serres and of Nicomedia, was working through the library of the Monastery of the Holy Sepulchre in the Phanar — the Jerusalem Patriarchate's dependency in the city — when he came across a Greek parchment codex. He published its two letters of Clement in 1875. The Didache he published in 1883, ten years after he first opened the book.

The manuscript is unusually well documented for its age, because the man who made it signed his work: a notary named Leon, who dated the subscription to 11 June 1056. It is 120 leaves, and the professional catalogue of Greek manuscripts lists seven works in it: a synopsis of Scripture attributed to John Chrysostom, the Epistle of Barnabas, 1 and 2 Clement, the Didache, the letters of Ignatius in their longer form, and a short piece on the genealogy of Christ through his legal father. The Didache occupies about six leaves in the middle. The codex is now in the Patriarchal Library of the Orthodox Patriarchate in Jerusalem as Panaghiou Taphou 54, and the Library of Congress microfilmed it: 123 public-domain images, freely viewable, of the only complete copy of this text that exists.

Almost everything else is scraps. Two small parchment fragments from Oxyrhynchus, catalogued P.Oxy. XV 1782 and dated to the late fourth century, carry parts of the first three chapters. A leaf in Coptic in the British Library carries part of chapters 10 to 12, and includes a thanksgiving over ointment that is in no Greek copy at all. There are Ethiopic and Latin pieces, and a Georgian version whose manuscript has since been lost and whose independence scholars dispute. For a document this important, the evidence is one book, dated 1056, plus fragments.

What a Church Actually Did

It opens with a catechism: "There are two ways, one of life and one of death, and there is a great difference between the two ways." Then it turns practical. Baptise in running water in the threefold name if you can; in other water if you cannot; in warm water if the cold is too much; and "if thou hast neither, then pour water on the head thrice in the name of the Father and of the Son and of the Holy Spirit." Both the person baptising and the person being baptised should fast first. Fast on Wednesdays and Fridays, not Mondays and Thursdays, because that is what the hypocrites do. Pray the Lord's Prayer three times a day — and note that the version given here closes with two clauses, "for Thine is the power and the glory for ever and ever," without the kingdom that later liturgy adds.

The eucharistic prayers come next, cup before bread, with the image of grain scattered on the hills and gathered into one loaf as a picture of the church, and the rule that only the baptised may share it. Then the part that gives the whole document its flavour: how to handle travelling apostles and prophets. Receive an apostle as you would the Lord — but he stays one night, two if he must, and "if he abide three days, he is a false prophet"; he leaves with bread and nothing else, and "if he ask money, he is a false prophet." A prophet who teaches the truth but does not do it is a false prophet. Genuine prophets, on the other hand, get the firstfruits, "for they are your chief-priests". Then a line of church order that is itself a dating argument: "Appoint for yourselves therefore bishops and deacons worthy of the Lord" — bishops and deacons, with no third order of presbyters between them. The last chapter turns to the end of the world, names a "world-deceiver" who will come as a son of God, and then simply stops: "Then shall the world see the Lord coming upon the clouds of heaven." There is no judgement scene. K. Niederwimmer and others hold that the original ending is lost.

This document is the single best answer to the claim that the early church suppressed books it did not like. It was not suppressed; it was read, valued, openly ranked, and openly excluded. Eusebius, around AD 325, lists it among books he does not treat as Scripture: "Among the rejected writings must be reckoned also the Acts of Paul, and the so-called Shepherd, and the Apocalypse of Peter, and in addition to these the extant epistle of Barnabas, and the so-called Teachings of the Apostles …" His word there is notha, and "rejected" reads harder in English than he meant it — he immediately adds that "all these may be reckoned among the disputed books", and he keeps them well clear of the heretics' forgeries, which he says are not to be placed even among these. Athanasius, in his festal letter of 367, is more explicit still, and this is the passage that settles the question: after listing the canon he adds that "there are other books besides these not indeed included in the Canon, but appointed by the Fathers to be read by those who newly join us", and names among them "that which is called the Teaching of the Apostles, and the Shepherd", in a list that also carries the Wisdom of Solomon, the Wisdom of Sirach, the book of Esther, Judith and Tobit. A named, approved middle tier: read in church, given to new converts, and not Scripture. That is a public filing system, not a bonfire.

Nobody Agrees When It Was Written

The proposals run from about AD 50 to well into the second century, and the range is not posturing — it follows from the fact that the document has no author, no addressee, no place and no date in it, and shows every sign of having been assembled in layers. A. Milavec argues for AD 50 to 70 and, unusually, treats the work as a deliberate unity rather than a patchwork. K. Niederwimmer's standard commentary places the final composition in the early second century, about AD 110 to 120, and analyses it into four earlier strata: a Jewish Two Ways tract, archaic liturgical material, a tradition about receiving travelling charismatics, and an apocalypse whose ending is gone. M. W. Holmes states the middle position: compiled in its present form perhaps as late as 150, though a date near the end of the first century seems likelier to many. This article has not opened Milavec's, Niederwimmer's or Holmes's volumes and reports all three positions at second hand, which is worth saying plainly rather than dressing paraphrase as quotation.

The arguments for an early date are internal and cumulative, and a reader can check most of them in the text: prophets are still itinerant and still a live pastoral problem; the ministry is bishops and deacons with no monarchical bishop above them; no heresy is named and no persecution is mentioned; the Christology is undeveloped. The arguments for a later date turn mostly on the relationship to Matthew, and that is the second real dispute. Roughly three camps: that the Didache draws on Matthew in its finished form, argued among others by C. M. Tuckett, whose case is that parts of it presuppose the editorial work of the evangelists themselves and not merely the traditions behind them — his articles are behind paywalls this article could not open, so his position is reported and not quoted; that both draw independently on a shared body of oral teaching, the position of J. A. Draper, J. S. Kloppenborg, Niederwimmer and others; and, from A. Garrow, that the dependence runs the other way for most of the book. Both of the first two camps concede the complication that makes the question hard: if the Didache was assembled in layers, the answer may differ layer by layer.

One thing the Dead Sea Scrolls did settle. The old three-cornered question — whether the Didache took the Two Ways from the Epistle of Barnabas, or the epistle from the Didache, or both from something older — has largely resolved in favour of a common Jewish source, because the Community Rule from Qumran turned out to contain the same two-ways teaching in a pre-Christian Jewish community. That is a good illustration of what this whole section is for. Two of the three positions people argued for a century were wrong, and the thing that settled it was a manuscript nobody knew existed.

A last note, and an honest one for an app written from an evangelical position: the pouring provision in chapter seven is a passage where evangelicals themselves genuinely differ, and have for a long time. The nineteenth-century editors of the Ante-Nicene Fathers said so in their own footnote, observing that the preceding verses point to immersion while this one permits pouring in certain cases. Neither side of that argument invented the sentence, and neither side is embarrassed by it as much as the other side hopes.

Scripture references

• Matthew 28:19-20 — The threefold baptismal name the Didache prescribes
• Matthew 6:9-13 — The Lord's Prayer as Matthew gives it; the Didache's version drops the kingdom from the doxology and orders it three times a day
• Matthew 7:15-20 — Testing false prophets by their fruit — the Didache turns this into a house rule
• Acts 2:42 — Teaching, fellowship, breaking of bread and prayer — the same four things, arranged as a manual

Sources

• Holding institutionPinakes (IRHT-CNRS): Jerusalem, Patriarchikê bibliothêkê, Panaghiou Taphou 054 — Institut de recherche et d'histoire des textes, CNRS · The standard professional catalogue of Greek manuscripts, and the source preferred here over the popular accounts. It gives the shelfmark, the date 1056, the scribe as Leon notarios, the note that the subscription follows on folio 120r, and the folio-by-folio contents list from which this article's account of the codex is taken, including the last item, which Pinakes titles the genealogy of Joseph, the betrothed of the Blessed Virgin Mary, and which this article renders as the genealogy of Christ through his legal father because the app's auto-linker cannot be given a bare "Joseph" without sending the reader to the patriarch. Popular sources and the current Wikipedia article disagree with it on the scribe's name and on whether the subscription is on 120r or 120v; Pinakes is followed hereSupports: The shelfmark, the date, the scribe, and the contents of the codex
• Holding institutionLibrary of Congress: Panagios Taphos 54, in Manuscripts in the Libraries of the Greek and Armenian Patriarchates in Jerusalem — Library of Congress · The catalogue record was read through the Library's own JSON interface, which confirms the title "Panagios Taphos 54. John Chrysostom: Synopsis of the Testaments; Apostolic Fathers. 1056 A.D. 120 f.", the date 1056, the note "Copied on vellum", 123 digitised images, and a rights statement placing the collection in the public domain with a requested credit line. The digitisation is of black-and-white microfilm, not colour. Note that the human-facing page refuses automated requests and has to be opened in a browserSupports: That a free facsimile of the only complete copy exists, and its extent
• Primary textThe Didache, translated by J. B. Lightfoot — Joseph Barber Lightfoot · Public domain, hosted at Wikisource under a pre-1931 tag, and read in full. Every phrase quoted from the Didache in this article was compared against it: the two ways, the pouring provision, the Wednesday and Friday fast, the two-clause doxology, the cup-before-bread eucharistic prayers, the three-day and the money tests for a false prophet, the firstfruits to the prophets, the appointment of bishops and deacons, and the abrupt endingSupports: Every quotation from the text of the Didache
• Primary textAthanasius, Festal Letter 39 (Nicene and Post-Nicene Fathers, series II, vol. IV) — Athanasius of Alexandria, in the NPNF translation · Public domain at Wikisource, read in full. The surviving fragment is printed there as chapter 25, not 39, which is why searching for chapter 39 fails. Source of the quotation naming the Teaching of the Apostles among books "not indeed included in the Canon, but appointed by the Fathers to be read by those who newly join us". The NPNF editors' own footnote raises a transmission question about this section, since it survives through the Coptic Life of Theodore; that caveat is theirs and is noted here rather than suppressedSupports: That the Didache was openly classified as read-but-not-canonical
• Primary textEusebius, Church History 3.25 (Nicene and Post-Nicene Fathers, series II, vol. I) — Eusebius of Caesarea, translated by A. C. McGiffert · Public domain at Wikisource, read in full, including the paragraphs either side of the one quoted — which is why this article flags that McGiffert's "rejected writings" renders notha more harshly than Eusebius's own next sentence supportsSupports: Eusebius's classification, and the qualification he immediately adds to it
• Holding institutionP.Oxy. XV 1782: Didache i-iii, University of Oxford — The Oxyrhynchus Papyri, University of Oxford · The holding institution's own record, DOI 10.25446/oxford.21132967.v2. It gives the material as parchment rather than papyrus despite the P.Oxy. series name, the format as codex, the date as late fourth century, and A. S. Hunt as editor. Its images are marked In Copyright, so they are linked here and not reproducedSupports: The fourth-century fragment, its material and its date
• ScholarshipK. Niederwimmer, The Didache: A Commentary; and A. Milavec, The Didache: Faith, Hope, and Life of the Earliest Christian Communities, 50-70 C.E. — Kurt Niederwimmer; Aaron Milavec · Niederwimmer, translated by L. M. Maloney, edited by H. W. Attridge, Hermeneia (Minneapolis: Fortress, 1998) — the standard commentary, source of the early-second-century dating, of the four-layer analysis, and of the position that the original ending is lost. Milavec (New York: Newman Press, 2003) is the sustained case for AD 50-70 and for reading the work as a unity rather than a patchwork. Print only, and this article has opened neither: both positions are reported at second hand, which is why the third section says so in its own words rather than quoting either manSupports: The two ends of the dating range, named to the scholarship that argues them
• ScholarshipH. van de Sandt and D. Flusser, The Didache: Its Jewish Sources and its Place in Early Judaism and Christianity; and J. A. Draper, ed., The Didache in Modern Research — Huub van de Sandt and David Flusser; Jonathan A. Draper · Van de Sandt and Flusser, Compendia Rerum Iudaicarum ad Novum Testamentum III/5 (Assen: Van Gorcum / Minneapolis: Fortress, 2002) is the standard reconstruction of the Jewish Two Ways source behind both the Didache and the Epistle of Barnabas, and is the scholarship behind this article’s third paragraph on the Community Rule. Draper’s collection, Arbeiten zur Geschichte des antiken Judentums und des Urchristentums 37 (Leiden: Brill, 1996), gathers the Didache-and-Matthew debate, including the oral-tradition position associated with Draper, J. S. Kloppenborg and Niederwimmer against the literary-dependence position of C. M. Tuckett and the reverse-dependence proposal of A. Garrow. Both print only; neither opened here. They are cited so that the named dispute points at the actual scholarship rather than at nothingSupports: The Two Ways source, and the three camps on the relationship to Matthew
• EncyclopediaWikipedia: Didache

Further reading

• Pinakes (IRHT-CNRS): Jerusalem, Patriarchikê bibliothêkê, Panaghiou Taphou 054
• Library of Congress: Panagios Taphos 54 (digitised)
• Didache, Lightfoot translation (public domain, Wikisource)