Bryennios's discovery of the Didache
held out the tantalizing promise of new insight into little known aspects of
early Christian belief and practice. However, ongoing failure to establish its
location within the web of early Christian life and literature means that its
full potential has yet to be realized.
The aim of this volume has been to
secure one dimension of the Didache's original context. I chose to examine its
relationship to Matthew's Gospel because some form of connection with this
gospel is all but undeniable. Of course, there is a world of difference between
establishing 'some form of connection' and pinpointing a connection of a
particular type. In this case, however, two factors promised to overcome the
usual problem of perpetually reversible dependency arguments: Matthew's Gospel
and the Didache share numerous unique similarities and the Didache is a
distinctively composite creation. I use the term 'distinctively composite' to
convey the sense, not only that the Didache was laid down in layers, but also
that it contains previously existing elements that have maintained their prior
form. To use a crude illustration, the Didache may be likened to a fish pie
within which whole fish and substantial pieces of fish, of distinctly different
types, may still be distinguished, (e.g. the Two Ways tractate, the Sayings
Onion, the Teknon unit, the Prophet Document, the closing Apocalypse, and so
on).
The composite character of the
Didache promises to provide a means of establishing its relationship to
Matthew's Gospel because points of connection between the two texts occur
within so many of the separate elements incorporated into the Didache, as well
as within the editorial comments of the Didache's direct contributors. This
combination of factors is best explained, I propose, if Matthew took small
pieces from the variety of separate whole traditions within the Didache's
'pie', as well as from the editorial comment that surrounds them, and mixed
them, with related resources from elsewhere, to make a dish in which whole
distinct traditions are no longer apparent. Rather, Matthew's creation is one
in which elements from different resources are thoroughly mixed together to
create a unified whole.[1]
While accepting the limitations of
this analogy it succeeds in holding simultaneously in view the distinctively
composite nature of the Didache as well as the presence of multiple contacts
between the two texts. If my 'fish pie' analogy is even approximately correct,
then the two mainstream explanations for the texts' relationship face
considerable difficulties.
Take, to begin with, the idea that
the Didache consistently depends upon Matthew's Gospel. This proposal requires
the various creators of the whole traditions that ultimately became
incorporated into the Didache, separately and in each case, to refer to
fragments of Matthew's Gospel in the process of creating these whole and
self-contained traditions. Not only that but they must also be supposed to have
selected fragments that are often widely dispersed in Matthew's Gospel and
closely conflated with material from other sources. In addition to the activity
of these indirect contributors it is also necessary to propose a repeated
homage to Matthew's Gospel by editors who directly contributed to the Didache,
most particularly the editor of the third redactional layer.[2] As a result, therefore,
it is necessary to imagine a queue of indirect contributors, each with diverse
backgrounds interests and styles, lining up over a period of time to draw
inspiration from Matthew's Gospel. Their work must then be gathered by direct
contributors who have a similar surgeon's instinct for selecting from the
gospel material that is 'from Matthew's special material or from the synoptic
traditions at points where Matthew's distinctive rendering is preferred'.[3] This picture is
incredible. I conclude, therefore, that the first three layers of the Didache
cannot have depended upon Matthew's Gospel.
The majority of scholars in the
contemporary debate agree that the Didache did not directly depend on Matthew's
Gospel.[4] They propose, instead,
that both texts held a number of separate traditions in common. This approach
overcomes two important difficulties. First, evidence to suggest the antiquity
of the Didache's traditions relative to Matthew's Gospel need not be
over-ridden, since it is possible that the Didache preserves the sources it
shares with Matthew more faithfully than did the evangelist. Second, this
reading avoids the need for a row of direct and indirect contributors who each
picked out fragments of Matthew's special material from Matthew's unified text.
Instead, it may be imagined that Matthew and the Didachist both fished in the
same pool of early traditions and gathered a number in common. Matthew, it may
then be proposed, conflated these with other related material to create
substantial discourses as well as shorter speeches of Jesus, while the
Didachist preserved them more or less in their original state.
The shared use of common traditions
provides an initially plausible explanation for the presence of similar
material in both the gospel and the
Didache's whole incorporated traditions. Even here, however, high levels of
coincidence must be appealed to. The first of which is that Matthew and the
Didache must have had access to a range of diverse traditions that are
distinctively similar to one another in comparison with those preserved in
other early Christian documents. This implies, not only that they fished at the
same lake, but that they fished the same part of that lake at a similar time
(without drawing directly from one another's catch). Other factors suggest,
however, that this is unlikely to have been the case. The compositional
analysis in Part I noted that the Didache's Base layer appears to have been
written by a Jew to a Gentile audience and by someone with no identifiable
interest in the ministry of prophets. The Prophet layer, by contrast, is solely
concerned with regulating, protecting and making viable the prophetic ministry.
Then again, the Modifying Teacher layer appears to have been written to an
audience with a Jewish, rather than a Gentile, background. Furthermore, this
author felt unable to modify the base text directly but was forced to recogize
and accept its authoritative status; an authority likely to have been
established over a period of time. These factors combine to suggest that these
layers were contributed by different editors at different times and in
different contexts. This observation decreases the likelihood that Matthew and
the Didachists fished the same part of the lake at the same time, thus
increasing the coincidence required to explain their shared knowledge of
otherwise unknown forms of tradition. On the other hand, this difficulty does
not arise if Matthew knew and used the Didache. In this case there would be
nothing surprising about the reappearance in Matthew's Gospel of elements from
several of the Didache's diverse range of incorporated traditions.
The independent use of shared
traditions requires an even greater level of coincidence when it is observed
that these separate traditions are sometimes similarly juxtaposed by both
Matthew and the Didache. For example, the concept of two ways is juxtaposed
with the golden rule in Did. 1.1 (TW)[5] and 1.2e (LS); while a similar
combination appears in Mt. 7.13 and 7.12. Instructions to be meek and merciful
occur in Did. 3.7 (TK) and 3.8 (TW); and also in Mt. 5.5 and 5.7. A call for
perfection is linked with loving beyond those who love you in Did. 1.4b(r)
(P/B) and 1.3c (SO.b); and also in Mt. 5.46,47 and 5.48. Separate traditions
set alongside one another in the Didache also re-appear in similar
juxtaposition in Matthew's account of Jesus's response to the rich young ruler.
Thus, Matthew combines the idea of entering life (Mt. 19.17) with a keeping of
the commandments (Mt. 19.18), including love of neighbour (Mt. 19.19), to
attain the goal of moral perfection (Mt. 19.21). A uniquely similar arrangement
appears in the Didache's combination of the 'way of life' (Did. 1.1 - TW) with
a demand to love the neighbour (Did. 1.2d - LS) to attain moral perfection
(Did. 1.4b(r) - P/B) and keep the commandments (Did. 2.2 - TW). On the other
hand, complex coincidence need not be appealed to if Matthew knew and used the
Didache directly. In this case the reappearance of separate traditions in
similar juxtaposition is entirely to be expected.[6]
Although demanding high levels of
coincidence, an appeal to the independent use of shared traditions provides a
theoretically possible explanation for similarities to Matthew's Gospel that
occur within one of the Didache's
incorporated units. However, this explanation is very much more difficult
to apply when the point of contact occurs within the Didache's redactional
material. For example, a point of contact falls within Did. 11.1-2 (=Mt.
5.17-20) which is highly likely to be an original creation of the modifying
teacher (cf. sections 11.D. and 7.D.). There are also strong reasons for
supposing that Did. 1.5b is an original composition, crafted for this specific
location, by the modifying teacher. First, it only makes sense in response to
the preceding instructions regarding giving. Second, resort is made to an
obscure proverb, in Did. 1.6, to support the authority of 1.5b. This suggests
that 1.5b has no previous authoritative pedigree of its own and thus, that the
reappearance of a reference to paying back the last penny in Mt. 5.26 is
unlikely to be due to the shared use of a common tradition.[7] A very similar
circumstance occurs in Did. 14.2, where the modifying teacher calls readers to
be reconciled before making sacrifice. Support for this instruction is sought,
in Did. 14.3, from the tangentially relevant Mal. 1.11. This suggests that Did.
14.2 did not have a prior pedigree of its own to which appeal could be made, a
circumstances that sits at odds with Matthew's apparent knowledge of a saying
of the Lord that directly addresses the relevant subject (Mt. 5.24). If both
texts depended on a common source, then why does Matthew present it as a saying
of Jesus, while the modifier mentions no authoritative source at all? This problem
does not arise, however, if Matthew knew and used Did. 14.2. Under these
circumstances the original contribution of the modifying teacher could be
understood as a Lord's word because it stood within a text entitled 'The
Teaching of the Lord, etc.'[8]
In all, therefore, an explanation
for the several similarities between Matthew's Gospel and the Didache which
proposes their independent use of common resources must overcome the following
difficulties: the coincidental knowledge, at different times and circumstances,
of several forms of tradition that are not known in any other text; the similar
juxtaposition, on a number of occasions, of those separate traditions; and
Matthew's consistent attribution of all this material to Jesus, even though the
original creativity of one of the Didache's redactors is evident on at least
three occasions. On the other hand, if Matthew directly used the Didache, then
none of these factors presents a difficulty. Under these circumstances it would
be natural to expect redactional as well as traditional elements within the
Didache to reappear in the gospel, sometimes in similar combination. Further,
it would be unsurprising to find this material consistently treated as the
direct speech of Jesus, since, from Matthew's point of view, it all appears
under the title, 'The Teaching of the Lord, by the Twelve Apostles, to the
Gentiles'.
So far I have pointed out
complications facing the two mainstream explanations for the connections
between the Didache and Matthew's Gospel, while at the same time seeking to
show that these difficulties do not apply if Matthew depended directly on the
Didache. If this latter solution is so simple, it may reasonably be asked, then
why has it not been seriously considered in the past? I offered an initial
response to this question in chapter one, but must now return to the crucial
matter of the presence of four appeals to 'the gospel' in Did. 8.2b, 11.3b and
15.3-4.
These references to a 'gospel'
present an obvious obstacle to the view that Matthew directly depended upon the
Didache. If, as is highly likely,[9] these lines were written
with knowledge of Matthew's Gospel, then a significant qualification must added
to any theory of Matthew's use of the Didache; namely, that the references to
'the gospel' were inserted after the Didache's earlier layers had been made use
of by Matthew.
At first sight, this assignment of the
gospel references in Did 8.2b; 11.3b and 15.3-4 to a wholly separate
redactional layer may seem a rather lame attempt to sidestep evidence that is
hostile to my overall thesis.[10] It is not the case,
however, that these gospel references present a puzzle for my argument alone,
or that they present fewer difficulties for the alternative hypotheses. This
can be illustrated by the crucial example of the Lord's Prayer in Did. 8. What
must be wrestled with here is that Did. 8.2b instructs its readers to pray 'as
the Lord commanded in his gospel', but the immediately ensuing prayer is
unlikely to have appeared in any manuscript of Matthew's work.[11] Those who propose the
Didache's dependence on Matthew at this point must provide some explanation for
the absence of Matthew's version of the prayer. A related problem is posed for those
who see the Didache as compiled by someone who sometimes reveals a direct
knowledge of Matthew's Gospel, but who prefers to quote that gospel's sources,
rather than the gospel itself.[12] These scholars must
explain why Jesus's direct instructions, as recorded in Matthew's Gospel, are
set up as a standard in 8.2b and then set aside in 8.2c.
Since the combination of an appeal
to the gospel in 8.2b and its non-quotation in 8.2c presents a puzzle for any
theory that sees these two lines as composed by the same author, it may be
preferable to see them as belonging to two different redactional layers. Under
these circumstances the disjunction between 8.2b and 8.2c need not be due to editorial
incompetence, but may be explained in terms of a later contributor's respect
for the basic text. Thus, a later interpolator may have wished to point readers
to Matthew's version of the Lord's Prayer, while at the same time being
unwilling to make direct alterations to the established document. The insertion
of 8.2b achieves this goal, even though the resulting text is somewhat awkward.
A similar effect is created when the modifying teacher avoids direct revision
of the host, but nonetheless manages to alter its force by inserting new
material. Here again the reading that results is sometimes awkward and
self-contradictory.
As well as providing an explanation
for the disjunction between Did. 8.2b and 8.2c, two further factors favour the proposal
that 8.2b; 11.3b and 15.3-4 were inserted to subordinate the Didache to Matthew's
Gospel. First, such an understanding allows each of the four references to 'the
gospel' to perform a similar function; something which cannot be achieved if
Did. 8.2b is seen as a quotation formula, since it is not possible for Did.
15.3,4 to operate in this way. Second, this proposal presents a credible motive
for the inclusion of appeals to 'the gospel'. Matthew's proposed conflationary
practice results in divergences between his text and related teaching in the
Didache, which in turn creates a need for guidance as to which set of instructions
should be followed. Such guidance is provided by appeal to 'the gospel' at
those points where the differences between the two texts are most likely to
cause confusion (8.2b; 11.3b and 15.3) as well as a catch-all indication that
Matthew's Gospel should be seen as having the final word (15.4). By contrast, it
is difficult to detect a credible motive for the inclusion of references to
'the gospel of the Lord' by someone who also added the Lord's Prayer or, for
example, rules regarding visitors. In this case the interpolator's additional contributions
would be immediately undermined by their simultaneous recognition of the ultimate
authority of Matthew's divergent instructions.
On closer inspection, therefore, the
assignment of Did. 8.2b; 11.3b and 15.3,4 to a separate redactional layer provides
the only explanation for their inclusion that does not require inconsistent or
eccentric behaviour on the part of their contributor. This in turn means that,
while there are severe obstacles to the view that the Didache depended on
Matthew's Gospel, or that both texts independently used common traditions,
there is no barrier to the conclusion that Matthew knew and used the first
three layers of the Didache. Indeed, the evidence points overwhelmingly in this
direction. The preceding discussion identified twenty-six points of contact
between the two texts (listed at the end of chapter fifteen). Of this total six
fall within material attributed to the original creativity of the Didache's
redactors, while the remainder occur within eleven of the fifteen passages
identified as separate traditions incorporated into the Didache. Individual
analyses of each of these points of contact found positive evidence for Matthew's
knowledge of the Didache in twenty-two of the twenty-six instances, and found
no evidence to preclude Matthew's direct knowledge of the Didache in the other four.
Even if my specific identification of layers and incorporated traditions within
the Didache is inaccurate or incomplete, it remains undeniable both, that the
Didache has a distinctively composite character, and, that it contains widely
dispersed and unique points of contact with Matthew's Gospel. I conclude that the
only credible explanation for this combination of factors is Matthew's direct
dependence on those elements of the Didache that predate the interpolation of the
four appeals to external authority of 'the gospel'.
When Archbishop Bryennios stumbled
across a manuscript of the Didache he came to believe that he had made a sensational
find. However, ongoing failure to establish its location within the web of
early Christian life and literature has severely weakened the impact of his
discovery. The above case for Matthew's direct knowledge of the Didache will, I
hope, lead to a secure appraisal of the Didache's place within the life of the
first Christians.[13] Once this location has
been established, the frustrated potential of Bryennios's remarkable discovery
may finally come to fruition.[14]
| Home Cube | Book information | Purchase Information |
[1] Analogies, while useful
in some respects, are seldom capable of conveying the full complexity of the
case. The analogy of the unified fish pie is, strictly speaking, most
appropriate when applied to Matthew's teaching discourses, such as the Sermon
on the Mount and the Apocalyptic Discourse.
Most scholars recognize a combination of literary craftsmanship combined
with a use of prior sources in the construction of Jesus's speeches. Carson,
while arguing for the unity of the text, bears witness to a general recognition
that this unity was forged, largely successfully, by the combination of prior
sources. Carson (1984: 17) thus states, 'The question of the unity of Matthew's
Gospel has little to do with source-critical questions. Instead it deals with
how well the evangelist has integrated his material to form cohesive pericopes
and a coherent whole. In sections very difficult to interpret (e.g., Matt 24),
it is sometimes argued that the evangelist has sewn together diverse traditions
that by nature are incapable of genuine coherence. Failing to understand the
material, he simply passed it on without recognizing that some of his sources
were mutually incompatible.
There are so many signs of
high literary craftsmanship in this Gospel that such skepticism is unjustified.'
[2] Whether or not my analysis of the compositional history of the Didache is correct in detail, it remains the case that points of contact with Matthew's Gospel occur within parts of the Didache that may, with some confidence, be attributed to both direct and indirect contributors.
[3] Court (1981: 111)
[4] It is interesting to note that the line of approach used is not based on a compositional analysis. Instead, these scholars tend to prefer 'development' arguments. That is to say, evidence is found to suggest that Matthew's version of a particular saying is, for some reason or another, more developed that the similar saying found in the Didache. A small selection of these types of observation have been made in the preceding discussion. In most cases, however, I have favoured a compositional approach in the belief that this type of argument is less easily reversed than the tracing of supposed developments. Nonetheless, it is noteworthy that arguments in favour of the antiquity of the Didache, in relation to Matthew's Gospel, may consistently be made where they share similar material - even if, on occasion, such arguments are not the only possible explanation for the differences between the two texts. Scholars who detect ancient traditions in the Didache, relative to their counterparts in Matthew's Gospel include, amongst others, Köster (1957); Rordorf (1991); Kloppenborg (1979) and van de Sandt (2002).
[5] The two letter codes
used in this discussion refer to separate traditions identified as having been
incorporated into the Didache during the process of its composition (cf. Part
I). A key to these codes may be found in the introduction to the full text of
the Didache at the beginning of this volume.
[6] Even though an extraordinary set of coincidences is required, it is still theoretically possible that Matthew and the Didachists made independent use of several previously existing traditions, perhaps already clustered in groups. However, to account for the spread of points of contact between the two texts any such clusters would need to be so large as to be all but indistinguishable from the Didache itself.
[7] For a full discussion of this point see section 11.B.
[8] For a full discussion of this point see section 11.A. Additional points of contact that fall within direct contributions to the Didache (on the basis of the analysis presented in Part I) include: the long title, the call to be perfect (1.4b(r)), and the linking verses prior to the closing apocalypse (16.1-2).
[9] Strictly speaking, if
the 'gospel' referred to was an oral or lost gospel, then it would not follow
that these references must have been written after the creation of Matthew's
text. However, as discussed in section 8.A. it appears highly likely that
Matthew's Gospel is in view here.
[10] Full details of my reasoning for the separation of Did. 8.2b; 11.3b and 15.3-4 from the rest of the Didache's compositional history may be found in Part I, especially in sections 7.C,E,G, chapter eight, and section 11.C.
[11] See section 11.B.1.
[12] See the comments of Niederwimmer and Köster quoted in section 8.A.3.
[13] The hypothesis of
Matthew's use of the Didache opens up a number of new questions, or places old
questions in a new light. Among these are the nature of the Didache's
relationship to the several other texts of the New Testament with which it
shares points of contact. A very partial exploration of this question has been
undertaken above in relation to Luke's and Mark's Gospels and, to an even
lesser extent, with regard to 1 Corinthians and Revelation. If the Didache is a
pre-Matthean text, then studies of its relationship to other first century texts
should serve, not only to confirm this conclusion, but also to assist in the
more specific identification of the Didache's dates, authorships and provenances.
[14] If Matthew knew the Didache then a great many implications follow, some of which are noted in the preceding discussion. One implication treated in partial detail is that the Didache provides evidence for a new solution to the Synoptic Problem; namely, that Matthew used Mark's Gospel to provide the spine of his account and supplemented this with relevant material from the Didache, Luke's Gospel (possibly Revelation) and other sources. This solution places the Didache in a 'Q-like' position in that there is evidence for Luke's and Matthew's combined knowledge of the Base layer of the Didache. However, the Didache also provides evidence for Matthew's direct knowledge of Luke's Gospel, thus obviating the need for a more extensive 'Q' to explain the remainder of the double tradition. This provides an example of how a new perspective on the Didache opens up the possibility of new solutions to otherwise unsolved puzzles. The Didache's ability to answer this and other questions will provide its own means of further identifying and confirming its place amongst the documents of the first Christians.