7
DETERMINING THE IDENTITY OF THE TEXT
By way of a framework for the following
discussion I shall use Burgon's seven "Notes of Truth." They are:
1. Antiquity, or Primitiveness;
2. Consent of Witnesses, or Number;
3. Variety of Evidence, or
Catholicity;
4. Respectability of Witnesses, or
Weight;
5. Continuity, or Unbroken
Tradition;
6. Evidence of the Entire Passage,
or Context;
7. Internal Considerations, or
Reasonableness.[1]
A reading, to be a serious candidate for
the original, should be old. If there is no attestation for a variant before
the middle ages it is unlikely to be genuine. A word of caution is required
here, however. Not only may age be demonstrated by a single early witness, but
also by the agreement of a number of later independent witnesses—their common
source would have to be a good deal older. Sturz has a good discussion of this
point.[2] But any reading that has wide late
attestation almost always has explicit early attestation as well.
To give a concrete definition to the idea
of "antiquity" I will take the year 400 A.D. as an arbitrary cut-off
point. Allowing only those witnesses who "spoke" before that date,
"antiquity" would include over seventy Fathers, Codices Aleph and B
and a number of fragmentary uncials, the early Papyri and the earliest
Versions. By way of specific illustration, ever since 1881 the word
"vinegar" in Matt. 27:34 has been despised as a "late,
Byzantine" reading—but what is the verdict of "antiquity"?
Against it are Codices Aleph and B, the Latin and Coptic versions, the
Apocryphal Acts of the Apostles, the Gospel of Nicodemus, and Macarius
Magnes—seven witnesses. In favor of it are the Gospel of Peter, Acta Philippi,
Barnabas, Irenaeus, Tertullian, Celsus, Origen, pseudo-Tatian, Athanasius,
Eusebius of Emesa, Theodore of Heraclea, Didymus, Gregory of Nyssa, Gregory
Nazianzus, Ephraem Syrus, Lactantius, Titus of Bostra and the Syriac
version—eighteen witnesses.[3] The witnesses for "vinegar" are
both older and more numerous than those for "wine."
Of course, age by itself is not enough. We have seen that most significant variants date to the second century. What we are after is the oldest reading, the original, and to judge between competing old readings we need other considerations.
A reading, to be a serious candidate for
the original, should be attested by a majority of the independent witnesses. Please
recall the discussion of weighing and counting given above. A reading attested
by only a few witnesses is unlikely to be genuine—the fewer the witnesses the
smaller the likelihood. Conversely, the greater the majority the more nearly
certain is the originality of the reading so attested. Wherever the text has
unanimous attestation the only reasonable conclusion is that it is certainly
original.[4]
Even Hort acknowledged the presumption
inherent in superior number. "A theoretical presumption indeed remains
that a majority of extant documents is more likely to represent a majority of
ancestral documents at each stage of transmission than vice versa."[5] The work of those who have done extensive
collating of MSS has tended to confirm this presumption. Thus Lake, Blake, and
New found only orphan children among the MSS they collated, and declared
further that there were almost no siblings—each MS is an "only
child."[6] This means they are independent witnesses,
in their own generation. In Burgon's words:
. . . hardly any have been copied from any of the rest. On the contrary,
they are discovered to differ among themselves in countless unimportant
particulars; and every here and there single copies exhibit idiosyncrasies
which are altogether startling and extraordinary. There has therefore demonstrably been no collusion—no
assimilation to an arbitrary standard,—no wholesale fraud. It is certain that
every one of them represents a MS., or a pedigree of MSS., older than itself;
and it is but fair to suppose that it exercises such representation with
tolerable accuracy.[7]
In accordance with good legal practice, it
is unfair to arbitrarily declare that the ancestors were not independent; some sort of
evidence must be produced. It has already been shown that Hort's
"genealogical evidence," with reference to MSS, is fictitious. But it
remains true that community of reading implies a common origin, unless it is
the type of mistake that several scribes might have made independently. What is
in view here is the common origin of individual readings, not of MSS, but where
several MSS share a large number of readings peculiar to themselves their claim
to independence is evidently compromised throughout. (The "Claremont
Profile Method"[8] gives promise of being an effective
instrument for plotting the relationship between MSS.)
However, there is one situation where
community of reading does not compromise independence. If the common origin of a reading is the
original, then the MSS that have it may not be disqualified; their claim to
independence remains unsullied. Of course, we do not know, at this stage in the
inquiry, which is the original reading, but some negative help is immediately
available. If one or more of the competing variants is an obvious mistake, then
those MSS which attest such variants are disqualified, at that one point
(recall that genealogy was supposed to be based upon community in error).
For the rest, the history of transmission
becomes an important factor, but to trace it with confidence we must take
account of at least two further considerations. In the meantime, the
independent status of MSS agreeing in readings that could be original must be
held in abeyance—there is not sufficient evidence to disqualify them, yet.
A reading, to be a serious candidate for
the original, should be attested by a wide variety of witnesses. By variety is
meant, in the first place, many geographical areas, but also different kinds of
witnesses—MSS, Fathers, Versions, and Lectionaries. The importance of
"variety" is well stated by Burgon.
Variety distinguishing witnesses massed together must needs constitute a
most powerful argument for believing such Evidence to be true. Witnesses of
different kinds; from different countries; speaking different
tongues:—witnesses who can never have met, and between whom it is incredible
that there should exist collusion of any kind:—such witnesses deserve to be
listened to most respectfully. Indeed, when witnesses of so varied a sort agree
in large numbers, they must needs be accounted worthy of even implicit
confidence. . . . Variety it is which imparts virtue to mere Number, prevents
the witness-box from being filled with packed deponents, ensures genuine
testimony. False witness is thus detected and condemned, because it agrees not
with the rest. Variety is the consent of independent witnesses, . . .
It is precisely this
consideration which constrains us to pay supreme attention to the combined
testimony of the Uncials and of the whole body of the Cursive Copies. They are
(a) dotted over at least 1000 years: (b) they evidently belong to so many
divers countries, —Greece, Constantinople, Asia Minor, Palestine, Syria,
Alexandria, and other parts of Africa, not to say Sicily, Southern Italy, Gaul,
England, and Ireland: (c) they exhibit so many strange characteristics and
peculiar sympathies: (d) they so clearly represent countless families of MSS.,
being in no single instance absolutely identical in their text, and certainly
not being copies of any other Codex in existence,—that their unanimous decision
I hold to be an absolutely irrefragable evidence of the Truth.[9]
Variety helps us evaluate the independence
of witnesses. If the witnesses which share a common reading come from only one
area, say Egypt, then their independence must be doubted. It seems quite
unreasonable to suppose that an original reading should survive in only one
limited locale. If the history of the transmission of the text was largely
normal, as I believe it was, then we must conclude that a reading found only in
one limited area cannot be original. It follows that witnesses supporting such
readings are disqualified, just like those supporting obvious mistakes—they are
not independent, at that point. They are disqualified as independent witnesses,
but their combined testimony still counts as one vote; their common ancestor is
still an independent witness.
As Burgon points out, it is variety that
lends validity to number, because variety implies independence. Conversely,
lack of variety implies dependence, which is why a reading that lacks variety
of attestation has little claim upon our confidence. It is an eloquent
testimony to the soporific effects of the W-H theory (with its
"genealogy") that subsequent scholarship has largely ignored the
factor of variety in attestation. There has been an occasional murmur of
disquiet,[10] but nothing approaching a recognition of
the true place that "variety" should have in the practice of New
Testament textual criticism. Burgon stated the obvious when he said:
Speaking generally, the consentient testimony of two, four, six, or more
witnesses, coming to us from widely sundered regions is weightier by far than
the same number of witnesses proceeding from one and the same locality, between
whom there probably exists some sort of sympathy, and possibly some degree of
collusion.[11]
Closely allied to
variety is the factor of continuity.
A reading, to be a serious candidate for
the original, should be attested throughout the ages of transmission, from
beginning to end. If the history of the transmission of the text was at all
normal, we would expect the original wording "to leave traces of its
existence and of its use all down the ages."[12] If a reading or tradition died out in the
fourth or fifth century we have the verdict of history against it. If a reading
has no attestation before the twelfth century, it is certainly a late
invention.[13]
Where there is variety there is almost
always continuity as well, but they are not identical considerations.
Continuity also helps us in evaluating the independence of witnesses. Readings
which form little eddies in the late "Byzantine" stream convict their
supporters of dependence at those points. Readings which enjoy both a wide
variety and continuity of attestation vindicate the independence of their
supporters. Apart from some objective demonstration to the contrary (such as
Hort claimed for "genealogy") it is not fair to reject the
independence of such witnesses. They must be allowed to vote. The punch line is
this: The majority of the extant MSS emerge as independent witnesses, in their
generation, and they must be counted until such a time as complete collations
permit an empiric grouping, like F. Wisse did in Luke 1, 10 and 20.[14]
Hort, followed by Zuntz and others,[15] rejected this consideration absolutely.
But the reader is now in some position to judge for himself. Since there was no
authoritative revision of the text in 300 A.D., or any other time, and since
the evidence indicates a reasonably normal history of transmission, how can the
validity of "continuity" as a "note of truth" be reasonably
denied? In my view, the factors of number, variety, and continuity form the
backbone of a sound methodology in textual criticism. They form a three-strand
rope, not easily broken. But there are several other considerations which are
helpful, on occasion and in their way.
Whereas the previous four
"notes" have centered on readings, this one centers on the witnesses.
Whereas the "notes" of number, variety, and continuity help us to
evaluate the independence of witnesses, this one is concerned with the
credibility of a witness judged by its own performance. “As to the Weight which
belongs to separate Copies, that must be determined mainly by watching their
evidence. If they go wrong continually, their character must be low. They are
governed in this respect by the rules which hold good in life.”[16]
The evidence offered above in the
discussion of the oldest MSS and of weighing versus counting must suffice to
illustrate both the importance and the applicability of this "note."
The oldest MSS can be objectively, statistically shown to be habitually wrong,
witnesses of very low character, therefore. Their respectability quotient
hovers near zero. Their great age only renders their behavior the more
reprehensible. (I am reminded of young King Henry's rebuke to Falstaff.)[17] In particular, I fail to see how anyone
can read Hoskier's Codex B and its Allies with attention and still
retain respect for B and Aleph as witnesses to the New Testament text—they have
been weighed and found wanting.
Since the modern critical and eclectic
texts are based precisely on B and Aleph and the other early MSS, blind guides
all, it is clear that modern scholars have severely ignored the consideration
of respectability, as an objective criterion. I submit that this "note of
truth" must be taken seriously; the result will be the complete overthrow
of the type of text currently in vogue.
The "context" spoken of here is
not what is usually understood by the word but is concerned with the behavior
of a given witness in the immediate vicinity of the problem being considered.
It is a specific and limited application of the previous "note."
As
regards the precise form of language employed, it will be found also a salutary
safeguard against error in every instance, to inspect with severe critical
exactness the entire context of the passage in dispute. If in certain Codexes
that context shall prove to be confessedly in a very corrupt state, then it
becomes even self-evident that those Codexes can only be admitted as witnesses
with considerable suspicion and reserve.[18]
An excellent illustration of the need for
this criterion is furnished by Codex D in the last three chapters of Luke—the
scene of Hort's famous "Western non-interpolations." After discussing
sixteen cases of omission (where W-H deleted material from the TR) in these
chapters, Burgon continues:
The sole authority for just half of the places above enumerated
[Luke 22:19-20; 24:3, 6, 9, 12, 36, 40, 52] is a single Greek codex,—and
that, the most depraved of all,—viz. Beza's D. It should further be stated that
the only allies discoverable for D are a few copies of the old Latin. . . .
When we reach down codex D from the shelf, we are reminded that, within the
space of the three chapters of S. Luke's Gospel now under consideration, there
are in all no less than 354 words omitted: of which, 250 are omitted by D
alone. May we have it explained to us why, of those 354 words, only 25 are
singled out by Drs. Westcott and Hort for permanent excision from the sacred
Text? Within the same compass, no less than 173 words have been added by
D to the commonly Received Text,—146, substituted,—243 transposed.
May we ask how it comes to pass that of those 562 words not one has been
promoted to their margin by the Revisionists?[19]
The focus here is upon Westcott and Hort.
According to their own judgment, codex D has omitted 329 words from the
genuine text of the last three chapters of Luke plus adding 173, substituting
146 and transposing 243. By their own admission the text of D here is in a
fantastically chaotic state, yet in eight places they omitted material from the
text on the sole authority of D! With the scribe on a wild omitting
spree, to say nothing of his other iniquities, how can any value be given to
the testimony of D in these chapters, much less prefer it above the united
voice of every other witness?!?!
This
Note of Truth has for its foundation the well-known law that mistakes have a
tendency to repeat themselves in the same or in other shapes. The carelessness,
or the vitiated atmosphere, that leads a copyist to misrepresent one word is sure
to lead him into error about another. The ill-ordered assiduity which prompted
one bad correction most probably did not rest there. And the errors committed
by a witness just before or just after the testimony which is being sifted was
given cannot but be held to be closely germane to the inquiry.[20]
Apart from the patent reasonableness of
Burgon's assertion, the studies of Colwell in P45, P66, and P75 have demonstrated it to be true. We have already
seen how Colwell was able, on the basis of the pattern of their mistakes, to
give a clear and different characterization to each of the three copyists.[21] Here again, this "note of
truth" seems to be completely ignored by current scholars. Why? Is its
validity not obvious?
This "note" has nothing to do
with the "internal evidence" of which we have heard so much. It is only rarely applicable and concerns
readings which are grammatically, logically, geographically, or scientifically
impossible. Burgon considered that pantwn, the
reading of B,D in Luke 19:37 is a grammatical impossibility; that kardiaiV, the reading of À,A,B,C,D, etc. in 2 Cor. 3:3 is a logical
impossibility; that ekaton
exhkonta, the reading of À,K,N,Q,P
in Luke 24:13 is a geographical impossibility; that eklipontoV, the reading of P75,À(B)C,L in Luke 23:45 is a scientific
impossibility (the Passover always coincides with a full moon, and a full moon
cannot eclipse the sun); and that autou, the
reading of À,B,D,L in Mark 6:22 is an historical impossibility
(it contradicts both Matthew and Josephus).[22]
I would offer oV, the reading of Aleph and three cursives in 1 Tim. 3:16 as a fine
example of a grammatical impossibility—it is a nominative relative pronoun with
no antecedent in the context; I regard the claim that it came from a primitive
hymn to be gratuitous, a desperate effort to save an obviously bad
reading. In the following section there
will be some further examples.
Although Burgon apparently limited the use
of this "note" to readings that he considered to be virtually
impossible, I will expand it in the direction of what is normally understood by
"reasonableness", namely the requirements of the context, which I
consider to be an important consideration. A variant that is at odds with the
context is suspect.
The first edition of this book was
criticized because it contained no examples to show how these principles apply
to specific cases. The first revision included appendices D and E, which
alleviated the criticism somewhat. In the intervening years my thinking on this
subject has matured considerably, in part because of significant research that
has become available in that interim, so I now propose to discuss some specific
examples—they each offer some difficulty that has theoretical implications, and
these will be discussed. One fundamental question for Majority Text theory is
this: "Is there a ceiling above which a reading may be considered 'safe'
or secure; that is, beyond reasonable challenge?" Personally, I have
tended to regard 80% as such a ceiling; I believe others would settle for 70%.
But what do we do if the attestation falls below 70% of the MSS, or below 60%,
or below 50%? I believe we must agree with Burgon that "majority"
cannot be the only criterion.
According to the International Greek New
Testament Project for Luke, some 60% of the Greek MSS insert tou Iwram between "Aram" and "Hesron".[23] But, out of 27 extant uncials only nine
include "Joram"; 18 do not and they are supported by the three
earliest Versions. ("Joram" was possibly an early corruption of Aram
[as per the ancestor of MS 1542] that was subsequently conflated with it; the
conflation survives in a large segment of the "Byzantine" tradition,
which is seriously divided here.)
"Joram" has a clear majority
attestation, albeit a weak one. However, the earliest MS to include it is from
the 8th century; all earlier MSS lack it. In terms of Burgon's "Notes of
Truth," Joram wins in "Number" but loses in
"Antiquity," "Variety" and "Continuity". I believe
Burgon would agree that "Joram" should be regarded as an
interpolation.
The Institut für neutestamentliche
Textforschung in Münster, Germany has published an almost complete
collation of the available MSS for selected variant sets in Acts. This permits a different statement of evidence
than one usually sees—this and the following examples from Acts are based on
that source.[24]
The evidence looks like this:
1) mellonteV --{TR}f18,lat,syr,sa 160MSS = 33.1%
2) mellonta --{HF} 130
" = 26.9%
3) mellontwn -- (3)82
" = 17.6%
4) mellwn --P74AB,bo 45 "
= 9.3%
5) mellon --{NU}À 36
" = 7.5%
6) mellontaV -- (1)25 " = 5.4%
(one other) --
1 “ =
.2%
483
Rather a dismaying picture—what to do? To
begin, the variants are all participial forms of the same verb. The key seems
to be the perceived referent or antecedent of the participle. Is it "the
Jews", "the Sanhedrin" or "the commander"? The best
answer from the point of view of the grammar is evidently "the Jews",
which would require a masculine, nominative, plural form—the only candidate is
variant 1). However, there were those
who took the referent to be "the Sanhedrin"—the Alexandrian MSS have sunedrion next to the participle, separated only by wV. The grammar requires a neuter, accusative, singular
form—variant 5). But, the Sanhedrin was made up of men, so perhaps some decided
it would be more appropriate to make it plural—variant 2); and maybe even
masculine besides—variant 6). Variant 3), being genitive, is really strange,
unless somehow someone thought that the commander intended to inquire of the
Sanhedrin, viewed as plural. Variant 4) presumably takes "the
commander" as the referent, but puts the form in the nominative, sort of ad
sensum since se is accusative. But variant 2) could also be referring
to the commander, precisely masculine, accusative, singular.
What are the requirements of the context?
"The commander" as referent does not fit. Not only was it not his idea, he sent Paul away that very night
to forestall the possibility. (That the Jews should attempt to tell the
commander what was in his mind is scarcely credible.) "The
Sanhedrin" as referent really doesn't fit either. to sunedrion appears in the text as the object of a
preposition, not as an initiating agent. It is "the Jews" that is
the Subject of the main verb, and therefore of the two infinitives, and our
participle is working with the second infinitive, "as ones intending to
inquire."
Conclusion: variant 1) is the only one
that really fits the context; it is also the best attested. Although it only musters 33.1% of the vote
(including f18), it
is also attested by the three ancient Versions—always weighty testimony.
Although the Majority Text is usually
attested by over 95% of the MSS, every so often we get an unpleasant surprise
where there is no majority reading at all. This case is as badly split as any I
have seen. And yet, our "notes of
truth" permit us to reach a convincing conclusion. "Number"
fails us, but "Antiquity", "Variety" and
"Continuity" do not. Although variants 4) and 5) are both ancient, so
is 1), and it wins in "Variety" and "Continuity"; it also
wins in "Reasonableness". So, I am cheerfully satisfied that mellonteV is the original reading.
The evidence looks like this:
1) oi peri ton paulon hlqon
--{TR,HF} (1)218MSS
= 46.3%
2) -- -- -- -- hlqomen --{NU}P74ÀA(B)Cf18,lat,syr,cop (4)180
" = 38.9%
3) oi peri ton paulon hlqomen -- (1)62
" = 13.3%
4) oi apostoloi apo turou hlqon -- [3](1)1 " = 1.1%
(one other reading) -- 2 " =
.4%
473
Variant 3) would appear to be a not very
felicitous conflation. Variant 2) best fits the context—since the beginning of
the chapter, and before, the main participants have been presented in the first
person plural. The closest finite verb on each side of the variant in question
is emeinamen, 1st plural. The information in variant
1) is unnecessary but not objectionable; if variant 1) were original there
would be no need to change it. Of course, if variant 2) were original there
would be no need to change it either, unless some felt it was time to remind
the reader who "we" was referring to. More likely it was the
influence of the Lectionaries, since they have precisely variant 1). Since the
MSS are quite evenly divided, the agreement of all three of the ancient
versions makes variant 2) the better attested. (Again f18 agrees with an ancient tradition.)
Once again we do not have a majority
reading, though the split is not quite so bad as in the prior case.
"Antiquity" and "Variety" are clearly with variant 2), and
so "Continuity" is presumably more with 2) than with 1), also. I
conclude that variant 2) has the best claim to be printed in the text.
The evidence looks like this:
1) de ek thV sunagwghV twn ioudaiwn --{TR,HF}f18
[1](1)285MSS = 60.2%
2) de autwn --{NU}P74ÀABCD,lat,syr,cop (1)77
" = 16.4%
3) de autwn ek thV sunagwghV twn ioudaiwn -- [13]98 " = 23.3%
(one other reading) -- 1
" = .2% 477
I believe this variant set must be
considered along with the presence of ta eqnh after parekaloun, but Aland's group did not include the
second set. However, from UBS3 it appears that virtually the same roster of
witnesses, including the three ancient versions (!), read variant 2) and
omit "the Gentiles". Where then is the Subject of the main verb parekaloun? Presumably for those witnesses it would be the Jews
and proselytes who had just heard Paul and wanted to hear it all over again the
next Sabbath. So why are they (Jews and proselytes) mentioned overtly again in
verse 43? And on what basis would the whole city show up the next week (v. 44)?
But to go back to verse 42, why would the first hearers want to hear the same
thing (ta rhmata tauta) again anyway? The really interested ones
stuck with Paul and Barnabas to learn more (v. 43), just as we would expect.
The witnesses to variants 1) and 3) join
in support of "the Gentiles", giving us a strong majority (over 80%).
So the Subject of parekaloun is ta eqnh—they
want a chance to hear the Gospel too, and the whole city turns out. It fits the
context perfectly. So, variant 3) appears to be a conflation and the basic
reading is variant 1). [If variant 3) is viewed as the original, variant 2)
could be the result of homoioteleuton, but not variant 1).] The witnesses to
variant 3), because they have "the Gentiles", are really on the side
of variant 1), not 2), so presumably 1) may be viewed as having 80% attestation.
For the witnesses to variant 1) the antecedent or referent of exiontwn must be Paul's group, since the Gentiles would
presumably address their request to the teacher.
In variant 2) autwn presumably serves as
Subject of both the participle and the main verb, but in that event the main
verb should take precedence and the pronoun should be nominative, not
genitive. However one might explain the
motivation for such a change—from 1) to 2) and deleting "the
Gentiles"—variant 2) is evidently wrong, even though attested by the three
ancient versions (which troubles me).
Perhaps someone faced with variant 1) took "of the Jews" to be
the referent of the participle instead of modifying "synagogue" (like
NKJV), and thought it should be Subject of the main verb as well—then, of
course, "the Gentiles" were in the way and were deleted. Then 1)
might have been shortened to 2) for "clarity".
This time we do have a majority reading,
although not as strong as we could wish. "Antiquity" and
"Variety" are with variant 2), although f18 confers "Antiquity" on variant
1) as well and therefore 1) wins in "Continuity". But,
"Context" (the performance of the MSS in the near context) comes into
play this time—it clearly favors variant 1), as does "Reasonableness"—it
enables us to say that the attestation for 3) really goes with 1), not 2), so
1) comes out with over 80%. In short, variant 1) has "Number",
"Continuity", "Context", "Reasonableness" and
"Antiquity"; variant 2) has "Antiquity" and
"Variety". I take it that the original text had: exiontwn de ek thV sunagwghV twn ioudaiwn
parekaloun ta eqnh, etc.
The evidence looks like this:
1) (without the long addition) --{HF,NU}P74ÀABf18,latpt,cop
280MSS = 58.1%
2) - 31): kai kata ton hmeteron nomon hqelhsamen krinai
parelqwn de lusiaV o ciliarcoV meta pollhV biaV ek twn ceirwn hmwn aphgagen
keleusaV touV kathgorouV autou ercesqai epi se. The five principle variations hinge on
the three underlined words; they are:
2) krinai . . .
epi se --latpt,syr (6)42MSS = 10% [7
variants]
9) krinai . . .
epi sou --
(15)26 " = 8.5% [8 variants]
17)
krinai . . . proV se -- (22)8 "
= 6.2% [9 variants]
26)
krinein . . . epi sou
--
(2)18 " = 4.1% [3 variants]
29)
krinein . . . epi se
--{TR} (5)4 " = 1.9% [3 variants]
32)
replaces aphgagen with five words, plus two other changes:
krinai . . . epi sou-- (6)15 "
= 4.4% [6 variants]
38)
completely rewrites the material:
krinai . . .
proV se-- (2)11 " = 2.7% [3
variants]
(twelve further variants)-- 20 " = 4.1%
482
Variant 2) presumably has the best claim
to be the standard form of the addition:
krinai beats krinein, epi beats proV, se beats sou. It is also
attested by syr and latpt.
However, although some form of the addition commands 41.9% of the MSS, there
are no less than 51 variants!
What about the context? The addition makes
good sense, and it fits nicely. But, it is not really necessary; that
information Felix already knew. The text reads quite well without the addition
also. I conclude that the short form was judged to be abrupt or incomplete,
giving rise to the addition; presumably the Autograph did not contain it. Since
Tertullian was an orator he may well have actually said what is in the
addition, plus a good deal more besides, but did Luke write it?
The external evidence, though divided, is
adequate to resolve this case: 58.1%
against a severely fragmented 41.9%. The ancient versions, being divided, do
not help us much this time. Although 58% is not a strong majority, by any
means, still, the severe fragmentation of the 42% sort of leaves variant 1)
without a worthy opponent. Variant 1) wins in "Antiquity",
"Number", "Variety" and "Continuity", so I have
no doubt that it is original. [The reading of the TR, variant 29), really has
little to commend it.]
The evidence looks like this:
1) - - -
-- -- - - - - - - - - - --{HF,NU}P74ÀABf18,syp,bo 339MSS = 70.5%
2) edoxen de tw Sila epimeinai autou --{TR}(itpt)syh?,sa
[2](12)83
" = 20.2%
3) “ “ “
“ “ autoqi --
(3)33 " = 7.5%
4) " " " " " autouV --C(D,latpt) (4)2
" = 1.2%
(three other readings) -- 3 " =
.6%
481
UBS and H-F agree that variant 1) is
correct, and indeed verse 33 seems to require that Silas returned to Jerusalem;
"they were sent back . . . to the apostles", and "they"
refers to Judas and Silas. The "problem" is that in verse 40 Paul
chooses Silas to accompany him, so he had to be in Antioch, not Jerusalem.
Accordingly the longer reading was created to solve the "problem".
The "some days" of verse 36 could well have been a month or two. From
Antioch to Jerusalem would be a trip of some 400 miles. Silas had time to go to Jerusalem and get
back to Antioch.
"Reasonableness" makes itself
felt here; variant 2) introduces a contradiction, which the TR unfortunately
perpetuates. Variant 1) also wins in "Number" and
"Continuity". "Antiquity" and "Variety" are
divided. Thus, with a majority of 70.5% variant 1) is the best candidate for
the original reading.
This is the last example from Acts, and
one that I consider to be especially difficult (it has the potential to be
damaging). The evidence looks like this (I arbitrarily neglect margins and
correctors, except for the early uncials):
1) eiV Ierousalhm --{HF,NU}ÀB(f18=30mss)
281MSS = 59.7%
- - - - - - - - - - -
-
2) apo Ierousalhm --D(f18=6mss)lat(syh) 51
" = 10.8%
3) ex Ierousalhm --{TR}P74A bo (syh) 16 " =