Internal Evidence

on John 8:1-11 (1917)

Exerpted from:H.J. Cadbury, A Possible Case of Lukan Authorship (Jn 7:53-8:11), 1917
Harvard Theological Review, Vol. 10, No.3 pg 237-244
for purposes of review and discussion

Page Index

Last Updated: Feb 21, 2009

Section 1: - Introduction
Section 2: - Cadbury on Lukan Style of Jn 8:1-11

Return to Index


The attempt to remove two large passages of the NT (Mark's Ending and Jn 8:1-11) hit the mainstream public in the 1880's, through the release of the 'Revised Version'. , the Christian public, this time a surprisingly literate and well educated public, demanded to know more about the evidences behind such drastic changes.

The academics were in fact caught with their pants down, as the two basic branches of research regarding authorship and text of John were completely at odds with one another as to the origin of John 8:1-11.

On the one hand, the textual evidence such as it was, and as understood at that time, made the Pericope de Adultera (John 7:53 -8:11) appear to be a later unauthorized addition to the Gospel of John.

On the other hand, the internal evidence as gathered and presented by Samuel Davidson made the passage look like a misplaced part of the Gospel of Luke. This was in part suggested by the unusual group of manuscripts called 'Family 13', which actually placed the passage after Luke 21:38.

Prof. Blass had previously suggested two editions of Luke (both issued by Luke the Evangelist), the second one issued later from Rome. This was an attempt to explain why some manuscripts (the majority) did not have the passage in Luke, but a few (Family 13) did.

The trouble was, the very early manuscripts that omitted the passage from John (Codex Aleph and B), also omitted the passage from Luke as well. Thus if the passage originated in Luke, the most ancient witnesses were wrong about the passage at least once, that is in Luke.

But if the early manuscripts Codex Vaticanus (B) and Sinaiticus (Aleph) were wrong in such a big way about Luke, (having incorrectly omitted the passage there), why could they not be equally wrong about John? The credibility of the oldest manuscripts was starting to crack under such analysis.

Cadbury reopens the case for Lukan authorship, and in doing so discusses both the circumstantial weaknesses and strengths of the case, and also the inherent weaknesses of any argument for Lukan authorship.

Cadbury himself appears to prefer the internal (stylistic) evidence to the textual evidence, but acknowledges some of the serious difficulties of such a position.

The papyrii (P66 and P75) and their significance to the question had not really come to light in 1917, and Cadbury's discussion is limited to the textual evidence prior to the publication of the papyrii, namely, the exhaustive collations and analysis of von Soden (1910-1912).

Samuel Davidson (1848) had published the first rather detailed examination of the internal evidence (i.e., vocabulary and style) of the passage, but some of Davidson's case was recognized as weak even by Tregelles (1860) who rejected the verses as a Johannine interpolation on textual grounds.

Cadbury brings the discussion of the style and diction of the passage up to date and puts it on a firmer footing with a more balanced presentation of evidences both for and against 'Lukan authorship'.

After acknowledging the weakness of the "Lukan" argument, Cadbury realises he has painted himself into a corner by rejecting Johannine authorship. All other options lead to a kind of apologetic agnosticism. The modern reader should be able to read between the lines here...

Return to Index

Cadbury on John 8:1-11

A Possible Case of Lukan Authorship

Henry J. Cadbury

In his Philology of the Gospels 1 Professor Blass referred somewhat casually to the Lukan style of the pericope adulterae. His theory of a Roman edition of Luke's works issued by the author himself, in connection with which his reference was made, has not received very wide acceptance, and so the linguistic phenomena to which he called attention were not made generally known. The motives of the present writer in bringing the subject forward again are not merely that the Lukan style of this passage impressed itself independantly upon him, as it might upon anyone familiar with Luke's style, but because von Soden's careful study of the text of the passage, 2 and Harnack's recent use of the style of the Lukan writings 3 make it desirable to give a fresh presentation of the evidence.

The Pericope Adulterae

With regard to the text of the pericope adulterae it must at once be confessed that it is one of the most uncertain passages in the whole N.T. The variants are extremely numerous, and as the section is entirely omitted by most of the great uncial MSS, a decision on readings cannot be made by the usual methods of valuation. Von Soden divided the authorities [MSS] into seven main groups, and attempted to appraise them and arrange them and so to recover the original text. To many however his reconstruction will not seem convincing. It will be safer therefore for us in considering the style of the passage to limit ourselves to no one form of the text but to include all variants, remembering constantly that some of the examples given are probably not the original readings. 4

The Distinct Style of Luke

The style of Luke, on the other hand, has become better known with the study of his writings. He has the most distinctive vocabulary of any NT writer, and a style so individual as to be recognizable in nearly every verse. Not matter what his subject or his source, these characteristics make themselves everywhere evident. Not merely in the nativity stories with their canticles at the beginning of his work, nor in the "we" passages at its close, may we find with Harnack abundant evidences of his style. Even the stories which he takes bodily from Mark are filled with his own peculiar ways of speech. 5 ; so homogenous is the style of the Lukan writings. It is therefore all the more striking that this brief passage - bearing as it does the evidence of antiquity and verisimilitude, yet certainly not part of the Fourth Gospel, as both its language and the MSS prove - should reveal nevertheless so many marks, some of them almost unmistakable, of Luke's style.

Evidence Against Lukan Authorship

First let us consider the negative evidence. There are of course some words in the passage that do not occur in Luke or Acts. They are:

  • αναμαρτητος
  • αυτοφωρω - αυτοφορω -αυτω τω φορω
  • διακελευω (μ1 )
  • καταγραφω (μ1 2 )
  • κατακυπτω - κατω κυπτω
  • κατηγορια 4(D) 6(μ2 3 4 6 7 ) a variant in Lk 6:7
  • μοιχεια (all MSS except D)

The first five of these are not found in any NT writer but are all compound words in the manner of Luke. Compare his use of απο-γραφω, ανα-κυπτω, συν-κυπτω, δια-τασσω , etc. 6

There are also some expressions which are less like Luke than like some other NT writer. The mention of the Jews in D and one or two other MSS, the use of "high priests and Pharisees" (μ1 ) in place of "scribes and Pharisees", are both variants that agree exactly with the manner of the Fourth Gospel. And the reading (μ4 5 ) μηκετι αμαρτανε without απο του νυν agrees exactly with John 5:14. For the reading εις καθ'εις (μ2 3 4 5 6 7 ) perhaps the nearest paralllel is in Mark 14:19 (εις κατα εις ).

Vocabulary Evidence For Lukan Authorship

Compare now with these possible linguistic affiliations to the other Gospels [above], the likeness of the passage to Luke-Acts [below:]. The following words or phrases occurring in this passage occur in Luke or Acts but in no other Gospel:

  • ανακυπτω
  • απο του νυν (μ1 2 3 6 7 )
  • εις εκαστος 8 ( μ5 6 ) 9 (μ1 )
  • επιμενω
  • εκπειραζω ( μ2 D ) 7
  • κατηγορος ( μ3 5 6 7 )
  • ορθρος - cf. ορθριζω, ορθρινος
  • προσποιεομαι ( μ5 7 )
  • συ ουν 8
  • συνειδησις ( μ5 7 )

More striking still is the list of words found in the pericope, which though not limited to Luke are more abundant in his work than in the other Gospels. From Hawkins' lists of Lukan phrases 9 we find in this passage:

  • αγω ( μ 1 2 3 5 6 7 )
  • ειπεν δε, ειπαν δε
  • ερωταω ( μ1 2 3 4 5 7 )
  • εχω , with infinitive ( μ5 )
  • λαος
  • νυν
  • πας , or πας ο λαος ( μ 1 3 5 6 7 )
  • παραγινομαι
  • πλην ( μ5 7 )
  • ων = 'when'

According to Harnack, 10 we are justified in marking as Lukan:

  • εις τον οικον ( μ2 3 5 6 7 )
  • εν μεσω
  • πορευομαι
  • αυτη η γυνη ( μ1 2 3 4 5 )
  • ως δε

Further examples may be gained from the Lexicon:

  • οι δε ακουσαντες ( μ2 3 4 5 6 7 ) (once in Mark)
  • αρχομαι απο (once in Matt.)
  • παραγινομαι εις ( μ1 2 3 5 7 ) (once in Matt.)
  • πορευου (twice in John)

Lukan Phraseology and Syntax Evidence

In the following cases there is a likeness of expression such as commonly exists between the different parts of Luke's writings:

John 7:53-8:11 and Lukan Parallels

Jn Ιησους δε επορευθη εις το ορος των Ελαιων
Lk. 22:39 και εχελθων οπορευθη κατα το εθος εις το ορος των Ελαιων
Jn ορθρου δε παλιν παρεγενετο εις το ιερον και πας ο λαος ηρχετο προς αυτον
Lk. 21:38 και πας ο λαος ωρθριζεν προς αυτον εν τω ιερω
Jn και στησαντες αυτην εν μεσω λεγουσιν αυτω
Ac. 4:7 και στησαντες αυτους εν μεσω επυνθανοντο
Jn ταυτεν ευρομεν επ' αυτωφωρω μοιχευομενην ( μ6 7 )
Lk. 23:2 τουτον ευραμεν διαστρεφοντα το εθνος ημων κ.τ.λ. ... cf. Ac.24:5
Jn ινα σχωσιν (εχωσιν, ευρωσιν) κατηγορειν (κατηγοριαν κατ')
Lk. 6:7 ινα ευρωσιν κατηγορειν (κατηγοριαν κατ') αυτου
Lk. 11:54( D al ) ινα ευρωσιν κατηγορησαι
Ac.28:19 εχων τι κατηγορειν
Jn ορθρου δε [βαθευς] παλιν ηλθεν ( μ6 )
Lk. 24:1 ορθρου βαθεως επι το μνημα ηλθαν
Jn και καθισας εδιδασκεν αυτους ( μ1 2 3 5 6 7 )
Lk. 5:3 καθισας δε [...]εδιδασκεν τους οχλους

Further examples of likeness are in construction and sentence structure:

With ( πρωτος βαλετω λιθον ) compare
Lk 2:2 ( αυτη απογραφη πρωτη εγενετο ) and other adverbial uses of the adjective in Luk 21:34, 24:18, 22, Act.20:6 (D), 28:13.

With the position of the pronoun in ( συ ουν τι λεγεις ) compare
Lk 16:7 συ δε ποσον οφειλεις ;
Ac.11:17 εγω τις ημην δυνατος κωλυσαι τον θεον ;
Ac.19:15 υμεις δε τινες εστε ;

With the use of the participle in (επεμενον ερωτωντες ) compare
Ac.12:16 επεμενεν κρουων
Lk.7:45 ου διελιπεν καταφιλουσα
Also Lk. 23:12, Acts 8:16

With ( κατελειφθη μονος ) compare
Lk. 10:40 μονην με κατελειπεν διακονειν

With the brief ουδεις in ουδεις ,κυριε of the woman's reply compare
Lk. 22:35 ... μη τινος υστηρησατε ; οι δε ειπαν, ουθενος

With κυριε in the same reply compare
Ac.10:14, 11:8 μηδαμως, κυριε
Lk.17:37 που, κυριε;
Ac.9:5, 22:8, 26:15 τις ει, κυριε;
Ac.10:4 τι εστιν, κυριε;

With πορευου απο του νυν μηκετι αμαρτανε compare
Lk. 5:10 με φοβου απο του νυν ανθρωπους εση ζωγρων

No other NT writing has such close parallels as those given from Luke and Acts.

Discussion of the Evidence

In view of the many misuses of the linguistic argument, especially in connection with Luke-Acts, it would be rash to assume at once from this evidence that the pericope adulterae is written by Luke. It is necessary to acknoweldge that there many limitations to the force of the examples given. First, few of them have unanimous textual support; second, many of them are not very unusual phrases in Greek literature. That no other NT writer uses a word is often an accident. But if NT standards are to be applied, there are a few unquestioned words that are really characteristic of Luke, as απο του νυν , αρχομαι απο , επιμενω , ειπεν δε, ως.

And while of course some of the variants must be rejected, any form of the text which we accept, even von Soden's, which is the shortest, will include more than half of our list of examples. It can safely be affirmed that the passage in its oldest form contained as much distinctively Lukan language as the average passage of equal brevity and simplicity in Luke's acknowledged works.

Against the theory of Lukan authorship the subject-matter and method of treatment offer no objection, but rather a confirmation. The third evangelist shows throughout a sympathy with women and with sinners that is congenial to this passage. Jesus' association with them is frequently criticised by the strict Pharisees in Luke. No further example is needed than the story of Simon and the sinner woman in Luke 7:36-50.

Textual evidence, however, does not encourage the hypothesis. As is well known, the best Greek MSS omit the passage entirely (Aleph, ABCLW et al.). It was known, however, in the West, as is shown by the Vulgate and perhaps some earlier Latin versions, by the references in Ambrose, Augustine, and Jerome, and by the early Greco-Latin codex D. Nearly all the authorities that contain or refer to it put it in chapter eight or at the end of the Gospel of John. The only exception is the Ferrar Group (Family 13) which places it after Luke 21:38. 11 There is therefore little textual reason to assign it to any of the canonical Gospels, and less for Luke than for John. 12

Dilemma Created by the Evidence

These facts bring us to a dilemma, either solution of which seems to contradict the current standards of NT criticism: either (1) the pericope adulterae is an original part of Luke's Gospel and was omitted without leaving any appreciable trace inthe MSS tradition of that Gospel, or (2) it is written by another than the third evangelist in a style that completely matches his own. 13

This paper aims not to solve the dilemma but to state it and to show its importance. For if the first solution is the correct one, then we must believe that in spite of their age, multiplicity, and agreement, our authorities for the NT text do not preclude such radical divergence from the autographs as the complete omission of a considerable section from one of the four Gospels. If this is possible, then certainly many of the most radical theories of interpolation and the most unsupported textual conjectures are also possible. Even radical scholars have often declared for the probable integrity of the best texts. Here, however, we should have a flagrant case of primitive tampering, for the omission could only be intentional. 14 And so our confidence in the transcriptional accuracy and in the doctrinal primitiveness of the earliest available text of the NT would be considerably shaken.

If, on the other hand, the passage is not from the pen of the auctor ad Theophilum, then someone, whether another author, a translator, or a scribe, intentionally or unintentionally, 15 wrote a style that is indistinguishable from the most distinctive of NT styles. In this case style proves to be a most unreliable criterion, and all critical arguments drawn from identity of style - such as the common authorship of John and 1 John, of Luke and Acts, of the Pauline letters, and even of the separate parts of a single work, - lose some of their weight. Especially such an argument as that often made concerning the Lukan style of the "we" passages must be reexamined in the light of this evidence. 16 For if in the pericope adulterae identity of style does not even prove final Lukan editing, it certainly cannot be used to prove in the "we" passages original Lukan authorship without sources. 17


Original Footnotes:

1. Blass, Philology of the Gospels, p. 159 (1898) with a reference to his edition of Luke, (1897), p. xlviii.

2. von Soden, Die Schriften des N.T. vol. I, pp. 486-524.

3. Especially Luke the Physician and The Date of the Acts.

4. All readings that are not found in all groups of MSS will be marked below with von Soden's symbols for the groups that contain them, e.g., μ1, μ2, etc. The numbers represent very nearly the order of preference given the groups by von Soden.

5. See Plummer, Luke, passim.

6. Plummer, Luke, p. 252: "Lk. is fond of compounds with dia-." There are over 50 words compounded with kata- which occur in Luke or Acts but not in Matt., Mark, or John.

7. The word occurs also in Mt. 4:7 (= Lk. 4:12) in a quote from Deut. 6:16.

8. According to Bruder only Lk. 4:7, 22:70; Act. 23:21.

9. Horae Synopticae, 2nd Ed. pp. 15-29

10. See Date of the Acts, pp. 5,6,9,15; Luke the Physician, pp. 40,50f.

11. Also Evangelistarium 435.

12. Of course its historicity is not dependant on its canonicity. Its internal character, agreeing as it does with the synoptic stories, bespeaks its genuineness as a tradition.

13. I omit as unlikely a third alternative - that it was part of a third (lost) work of the third evangelist. Blass's view that it was from a second edition of the Third Gospel issued by the author himself combines the difficulties of this view with those of (1) above.

14. The motive would probably be the fear that the section would be abused to condone looseness in sexual relations.

15. The decision between these alternatives and concerning the actual origin of the section if not from Luke forms a most interesting problem, but does not affect the implications of the main dilemma. Eus. H. E. III. 39:16 suggests two possible 2nd century sources. He says: "(Papias) relates another story of a woman, who was accused of many sins before the Lord, which is contained in the Gospel according to the Hebrews." The evidence of the story's western circulation and the variety of its readings may suggest that it was translated into Greek from the Latin. That the later scribes wrote a style like Luke's is not improbable. Blass, Evangelium secundum Lucam, 1897, pp. lviiff., has given some interesting cases from variants in Mark, and unless one accepts his hypothesis of two editions by Luke, his evidence for the Lukan style of the "Western" text of Luke and Acts (cf. his Professor Harnack und die Schritften des Lukas, 1907) will point in the same direction. That this "Lukanizing" is intentional is improbable. Perhaps the style of Luke was the most familiar to the scribes and probably it was the most congenial to them on account of its literary quality. Many of Luke's minor changes in Mark are made independantly by scribes of Mark, e.g., in D αγω for φερω.

16. The argument that the "we" passages are so distinctly Lukan in style that the author cannot be using a source is presented most fully by Harnack, Luke the Physician (1906), pp 40-120; Date of the Acts (1911), pp. 1-28; cf. also Hawkins, Horae Synopticae, 2nd ed. pp 182 ff. The inference of these scholars is that therefore Luke and Acts were written by a companion of Paul, presumably Luke.

17. Since the foregoing article was written there has come to hand H. McLachlan's St. Luke Evangelist and Historian (1912), with its full and independant argument for the Lukan authorship of the pericope adulterae (pp 94-126).