Monday, January 25, 2010

'in his house' Hebrews 3.2 (incl P46)

27
This is a tricky variant, not necessarily because a great deal hinges on it, but because the issues seem fairly evenly weighted.

Firstly, on the manuscript evidence, it does seem pretty clear that P46 lacked OLW. There is some damage, but there is no space for it. Here is the bottom of the relevant page:






And here is a bit more detail of the text as E[N T]W OI[KW:









The combination of P46, P13, B and the Coptic suggests this was an influential early reading (certainly in Egypt, evidence beyond that is limited). Since this reading is not the reading of Numbers 12.7 (quoted in Heb 3.5 and anticipated here in 3.2); and yet makes perfect sense here at 3.2 I think there is good reason to prefer the shorter, unharmonised, earliest attested text at this point. What do you think?

27 comments

  1. I agree. It is more likely that a scribe or scribes harmonised rather than disharmonised the text, and hence the reading in P13 P46 B Coptic is likely the original.

    ReplyDelete
  2. Timo, what empirical basis is there for a general claim that it is more likely for (a) scribe(s) to harmonize than to disharmonize?

    Interestingly, neither in Codex Vaticanus nor in Codex Alexandrinus Heb 3:2 or 3:5 has been marked as an OT citation by the scribes/readers who added the diple markers to the margins of said mss. The sole intention of the diple markers was to highlight scriptural quotations. And the assumption is that the persons who added them where actively searching for scriptural quotations. If they did not spot Heb 3:2 nor 3:5, what are the odds for scribes to spot the scriptural basis on the fly (while transcribing, that is)?

    Moreover, how do we rate an anticipated harmonization as in this case?

    Finally, what about a simple mistake due to homoioteleuton ([OLW] TW OIKW)?

    ReplyDelete
  3. Is provenance of reading so unimportant that the earliest attested Greek reading from every location but one should be so easily neglected?

    Jonathan C. Borland

    ReplyDelete
  4. Ulrich,
    I'm not sure it counts as 'empirical', since it presumes a text-form to begin with, but there is a lot of evidence gathered in Alan H. Cadwallader, 'The Correction of the Text of Hebrews towards the LXX' Novum Testamentum, Vol. 34, Fasc. 3 (Jul., 1992), pp. 257-292.
    Do you have a view on this?

    Jonathan,
    Your question poses perhaps the main factor on the other side of the issue: the variety of witnesses (albeit much from a significantly later date) supporting OLW. I would say this is an important consideration, but not decisive in view of the other factors.

    ReplyDelete
  5. Is provenance of reading so unimportant that the earliest attested Greek reading from every location but one should be so easily neglected?

    What's the location of B?

    ReplyDelete
  6. Thx Peter, for providing the name of the article I was trying to remember :)

    Ulrich, I don't think that a lack of a diple automatically invalidates the possibility that a scribe recognized the text as a quote and "improved" it. What evidence is there that the *same* scribe made the diples and "improved" the texts, or that the lack of diple indicates the text was not taken as a quote or that "improvements" were made only on those texts that have a diple?

    As to the case of homoioteleuton, that is possible but it is probable?

    ReplyDelete
  7. sorry.. should have written ".. but is it probable?"

    ReplyDelete
  8. What's the location of B?

    I surmise that its text is from Egypt.

    ReplyDelete
  9. This comment has been removed by the author.

    ReplyDelete
  10. Why is Egypt your surmise?

    Skeat, e.g., put the writing of B in Caesarea, which, if true, would nicely negate your "but one location" point.

    ReplyDelete
  11. Peter and Timo, I have no stake in deciding which of the Heb 3:2 readings is the "Ausgangstext". My intention was to question the unquestioned basic assumtion that scribes were more likely inclined to harmonization towards a OT source text than disharmonization.

    In my experience this is much too easily invoked on the basis of the selection of texts we have at our disposal. But you are in good company. The Alands and Metzger think so too.

    BTW- 20th c. editors of the LXX tended to assume the same thing, but from the other side. In Rahlfs' view, e.g., scribes of LXX mss were guilty of harmonization towards the NT version of the source text.

    I still wrestle with reconciling theoretically both perspectives as pervasive phenomena in operation at the same time. Moreover, there is evidence to the contrary, i.e. that harmonization was emphatically a none issue with some corrections in Sinaiticus and Vaticanus, which makes me wonder what an even-handed perspective on the issue of alleged harmonizations would look like, if we included LXX textual information as well.

    ReplyDelete
  12. Even were Vaticanus copied in Caesarea under the care of Eusebius, such would not make its readings Caesarean. Its high coincidence of readings with Clement, Origen, p46, p75, Sahidic, etc., still suggest Egypt as its primary provenance of reading. I do not doubt that Origen's weight as the most prolific scholar of his day would have helped the spread of his native land's critically famous texts, and we see the probability of this on many occasions, but certainly not at Heb 3:2 (or 3:6 with the omission of μεχρι τελους βεβαιαν [MECRI TELOUS BEBAIAN])!

    Jonathan C. Borland

    ReplyDelete
  13. Let's see. Despite the argument about provenance and the importance of "but one location," what now seems more important to you is "primary provenance." And "primary provenance" is determined by not by location (Caesarea is not in Egypt, after all) but by "high coincidence of readings."

    So then, what's the point of arguing about location, when it ultimately gets trumped by a high coincidence of readings?

    ReplyDelete
  14. Ulrich,
    Good point. It would be good to study both types of phenomenon together.

    ReplyDelete
  15. It is not at all certain that Caesarea was where Vaticanus was copied, but even if it were, I thought provenance was determined by probabilities combined with comparison with orthography, Fathers, and Versions. Thus, for reasons mentioned in my last post, two papyrus mss with Vaticanus as the sole Greek attestation of a reading, if they convince me of anything, do not convince me of much of a provenance outside of Egypt.

    ReplyDelete
  16. Hi All,

    If you are ever interested in comments on textual variation in Hebrews, I survey comments from a number of sources (Zuntz, Metzger, Attridge, and other commentaries) relating to the 44 variation units in the UBS4 apparatus in part two of my PhD dissertation. The relevant survey for OLW begins on page 293 of the PDF, which can be obtained here: http://www.tfinney.net/PhD/PDF/part2.pdf. If you look there, you will see that Attridge agrees with Ulrich's observation concerning homoioteleuton.

    You might also find relevant comments by looking at the transcription notes relating to the manuscripts which have the verse in question. In this case you could search the PDF for the string "". If you do this you will find that Henry A. Sanders said something about OLW in _A Third-Century Papyrus Codex of the
    Epistles of Paul_ (Ann Arbor: University of Michigan Press, 1935), 61.

    Concerning Stephen Carlson's question about where B is from, "spelling" maps produced by multidimensional scaling of the orthographic agreements among all majuscule and papyrus manuscripts of Hebrews known in 1999 point to Egypt, as do "textual" maps based on agreements among orthographically levelled transcriptions of the same manuscripts.

    Best,

    Tim Finney

    ReplyDelete
  17. Hi All,

    I noticed that the string I said to search for didn't get through, no doubt because it used angle brackets. The relevant string is "<ch 3><v 2>".

    Best,

    Tim Finney

    ReplyDelete
  18. This is how the argument is looking to me:

    You made an argument against the quality of the text of Vaticanus at a particular variation unit based on the location of the witnesses. There is, however, no direct evidence for the location of B's copying, so you go to indirect evidence. Yet, the indirect evidence for B's location is its quality of the text.

    It seems to me that, this "but one location" argument, as applied to B, is just begging the question.

    I'm sure there are good arguments against B's reading at Heb 3:2, but the "but one location" argument isn't one of them.

    ReplyDelete
  19. Thanks Tim.

    Re the Timo/Ulrich comments: I was musing to myself this morning while cycling to work that precisely the presence or absence of OLW seems fairly decisive in the degree to which there is a clear allusion to Num 12.7 here in 3.2. So perhaps it goes beyond a harmonisation to the text of the LXX, since the (hypothetical) addition of OLW then constitutes 3.2 as a strong allusion to scripture.

    ReplyDelete
  20. Peter Head:
    Re the Timo/Ulrich comments: I was musing to myself this morning while cycling to work that precisely the presence or absence of OLW seems fairly decisive in the degree to which there is a clear allusion to Num 12.7 here in 3.2. So perhaps it goes beyond a harmonisation to the text of the LXX, since the (hypothetical) addition of OLW then constitutes 3.2 as a strong allusion to scripture.

    The gist of your musings, Peter, suggest to me that the (strong) allusion with OLW has to be credited to someone, either a scribe or the author. Take your pick.

    But, if you think this is a valid line of argument, you have to pick the author. This is clearly a no-brainer. Don't disappoint me, Peter ;-)

    ReplyDelete
  21. Stephen,

    Thanks for the helpful dialogue. My original question was based on the assumption, as you rightly pointed out, that the Alexandrian nature of Vaticanus actually indicates that it might have proceeded from there, and that the two papyrus manuscripts (p13 p46) found in Egypt might actually have been produced and used in that location. My further assumption was that three Greek mss alone (p13 p46 B) out of the 560 or so mss we have of Hebrews does rather indicate the antithesis of what we would normally consider widespread in the GNT copying tradition. My yet further assumption was that studies like Zuntz' indicating the relationship of p46 and B as a group would tend to suggest their limited local-text nature when not supported by other mss with a decidedly different textual history.

    So perhaps I should recast my original question thus: Is provenance of reading so unimportant that the earliest attested Greek reading from every Greek manuscript everywhere but three, two of which were found in Egypt and one of which is the leading representative of what are called "Alexandrian" readings, should be so easily neglected?

    Jonathan C. Borland

    ReplyDelete
  22. RE harmonization.

    It seems to my 'his whole house' is a pretty common statement in the OT. It doesn't seem to be a stretch that the scribe harmonized to that phrase.

    Do we know if Greek speaking scribes experienced words in their exact case, or just words (That is do we really need to find a case where the phrase is 'in his whole house' as an exact harmonization?).

    I know as an english speaker, I don't treat verbs in different tenses or nouns in the singular or plural as different words, but I'm not sure how much it would associate with a recognizable phrase.

    ReplyDelete
  23. Is it not more probable to suggest that perhaps a common ancestor of three Alexandrian mss accidentally skipped three letters for some reason, perhaps since wherever EN (εν) occurs in the NT it is followed by the letter T (τ) nearly half the time? Besides, such small omissions are characteristic of scribal habits among the papyrii anyway, as P. M. Head and others have well noted, no?

    Jonathan C. Borland

    ReplyDelete
  24. I agree with Jonathan that this looks like a local variation of the text. It looks to me like an early copyist omitted OLW from a copy destined to influence later texts which seem to be Egyptian (P13, P46, B, Coptic). I don't know whether the omission was by accident (parablepsis) or design (the scribe didn't think OLW should be there). I'm inclined to vote against parablepsis as I imagine a scribe would read and copy a phrase like EN OLW TW OIKW AUTOU as a unit.

    Tim Finney

    ReplyDelete
  25. Thinking about it a bit more, I was too quick to dismiss the possibility that P13, P46, B, the Coptics, and Ambrose preserve the initial text and that the rest carry corruptions at this point. It is easy to see how this could happen: EN OLW TW OIKW is quoted from the LXX a bit further on (Heb 3.5). The author of Hebrews may well have written EN TW OIKW instead of EN OLW TW OIKW as OLW is superfluous to his argument. However, adding OLW at 3.2 would be such a tempting harmonization as to be unstoppable, almost. The errant scribes need never have looked at a copy of the LXX.

    This doesn't change my view that the witnesses which omit OLW are all Egyptian, apart from Ambrose and one Vulgate manuscript. Interestingly, multidimensional scaling results based on the 44 variation units of the UBS4 apparatus of Hebrews place the text of Ambrose and P46 near each other.

    Tim Finney

    ReplyDelete
  26. Thanks Tim for pointing two relevant points.

    It seems to me that the likeliest explanation is this. The original text was EN TW OIKW AUTOU (omitting OLW) in Heb 3.2. Homoioteleuton seems less likely than a harmonization, given that the text is a unit. Later on a scribe harmonized Heb 3.2 to Heb 3.5, which contains OLW and is a quote from LXX. Royse has shown quite convincingly that scribes harmonized their texts to their immediate context and in this light adding OLW to Heb 3.2 makes sense.

    Hence, I stand by my original assessment that P13 P46 B Coptic Vg-ms Ambrose Cyril (1/3) most likely contain the original reading, even *if* this text is taken as coming from one locale. Such assessment remains in some doubt because of the lack of OLW in Ambrose and Vg-ms, even though Ambrose is close to P46 in these instances, as Tim indicated.

    ReplyDelete
  27. This comment has been removed by the author.

    ReplyDelete