Loading ...
Sorry, an error occurred while loading the content.

"Brother Timothy" vs. ἡμῶν in Hebrews 1 3:23

Expand Messages
  • Mr. Buck
    Current editions of the Cambridge and Oxford KJV, as well as Webster s 1833 edition and even the NKJV, have ourin italics: Know ye, that our brother Timothy
    Message 1 of 6 , Dec 9, 2013
    • 0 Attachment
      Current editions of the Cambridge and Oxford KJV, as well as Webster's 1833 edition and even the NKJV, have our in italics:
      "Know ye, that our brother Timothy is set at liberty"

      This is as it should be, since neither the TR nor the MT contain HMWN.  But going back in time, the 1611 KJV didn't italicize the word:
      "Know yee, that our brother Timothie is set at libertie"

      This is yet another case in which the Clementine Vulgate served as a Vorlage for the KJV:
      "Cognoscite fratrem nostrum Timotheum dimissum"

      But it's not just the Vulgate, Geneva, Rheims, and Wycliffe that have 'our.' It's in both the Syriac and Coptic versions. In fact, I haven't found a single ancient version to testify to the omission of HMWN (although there appear to be a fair share of mss that demonstrate the common scribal confusion with UMWN, thus giving further testimony against the omission).

      This variant also represents an amazing confluence of peace and goodwill in the field of textual criticism. Vaticanus' voice is silent here, as is, of course, Bezae's. Sinaiticus goes both ways, and p46--notorious for its omissions where it isn't already fragmentary--actually has the word in prima manu. And it's not even an issue for the KJO's, now that the Pure Cambridge Edition has forever relegated our to italics. So regardless of our presuppositions, we should be able to find common ground here, during this most holy time of the year. Well, except for those proponents of the Majority Text, who are invited to make their case in this case.

      The scribal habits research of Colwell, Royse, Head, Hernandez, and Hurtado (I'd like to include Robinson and Wilson, but AFAIK their studies haven't been published yet) conclusively demonstrates that scribes (as opposed to correctors) are more likely to omit than to add. Thus any omission, be it however widespread in the later manuscripts, which is extremely rare to unknown in any of the early manuscripts, or in any versions translated from early manuscripts now lost, is highly unlikely to be original.

      Tim Finney, any insights on this variant from your textual study of Hebrews?

       
      Daniel Buck 
    • yennifmit
      Hi Daniel, This presents an opportunity to show one of the approaches I have been experimenting with lately. An MSW plot suggests that a ten-way partition is
      Message 2 of 6 , Dec 11, 2013
      • 0 Attachment

        Hi Daniel,


        This presents an opportunity to show one of the approaches I have been experimenting with lately. An MSW plot suggests that a ten-way partition is among the most preferable for the UBS4 data set for Hebrews. (There is a peak corresponding to ten groups.)


        http://www.tfinney.net/Views/msw/Heb-UBS4.15.SMD.png


        Using PAM (partitioning around medoids) to divide the data set into ten groups produces this partition:


        33: UBS Aleph A 33 cop-sa

        P46: P46

        2464: C P Psi 0150 81 104 365 436 459 1912 2464 arm slav

        it-d: D it-d

        Byz: K L 075 424 1241 1319 1852 1962 2200 Byz Lect Chrysostom

        1881: 6 1739 1881

        2127: 256 263 1573 2127 syr-h cop-bo

        vg-st: it-ar it-b it-comp vg-cl vg-ww vg-st

        geo-2: syr-p geo-1 geo-2

        eth-TH: eth-pp eth-TH


        (The following are poorly classified though: arm 0150 Psi syr-p A slav cop-sa cop-bo 459 104. it-*, vg-*, geo-*, eth-* = witnesses or versions of the Old Latin, Vulgate, Georgian, Ethiopic presented separately in the UBS4 apparatus.)


        The witnesses placed before colons are the medoids of the respective groups. They act as group representatives. Using bracketed medoid labels to represent groups we get the following based on which group has which reading using the NA26 apparatus as a guide: (Apologies for using the older ed.)


        With HMWN: [P46] [2464] [it-d] [1881] [2127] [vg-st]

        Without HMWN: [Byz]

        Can't say: [eth-TH] [geo-2] (I can't tell from a quick look at the NA26 app.)

        Divided: [2464] because it contains Psi. Whichever groups contain Aleph-2 and D-2 would be divided too.


        Taking the groups as diverse witnesses to the initial text in this case comes down strongly in favour of HMWN.


        There is an immediate objection which is that some of the groups may be descended from others and therefore do not count as independent witnesses to the initial text. To get around that one can look at the NJ analysis result which I will do in a subsequent post.


        As an aside, cases where some venerable witnesses side with [Byz] against most others might be taken as evidence for awareness of [Byz] readings where and when those venerables were produced.


        Best,


        Tim Finney


      • yennifmit
        Hi Daniel, In my other post I demonstrated a PAM-based approach to deciding what is the initial text at this variation site. One problem with that approach is
        Message 3 of 6 , Dec 12, 2013
        • 0 Attachment

          Hi Daniel,


          In my other post I demonstrated a PAM-based approach to deciding what is the initial text at this variation site. One problem with that approach is that some clusters may be dependent on others, thus giving some readings greater weight than deserved. I will now use an NJ result for the UBS4 Hebrews data set to show another approach.


          Applying NJ to the data set produces this tree:


          http://www.tfinney.net/Views/phylo/nj/Heb-UBS4.15.SMD.png


          B is not included as it falls victim to a vetting process designed to eliminate data points based on small numbers of variation sites. There is a way to force B to be included but the resulting tree is based on less data:


          http://www.tfinney.net/Views/phylo/nj/Heb-UBS4.B.15.SMD.png


          I will use the first NJ tree from here on. Some caveats should first be issued. The tree is produced by the neighbour-joining algorithm -- a phenetic rather than phylogenetic technique. It should be regarded as provisional -- a more comprehensive data set will no doubt produce a different NJ tree, though one hopes the tree presented here is not too far off the mark. (I think it's better than one based on a data set that does not include diverse classes of witnesses including ancient versions.) It is also possible that a phylogenetic analysis of the data matrix underlying the distance matrix used to produce the NJ result will produce a different tree which better represents the actual topology of the family tree of New Testament witnesses of the Book of Hebrews. Also, mixture can make things fuzzy and too much mixture would make analysis of this kind pointless. Thanks to Stephen Carlson for his counsel which has prompted me to issue these caveats.


          The NJ tree is unrooted, meaning that any node (i.e. branching point) or leaf (i.e. end point) could be closest to what we seek, namely the text from which all others we have descend. How then does one identify the location of the root? Phylogenetic analysis will identify a root, NJ won't. I therefore need to do additional detective work.


          Is there other evidence to help identify the location of the root? I would say yes. The texts were produced for Christian populations and given the ubiquitous phenomenon of textual variation I expect to see textual varieties emerge which are associated with those populations. On the assumption that the earliest phases of the text's diffusion were characterised by relative isolation of the major Christian population centres, I expect that each centre developed its own flavour of the text. In my view, the major branches of a phenetic tree correspond to early Christian population centres such as the Latin-speaking West, Asia Minor, Syria, Palestine, and Egypt. (Streeter and others said as much a long time ago.)


          It seems to me that the NJ result accords with this view (as do results produced by CMDS, DC, and PAM). One half of the NJ tree for the UBS4 Hebrews data includes the Latin, Armenian, Georgian, Coptic, and Peshitta Syriac versions. This half includes Western and Alexandrian witnesses. Whether the branch comprised of Aleph, A, C, 33, 81, 436, 2464, and the Armenian should be associated with a particular geographical location is left as an exercise for the reader. (One needs to be a bit careful as branches and witnesses can leap from place to place given slight changes in the data sets. Compare, for example, the with-B and without-B trees. Mixed texts are particularly prone to jumping from branch to branch. As it happens, the Aleph, A, C, ... group is seen in CMDS and DC results too.)


          According to the NJ tree, which could be wrong, this half of the tree goes back to a common node. One branch, let's call it X, includes the Latin witnesses (except it-d) and a branch which includes Aleph, A, C, ... If this is basically right, when would we expect these to have diverged? I would say "the date of the common ancestor of these branches" which would make it earlier than the creation of the Latin tradition. Somewhere in the second century seems a reasonable date for the initial Latin translation of Paul's letters (including Hebrews). (I am inclined towards early rather than late second century because of the missionary impulse of the early Church. Why wait? I think the same way in relation to the Syriac and Coptic.) And what would be the date corresponding to the node where branch X splits from the common ancestor of the Georgian, Coptic, Syriac Peshitta, D, it-D, and P46? Earlier still. This is where I think we should be looking for the nearest extant relatives of the initial text. It would seem that the Byzantine branches also go back to a text in this vicinity.


          The NJ tree indicates that texts such as 6, 1881, 0150, Psi, and the Harclean Syriac stand in this neck of the woods. (Thomas of Harkel would seem to have done a pretty good job at Enaton in 616.) The NA26 apparatus (all I have handy) says that Psi omits HMWN and 1881 includes it. (Can anyone tell us what the others have?) On the principle that textual novelties are unlikely to propagate to a majority of child texts, I would say that the most frequent reading among these witnesses is a reasonable estimate of the initial text at this variation site, given the sample presented in the UBS4 apparatus.


          Best,


          Tim Finney


        • Stephen Carlson
          ... Thanks for that. I find it odd that the position of 1739 changes so much between the two trees by including or excluding B. It seems that, without B, 1739
          Message 4 of 6 , Dec 17, 2013
          • 0 Attachment
            On Fri, Dec 13, 2013 at 7:11 AM, <tjf@...> wrote:

            Applying NJ to the data set produces this tree:

            http://www.tfinney.net/Views/phylo/nj/Heb-UBS4.15.SMD.png

            B is not included as it falls victim to a vetting process designed to eliminate data points based on small numbers of variation sites. There is a way to force B to be included but the resulting tree is based on less data:

            http://www.tfinney.net/Views/phylo/nj/Heb-UBS4.B.15.SMD.png

            I will use the first NJ tree from here on.

            Thanks for that. I find it odd that the position of 1739 changes so much between the two trees by including or excluding B. It seems that, without B, 1739 is attracted to P46. For what it's worth, your second tree is closer to one I produced for Galatians.  In Galatians, I rooted the tree on the branch that separates P46-B-D-F-G-d on one side with everything else on the other side. 

            Stephen
            --
            Stephen C. Carlson, Ph.D. (Duke)
            Post-Doctoral Fellow, Theology, Uppsala
          • yennifmit
            Hi Stephen, Thanks for your reply. Some witnesses tend to jump from branch to branch with slight changes to the underlying data set. The different locations of
            Message 5 of 6 , Dec 19, 2013
            • 0 Attachment

              Hi Stephen,


              Thanks for your reply. Some witnesses tend to jump from branch to branch with slight changes to the underlying data set. The different locations of 1739 in the with- and without-B NJ results is a perfect example. DC, NJ, and PAM all do the same kind of thing, placing something in a branch or category even when it might not be a good fit. One reason this can happen is that the text is mixed, somewhere near the watershed between two taxa. Another can be that the text doesn't fit into categories populated by later texts, and this is the category where I would place 1739.


              Looking at the CMDS map is often helpful in such cases. 1739 is between P46 on one hand and 1881 (and 6) on the other. It's in the region of textual space I've been pointing to as a good place to look for extant near relatives of the initial text.


              How did you decide to root your tree where you did?


              Best,


              Tim Finney



            • Stephen Carlson
              ... Basically, internal evidence. The phylogenetic method I chose enables me to reconstruct the state of every inferred ancestor, so I can see what changes are
              Message 6 of 6 , Dec 20, 2013
              • 0 Attachment
                On Fri, Dec 20, 2013 at 3:20 AM, <tjf@...> wrote:

                How did you decide to root your tree where you did?

                Basically, internal evidence. The phylogenetic method I chose enables me to reconstruct the state of every inferred ancestor, so I can see what changes are inferred to have happened on each branch and evaluate those changes based on the traditional use of intrinsic and transcriptional probabilities. What I've found was that many Western readings, particularly those supported by P46 and/or B, were fairly strong in terms of internal evidence. But that's what Zuntz said for 1 Cor and Hebrews.

                Stephen 
                --
                Stephen C. Carlson, Ph.D. (Duke)
                Post-Doctoral Fellow, Theology, Uppsala
              Your message has been successfully submitted and would be delivered to recipients shortly.