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

NT Stemmatics

Expand Messages
  • Jonathan Borland
    I have a few questions regarding NT stemmatics. Stephen Carlson has a short article with some links found here: http://www.mindspring.com/~scarlson/tc/
    Message 1 of 7 , May 20 5:03 AM
    • 0 Attachment
      I have a few questions regarding NT stemmatics. Stephen Carlson has a
      short article with some links found here:

      http://www.mindspring.com/~scarlson/tc/

      Carlson's definition of NT stemmatics in a nutshell: "Cladistics
      proceeds by examining each point of variation to find the 'optimum
      tree' with the fewest changes" (§14). He goes on to say that
      "cladistics does not require prior knowledge of the original reading,"
      which is nice, because we wouldn't need cladistics if we knew the
      original reading.

      So my questions are:

      1. To find the optimum tree with the fewest changes, does this mean we
      should start with the evidence of all MSS from all ages and see what
      results as the mammoth perpetuation of original readings from which all
      erroneous readings emanated?

      2. Do dates of MSS not really matter in cladistics, since basically all
      MSS had to be copied from earlier MSS, and it can be proven that some
      late MSS, like, e.g. 33, are related to some of the earliest MSS so far
      found?

      3. Does cladistics give more credibility or "weight" to MSS that have a
      more continuous succession of non-deviation from the "optimum tree"
      than those that do not? In other words, do MSS that perpetually
      deviate from the "optimum tree" receive a poor credibility rating
      (contra modern eclecticism), or is modern eclecticism allowed to stand?
      If the latter, I'm wondering if NT stemmatics can maintain any
      reputation at all as a technical solution to NT TC.

      Jonathan Borland
      Lakeland, FL
    • Stephen C. Carlson
      ... The determination of whether a reading is original or an error occurs after the cladistic analysis rather than before. It turns out that the number of
      Message 2 of 7 , May 20 10:26 PM
      • 0 Attachment
        Jonathan Borland wrote:
        >1. To find the optimum tree with the fewest changes, does this mean we
        >should start with the evidence of all MSS from all ages and see what
        >results as the mammoth perpetuation of original readings from which all
        >erroneous readings emanated?

        The determination of whether a reading is original or an error occurs
        after the cladistic analysis rather than before. It turns out that the
        number of changes on a tree remains the same regardless of which
        part of the tree is oriented to be the root (archetype). However, the
        number of changes on the tree is affected by the *shape* of the tree.
        For example, a tree in which 33 is close to L, say, has fewer changes
        on it (other things being equal) than one in which 33 is distantly
        separated from L due to the relatively high number of agreements of
        33 and L against other MSS. If 33 and L are close together, then each of
        their agreements need only imply one change on the tree rather than
        two independent and coincidental changes on two different parts of
        the tree.

        >2. Do dates of MSS not really matter in cladistics, since basically all
        >MSS had to be copied from earlier MSS, and it can be proven that some
        >late MSS, like, e.g. 33, are related to some of the earliest MSS so far
        >found?

        Since the number of changes on a tree remains the same regardless
        of how the tree is oriented, cladistics does not care about the dates
        of MSS. I suppose that if the dates of MSS closely match the shape
        of the optimal tree (e.g. the tree shows a progression from early to
        late MSS), one could use that information to decide fairly confidently
        how the tree should be oriented, but the shape of the text is going to
        be determined by the patterns of readings in the MSS, not by their
        dates.

        >3. Does cladistics give more credibility or "weight" to MSS that have a
        >more continuous succession of non-deviation from the "optimum tree"
        >than those that do not? In other words, do MSS that perpetually
        >deviate from the "optimum tree" receive a poor credibility rating
        >(contra modern eclecticism), or is modern eclecticism allowed to stand?
        > If the latter, I'm wondering if NT stemmatics can maintain any
        >reputation at all as a technical solution to NT TC.

        Cladistics does not weigh MSS; rather, it seeks to find an arrangement
        of MSS (i.e. a tree) that minimizes the number of changes implied by
        the tree. Once an optimal (unrooted) tree is found, one branch of it
        can be selected for the location of the "root" or the Archetype. After
        the tree is rooted, it can be treated as a stemma and the stemmatic
        history will, a posteriori, imply a certain weighting to the witnesses.

        Cladistics tells us the shape of the optimal tree; other methods
        have to be used to orient the tree so that the branch closest to
        the autograph is on top. The latter operation is not technically
        part of the cladistic maximum parsimony analysis. I suppose a
        Byzantine prioritist and one who weighs the MSS in accordance
        with modern eclectic praxis might disagree over where to locate
        the root of the tree (i.e., near the Byzantine and Alexandrian
        hyparchetypes, respectively), but at least they will be arguing
        over the same basic shape of the transmissional history rather
        than over completely different conceptions of it.

        Cladistics is a method for producing a stemma and for that reason
        can be thought of a subdiscipline within stemmatics. Stemmatics
        differs from modern eclecticism in that it postulates an explicit
        history of the text (the stemma or family tree) from which readings
        of ancestors can be reconstructed, and the choice of readings at
        any place in the tree is constrained by that history. This is how
        stemmatics avoids the risk, raised by Maurice Robinson, inherent
        in the approach of considering each variation unit in isolation of
        each other, which could produce a text (a sequence of readings)
        divorced from any historical pattern. In actual practice, though,
        the choice of readings in modern eclecticism seems to follow B
        and its relatives so closely over large stretches of text that the
        risk of producing a historically unrealistic text is remote. (Whether
        that text constitutes the Archetype or merely a proto-Alexandrian
        hyparchetype is a different question!)

        Stephen Carlson

        --
        Stephen C. Carlson,
        mailto:scarlson@...
        "Poetry speaks of aspirations, and songs chant the words." Shujing 2.35
      • Wieland Willker
        Thought of the day: What if the amount of information is not sufficient enough to create the correct stemma? Of course we can only use what we have, but if we
        Message 3 of 7 , May 25 6:19 AM
        • 0 Attachment
          Thought of the day:
          What if the amount of information is not sufficient enough to create the
          correct stemma? Of course we can only use what we have, but if we have
          not enough? If the input data are too fragmentary and incomplete then
          the resultant stemma probably has not much in common with
          reality/history.
          It could be checked using artificial data to show how much you need to
          get near the truth (how many "missing links" are allowed). The problem
          with NT TC is that we don't even know how much is lost.

          Best wishes
          Wieland
          <><
          ------------------------------------------------
          Wieland Willker, Bremen, Germany
          mailto:willker@...-bremen.de
          http://www.uni-bremen.de/~wie
          Textcritical commentary:
          http://www.uni-bremen.de/~wie/TCG/index.html
        • Jonathan Borland
          I think I understand the question. How do we know that what exists is an accurate representation of the original? I guess one explanation goes like this. 1.
          Message 4 of 7 , May 25 7:24 AM
          • 0 Attachment
            I think I understand the question. How do we know that what exists is
            an accurate representation of the original? I guess one explanation
            goes like this.

            1. All copies came from somewhere.

            2. When no variants occur, the probability is that all copies have
            transmitted the original reading. If this probability is not granted,
            then NT Stemmatics really is a worthless venture, and so is all NT TC
            for that matter.

            3. When only one variant occurs against all the rest, the probability
            is that it occurred in isolation and relatively late in the MS
            tradition, so late that its descendants (also not extant), if any, were
            unable to multiply the variant reading enough to impact the growing
            numbers of the copies that contained the original reading.

            4. If this kind of transmissional probability is granted, then
            analyzing the evolution of variant units back to the optimal tree trunk
            from which all branches and variant units emerged seems like a
            worthwhile task.

            And hopefully NT Stemmatics will start with all the branches and not
            just a select few that are presupposed to be the best. Otherwise it is
            not a scientific exercise at all.

            Jonathan Borland
            Lakeland, FL


            On May 25, 2004, at 9:19 PM, Wieland Willker wrote:

            > Thought of the day:
            > What if the amount of information is not sufficient enough to create
            > the
            > correct stemma? Of course we can only use what we have, but if we have
            > not enough? If the input data are too fragmentary and incomplete then
            > the resultant stemma probably has not much in common with
            > reality/history.
            > It could be checked using artificial data to show how much you need to
            > get near the truth (how many "missing links" are allowed). The problem
            > with NT TC is that we don't even know how much is lost.
            >
            > Best wishes
            >     Wieland
            >        <><
            > ------------------------------------------------
            > Wieland Willker, Bremen, Germany
            > mailto:willker@...-bremen.de
            > http://www.uni-bremen.de/~wie
            > Textcritical commentary:
            > http://www.uni-bremen.de/~wie/TCG/index.html
            >
            >
            >
            >
            >
            > Yahoo! Groups Sponsor
            >
            > ADVERTISEMENT
            > <lrec_companion_043004.gif>
            > <l.gif>
            >
            > Yahoo! Groups Links
            >
            > • To visit your group on the web, go to:
            > http://groups.yahoo.com/group/textualcriticism/
            >  
            > • To unsubscribe from this group, send an email to:
            > textualcriticism-unsubscribe@yahoogroups.com
            >  
            > • Your use of Yahoo! Groups is subject to the Yahoo! Terms of
            > Service.
            >
            >
          • Stephen C. Carlson
            ... Lack of data in cladistics usually leads to lack of resolution. The maximal lack of resolution is the bush pattern in which all extant witnesses are
            Message 5 of 7 , May 25 9:51 AM
            • 0 Attachment
              At 03:19 PM 5/25/2004 +0200, Wieland Willker wrote:
              >Thought of the day:
              >What if the amount of information is not sufficient enough to create the
              >correct stemma? Of course we can only use what we have, but if we have
              >not enough? If the input data are too fragmentary and incomplete then
              >the resultant stemma probably has not much in common with
              >reality/history.

              Lack of data in cladistics usually leads to lack of resolution. The
              maximal lack of resolution is the bush pattern in which all extant
              witnesses are believed to be descendent of some common ancestor but
              there is not enough evidence to conclude that any two are more closely
              related to each other. For example, an unrealized bush with four
              witnesses, A, B, C, and D, looks like:

              x
              / | | \
              A B C D

              If a polychotomy in the stemma (i.e., a MS in the stemma with multiple
              descendents) is interpreted as "soft", then the stemma is considered
              correct as long as A, B, C, and D are actual descendents of x no matter
              how they are further related to each other. Thus, lack of data that
              can specify in more detail the relationships among A, B, C, and D
              would lead to a soft polychotomy.

              However, if it can be determined that A and B agree in error more than
              any other pair, the relationships in the stemma become more resolved
              or distinct, such as:


              x
              / | \
              y C D
              / \
              A B

              There are well-accepted statistical techniques, namely the "bootstrap,"
              that are commonly used for judging how much resolution in the stemma
              is actually supported by the data.

              >It could be checked using artificial data to show how much you need to
              >get near the truth (how many "missing links" are allowed). The problem
              >with NT TC is that we don't even know how much is lost.

              People have done this. Check out the last several years worth of issues
              in the journals of _Systematic Biology_, _Cladistics_, etc. There are
              dozens and dozens of article on such statistical techniques.

              Joe Felsenstein recently came out with a very nice textbook called INFERRING
              PHYLOGENIES, which explains a lot of the theory and practice behind cladistics
              (without getting too doctrinaire about it). If your university supports a
              decent biology/systematics/bioinformatics program, your library may already
              have it.

              Stephen Carlson
              --
              Stephen C. Carlson mailto:scarlson@...
              Weblog: http://www.mindspring.com/~scarlson/hypotyposeis/blogger.html
              "Poetry speaks of aspirations, and songs chant the words." Shujing 2.35
            • Wieland Willker
              ... No, not original , but history . How do we know that what exists is a (tolerably) accurate representation of the textual history? I think, simplified, we
              Message 6 of 7 , May 25 9:51 AM
              • 0 Attachment
                > I think I understand the question. How do we know that what exists
                > is an accurate representation of the original?

                No, not "original", but "history". How do we know that what exists is a
                (tolerably) accurate representation of the textual history?
                I think, simplified, we have a better understanding of what the original
                is than what the history of the text is. And the creation of a stemma
                is, IMHO, not so much an attempt to reconstruct the original, but to
                reconstruct the HISTORY of the text and so to get a better justification
                for the text we claim to be "the original".

                Best wishes
                Wieland
                <><
                ------------------------------------------------
                Wieland Willker, Bremen, Germany
                mailto:willker@...-bremen.de
                http://www.uni-bremen.de/~wie
                Textcritical commentary:
                http://www.uni-bremen.de/~wie/TCG/index.html
              • Stephen C. Carlson
                ... No, the question was about the stemma (the pattern of relationships among the MSS), not the text of the archetype. Stephen Carlson -- Stephen C. Carlson
                Message 7 of 7 , May 25 9:52 AM
                • 0 Attachment
                  At 10:24 PM 5/25/2004 +0800, Jonathan Borland wrote:
                  >I think I understand the question. How do we know that what exists is
                  >an accurate representation of the original? I guess one explanation
                  >goes like this.

                  No, the question was about the stemma (the pattern of relationships
                  among the MSS), not the text of the archetype.

                  Stephen Carlson
                  --
                  Stephen C. Carlson mailto:scarlson@...
                  Weblog: http://www.mindspring.com/~scarlson/hypotyposeis/blogger.html
                  "Poetry speaks of aspirations, and songs chant the words." Shujing 2.35
                Your message has been successfully submitted and would be delivered to recipients shortly.