What is Relatedness?

[X99Lynx at aol.com]

I wrote:
<<What if five lexical attributes appear as shared innovations between IE
languages A and B and therefore evidence a greater relatedness between the
two than a language C - which has none of them.  What if however Language C
shares a totally valid morphological innovation with B not found in Language
A?  How can this be reconciled?>>

In a message dated 1/28/00 3:15:15 AM, JoatSimeon at aol.com wrote:
<<-- these are known as isogloss boundaries, and present no fundamental
problems.>>

Good solution.  Wrong methodology.  I had been talking about about the UPenn
tree in these posts.

Isoglosses apply to maps.  Basically an isogloss is nothing but a boundary
line.  And in simple location or wave or other diachronic maps, isoglosses
can be used to help untangle what would appear to be conflicting evidence and
make it subjectable to explanation, as with cultural explanations given for
the Benrath and Urdingen isoglosses - the "Rhenish Fan."  Maps add those
kinds of dimensions to the analysis.

HOWEVER, these kinds of conflicts need to be either resolved, avoided or
disregarded BEFORE a tree is drawn, because there is no way to portray these
variables in a SINGLE tree.  The 'pure phylogeny' approach used in the UPenn
tree is to draw MANY trees to arrive at the ones that minimize these kinds of
conflicts (called I believe convexity in this approach). So in effect you are
given the trees that avoid the conflicts best.  Conflicts not reconciliable
by realignment of the branching do not show up as conflicts in the final tree
or trees.

Regards,
Steve Long