Germanic again

From: Alexander Stolbov
Message: 326
Date: 1999-11-23

cybalist message #142cybalist: Odp: Cowboys on Horseback
Dear Piotr,
 
Please excuse my rather late reply on the materials concerning the position of Germanic in the IE dialect tree.
 
I used to relay on the scheme of IE dialects by T.Gamkrelidze & V.Ivanov developed by Cyril Babaev. It can be represented in the following form:
 
                                                 PIE
                                                /      \
                          Non-Anatolian IE     Anatolian
                           /                    \
          (Western block)           (Eastern block)
               /              \                  \       \       \
(Non-Tocharian W.)  Tocharian       \       \       Greek-Thraco-Phrygo-Armenian
 \      \       \        \                           \       Indo-Iranian
   \      \       \        Celtic                   Germano-Balto-Slavic
     \      \        Italic                                   /                \
        \        Venetic                                            Germanic        Balto-Slavic
        Illyrian
 
 
To my mind this scheme remains "correct" , i.e. inwardly non-contradictory (at least in the aspect of Germanic), if we take into account only morphological evedences and ignore phonological ones (which are very strong indeed: satem/centum division, RUKI rule).
 
 
I think your scenario given in the message of November, 2 can be shown in the analogues form (please correct me if I represent it not quite precisely):
 
                                                 PIE
                                                /      \
                                              \       Anatolian
          (Non-Anat.-non-Toch. IE)    Tocharian*
                     /                        \
  Western block                      Eastern block                
    \        \        \                              /          \
      \        \      Celto-Italic          Hellenic    "Satemic"
        \    "Pannonian" (incl.Illyrian)               /           \
         Northern                          "Transylvanian"    "North Pontic"          
          /        \                                                          /        \ 
Germanic     extinct group                            Balto-Slavic    Aryan
                ("Old European") 
                       
 
*The position of Tocharian should be the subject of a special discussion.
 
Again, this scheme is absolutely "correct" (inwardly non-contradictory) if we take one sort of evidences (phonological in this case) and ignore another one (morphological: *m/*bh division and other). 
 
 
What to do in such a situation? To my mind:
 
1. To try to prove theoretically that one sort of evidences is principally more conclusive for problems of genetic kinship than another one. For example: "morphological evidences should be considered only if they don't contradict to phonological evidences" or vice versa. We treat so lexical evidences now - as the weakest argument for genetic relations. If this question is solved theoretically we get the hierarchy of arguments (in order of decreasing the conclusive ability): phonological-morphological-lexical or morphological-phonological-lexical.
 
2. If the question formulated above principally has no general solution we need to consider contradicting evidences together and to find a compromise solution of the highest probability.
 
Piotr, you have suggested a variant where Germanic is an independent (non-West and non-East) branch of IE family. It can be shown like this:
                                       
                         PIE
                        /     \
                       \     Anatolian
                     / \    Tocharian
                   /    Germanic 
                  / \
      Western   Eastern
         block      block
 
Looks like "the golden mean". However in this case it is necessary to demonstrate some archaic features of Germanic and to find innovations, which are shared by the languages of both the Western and the Eastern block contrary to Germanic+Tocharian+Anatolian (or maybe without Tocharian - only Germanic+Anatolian). 
 
Alexander