Celtic closer to Anatolian?

[ECOLING at aol.com]

For Celtic,
is there any argument that on balance
the geographic position of Celtic in the earliest
stages of PIE dialect network would put it
closer to Anatolian, or Tocharian, or Armenian, or etc.,
because of a few shared isoglosses with those
which might be common innovations?
Or are any sharings retentions?
I am trying to dredge up some old memories.

Lloyd Anderson
Ecological Linguistics