Correction: I mean accidental
"(near-)homosemy", i.e. unmotivated similarity of meaning between unrelated
roots.
Before Miguel responds, I'd like to mention
a pattern that doesn't conform to the usual rules of PIE allomorphy but looks
real enough:
*gWem- : *gWah2- 'come, walk'
*men- : *mah2- 'think, be mentally
roused'
*bHen- : *bHah2- 'say'
And what's the function of "determinants"
in roots like *drem-/*drah2-/*dreu- 'run'? Unextended **der- 'run' is not
attested, but then word-coining need not start with a real "base"; as historical
examples show, we can have all sorts of backformations based on truncated roots
treated as bound morphemes (say, if originally simplex *stah2- 'stand' is
reanalysed as *st-ah2-, this opens the way for *-ah2- to be substituted with
other "determinants", producing a constellation of alliterating roots that
cannot be related in the regular way).
Piotr
----- Original Message -----
Sent: Wednesday, January 24, 2001 12:51 PM
Subject: Doublets in PIE
First
possibility (already discussed ad nauseam): purely accidental homophony or
near-homophony. Hardly exciting but common.