Re: -tlo- vs -tro- (was: rtl)

I will use conventional PIE forms even though telling some of my
changes may make some points clearer later.

*skWetlos 'tale' > OI scél; Welsh chwedl

Why not **sekWtlos? This isn't an aberration; many roots of CVC
when the second C is a stop show:

*sghetlos > Skt ks.atrá-; Av xs^aTra- 'dominion'

*sghetlyos > Skt ks.atríya- 'sovereign'; Hom. G skhétlios '~ forceful,
heedless'

This seems the same as:

*sgWhetos 'holding in (participle)', *n,+ 'without _, not (able to) be
held in' > Av azgata-; G áskhetos

This is opposed to:

*segYhtor+ > Hékto:r

And this is not a result of the following C, since:

*sgWhenos 'holding down/in (adj.), winning (in wrestling)' > G sthénos
'strength'

even though there are plenty of adj. in C-no-.

This is part of the ev. showing a series of PIE forms in *-etos,
*-enos, *-etlos, with a particular group of meanings. The form *-etos
makes participles of verbs of movement (or not moving, specifically);
*-etlos is the causative. *-enos is the adjective, but with the
relation between noun and adj. in PIE it can often form nouns.

Stops ending the root stay before *-etos, others move so *-eCtos is
formed. Particular combinations may show other forms; it's
complicated by both full- and 0-grade existing, not always showing
both, some unclear as to origin.

In conventional PIE terms, CVGC appear as CíCetos or CúCetos as:

*gWíxWotos *gWíxWt(e)+ > G bíotos; L vi:ta

*bhúxatlos bhúxtl(e)+ > OE botl; Slavic *bu:dlod > Czech bydlo
'dwelling', etc.

This is completely diff. than the treatment of *-tor- and *-tro-.