Re: *-tro-/*-tlo-

--- Piotr Gasiorowski <gpiotr@...> wrote:

> >> --- In cybalist@yahoogroups.com, "stlatos"
> <stlatos@> wrote:
>
> > I do not believe the PIE forms ending in *-tlo-
> and *-tro- have the
> > same origin. There are many reasons for this, but
> first consider:
> >
> > *staxtlos stx,tl(e)+ > OPr stakle 'support'; G
> sté:le: 'block'
>
> I'd exclude the Greek word as possibly derived from
> PIE stel- 'support,
> put up' (perhaps *stl.-nah2, cf. Gmc. *stullan-
> 'pillar' and *stel-an-
> 'stalk').

There are no certain cases of ln>:l. All def. ones show ln>ll in
all dialects (or ln if n is restored by analogy). The case of ou^los
'curly' is from *wlxwos and shows R+tone X > oR/Ro before m/w and similar:

*tl,xmn,+ > tólma

*tr,xYmo+ > tórmos

*pr,xWmo+ > prómos

The dialects do not show anything favoring *ln but allow that
derivation in *sl as Sihler says in his _New Comparative Grammar of
Latin and Greek_ (224.2.b):

A widely-held view, however, is that these continue original *-ln-,
particularly in formally opaque formations (hence not subject to
recomposition). In fact, the few forms thus analyzed that are not
profoundly obscure admit of better explanations. One of the clearer
among the latter is G sté:le: 'block, slab', Dor. stá:la:, Lesb.
stálla:, conventionally traced to *stalna: < stl,-neH2-, from PIE
*stel- (the root seen in stéllo: 'put in order'). However, on both
formal and semantic grounds the form is better seen as a noun in
*-sleH2- from a different root, *steH2- 'stand':

He doesn't allow Ktl>Ksl, so this *-slo- would really be PIE *-tlo-
in meaning. The rules for Greek include:

t>th/x+syl_R
x-syl t > s/_l
x()-syl t > th/_R

> > *tlaxtlos > *taxtlos tx,tl(e)+ > L tabula; G
> te:lía 'board'
>
> Why not *tl.h2-tlah2- > *t&2-tHlah2? What do you
> need the full grade
> for?

You said that *ah2t>ath but not *h2,t; I say Greek can be explained
only with full grade. Either way it seems needed.

> > *xwexYtlos > G áethlon 'prize of a contest'
>
> OK, but why masculine? (the same concerns the
> examples above and below)

This doesn't matter for this section; I'll explain why later but it
doesn't seem very important since all these -tlo- forms may have any
gender in IE languages.

> > *bhuuxtlos > Slavic *bu:dlod > Czech bydlo
> 'dwelling', etc.
>
> Slavic evidence proves nothing about the original
> distribution of
> *-tl/ro- variants. They have been levelled out
> completely there.

What? How about Germanic *buþla-? I didn't think my choice of
which example to use would matter.

> > then compare them to:
> >
> > *mexY-trom mxY,-tr(e)+ > Skt má:tra:-, G métron
> 'measure'
>
> The Greek word may be *méd-tro-, with the same root
> as in Gmc. *met-a/i-
> 'mete out, measure' treatment of *-d-tr/lo- as in
> Gmc. *seTla- <
> *séd-tlo- and *Billa- 'sword' < *BiDla- <
> *bHid-tló-. That would explain
> the short vowel _and_ the absence of aspiration
> induced by *h1.

What part are you concerned about? Greek doesn't show tr>thr after
xY, (*ter-xY-trom > G téretron). And isn't bill "a particular kind"
of sword?

> > *xYer-xY-trom xYr,xY-tr(e)+ > Skt arítra- 'oar';
> Lith irklas
> >
> > *xar-xW-trom > G árotron; L ara:trum 'plow'
> >
> > *ter-xY-trom > G téretron; L terebra 'auger'
> >
> > PIE *gWer-xW-trom 'throat' > Lith gerkle:; Grk
> *bérathrom > bérethron
> > / bárathron 'pit'
> >
> > *kWen.-x-tro+ > Skt khanítra- 'spade'
> >
> > *pew-x-tro+ > Skt pavítra- 'filter, etc.'
>
> How do you know that the last two have Skt. -tra-
> from *-tro- rather
> than *-tlo-? The others above involve roots with *r,
> which explains the
> preference for *-tro-.

Because they show exactly the same odd form as the others. PIE
*snexYtlo+ > ON snæ:lda 'spindle' not **sen-xY-tlo+, etc.

> > From this it seems that root of CVX are the same
> in both, but CCVX
> > appear as CCVX-tlo- but CVC-X-tro- and PwVx as
> Pux-tlo- but
> > PVw-x-tro-.
>
> How about the 'oar' word, which shows both the full
> grade (*h1érh1-) and
> the zero grade (*h1r.'h1-) without any semantic
> difference?

The diff. between -tlo- and -tro- is in the full grades. I gave
both grades, the 0-grades can be the same, but the full are different:
no *xYrexY-tro+, *kWnax-tro+, etc.

I'll be giving additonal ev. later.

> > Therefore, no *tel-x-tlo+ even though *tel-x-mon.+
> existed, no
> > *bhew-x-tlo+ even though *bhew-x-to+, etc. This
> could most easily be
> > from an older difference in syllabification as
> C-tr but Ct-l never
> > C-tl. Of course, other ev. shows me that PIE was a
> tonal language
> > with multiple high, mid, low assigned for various
> meanings, but the
> > exact reason doesn't matter here.
>
> The different syllable division may explain the
> preference for *CRX-tlo-
> over *CeRX-tlo- even if the root syllable was
> accented.

There is no pref. for 0- over full-grade; each PIE word had both and
any IE language may preserve one or the other, as in many kinds of words.

> (1) Nomina actionis, typically with final accent, as
> in Gk. loetrón
> (Myc. <re-wo-to-ro>) from *leuh3-tró-m, or Skt.
> hotrá- 'offering, the
> office of a priest' < *g^Heu-tró- (or *-tló-). So
> <hotrá> is the job or
> duty of a <hótar->: *[g^Héu-tor]-ó- -->
> *g^Heutr/ló-.

The PIE is *Low-xY-trom > *lau-þran > ON lauðr, OE le:aðor 'foam' >
lather. There are many different changes in sequences of e-o in
various Greek dia.

The noun is from the caus. *LowexYe+ > *Loweye+. Similar forms
existed which also created *dholbh-x()-tlo+ > *dolpto > Russian doloto
'chisel', etc.