Etruscans (was: minimal pairs)

[Miguel Carrasquer Vidal]

On Sun, 18 Feb 2001 23:04:34 -0000, "Douglas G Kilday"
<acnasvers at hotmail.com> wrote:

>Miguel Carrasquer Vidal (11 Feb 2001) wrote:

>>I'm following Beekes and v.d. Meer here, who reconstruct:

>>s-gen.  *-si
>>l-gen.  *-la
>>loc.    *-i

>The oldest Etruscan inscriptions (700 BCE) have s-genitives in -s. In some
>of the grammatical literature, datives in -si/s'i (the distinction is
>orthographic South/North) are confounded with genitives in -s/s' or regarded
>as "emphatic" genitives, and the derivative suffix -s'i/si adds to the
>confusion. I don't know any basis for reconstructing the s-gen. as *-si.

>The genitive in -la is characteristic of enclitic demonstratives (-cla,
>-tla, -s'la/sla, -s'vla/svla from -ca, -ta, etc.). It is not clear that the
>dative in -le originated from *-la + -i. Early Archaic nouns taking this
>inflection are typically declined thus: nom. Arath, gen. Arathia, dat.
>Arathiale. Late Archaic (Orvieto) has gen. Aranthia and Aranthial; Recent
>dialects have usually Arntheal, Arnthial, or Arnthal. Whether -l was sounded
>but not written in Ara(n)thia is an open question: was it always there
>(reduced from *-la), did it excresce, or was it "restored" by analogy?
>Lemnian final -l suggests that it was there in Arch. Etr. Since final -a of
>*noun-stems* does not contract with -i in Archaic, the vowel of Proto-Etr.
>*-la must have been half-short or a schwa, i.e. *-l at .

>>abl = gen + gen [ *-la-si > -las > -ls; *-si-si > *-sis > -is]
>>dat = gen + loc [ *-la-i > -le; *-si-i > -si]

>The l-ablative can be explained as -ls < -las < *-l@ + -s. The s-abl. cannot
>be decomposed this way. First, gen. *-si is a figment; second, medial /s/
>doesn't just vanish in Etruscan. I prefer to regard -is as unitary.
>Compulsive atomists, of course, are free to "derive" it from -i + -s.

I'd better quote (translate) Beekes & v.d. Meer in full:

[paradigms:]

stems in: -a               -u     -e      -i     -C     plural

nom.      -a,-0            -u,-0  -e,-0   -i,-0  -C     -r
s-gen.    -as              -us    -es     -is    -Cs    -ras
s-abl.    -es              -uis   -e(i)s  -is    -Cs
s-dat.    -asi             -usi                  -Csi   -rasi
l-gen.    -al              -ul    -el            -Cl
l-abl.    -al(a)s
l-dat.    -ale,-althi
loc.      -e                      -e(i)
loc.+thi  -ethi,-aithi            -ethi                 -rthi

The _genitive_ was treated above [discussion about distribution of -s
and -l genitives].

An _ablative_ was formed by adding the gen. -s to the genitive. With
the l-gen. this gave -l-s, archaic -las (so the gen. -l is probably
from *-la).  This is the so-called double genitive.  With the gen. in
-s, that probably came from *-si, this gave *-si-s; syncope gave -s
with umlaut, e.g. -uis; -ais became -es.

Another form, which Rix calls pertinentivus, but most _dative_,
originated by adding the locative ending -i to the genitive; so -s-i,
but -la-i > -le.

These last two forms (ablative and dative) are easily understood if
the genitives in -s and -l were originally adjectives (so not "of X"
but "X-ish" [Du. "dus niet 'van de school' maar 'schools'"]).  Cases
of cases are not unusual, especially with the genitive (e.g. in the
Caucasus).

The _locative_-ending was -i.  With -a this gave -ai > -ei > -e.  -thi
and t(e) were postpositions, that could be added after the ending -i,
e.g. -aithi > -ethi.

The plural was marked by an -r after the stem; ais pl. ais-e-r "gods".
After that came the same endings as in the singular.  Note that before
the genitive -s, an -a- appears; this probably belonged with the r, so
-r < [*]-ra.

<end quote>

>>The locative in -i (for a-stems: *-a-i > -e) could optionally be
>>extended with the postposition -thi (-ethi < *-a-i-thi).

>This explanation of -thi makes no sense. The notion that any spoken language
>can afford the luxury of "optional" syllabic morphemes (i.e. arbitrary and
>non-functional) is absurd.

Is it?

> The notion that inscriptions, which generally
>economize words and letters, would bother representing superfluous morphemes
>is even more absurd. Whenever functionless morphemes are proposed, it's a
>good bet that the proponents don't have a functional grasp of the language
>they are analyzing, and their resulting grammar will be dysfunctional.

>The noun <zilc> 'type of office, zilacate' offers a clear example of
>contrast in usage between locative <zilcti> and comitative <zilci>:

>(1) zilachnu ciz zilcti purts'vavcti 'served as zilac thrice in the
>purtsva-zilacate' (i.e. he served in *his own* zilacate)
>(2) zilci vel[u]s[i] hulchniesi 'during the zilacate of Vel Hulchnies'
>(comitative absolute indicating the *year* in which Larth Velchas
>consecrated offerings in the Velcha family-tomb)
>(3) zilci larthal cusus' titinal larisalc salinis' aulesla 'during the
>zilacate of Larth Cusu son of Titinei and Laris Salini son of Aule' (com.
>abs. indicating year; dependent construction has replaced earlier
>hemiparataxis of entire phrase in com. abs.)

I wouldn't call that a "comitative".  It's simply a generalized
locative (here in a temporal sense).  It's quite possible that the
"optional" postposition -thi was added to emphasize a _local_ locative
("in", not "during", "on" etc.)

>>So I would analyze <Holaiesi phokias'iale> as:

>>Dat. holaie-si-i > Holaiesi "For Holaie"
>>Gen. phokia-s(i)  "of Phokaia" + dat. phokia-si-ala-i > phokias'iale
>>"for the Phokaian", with palatalization of -si- (> -s finally) when
>>followed by the genitive suffix -ala-, and -ai > -e, as in (later)
>>Etruscan.  Cf. Vanalas'ial, which is a double genitive: "of (that) of
>>*Vanala" [or an ablative "from *Vanala", although in Etruscan we only
>>have *-(a)lasi > -(a)ls, not *-si(a)la].

>I see we are now using opposite sibilant conventions.

Yes.  I'm sorry, I didn't explicitly announce that change.  It's
partially laziness (sigma appears only <evistho>, <-asial(e)> and
<sialchv(e)is'>), partially a theory about actual pronunciation (given
that <sialchveis'> matches Etruscan <s'a>, <s'ealch>).

>I prefer to use <s>
>for sigma, <s'> for zig-zag. Since Lemnian employs zig-zag for the genitive,
>my transcription is equivalent to traditional North Etruscan with gen. in
><s'>.

This can actually be taken as another argument in favour of *-si (-i
was dropped in S. Etruscan, but palatalized the sibilant to -s' in N.
Etruscan).

>(This has no significance for theories about origins.) It is important
>to pay attention to the sibilants. The one in <vanalasial> and <phokiasiale>
>is *not* the sibilant found in the gen. and dat. suffixes. Both frontal and
>lateral inscriptions on the stele are fully consistent in distinguishing the
>sibilants. Palatalization cannot be invoked, as both <sialchveis'> and
><s'ias'i> occur on the stele. Therefore, the -si- in these two words is not
>inflectional but derivational. I agree that <holaies'i phokiasiale> is the
>dative of 'Holaie the Phocaean', but I don't regard -sial and -siale as
>multiple case-suffixes; -si (as opposed to -s'i) is not an inflection.

If *-si-al(a) (double genitive) was common enough, it might have been
palatalized to /-Sial(a)/ even where this would not normally have been
the case (anyway, I'm half inclined to read <s'ias'i> as <[av]is'
ais'>).

>>In Etruscan there are certainly cases that remind one of ablaut.  Take
>>the root <tev-> "to show, (to put?)", which appears as <tv-> in the
>>mirror-inscription: "eca sren tva ichnac hercle unial clan thra sce"
>>(this image shows how Hercules Juno's [adopted?] son [became?]").
>>>From the same root we have <tevarath> "referee, judge", and maybe in
>>Lemnian the two words <toverona[rom]> and <tavarsio>.  That would make
>>sense if "Sivai"'s function was indeed that of "judge" (evistho < Grk.
>>eu-histo:r [?])

>This mirror-inscription is written on a rectangular tabella held over Juno's
>head by Jove. It contains five lines of exactly seven letters each (<ch> and
><th> of course are single letters). I suspect the forms <tva>, <thra>, and
><sce> are shortened from *teva, *thura, and *sece in order to fit the
>message into the enclosed space (or the 5x7 scheme, which may have some
>obscure significance) with minimal disruption. I believe your translation is
>essentially correct. I would connect *thura with the suffix -thur(a) 'member
>of a family, religious brotherhood, etc.'

I've entertained the thought of reading clan:thra as one word (despite
the dots given in the transcription [which I can't make out on my copy
of the inscription]), "adoptive son" (cf. Lat. mater-tera, matr-aster
?), leaving only <s-ce> as "he became" (and of course a hint of PIE
*h1(e)s- "to be").

=======================
Miguel Carrasquer Vidal
mcv at wxs.nl