--- In
cybalist@yahoogroups.com, "Arnaud Fournet"
<fournet.arnaud@...> wrote:
>
> ----- Original Message -----
> From: "dgkilday57" <dgkilday57@...>
> >
> > The problem I have with that is that Old European river names
are so
> > uniform over such a large area that they must have been spread
by a
> > water-people, and the IIr.'s are land-people.
>
> A subset of these landlubbers, Iranian-speakers, are known to have
> ranged from the Arctic Ocean to the Indian Ocean,
> DGK
> ==================
> What supports the claim that Iranian-speakers are attested as
north as the
> Artic Ocean ?
> Arnaud
> ========
Whoops. I have no evidence indicating they actually bathed in the
Arctic Ocean. However, Iranian loanwords in Finno-Ugric suggest
that they got reasonably close. The point is that "landlubbers" can
expand over a vast area under the right conditions.
> Also, I disagree about the alleged uniformity of the river-names
> assigned to the Alteuropäisch system. Apart from later names
(mostly
> Celtic and Baltic) included by mistake, Alteuropäisch has river-
names
> of both Veneto-Illyrian and Indo-Iranian types. In the Northwest
of
> Europe, I see a core of Veneto-Illyrian with an intermittent
overlay
> of Indo-Iranian. In the Southwest, Indo-Iranian forms the oldest
> Indo-European stratum, overlain in turn by Lusitanian (which I
> consider to be a Veneto-Illyrian language) and Celtiberian.
> ========
> How do you describe Veneto-Illyrian ?
> How to recognize a Veneto-Illyrian word ?
> Arnaud
> =======
Thanks very much for asking this question. Like Pokorny and
Hubschmid, I had blithely assumed that the attested Venetic language
of the Po Valley belonged with Illyrian in a "Veneto-Illyrian"
branch of Indo-European. Careful examination of the Venetic corpus
has now shown me that Venetic cannot be shoehorned into the same IE
branch as Illyrian. However, the affinities between Lusitanian and
Messapic are too striking to ignore, and the connections between
Messapic itself and the poorly known languages of Illyria proper are
clear. Therefore I must postulate an Illyro-Lusitanian branch of IE
which excludes Venetic, as explained below.
The principal Illyro-Lusitanian diagnostics are gemination of
postvocalic consonants followed by semivowels, accent generally
recessive but falling on the last syllable of the first part of a
compound noun, and monophthongization of long diphthongs. Within
this branch of Indo-European, Lusitanian retains inherited /o/ and
final /m/, while the Illyrian group changes them to /a/ and /n/.
The latter group includes Q-languages and P-languages based on the
treatment of labiovelars. Messapic and Japygian are Q-Illyrian
languages, while Macedonian in my view is a P-Illyrian language.
All these languages have merged the inherited voiced aspirates with
the simple voiced stops, and each shows some idiosyncratic secondary
developments with its consonants.
M. Durante, in "Il nome di Pesaro e l'accento iniziale in Illirico"
(AION-L 1:35-45, 1959), reconstructs a South Illyrian *luk{w}o-
'wolf' on the basis of the Paeonian (i.e. North Macedonian)
personal names <Lukkeios> and <Luppeios>, which he contrasts with
Proto-Albanian *ulk{w}o- (Alb. <ujk> 'wolf'). He regards the
etymology of the Paeonian names as certain, on the grounds that
<Lupus> and <Lupa> are common personal names in Latin inscriptions
of the Balkan peninsula, and a labiovelar can go either way. I
note, however, that a simple labiovelar would normally have
produced -k- or -p-, if not retained as -k{w}-. The (South)
Illyrian for 'wolf' must have been *lukk{w}o-, with the gemination
occurring either in Illyrian or in its parent Indo-European
language, since the PIE form itself is reconstructed as *wl.kwo-. A
similar gemination occurred independently in some Italian and
insular dialects of Late Latin. Probus prescribes "aqua non acqua",
and the futility of his prescriptivism is shown by Modern Italian
<acqua>, <acquidotto>, etc. We also have Logudorese <abba>, with
Nuorese <abbatiza> and Log. <abbaduza> from Lat. <aquatilia>, Log.
<abbagotta> 'glue' i.e. 'aqua cocta', Corsican <strakkwu> 'stranded
goods' i.e. '(bona) extra aquam', Log. <abbardzu>, Sicilian
<akkwaloru>, and Apulian <akkwara> from Lat. <aqua:rium>, <-a>, and
Sic. <akkwattsu> from Lat. <aqua:tio:> (REW 570-588).
Attic Greek <híppos> 'horse' can be regarded as an inherited reflex
of PIE *H1ek^wo- only through strained ad-hoc assumptions. It forms
both first and second elements of compound personal names. In
Macedonian, <-ippos> occurs as a second element; <Bílippos> is
Hellenized as <Phílippos>. In Epidaurian and Tarentine Greek proper
names, <-ikkos> is found. As with the Paeonian names, it is evident
that we are dealing with Q-Illyrian and P-Illyrian material. Since
Illyrian is a centum group (pace Hamp et al.), it makes no
distinction between inherited */k/ and */k^/, and reflects */k^w/
the same as */kw/, producing a geminate in intervocalic position. P-
Illyrian has /pp/, and Q-Illyrian including the Japygian extension
into Italy (with Messapic) has /kk/, Tarentine having borrowed <-
ikkos> from Messapic. Mycenaean Greek has no compound personal
names formed from 'horse', indicating this word's novelty, and the
attested forms, nom. sg. <i-qo> and gen. sg. <i-qo-jo>, show that a
labiovelar was still in use (/(h)ikk{w}os/, /(h)ikk{w}oyo/, cf. Epic
Grk. gen. <híppoio>) in the Illyrian dialect from which Mycenaean
borrowed. I see no reason to suppose that Attic-Epic <híppos>
continues the Mycenaean form, with */kk{w}/ developing into /pp/ in
Proto-Attic. More likely the Attic form is borrowed directly from
Macedonian or other P-Illyrian. The inherited Greek root likely
appears in <Epeiós> (builder of the Trojan Horse, Il. 23:665, Od.
8:493, etc.) from an extinct *epos 'horse'.
Labiovelars in word-initial position or preceded by other consonants
evidently did not undergo gemination. J. Whatmough (_Prae-Italic
Dialects_ v. 2, p. 606, 1933) regards the Messapic onomastic forms
<Penkeos> and <Penkaheh[e]> as derived from *penkwe 'five' (the
forms are genn. sgg. of an /e:w/-stem and an /ayo/-stem). Latin
<gurdus> can be explained as a loan from Lusitanian, reflecting
*gwr.H-do- (see Cyb. #59524).
J.U. Hubschmied (Vox Rom. 1:102-3) refers Lower Engadine
<gioc> 'juniper' to an Illyrian *yukk{w}o-. A protoform *yuppo- is
required by Old Upper Engadine <giop> 'id.' (see J. Hubschmid,
_Praeromanica_, p.32, 1949). Both these and some awkwardly related
words for 'juniper' in other languages can be explained if we
postulate a Western IE *yukw- which was extended in Celtic as *yuk{w}
i-, in Illyro-Lusitanian as *yukk{w}o- yielding Q-Ill. *yukko- and P-
Ill. *yuppo-, both represented in the Illyrian substrate of the East
Alps. Ligurian <jupikellos> (Diosc.) must have its first element
borrowed from Gaulish *yupi-, and the modern Alpine toponym
<Giubiasco> likewise has a Ligurian suffix on a Gaulish base, since
Ligurian did not labialize inherited */kw/. Latin <ju:niperus>, <-
i:> (f.) has evidently been altered to conform to other tree-names
from *ju:nipera 'juniper-berry-producing', from *ju:num 'juniper-
berry', from Old Lat. *juxnom < *yukw-snom.
Despite its retention of inherited /o/ and final /m/, Lusitanian
shows striking affinities to Messapic. In the inscription of Cabeço
das Fráguas, <Iccona> appears to be the name of the goddess
equivalent to the Gaulish Epona, in the dative case. Thus
Lusitanian has *ikko- from PIE *H1ek^wo- 'horse', just as Messapic
and other Q-Illyrian have *(h)ikko-. If we assume that PIE *H1e-
was reflected in Illyro-Lusitanian as *hi-, we can explain not
only 'horse' but several other words in Messapic and Lusitanian.
Mess. <hipades> is equivalent in sense to Grk. <epéthe:ke> 'placed'
(something) 'upon' (someone), in practice 'dedicated'
(something) 'to' (a deity). Mess. <hipa-> corresponds to Grk. <epi-
> if both reflect *H1ep- with different grades and extensions, <hipa-
> reflecting *H1epo- and <epi-> *H1pi-. The simple Mess. aorist
<des> 'placed', presumably /de:s/, lacks the Greek augment and
kappatic extension of <éthe:ke>, corresponding instead in form to
Epic <bê> 'went' (with augment <ébe:>), where Greek has lost the
3sg. secondary ending -t, and Messapic has apparently substituted
the primary ending -ti with assibilation to -s (for this subst. cf.
Lat. <fe:cit>, <siet> (Cato), Old Lat. <fe:ced>, <sied> with sec. -
d). Loss of initial /h/ in Lusitanian is trivial, and <ipadem> (my
correction for the v.l. <ifadem> which makes no principled sense)
can be understood as the 1sg. of the same verb, 'I dedicated' (see
my analysis of the CdF insc. in Cyb. #58889). Lus.
<indi> 'thereupon, and then' in my view represents Ill.-Lus. *hindi
from PIE *H1en-dhi, cognate with Lat. <inde>. Messapic also
has /a:/-stem datives in <-a> (presumably /-a:/ from */-a:i/, cf.
Whatmough, op. cit. p. 600), e.g. <Aprodita> in several dedicatory
inscriptions. Furthermore, Messapic regularly produces a geminate
when a consonant is followed by the palatal approximant [y] plus a
vowel; thus Mess. <Orra> on coins, Lat. <Uria>, Grk. <Ouría>, and
various gentilicia in -CCes from *-Cyos, e.g. <Platorres> nom. sg.
from *Platorios, <Blatthes> from *Blatios (Whatmough, op. cit., p.
603). In the CdF text we have <loiminna> following <Iccona>,
evidently an adjective in the fem. dat. sg. agreeing with it, very
plausibly from *loiminya:i, and <labbo> (the v.l. <laebo> is
unlikely since the following word is spelled <comaiam>) can be
understood as a neut. dat. sg. with instrumental force, earlier
*labyo:i.
I take the Lus. forms in -ai, -oi from the text of Lamas de Moledo
as nomm. pll. with short diphthongs, not datt. sgg. with long ones.
Lusitanian generalizes /o/ as the connecting vowel of compounds as
we see in <Trebopala> dat. sg. 'to the Protector of the Village' vel
sim. and *Toutopalanta:ikos 'of the People of the Palanta (river)'
extracted from Latinized Celtiberian <Toudopalandaigae> dat. sg.
(see Cyb. #58889); this vowel is apparently elided in the dat. sg.
<Trebarune>. Latin <andabata> 'gladiator who fought with an eyeless
helmet' has been etymologized as borrowed from a Gaulish compound
*andobatta 'blind beater' but the alteration is unexplained. More
likely this word is from Messapic with secondary /a/ as the
connecting vowel. The word could have reached Latin through Old
Oscan, before the assimilation of -nd- to -nn- occurred in Oscan.
Gladiatorial games came to the Romans from the Samnites.
Illyrian accent is inferred primarily from place-names and river-
names whose modern forms require an ancient accent more recessive
than that of the Latin penultimate law. In Illyria proper, Albanian
<Durrës> reflects <Dýrrachium> (although Italian <Durázzo> and
Serbian <Drac^> require a normalized <Dyrráchium>), <Drisht(i)>
reflects <Drívastum>, and <Ishm> reflects <Ísamnus>. In Messapia,
<Bríndisi> requires <Brúndisium>, <Bréntesion>, or the like (several
ancient forms occur, but the native name must have had initial
accent). In Picenum, <Pésaro> requires <Písaurum>. Durante (op.
cit.) argues that the correct name of the river upon which Pesaro
sits (now the Foglia) was <Ísaurus>, with <Písaurum> a prepositional
compound using the zero-grade *H1pi-, Ill. *pi-. He cites
<Parupion> (Geog. Rav.) and <Pasinum> (Plin.) as parallel examples,
invoking elision of *pi- to *p'-, but the Ill. combining form 'upon-
, on-' could just as well have been *p- from the unextended zero-
grade *H1p-. <Terni>, <Téramo>, and <Térmoli> require <Intéramn(i)
a>, showing that Illyrian accent was not strictly word-initial, but
immediately preceded the second element of nominal compounds. The
actual Ill. form was likely *Antérapna from *n.ter-ap-na: '(town)
between waters', the /in/ and /mn/ being Latinizations.
Turning now to Venetic we find that inherited long diphthongs are
preserved as diphthongs in dat. sg. endings, e.g. <Vhrutana.i.> 'to
Frutana', <Gra.i.ko.i.> 'to Graikos'. Inherited */kw/ remains as
such in <-kve> 'and'. Forms like <Vhugia> and <Re.i.tiia.i.> show
no evidence of /y/-induced gemination. Intervocalic */bh/ becomes -
b-, but initial */bh/ becomes f- (usually written <vh>) as in
Latin. (Indeed, Venetic treatment of voiced aspirates is so similar
to that of Latin, which is peculiar within Italic, that I am tempted
to posit a Venetic substrate for Latin. This is supported by the
names <Praeneste> and <Venetulani> (Plin.) as well as the use
of 'free ones' in the sense 'children' (Lat. <li:beri:>, Ven. dat.
pl. <louderobos>). But I digress.) The accent of <Triéste>
requires ancient <Tergéste>, and <Odérzo> requires <Opitérgium>, in
contrast to the accent of the Illyrian type (*Térgeste,
*Opítergium). Both of these are native Venetic formations based on
*terg- 'marketplace', which has been borrowed into South Slavic
(Serbian <trg> 'town square'). In Venetic inscriptions, the noun
usually written <ekupetaris> (once <ekvopetars>) 'funerary monument,
commemorative gift, memorial' vel sim. has a syncopated form <ep(p)
etaris>, indicating a secondary accent on the first, not last
syllable of the first element of a compound in Venetic. Thus
Venetic has too many differences from Illyro-Lusitanian languages to
be included in the group. Now, while Venetian <Trevíso> points to
<Tarvísium>, Friulian <Tárvis> requires <Tárvisium>. Since this is
formed like <Brúndisium>, it appears to be an Illyrian toponym whose
accent was preserved in the East Alps, but not by the Veneti who
settled on the plain later. Another indication of the priority of
Illyrian here is the river-name <Plavis>, now <Piave>. Venetic
maintains inherited /o/ and the river would have been *Plovis (from
the /o/-grade of *pleu-) if the Veneti had named it.
Latin <accipenser> 'sturgeon' (which has several variants) can be
understood with some difficulty as an indirect loan from Illyrian
through Venetic. Venetian-Paduan-Triestine <kópeze> shows that the
original accent was on the second syllable, while Vicentine-Fiuman
<kopéze> reflects a normalized Latin form with penultimate accent.
Meyer-Lübke takes the correct form as <acupe:(n)ser> (REW 129).
However, the Hellenized form <akkipé:sios> indicates that the
geminate was real. The original accent is consistent with an
Illyrian compound; borrowed into Venetic, the word would not have
been recognized as a compound, and antepenultimate accent would have
been acceptable, as in <Tarvísium>. I take the first element as
Ill. *akka 'river' from *akwa:, the second as *penser(i)s vel sim.
of undetermined origin but referring to some large animal, so that
the sturgeon was called 'river-ox' or 'river-beast' or something
similar. We might expect the connecting vowel to be secondary /a/,
since Lusitanian as we have seen generalizes /o/. The Romance forms
however demand *akku-penser(i)s. It appears that in this position
the Japygian dialect at the head of the Adriatic had /u/ before a
labial rather than secondary /a/. If my view of <andabata> is
correct, Messapic had /a/ even before a labial. H. Krahe (IF
58:143) similarly observed that Illyrian names sometimes have /u/
before a labial when secondary /a/ from */o/ is to be expected.
> In developing his theory of Alteuropäisch, Hans Krahe focused too
> much attention on suffixes, too little attention on the relation
> between root-grade and suffix, and far too little attention on the
> semantics of the complete names. His successors have offered some
> partial remediation, but to my knowledge no revision of the theory
> has been published which addresses all the problems. What I
propose
> is, of course, a radical revision. Resulting from the conflation
of
> distinct strata, the Alteuropäisch system as such is a mirage, not
> the product of a uniform Proto-Western IE as Krahe thought, much
less
> of PIE itself as W.P. Schmid insisted.
> =======
> Why not ?
> I would rather agree with the revision,
> but what supports it ?
> Arnaud
> =====
Work in progress ...
DGK