Re: Substrates in Latin and Germanic [was: The reason for Caesar's o

From: Torsten
Message: 68712
Date: 2012-03-02

--- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy <bhrihstlobhrouzghdhroy@...> wrote:
>
> 2012/3/1, Torsten <tgpedersen@...>:
>
> > You misunderstand. I haven't claimed to have disproven the
> > standard theory of a direct decence of the 'mots populaires' from
> > PIE, only that I *prefer* to explain them as loanwords. This I do
> > because of the greater explaining power of that theory since it
> > explains the systematically (not random) skewed distribution of
> > the semantics of those words.

> Both hereditary and substrate etymology can make use of
> 11,449,000 billions of Indo-European words (resulting from 2140
> roots, 100 suffixes, five ablaut grades, and word-composition
> rules). If a language derives from an earlier phase, it has always a
> diachronic phonology: this is what hereditary etymology can utilize
> without positing anything else.
> Substrate etymology has to posit an additional diachronic
> phonology. Since both hereditary and substrate etymology are ALWAYS
> possible (provided they are lexically and morphologically correct
> and phonologically coherent),

Fail. You assume here that a given word always can be given derivations in both the basic and the donor language. However, donor languages are only posited if there is a set of words in the given language which can not be explained (reasonably, in casu without zero-grade h2's) by a derivation in the basic language.

> this additional phonology is based on
> etymologies that are NEVER compelling (note: complelling; maybe they
> are indeed true, but not compelling) because they have always
> another equally possible hereditary etymology beside them.

That is exactly what they don't have.

> For this reason only, substrate etymology, although more than
> possible, is always flanked by hereditary etymology. They are
> equally powerful, the latter is more economic

Your flawed premise renders your conclusion invalid.

> 2012/3/1, Torsten <tgpedersen@...>:
> >> Note that I'm not excluding that they are loans, I'm just
> >> claiming that the hereditary hypothesis is at least at the same
> >> level of probability
> >
> > Except for explaining the skewed distribution of their semantics.
> >

> As for semantics, You can pick any semantic group and find out
> which phoneme is prevailing and then attribute the origin of all
> other occurrences of that phoneme to a substrate.

Yes I can, if I can find such a prevailing phoneme.

> 'Popular' is too vague a concept; moreover, with the addition of
> 'religious'! What is left then?

There are things in this world which are neither popular nor religious.

> You have simply taken Latin lexicon and attributed to Coastal
> Ausonian.

No, I have taken the 'mots populaires' with roots in -a- (and -ae- and -au-) and attributed them to Opscan/Ausonian. Since that language surely had other vowels than -a-, many other, but not all Latin words may be Plebeian loans from that language.

> 2012/3/1, Torsten <tgpedersen@...>:
> >> and morevoer doesn't have to postulate a substrate presence in
> >> Rome (not otherwise documented except for these controversial
> >> words).
> >
> > Not true. I wish linguists would be more aware of ancient sources
> > and of archaeology.

> I can't understand Your citations. What have they to do with the
> documentation of substrates in Latin *except these very words*?

I'm not sure what your question is about. As far as I can see, documenting a substrate is done on a word-by-word basis.


> 2012/3/1, Torsten <tgpedersen@...>:>
> http://tech.groups.yahoo.com/group/cybalist/message/29491?var=0&l=1
> > BTW I think Venetic had sg. -sk-/ pl. -st- alternation (cf eg.
> > Polish)
http://tech.groups.yahoo.com/group/cybalist/message/59166?var=0&l=1
http://tech.groups.yahoo.com/group/cybalist/message/66676?var=0&l=1
http://tech.groups.yahoo.com/group/cybalist/message/66683?var=0&l=1
> > The source for the Roman Plebeian 'a-language' would then be the
> > Opici/Ausones
> > http://en.wikipedia.org/wiki/Opici
> > http://en.wikipedia.org/wiki/Ausones
> > (from *op-/*ow- "mouth of a river", cf
> > http://en.wikipedia.org/wiki/Ostia_Antica
> > http://en.wikipedia.org/wiki/Osismi
> > http://en.wikipedia.org/wiki/Aesti
> > http://en.wikipedia.org/wiki/Swinoujscie
> > http://en.wikipedia.org/wiki/Aarhus#Name
> > cf
> > http://www.angelfire.com/rant/tgpedersen/Op.html
> > Semitic A-p- "mouth of a river"
> > ) ->
> > http://en.wikipedia.org/wiki/Sabines
> > ->
> > http://en.wikipedia.org/wiki/Plebeians

> This is an unduly expanded version of Meillet's and Peruzzi's
> ideas.

Which are? I'm afraid I haven't heard them.

> Your Ausonians are Latin, body and soul.

I'm sure the Romans eventually convinced them of that. It took some time for them to catch on to the idea.
http://en.wikipedia.org/wiki/Ausones
So, no.

> What have gained in splitting Latin into two languages?

Apart from annoying you? A better understanding of Roman origins
http://en.wikipedia.org/wiki/Patrician_(ancient_Rome)
http://en.wikipedia.org/wiki/Plebeians

> 2012/3/1, Torsten <tgpedersen@...>:
>
> >> Alas the very existence of such component people is product
> >> of a linguistic hypothesis and therefore cannot be the base for
> >> further arguments: it's simply one and the same argument - a good
> >> hypothesis, but not better than the hypothesis of the absence of
> >> the /a/-substrate of Latin
> >
> > No, see above; we have to assume the existence of that component
> > on historico-sociological grounds anyway, so Occam doesn't apply
> > here.
> >

> It's obvious that a popular component did exist in Rome. The
> problem is in linking it with /a/ < */o/

Where does that /a/ < */o/ rule enter the picture?


> 2012/3/1, Torsten <tgpedersen@...>:
>
> > I know that that is what orthodoxy teaches us, so imagine my
> > surprise when I checked for Celtic cognates of NWB words in p-,
> > eg.
http://www.angelfire.com/rant/tgpedersen/KuhnText/01paik-betr_gen.html
http://www.angelfire.com/rant/tgpedersen/KuhnText/02pal-steif.html
http://www.angelfire.com/rant/tgpedersen/KuhnText/03palm-fassen.html
http://www.angelfire.com/rant/tgpedersen/KuhnText/04palt-lappen.html
http://www.angelfire.com/rant/tgpedersen/KuhnText/05par-sichtbar.html
> > and check further for yourself here:
http://www.angelfire.com/rant/tgpedersen/KuhnText/list.html
> >

> Are You joking? Why should these Celtic words be other than
> Latin peccatum?

That's exactly the word the attribution of which to either Latin or NWB caused scores of postings here.

> Anyway, I'm deeply anti-Latinist in Celtic etymology (as You can
> easily imagine), so if You find that presumed Latinisms in Celtic
> are of IE origin, well done! Latin loan or NWB loan? Good
> alternative.
> I don't agree when You connect them with Germanic, because in
> that case You are denying a hereditary derivation from *beig'-

I can make sense of what you wrote if I emend it to
'I don't agree when You don't connect them with Germanic...'
since disconneting them from Germanic and its Grimm-shift is exactly what I do when I attribute them to NWB.
However, Kuhn seems to assume that the NWB loans in Germanic are post-Grimm; if they were pre-Grimm they would have been loaned in the form you cite for 'Indic'. Another indication of this may be the fact that the NWB words in p- identified by Kuhn often have variants in b-; a similar phenomenon appears in Jutland
http://tech.groups.yahoo.com/group/cybalist/message/30336

> (cp. Indic *bid.d.a-, *bed.d.a- 'defective', Turner 9238,

Reference? Turner?

> exactly from *big'-do- and *boig'-do-, the protoforms of Germanic
> *pik- and *paik-)


> 2012/3/1, Torsten <tgpedersen@...>:
>
> > Actually I think Old European is Venetic, in order to please
> > Occam.
> > http://en.wikipedia.org/wiki/Veneti_(Gaul)
> > http://en.wikipedia.org/wiki/Vendsyssel#Etymology
> > Note that they are a sea people, which explains the wide
> > distribution of Venetic/Old European hydronyms, cf the
> > distribution of Dutch hydronyms in North Europe (North Sea, etc).

> PIE *bh- > Venetic f- word-initially, Old European b-
cf Belgae/Villigst
http://tech.groups.yahoo.com/group/cybalist/message/60861?var=0&l=1
http://tech.groups.yahoo.com/group/cybalist/message/65302?var=0&l=1
> PIE */o/ > Venetic /o/, Old European (allegedly) /a/ (I think
> rather /a/ where local language has /a/ < /o/, otherwise /o/).
So no problem here.

> IMHO, as in Krahe's opinion, Old European is not a separate
> language, but a protohistorical phase of historical IE languages, so
> Ockham is still more pleased

You are entitled to your opinion. It entails the idea of a Europe once speaking a homogenous IE dialect.

> 2012/3/1, Torsten <tgpedersen@...>:
> >> In this particular case I have doubts that Daco-Misian had a
> >> Lautverschiebung, while I find quite convincing that Thracian had
> >> one (Georgiev, Duridanov).
> >
> > I haven't read them, you'd have to quote their arguments to
> > convince me. I don't remember seeing any signs of LV in the
> > glosses in Detschew's "Die thrakischen Sprachresten'.
>
> Since */bh/ */dh/ */gh/ > /b/ /d/ /g/ aren't diagnostic, I list
> only */b/ */d/ */g/ > /p/ /t/ /k/:
> Skalpe:nos (place name): lit. Skalbupis,
cf. the wildly varying *xalp-(?) "slave" word
http://tech.groups.yahoo.com/group/cybalist/message/66821?var=0&l=1
I wouldn't want to base too much on the distribution of b/p in that.

> Utus (river name) < *udo-s (*wed- 'water'),
A river called 'water'? Hmmm.

> Kikones < *gwig-on-es : quick
The quick ones? Why?

> and for */p/ > /ph/
> Ostaphos (place name) : *h2ap- 'water',
> rhomphaia (if Thracian) : Lat. rumpo

That's not much.


> 2012/3/1, Torsten <tgpedersen@...>:
> >> OK, very good. This is a good argument. Until we don't know
> >> something more about alternative etymologies and explanations of
> >> the name (I could propose some of them, but now it's really
> >> irrelevant, so I omit to do it), a simple ending -daua is a weak
> >> piece of evidence, but still it IS a piece of evidence.
> >
> > Sorry, you can't do that. The
> > http://en.wikipedia.org/wiki/Dava_(Dacian)
> > names have been used to separate Dacians from Thracians
...
> > so you can't give up the ethnic connotation in the case of
> > Setidava without losing the distinction between Dacians and
> > Thracians.

> Sorry, I haven't understood. What should I do and why?

http://en.wikipedia.org/wiki/Dacian_language#Thracian
'Georgiev was the first scholar to discover a linguistically significant toponymic fact: Daco-Moesian placenames generally end in -DAVA (variants: -daba, -deva: "town" or "stronghold"). But placenames in Thrace proper (i.e. South of the Haemus range - Balkan mountains) usually end in -PARA (variant: -pera: "village" or "settlement":[161] cf Hindi suffix -pur = "town" e.g. Udaipur),[original research?] or, in fewer cases, in -BRIA ("town") or -DIZA (or -dizos: "stronghold")[162][163] But Papazoglu (1978) and Tacheva (1997) reject the argument that such different placename-endings imply different languages[164][165] (although, in historical linguistics, changes in placename suffixes is generally regarded as strong evidence of changes in prevalent language).[166]'
http://en.wikipedia.org/wiki/Dacian_language#2nd_century_AD

If you give up the claim that -dava names are indicative of Dacian presence, you lose of of the main criteria for distinguishing linguistically between Dacian and Thracian settlements.


> 2012/3/1, Torsten <tgpedersen@...>:
> >> So, for the sake of the argument, let's state that (I'm
> >> quoting) "a Dacian outpost in North Central Europe" was in
> >> linguistic contact with Proto-Germans and that words could flow
> >> from there up to Scandinavia (not just into Scirian). This is a
> >> further, but possible, hypothesis.
> > Yes
http://tech.groups.yahoo.com/group/cybalist/message/67598?var=0&l=1
http://tech.groups.yahoo.com/group/cybalist/message/67603?var=0&l=1
> > cf
http://en.wikipedia.org/wiki/Dauciones
http://tech.groups.yahoo.com/group/cybalist/message/66976?var=0&l=1
> > who I think are better explained as *daN-k-io-.
> >
> > On arriving in Denmark, cf.
http://tech.groups.yahoo.com/group/cybalist/message/67106?var=0&l=1
http://tech.groups.yahoo.com/group/cybalist/message/66612?var=0&l=1
> >

> From what do You get they were Dacians?

From the name.

> They could very well be Germanic people as well!

By that time they probably spoke some Germanic dialect. Note however one common characteristic of North Germanic
http://en.wikipedia.org/wiki/Danish_tongue
and the presumed substrate of the Balkan Sprachbund
http://en.wikipedia.org/wiki/Old_Albanian#.28Old.29_Albanian
namely the suffixed definite article.


> 2012/3/1, Torsten <tgpedersen@...>:
> >> All of Kuhn's etymologies (which I like, by the way) are less
> >> than a single inscription. In order to be sure that a language
> >> has been spoken in a region, one needs inscriptions.
> >
> > I disagree. If a sufficiently large part of the vocabulary of a
> > language can't be matched with a standard derivation from its
> > supposed, you have to suppose a different ancestor. That goes for
> > mixed language as well as for mixed creatures.
> >
> But every part of Germanic and Latin vocabulary CAN be matched
> with a standard derivation from their parent language

No. That's why those substrates were proposed.

> (the text reads "from its supposed": what "supposed"? A parent
> language?).

Everything in historical linguistics is 'supposed'.


> There's ample room for substrates only in MEDIAEVAL and MODERN
> languages, especially Romance languages, Souther Slavonic, and
> English (and of course Hungarian, Turkish, Maltese and so on and so
> forth).

Nonsense. Substrates is not a recent thing.


> Most uncertain etymologies of Romance words are in fact easily
> traceable to PIE through local Old Celtic; similarly for Roumanian
> through Dacian and so on. Ancient IE languages are, on the contrary,
> always provided with PIE etymologies. Always. And regular ones, too.
> You can try to add alternative substrate etymologies, but You can
> never find anything better. Only at the same level

You can only maintain that position if you posit that the process of language genesis has changed fundamentally in the meanwhile.


Torsten