I would also like to share an article which may
look quite peculiar. But why not to read it and to examine its author's
arguments? I mean a John Dunn's work on Proto-Tsimshian (a West Coast
Indian/Native American language, sometimes mentioned within the Penutian
macrophylum). The author (of Univ. of Oklahoma) states that the Tsimshian(ic)
family developed as a result of an old, unknown pre-Indo-European migration to
the New World long before Columbus' time, and that proto-Tsimshian was a sister
language of PIE (and especially related to the Tocharian branch). He also
dicusses (in the PS of the paper) some archaelogical correlates that make his
hypothesis more plausible.
The paper was available on Dunn's website and
freely available once - but now his website has been closed. Fortunately I
managed to download the paper (publicated as a set of images), and now I can
show it here. The pdf is made by me personally of those images, and I hope it is
enough readible.
Here you are the Dunn's paper on IE-Tsimshian
possible connection:
Please excuse me for some personal thoughts on
Tsimshian languages. I know that this is not a discussion group on Amerindian
languages :-) (but if Dunn was right, Tsimshian might count as one of a
sub-group of Indo-European sensu lato). In my humble opinion the presented
material that contains ca. 150 IE-like Tsimshian roots
may show results of PIE borrowings into a Penutian language which was
proto-Tsimshian (PT) (it means: I doubt if PT is closely genetically
related to PIE). But even then a PIE migration should be regarded in order to
have the Tsimshian-IE similarities explained.
Some proposed etymologies are enough interesting.
For example PT *tal talk (to), praise, answer = PIE *del- recount (to)
(Dunn). Should not rather be talk for PIE as well? But Eng. "tell" (as well as
"talk") seems to be derived from this root. (IMHO also the IE word for
"language/tongue" is also based on the same root, and the reconstruction like
*dng'uH- is just false, should be *dlng'uH-. This initial *dl- cluster
yielded zero in Slavic, hence *eNzy- (for both tongue and language). It was
spelt as D in archaic Latin, hence DINGVA, and then it merged with L, hence
"lingua" in the classical language. In Armenian this *dl- yielded *tl-
regularily, and then *tr- > *rt-, and finally art- (attested). Besides, *del-
-like roots are present in Nostratic languages (cf. in Turkish) and denote the
act of speaking or a language. But this all is another story.
Another thought is on laryngeals. They all have
been preserved well in PT, but we can see more laryngeals there than in PIE. It
means that proto-IE had even more laryngeals than the reconstructed PIE (it is
my deduction, based on the assumption that there really was a proto-IE migration
into Tsimshianic Coast). Dunn compares TS *peHl tear (to), break in two with a
laryngeal, and PIE *spel- split (to), break off without a laryngeal. The s-
here seems to be movable: Polish po'l/ half < Slavic *polU may be derived
of the same root. A linguistic puzzle is Finnish puoli half (< *po:l-). If
it had been a Slavic borrowing, it would not have had the long
vowel. So, maybe this word is older? It may be an Uralic cognate of the PIE term
(a common heredity from Nostratic) or a borrowing from a very old stage of PIE,
when more laryngeals were still present that may be reconstructible now with the
normal way. If it was so (and if Dunn is right), both the presence of the
long vowel in proto-Finnish and of the laryngeal in proto-Tsimshian would be
explained.
I am not a fanatic Dunn's follower and I
analyse the matter as a possibility. I just consider if he may be right (and if
yes, to what degree). Telling the truth, as for me, Tsimshianic languages are
very odd even without their possible IE connections. They use complex
consonantal clusters, they have odd ejective-nasal consonants (AFAIK confusing
phoneticians). Besides, many years ago Tadeusz Milewski, my native
(so, Polish), stated that some Tsimshianic languages have the
highest typological peculiarities. His work is little known in so called
western countries, perhaps because he did not publish it in English. Now authors
of world typological literature do not even mention him, and this is a pity,
really (in the same time he is widely quoted in Polish
literature...).
In short, he cannot have agreed with the dual
division into accusative and ergative languages (tripartite and active lngs were
little known in his time). Instead, he proposed a division into six classes (now
completely forgotten without having been discussed first). Namely, he noticed
that the basic roles of intransitive subject, agent and patient may
be played not only by nominative, accusative, ergative and absolutive, but
also by genitive. Hopi and Indonesian are known of not having accusative and
using genitive instead (for patient; btw. Slavic lngs use genitive instead of
accusative in many constructions - it would be worth discussing if this is an
original IE feature or a Slavic invention). Eskimo languages merged ergative and
genitive instead (so, there genitive is used for denoting agent). But among
Tsimshianic languages are the most rare ones. In Nisga'a (aka Nass aka Nishga
aka Nass River dialect), at least in some construction, nominative merged with
genitive, and this is the only language with this peculiarity all over the
world (genitive, beside for the nominal attribute, can be used for any subject,
both intransitive and transitive (=agent)). (In fact, Nisga'a uses also
other constructions without genitive, so it is not a "clean" typological
oddity.) Which is even more astonishing, the genitive formant (aka marker)
is -L (the voiceless lateral spirant) which is the
etymological counterpart of IE *s according to Dunn.
Just think, both nominative and genitive
(singular) have (may have) the -s (-os) ending (esp. in Hittite). And what
if it was really one form in the beginning, and only then it differentiated
(among others, by stress moving)? If yes, both the -L ending in Nisga'a, and the
typological oddity of this language, may have IE roots... Was the proto-IE
language a similar typological oddity as well?
(The last, 6th typological group in Milewski's
classification, have genitive used as absolutive, i.e. for
both intransitive subject and patient. Tsimshianic languages other
than Nisga'a may belong here, as well as various (Amer)Indian languages, esp.
those from South America.)
Please do not blame for this post, perhaps some
people will acknowledge the discussed problems interested.
Grzegorz Jagodziński