On Mon, 9 Dec 2002 14:56:17 -0800 (PST), Juha Savolainen
<
juhavs@...> wrote:
>First, it seems to me that you basically agree that
>the study of isoglosses shared by Indo-European
>languages that are each others neighbours
>geographically (and presumably were close neighbours
>in the past) can provide us valuable information about
>the dispersal history of the Indo-European languages
>and hence probably about their common home as well.
Sure.
>Second, while you seem to accept the idea of using
>shared isoglosses as a way to go forward, you do not
>seem to be very impressed by the choice of these
>purported isoglosses (whether by Hock or Elst): the
>isoglosses suitable for arguing one´s case here
>(either way) must not be mere convergencies.
>
>Summing up, I take your comments as saying that yes,
>you can study the dispersal history by means of
>isoglosses shared by geographical neighbours, but the
>discussion cited by you (i.e. by JS) does not
>accomplish much in this respect
Did I get your
>meaning right here?
I thought some of the things you mentioned deserved some comment.
The fact is that (unfortunately) linguistics is not an exact science,
and the same facts can be interpreted in different ways by different
people. Also, things are rarely as simple as they seem.
On the other hand, we would prefer them to be as simple as possible.
The study of language change is one of trying to trace back the
transformations that led to the attested languages A, B, C, ... from
an unknown common proto-language X. The problem is that,
mathematically, X can be *anything*, because there will always be an
infinite set of [sets of] transformations that lead from any X to A,
B, C etc. [this can easily be verified by realizing that even
something as simple as *a > a has an infinite number of expansions: *a
-> a; *a -> o -> a; *a -> o -> a -> o -> a; etc.]. The trick is
finding the X that leads to A, B, C... in the smallest number of steps
and with the smallest amount of _implausible_ steps (there are
probably some additional requirements, e.g. maximizing the amount of
inherited material: we can postulate an X with features F1 .. Fn,
where *all* the features were lost and replaced by new features in the
daughter languages A, B, C ..., but such a reconstruction would not be
very useful).
>Mallory sets five principles which are required of any
>solution to the IE homeland problem for it to be at
>least a plausible solution (explanation for less
>obvious principles in commas)
>
>(1) Temporal-spatial plausibility
Of course. PIE must be younger than say 10,000 years, and it's
definitely older than 4,000 years (3500 years ago Hittite, Greek and
Sanskrit were different enough to warrant at least half a millennium
of separation, probably much more). Spatial plausability excludes I
suppose the Americas, the Arctic, SE Asia or Sub-Saharan Africa.
>(2) Exclusion principle (it is unlikely that the IE
>homeland lay in a territory already occupied by a
>non-IE language)
This is more doubtful. Perhaps it's unlikely, but it's not
impossible.
>(3) Relationship principle (the IE homeland solution
>must accommodate the inter-group relationships of the
>IE family)
Yes. But there are different possible interpretations here,
obviously, and different temporal layers of mutual influence can have
occurred. In principle, however, there is hope that most of it can be
sorted out eventually.
>(4) Total distribution principle (the solution to the
>IE home problem must explain all the languages
>belonging to the IE family)
Of course.
>(5) Archaeological plausibility
A more doubtful one, again. Pots and graves just don't tell us what
people spoke. It would be nice if the expansions of the IE langauges
could be linked to the spread of archaeological horizons or
technological advances, but it's hard to come up with proof positive.
>Do you accept Mallory´s set of five principles? And if
>you do, what would be your best guesstimate for the IE
>homeland?
My best guesstimate for the homeland is the Balkan peninsula,
6500-5500 BC (calibrated). The Linear Ware (LBK) expansion into the
Northern European Lowlands (from Hungary to Denmark, from Holland to
Poland, and further into the Ukranian steppelands), marks the
separation of Anatolian (stay-behinds in the Balkans) from the rest of
PIE (the Linear Pottery folk). Next to break away was Pre-Tocharian
(eastwards into the steppes), and then, more gradually, peripheral
groups began to differentiate to the north (Germanic) and south
(Armenian) of the central area (from where Italo-Celtic,
Balto-Slavic(-Albanian), Greek and Indo-Iranian expanded subsequently,
after 3500 BC).
=======================
Miguel Carrasquer Vidal
mcv@...