The Indo-European Hypothesis [was Re: The Neolithic Hypothesis]

[Larry Trask]

On Sat, 3 Apr 1999, Ray Hendon wrote:

> Thank you for responding so thoroughly and informatively to my
> poorly phrased and uninformed questions.  I was most reluctant to
> pose any question about language, given my utter ignorance in the
> field.  I certainly had no intention of sounding a challenge to
> those in the field of historical linguistics about the veracity or
> accuracy of IE.  Being professionally skeptical, however, I did want
> to consider the possibility that IE, as an hypothesis, would be
> subject to the rules of all hypotheses, and vulnerable to evidence
> that would just as convincingly point to the opposite conclusion as
> to the expected conclusion.

Actually, I did not get the impression that you were challenging
anything.  But I do think IE is an exceptionally good case of a family
which is well modeled by the family-tree model.

Two other respondents have pointed out that our family tree for IE is
muddled by the usual effects of contact and convergence.  Indeed it is,
and we certainly can't draw a single unambiguous tree covering
everything from PIE down to the modern dialects.

One of my favorite examples is Norwegian and Icelandic.  Iceland was
settled from Norway at a time when North Germanic (Scandinavian) already
exhibited noticeable regional variation.  Accordingly, we would expect
Icelandic and Norwegian to form a single node within Germanic.  As it
happens, however, Norwegian has largely developed in contact with its
neighbors Danish and Swedish, and modern Norwegian is much closer to
Danish and Swedish -- with which it is at least partly mutually
comprehensible -- than it is to Icelandic -- with which it is not
mutually comprehensible at all.  Consequently, our family tree today
usually puts Icelandic (and Faroese) off on a separate branch from the
three continental languages, in spite of the historical position.  And
this example is not isolated.

We also have a major problem with the subgrouping of IE: our standard
tree posits ten or twelve coordinate branches of IE, all of them
seemingly derived from separate daughters of PIE.  But nobody considers
this realistic.  The Pennsylvania group have recently proposed a more
normal-looking branching tree for IE, but I haven't yet seen many
comments on this.

> Your comments and explanation of the family tree model was also most
> helpful.  Certainly in Europe, where the dominate languages share
> such obvious roots, a family tree model would be the handiest and
> most logical model to explain divergence.  I wonder if the Asian
> languages of China, Korea and Japan share a similar background of
> divergence due to isolation.

Almost everybody seems to accept that the Chinese languages form a
branch of a much larger Sino-Tibetan family. But Japanese and Korean
continue to be puzzles.  There are attempts underway to link these last
two to each other, to Tungusic, and to Altaic generally (even though the
reality of Altaic is disputed).  Nobody knows.

> I must confess that your last point relative to the possibility that
> IE developed from more than one language, did cross my mind as I was
> investigating the issue. So, I felt it worthwile to ask about it.
> I am convinced now, that the IE model has stood the test of time,
> analysis and criticism, and accept your assertion that while IE may
> not be the only accepted model of linguistic development, it does
> the best job of explaining how most European languages developed.

Yes, I think the family-tree model works reasonably well for IE.
But the general validity of that model is very much under debate these
days.

> The only question I have now is, where does the Africian languages
> (Hebrew and Arabic, primarily) enter the IE equation?  Is it assumed
> that prior to PIE, the Asian and Africian languages that were non
> PIE were influencial in the ultimate development of PIE?  Surely
> there were many words that came from these sources, given the
> importance of the religious vocabulary available to the Hebraic
> people.

Hebrew and Arabic are unquestionably Semitic, and the validity of the
Semitic family is doubted by no one.  But Semitic is a good example of a
family to which the family-tree model is of questionable applicability.
The subgrouping of Semitic has proved to be a headache, and some
specialists are now leaning toward the view that Semitic existed for a
long time as a large dialect continuum, from which the individual
languages that we know crystallized only slowly, more or less in the
manner defended by Uriel Weinreich, Bob Le Page, and others.

The further membership of Semitic in a much larger Afro-Asiatic family
is likewise more or less universally accepted, chiefly on morphological
grounds, though reconstruction of Proto-AA has proved difficult, and no
proposed reconstruction has as yet won any great degree of acceptance.

A possible very remote genetic link between IE and AA forms part of the
Nostratic hypothesis, but is widely considered to be one of the weakest
parts of that hypothesis.  Very many people accept the existence of
early loan words between IE and Semitic, but the details are debated.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk