the family evolved in southern and western Anatolia, then we must suppose that
the Indo-European side of the family evolved somewhere else. And if we agree
that the language spoken in Anatolia in the eighth or seventh millennium BCwas
Proto-Indo-Hittite, then we can be certain that a long time (not several centuries,
but several millennia) elapsed before the maturation of PIE.
From the certain and the near-certain we must venture into the uncertain but
probable. A wave of advance cannot account for the evolution from Proto-Indo-
Hittite to PIE. Here we need not a gradual extension, a day’s walk or a day’s sail
at a time, but a clean break: a hiving-off or colonization to a distant land, resulting
in a considerable geographical separation from the originating population. An
oversea colonization, probably before the end of the sixth millennium BC, must
have established settlements of Indo-Hittite speakers far from western Anatolia
and the Balkans, and the colonists or their offspring must have proceeded inland
to a place still farther away: far enough away, that is, that the settlers were no
longer connected to lands already inhabited by Indo-Hittite speakers. In their new
home the colonists would instead have been in close contact with the local hunter-
gatherers. Interaction between the Neolithic settlers and the native hunter-gatherers
would have led to far-reaching innovations in the Indo-Hittite that the settlers
brought with them, and so to the evolution of PIE. This borrows from Marek
Zvelebil’s “creolization hypothesis,” except that Zvelebil speculated that
creolization accounted for the divergence of the several subgroups of PIE,^42 and
here the speculation is that PIE was itself a creole language: contact-induced
language change, that is, was a very important factor in the substantial divergence
of PIE from its Indo-Hittite roots. Surely Proto-Indo-Hittite colonists, when they
reached an arable land far from their point of departure, must have mixed
intensively with the Mesolithic inhabitants of the place.
That such a place was northern or central Europe is very unlikely. A hundred
years ago, mostly for racial reasons, northern Europe was widely considered to have
been the homeland of the Proto-Indo-Europeans. A similar scenario, although on
different grounds, has been advocated since the early 1980s by Alexander Häusler,
who believes that PIE was spoken over most of northern, central and eastern Europe.
More specifically, the Baltic, Keltic and Germanic branches of the Indo-European
language family were rooted, according to Häusler, in the Funnel-beaker culture
(Trichterbecherkultur. or TBK, in German, and often TRB Culture in English). This
was an Early Neolithic culture, which in the fourth millennium BCstretched from
the Netherlands to the Baltic and south to the headwaters of the Elbe.^43
A northern European homeland for PIE can hardly be squared with the Indo-
Hittite theory, and with the near-certainty that western Anatolia was the original
locus of Proto-Indo-Hittite. How, ca. 6000 or 5000 BC, could Proto-Indo-Hittite
speakers have hived off from Anatolia to northern Europe, and why would they
have chosen to do so when there were equally fertile lands much closer to home?
The required isolation from the world of Indo-Hittite speakers is difficult to
imagine in Europe, where a wave of advance should have created an Indo-Hittite
continuum from the Bosporus and the Dardanelles to whatever point the wave
had reached on the ground.
Origins and spread of Proto-Indo-European 11