Examining the farming language dispersal hypothesis

"Section: BOOKS et al.
ARCHAEOLOGY

The circa 6000 extant languages are classified into a series of major
families that have wide geographical distributions. All the way from
Madagascar to Easter Island, for example, people speak languages that
fall into the Austronesian family. People from Ireland to Bangladesh
speak languages of the Indo-European family. Niger-Congo (mainly
Bantu) languages are heard from Cameroon to South Africa, and the Uto-
Aztecan family is found from southern Mexico to Idaho. This
distribution pattern needs to be explained. In a series of
independent publications focusing on quite different parts of the
world, archaeologists Colin Renfrew and Peter Bellwood have proposed
that the pattern reflects the dispersal of the first farming
communities from the multiple centers of agricultural origins. Each
such center, they suggested, was also the homeland for a now-
extinct "proto-language" that gave rise to one of the language
families. Renfrew has argued that Proto-Indo-European (PIE) was
originally spoken in Anatolia around 7000 B.C. and was spread into
Europe and southern Asia by immigrant Neolithic people who brought
farming into those continents and displaced the indigenous hunter-
gatherers. Bellwood has offered a similar hypothesis for the spread
of Austronesian languages from southern China, through Taiwan, and on
throughout Polynesia.

The farming-language dispersal hypothesis has generated intense
debate, and in August 2001 Bellwood and Renfrew hosted a conference
at the McDonald Institute for Archaeological Research (Cambridge,
England) to explore the topic. They invited not only archaeologists
and linguists but also geneticists whose study of modern human
diversity provides an additional means to establish the pattern and
chronology of human dispersal. The meeting epitomized the ongoing
integration of archaeology, linguistics, and genetics — what Renfrew
correctly describes as a new synthesis — that is fundamentally
changing how we reconstruct our past.

Examining the Farming/Language Dispersal Hypothesis presents revised
versions of the 36 papers, from 57 contributors, that were discussed
at the conference. The chapters are evenly balanced among the three
major disciplines, and they cover most of the globe (in four regional
sections on Western Asia and North Africa, Asia and Oceania,
Mesoamerica and southwestern United States, and Europe). The
excellently produced volume includes numerous informative maps (often
complex, but very well executed).

Some contributors keep to their own discipline and provide no help to
those unfamiliar with their territory. Others attempt their own
syntheses and demonstrate an impressive command of the latest
evidence from the three fields. Notable chapters include those by
Dorian Fuller (on India), David Phillipson (on sub-Saharan Africa),
and Marek Zvelebil (on Europe). What we find is not unexpected: the
history of the world is more complex than the farming-language
dispersal hypothesis suggests. Scratch the archaeology of any one
continent and what had once been argued as a simple case of Neolithic
dispersal becomes a mosaic of multiple processes that may have
produced quite independent patterns of economic, linguistic, and
biological change within the populations concerned.

Similarly, scratching the surface of any one discipline exposes
methodological difficulties. The key shortcoming of prehistoric
archaeology is well known: It struggles to determine whether cultural
change originates in the spread of people or of new ideas and
technologies. But at least archaeology provides a chronological
framework for tracking change — the fundamental element missing from
historical linguistics. By identifying the words that are shared by
languages within a family, linguists attempt to reconstruct the
lexicon of its proto-language. They argue that the presence of words
for domestic animals, crops, and agricultural techniques indicates a
proto-language spoken by farmers. Some words may have chronological
implications: Bernard Comrie is confident that PIE contained the
verb "to plough," which would suggest that it could not have been
associated with the earliest farmers in Europe because ploughs were a
later introduction. But such cases are rare and attempts to determine
the locations and dates of now-extinct languages or people by
correlating with archaeological evidence seem spurious or (as in
Christopher Ehret's contribution regarding East Africa) rather
fanciful. The lack of a means to establish a chronological framework
for language evolution severely constrains the utility of historical
linguistics. The discipline formerly had "glottochronology" — the
idea that words were lost from a language at a constant rate — but
this has now been widely discredited, leaving the field in much the
same position as prehistoric archaeology before the discovery of
radiocarbon dating.

Molecular genetics has already revolutionized the field of human
evolution, and it appears set to do the same for prehistoric
archaeology by allowing us to map the dispersal patterns of modern
humans. Geneticists identify where similar patterns of mutations are
found within extant groups of people and then assume that such
patterns had a common origin. They use mitochondrial DNA for tracing
the histories of female populations and the Y chromosome for those of
males. The volume indicates that molecular methods have a greater
potential than historical linguistics for determining the patterns of
human dispersal. But the approach has its own set of unresolved
methodological issues, which are very usefully described by the
contributors. The authors also recognize that the current number of
samples is clearly too small for the conclusions they wish to draw.

The ideal approach is to use genetic distribution patterns in modern
humans to test hypotheses regarding prehistoric dispersals derived
from archaeology and linguistics. Taking this approach, several
contributors have produced results that conflict with the hypotheses
of Renfrew and Bellwood by suggesting that the predominant patterns
of genetic diversity in both Europe and Polynesia arose long before
the supposed spread of the first farming peoples. But precisely how
current patterns of genetic diversity relate to several millennia of
population history remains to be fully established.

Renfrew's concluding comments express the frustration that will be
felt by many reading this volume: When we look at the wide
geographical distribution of languages, there is something to be
explained, and some general processes should be at work. But as soon
as we examine one language family or one region, we are immersed in
historical particulars. When considering the explanation for these,
we find not only the absence of an interdisciplinary consensus but
the lack of agreement among practitioners within a single field.

Although this volume neither validates nor falsifies the farming-
language dispersal hypothesis, it unquestionably constitutes a key
milestone in the creation of a new synthesis of human prehistory. As
such, it is required reading for anyone concerned with understanding
our present circumstances as well as our past.

Examining the Farming/Language Dispersal Hypothesis
Peter Bellwood and Colin Renfrew, Eds.

McDonald Institute for Archaeological Research, Cambridge, 2002. 519
pp. $85, £50. ISBN 1-902937-20-1.

~~~~~~~~

By Steven Mithen"