> > But then what is a real cognate ?
> >
> > Then why should we consider PIE reconstruction is acceptable ?
> > Your reasoning is like saying everything is wrong or dubious.
> > Why should we consider IE more acceptable than Nostratic ?
> > If both apply the same methods, why should they not be equally
> > acceptable ?
>
> This is why I think Nostratic is dubious:
>
> I think the Peter Bellwood theory is basically correct, that the
> large language families are the result of small hunter-gatherer
> groups learning the (Wörter und) Sachen of agriculture, then going
> on to fill out the landscape around them.
> Torsten
> =========
> I guess I saw M. Bellwood last saturday,
> They had organized conferences in PAris about Neolithization of
> Europe.
> He was there.
> The major pb of his theory is that it's based on the example of
> Autronesian,
> which expanded in a human vacuum in the Pacific.
Not quite accurate:
Peter Bellwood:
Examining the farming/language dispersal hypothesis in
the East Asian context, pp. 17-19, in
The Peopling of East Asia
'The Farming/Language Dispersal Hypothesis (Bellwood 2001a; Bellwood
and Renfrew 2003; Renfrew 1996) suggests that the foundation
dispersals of many of the major language families of tropical and
temperate latitudes (e.g. Indo-European, Afro-Asiatic, ST, AA, AN,
Uto-Aztecan) occurred consequent upon the establishment of reliable
agricultural (and especially agropastoral) economies and increasing
population densities in and around agricultural homeland areas. As a
result of these increasing population densities, some degree of
centrifugal movement would have been inevitable in non-circumscribed
situations. The hypothesis has been applied to the geographical region
termed 'China' on several occasions (e.g. Bellwood 1994, 1995,
1997a,b; see also Higham 1996, 2003; Reid 1996), especially for the
ST, AA, Tai (Thai-Kadai, Daic, Kd), HM (Miao-Yao) and AN language
families. Suffice it to say that recent developments in linguistics
and archaeology do not seem to negate the hypothesis in any major way,
insofar as it applies to the agricultural homeland regions of China -
Manchuria, Mesoamerica or Southwest Asia. However, like all good
historical hypotheses which attempt to integrate data from
archaeology, linguistics and genetics, this one is not and probably
never will be subject to positive proof or disproof. In the following
text, the hypothesis will be qualified with respect to certain aspects
which sometimes give false impressions of absolutism; it is not
intended to explain all language distributions in all periods of the
human past and it is highly sensitive to situational factors.
2. The foundation spreads 1 of language families, in many cases
occurring long before history and over vast extents, and in
sociocultural situations of small-scale preliterate farming societies,
required population movement as their major driving force. Language
shift doubtless worked to a degree on a local scale, but it could
never have propelled foundation Indo-European languages across the
vast stretch of territory from Anatolia or the Ukraine to Western
Europe and Bangladesh, or AN languages across the even vaster extent
of ocean and islands from Taiwan to Madagascar and Easter Island. The
corpus of recorded language-spread situations in history is extremely
large, and supports this perspective strongly (discussed to some
degree in Bellwood 2001b, 2003). There are no recorded situations of
language shift, whether through elite dominance or any other
mechanism, that could conceivably explain such large-scale dispersals
in the absence of any substantial factor of population movement.
3 Such outward flows from agricultural heartland areas will have
tended to continue as long as demographic gradients falling off
centrifugally were maintained, even though
antecedent populations, whether hunter-gatherers or other
preceding groups of less numerous/less aggressive farmers, can always
be expected to have given rise to at least some substratum effects.
antecedent hunter-gatherers sometimes adopted agriculture and
the languages of incoming farmers, and then might have undergone
expansion in their own right. Preceding groups of agriculturalists
could also have adopted the languages of different incoming farming
populations, as must presumably have happened amongst some lowland
Melanesian Papuan-speaking populations who adopted AN languages (such
populations could have been either gardeners or hunter-gatherers). In
this regard it is important to note the high degree of biological
variation amongst populations who belong to some of the major language
families, for example northern Indians and Scandinavians
(Indo-European), Filipinos and Solomon Islanders (AN), Arabs and
Ethiopians (Afro-Asiatic), Mongolians and Turks ('Altaic', if one
accepts the existence of this grouping). It seems most unlikely that
such variation could be due to natural selection alone working on
common base populations in the short time spans available since the
Neolithic or since the relevant protolanguages existed, and obviously
one needs to incorporate concepts of language shift and
contact-induced change in any global class of explanation, such as
that represented by the farming/language dispersal hypothesis. But
these concepts alone cannot explain everything.
following on from the earlier text, we cannot expect genetic
outcomes to mirror exactly those of archaeology and linguistics
(people intermarry, but languages find it difficult to do so on the
50:50 level characteristic of recombining chromosomes, at least not
anew in every generation!). However, current tendencies within the
anthropological literature to state that geographical patterns in
languages, cultures and genes always vary completely independently of
each other are not helpful for historical understanding and seem to
reflect more an ethical statement about how our present troubled and
ethnically-divided world should function, rather than any informed
wisdom about how it might have functioned in the past.
the actual spreads of ancestral languages within specific
families have been layered through quite long time spans (4,000 years
in the case of AN and doubtless longer in the case of Indo-European).
Our interest in this article is mainly in the primary and very
extensive foundation layers of language family dispersal.
I should also add three provisos. First, no claim is being made here
that only agriculturalist language families ever spread; we also have
several language families which originated and spread amongst
hunter-gatherers, such as Uralic, Eskimo-Aleut, Athabaskan, Algonquian
in Canada, and maybe even the much debated Pama-Nyungan. These need to
be explained too, and population movement is doubtless as significant
here as in the spreads of the agriculturalist language families.
Second, I do not wish to suggest that agriculturalist dispersal goes
back to the very roots of all language families which are currently
agriculturalist. It is possible, for instance, that both Niger-Congo
and Afro-Asiatic had already undergone some dispersal prior to the
development of agriculture, although in both these cases the evidence
is by no means clear since it is difficult to reconstruct with
absolute certainty the economic basis of the period represented by the
basal proto-language (e.g. Ehret 2003 vs Militarev 2003 for
Proto-Afro-Asiatic). Third, not all agriculturalist language
families/subgroups underwent spread - the list of stay-at-homes,
doubtless for reasons connected with circumscription and successful
intensification of production, is long (Egyptian, Sumerian,
Mixe-Zoque, the Caucasian language families...).'
> IE languages expanded partly on other IE languages
> and languages that may have been close to PIE
> so the case is different.
> Arnaud
> ============
As you can see from Bellwood's paper, it makes no difference to the
basic scenario.
> That means the languages of those
> families began to diverge < 10 k years ago, which is the max. age of
> agriculture.
> T
> ========
> No,
> PAA is clearly older than agriculture
> and I can prove it with Berber,
> which was already separated from the rest
> then a new layer of words reached it.
> Arnaud
> ===========
See Bellwood's article.
> The extra-African language families themselves, however,
> must have split up approx 80 k years ago, for that was the time when
> the present extra-African population colonized the world. The only
> way for that split-up to have been younger is if some of the founder
> population had changed language at some time in the past. In other
> words, the reconstruction of Nostratic should be ca. 8 times harder
> than reconstruction of any of the component families, not to mention
> the fact that the reconstruction by necessity will be based on
> derived, not primary data.
> Torsten
>
> ================
> The factor 8 is probably exagerated.
> Maybe 2 or 3 is better.
Please present an argument for that claim. We are not in a bargaining
situation.
> I don't understand : the reconstruction by necessity will be based
> on derived, not primary data.
And that is why you reach down, as Brian calls it, and are not aware
of the pitfalls of that method. Watch me in Arnaud mode:
[arnaud]
Fr. téléphone, Grm. Telephon, Da. telefon, Est. telefon. I reconstruct
Nostratic *telepon which was spoken 80,000 years ago.
[/arnaud]
Do you see the problem?
Torsten