Worth a look, this from
Andrew Garrett: "
Convergence in the formation of Indo-European subgroups: Phylogeny and chronology", in
Phylogenetic methods and the prehistory of languages, ed. by Peter Forster and Colin Renfrew (Cambridge: McDonald Institute for Archaeological Research, 2006), pp. 139-151
The technicalities all go soaring way over my head. We have Mycenaean Greek, in the Linear B texts, from around 1400 BC, and then that culture collapses around 1200 BC and the next Greek works come from around 800 BC. To quote James Clackson, (Indo-European Linguistics: An Introduction, 2007):
"Garrett draws up a set of features which could be assumed for Proto-Greek on the basis of the later Greek works....for Greek, there is the advantage that the features assumed for Proto-Greek can actually be compared with a language of the second millenium BC, Mycenaean Greek. We know that Mycenaean cannot be equated with Proto-Greek, since it has undergone some changes shared only with some later Greek dialects, and so it must be later. Yet all of the distinct morphological features and many of the distinct phonological features, which are assumed to be distinctive for Proto-Greek can be shown not to have take place at the time of Mycenaean. Wherever later Greek dialects have made innovations in morphology from PIE, Mycenaean Greek appears not to have participated in that innovation. In other words, the distinctive aspects of the later Greek dialects (which they all share) arose across a number of varieties which already were distinguished one from another. It is not possible, using the shared morphological innovation criterion, to construct a unified invariant entity such as "Proto-Greek" which is distinguishable from PIE......if we had more evidence for other IE languages other than Anatolian contemporary with Mycenaean, we might not be able to separate out what was 'Greek' about Mycenaean from its neighbours. The Greek sub-group was only truly formed in the period after the Mycenaean when convergence between the different dialects of Greek took place, in part related to social changes coupled with a strong sense of Greek ethnic identity."
To quote Garrett:
In
sum,
especially
if
we
allow
that
at
least
a
few
post‐Proto‐Greek
changes
must
already
have
affected
Mycenaean
before
its
attestation
(it
is
after
all
a
Greek
dialect),
detailed
analysis
reduces
the
dossier
of
demonstrable
and
uniquely
Proto‐Greek
innovations
in
phonology
and
inflectional
morphology
to
nearly
zero.
Proto‐Greek
retained
the
basic
NIE
noun
system,
verb
system,
segment
inventory,
syllable
structure,
and
arguably
phonological
word
structure.
In
all
these
areas
of
linguistic
structure,
Greek
was
not
yet
Greek
early
in
the
second
millennium.
But
if
so,
it
hardly
makes
sense
to
reconstruct
Proto-Greek
as
such:
a
coherent
IE
dialect,
spoken
by
some
IE
speech
community,
ancestral
to
all
the
later
Greek
dialects.
It
is
just
as
likely
that
Greek
was
formed
by
the
coalescence of
dialects
that
originally
formed
part
of
a
continuum
with
other
NIE
dialects,
including
some
that
went
on
to
participate
in
the
formation
of
other
IE
branches.
If
this
framework
is
appropriate
for
IE
branches
generally,
we
cannot
regard
IE
‘subgroups’
as
sub‐groups
in
a
classical
sense.
Rather,
the
loss
or
‘pruning’
of
intermediate
dialects,
together
with
convergence
in
situ
among
the
dialects
that
were
to
become
Greek,
Italic,
Celtic,
and
so
on,
have
in
tandem
created
the
appearance
of
a
tree
with
discrete
branches.
But
the
true
historical
filiation
of
the
IE
family
is
unknown,
and
it
may
be
unknowable.
We cannot check whether Garrett is right about IE branches in general, because in Greek we have the unique situation of having sufficient texts from two eras. That is why "it may be unknowable".
This idea that the language around 2000 BC around Greece was indistinguishable from PIE suggests a late date for PIE and weighs against the Paleolithic Continuity Theory. (Garrett's argument with respect to Renfrew's
PIE-spread-with-agriculture theory is: "the model requires the
unscientific assumption that linguistic change in the period for which
we have no direct evidence was radically different from change we can
study directly".) On the other hand, don't forget
Clackson's own metaphor, that conveys the idea that the PIE includes in itself reconstructions that belong to very different time periods, so what does it mean to be "indistinguishable from PIE"?
One more interesting thing is that Alinei has, to my reading, a very similar idea of how languages arose.
Mario Alinei, The Problem of Dating in Linguistic, "Quaderni di semantica" 25, 2004, pp. 211-232.
As I have already pointed out, written languages imply, by the very fact that they are expressions of
dominant groups, the existence of dialects of subordinate groups, which, though not attested, are
nevertheless as real as the invisible face of the moon. Precisely because a written norm represents
one of the geovariants or sociovariants promoted to the dominant norm, it reveals, ex silentio, other
norms, which remain necessarily excluded from written evidence, with the possible exception of
some traces surviving in the chosen koiné (common language).
From the structural point of view,
then, the appearance of a written language is also direct testimony of the emergence into 'history' of
the elite group which has seized power, and indirect testimony of the loss of power by other groups,
in regard to whom the new 'literates' assert themselves as the owners of the surplus product, as
ideological leaders and as rulers. Each written language represents, accordingly, a cluster of
dialects, still without voice, but in fact rightly present within the framework of the new social
relations consecrated by the written language.
We must, therefore, bear in mind that these dialects do exist, although we do not see them,
and we must take them into account in our theoretical interpretation. Since, for example, some IE
languages appear in the Mediterranean basin in their written form in the 2nd millennium, two
conclusions can be inferred from that fact alone:
(a) in the areas where there is definite evidence of
written languages we may be sure that the sociolinguistic stratification already reached Gordon
Childe's 'urban' level;
(b) in other areas, where the Metal ages cultures appear, we may assume that
social stratification was already at a considerably advanced stage.
There is, besides, another factor which should be taken into account. As I have already
noted, written norm is usually not equivalent to a 'pure' geovariant, but it is a koiné, implying an
admixture of elements from other geovariants (borrowings, morphological variants, and the like).
Mycenean Greek, for example, is regarded, as we have already seen, as a koiné. Even in the modern
world we can notice this intermingling in the process of the formation of a new written language -
in the case of Basque and Catalan, for example.
The formation of a written koiné implies, in short,
three different innovative aspects:
(1) a koiné, precisely because it is a mixture of one dialect with
elements of other dialects, represents a novum which did not exist previously; in other words, a
written norm, being a 'mixture', is as a rule more recent and less genuine than the norms of the
subordinated groups which have remained completely or partly in the dark;
(2) the elements of
other dialects accepted by the koiné become levelled with the dominant system and lose some of
their traits;
(3) other geovariants do not cease to exist at the moment a koiné is established, but they
become, or revert to, 'dialects', with the only difference that from that time on they undergo the
levelling influence of the new dominant language.
In the light of these considerations, the earliest written attestations of European languages,
either classical or mediaeval, cannot not seen as monolithic expressions of undifferentiated ethnic
groups, from which all that comes 'after' must be mechanically derived. Inverting the traditional
hierarchy, the first written norms must now be seen as the most fortunate representatives of a
dialectal continuum which despite the successive levelling has survived to the present day, and
which is the only source of our knowledge of the hidden face of the moon.
Just as in the Middle
Ages the earliest attestations of the dialects destined to become national norms are combined with
attestations of numerous other dialects, which prove that the modern dialectal continuum actually
existed already at that time, and probably also in the preceding centuries (for which geolinguistic
evidence is much scarcer), so Scandinavian runes, Irish oghams, Gothic, Norren, old Slavic, and so
on, must be interpreted as the mixed and most fortunate geovariants of a dialectal continuum
equally rich and articulated as the modern one. They must not be seen as its matrices, nor,
obviously, as unique offshoots of reconstructed proto-Germanic, proto-Celtic and proto-Slavic.
In
fact, whatever appears after the emergence of the written language did not come after, but was pre-existent to the written language. According to this new view, the current dialects are not derivatives
of the ancient written languages, as traditionally thought, but developments, in the course of
subsequent millennia, of those earlier geovariants which were parallel with and pre-existent to the
written languages. And the new dialectology, according to this view, becomes an integrating part of
the renewed historical linguistics, as the study, as it were, of the hidden face of the moon, that is of
the speeches of those social groups which became subordinated to the new elites in the Metal ages,
but which were obviously pre-existent to the Metal age itself.
In the case of a written language there is, then, only one birth to register in addition to the
birth of the written language as such, and that is the birth of the dominant group. The ethnic group,
or its part subjugated by the dominant elite, is millennia older than these events.
Clackson's constellations, Alinei's hidden face of the moon - I love these metaphors, and the way they illuminate the meaning of the PIE reconstruction.
PS: There is a simple answer to Garrett's argument : "the model requires the
unscientific assumption that linguistic change in the period for which
we have no direct evidence was radically different from change we can
study directly". Namely, linguistic change before writing follows different rules from linguistic changes after writing.
