Contributions by Steve Long (long ;-)

[Ralf-Stefan Georg]

LT:

>Well, Steve Long will doubtless be bemused to find himself the subject
>of discussion on this list.  But Lloyd Anderson has made it so, and,
>since Lloyd has quoted me repeatedly in his posting, I guess I might
>respond to a few of his points -- especially since I think that Steve
>and Lloyd are both gravely wrong on certain fundamental points.

>I think both have fallen badly into the reification fallacy.  The
>reification fallacy lies in inventing a name, and then concluding that,
>since we have a name, there must exist something "out there" for the
>name to refer to.

>In the linguistic case, the fallacy lies in assuming that names like
>`English', `German' and `Italian' must designate actual entities in the
>real world, because the names exist.

Once in a while one should say when some posting really hits the bull's
eye, and this statement does. Excellent. Like what followed.

>[Pete Grey]

>>> Imagine a situation where a group of spakers of a language go and settle
>>> elsewhere, where substrate and other factors make their language change
>>> swiftly, while those who stayed at home enjoy a very much slower rate of
>>> change.    After some years, and political upheaval, we can see a situation
>>> where the settlers are deemed to be speaking a different language from that
>>> which they brought with them years before.

I confess that I, learning about the thread entitled "Can parent and
Daughter coexist [which I haven't been able to follow], intuitively thought
about such a scenario to say "yes", but I'm happy to admit that I'm glad I
haven't.

LT:
>Is the English of 1999 "the same language" as the English of 1998?

In the spirit of your argument: no, 'course not. But one
attitudinal/sociological factor should not be overlooked: Though changes
have taken place, 1999 English speakers *think* they are using the "same"
language as they were doing a year before. And, of course, the changes in
any individual's ideolect will indeed be infinitesimal (but some speakers
will have died, taking their ideolects with them into their graves,
mass-media will have introduced new lexical items into the speech of many
people aso. aso.). I simply suspect that there are two meanings of "same"
here. Maybe it will be useful to speak of a more mathematical "sameness"
(never observable between even the speech of two individuals from one
family, or maybe even between two days for one individual) and a more
"fuzzy" notion of "prototypical" sameness. Of course, 1998 and 1999 English
are not "same-1", but they are of course more "same-2" than 1999-English
and 1799-English, or English and Scots, which in turn are more "same-2"
than English and Dutch, next English and Danish, E. & Gothic, E. &
Lithuanian etc.
To require the notion "same" only for the idealized situation ("same-1") is
one extreme, the other extreme being to use "same" for English and
Tokharian.
For the latter "historical identity" will be a term which may be put to
some usefulness, provided it is never confused with "same-1", but, then,
who will ever do so ?

LT:

>Take a real case in this vein.  Is modern Greek "the same language" as
>ancient Greek?  If not, where's the cutoff point?  And why do we call
>both of them `Greek'?  But, if you think they are the same language,
>then what's your response to the observation that Pericles and an
>Athenian taxi-driver couldn't understand each other at all?

Correct. Straightforward identity is not the thing. But "historical
identity". In the Greek case, we can illustrate this with the following:
ever since the other Ancient Greek "dialects" were ousted by Attic (yes,
there is Tsakonian, but this is only an example), the language has been
transmitted in the "normal" way, i.e. from parents to child (simplified !),
with every new generation *intending* to speak the "language" (it's time to
introduce the quotes here ...) of the preceding one, *aiming* at their
norm, yet introducing change after change etc., we all know that.
In the Greek case, no sizable group of speakers worth speaking of has ever
seperated from this "mainstream" of vertical descent with every generation
being convinced to speak just the language of the older one(s). When such a
seperation does occur, it doesn't have to involve physical seperation of
the migration type, cultural reassignments of subgroups without physical
dislocation of whole groups may suffice. Such a seperation, however it does
come about (it doesn't have to be a historical "catastrophe", contact may
simply peter out for hosts of reasons), will lead to speakers interpreting
their we-group as different from "them", gradually leading to an
interpretation of any differences in linguistic usage between "us" and
"them" as marking "different languages", however close their lects may be
in purely linguistic terms (this is currently happening in Yugoslavia, as
Larry noted; the opposite does also occur: witness the overnight extinction
of language with 2 1/2 Mio. speakers recently. Cannot happen ? Has
happened: Moldavian; I'm not properly informed whether this has happened to
Tadzhik yet).
This has, of course, also linguistic consequences. Innovations spreading in
"their" variety are no longer copied in "our" one, since "we" simply don't
talk to "them" any longer, or less often aso. (This is, btw., the gist of
the wave-model, so hotly debated long ago, but in reality not in opposition
to any stammbaum-model: they are just two aspects of the same thing).
Now, in the first scenario, separation-free vertical continuity with
maintained group-identity, it is hard to avoid some "sameness"-metaphor,
for which "historical identity" might well do the job.
What about "sameness" after seperation ? Of course, speakers of seperated
varieties will foster the same ideology, namely that they simply continue
to speak like their ancestors did, mostly blaming "them" for not to. Of
course, there is also some degree of "sameness", but for this the
"relatedness"-metaphor is generally used, involving at least the following
assumptions:
There is "historical identity" between the lects under discussion; however,
since more than one variety ("language", i.e. the conglomerate of
linguistic, but mostly also political, cultural and ideological notions
which leads people to single out such a construct, giving it a single name)
may claim the right for this kind of "sameness", and the coexisting
varieties may thus not be chronologically ordered in the way we can treat
the speech of Pericles and our Athens taxi driver (nobody could understand
Pericles today without learning his lg.).

This doesn not answer the cardinal question whether a parent language and
its daughters may coexist, yet, since every language of thursday is a
changed variety of wednesday's language, this appears meaningless. But not
quite.
If we, for a moment, abstain from any attempt to "define" a "language",
which seems to be as impossible as to define any term of everyday language
in scientific terms, a different approach might bring about that there are
possibilities where to speak of coexisting parents and daughters  could
after all make some sense.
I have in mind not "languages", but linguistic *items*. By "item" (a term
used in David Nettle's interesting new book on "Linguistic Diversity") I
understand, very loosely "defined", just everything languages are made up
of (I know that views of "languages" as "sets" consisting of "elements" are
grossly simplified and not at all adequate, but few people will deny, I
hope, that some aspects of language may well be treated this way; certainly
the lexicon is by and large describable as a set with elements), words,
morphemes, accent patterns, syntactic rules, everything one could describe
and put a name on. Now, I do maintain that, with linguistic *items*, parent
and offspring can coexist, and often do. A simple example: the Germanic
words for "king" are derived from a reconstructable Proto-Germanic item
*kuningaz, which, by definition "no longer exists". Yet it was borrowed
early into Finnish, where it survives to this day as /kuningas/.
Discounting the small difference this example still presents, I hope the
picture is clear: a parent and its offspring may coexist, and of course
there are more and better examples where, in a given family, forms may
exist, which simply haven't underwent any significant change since earliest
times. Or take Italian /akwa/ against, say, French /o/. Of course, the
French does not go back to the Italian (neither the "languages", nor these
"items", but the parent form of French /o/ was, at times, /akwa/; this
remained unchanged in Italian; the items are certainly not "same-1", but
"same-2" they are, and Italian speakers still copy the form, which French
speakers have changed so dramatically over the centuries. Of course, one
could insist on our inability to assert "sameness" on the level of each and
every phonetic detail of vowel openness or V-1 time, or prosody and stuff,
but there is a level of abstraction (the phonological one), where the
notion "same" for Latin aqua and Italian /akwa/ makes sense.
So, on the item level, parent and offspring can coexist, and they often do
(in some close-knit language families, like e.g. Turkic, they very often
do, i.e. what has to be reconstructed for the proto-lg. is very often
attested largely or totally unchanged in some modern language).

To sum up: since the notion "language x" is a fuzzy notion, discredited by
the very fact of the inevitability of constant day-to-day language change,
more so by the nonlinguistic motives which often lead to the reification,
as Larry said so brilliantly, of "languages", the question of whether
parent and daughter *language* can coexist is meaningless or, consequently,
to be answered in the negative, as Larry did.
On the level of single linguistic items, this is not necessarily so, only
if one insists on /aqua/ being uttered by Cato the Elder and /acqua/
uttered by Luca Cavalli Sforza have no right whatsoever to being treated as
"the same" item on whatever level of abstraction.
The question, whether "language" A and B, sharing this and that set of "the
same" items are the "same language" or not, or after which precise amount
of divergence-increasing changes they cease to be is not really a
meaningful question of linguistics, it may even not be a meaningful
question for philosophers. Constructivism takes care of most of these
pseudo-problems.

St.G.

[ moderator snip ]

Stefan Georg
Heerstrasse 7
D-53111 Bonn
FRG
+49-228-69-13-32