PIE syntax and word-order

Thomas McFadden tmcfadde at babel.ling.upenn.edu
Thu Jun 7 19:50:54 UTC 2001


> You suspect wrongly. Here is your sentence :

> 1/ Hier j'ai achete une maison tres grande a Vaucresson

> 2/ Hier j'ai achete une tres grande maison a Vaucresson

> Absolutely no difference in meaning, style or emphasis. I am sorry.

this proves nothing.  as i think i said a few days ago, the type of
difference we expect in this type of variation can be very difficult to
measure based on a speaker's intuitions about a particular sentence.  they
are best studied by looking for patterns of how the variants are
actually used in texts.  note that i am insisting on consulting actual
texts here, in spite of what you imply in your remarks below.  but, as i
have also mentioned before, even if it turns out that you are right in
this instance, that there is absolutely no difference in the meaning,
implication or contextual conditions on the usage of the two variants you
list above, that still does not bear on the question of whether the
variation between them is regulated by the grammar.  no matter what they
mean, they are specified as possibilities in French, while a number of
other imaginable word orders are NOT possible in French.  this can NOT be
reduced to some vague notion of stylistics.  and the fact remains that for
the most part, word order variation within a language CAN be connected
with variation in meaning, or implication, or discourse context, at least
variation on the level involving subject verb and object.

> But this is not PIE, only a late IE language with a fairly fixed word-order,
> with some pockets of freedom in this matter.

of course.  it is obvious that word order is grammatically determined, it
is less obvious in Greek, Latin, Sanskrit, etc.

> I am convinced that a more intimate practice of old IE languages like Latin,
> Greek, Sanskrit,or Lithuanian would convince you that word-order is not an
> essential part of their grammar.

and i am equally convinced that a detailed study of actual texts would
reveal consistent patterns of word order correlating in some way with the
discourse context.  hopefully i will be able to offer references to such
work very soon, so that i can do more than claim that i'm right.

> In the previous posting, I made the (nasty) supposition that the obsession
> for word-order in some circles was due to the mother-tongue of their
> scholars. In fact I suspect that the reason is of a more general type : the
> overinterest for general linguistics with its theoretical fashions,
> generativism among others. It is always easier to build a general theory on
> something (e.g. word-order) than to be confronted with facts, i.e. existing
> languages.

i let the original comment about obsession with word order pass, but you
continue to make prejudiced blanket statements that are non-arguments and
actually incorrect, so now i'm forced to respond.  argue against
generativism if you like, i won't mind, but there's no need to call it a
`theoretical fashion.'  it's been around too long and proved too useful
for that.  there is no question that a common flaw in generative work is
an exaggerated concern with terminology and the inner-workings of a
particular theory, and a concommitant relative ignorance of real
linguistic data.  but it's simply inaccurate to assume that this
characterizes all of generativism, or even most of it.  good
generative work constructs hypotheses about how language works based
on extensive real language data.  other theoretical
leanings have their own shortcomings too.  but what i have been talking
about here is not guilty of ignoring data.  (and by the way, i must
admit to being confused by your statement `It is always easier to build
a general theory on something (e.g. word-order) than to be confronted with
facts, i.e. existing languages.'  in what way are attested word-order
patterns from existing languages not facts??) i'm making no arguments
based on a specific theory, and in fact my position is not specifically a
generative one.   and i am insisting that what i am arguing for needs to
be tested on actual textual data, because what i know about effects on
word order in `free word order' languages comes from corpus-based studies,
not on native-speaker intuitions and not on argumentation from a
pre-existing theory.

please argue against what i am actually saying rather than slinging mud at
a distorted picture of a theory that you have associated me with.  if you
can convince me that Latin (or any of those other languages) does not have
its own language-specific patterns of word order that are conditioned by
semantics/pragmatics etc., then i'll admit i was wrong.

(i do realize that the burden of proof is on my side, since i'm the one
who's arguing for something.  i'm working on it.)

Tom McFadden



More information about the Indo-european mailing list