synchronic vs diachronic causes of language changes

[bwald]

----------------------------Original message----------------------------
Ireena Lifshitc-Fufajeva makes a valid distinction between "how?" and
"why?" with respect to any particular linguistic change.  However, her
assumption that "why?" is answered by "need" leaves something to be
desired.  She writes with regard to the shift from fixed to movable stress
for certain Russian neuter nouns:
 
27 Russian nouns o-neutra total
>changed from 16 centure fixed stress to movable, some from
>them in literare Russian. I see in this fact an interesting
>example of language self-regulation under impact of mighty need
>of speakers in the short (grammar) expression of number.
 
No doubt this change is motivated by the *desire* to restore an audible
distinction between the singular and the plural of the nouns in question
(following a/o phonetic merger in unstressed positions, where -a is the
plural counterpart of -o as Russian suffixes) .  However, she spoils her
explanation with the parenthetical comment:
 
>(This category is almost universal as you know.)
 
Here she assigns languages which do not have plural as an obligatory
category to the realm of "unnatural" curiosities, an ethnocentric position
that few of "you" will accept.  The real explanation does not have to do
with "need" in such a universal sense.  It has to do with the grammatical
nature of Russian.  In Russian, number is an obligatory category for
(count) nouns.  Speakers feel a discomfort with losing the distinction for
what seem to them to be an arbitrary set of nouns (based on their form, not
their semantics).  Therefore, they act to restore the distinction and
regularise (or, as she says, "regulate") singular:plural marking when there
is no semantic motivation for suspending it.
 
If the restoration was based on "communicative" need (to dismiss the notion
of "universal need"), then it would not occur when a modifier is present
which unambiguously indicates the number of the noun, as in the case of
Uralic or Altaic (with a number, or, in Russian, when the modifier already
has movable stress).  But I doubt that this is the case for the Russian
innovation.  However, it is the case for the English "plurals" of "you".
It is not always clear pragmatically whether "you all/guys/etc" is
necessary to clarify that some single addressee is not being picked out of
a group, but it is clear that "you" indicating "one", as in "*you* can't
make a silk purse out of a sow's ear" is rhetorically closer to the
"plural" as a "generic" generalising to "everybody", than it is to a
specific single addressee.  And yet it is unheard of to say "*you guys*
can't make a silk purse..." to indicate "*one* cannot make a silk purse..."
("*you can't", in this sense, implies "nobody can", "you guys can't"
allows "somebody else can")
 
As I mentioned last time this subject came up, I am not convinced that "you
guys/etc" is a paradigmatic "plural" of "you" (and we don't have the same
option of suspending number marking for "we" and "they" that we do for
"you"), rather than an optional phrase, producing further specificity.
Similarly, I would expect someone whose language distinguishes an
inclusive/exclusive first plural to be either be amazed that IE languages
(among others) don't make it, or to analyse "me-and-you / you-and-I" as the
"inclusive 1p" in English.  I guess in the next stage we also get a "plural
inclusive" 1p, "me-and-you guys/all/etc", etc.  distinguished from the
"dual" 'me-and-you'.
 
NB.  Shocking as some of you more conservative speakers may find it
(despite your self-discipline as linguists), 'you-and-I' has indeed become
a unit for some American speakers.  They spontaneously say, "I wonder why
he didn't invite you and I" as well as the usual suspects, e.g.,  following
prepositions and in pivot contexts like: 'he wants you and I to come'. --
'me-and-you' as a unit is less of a problem, since the only innovation is
generalisation to subject contexts, but that also makes it less obvious
that it is indeed a pronominal unit rather than the result of loss of case
distinctions in favor of "me" as "unmarked" on the basis of its already
superior privileges of occurrence, as Sapir suggested in his comments on
'drift'.