Re: THEORY: Mixed erg/acc

At 05:47 PM 3/10/2000 -0600, Matt Pearson wrote:
>Mahajan's story is actually more of a series of observations than
>a full-blown analysis.  It goes kind of like this:
>
>(1)  Languages may express possession either by means of a
>verb "have", or by using the copula "be" with the possessor
>marked by an oblique case morpheme or adposition (typically
>"to" or "with" or "by").  Hindi belongs to the latter camp,
>so "I have a book" is "me by book COP" (or "by me is a book").
>
>(2)  The perfective in languages like English is formed by
>combining the past participle with "have".  The perfective in
>Hindi is formed by combining the past participle with "be".
Latin has both kinds.  Alongside the English-like construction with "have"
(Petrus librum habet = "Peter has a/the book") is an alternative
construction with the so-called "dative of possessor" (liber Petro est =
"a/the book is to Peter").  I don't know what the functional difference is
between them, but I assume there must be one; I would guess that it has
something to do with whether the possessor or the possessum is the topic
and/or with whether the possessum is definite or indefinite.
>
>(3)  Some linguists have proposed that possessive "have" is
>derived (synchronically) by combining "be" with an adposition.
>Mahajan argues that this analysis should be extended to the
>auxiliary "have" used in perfective constructions.
>
>(4)  Languages can chose either to combine "be" with the
>adposition to give "have", or to express "be" and the
>adposition as separate morphemes.  English exploits the
>former strategy while Hindi exploits the latter strategy.
>In the case of the perfective in Hindi, the adposition is
>identified as the ergative case suffix.  Thus, schematically
>we have:
>
>Possessive constructions:
>
>Hindi:  BY-him  a book  IS
>English:        he  HAS  a book
>
>Perfective constructions:
>
>Hindi:  BY-him  a book  WRITTEN IS     (BY = ergative case)
>English:        he  HAS WRITTEN  a book
>
>Under this story, it's no accident that perfectives in English
>are formed using "have" rather than "be".  This analysis also
>suggests close structural parallels between perfectives and
>passives, which both use the same participle form in most
>Indo-European languages:
>
>        John   HAS READ   the book
>        the book   IS READ   BY John
>
>Both are derived from the deep structure "is read the book
>by John".  In the passive construction, "the book" raises to
>become the derived subject, giving "The book is read by John".
>In the perfective construction, the preposition "by" incorporates
>into "is", creating "has".  Deprived of its preposition, "John"
>then raises to become the derived subject, giving "John has
>read the book".  (This latter analysis was also independently
>proposed by the Dutch linguist Teun Hoekstra.)
>
>There's more to the story than that, but that's all I remember...
>
>Matt.
This makes perfect sense to me.  But I can see some possible complications,
especially with ditransitive verbs.  If the case affix or adposition that
ends up in the ergative role was originally dative, you'd end up with
ambiguous sentences like "to John is written the letter to Mary".  I
suppose that would lead to either fixed word order or a new
affix/adposition taking over the dative role.

One of the conlangs that I've been thinking about lately has exactly this
kind of mixed-case system evolving from a mostly isolating grammar with VO,
head-modifier syntax and prepositions (unlike the Indo-Iranian languages,
which are generally OV, modifier-head, with postpositions and/or case
suffixes).  The idea is that most finite verbs are replaced by combinations
of the copula (which can be zero in present indicative main clauses) with
non-finite verb forms (gerunds and participles): for perfective verbs, a
past participle, which is passive for transitive verbs ("by John (is)
written the letter"), and for imperfective, a prepositional phrase with a
gerund as the object, and with the patient marked formally as a possessor
("John (is) at writing of the letter").  Thus, besides turning "by" into an
ergative case marker, you also turn "of" into an accusative case marker.
This allows for a lot of freedom of word order despite the lack of case
affixes.

(Where I've gotten stuck so far is in coming up with a new genitive
construction to show "real" nominal possession, to replace the old genitive
that's been co-opted as an accusative.  There are lots of possibilities --
one of which is simply continuing to use the same construction in both
functions, which I suppose wouldn't cause any more ambiguity than the fact
that for animate nouns in Russian the genitive and accusative are the same
-- but none of them quite works the way I want it to in all situations.
The real problem is that I want not only freedom of major constituent order
within the clause, but also freedom of word order within phrases, so that,
for instance, genitives or attributive adjectives could either precede or
follow their head nouns, depending on which is more emphasized.)

- Tim