Re: THEORY: Non-nom Subj & Nom Obj -- Quirky OVS Word Order Or Quirky Case?

From:    Christopher Wright <dhasenan@...>
> Tom C. wrote:
> > Here's the thing; Why aren't these sentences in an admittedly unusual
> > OVS word order, rather than having quirky cases?
>
> In order for that to happen, you'd need a few things to happen:
> 1. Nominative case would have to be assigned to a subject in its base position.
> 2. Either the base position for subjects would have to be below*
>   (c-commanded by) the verb, or the verb would have to move above the base
>    position for the subject.
> 3. The object would have to be assigned case from the verb as usual.
I think probably this is a little too much couched in framework-specific
terms to be useful for most people.  In particular, GB/PP/Minimal-oid
frameworks tend to be either so restrictive that they are immediately
falsified by vast quantities of data (most forms), or they are too
unrestricted (e.g. Kayne's work on the underlying universality of
word order) and end up being circular or purely descriptive or both.

In this case, for Tom C's purposes, it is sufficient to provide the
basic *empirical* tests that are supposed to underly theories.  In
English, we say *anything* is a subject because it has a number of
formal, empirical properties; what gives rise to these properties is
framework dependent.  Among such properties are:

  -- Control of agreement, if only one argument agrees
  -- Control of reflexivization
  -- control constructions: does the subject of the matrix
     verb control coreference with the subject of the embedded
     verb
  -- behavior under causativization:  in some languages, such as
     Malayalam, any surface subject becomes an object when the base
     verb has been causativized
  -- Passivization:  does the object of the active verb acquire different
     properties when passivized?

The fact that these properties may diverge was one of the main motivations
for the multistratal analysis of generative grammar to this day, but
monostratal analyses find ways around this problem, chiefly by
reference to other modules such as some kind of formally separate
argument structure.

> * 'Up' is to the left in surface order, and 'down' is to the right. These
> days, it's rare to suggest that something moves down, due to theoretical
> concerns.
In fact, before spellout, it's nearly unheard of.  However, it's the
norm in reconstruction situations where you need to move the XP back
down to its original position to receive the right scope at LF.

> It's possible--compare it to stylistic inversion. It would have to be
> optional, though--no theory that I know of allows a verb to decide the word
> order of a sentence at so basic a level. So you could simply say 'If it's
> optional, it's OVS surface word order from stylistic inversion; otherwise,
> it's quirky case.'
That's partially true, but syntactically ergative languages, such as
those which have V-S order with intransitives and A-V-P with transitives,
are well attested, if rare, and they don't fit into that pattern well.

> I'm not familiar with the Icelandic data, but it's probably a matter of
> optionality. Languages just don't switch between SVO and OVS for certain
> verbs;
Not so, as I stated above.  And if you're a card-carrying Minimalist,
optionality is something to be taken as a last resort.  In fact, the
reason the Icelandic data is taken not to be inversion is that the
oblique arguments have properties that fronted arguments (which also
exist in Icelandic) do not have, such as that they can be omitted
only under identity with the matrix subject.  Apparently this is also
true for modern German:

<http://www.ling.lu.se/conference/Grammatik_i_fokus/Eythorsson_Barddal.html>

I'm obliged to refer all to Johanna's abstract, since I had a beer with
her a few years ago and saw her drunk. :)  That link also has references
for more information about Germanic oblique subjects.


 =========================================================================
Thomas Wier	       "I find it useful to meet my subjects personally,
Dept. of Linguistics    because our secret police don't get it right
University of Chicago   half the time." -- octogenarian Sheikh Zayed of
1010 E. 59th Street     Abu Dhabi, to a French reporter.
Chicago, IL 60637