Re: THEORY: branchedness [was Re: Word order]

On Wed, 7 Aug 2002, Thomas R. Wier wrote:

> This answer is a little more complicated, and it is my impression
> that the answer is very theory-dependent.
Yes. Very much so.

>  In Government-Binding theory,
> some claim that at deep-structure, all languages are actually VSO,
> and move-alpha rules raise the subject in SVO languages, and both the
> subject and object in SOV languages. For these people, then, a VSO
> language is simply one where neither the subject nor the object is
> raised.  This would seem to suggest that VSO languages are basically
> right-branching.  In any event, the verb will be the head of the VP,
> and the VP will be the dominant phrase in the sentence.
This is sort of accurate, but not quite. Classical GB theory (mid- to late
80s), held that all languages were either SVO or SOV in deep structure,
depending on the language. The verb in many cases, such as Spanish,
French, and many others, but in other cases it did not, such as English.
If nothing else moved, this would give you VSO and SOV langauges. In some
languages, like English and French, the subject raised, which is why
English and French are predominately SVO: the subject is raised. Languages
like Zapotec or Irish don't raise their subjects, so the verb is initial:
VSO. With underlying SOV languages, it didn't matter much what moved and
what didn't, SOV was always the resulting order.

In the early 90s, Antisymmetry Theory claimed that all languages were
underlying SVO (by stipulation). To get an SVO language, the subject
raised and the verb could have raised or stayed put (French vs. English).
VSO languages were the result of verbs raising, but subjects not moving
(or moving less than the verb did). SOV was formed by the subject and
object both raising above the verb.

Marcus