Print

Print


Hi!

tomhchappell <[log in to unmask]> writes:
>...
> BTW I and at least one other poster were
> trying to make the point that the algorithm can, specifically, handle
> context-sensitive grammars; it is still somewhat fuzzy, Henrik,
> whether or not you got that point.
>...

Fuzzy?  Sorry! :-) I got that point.  But for Dutch you need at least
rules like

   a1 b1 ... an bn -> a1 .. an b1 ... bn

I don't see how this is handled for an unlimited n.

>...
> Do you think Dutch, or any other natlang, is not tree-like?
>...

I'm not sure whether I understand this question correctly.

I'm sure you can draw a tree for a given Dutch sentence, but the
formalism that generates or analyses such a tree in the general case
needs to be at least Tree Adjoining Grammars due to the syntax
structure of Dutch (and as usual for natlangs, such a simple formalism
is still limited, but let's neglect that).

But I'm quite sure Dutch is configurational, but it may depend on what
you require.  As I tried to say in my other post, intuitively I don't
see the Dutch cross-serial dependency (what I called 'funny verb
word') as a movement operation or any other complex 'post-pass' fix-up
of the syntax (I tried to express this my saying that standard
formalisms are inappropriate and that the brain must work somewhat
differently, because it's so easy to parse (!personal view)).  So the
Tree Adjoining Grammar approach pleases me because it shows that the
structure can be generated hierarchically with quite a simple
formalism.

**Henrik