[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lojban] Re: X-bar, chomsky and lojban

Jorge Llambías, On 17/05/2006 03:26:

On 5/16/06, And Rosta <and.rosta@gmail.com> wrote:


For syntactic reasons. I.e. KU is head of KU-phrase. LE
is head of LE-phrase, which is complement of KU-phrase.
Selbri is head of Selbri-Phrase, which is complement of
LE-phrase.

The rationale is that KU-phrase = the distributionally defined
class normally called 'sumti'. Then KU-phrase can have any
of a range of complements (LE, LA, KOhA, etc. etc.).


KOhA and LA CMENE cannot take a {ku} though. Is the
idea that in those cases the KU-phrase is always realized
as null?


In the alternative analysis (in which KU is complement
of LE), you would not be able to dispense with the
class 'Sumti' and you'd have to define it disjunctively,
as "LE-phrase, or LA-phrase, or KohA-phrase, or ...".
Which is much more inelegant.


Here's an argument for the alternative analysis in which
Selbri-phrase is complement of KU-phrase and KU-phrase
is complement of LE-phrase:

Relative clauses can appear in three positions in a sumti,
as marked by the stars:

    le * broda * ku *

If we take the relative-clause as a complement of the
KU-phrase then all three can be accounted for if the
Selbri-phrase is also a complement of the KU-phrase.
But if the Selbri-phrase is a comp[lement of the LE-phrase,
then relative clauses have to be sometimes complement
of the KU-phrase and sometimes a complement of
something else (either of the LE-phrase or of the
Selbri-phrase.


A reasonable argument, though it loses some of its force if relative clauses can occur in other loci besides this three places -- i.e. in places other than within LE/KU phrases. This is because some other generalization is anyway needed to account for the distribution of relative clauses.

Note, BTW, that multiplicable elements such as relative clauses would count as adjuncts rather than complements. I.e. an XP that contains an XP and an adjunct. This entails that in the normal case, adjuncts occur further from the head than complements. (A potential problem for treating the pre-ku relative clause as directly belonging to the KU phrase.) In some apparent exceptions to this generalization, it is sometimes the case that a complement -- typically when long -- has been shifted to an adjuct position (as in "I read _ today [every single one of the books you lent me]").


> I don't know if X-bar allows the head and the complement
> to be infixed/circumfixed to each other, but if it does then
> I think LE.../KU/ should be considered as one element and
> the inner selbri as the other element.

No, this is not possible.


I don't like X-bar for Lojban then. :)


X-bar, at the level of genericity I've been taking it at, seems to be a (bogstandard & unesoteric) universal of natural language.

--And.