Re: [lojban] Re: Regular Language

[Robin Lee Powell <rlpowell@digitalkingdom.org>]

On Sun, Sep 04, 2011 at 12:51:16PM -0700, ianek wrote:
> 
> 
> On Sep 3, 10:04 pm, Robin Lee Powell <rlpow...@digitalkingdom.org>
> wrote:
> >
> > Wait, what?  No.  No human language is context free, they are
> > all context sensitive in the chomsky hieararchy, if not actually
> > unrestricted.
> >  Seehttp://en.wikipedia.org/wiki/Context-free_grammar#Linguistic_applicat...
> > for some citations on that issue.
> 
> The article you cite says:  "Gerald Gazdar and Geoffrey Pullum
> have argued that despite a few non-context-free constructions in
> natural language (such as cross-serial dependencies in Swiss
> German[3] and reduplication in Bambara[5]), the vast majority of
> forms in natural language are indeed context-free." Also it says
> that Chomsky's arguments were disproved. Now the only argument for
> non-CFG-ness of natural languages is 'cause Chomsky said so. But
> he's not omniscient.

OK, umm, I think we're all talking about something very different.

Lojban parsing *includes part of speech information*.

When I say that "Lojban is PEG parseable", I mean that I can parse a
sentence and give you complete information about the role of each
word in the sentence, unambigiously.  You can even do that with a
CFG on subsets of Lojban.

This is *obviously* not the case with English; "time flies like an
arrow" cannot *remotely* be parsed down to the part-of-speech level
with anything short of a CSG that has access to the entire
conversation, and maybe not even then.

If what you and Graham are talking about is "I can create a CFG that
will generate that English sentence structure", then sure, I can
accept that English is generatively a CFG in that sense.

It's just that I don't care, at all.  :D

I'm interested meaningful parsing; formal generative grammar work
has never been interesting or relevant to me, even when I was taking
it in school.

So, if I hijacked the conversation, as I now think I probably did,
than I apologize.

Lojban probably is generatively a CFG, but since the resulting
parses would be completely ambiguous in the presence of any
terminator elision, I don't think that's something anyone has
actually tried or tested.  If someone wants to demonstrate, I'd be
vaguely curious.

> On Sep 4, 8:39 am, Robin Lee Powell <rlpow...@digitalkingdom.org>
> wrote:
> >
> > That's fascinating, because people seem to have no problem
> > internalizing Lojban's elidable terminators, which as I'd said
> > don't seem to be CFG-able (at least in a sane number of rules).
> 
> But consider that they're highly confusing at times, eg. when it
> comes to "kei kei kei". Personally I think that's because people
> are accustomed to use a constant memory (so there's no stack) for
> language processing, and let me remind you what class of languages
> is parsable in constant memory.

I don't have any reason to believe that limiting the depth would
change the problem; you're welcome to attempt to produce a
non-ambiguous CFG with that limitation, though.

-Robin

-- 
http://singinst.org/ :  Our last, best hope for a fantastic future.
Lojban (http://www.lojban.org/): The language in which "this parrot
is dead" is "ti poi spitaki cu morsi", but "this sentence is false"
is "na nei".   My personal page: http://www.digitalkingdom.org/rlp/

-- 
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to lojban@googlegroups.com.
To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.