[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >

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

Re: [engelang] reformulating the core grammar


On Mon, Oct 8, 2012 at 6:50 PM, selpa'i <seladwa@hidden.email> wrote:

Am 08.10.2012 23:53, schrieb Mike S.:

> On Mon, Oct 8, 2012 at 9:00 AM, selpa'i <seladwa@hidden.email
> <mailto:seladwa@hidden.email>> wrote:
>
> Yesterday I tried writing a parser for Xorban using the
> documentation on
> wordpress. I noticed some things that confused me a little bit.
>
> Nice, what are you writing it in?

Prolog. It's working fine, but so far I can only run it on my own
computer. It would require some effort to get it to run online or to
make an executable file out of it because those are areas I'm not very
experienced in.

Hmm, if you post it somewhere, maybe I'll port it to Java. Maybe...
 

> Despite all that, the parser is working, albeit it's not yet the way
> I'd
> want it to be. I personally prefer strictly binary syntax trees, and
> the
> current BNF is not at all binary. So I'd have to rewrite some of it.
>
> I'd be happy to help rewrite the grammar if you wanted me to do so.

Well, I'm not sure. It's not necessarily difficult to rewrite it
sloppily :). I already had to rewrite some small bits to make it fit the
prolog syntax. And I added a "discourse" unit to allow for parsing texts.
Eventually, what I want to see is an actual syntax tree of xorban a
utterance that is purely binary, just like in standard linguistics,
using X'-theory. In some cases, this won't be so easy I think because
you have to decide what binds with what and what has presedence.
I uploaded an image of a simple parse tree that I just improvised at
http://www.selpahi.de/xorban_tree1.png
As you can see, I had to invent a new type of phrase that the binary
operator could then take as its complement. Such things need to be done
for each and every rule. And finding good names for these phrases is
also not so easy :P


mu'o mi'e la selpa'i

Nice.  Is there any difference between a sentence and a discourse unit?

I created a binary version of the syntax based on how I prefer to view the grammar here: http://loglang.wordpress.com/xorban/grammar/

I didn't bother to make the nested parentheticals conform to the binary syntax, though.  If you want to do that, you'd need to make a separate production for every unique place they appear; e.g.

TermP := Term P*

That'd about double the size of the syntax section of the grammar.  Jorge has gotten rid of the old "modifier" production which would have kept his version of the grammar more binary:

modifier := NA | LA formula

In my grammar that's been renamed a term, but it amounts to the same thing.  I do think that it's very helpful to view {binary operator + first complement} as a meaningful constituent that behaves in a manner similar to a unary operator.  It's a construction that has a fair number of reasonably close analogs in natural languages: objects/sumti, adjectives, adverbs, case tags.  Two formulas in a row doesn't really have any constituent-like analog in natural languages.

--
co ma'a mke

Xorban blog: Xorban.wordpress.com
My LL blog: Loglang.wordpress.com