English grammar in BNF?
Guido Reina
I am seeking the web searching a not elementary english grammar in BNF, all
the things I have seen are very basic. Could anybody tell me some link?.
Thanks in advance.
Bob
"Guido Reina" <guido...@tecnet-sports.com> wrote in message
news:9d6e3h$6n0...@news1s.iddeo2.es...
> Hi all.
>
> I am seeking the web searching a not elementary english grammar in BNF,
> all the things I have seen are very basic. Could anybody tell me some
link?.
There probably isn't one.
BNF only handles context free grammars. BNF is adequate for computer
languages, but not for natural languages.
English and all other natural languages are context sensitive and require a
more powerful means of describing the language. This was formally postulated
in 1957. However nobody, as yet, has devised such a grammar.
Bob
Charles
> "Guido Reina" <guido...@tecnet-sports.com> wrote
> > I am seeking the web searching a not elementary english grammar in BNF,
> > all the things I have seen are very basic. Could anybody tell me some link?.
> BNF only handles context free grammars. BNF is adequate for computer
> languages, but not for natural languages.
>
> English and all other natural languages are context sensitive and require a
> more powerful means of describing the language. This was formally postulated
> in 1957. However nobody, as yet, has devised such a grammar.
Right. So, for something more formal, there's XTAG, etc.
For BNF, if you insist ... http://www.srv.net/~ram/syntax.html
For systematic useful examples of this exotic language:
http://email.uwp.edu/~canary/eng382/
321450001
<guido...@tecnet-sports.com> skribis:
Ciao Guido, comme va?
Some time ago I found a formal grammar of English in the www, though I
think it was not BNF but another formalism. I still have the address:
http://www.cs.kun.nl/agfl/download.html
See under MEKO-grammar. I hope you find it useful.
Regards,
Marcos
Guido Reina
Yesterday, by accident, I found "xhpsg", which had a good PDF and many
files, now I have a good material for starting. I also have downloaded the
"agfl".
Thanks again.
Guido
Mike Maxwell
languages) floating around, of more or less completeness. None will be BNF.
Many are proprietary.
Fernando Pereira wrote a smallish one in Prolog in the early '80s for his
Chat-80 project (which was his PhD project). You can probably find Chat-80
at some AI repository, and it should be fairly portable. The grammar is in
the form of Prolog productions, as I recall.
There's a fairly high coverage, and equally less portable :-), grammar in
the appendix of a PhD dissertation from the late '80s. (Disclaimer: I wrote
most of the grammar, thought not the dissertation.) You can get it from
Univ Microfilms, I imagine:
Harrison, Philip.1988. "A New Algorithm for Parsing
Generalized Phrase Structure Grammar." PhD diss,
U of Washington.
As the title suggests, the grammar is (an engineer's version of) the GPSG
formalism.
I don't think the real problem with using BNF is the non-context freeness of
English (there's actually very little of that), but rather with things like
gaps left by wh-"movement." Also the fact that to do a grammar compactly
(500 rules or less, say) requires the use of morphosyntactic features
(person, tense, number...), along with feature unification (or s.t. similar)
at run time to enforce feature restrictions. But to use BNF, I believe you
have to multiply out those features at grammar specification time, which
would make for an incredibly large grammar.
And don't forget that once you get the syntax and lexicon down, the hard
problems (semantics and pragmatics) are still ahead of you.
--
Mike Maxwell
Summer Institute of Linguistics
Mike_M...@sil.org
Alpo Lind
http://pointti.vip.fi
Alpo Lind
Guido Reina <guido...@tecnet-sports.com> kirjoitti
viestissä:9d6e3h$6n0...@news1s.iddeo2.es...