John wrote: "In any case, N-N patterns are among the most difficult
syntactic constructions to interpret for NLP systems, primarily because
they are so idiosyncratic."
Agree.
Compound nouns are too peculiar sometimes, regardless being the ordered
sequences where the anteceding word modifies or describes the consequent
one, thus indicating what kind of thing (object, state, action or relation)
it is.
Seemingly the peculiarity comes from the rule that a compound noun could be
formed in many/any ways:
Noun +Noun; Verb + Noun; Adjective + Noun; Preposition + Noun; Noun + Verb;
Noun + Preposition;
Adjective + Verb (dry-cleaning); Preposition + Verb (output or input) (01)
I stopped looking for NLP systems some time ago. Now wonder if any
successful research in the language machines programmed to identify a type
of noun, if it is abstract or concrete, common or collective, countable or
uncountable, simple, complex or compound. Seemingly, there must be something
for the proper names as named entities, due to capital letters. And if any
LM, say, capable to discern collective nouns as naming groups of things,
people, animals, or a group of specific things, like furniture, crockery,
etc. (02)
Azamat
PS: It was a similar thread on the SW forum, "Naming convention for deeply
structured classes and subclasses", interesting to look:
http://lists.w3.org/Archives/Public/semantic-web/2009Jul/0140.html. (03)
----- Original Message -----
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
To: <ontolog-forum@xxxxxxxxxxxxxxxx>
Sent: Friday, June 03, 2011 12:46 AM
Subject: Re: [ontolog-forum] Bad language - no biscuit. (04)
> Leo and Simon,
>
> As Leo pointed out, the phrase "attorney general" is actually
> a noun-adjective phrase, using a grammatical pattern that is
> based on French word order.
>
> Leo>
>> In general, hyphenated or string-concatenated noun-noun compounds
>> are clearer, but because they are lexicalized, they often have
>> idiosyncratic meanings. Example: broomstick, egghead, night-rider.
>
> Actually, even those compounds follow the pattern that the last
> noun is the head and the preceding noun modifies it:
>
> broomstick -- a stick of a broom
>
> egghead -- this is a combination of a metonym (using 'head' to
> refer to a person with a focus on the head) and using the
> word 'egg' as a deprecating modifier of the head of that person.
>
> night-rider -- a rider who rides at night (possibly with other
> connections to the night).
>
> In any case, N-N patterns are among the most difficult syntactic
> constructions to interpret for NLP systems, primarily because
> they are so idiosyncratic.
>
> John
>
>
>
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> (05)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (06)
|