[Corpora-List] Moving Lexical Semantics from Alchemy to Science

[Ted Pedersen]

What a fun thread. :)

On another source of compounds - WordNet actually has a rather large
number of compounds - there are, I believe*, 64,621 compounds found in
WordNet, and the vast majority of them just have one sense. There are
2,701 ambiguous compounds in WordNet, of which 1,895 are nouns.

One conclusion you can (and should) draw from this is if you are
designing a WSD system using WordNet senses, it pays to identify
compounds correctly. :)

In any case, I put some lists up that might be fun to look at....

http://marimba.d.umn.edu/WordNet_Compounds/

There are three files that include compounds and their glosses....

wn-ambig-compound-all.txt	- all the ambiguous compounds (more than 1 sense)
wn-ambig-compound-nouns.txt - all the ambiguous nouns
wn-compounds-all.txt - all the compounds

* I gathered this data very quickly, so please don't rely on it for
anything more than conversation. I'm happy to double check or explain
how I did this to anyone who might have more serious interests.

And of course, WordNet compounds is a rather generic category - names,
phrases, idioms, etc. are all included, and often time the
distinctions made in the ambiguous cases are razor thin....

But, it's still kind of fun to look at I would say.

Enjoy,
Ted

On Thu, Jan 27, 2011 at 5:57 AM, John Williams
<j0hnwh0ever.corpora at gmail.com> wrote:
>
>
>> ... P.S. Anyone have some other ambiguous open compounds they are familiar
>> with, besides 'solar system'?
>
> 'golf club' springs to mind
> j0hn
>
> -----------
>
> John Williams
> Lecturer in English Language and Linguistics
> School of Languages and Area Studies
> PK 2.18, University of Portsmouth
> Portsmouth PO1 2DZ
> Tel: (0239 284) 2162
> Email: john.x.williams at port.ac.uk
> _______________________________________________
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
>
>



-- 
Ted Pedersen
http://www.d.umn.edu/~tpederse

_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora