Re: Lexicalization as collection of senses/onomasiological lexicon

+1 

by the way this model allows multilingual lexicalization, which is 
something we are interest in.

I guess that cardinalities are 0-to-many for both 'sense' and 'reference', 
right?

>From this model, it follows that ontology axioms (e.g. inclusion) are not 
reflected by lexical relations (e.g. hyponymy), which, I guess, hold among 
lexical senses. For instance, you may have synonyms referring to disjoint 
concepts. Is it correct \ desired (by others than me)?

Regards,

Guido Vetere
Manager, Center for Advanced Studies IBM Italia
_________________________________________________
Rome                                     Trento
Via Sciangai 53                       Piazza Manci 12
00144 Roma, Italy                   38123 Povo in Trento
+39 (0)6 59662137                 +39 (0)461 312345

Mobile: +39 3357454658
_________________________________________________



From:   "John P. McCrae" <jmccrae@cit-ec.uni-bielefeld.de>
To:     public-ontolex <public-ontolex@w3.org>, 
Date:   10/04/2014 13:29
Subject:        Lexicalization as collection of senses/onomasiological 
lexicon
Sent by:        johnmccrae@gmail.com



Hi all,

So there was an interesting discussion on the telco last week about the 
nature of "lexicalization"... I will try to make a summary/proposal.

We can currently represent the data as a collection of words (lexical 
entries) by means of the Lexicon object, and as a set of concepts as an 
OWL ontology, however there is no object for describing how a single 
lexicon lexicalizes a single ontology. This would be useful for metadata 
so that we can say how much coverage a lexicon gives relative to the 
ontology.

This "lexicalization" object that has proposed by Armando, is an object 
that describes the connection between an ontology and a lexicon, is a 
collection of pairs (Ontology Entity, Lexical Entry) or as we know them 
better Lexical Senses! Thus we can define the Lexicalization as a 
collection of senses.

Related to this is the fact that some lexicons (e.g., SALDO and arguably 
WordNet) are based around senses not words, and are thus onomasiological 
lexicons, that is the lexicon as a collection of senses, as opposed to a 
collection of words. To this end it may make sense to name the 
lexicalization something else, such as SenseLexicon, as its role as a 
lexicon of senses. In practice then the proposal is to have something like 
this:

Lexicon          o===== LexicalEntry
                             ||
                            sense
                             vv
Lexicalization   o===== LexicalSense
                             ||
                          reference
                             vv
Ontology    o===== Class/Property/Individual
       
Does this seem reasonable?

Regards,
John

IBM Italia S.p.A.
Sede Legale: Circonvallazione Idroscalo - 20090 Segrate (MI) 
Cap. Soc. euro 347.256.998,80
C. F. e Reg. Imprese MI 01442240030 - Partita IVA 10914660153
Società con unico azionista
Società soggetta all?attività di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
above)