W3C home > Mailing lists > Public > public-ontolex@w3.org > November 2012

Re: input document for discussion on Friday

From: Guido Vetere <gvetere@it.ibm.com>
Date: Thu, 1 Nov 2012 11:43:24 +0100
To: Philipp Cimiano <cimiano@cit-ec.uni-bielefeld.de>
Cc: "public-ontolex@w3.org" <public-ontolex@w3.org>
Message-ID: <OF57CB2333.1C4B194D-ONC1257AA9.00313303-C1257AA9.003AEBF1@it.ibm.com>
 Am 31.10.12 13:35, schrieb Guido Vetere: 
> the model of Sense as (sub)classes that I've recommended (as is 
> implemented in Senso Comune) would look like the following: 
> ex:lemon rdf:type ontolex:Lex. 
> ex:lemon ontolex:hasSense lemon_1. 
> ex:lemon_1 rdf:type (ontolex:Sense AND ontolex:denotes ONLY <http://
> dbpedia.org/page/Lemon>). 
> Of course, if you want you can introduce a named class, like 
> mylex:Lemon_as_fruit owl:equivalentClass (ontolex:Sense AND 
> ontolex:denotes ONLY <http://dbpedia.org/page/Lemon>) 
> to have: 
> ex:lemon_1 rdf:type mylex:Lemon_as_fruit 
> Can we consider this as another option? 

Philipp Cimiano <cimiano@cit-ec.uni-bielefeld.de> wrote on 31/10/2012 

> Guido,
>  yes, of course, this is an option. But where do you see the 
> advantage compared to directly saying that the sense is a subclass 
> of the class in question?


I think it depends on what ontology concepts and sense, respectively, 
represent in your system. If they are both representations of the same 
kind of things, then maybe there's not a clear advantage in the model I 
suggest. Actually, in many IT artifacts called 'ontology', concepts are 
indistinguishable from linguistic senses, so I understand where the 
question comes from.

On the other hand, if the ontology that you want to map with your lexicon 
is intended to be a 'theory of the reality' independent from any language, 
then you wouldn't say, for example, that the sense 'cat' in English is an 
instance of the concept 'cat' (or a subclass) and therefore is expected to 
have four legs and a tail. In other words, if you want 'Sense' to stand 
for a specific kind of things, then you need to be 'multiplicative', i.e. 
introduce an entity 'Sense' besides the entity you want it to refer to, in 
order to keep predication on linguistic facts in a distinguished place. 
Once again, if you are happy with saying that cats don't have four legs, 
but they are just said to have four legs (in this view, mereology is just 
another name of meronimy) then the distinction between senses and other 
classes may appear to you as a useless sophistry (if not a dangerous 

I can provide several arguments in favour of a multiplicative approach. 
One of them, as discussed some time ago, has to do with vagueness. But I 
think that we should be liberal with respect to different views of what 
senses are from an ontological standpoint. If I understand the discussion 
we have been doing so far correctly, we agreed to reify senses, which 
means, at least, having them as mediating elements in data structures that 
bring words and concepts together. Let's give a shape to this structure 
and allow different formal ontological interpretations for it. Then we may 
discuss pros and cons of each of them. 

Kind regards,

Guido Vetere
Manager, Center for Advanced Studies IBM Italia
Rome                                     Trento
Via Sciangai 53                       Via Sommarive 18
00144 Roma, Italy                   38123 Povo in Trento, Italy
+39 (0)6 59662137                 +39 (0)461 312312

Mobile: +39 3357454658


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
Societą con unico azionista
Societą soggetta all?attivitą di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
Received on Thursday, 1 November 2012 11:49:04 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:36:27 UTC