RE: summary of state-of-play

Dear all, sorry for the silence of these last days..

 

...I would make things more complicate by adding a few things on the
discussion :-)

 

First of all, I would reply on the commitsTo. Agree on changing it.

Also, a part from the fact that we may have some inverse properties, a few
things which I would maintain for sake of understandability and ease-of-use
are:

 

1.       The directed path: Word-->Lexicalsense-->Lexicalconcept-->OntEntity
(so, beyond any name’s choice,  we need an inverse of subsumes, and to me it
is more important and of common use than the opposite direction)

2.       OntEntity --> LexicalConcept . (As I still imagine people willing
to simply decorate ontentities with lexical concepts from a resource like
wordnet)

 

Now, the added things:

 

1.       IMHO, we should be more clear about what a Lexical Concept is.
While I myself initially pushed for this name, as I felt this was the most
ideal subsumer of things like synsets, I wanted to at least try a step back
in understanding what is this real difference from simple skos:Concept (I
sent an email on 14/06), as we should motivate its existence, and other
“surrounding” things would come out easily.. 

a.       I read in a past email (maybe from Philipp?) that a Lexical Concept
is a concept which is lexicalized, but I would avoid such a definition. In
these terms, almost any skos dataset that I know ok, would contain
LexicalConcepts instead of skos:Concepts. Lexical Concepts are instead a
sort of semantic backbone of language. See, apart from the “5 papers on
WordNet“, which just hints at them, also [1] and [2] (in particular start of
section 3 of 2, and also this extract: “Rather, they are units of linguistic
knowledge abstracted from across usage events (i.e., utterances) that encode
linguistic content and facilitate access to conceptual (i.e.,
non-linguistic) knowledge” ).

b.      If we feel clear any difference, then LexicalConcepts are ok (and we
should then understand the differences with respect to, let’s say, highly
lexicalized conceptual structures such as Agrovoc, Eurovoc or GEMET),
otherwise, it is difficult to motivate their existence (that is, it is not a
naming problem).

2.       I’m still really doubtful about “LexicalSense”. I would strongly
vote for “Sense” alone. Almost all the literature I read on this (but I’m
not a linguist!) speaks about senses and it is also nn clear to me what
adding Lexical means. Btw, not willing to create more entropy than
clarifications, so if it “frozen” now, I can live with it!



I’m not voting for the other properties…still thinking about them..

 

Best,

 

Armando

 

[1] VYVYAN EVANS  "Lexical concepts, cognitive models and
meaning-construction",2006,"Cognitive Linguistics"

[2]
<http://www.vyvevans.net/On%20the%20nature%20of%20lexical%20concepts.pdf>
http://www.vyvevans.net/On%20the%20nature%20of%20lexical%20concepts.pdf

 

 

From: Philipp Cimiano [mailto:cimiano@cit-ec.uni-bielefeld.de] 
Sent: Thursday, June 27, 2013 4:26 PM
To: Aldo Gangemi
Cc: John McCrae; Guido Vetere; public-ontolex@w3.org
Subject: Re: summary of state-of-play

 

Dear all,

     I see three issues to be discussed tomorrow:

1) commitsTo (in which direction to use it?); actually, given the discussion
that we had about this, I would propose to replace it by something less
controversial
2) evokes Relation: how to name it, please make up your mind
3) subsumes: that was meant by mean as sth. like "contains" indeed. I agree
with John's statement. The LexicalSense is a particular sense of a word and
the other (Lexical Concept) is a collection of lexical senses. A single
lexical sense can not be a collection of senses at the same time, right? A
singleton set is clearly a set, but an element can not be a set, it is
contained in it. 

Btw. we can not use refers between Lexical Entry and Lexical Concept. We
have always said that "reference" should have an extensional object as
range. Lexical Concepts are not extensional. 

I feel that if we fix these things we are mostly done with the core model.

I will add some definitions to the ontology axioms and definitions tonight
to the ontology and send it around for tomorrow.

Best regards,

Philipp.

Am 26.06.13 00:23, schrieb Aldo Gangemi:

Dear John, I simply assumed the "subsumes" relation as the inverse of
subClassOf, as with the typical meaning of subsumption. In that case,
LexicalSense and LexicalConcept cannot be disjoint.

 

Now, you're pointing to a different requirement, i.e. that the meaning of
one word (a word sense) is inherently different from the meaning of an
equivalence class (not just a collection) of words (a synset). 

However, I do not see any reason for disjointness. A lexical sense can be
easily seen as an extreme case of a lexical concept, where the equivalence
class is constituted just by one word (actually a lemma). 

In this way, the axiom <LexicalSense subClassOf LexicalConcept> is perfectly
valid, as well as the derivative axiom <sense subPropertyOf evokes>.

 

I think this view simplifies the model, but if you have counterexamples or
conflicting requirements, please let's discuss it.

 

Aldo

 

On Jun 25, 2013, at 6:35:58 PM , John McCrae
<jmccrae@cit-ec.uni-bielefeld.de <mailto:jmccrae@cit-ec.uni-bielefeld.de> >
wrote:





Hi, 

 

Quickly I agree that the commits to should be pointed from the concept to
the ontology.

 

Aldo suggests that "sense" is a subproperty of "evokes"... I am puzzled as
this would lead to a contradictory ontology as the range of "sense" is
LexicalSense and the "range" of "evokes" is LexicalConcept, but LexicalSense
and LexicalConcepts should be disjoint as a LexicalSense is a particular
meaning of a single word, where as LexicalConcept is the meaning of a
collection of words (i.e., a synset). We should avoid creating any confusion
between lexical sense and lexical concepts as they are quite different
objects with different roles in the lexicon-ontology model.

 

Regards,

John

 

 

On Tue, Jun 25, 2013 at 6:14 PM, Aldo Gangemi <aldo.gangemi@gmail.com
<mailto:aldo.gangemi@gmail.com> > wrote:

Dear all, "ontological commitment" means that someone commits to the
existence (in some universe of discourse) of certain entities whose type is
given by a name. Therefore I agree with Guido here: if we have to use
"commits to", the direction should be reversed.

On the other hand, the notion is quite controversial and laden with
philosophical debates about ways to establish the actual existence of
committed entities, and I suggest we ignore it here.

I'd just delete it: the "reference" relation is enough I guess, and can be
assumed to hold between any kind of intensional entity and (extensionally
seen) ontology entities.

 

One more thing: we probably need to make "sense" a subproperty of "evokes".

 

Aldo

sent by aldo from a mobile


On 25/giu/2013, at 17:19, Guido Vetere <gvetere@it.ibm.com
<mailto:gvetere@it.ibm.com> > wrote:

Philipp, 

In my view (but we may ask) Guarino et al (following Quine) talk of the
specification of the commitment for a vocabulary of predicates, which are
substantially logic-linguistic symbols (as is in the tradition of analytic
philosophy). According to authors, such a vocabulary comes with an implicit
ontology, but due to polysemy, vagueness, etc, of the linguistic rendering,
the intended models of such vocabularies should be (case by case) specified
by a set of suitable constraints. The specification of such constraints is
what they refer to as the 'formalization of an ontological commitment'. 

Now, I think that in Guarino's work, Ontology Entity and Lexical Concept are
melted together in the logic vocabulary, so we cannot draw a clear
conclusion from  there. If I had to choose a direction for 'commitsTo'
between Ontology Entity, Lexical Concept, I would say that a Lexical Concept
commits to an Ontology Entity. The other way around wouldn't make sense to
me. 

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
+39 (0)6 59662137 <tel:%2B39%20%280%296%2059662137>                  

Mobile: +39 3357454658 <tel:%2B39%203357454658> 
_________________________________________________ 




Philipp Cimiano <cimiano@cit-ec.uni-bielefeld.de
<mailto:cimiano@cit-ec.uni-bielefeld.de> > 

25/06/2013 15:43 


To

Guido Vetere/Italy/IBM@IBMIT 


cc

public-ontolex@w3.org <mailto:public-ontolex@w3.org>  


Subject

Re: summary of state-of-play

 

		




Guido, all,

in his 1994 AAAI Paper
(http://www.mit.bme.hu/system/iles/oktatas/targyak/7412/Formalizing_Ontologi
cal_Commitments.pdf
<http://www.mit.bme.hu/system/files/oktatas/targyak/7412/Formalizing_Ontolog
ical_Commitments.pdf> ) Guarino talks about " an ontological commitment for
L" where L is a logical language. For me, it thus seems natural to see the
ontological commitment as a "property" of language L. Under this view, it is
the vocabulary that is in the domain of the commitsTo property and the
"conceptual relation" is in the range. 

But of course this is quite arbitrary. We need to define it properly I
agree.

See below...


Am 25.06.13 15:30, schrieb Guido Vetere: 
Philipp, 

If I remember well, the notion of 'ontological commitment' is also known in
Quine's philosophy, denoting the kind of thing that must exist in order for
an expression to denote something.  If this is also our notion, then I think
that the arrow should lead from the lexical class to the ontological one,
not the other way around. 

Some question about the model. 

Is 'denotes' equivalent to sense°reference? If yes, it should be noted
somehow. 

Yes 

The relation 'subsumes' is obscure to me: is it the inverse of is-a? 

No, it means that a particular lexical concept (e.g. a synset) subsumes or
includes the particular sense of a word. If you have a better way of naming
this, please say so! I feel we do not yet have the ideal name for it. For
example, a synset (as a lexical concept) includes not really a word, but a
sense of a word. 
Is 'evokes' (whatever it means) related to sense°inverse-of-subsumes? 

Yes, it is equivalent to sense o inverse-of-subsumes 

Thank you and apologize if the answer is already there .. 


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
+39 (0)6 59662137 <tel:%2B39%20%280%296%2059662137>                  

Mobile: +39 3357454658 <tel:%2B39%203357454658> 
_________________________________________________ 


Philipp Cimiano  <mailto:cimiano@cit-ec.uni-bielefeld.de>
<cimiano@cit-ec.uni-bielefeld.de> 

25/06/2013 15:04 

 


To

 <mailto:public-ontolex@w3.org> public-ontolex@w3.org 


cc

	

Subject

Re: summary of state-of-play

 

		





Elena, all,

well, I used "commitsTo" in the sense of Guarino in order to say that a
certain symbol in an ontological vocabulary refers to (commits to) some
conceptual relation in a conceptualization, the conceptualization being
essentially "intensional" and not directly accessible (e.g. in the head of
someone, implicit in a certain community).

I used commitTo to avoid using again something like "reference" which would
otherwise become quite overloaded.

Aldo can elaborate on this much more than me, but I hope the intuition
behind using commitsTo is clear now.

Along these lines, commitsTo can also be established between an ontological
entity (extensional) and a skos:Concept (intensional)

But I agree with Aldo that skos:Concept is the more general class and that
skos:Concepts need not be lexicalized. Under this understanding
ontolex:LexicalConcept is a subclass of skos:Concept in the sense of being a
special skos:Concept that is lexicalized.

Hope this clarifies my intuitions.

Best regards,

Philipp.

Am 25.06.13 13:40, schrieb Aldo Gangemi: 
Hi Elena, 

On Jun 25, 2013, at 1:19:49 PM , Elena Montiel Ponsoda <elemontiel@gmail.com
<mailto:elemontiel@gmail.com> > wrote: 

Dear Philipp, all,

Thanks for the "state-of-play" document and the summary of the document at
http://www.w3.org/community/ontolex/wiki/Specification_of_Core_Model 

I just went through it and in general I agree with the model proposed. 
I have two comments that we may discuss on Friday. 

*	what is the meaning of the "commitsTo" relation? Could it also be
established between an OntologyEntity and a skos:Concept? 
*	I am not sure I fully understand the relation between LexicalConcept
and skos:Concept (sorry if you already discussed it!!). Wouldn't a
LexicalConcept be also subsuming a skos:Concept? I think a LexicalConcept is
somehow more general, am I mistaken?

Quickly: I think not. SKOS is very general and includes all sorts of
concepts, be them lexical or not. 
Aldo 
Talk to you on Friday!
Elena 

El 21/06/2013 15:30, Philipp Cimiano escribió: 
Dear all, 

we had a very short meeting today. Apologies for the very late announcement
on my side. I will announce the meeting earlier next week. 

In any case, we agreed that it is good that the model as it stands can
accomodate both the view of Frames as Extensional Entitites / Class (i.e.
sets of situations) and the view as intensional/cognitive Lexical Concepts. 

I feel that we need not to adopt any strong position towards any of these
ends as FrameNet has been anyway modelled by different people in OWL/RDF
already (Aldo, Alessandro, etc.) and it is certainly not the main use of the
ontolex model. 

In any case, the (short) minutes from today are here:
http://www.w3.org/2013/06/21-ontolex-minutes.html 

We will talk again next week at the usual time slot. 

Please all read my document and inspect the OWL ontology. We will decide on
this core very soon ;-) 

Have a good weekend, 

Philipp. 


-- 
Elena Montiel-Ponsoda
Ontology Engineering Group (OEG)
Departamento de Inteligencia Artificial
Facultad de Informática
Campus de Montegancedo s/n
Boadilla del Monte-28660 Madrid, España
 <http://www.oeg-upm.net/> www.oeg-upm.net
Tel. (+34) 91 336 36 70 <tel:%28%2B34%29%2091%20336%2036%2070> 
Fax  (+34) 91 352 48 19 <tel:%28%2B34%29%2091%20352%2048%2019>  



-- 
Prof. Dr. Philipp Cimiano
Semantic Computing Group
Excellence Cluster - Cognitive Interaction Technology (CITEC)
University of Bielefeld

Phone: +49 521 106 12249 <tel:%2B49%20521%20106%2012249> 
Fax: +49 521 106 12412 <tel:%2B49%20521%20106%2012412> 
Mail:  <mailto:cimiano@cit-ec.uni-bielefeld.de>
cimiano@cit-ec.uni-bielefeld.de

Room H-127
Morgenbreede 39
33615 Bielefeld 

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) 


-- 
Prof. Dr. Philipp Cimiano
Semantic Computing Group
Excellence Cluster - Cognitive Interaction Technology (CITEC)
University of Bielefeld

Phone: +49 521 106 12249 <tel:%2B49%20521%20106%2012249> 
Fax: +49 521 106 12412 <tel:%2B49%20521%20106%2012412> 
Mail:  <mailto:cimiano@cit-ec.uni-bielefeld.de>
cimiano@cit-ec.uni-bielefeld.de

Room H-127
Morgenbreede 39
33615 Bielefeld 

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)

 

 






-- 
Prof. Dr. Philipp Cimiano
Semantic Computing Group
Excellence Cluster - Cognitive Interaction Technology (CITEC)
University of Bielefeld
 
Phone: +49 521 106 12249
Fax: +49 521 106 12412
Mail: cimiano@cit-ec.uni-bielefeld.de
<mailto:cimiano@cit-ec.uni-bielefeld.de> 
 
Room H-127
Morgenbreede 39
33615 Bielefeld