Re: Linking to SKOS

Hi,

Some quick comments


On Fri, Oct 25, 2013 at 9:46 AM, Aldo Gangemi <aldo.gangemi@cnr.it> wrote:

> Some comments below
>
> On Oct 18, 2013, at 6:05:38 PM , Philipp Cimiano <
> cimiano@cit-ec.uni-bielefeld.de> wrote:
>
>  Dear all,
>
> before the summer break, we discussed some problems in linking the ontolex
> model to SKOS. I would like to make an initial proposal along three lines:
>
> 1) SKOS says: " The class skosxl:Label is a special class of lexical
> entities." (see http://www.w3.org/TR/skos-reference/#xl-Label). The
> particularity is that skosxl:Labels have a only 1 restriction on
> "literalForm", i.e. there is exactly one literal form for skosxl:Labels.
> Clearly, such a restriction is compatible with our model. Nevertheless, we
> could state that "skosxl:Label" is a subClassOf ontolex:LexicalEntry. It
> conforms to the first sentence at
> http://www.w3.org/TR/skos-reference/#xl-Label and documents the fact that
> skosxl:Labels are very specific kinds of lexical entities (entries) that
> make a number of assumptions that our more general class "LexicalEntry"
> does not make. As we say that skosxl:Label is a SubClass we do not have any
> implications and there are no implications from our side. The implication
> is more on the side of the people who use both the SKOS and the ontolex
> vocabulary. But the effect is nice I think, i.e. all skosxl:Labels become
> ontolex:LexicalEntries. In practice, we could also write a converter from
> SKOS to ontolex that converts every Label into exactly one LexicalEntry.
> Not ideal, but useful to create a bridge between both models.
>
> You mean ontolex:Form, right? A lexical entry can have multiple forms so
is not compatible with skosxl:Label (I believe this was already discussed
in a previous telco).

>
> Seems fine and appropriate
>
>
> 2) ontolex:LexicalConcept is defined as "the mental representation of the
> shared meaning of a collection of senses that can be exchanged in many
> contexts without substantially changing the meaning". We could definitely
> argue that ontolex:LexicalConcept is thus a subClassOf skos:Concept, which
> is "an idea or notion; a unit of thought.". I see no overcommitment here.
>
>
> Same with me. On another ground, I tend to see also lexical concepts and
> concepts in general as subclasses of semiotics:Meaning, but it's not
> relevant to define such link in the lemon-ontolex context.
>
This is already in the model... what is the issue here?

>
>
> 3) On skosxl:altLabel, skosxl:prefLabel and skosxl:hiddenLabel, what we
> could do is introduce an underspecified property ontolex:label
>
> and than say that skosxl:altLabel, skosxl:prefLabel and skosxl:hiddenLabel
> are subclasses of this underspecified label class that we introduce.
>
> Then we could add the following chain:
>
>  contains^- o sense^- o form o writtenForm subPropertyOf ontolex:label
>
> With that we would make clear that we can map our writtenForms to labels
> of some kind, but that the decision whether this is a prefLabel, altLabel
> and hiddenLabel is underspecified.
>
> Pragmatically we could thus even write a converter form ontolex to SKOS
> that uses this underspecified property ontolex:label.
>
> Assuming we agree to the two points above, it would be possible to just
use the SKOS-XL properties still, as we have that LexicalConcept ⊑
skos:Concept and Form ⊑ skosxl:Label.

If we wanted to model the semantics of SKOS's pref/hidden/alt label the
more obvious way to do this would be on the sense as the sense represents a
term used with a particular meaning, hence is the perfect location for
selectional restrictions such as this.

Also, the chain you propose would not work, writtenForm is a datatype
property

Regards,
John

>
> This should work
>
>
> OK, all from me for now. The emails I sent are long, but I think that the
> issue warrants a deeper discussion.
>
> Talk to all of you this afternoon at 15:00 (CEST).
>
>
> Hope I could make it. I'm in Australia at a dinner and for me that will be
> midnight.
> Aldo
>
>
> Best regards,
>
> Philipp.
>
> --
>
> Prof. Dr. Philipp Cimiano
>
> Phone: +49 521 106 12249
> Fax: +49 521 106 12412
> Mail: cimiano@cit-ec.uni-bielefeld.de
>
> Forschungsbau Intelligente Systeme (FBIIS)
> Raum 2.307
> Universität Bielefeld
> Inspiration 1
> 33619 Bielefeld
>
>
>