Re: FrameNet and ontolex

As Alessandro says, there is already a FrameNet-OWL dataset, developed by STLab, as described in:

Nuzzolese A.G., Gangemi A., Presutti V. Gathering Lexical Linked Data and Knowledge Patterns from FrameNet. O. Corcho, M. Musen (eds.): Proceedings of K-CAP 2011, the Sixth International Conference on Knowledge Capture, ACM, 2011

Our model supersedes previous ones, and has been agreed with Berkeley people.

The OWL schema is at:

http://www.ontologydesignpatterns.org/ont/framenet/tbox/schema.owl

Concerning the mapping to OntoLex, I'll be more detailed in a next message, but:

On Jun 19, 2013, at 5:39:22 PM , Alessandro Oltramari <aoltrama@andrew.cmu.edu> wrote:

> Dear All,
> 
> I have been thinking about modeling FrameNet with the W3C Ontology-Lexicon Model (as formalized in the PDF document sent by Philipp on  June 7th). 
> 
> I came to the conclusion that FrameNet doesn't include any peculiarity we should worry about. 
> 
> Frames (e.g. "Motion") and their constituent Frame Elements (=semantic roles e.g.  Goal, Source, Path, Direction ,Theme etc.) need to be represented at the ontological level. In principle, they are kind of ONTOLOGY ENTITY in our model. Of course one could argue about the "place" of frames in a (reference) ontology. I will skip this discussion here, though I'd point you to Aldo's work on "OntoFrameNet" (and knowledge patterns) as well as to Jeff Scheffczyk's effort on porting FrameNet in OWL DL (https://framenet.icsi.berkeley.edu/fndrupal/fnbibliography/author/143) 

I am reluctant to do that: as explained in the paper, frames and frame elements are lexical senses, similar to WordNet synsets and senses. Certainly we can decide to interpret frames as classes and frame elements as relations (and we suggest a recipe in the paper), but this is beyond the OntoLex scope, which focuses on porting linguistic resources to linked data.

> 
> Frames and Frame Elements "evoke" Lexical Units (e.g. the frame "Motion" evokes the verb "travel" - and several others, of couse): accordingly, lexical units would simply map to LEXICAL ENTRY in the ontolex model. LEXICAL SENSE would represent the disambiguated sense of LEXICAL UNIT, e.g. travel#1 (assuming WordNet as lexical resource). Closing the loop, this sense would then be contained in the synset:	
> 
> • S: (v) travel#1, go#1, move#1, locomote#1 (change location; move, travel, or proceed, also metaphorically) "How fast does your new car go?"; "We travelled from Rome to Naples by bus"; "The policemen went from door to door looking for the suspect"; "The soldiers moved towards the city in an attempt to take it before night fell"; "news travelled fast"
> 

Actually it's quite different: lexemes evoke frames (or frame elements), while lexical units are lexico-semantic units that localize frames into one lexical citation form. Evoking is however a relation used only in literature. The FrameNet database only contains relations between lexemes and lexical units, and between lexical units and frames.

From the FN book:

"A lexical unit (LU) is a pairing of a word with a meaning. Typically, each sense of a polysemous word belongs to a different semantic frame, a script-like conceptual structure that describes a particular type of situation, object, or event along with its participants and props."

Which makes LU quite similar to Word Sense in WordNet.

> I think this example is pretty straightforward, though an interesting issue may raise: how do we model the "evoke" relation? 

If we want that, we just need to make a property chain: [(fnschema:lexemeLU O fnschema:lexUnitFrame) rdfs:subPropertyOf fnschema:evokes]. However, Berkeley sometimes accepts that also LUs evoke frames. Probably not clear enough about that.

> 
> If we accept that Frames are Ontological Entity, do we want to state that "evoke" is a special relationship holding between Frames and Lexical Entries or does it make sense to generalize and claim that every Ontology Entity evoke a Lexical Entry? If we go for the second option, then "evoke" will correspond to  the inverse of "denotes". 
> 
> 

As I said, evoking is no denotation. As FrameNet book says:

"Each lexical unit is linked to a semantic frame, and hence to the other words which evoke that frame. This makes the FrameNet database similar to a thesaurus, grouping together semantically similar words."

In summary (if I remember well the current OntoLex status):

fnschema:Frame rdfs:subClassOf ontolex:LexicalMeaning
fnschema:FrameElement rdfs:subClassOf ontolex:LexicalMeaning
fnschema:LexUnit rdfs:subClassOf ontolex:LexicalSense
fnschema:Lexeme rdfs:subClassOf ontolex:Word (Expression?)

Ciao
Aldo

> Best,
> 
> Alessandro
> 
> 
> 
> 
> Alessandro Oltramari
> Research Associate
> Psychology Department, Carnegie Mellon University
> 5000 Forbes Avenue, Pittsburgh PA 15213
> Tel.:  +1-412-268-6284 	Fax.: +1-412-268-2798     Mobile: +1-412-689-1514
> Homepage: http://fms.psy.cmu.edu/member/aoltrama 
> Twitter/Skype: oltramale 
> "There’s no such thing as the unknown– only things temporarily hidden, temporarily not understood.” (Capt. J.T. Kirk)
> "To dare is to lose one's footing momentarily. Not to dare is to lose oneself." (S. Kierkegaard)
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
>