Re: Term decomposition in OntoLex

Hi John,

I like it!

Is there anything that disallows (or discourages) "identifies" to link 
components to arguments as well? As you know, I think it's important 
that it's possible to capture discontinuous decompositions like "has 
<#arg> inhabitants", e.g. as follows:

haveInhabitants#Root syntax:constituent
   [ a :VP ; syntax:constituent
     [ a :V ; syntax:identifies :have ] ,
     [ a :NUM ; syntax:identifies <#arg> ] ,
     [ a :N ; lexinfo:number lexinfo:plural ; syntax:identifies 
:inhabitant ]
   ] .

Or should there be another way to do this?

Regards,
Christina


On 01/17/2014 06:27 PM, John P. McCrae wrote:
> Hi all,
>
> I wrote a section on representation of term decomposition here
>
> http://www.w3.org/community/ontolex/wiki/Syntax_and_Semantics_Module#Term_Decomposition
>
> The summary is that I propose we introduce a class called /Component/, 
> which represents a part of a lexical entry. We link lexical entries to 
> components by means of a /constituent/ property, and if the component 
> can be represented by a lexical entry we say the component /identifies 
> (is identified by?) /that entry. For convenience, there is a property 
> /subterm /that shortcuts the chain to allow an entry to entry linking. 
> Finally we allow components to be constituents of other components to 
> enable the representation of general phrase structures.
>
> Any comments?
>
> Regards,
> John


-- 
Christina Unger, PhD
AG Semantic Computing
Universität Bielefeld

Office: CITEC 2.311
Phone: 0521 106 12224