Philipp, Guido,

Guido would like to transform a sense lexicon "on the fly", making it a full-fledged axiomatized ontology just because it is mapped to an axiomatized ontology.

Now, we can obtain the same result with a two-step process, which I recommend usually when such problems arise with heterogeneous knowledge:

a) use a simple model to query (and modestly reason) over ontology+lexicon: with (a) most ontolex use cases will be satisfied, including Guido actually (ehm)

b) make a refactoring of the lexicon as an axiomatized ontology by using a "copycat" process: substitute senses to ontology entities (or make them equivalent to), and reason in regular DL mode

For (a), that's my attempt: I designed a simple OWL ontology (attached) for the example of Guido's, in a way that differs from both Guido's examples (can be considered an extension of example 1). It implements the basic pattern we are talking about (LexicalItem-hasSense->Sense-representedBy->OntologyEntity), augmented with a DomainEntity class and some properties between senses: specializes, and its subproperty meronymicallySpecializes. Now, instead of attempting any HODL reasoning, more modestly I show how to get a semantic field with a simple SPARQL query:

Now the ball passes to Guido: when should we need (b), if your use case is satisfied already with (a)?

Ciao

Aldo