Multilingual thesaurus or ontologies and Topic Maps

For italian people see also the original post and the review of this thesis.

I dedicate this post to the Joint Research Centre – European Commission of Ispra and to Carlo Ferigato who selected for this work one of my fellow student…
Well, she is certainly a skilful cataloguer, but I don’t know if she is able to manage thesaurus or ontologies.
I must say great game and good luck mates 🙂
Maybe this post could also help her (and them) ;-).

This post is about multilingual thesaurus or ontologies.

The problem is underlined in several translations of Vanda Broughton’s manual Essential Thesaurus Construction: hierarchical relations (narrower term  and broader term) could be different according to the language.

A simple example.

“Rabbits” in an Italian thesaurus will be under farm animals (“animali da allevamento” in Italian), while in an English thesaurus it will be under pets (“animali da compagnia” in Italian)

So the translation it’s not limited to the term (rabbits [English] – conigli [Italian]), but involves also the structure itself of the thesaurus.

This is a problem partially different from that on asynchronous multilingual thesaurus underlined on the Guidelines for Multilingual Thesauri by IFLA.

With Topic Maps is quite simple managing this problem, using scope construct that you can apply both on term’s names (Topic names) and also on hierarchical relations (association in Topic Maps language), creating in this way real linguistics facets that determine both the translation of terms and also the relations themselves.

It will be a matter of the graphic interface to enhance this kind of browsing.

This is not a banal poly-hierarchy (deprecated by some thesaurus such the EuroVoc), but we have to link the structure itself of the thesaurus with the language facet.

You can model this behavior with a software as Onotoa:

Obviously it isn’t a complete model of a thesaurus (it lacks associative relationships, reject forms and so on)

Who doesn’t like UML style of Onotoa can watch following example expressed in gtmalpha and the corresponding ontology expressed in a new syntax that I called gtmalpha+ (my proposal to adapt original notation of gtmalpha to the grade 1 of GTM, that is to the ontological level).

ONTOLOGY IN GTMALPHA+

EXAMPLE IN GTMALPHA

A software as Topincs (even if it’s not build specifically to manage thesaurus) it’s a good starting point for similar experimentations.