On 19. feb. 2014, at 15:38, Richard Light wrote:
> On 18/02/2014 11:22, Řyvind Eide wrote:
>> Is it correct to say that we know of an event (a funeral) through the inscription, and that Fabius Honoratus played a role in that event? Or is it more to the point to say that the inscription claims that such an event happened? Or does the idea of an event in this case sound strange to the scholarly thinking around such inscriptions?
>> The good thing about event based modelling (and encoding) is that it makes it easy to clarify some of the role issues. The problem is that it may be a bit cumbersome to encode.
> Looking at <event> , I notice that it is a second-class citizen, in that it has to occur within a <person> or <place> (or a list of events). Surely this limits the extent to which one could do "event-based modelling" in TEI.
True. But in the context of a text (the context where I would use persName rather than person) I would use an <rs type=event>.
It may very well be that in a real use case I would use rs and various versions of the name element in the text and then do the event based modelling (whether it is stored in the TEI header or elsewhere) in another formalism. It all depends on the specific context of the work at hand.
In order to connect such a model (whatever formalism it is expressed in) to the text, the use of TEI elements to anchor the model to the textual source on which it is based is the main point, not how the linking is done.