Re: Consolidating triple/edges -- named occurrence version

> On 5. Jan 2024, at 14:35, Niklas Lindström <lindstream@gmail.com> wrote:
> 
> Thank you for the further examples, and for this reference (timely for
> my freshly brewed cup of tea)! This is quite elucidating!
> 
> As far as I can see this confirms much of my intuitions; such that our
> "occurrences of triples" are really that which are "in virtue of" (or
> claims/entails/necessitates/projects//implicates/makes for) triples
> (one or many); such as events and any other "(sub)states of affairs"
> (but now I need to be careful again to avoid overreaching with words
> that may imply too much).
> 
> There may be something to these truth-makers which might be too strong
> though -- if I describe something as a truth-maker, do I state that it
> exists, and thus its implications? I suppose it depends on the nature
> of the thing (a document or an event for instance)? Do we want only
> the weakest form of them in the core semantics, and leave the stronger
> forms (I state an event thus what truths it makes) to additional
> semantics? (I may need more tea...)

This is going in strange directions and I’d like to try to ground the discussion a bit. I see one fundamental dichotomy, namely if a statement is
- asserted or 
- not asserted.

The possible semantics of these two options are manifold: 
- a statement may be asserted, but further statements may add detail to it, even crucial detail that make its assertion more useful in one context and a nuisance in another
- a statement may not be asserted, but further statements may describe under what circumstances it should be considered as asserted. Such circumstances might be Bob's worldview, the latest version, a certain validity timeframe, a credibility rating treshold, etc.

So essentially what the difference of "asserted" and "not asserted" boils down to is a default state of visbility, which may be enlarged or constrained by additional statements (and of course applications that make sense of such additional information).

RDF being a KR technology to integrate facts (i.e. data that is considered to be universally true) and focused on ease of use and simplicity, we should be careful not to introduce unessential vagueness into the core mechanism. Of course, understanding more precisely the different ways in which something may be unasserted but useful, and how to express such variations, is itself useful, and I’m all for appropriate tooling (vocabs and syntax)!

Sorry if I was just stating the obvious!

Best,
Thomas

> 
> Cheers,
> Niklas
> 
> On Fri, Jan 5, 2024 at 12:34 PM Franconi Enrico <franconi@inf.unibz.it> wrote:
>> 
>> Some reference from philosophy - if you want to get lost in it :-)
>> 
>> Guarino, N., Sales, T.P., Guizzardi, G. (2018).
>> Reification and Truthmaking Patterns.
>> In: Trujillo, J., et al. Conceptual Modeling - ER 2018. Lecture Notes in Computer Science, vol 11157. Springer.
>> https://doi.org/10.1007/978-3-030-00847-5_13
>> (attached)
>> 
>

Received on Friday, 5 January 2024 14:32:59 UTC