Definition of I4 Proposition Set and what an instance of I2 Belief is about

Post by Martin Doerr (17 September 2023)

Dear All,

I made a mistake with the Issue number: The below message is actually a proposal for issue 614! So far, it has been formulated as:
 

"Some ideas:

• An Attribute Assignment essentially talks about a single property instance, which forms a parallel to pointing to a named graph that contains one property instance. 
• I1 Argumentation which results in an instance of I2 Belief cannot be a subclass of E13.
• There is also a problem with S4 Observation if the reification construct deals with more than one property simultaneously (what is now referred to as Situation). 
• The CRM set of properties that are equivalent to a named graph can be represented as a set of propositions and the connection between them in FOL. One can always name that and say that the predicate “X” stands for a proposition, in CRM an instance of E89. On the other hand, the scope note of I4 Proposition Set explicitly refers to binary propositions and formal ontology concepts, which seems too restrictive. Logical constructs do not specify the order or mode of the logical system the statements are expressed in. This could yield propositions that are incompatible with the CRM. The scope note needs redrafting.
• The idea is that CRMinf can be connected to a CRM compatible knowledge base through such statements. As CRMinf stands now, it seems that many things that we regard as premises and conclusions won’t be formulated in properties that have been defined in a formal ontology. In general, the scope of CRMinf should be broader than what can now be expressed. The reference to name graphs should explicitly only leave room for named graphs that contain 1+ property instances only (rather than instances of properties and classes or just the one property instance permitted by E13). 
• If there is any formalization in FOL that can be used to declare the E13 reification of a single property as a specific case of a named graph, it would be interesting to look at it. "

I hence reformulate:

1. We need a connection of an I4 Proposition Set (represented in a KB via a URI) to a content model.
2. If we follow the logic of P190, we can connect the proposition set to a materialized copy of an information object. The latter, in contrast to P190, will not be interpreted as representing an immaterial symbolic form of a certain symbolic level, but as representing an immaterial propositional form (see examples for E89 Propositional Object). This requires the latter to be either a set of statements in a KR language, OR in sufficiently unambiguous textual form (in the context and for the purpose of reference).
3. We need 3 kinds of links: To E52 String, to a symbolic object, and to an external file, be it in RDF or clear text.

• I propose "Jxx1,2,3 has model representation (represents model of): E52 String / E73 Information Object / E24 Human-Made Feature 
  • Jxx2 would be used for KR file contents, but also unambiguous texts, (P2 has type, such as “the Nebra Sky Disc dates to the Early Bronze Age” (Pernicka et al. 2020)"). Ontologically, the  "{}" connecting a URI in TRIG with a Named Graph corresponds to the Jxx2 link.
  • Jxx3 may point to a URL or archival identifier. Jxx3 is a shortcut of Jxx2 - Pxxx has representative copy.

An instance of I4 Proposition Set should be regarded to be neutral to the relationship to reality, as any Information Object. The relationship to reality is determined by the link using the proposition set: 

1. In the context of I2 Belief, the belief value of "TRUE" means "real", if the propositions can be related to reality (i.e., are about real world items). "FALSE" means at least one of the propositions in the set are regarded to be not real. NOTE that in a set of inconsistent propositions, the false one may not be known! Belief values expressing possibility or probability mean "possibly real" if the propositions can be related to reality.
2. In the context of observable situations, the link between the observation activity and the situation implies the degree of reality assumed. It makes sense to observe that a situation did not hold (in its specified area and time).
3. Only proposition set held to be true by the maintainers of a KB can be introduced as Named Graphs, because they become part of the stated knowledge.
4. Proposition Sets held to be possibly true by the maintainers of a KB can be introduced as Named Graphs, if the operation of the KB foresees filtering by provenance and likelihood. In this case, Named Graphs are  particularly effective.
5. implementation rule/ constraints for typing Named Graph URIs are required.
6. Finally, observing or assigning a single property (E13) can be described as a shortcut to a content model held to be true....

I propose to decide, if this is the way to go forward, and then

• Spell out the properties
• Adjust scope notes of I4, Observation etc.
• write the FOL connection between single property assignment and I4.

Best,

Martin

In the 57th CIDOC CRM & 50th FRBR/LRMoo SIG Meeting, MD walked the SIG through a proposal on how to connect an I4 Proposition Set to a content model. 
His proposal involved the following: 

• Introduce a set of properties pointing from I4 Proposition Set to E62 String, E73 Information Object, E25 Human-Made Feature and provide their definitions. 
• Define the relationship of an I4 to reality through I2 Belief. 
• Define E13, which forms a single property assignment, as a specific case of I4 Proposition Set through FOL. 
• Define S4, which forms a single observation, as a specific case of I11 Situation through FOL.
• Adjust the scope notes for I4, S4

Discussion points:

• Concerning the relation to reality: the KB should allow one to express statements that were previously considered to be true, but are now considered to be false by its maintainers, as this allows one to observe the evolution of the statements within the KB. The alternative of creating snapshots of the KB (instances of E73) and linking to them through the content model is OK, but should not be the only choice. 
• Content model seems a bit obscure as a term. It needs to be properly defined. 

Decisions:
The SIG voted in favor of introducing a content model (analogous to P190) and discuss the I4 Proposition Set’s relationship to reality as sketched in MDs HW and provide the scope notes for the properties. 

HW: PF (to lead), MD, GH, AGr, & SdS (to proofread).
 

In the 58^th CIDOC CRM SIG and 51^st FRBR/LRMoo SIG meeting, PF presented a proposal for a new content model concerning I4 Proposition Set and Ixx Single Proposition Set, plus how it connects to the CRMbase model. For the details of the proposal (joint HW by MD and PF) see here.

The SIG proceeded with reviewing the proposed definitions for I4 Proposition Set (redrafting) and Ixx1 One-Proposition Set. Details of the proposal and the SIG’s feedback can be found in the attached document.

Overall discussion points:

• Property needs relabeling: Jxx2 is formal meaning of ⇒ Jxx2 is a formal interpretation of. One cannot ascribe one single meaning to an information object.
• The purpose of introducing the  Ixx1 One-Proposition Set modeling construct and connecting it with both E13 Attribute Assignment and I4 Proposition Set is twofold: first to allow single statements to be the content of an instance of I2 Belief and ascribe them a truth value.
  
  • Difference of opinion registered wrt the necessity to introduce Ixx1 One-Proposition Set and properties to talk about single attributions and why not E13 is adequate for that in and of itself.
  • The construct has not been grounded on some particular practical example; it was rather an attempt at resolving the incompatibility of the reification construct (which only supports a single statement) with knowledge graphs.
    
    • Nb. The scope-note for Ixx1 One-Proposition Set (see appendix) makes it clear that the E13.P140|P141: E1 and E13.P177:E55 typically shortcut over the full path will be defined in a knowledge base.
• Second, to indicate that an instance of E13 Attribute Assignment forms a particular kind of argument, i.e., one that is considered to be true. The shortcut is used, when the belief value of the singleton proposition set is true.

Summary of decisions:

1. New scope note for I4 Proposition Set, needs redrafting to read nicer.
2. Examples for Ixx1 One-Proposition Set need to be added –to consult the examples that AG provided. See here for slides and paper
3. Properties need examples before they get admitted in the model.

The HW is assigned to the same group (MD, PF, GH; AG to collaborate for the examples; SdS to proofread).

Paris, March 2024

Post by Eleni Tsouloucha (6 September 2024)

Dear all,

In view of the massive reorganization that CRMinf has been undergoing, and the need to issue a stable version of it sooner than later, I’m sending you
the proposed changes we’ll be discussing and voting for at the SIG meeting. 

You can refer to the (updated [1]) specification document for a better grasp of the ensuing model. It goes without saying that any
comments you have are welcome.

Α summary of the proposed changes can be found below: 

   1. Introduce class One-Proposition Set
   2. Introduce property is encoded by
   3. Introduce property has unambiguous description (describes the formal meaning of)
   4. Introduce property that the formal meaning of (has a meaning belief)
   5. Introduce property contains entity (is contained in) 
   6. Introduce property contains property type (is property type in)
   7. Introduce property has domain (is domain of)
   8. Introduce property has range (is range of) 
   9. Introduce property has property type (is property type of)
   10. Introduce property assigned proposition (is assigned by) 
   11. New example for I2 Belief
   12. Introduction (Scope update, usage examples)

[1] You'll see that classes and properties have been assigned numeric
identifiers in the document, which serves the overall ease of presentation.
They have been clearly marked as under discussion and form part of the
things to be decided on.

All the best,