Issue 600: Class/Property labels are not definitions

ID: 
600
Starting Date: 
2022-06-02
Working Group: 
3
Status: 
Done
Background: 

In the 53rd CIDOC CRM & 46th FRBRoo SIG meeting, it was decided to start a new issue re the drafting of a clause stating that the labels of classes and properties of the CRMbase and family models do not stand in lieu of their definition. They only form mnemonic devices that allow one to refer to any given class/property. 

This is the issue to discuss drafting the clause. 

May 2022

Current Proposal: 

Post by Thanasis Velios (14 July 2022)

Dear all,

As part of the homework for issue 600 on the role of labels of classes and properties, Erin and I drafted the following text to be added at the end of the "Naming Conventions" section.

"The nominal groups used to identify classes and properties are also referred to as ‘labels’. While the goal of a label may be to indicate the meaning of the class or property in a few words, it would be impossible to distil the nuance of the represented concept into a sentence fragment; relying solely on the label to infer the meaning of the class or property introduces significant potential for misunderstanding. Therefore, these labels should instead function as reminders of the corresponding scope notes. As the labels cannot encapsulate the intension of classes and properties, no assumption about the meaning of a class or property should be made from its label alone."

Happy to hear your comments and prepare a WD for Rome.

All the best,

Thanasis

Post by Martin Doerr (14 July 2022)

Dear Thanasi,

A suggest minor modifications, more explicit:

"The nominal groups used to identify classes and properties are also referred to as 'labels'. While the goal of a label is to approximate the meaning of the class or property in a few words, it is in general impossible to distil the nuance of the represented concept into a sentence fragment with the limited terms of one particular natural language; relying (del:solely) on the label to infer the meaning of the class or property introduces significant potential for misunderstanding. Therefore, these labels function instead as reminders of the corresponding scope notes only. As the labels cannot encapsulate the intension of classes and properties, no assumption about the meaning of a class or property should be made from its label alone."

Opinions?

Best,

Martin

Outcome: 

In the 54th CIDOC CRM & 47th FRBR/LRMoo SIG meeting, the SIG discussed the clause provided by Erin Canning, Thanasis Velios and Martin Doerr texpanding on the idea that class/property labels are not definitions and accepted it with minor modifications. The text will appear at the end of "Naming Conventions" in the introduction to CIDOC CRM (right before the section "Inheritance and Transitivity"). 

The reformulated text can be found below:

Naming Conventions

  • […]
  • The names used to identify classes and properties are also referred to as ‘labels’. While the goal of a label is to approximate the meaning of the class or property in a few words, it is in general impossible to distil the nuance of the represented concept into a sentence fragment with the limited terms of one particular natural language. Relying on the label to infer the meaning of the class or property introduces significant potential for misunderstanding. Therefore, these labels should instead function as reminders of the corresponding scope notes only. As the labels cannot encapsulate the intension of classes and properties, no assumption about the meaning of a class or property should be made from its label alone

Inheritance and Transitivity

 Issue Closed