Issue 345: properties having domain or range deprecated classes

Starting Date: 
Working Group: 

Posted by Robert Sanderson on 12/5/2017

Dear all,

In the (latest?) version of 6.2.2, E46 Section Definition is deprecated in favor of Appellations (and, one imagines, Features). However P58 still references it as a domain.  I expect that P58 should also be deprecated?

Posted by Martin 13/5/2017

Dear Robert,

Yes, sure, thank you!

Posted by Phil Carlisle on 21/6/2017

Hi all,

I’m currently creating new resource models for the Arches project and looking at Actors.

Since E82 has been deprecated in favour of E41 then surely P131 should also be deprecated along with the text in the introduction which refers to these in CRM Compatibility of Information Systems paragraph 3 – “Note that there is no minimum requirement for the classes and properties that must be present in the exported user data. Therefore it is possible that the data may pertain to instances of just a single property, such as E21 Person. P131 is identified by: E82 Actor Appellation.”.

Plus the references in the Property hierarchy and the entries for P1 and E39

Posted by Robert on 21/6/2017

Hi Philip,

Indeed,  _all_ of the subclasses of E41 other than E42 should probably be deprecated.  They add nothing new, semantically, compared to Appellation.
Also the subclasses of E13 … (E14-E17) are overly specific, and insufficient. There’s no Assignment for Appellations, for example, yet naming things occurs all the time.

Even if they’re not officially deprecated, organizations actually producing data should simply not use them.

For interest’s sake, our analysis of which classes in the CRM are useful:

Posted by Franco on 21/6/2017

Dear Robert

I had a look at the page. Interesting, at first sight I agree with most of the statements there, but it will require more attention to comment. Anyway, there’s one thing I can say now.

I think that the analysis of E34 Inscription is a bit superficial. An inscription, as epigraphists know well, has several dimensions and the proposed use of E33 captures only the text one. Instead, it is important also to consider the graphic component and the material one.

Example 1: the Phaistos disk has an as yet undeciphered linguistic content - we could imagine that the creators made bizarre figures just to puzzle future archaeologists: it actually could not be a linguistic object, and I would be scared to define it only this way.

Example 2: in the Sherlock Holmes novel “The adventure of the Dancing Men” in the collection “The Return of Sherlock Holmes" the paper messages are inscriptions on paper. They become linguistic objects only when deciphered by the Consulting Detective - if you don’t know the story you may wish to read it.

My colleague Achille Felicetti has developed a CRM extension for epigraphy which aims to manage all these aspects and he may wish to explain better than me how they propose to do it, or refer to a paper where he has detailed the whole model.

Posted by Robert on 21/6/2017

Thank you, Franco.

I completely agree with your assessment here.  The concern with the current model, of course, is that E34 Inscription is a subclass of E33 Linguistic Object (along with E37 Mark) meaning that every Inscription is also a Linguistic Object.  For your examples, before the content can be ascertained to actually have linguistic content, I would imagine them to be only E36 Visual Items.  I’ll update the analysis with this clarification.

And also, for clarity about the scope, this work does not take into account any of the extensions to CRM, only the core model as expressed in RDF.

Thank you again! 

Posted by Achille on 22/6/2017

Dear Robert,

A quick note on the E34 Inscription class, following Franco’s considerations (and thanks to him for pulling me in :-))

Some time ago, in a paper illustrating a possible CIDOC CRM application to the epigraphic world, we started investigating the degree of applicability of E34 to the semantic description of an epigraph. Very soon we discovered that:

1. the scope notes of the E34 class state that it “comprises recognisable, short texts attached to instances of E24 Physical Man-Made Thing”; from an epigraphic perspective, although many inscriptions bear short texts, the brevity or length of an inscription is not among its main characteristics. In fact, there are inscriptions occupying entire walls (the Gortyn Law Code or the Res Gestae Divi Augusti, for example) and in any case the “short text” of the E34 class remained too vague and undefined for the purposes of our investigation.

2. as it is currently defined within the CIDOC CRM, E34 class belongs to the “conceptual objects” entities, which in turn are defined as “non-material products of our minds and other human produced data” something that does not takes into any account the “materiality” of an epigraph, a fundamental and essential component of its identity.

During the investigation and study of an epigraph, scholars typically move from the analysis of its physical features before getting to their archaeological, palaeographic, linguistic and historical meaning. In particular, an inscription defined only as a conceptual object, does not seem to fully capture the very nature of an epigraph, consisting of a physical part as well as of the conceptual component the E34 class points to. The etymology of the word “epigraph” itself (i.e.: “written on something”) seems to suggest such inextricable relation. In all these aspects, an inscription presents an immediate resemblance to the E25 Man-Made Feature class of the core model. The conceptual object is also present, but it is something referred by the inscription, not the inscription itself: this relation(s) could be further detailed by using specific properties. In our paper we have tried to outline this issue by also proposing the modification of the E34 class in the core model itself.

You could read the rest of the story in the paper, if you like, just following this link:

This paper laid the groundwork for the development of the extension for ancient texts we are currently working on.
Indeed, we have further expanded the idea of a text being something halfway between the physical and the conceptual world and created new classes that could be used to describe a huge variety of written texts, not only epigraphic inscriptions. But this is another story. Which, by the way, you can read here :-)

I hope this helps.

Current Proposal: 

In 39th crm-sig meeting, the crm-sig discussed about what to do  with the properties that their domain or range are deprecated classes. The decisions are

- Martin, will make a proposal to delete the P58, to be decided by email vote,
- Homework assigned to Martin and Christian Emil to go over all deprecated classes and Martin will formulate a list with properties that have problem and CEO will check
- Martin will coordinate this issue.
Heraklion, October 2017

Posted by Christian Emil on 4/1/2018


The following classes and properties are depricated: E46, E50, E75, E82, P177, P178 in 6.2.2

Only 2 properties have a depricated range:

P58 has section definition (defines section)
P149 is identified by (identifies)

P58:  E18 Physical Thing to E41 Appellation

P149 will simply become equal to  P1 Identified by (identifies) restricted to E28 Conceptual Objects and should be deleted/ deprecated.

I also searched the document (6.2.2) for references to the depricated classes and propierties. In the attached document I inserted comments, text to be inserted is marked with yellow and to be deleted with red.  I also checked the use if 'instance of' and have marked inserted missing 'instances of' in yellow. I could have used track changes, and may do this. But please, check the documetn first.

Meetings discussed: