Issue 287: FRBRoo / CRM for prints?

Starting Date: 
Working Group: 
Closing Date: 

Posted by Regina Stein on 30/7/2015


Did anybody go into depth with FRBRoo / CRM modelling or mapping for prints (visual works) (-> ;-)) , or can point us to respective work on this kind of material? 

Posted by Christian Emil on 30/7/2015

Hi Regine
If I understand AAT correctly, it is a thesaurus and is as such a hierarchy of concepts and can be seen as a incarnation of a hierarchy under the E55 Type.
In a CRM/FRBRoo context a print is a physical object (one of the items of  a series), for instance a lithography, a  paper carrying an image  or more. A lithography would usually  be given the AAT type 'print'  (or belong to this type/be a member of the set of objects that can be said to belong to the hypothetical set of all prints).

This may not be what you have in mind?

Posted by Martin on 30/7/2015

Dear Regine,

There has been an implicit discussion in the CRM about prints as production with particular tools.
E12 Production

Subclass of:         E11 Modification

                              E63 Beginning of Existence

Scope note:         This class comprises activities that are designed to, and succeed in, creating one or more new items.

It specializes the notion of modification into production. The decision as to whether or not an object is regarded as new is context sensitive. Normally, items are considered “new” if there is no obvious overall similarity between them and the consumed items and material used in their production. In other cases, an item is considered “new” because it becomes relevant to documentation by a modification. For example, the scribbling of a name on a potsherd may make it a voting token. The original potsherd may not be worth documenting, in contrast to the inscribed one.

This entity can be collective: the printing of a thousand books, for example, would normally be considered a single event.

An event should also be documented using E81 Transformation if it results in the destruction of one or more objects and the simultaneous production of others using parts or material from the originals. In this case, the new items have separate identities and matter is preserved, but identity is not.


§  the construction of the SS Great Britain

§  the first casting of the Little Mermaid from the harbour of Copenhagen

§  Rembrandt’s creating of the seventh state of his etching “Woman sitting half dressed beside a stove”, 1658, identified by Bartsch Number 197 (E12,E65,E81)
So, the print plate undergoes "transformations" and implies the creation of an information object being present on the plate(s) and the prints, or, in more creative techniques, the information content of the plate is "incorporated" in the prints.

The print plate is "used specific object" in the printing process, but a specialization of E12 may be adequate to fix the
specific kind of use and its consequences of information transfer to the copies.

We are also discussing a generalization of
F32 Carrier Production Event

Subclass of:        E12 Production

Scope note:         This class comprises activities that result in instances of F54 Utilized Information Carrier coming into existence. Both the production of a series of physical objects (printed books, scores, CDs, DVDs, CD-ROMS, etc.) and the creation of a new copy of a file on an electronic carrier are regarded as instances of F32 Carrier Production Event.

Typically, the production of copies of a publication (no matter whether it is a book, a sound recording, a DVD, a cartographic resource, etc.) strives to produce items all as similar as possible to a prototype that displays all the features that all the copies of the publication should also display, which is reflected in property R27 used as source material F24 Publication Expression.

into industrial production, of cars, tools, coins and whatever.

Artistic prints with limited copies etc may not be regarded as producing "things of type XXX".

A CRM extension into the world of artitstic printing may be interesting.

If its only about using AAT vocabulary, Christian-Emil's remark's should be sufficient.

I do not know if the AAT differentiates the plate as museum object from the copy.

Posted by Regine Stein on 31/7/2015

Dear Christian-Emil, dear Martin,

Many thanks for your responses!

Our interest is indeed to understand the whole process of artistic printing (firstly in early modern Europe), how the idea of a visual work evolved in the process, what are the relationships between the various (conceptual and physical) objects involved in the process.

Making a printing plate always starts with a drawing which may either be a copy of an existing visual work, a painting, or may be intentionally designed for a print, either by the printmaker or by another artist. In German language we typically use in documentation the roles "Inventor" and "Stecher", according to the Latin "INVENIT" and "FECIT" which one can often find in inscriptions (see e.g. engravings by Marcantonio Raimondi with inscription "RAPHA URBI INVEN / MAF" - Raphael invented it / Marcantonio fecit = Marcantonio made it). Then, the same visual idea may be realized in several printing plates. Then, we have prints from different states of this same printing plate, and they are sometimes considered as a new visual work, sometimes just as modification. In order to limit the number of copies a printing plate may be scratched.  If we are lucky the printing plate still exists somewhere but obviously in its last state, and earlier states are only known through the prints. Then prints may be compiled into series and so on.

As we are dealing with multiples we wonder if FRBRoo is appropriate to approach this, our questions include:
- Should we consider multiple realizations of the same drawing in various printing plates as multiple F2 Expression (F24 Publication Expression) of the same F1 Work (F14 Individual Work)? Or are they all different works?
- Should we consider different states of one printing plate as F3 Manifestation Product Type?
- How to reflect the different states of the printing plate as "used specific object" in E12 Production?
- Analysis of what is typically recorded in the documentation in a museum holding one (or multiple) print(s): which information pertains to the Work / Expression (e.g. the subject /  P62 depicts), which to the printing plate / F3 Manifestation Product Type (e.g. the state), which to the actual museum object / F4 Item?

We are grateful for further comments on this, and will certainly be back to the group as we move on.

Btw, the AAT of course differentiates
"prints (visual works)" ->
- or more specifically "engravings (prints)" ->
"printing plates" ->
as well as the process of "engraving (printing process)" ->

Posted by Martin on 3/8/2015

Dear Regine,

I think this is a nice application of FRBRoo to a more museum-like context. The Work discussion is an old one. Both positions are justified. Therefore FRBRoo defines an "Individual Work" in contrast to a "Complex Work". In practice, i'd suggest instantiating the "Individual Work" can be ommitted, once it can be inferred, as we had proposed to Europeana.

I'd argue, the distinction "they are sometimes considered as a new visual work, sometimes just as modification" is pointless, because there is a new artistic contribution. The expression is new. The Work gets another realization. The plate is
physically modified. The feature on it, the scratches, are new. An Expression cannot be modified.

The "invenit" can be associated with the work conception, but is an expression in its own right. The "FECIT" has its own artistic contribution, I'd argue it is a derivative.

But there are better FRBR experts on this list!

I think it will be good to create a guideline for this case. I believe it should be discussed together with the bronze casting

Posted by Jim Salmons on 7/8/2015

Regine, C.E.S., and Martin,

In the context of FRBRoo as a DSL/extension of #cidocCRM with particular reference to the “serialization” character of print production, it may be useful to consider the’s contribution of PRESSoo, an extension/harmonization that brings serialization and continuation semantics to FRBRoo.

The reference document is found here:


and Patrick Le Boeuf’s presentation on behalf of the ISSN working group which developed this valuable contribution is here:


Posted by Martin on 9/8/2015

Hi Jim,

A serialization character of prints is not so obvious to me. The reworking of print plates appears to me to be a rather peculiar process of modification of a material object, not so much an editorial series. The actual printing process is either a mechanical one, or a combination of manual creative work and mechanical work, which I'd see more similar to book printing on one side, and mold-based techniques, such as in ceramics or bronze? 

Posted by Patrick  on 10/8/2015

Hi everyone,
I completely agree with Martin. There is no point in dealing with art prints as continuing resources. PRESSoo is inadequate in this context. I think there is enough stuff in a combination of FRBRoo and CIDOC CRM to cope with at least the main issues raised by art prints. Surely an extension of FRBRoo could bring more refinements in the treatment of such products, but FRBRoo has to remain a high-level conceptual model.

Posted by Simon Spero on 10/8/2015

To add a bit more confusion to the discussion,  one can consider Blake's hand printed books.
Every plate was coloured individually, and differently, and are some scholars consider each copy to be a distinct [thing more abstract than Manifestation].

There is some justification for this,  as there is distinct intellectual content (produced by the original author, even). This seems to fall into the awkward place in FRBR where expression and work overlap.

This happens a lot in different places in FRBR;  it's difficult to handle  without using a second order logic,  or at least a first order logic which allows for quantifying over known predicates. 

Posted by Jim Salmons on 10/8/2015

Regine, Martin, and Patrick,

I agree with Patrick and Martin that PRESSoo is not _directly_ applicable to the challenge of modeling prints. I can only speak here as a software designer/developer and NOT as a humanities specialist with knowledge of the subtleties of historic/artistic print production, etc.. But given my decades of advanced object-oriented software design experience, I strongly believe in studying closely-aligned domain models for inspiration and insight when facing your own new modeling challenge.

In this sense, the most valuable lesson that print-modeling folks might gather from the SIG members who developed PRESSoo is to look at the "deep dive" into a domain-specific language (DSL) extension of #cidocCRM's 'E29 Design or Procedure' (e.g. pages 17-22 being an interesting overview,

In both the case of a serial publication and a complex print-production, the concept of some "process over time" is central with a notion of the creator's intentional plan to produce the work... the 'Z12 Issuing Rule' of PRESSoo, the idea of 'Z1 Serial Transformation' and with this the decomposition of procedure to make Properties explicit such as 'absorbed', 'enhanced', 'merged', 'separated', etc. Given a "blind concept-matching test," many of these model elements could be mistaken for their applicability to print production.

It is not that this _specific_ complete DSL (PRESSoo) is an "off the shelf" solution to modeling prints, but rather that the territory covered by the PRESSoo SIG in its ruminations to produce their domain-specific model have much to contribute as a 'modeling resource' for Regine's team to consider as they decide on their own requirements and #cidocCRM-compliant extension. The definition document is part definition and part ‘project post mortem’ of insights about their modeling decisions.

If nothing else, a good read of the PRESSoo definition document could help inform the modeling decisions to be made by Regine's team. I found the PRESSoo document full of helpful insightful discussion and rich with 'use case' diagrams that reflect the broad thinking of the PRESSoo SIG's work. I think it's a good example of #cidocCRM DSL extension documentation regardless of your domain of interest.

Pardon me if these comments are a re-hash of well-worn conversation over domain-specific issues resolved long ago. My intent in commenting has more to do with sharing insights about modeling best practice – “Borrow, don’t re-invent…” for example – rather than being specific to a particular model.

Posted by Martin on 12/8/2015

Dear Simon,

Interesting thought. I'd say however, I am not confused by this. In my opinion, the problem lays in the human mind, which tries to stretch abstractions over incompatible evidence.

From a CRM point of view, every copy of such a process is an individual physical object. Why should we call it a Manifestation, or even something more abstract, if it does not fit to the book market process this concept was designed for?

If there is an individual artistic contribution each time (as I mentioned in my last message), then we talk about new features on the plate each time, which correspond to a new symbolic content ("Expression"). There are two cases: Either it is an overall derivative, or there is an "incorporation" of content by virtue of the fixed features on the plate. So, I'd say the individually colored copy "incorporates" the plate optical content and adds a new creation process to it.

The question of second order logic arises when we want to talk about types of types of things which inherit fixed properties from their class, such as the number of pages of a book.
If regarding such copies is better modelled as instances of a type or not is for me the primary question, not the kind of logic to be used. The fixed features on the individually colored plate would hardly be expanded into discrete properties, besides itself being on the plate.

Also, to my understanding, the second order logic features needed to describe inheritance of properties from types of types or classes is completely decidable and no technical problem except for the current reluctance of IT to think in such terms .
Interesting are the requirements which kind of logic cultural-historical data need, rather than, which logic to avoid. From CRM-SIG point of view, we require since a long time more powerful forms of logic than standard DL.

In  34nd joined meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9 and the 27th FRBR - CIDOC CRM, it is decided a guideline (FAQ)t o be produced for modelling visual works with FRBRoo. It is assigned to Pat.

Heraklion, October 2015

posted by Pat on 31/7/2016

Hello everyone,

Here is my draft of a FAQ for issue 287. The text has benefitted from input from Marie-Chantal L'Ecuyer-Coelho, chair of the ARLIS/NA Cataloguing Committee.

posted by Pierre Choffee on 31/7/2016

Hi Pat, hello everyone,

Starting from the initial sketch or drawing, and ending with the final prints, what would you think of this FRBRoo process :

(1) Drawing
The initial sketch or drawing may be a piece of art in itself, kept in a museum, we could describe its creation with the usual F28 Expression Creation activity, creating an F22 Self-Contained Expression which realises an F14 Individual Work. The same F28 Expression Creation r18 created an F4 Manifestation Singleton which is the object (the drawing).

(2) Production of the plate
This drawing will be used to produce a plate. To make it short, we could use an E12 Production activity which would have p16 used a specific object F4 Manifestation Singleton (the drawing, F4 is a subclass of E24 Physical Man-Made Thing) and p31 produced an E22 Man-Made Object, the plate.
If we consider this plate (and its possible modifications during the process) as expressions of a work, we can model the creation process with F28-F22-F14 creating an F4 (a plate), and F28 p16 used a specific object F4 (the drawing).
Should the plate be modified, the process is repeated with F28-F22-F14 creating an F4 (a new plate), and F28 p16 used a specific object F4 (the precedent plate).
F28 is a subclass of E65 Creation, but also of E12 Production which itself is a subclass of E11 Modification.
The various F22 Self-Contained Expression resulting from these multiple production/modification activities are various forms that the same work takes each time it is realised, so they can be grouped inside an F15 Complex Work which they "realise" : F15 r3 is realised in F22.

(3) Test prints
The first print(s) are generally not part of a publication process, they are proofs.They can be modelled via an E12 Production activity which would have p16 used a specific object F4 Manifestation Singleton (the plate) and p31 produced one or more E22 Man-Made Object, the test print(s).
If we want to use F28 Expression Creation here, since F4s are supposed to be unique, we need to use as many F38, producing as many F22 as there are F4s, which makes sense as the "épreuves d'artiste" have a unique number.

(4) Publication
The 'industrial' process of publication is the usual one in FRBRoo, F30 Publication Event creating an F24 Publication Expression which realises an F19 Publication Work. The F32 Carrier Production Event p16 used specific object F4 Manifestation Singleton (the plate) and p31 produced an F5 Item.
The F24 Publication Expression r6i is carried by the F5 Item and r27i was used as source material by the F32 Carrier Production Event.

There may be complex printing processes (e.g. various prints for the different colours of the same print), but this would be a variant of the above.

posted by Pat on 2/8/2016

Hello Pierre,

Just to say that I agree with you detailed analysis of the stages.

Depending on need, some of the stages would not be recorded, and in some techniques some of the steps might not be used (I believe that in some techniques there might not be a preliminary drawing, rather the print matrix may be created directly, and then the rest of the process applies).


Current Proposal: 

In the 37th joined meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9 and the 30th   FRBR - CIDOC CRM Harmonization meeting, the crm-sig reviewed Pat's proposal and decided  that Modification would not be strong enough, since in modification the physical thing’s identity doesn’t change, there is nothing new. Sig proposes to Pat to make use of transformation (with implied destruction) instead of modification.  If Pat accept and the notion of transformation is correctly interpreted, then the FAQ answer can be accepted, otherwise the issue should be discussed.

Also a related issue has been raised about  a specific type of ‘used’ relation in a production that would allow tracking of transfer of information.
Berlin, December 2016



sent by Pat    on 13/1/2017

I agree, E81 Transformation is certainly better than E11 Modification because the matrix has a distinct identity both before and after the change, and there can be considerable physical change to the plate (or it can just be a few lines), but E81 is more general. So I have updated the FAQ to use E81 and P124 (which works on E77 a superclass of E24), the new text is attached.

Reference to Issues:

Meetings discussed: