Frames on paintings

14 views
Skip to first unread message

Rupert Shepherd

unread,
Jul 21, 2025, 6:39:15 AMJul 21
to Linked Art
Dear colleagues

Richard Light and I are discussing how best to model picture frames in relation to the paintings they frame; let's begin by starting from the painting record.

Some frames are physically part of the painting: carved from the same piece of wood, or nailed / glued / screwed in place. These seem easy enough, I think: treat as Parts, with classififedAs identifying the type of frame (integral / engaged etc.)

Others, however, are independent objects, usually made by different people, often made at different times,  that may be attached to or removed from one or more paintings as need, taste, etc. dictate. Semantically, I don't think we can consider them to be parts of one particular object, no matter how close the relationship with a specific painting may be. It looks to me as though we're talking about Related Objects as modelled by a Related Entity Attribute Assignment; does that sound right?

We would also want to indicate whether a frame is currently attached to a painting, or has been in the past - I'm assuming we could add AAT 'current (information status)' (https://vocab.getty.edu/aat/300456596) or 'former (information indicator)' (https://vocab.getty.edu/aat/300404776) to our attribute assignment?

All advice gratefully received before we progress much further with this!

Thanks, as always,

Rupert

Rupert Shepherd, PhD FSA (he/him)

Collection Information Manager

E rupert....@nationalgallery.org.uk

T +44 (0)20 7747 5921

 

Mostly working remotely; calls to the number above should still reach me.

 

José María Velasco: A View of Mexico

Robert Sanderson

unread,
Jul 21, 2025, 9:37:55 AMJul 21
to Rupert Shepherd, Linked Art

Hi Rupert, all,

We don't make a distinction about the separability of parts in the model. A chess piece is part_of the chess set, even though it's not attached to the board or other pieces in anyway. The arm (if there's a need to identify it separately) is part_of the statue, even though removing it would be entirely destructive. The drawer is part_of the desk, and could likely be removed but perhaps not without undoing some screws. So the frame can just be part_of the painting, regardless of whether it's integral or not, who it's made by, or any other feature. 

If it's just a frame with no artwork, then agreed it's not part of anything. It's just a stand alone object. If it is currently part of an artwork (no judgement as to the value of the image vs the frame, or how it's used in the artwork) then personally I would just use part_of. You could go the AttributeAssignment route but it will open up the can of worms of when to use the AA route and when to use part_of.

For previously part of, you could use the `removed_by` property on the frame with a PartRemoval that then `diminishes` the painting it was part of: 

That would allow the PartRemoval to handle who, when, where, why, and how it was removed.

In CRM there's also `augmented_by` a PartAddition that `added` a part, but we haven't had a use case for it with actual data to support its inclusion in the model.
With it you could say that:

{
  "id": "painting-uri",
  "type": "HumanMadeObject",
  "_label": "Full Painting",
  "augmented_by": [
    {
      "type": "PartAddition",
      "added": {"id": "frame-uri", "type": "HumanMadeObject", "_label": "Frame"
    }
  ]
}

Hope that helps!

Rob

--
You received this message because you are subscribed to the Google Groups "Linked Art" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linked-art+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/linked-art/CWLP265MB08346FCCCF1E90A890A08123C75DA%40CWLP265MB0834.GBRP265.PROD.OUTLOOK.COM.


--
Rob Sanderson
Senior Director for Digital Cultural Heritage
Yale University

Athanasios Velios

unread,
Jul 21, 2025, 10:00:15 AMJul 21
to Linked Art
Another way I think about similar situations is that we really have three things: 1) the painting (without any frame), 2) the frame and 3) the painting with the frame. (1) and (2) are parts of (3). (3) seizes to exist if the painting is removed from the frame and another entity is created when a different frame is attached. The painting (1) remains as is.

All the best,

Thanasis

Rupert Shepherd

unread,
Jul 21, 2025, 11:26:28 AMJul 21
to Athanasios Velios, Linked Art
Thanasis, Rob, thanks! Yes, all our frames are also recorded as objects in their own right; so this would be a reciprocal relationship: frame is part of painting, and conversely painting is part of frame. Happy to avoid AttributeAssignments if you feel it doesn't cause semantic problems.

Have I just given you a use case for augmented_by? We do reframe quiet frequently, and we record the date a new frame was added to  a painting ...

Best wishes

Rupert


From: linke...@googlegroups.com <linke...@googlegroups.com> on behalf of Athanasios Velios <a.ve...@gmail.com>
Sent: 21 July 2025 15:00
To: Linked Art <linke...@googlegroups.com>
Subject: [EXTERNAL] Re: [linked-art] Frames on paintings
 
This Message originated outside your organization.
Another way I think about similar situations is that we really have three things: 1) the painting (without any frame), 2) the frame and 3) the painting with the frame. (1) and (2) are parts of (3). (3) seizes to exist if the painting is removed from the frame and another entity is created when a different frame is attached. The painting (1) remains as is.

All the best,

Thanasis

On Monday, 21 July 2025 at 14:37:55 UTC+1 azar...@gmail.com wrote:

Hi Rupert, all,

We don't make a distinction about the separability of parts in the model. A chess piece is part_of the chess set, even though it's not attached to the board or other pieces in anyway. The arm (if there's a need to identify it separately) is part_of the statue, even though removing it would be entirely destructive. The drawer is part_of the desk, and could likely be removed but perhaps not without undoing some screws. So the frame can just be part_of the painting, regardless of whether it's integral or not, who it's made by, or any other feature. 

If it's just a frame with no artwork, then agreed it's not part of anything. It's just a stand alone object. If it is currently part of an artwork (no judgement as to the value of the image vs the frame, or how it's used in the artwork) then personally I would just use part_of. You could go the AttributeAssignment route but it will open up the can of worms of when to use the AA route and when to use part_of.

For previously part of, you could use the `removed_by` property on the frame with a PartRemoval that then `diminishes` the painting it was part of: 

That would allow the PartRemoval to handle who, when, where, why, and how it was removed.

In CRM there's also `augmented_by` a PartAddition that `added` a part, but we haven't had a use case for it with actual data to support its inclusion in the model.
With it you could say that:

{
  "id": "painting-uri",
  "type": "HumanMadeObject",
  "_label": "Full Painting",
  "augmented_by": [
    {
      "type": "PartAddition",
      "added": {"id": "frame-uri", "type": "HumanMadeObject", "_label": "Frame"
    }
  ]
}

Hope that helps!

Rob

On Mon, Jul 21, 2025 at 6:39 AM 'Rupert Shepherd' via Linked Art <https://url.uk.m.mimecastprotect.com/s/TfA4ClRgLFXn8GZUyi3izi1dt?domain=googlegroups.com> wrote:
Dear colleagues

Richard Light and I are discussing how best to model picture frames in relation to the paintings they frame; let's begin by starting from the painting record.

Some frames are physically part of the painting: carved from the same piece of wood, or nailed / glued / screwed in place. These seem easy enough, I think: treat as Parts, with classififedAs identifying the type of frame (integral / engaged etc.)

Others, however, are independent objects, usually made by different people, often made at different times,  that may be attached to or removed from one or more paintings as need, taste, etc. dictate. Semantically, I don't think we can consider them to be parts of one particular object, no matter how close the relationship with a specific painting may be. It looks to me as though we're talking about Related Objects as modelled by a Related Entity Attribute Assignment; does that sound right?

We would also want to indicate whether a frame is currently attached to a painting, or has been in the past - I'm assuming we could add AAT 'current (information status)' (https://vocab.getty.edu/aat/300456596) or 'former (information indicator)' (https://vocab.getty.edu/aat/300404776) to our attribute assignment?

All advice gratefully received before we progress much further with this!

Thanks, as always,

Rupert

Rupert Shepherd, PhD FSA (he/him)

Collection Information Manager

T +44 (0)20 7747 5921

 

Mostly working remotely; calls to the number above should still reach me.

 

José María Velasco: A View of Mexico
--
You received this message because you are subscribed to the Google Groups "Linked Art" group.
To unsubscribe from this group and stop receiving emails from it, send an email to https://url.uk.m.mimecastprotect.com/s/SXxQCqQn5sX03B5UvFqiEfS21?domain=googlegroups.com.


--
Rob Sanderson
Senior Director for Digital Cultural Heritage
Yale University

--
You received this message because you are subscribed to the Google Groups "Linked Art" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linked-art+...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages