Texts and Documents

Table of Contents


Introduction

This section documents the model for documents that contain text, including artworks such as medieval manuscripts, archival material such as letters, ledgers or diaries, scholarly communication such as journals, articles and monographs, digital objects such as web pages, or any other sort of written communication.

The intent is not to be an all-encompassing model that would be suitable for a graph based library management system, archival finding aids, or catalog of any digital resource, but instead to provide sufficient description that the object is identified, understandable and able to be referenced within other more specific systems and ontologies. This model is intended to be enough to use for the basic use cases of referencing texts that are related to artwork, regardless of whether they are held in a library, archive, museum, or on the internet.

Notably, it does not attempt to reproduce the formalisms of FRBR, as manifested by FRBRoo or BibFrame, or other conceptual hierarchies, but instead provide as simple as possible a model to accomplish core bibliographic reference tasks.

Physical Objects, Conceptual Texts

The first distinction that is needed is between the physical carrier of the text, and the text itself. Like the VisualItem pattern for the artwork's visual content shown by objects, the LinguisticObject that represents the text of a work can be carried by many ManMadeObjects. In this way, all of the copies of a particular book carry the same information content, that only needs to be described once and can act as a single connection point.

Like other LinguisticObjects, it can have a value for the actual text of the work, classifications, languages and so forth.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/object/23", 
  "type": "ManMadeObject", 
  "label": "Physical Book", 
  "classified_as": [{"id": "aat:300028051","type": "Type","label": "Book"}], 
  "carries": [
    {
      "id": "https://linked.art/example/text/8", 
      "type": "LinguisticObject", 
      "label": "Textual content of the Book", 
      "value": "Once upon a time, ...", 
      "language": [{"id": "aat:300388277","type": "Language","label": "English"}]
    }
  ]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Core Features

The same core features of all other resources are also applicable to the LinguisticObjects used to represent texts. They must have an id and type, they should have a label, they should have a classified_as relationship to a Type that further describes the sort of object, and so forth. They may also have a language property, that references a Language resource.

The title of a text is captured using the Name pattern, and assigned identifiers using the Identifier pattern, referenced with the identified_by relationship.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/9", 
  "type": "LinguisticObject", 
  "label": "Example Article", 
  "identified_by": [
    {
      "id": "https://linked.art/example/name/13", 
      "type": "Name", 
      "value": "Title of Article"
    }, 
    {
      "id": "https://linked.art/example/identifier/5", 
      "type": "Identifier", 
      "label": "DOI of Article", 
      "value": "10.1000/1011"
    }
  ], 
  "classified_as": [
    {"id": "aat:300048715","type": "Type","label": "Article"}
  ], 
  "language": [{"id": "aat:300388277","type": "Language","label": "English"}]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Creation and Publication

The production of the physical carriers of texts uses the same model as for other physical objects, and may be of interest for manuscripts, very early printed works (incunabula), letters or other similar documents, however the factory details for a specific modern book are likely of much less importance to capture.

There are two primary text-specific activities that are captured -- the creation of the text, being the conceptual LinguisticObject and its publication. The text is created by a Creation activity of the author, but then there is a publishing Activity (aat:300054686) carried out by the publishing organization, for the same resource.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/10", 
  "type": "LinguisticObject", 
  "label": "Example Journal Article", 
  "classified_as": [
    {"id": "aat:300048715","type": "Type","label": "Article"}
  ], 
  "used_for": [
    {
      "id": "https://linked.art/example/activity/12", 
      "type": "Activity", 
      "classified_as": [
        {"id": "aat:300054686","type": "Type","label": "Publishing"}
      ], 
      "carried_out_by": [
        {
          "id": "https://linked.art/example/group/11", 
          "type": "Group", 
          "label": "Publishing House"
        }
      ]
    }
  ], 
  "created_by": {
    "id": "https://linked.art/example/Creation/5", 
    "type": "Creation", 
    "label": "Creation of the Article Content", 
    "carried_out_by": [
      {
        "id": "https://linked.art/example/person/19", 
        "type": "Person", 
        "label": "Author"
      }
    ]
  }
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Structure

The textual structure can be modeled with the partitioning of LinguisticObjects via the includes property, in the same way that the parts of a physical object can be described using the part property. Thus the content of an Article can be included in the containing Issue, which is included within a Volume, which is within the Journal or other periodical. Similarly Chapters can be included within a Book or Proceedings, particular entries within a catalog, and so forth.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/11", 
  "type": "LinguisticObject", 
  "label": "Example Monograph Text", 
  "classified_as": [
    {"id": "aat:300060417","type": "Type","label": "Monograph"}
  ], 
  "part": [
    {
      "id": "https://linked.art/example/text/12", 
      "type": "LinguisticObject", 
      "label": "Chapter 1", 
      "classified_as": [{"id": "aat:300311699","type": "Type","label": "Chapter"}]
    }, 
    {
      "id": "https://linked.art/example/text/13", 
      "type": "LinguisticObject", 
      "label": "Chapter 2", 
      "classified_as": [{"id": "aat:300311699","type": "Type","label": "Chapter"}]
    }
  ]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Pages

Textual content is typically presented on pages or folios. As there might be many physical copies with the same structure, it is common to describe the pagination of the content as it applies to the content in general, rather than the many physical objects that carry that content. Pagination (aat:300200294) or Foliation (aat:300200662) statements are the most common way to represent this, as simple descriptive fields following the core statement pattern. For uncomplicated, machine readable pagination values, it is possible create a pair of start and end Dimensions.

Pagination Statement

The easiest way to describe the page range of a particular section of a larger collection is with a pagination statement. The disadvantage of the approach is that it is not computationally available for processing.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/14", 
  "type": "LinguisticObject", 
  "label": "Example 10 page Article", 
  "classified_as": [
    {"id": "aat:300048715","type": "Type","label": "Article"}
  ], 
  "referred_to_by": [
    {
      "id": "https://linked.art/example/text/15", 
      "type": "LinguisticObject", 
      "value": "125-135", 
      "classified_as": [
        {"id": "aat:300200294","type": "Type","label": "Pagination Statement"}, 
        {"id": "aat:300418049","type": "Type","label": "Brief Text"}
      ]
    }
  ]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Page Dimensions

Alternatively, if the pagination is a single range with known start and end page numbers, it can be expressed as two Dimension objects. Dimensions are the only way to express non time-based values in CRM, so while this may feel like an unnatural fit, the alternative is custom classes and relationships. It must be noted that neither FRBRoo nor BibFrame have the concept of pagination in a machine readable context. If the text is non-contiguous (such as an article with a supplemental at the end of the issue), then there should be LinguisticObjects created for both sections, each of which has their own pair of page numbers.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/16", 
  "type": "LinguisticObject", 
  "label": "Example 10 page Article", 
  "classified_as": [
    {"id": "aat:300048715","type": "Type","label": "Article"}
  ], 
  "dimension": [
    {
      "id": "https://linked.art/example/value/4", 
      "type": "Dimension", 
      "value": 125, 
      "classified_as": [{"id": "aat:300404050","type": "Type","label": "First"}], 
      "unit": {
        "id": "https://linked.art/example/MeasurementUnit/0", 
        "type": "MeasurementUnit", 
        "label": "Page Numbers"
      }
    }, 
    {
      "id": "https://linked.art/example/value/5", 
      "type": "Dimension", 
      "value": 135, 
      "classified_as": [
        {
          "id": "https://linked.art/example/Type/1", 
          "type": "Type", 
          "label": "Last -- waiting for AAT id"
        }
      ], 
      "unit": {
        "id": "https://linked.art/example/MeasurementUnit/0", 
        "type": "MeasurementUnit", 
        "label": "Page Numbers"
      }
    }
  ]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Vocabulary Needed

As this is not something that has typically been part of CRM, nor other art related ontologies, the Dimension class is not ideal for the mapping and there aren't obvious classifications for the start or end Dimensions, nor the unit of page numbers. This part of the model should be considered at risk for change or removal. An alternative model would be to allow parts to be ordered, and then model the first page, plus a number of parts that would follow that. The first page would then have an Identifier of the page number, from which the end page could be calculated.

Abstracts

While the full text of the LinguisticObject may be either too long to provide, or protected by publication and access restrictions, it is common for an abstract to be published and circulated openly. This abstract might be provided by some other actor than the author of the content, such as via an abstracting and indexing service. In this case, the Abstract would have its own separate Creation activity from that of the abstracted content.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/17", 
  "type": "LinguisticObject", 
  "label": "Abstract of Example Article", 
  "value": "This is the abstract of the example article, which is excellent.", 
  "classified_as": [
    {"id": "aat:300026032","type": "Type","label": "Abstract"}, 
    {"id": "aat:300418049","type": "Type","label": "Brief Text"}
  ], 
  "refers_to": [
    {
      "id": "https://linked.art/example/text/18", 
      "type": "LinguisticObject", 
      "label": "Example Article", 
      "classified_as": [{"id": "aat:300048715","type": "Type","label": "Article"}]
    }
  ], 
  "created_by": {
    "id": "https://linked.art/example/Creation/6", 
    "type": "Creation", 
    "carried_out_by": [
      {
        "id": "https://linked.art/example/person/20", 
        "type": "Person", 
        "label": "Example Abstracter"
      }
    ]
  }
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

References to other Resources

Textual content can refer_to other resources in the overall model. These references might be just that the content discusses the resource, such as a mention of a particular painting. Equally, the reference might be that the content was used as the evidence for the resource, such as a Materials Statement being the evidence for the Materials in the model that the painting is asserted to be made_of.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/19", 
  "type": "LinguisticObject", 
  "label": "Article about Example Painting", 
  "classified_as": [
    {"id": "aat:300048715","type": "Type","label": "Article"}
  ], 
  "refers_to": [
    {
      "id": "https://linked.art/example/object/24", 
      "type": "ManMadeObject", 
      "label": "Example Painting", 
      "classified_as": [
        {"id": "aat:300033618","type": "Type","label": "Painting"}, 
        {"id": "aat:300133025","type": "Type","label": "Artwork"}
      ]
    }
  ]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)

Translations

One Linguistic Object, be it captured in the model as a Statement or as the full text content carried by a physical object, can be the translation of another. For example an article can be the translation of an original article in a different language. This is represented using the translation_of relationship.

{
  "@context": "https://linked.art/ns/v1/linked-art.json", 
  "id": "https://linked.art/example/text/20", 
  "type": "LinguisticObject", 
  "label": "The Birth of Artistic Conception in China", 
  "classified_as": [
    {"id": "aat:300048715","type": "Type","label": "Article"}
  ], 
  "language": [{"id": "aat:300388277","type": "Language","label": "English"}], 
  "translation_of": [
    {
      "id": "https://linked.art/example/text/21", 
      "type": "LinguisticObject", 
      "label": "Zhongguo yishu yijing zhi dansheng", 
      "classified_as": [
        {"id": "aat:300048715","type": "Type","label": "Article"}
      ], 
      "language": [{"id": "aat:300388113","type": "Language","label": "Chinese"}]
    }
  ]
}

JSON-LD (Raw) | JSON-LD (Playground) | Turtle (Raw) | Turtle (Styled)