Skip to content

Relationship Module

Jump to bottom
LuotongCheng edited this page Sep 26, 2024 · 7 revisions

Relationship Module

Overview

This module is designed to model social relationships among characters, utilizing concepts from DOLCE Lite Plus (DLP). A social relationship is defined as an interaction between agents, with roles assigned within these interactions (DOLCE Lite Plus--Social Relationship).

Relationship Characteristics

The module incorporates the following characteristics of social relationships (Mika & Gangemi, 2016):

  • History: The events (durative or temporal) that contribute to the development and changes in the relationship over time.
  • Role: The various roles that entities can assume within the relationship.

Relationship Evolution

Relationships are dynamic rather than fixed (Chaturvedi et al., 2015). This module is also designed to capture the dynamic and evolving nature of relationships in a narrative.

Scope

The module provides a framework for modeling social relationships in ontologies, with a focus on relationship roles, and relationship events, as well as the dynamic nature of relationship. However, the module does not include other aspects like sign (valence), strength, and frequency.

Ontology (diagram)

Class

  • dlp:social-relationship: Represents the social relationship.
  • gc:G4_Relationship_Role: Represents the roles required by the social relationship.
  • gc:G5_Narrative_Event: Represents events or statives within the relationship history.
  • gc:G1_Character: Represents the characters involved in the relationship.

Property

  • dlp:involves: Identifies which characters are involved in the relationship.
  • d-uses: Indicates that the relationship uses or requires other elements, such as roles and relationship history.
  • dlp:generically-dependent-on: Indicates that a relationship dependent on G5_Narrative_Event.
  • dlp:plays: Indicates the role that a character assumes within the relationship, specifying the nature of the character's involvement (e.g., friend, lover).
  • dlp:participant-in: Indicates a character’s involvement in events that contribute to the development or evolution of the relationship.

Example: Demonstrating the Ontology with "Harry Potter and the Deathly Hallows" (diagram)

Relationship Events (chronological order)

  • Initial friendship: At the Burrow, Harry, Ron, and Hermione exhibit mutual support and shared resolve as they prepare to fight Voldemort. This represents their friendship with the role of friend.

  • Evolving relationship:

    1. Rescue attempt: Ron attempts to rescue Hermione while Bellatrix Lestrange tortures her with the Cruciatus Curse. Ron’s emotional breakdown and rescue attempt illustrate both friendship and the beginning of romantic love.
    2. Kiss in safe house: After escaping from Malfoy Manor, Ron and Hermione share a kiss in their safe house, marking the evolution to romantic love.
    3. Kiss in the chamber of secrets: The kiss in the Chamber of Secrets solidifies their romantic love.
    4. The battle of Hogwarts: Showcase the friendship among the three characters. The kiss in the Chamber of Secrets falls within the context of the Battle of Hogwarts, demonstrating the dynamic nature of their relationship.

In this example, we model two types of relationships and their evolution:

  1. Romantic love evolving from friendship: We represent the romantic love between Ron and Hermione as an evolution from their initial friendship.

  2. Roles in different relationships:

    • Friendship: In their friendship with Harry, Ron and Hermione both assume the role of friend.
    • Romantic love: In their romantic relationship, Ron and Hermione take on the roles of lover and beloved.

For details on how events relate to each other, see the event module.

Advantages

This ontology addresses several complexities in modeling relationships:

  • Handling complex relationships: The ontology addresses complex relationships involving more than two characters, such as circular or triangular relationships. It divides relationships into two categories:

    • Homogeneous relationships: Relationships where all characters play the same role (e.g., all friends). For these, we use dlp:involves to link characters to the relationship, with all characters having the same role (e.g., friend).
    • Heterogeneous relationships: Relationships where characters play different roles (e.g., a love triangle). These are divided into pairs, allowing different relationships and roles to be represented. For example:
      • Relationship A involves characters A and B, where A and B are lovers.
      • Relationship B involves characters B and C, where B and C are rivals.

    This approach allows modeling nuanced aspects of relationships and distinguishing different roles in the same or different relationships. By using the role property, only one property (dlp:involves) is needed to link characters to relationships, and roles can be defined as crm:E55_Type using a controlled vocabulary.

  • Linking relationships to events: The ontology links relationships to events. The relationship history is a sequence of events, which could be modelled by the relationship between events (see event module). Instead of making relationship history as an class using dlp:course, as Mika & Gangemi (2016) did, we use the property dlp:generically-dependent-on` to directly link relationship and event. In this way, we can simplify the model. Also, this allows modeling the evolution and dynamics of relationships, as events or statives are significantly contribute to changes in relationships. By connecting events to relationships and participants, we can model how relationships evolve over time.

Limitations

Sign and strength aspects: Currently, the ontology model does not include aspects related to the sign (positive or negative) and strength (strong or weak) of relationships. These aspects contribute to a more detailed understanding of relationship semantics. Characters within the same relationship may have different attitudes (sign) and varying strengths of identification, which are not captured in this model.

References

Chaturvedi, S., Srivastava, S., Daume III, H., & Dyer, C. (2015). Modeling dynamic relationships between characters in literary novels. arXiv preprint arXiv:1511.09376. http://arxiv.org/abs/1511.09376

Mika, P., & Gangemi, A. (2016). Descriptions of social relations. Benefits, 1, 14. https://doi.org/10.3390/benefits1010014

Clone this wiki locally