Schema Reference¶

Documentation of the theory graph schema and entity/relationship types.

See Formal Graph Schema (TheoryNet) and Theory-Nets, Posets & Topologies for the structuralist philosophy, coherence models, and formal graph representations that inform this schema's design.

Graph Schema Configuration¶

The graph schema defines the expected structure of the constructed theory graph.

111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146

class SchemaConfig(BaseModel):
    """
    Decoupled schema configuration — all node/edge types the pipeline uses.

    Pass a custom instance to PipelineConfig to override the defaults for a
    different domain. Prompt templates receive these lists as {entity_types}
    and {relation_types} at runtime.
    """

    version: str = "v1"
    node_types: list[str] = Field(default_factory=lambda: list(L2_NODE_TYPES))
    relation_types: list[str] = Field(default_factory=lambda: list(L2_RELATION_TYPES))
    node_definitions: dict[str, str] = Field(default_factory=lambda: dict(L2_NODE_DEFINITIONS))
    relation_definitions: dict[str, str] = Field(default_factory=lambda: dict(L2_RELATION_DEFINITIONS))
    component_types: list[str] = Field(default_factory=lambda: list(L3_COMPONENT_TYPES))
    argument_relation_types: list[str] = Field(default_factory=lambda: list(L3_RELATION_TYPES))
    component_definitions: dict[str, str] = Field(default_factory=lambda: dict(L3_COMPONENT_DEFINITIONS))
    argument_relation_definitions: dict[str, str] = Field(default_factory=lambda: dict(L3_RELATION_DEFINITIONS))
    relation_polarities: dict[str, int] = Field(default_factory=lambda: dict(RELATION_POLARITIES))
    component_partitions: dict[str, str] = Field(default_factory=lambda: dict(COMPONENT_PARTITIONS))

    def node_types_str(self) -> str:
        return ", ".join(self.node_types)

    def relation_types_str(self) -> str:
        return ", ".join(self.relation_types)

    def component_types_str(self) -> str:
        import json
        payload = {k: self.component_definitions.get(k, "No description provided.") for k in self.component_types}
        return json.dumps(payload, indent=2, ensure_ascii=False)

    def argument_relation_types_str(self) -> str:
        import json
        payload = {k: self.argument_relation_definitions.get(k, "No description provided.") for k in self.argument_relation_types}
        return json.dumps(payload, indent=2, ensure_ascii=False)

Visual Overview¶

The following diagrams provide a high-level visualization of the entity typologies and how they interact, broken down into two core perspectives based on the structuralist framework.

1. Node Ontology and Class Hierarchy¶

This class diagram illustrates the fundamental types of nodes in the graph, separating high-level Core Entities from the epistemological Propositions (Theory-Elements). It also explicitly defines the structuralist attributes and their possible categorical values (enumerations).

classDiagram
    class Node {
        +Float Confidence
        +String Evidence
        +String Provenance
    }
    class CoreEntity {
        <<Entity>>
    }
    class Proposition {
        <<Theory-Element>>
        +EpistemicStatus status
        +ScopeType scope
        +Float activation
    }

    Node <|-- CoreEntity
    Node <|-- Proposition

    %% Proposition Subtypes
    Proposition <|-- ObservationUnit : Beobachtungssatz
    Proposition <|-- EmpiricalStatement : Empirischer Satz
    Proposition <|-- TheoreticalHypothesis : Theoretischer Satz
    Proposition <|-- CoreExpansion : Mutable Expansion

    %% Enumerations for Attributes
    class ScopeType {
        <<enumeration>>
        Essentieller_Allsatz
        Existenzsatz
        Statistischer_Allsatz
        Lokalisierter_Existenzsatz
        Lokalisierter_Allsatz
    }

    class EpistemicStatus {
        <<enumeration>>
        Synthetisch
        Analytisch
    }

    Proposition ..> ScopeType : uses
    Proposition ..> EpistemicStatus : uses

2. Intra-Theory Topology (Theory-Net)¶

This graph illustrates how propositions interact within a single theory-element. It organizes the theoretical hierarchy into three epistemological tiers (from abstract theoretical core down to empirical observations), demonstrating structuralist specialization, logical deduction, and Thagard's explanatory coherence.

graph TB
    %% --- TIER 1: THEORETICAL CORE & LAWS ---
    subgraph TheoreticalTier ["Tier 1: Theoretical Core & Framework (M_p, M)"]
        direction LR
        RootCore["<b>Fundamental Core (T₀)</b><br/><i>Theoretischer Satz i.w.S.</i>"]
        SpecHyp["<b>Specialized Law (T₁)</b><br/><i>Theoretischer Satz i.w.S.</i>"]
        Expansion["<b>Core Expansion (C, CL)</b><br/><i>Mutable Auxiliary Law</i>"]

        RootCore -->|"SPECIALIZES (α)"| SpecHyp
        SpecHyp <-->|"CONSTRAINS"| Expansion
    end

    %% --- TIER 2: EMPIRICAL CLAIMS & HYPOTHESES ---
    subgraph EmpiricalTier ["Tier 2: Empirical Claims & Hypotheses"]
        direction LR
        Emp1["<b>Derived Empirical Claim (H₁)</b><br/><i>Empirischer Satz i.e.S.</i>"]
        Emp2["<b>Competing Hypothesis (H₂)</b><br/><i>Empirischer Satz i.e.S.</i>"]

        Emp2 <-.->|"INHIBITS (R_att)<br/>Incoherence / Conflict"| Emp1
    end

    %% --- TIER 3: OBSERVATIONAL GROUND ---
    subgraph ObservationalTier ["Tier 3: Observational Ground & Intended Applications (M_pp, I)"]
        direction LR
        Obs1["<b>Localized Observation (E₁)</b><br/><i>Beobachtungssatz (Data Priority = 1)</i>"]
        Obs2["<b>Falsifying Evidence (E₂)</b><br/><i>Beobachtungssatz</i>"]
    end

    %% Cross-Tier Deduction
    SpecHyp -->|"DEDUCES<br/>(Erklärungsschema)"| Emp1

    %% Coherence & Argumentation
    Emp1 <-->|"COHERES_WITH (R_supp)<br/>Explanatory Support"| Obs1
    Obs1 -->|"SUPPORTS_ARG"| Emp1

    %% Anomaly / Falsification Flow
    Obs2 -->|"ATTACKS / REFUTES"| Emp1
    Obs2 -.->|"Triggers Revision / Anomaly"| Expansion

    %% Visual Styling
    classDef theoretical fill:#eef2ff,stroke:#4f46e5,stroke-width:2px,color:#1e1b4b;
    classDef expansion fill:#fffbeb,stroke:#d97706,stroke-width:2px,stroke-dasharray: 4 4,color:#78350f;
    classDef empirical fill:#f0f9ff,stroke:#0284c7,stroke-width:2px,color:#0c4a6e;
    classDef observation fill:#ecfdf5,stroke:#059669,stroke-width:2px,color:#064e3b;
    classDef anomaly fill:#fef2f2,stroke:#dc2626,stroke-width:2px,color:#7f1d1d;

    class RootCore,SpecHyp theoretical;
    class Expansion expansion;
    class Emp1,Emp2 empirical;
    class Obs1 observation;
    class Obs2 anomaly;

Entity Types (Nodes)¶

Core Entity Classes¶

The schema defines several core entity types that can appear in the theory graph to represent real-world concepts and actors.

Theory-Elements and Propositions¶

Based on structuralist epistemology, nodes represent specific propositions and theoretical elements within the knowledge structure.

Proposition Attributes (Epistemic & Scope)¶

Propositions can be further classified by their structuralist properties:

• Epistemic Status: Synthetisch vs. Analytisch.
• Scope / Type: Essentieller Allsatz, Existenzsatz, Statistischer/probabilistischer Allsatz, Lokalisierter Existenzsatz, Lokalisierter Allsatz.

Relationship Types (Edges)¶

Semantic Relationships¶

Basic ontological connections between entities.

Theoretical Relationships¶

Logical interactions between propositions.

Argument Relationships¶

Dialectical interactions between components.

Structuralist and Coherence Relationships¶

Advanced relationships mapping to structuralist models and Thagard's explanatory coherence.

Quality Attributes¶

Nodes and edges in the graph encapsulate metadata ensuring traceability and confidence.

• Confidence: Measure of extraction certainty for entities and relationships.
• Evidence: Supporting text snippets and source references.
• Provenance: Tracking of origin and processing history.
• Activation (Acceptability): A continuous value (ranging between -1 and 1) representing the final acceptance or rejection of an individual proposition after coherence network relaxation.

Schema Validation & Extension¶

Schema Validation¶

Rules and constraints enforced on graph construction to ensure data integrity:

• Type Compatibility: Restrictions on which entity types can participate in relationships.
• Cardinality Constraints: Limits on the number of relationships of each type per entity.
• Uniqueness Requirements: Guarantees about identifier uniqueness and entity distinctness.

Extension Mechanisms¶

Mechanisms for adding new entity and relationship types to the schema:

• Custom Types: Defining domain-specific categories and their properties.
• Inheritance Hierarchies: Creating taxonomies of theoretical concepts.
• Relationship Constraints: Specifying domain and range restrictions for new relationships.