Skip to main content

Claim Relations User Guide

Overview​

Claim relations create typed connections between claims, allowing you to model complex relationships like support, contradiction, causation, and more. This guide explains how to create and manage these relationships.

What Are Claim Relations?​

A claim relation connects two claims with a specific relationship type:

  • Source Claim: The claim making the assertion
  • Relation Type: How they're related (supports, conflicts, causes, etc.)
  • Target Claim: The claim being referenced
  • Confidence: How certain you are about this relationship (optional)
  • Notes: Additional context (optional)

Viewing Relations​

Accessing the Relations View​

  1. Navigate to any claim in the tree view
  2. Click the Relations icon (network/tree icon)
  3. The relations panel expands below the claim

Understanding the Display​

Outgoing Relations

  • Relations where this claim is the source
  • Shows: Relation type → Target claim
  • Example: "Baseball is popular" supports → "Sports attract audiences"

Incoming Relations

  • Relations where this claim is the target
  • Shows: Source claim → Relation type
  • Example: "MLB exists" → supports "Baseball is popular"

Creating Relations​

Step-by-Step Process​

  1. Select the source claim

  2. Open its relations view

  3. Click Add Relation

  4. The relation editor dialog opens:

    • Source Claim (read-only): The claim you started from
    • Relation Type: Choose from available types
    • Target Claim: Select the target from dropdown
    • Confidence: Set your certainty (default: 80%)
    • Notes: Add optional explanation

  5. Click Save Relation

Choosing a Relation Type​

Your available relation types come from your ontology. Common types include:

Epistemic Relations (about knowledge/belief)

  • supports: Provides evidence for
  • conflicts: Contradicts or opposes
  • refutes: Directly disproves
  • questions: Raises doubt about

Logical Relations

  • implies: Logically entails
  • presupposes: Assumes as prerequisite
  • follows_from: Is a consequence of

Causal Relations

  • causes: Brings about
  • enables: Makes possible
  • prevents: Stops from happening

Temporal Relations

  • precedes: Comes before
  • follows: Comes after
  • coincides_with: Happens at same time

Creating Compatible Relations​

⚠️ Not all relation types work with claims. The system filters to show only claim-compatible types.

If you see "No compatible relation types":

  1. Go to Ontology Workspace
  2. Create or edit relation types
  3. Ensure sourceTypes includes "claim"
  4. Ensure targetTypes includes "claim"

Setting Confidence​

Confidence indicates how certain you are about the relationship:

  • 90-100%: Very strong connection, well-established
  • 70-89%: Strong connection, good evidence
  • 50-69%: Moderate connection, some uncertainty
  • 30-49%: Weak connection, tentative
  • 0-29%: Very weak, speculative

When to Use Lower Confidence​

  • Indirect or complex relationships
  • Interpretative connections
  • Preliminary analysis
  • Debatable associations

When to Use Higher Confidence​

  • Direct logical implications
  • Clear causal relationships
  • Well-documented connections
  • Unambiguous support/conflict

Adding Context Notes​

Notes help explain non-obvious relationships:

Good Note Examples​

✅ "This supports the main thesis because..." ✅ "Conflicts on the basis that X assumes Y, but..." ✅ "Causal connection demonstrated by studies A, B, C" ✅ "Temporal ordering established by video timestamps"

When Notes Are Helpful​

  • Complex or subtle relationships
  • Multiple competing interpretations
  • Relationships requiring domain knowledge
  • Collaborative work (explain reasoning to others)

Managing Relations​

Deleting a Relation​

  1. Open the relations view for a claim
  2. Find the relation to delete
  3. Click the Delete icon (trash)
  4. Confirm deletion

Note: This only deletes the relationship, not the claims themselves.

Reviewing Relations​

Use the relations view to:

  • Audit claim networks
  • Find contradictions
  • Identify support structures
  • Trace logical chains
  • Verify relationship validity

Relation Patterns​

Support Chains​

Claim A supports Claim B supports Claim C

Build evidential hierarchies where each claim supports the next.

Mutual Support​

Claim A supports Claim B
Claim B supports Claim A

Useful for reciprocal relationships, but watch for circular reasoning.

Conflict Networks​

Claim A conflicts Claim B
Claim A conflicts Claim C
Claim B conflicts Claim C

Model mutually exclusive alternatives or contradictory viewpoints.

Causal Chains​

Claim A causes Claim B causes Claim C

Map cause-effect relationships through multiple steps.

Best Practices​

Relationship Quality​

Do:

  • Create specific, meaningful relations
  • Use appropriate relation types
  • Add notes for complex relationships
  • Set realistic confidence scores
  • Review relations periodically

Don't:

  • Create relations just because claims mention similar topics
  • Use generic "related_to" for everything (be specific)
  • Set maximum confidence without strong justification
  • Create circular reasoning loops
  • Leave ambiguous relations without notes

Organizing Relations​

Start Simple

  • Begin with obvious support/conflict relations
  • Add more nuanced relations later
  • Focus on key claims first

Work Systematically

  • Process one claim at a time
  • Check both incoming and outgoing relations
  • Look for missing connections

Maintain Consistency

  • Use relation types consistently
  • Apply similar confidence standards
  • Follow team conventions

Advanced Use Cases​

Argument Mapping​

Use relations to model complete arguments:

  1. Main conclusion (root claim)
  2. Premises (supporting claims)
  3. Objections (conflicting claims)
  4. Rebuttals (conflicts to objections)

Belief Networks​

Model how evidence accumulates:

  • Multiple weak supports = moderate confidence
  • Single strong support = high confidence
  • Conflicting evidence = uncertainty

Comparative Analysis​

Compare multiple viewpoints:

  • Claims from different sources
  • Relations show agreements/disagreements
  • Build neutral claim network

Temporal Narratives​

Track how claims evolve:

  • Use precedes/follows relations
  • Model belief changes over time
  • Show historical development

Troubleshooting​

Can't Find Target Claim

  • Verify the claim exists
  • Check you're not trying to relate to the source itself
  • Use search/filter in the claim dropdown

Relation Type Unavailable

  • Check ontology settings
  • Verify relation type supports claim↔claim
  • Create new relation type if needed

Confidence Scores Unclear

  • Review your confidence criteria
  • Consult with team members
  • Use notes to explain your reasoning

Too Many Relations

  • Focus on most important connections
  • Remove redundant relations
  • Use hierarchy instead of many flat relations

Integration with Other Features​

With Subclaims​

  • Parent-child is structural (not a relation)
  • Relations are semantic connections
  • Can relate claims at different hierarchy levels

With Annotations​

  • Relate claims that reference same objects
  • Build networks around key events
  • Connect temporal claims

With Ontology​

  • Relation types come from ontology
  • Customize types for your domain
  • Define clear semantics

Next Steps​

  • Build argument maps with your claims
  • Analyze claim networks for insights
  • Export relation graphs
  • Integrate with analysis tools

See Also​