Burden of Proof Shifting

The initial threshold that activates the shifting mechanism. A party must establish a prima facie case by producing sufficient evidence for each element of their claim.

• Computational Model: A state machine transitions from INITIAL to BURDEN_ON_RESPONDENT when a claim's evidentiary score exceeds a predefined threshold.
• Key Challenge: Determining the sufficiency threshold algorithmically, often requiring domain-specific calibration on historical case data.
• Example: In discrimination law, a plaintiff shows membership in a protected class, adverse action, and circumstances giving rise to an inference of discrimination.

A critical distinction in burden shifting models. The burden of production (rebuttal) requires a party to introduce evidence, while the burden of persuasion remains fixed on the original proponent.

• Production Burden: A lightweight obligation to merely introduce some contrary evidence. Failure results in a directed verdict.
• Persuasion Burden: The heavy obligation to convince the trier of fact. This typically never shifts.
• Modeling Implication: Systems must maintain two separate state variables—one tracking who must speak next, and one tracking who must ultimately prove the case.

The paradigmatic three-stage burden shifting protocol from employment discrimination law, widely used as a benchmark for computational models.

• Stage 1: Plaintiff establishes a prima facie case.
• Stage 2: Defendant articulates a legitimate, non-discriminatory reason for the adverse action.
• Stage 3: Plaintiff demonstrates that the proffered reason is pretextual.
• Formalization: This maps cleanly to a finite state automaton with transitions triggered by satisfaction of each stage's evidentiary predicate.

A mechanism where a legal presumption automatically shifts the burden. Unlike the McDonnell Douglas framework, this shift is triggered by a rule, not a factual showing.

• Rebuttable Presumption: Shifts the burden of production to the opposing party. Example: a person missing for seven years is presumed dead.
• Conclusive Presumption: A rule of substantive law that cannot be rebutted. No burden shift occurs; the fact is simply established.
• Computational Representation: Modeled as an IF-THEN production rule that modifies the argument graph's obligation nodes.

A distinct burden shift where the respondent introduces new matter that, if proven, defeats the claim regardless of its truth. The burden of persuasion uniquely shifts to the respondent for this new issue.

• Examples: Statute of limitations, self-defense, contributory negligence.
• Argument Graph Impact: The defense creates a new sub-argument tree where the respondent is the proponent, requiring the system to instantiate a new proof obligation.
• Detection Task: Classifying a text span as an affirmative defense rather than a denial, which triggers different procedural consequences.

The core computational task of monitoring which party currently bears the burden for each discrete issue in a multi-issue case.

• Issue-Level Granularity: Burden is tracked per legal issue, not globally. Party A may bear the burden on Issue 1 while Party B bears it on Issue 2.
• Temporal Dynamics: The burden state evolves as new pleadings, motions, and evidence enter the docket.
• Implementation: A burden ledger data structure mapping (issue_id, party_id) -> obligation_type that updates with each new document in the case lifecycle.

The formal representation of legal arguments that can be invalidated by exceptions or contrary evidence. This non-monotonic logic is essential for burden shifting, as a valid claim can be defeated by new evidence, shifting the burden back to the proponent to introduce a stronger argument or exception.

• Captures the revisable nature of legal conclusions
• Models how presumptions are defeated
• Uses default logic and answer set programming

The binary or multi-class task of determining whether one legal argument component strengthens, weakens, or is neutral toward another. Burden shifting is triggered when an attack relation is successfully classified, indicating the respondent has met their burden of production and the proponent must now respond.

• Binary: Support vs. Attack
• Multi-class: Strong attack, weak attack, neutral
• Foundational for automated burden tracking

A foundational mathematical framework that models arguments as abstract nodes in a directed graph, focusing solely on attack relations to determine acceptable sets of claims. Burden shifting corresponds to computing new preferred extensions or grounded semantics after introducing a counter-argument.

• Abstract nodes ignore internal structure
• Attack relations define conflict
• Semantics compute justified positions

The automated synthesis of a plausible opposing legal argument to a given claim. In burden-shifting systems, this capability simulates the respondent's move, generating the evidence or reasoning that, if accepted, shifts the burden of persuasion back to the original proponent.

• Used for case strategy stress-testing
• Generates rebuttals based on precedent
• Tests the resilience of a prima facie case

A metric that quantifies the logical consistency and internal connectivity of a set of legal arguments. After a burden shift occurs, the system must re-score the entire argument graph to ensure the new state is not self-contradictory and that the shifted burden has been properly discharged.

• Detects circular reasoning
• Measures graph density and connectivity
• Ensures post-shift argument validity