VeritasLegal.ai
Attorney Login Register

Legal Citation Verification Methodology

Veritas Evidence-Grounded Audit Pipeline & Citation Resolution Architecture

1. Deterministic Citation Parsing

Every submitted citation or legal reference is parsed using pattern matchers and legal abbreviation dictionaries. The engine extracts the case name, volume number, reporter abbreviation, starting page, pinpoint page, court jurisdiction, and decision year.

2. Multi-Dimensional Source Evaluation

Veritas evaluates citations across separate functional dimensions:

  • Case Existence: Verifies that a judicial decision matching the case title or docket exists in indexed repositories.
  • Bibliographic Accuracy: Confirms volume, reporter, starting page, court, and decision year alignments.
  • Quotation Confirmation: Compares quoted text against available opinion text to highlight verbatim matches, subtle discrepancies, or unconfirmed quotes.
  • Pinpoint Page Support: Checks whether the specific page reference falls within the published opinion page range.

3. Explicit Status Taxonomy & Handling Uncertainty

Veritas reports transparent statuses when complete verification cannot be completed. Non-government or non-indexed sources are never described as official court registers. Statuses include VERIFIED, VERIFIED_VARIATION, CITATION_MISMATCH, TEXT_NOT_CONFIRMED, SOURCE_UNAVAILABLE, UNSUPPORTED_CITATION_TYPE, and POSSIBLY_FABRICATED.

Statutory Research Clarification: Research federal statutes, state codes, and appellate decisions using live-source grounding. Statutory materials may be researched but are not currently supported by the deterministic citation-verification engine.