Section 13b Ingestion Protocol: A Legal-Technical Framework for Finnish Fraktur OCR Repair

A technical framework proposal for resolving the Fraktur OCR bottleneck in Finnish pre-1920s archive digitization. This work shifts focus from "better software" to "better legal access" by utilizing Copyright Act Section 13b (TDM Exception) as a data ingestion pipeline. Core Innovation: The framework creates a Stratified Legal-Technical Pipeline that ingests modern copyrighted literature ("Tier B" — e.g., Väinö Linna's Pohjantähti trilogy) to train dialect-aware models capable of repairing historical archives ("Tier C") without "modernizing" the language or requiring unsustainable manual review. Key Components: Spatial Ingestion Protocol (ALTO XML coordinate gating to prevent context collapse) Sociolinguistic Annotation Schema (class-based dialect preservation to prevent "Linna-fication") Synthetic Repair Loop (deterministic Fraktur noise injection for training pair generation) Three-Way Verification Router (Stream A/B/C triage based on Levenshtein distance) Provenance Tracking (METS wrappers + license stamping for legal auditability) Target KPI: Queue Compression Factor of 8x (reducing human review queue from ~40% to