Temporal Misgrounding in Legal RAG: A Versioned-Corpus Benchmark for French Tax Law
2026-08-10 • Computation and Language
Computation and LanguageArtificial IntelligenceInformation Retrieval
AI summaryⓘ
The authors studied a problem called temporal misgrounding, where legal systems mistakenly use the current version of a law instead of the correct past or future version that applies. They created a special dataset called FiscalQA Pro using many versions of French tax laws spanning from 1938 to 2031, along with challenging questions that require understanding the correct law version by date. They found that existing models usually pick the wrong version and perform poorly, but their new retrieval method that considers multiple versions works very well. They also provided resources like citation data and code to help future research in version-aware legal question answering.
temporal misgroundinglegal retrievalversioned corpusFrench tax codetemporal reasoningretriever-augmented generation (RAG)legal question answeringjurisprudence citationsdataset benchmarkmodel evaluation
Authors
Rose Cymbler, Daniel Guez, Laurent Fabre
Abstract
We identify and quantify temporal misgrounding: the systematic retrieval and citation of the currently in-force version of a legal article when the applicable version is an earlier or future one. Standard legal RAG treats the corpus as static; we argue legal question answering is a temporally-indexed retrieval problem. We introduce FiscalQA Pro, pairing a versioned corpus of 32,436 article-versions of the French tax code (93 years, 1938-2031) with an all-model-hard temporal-reasoning track: 209 scored, expert-reviewed questions across 33 CGI articles (221 released; twelve flagged out of the answerable scope). At selection time, no evaluated model recovered its date-applicable answer closed-book in any of four sampling draws, and the currently in-force text lacks the gold value for all but one of the scored questions. Answers are scored deterministically via atomic ground-truth "nuggets" (regex and numeric-with-tolerance), never LLM-as-judge: an LLM judge would inherit the temporal bias it is meant to score. Across eleven models (five frontier closed-API systems plus Gemini 2.5 Pro as a substitute entry, and five open-weight), parametric knowledge yields 3.0% mean strict accuracy and RAG over a static current-version corpus 2.7%. Static RAG retrieves the date-applicable version 0% of the time, confidently citing a real but inapplicable version. Our end-to-end retriever over a multi-version index, with no oracle, reaches 98.3% mean strict; an oracle-article ablation reaches 99.1%, locating the residual gap in first-stage recall, not version selection. We additionally release a version-aware jurisprudence dataset of 69,208 citation links, together with the corpus, benchmark, model responses, and pipeline code.