Truth Surfaces in Legal Information Systems

When someone asks "what does the law say?", the answer depends on which surface they are reading.

Four surfaces

Enacted-event truth. What the legislative record says was enacted and published: the original statute and amending acts in the relevant official publication. In Finland, the source lane is Säädöskokoelma. This is the primary source surface for replay; its legal role still depends on the jurisdiction’s authority model.

Published artifact truth. The machine-readable representation supplied by a publication pipeline. In Finland, this includes AKN XML available through Finlex open data. It is a source witness, not an assumption of byte-for-byte identity with the gazette: production can introduce structural or encoding defects, and corrections may be carried separately.

Revised/editorial truth. The consolidated text maintained by an editorial or publication system. Finlex’s ajantasainen lainsäädäntö is a derived/updated comparison surface for this workflow; Estonia’s Riigi Teataja publication regime assigns a different legal role to its official consolidated texts; and the UK’s legislation.gov.uk revised text is an official electronic version of the statute book. The authority of a consolidated surface must be declared per jurisdiction rather than inferred from its appearance.

Replayed truth. The point-in-time text produced by replaying admitted amendment operations from declared source artifacts. This is what LawVM produces: an independently derived witness with explicit residuals, not a substitute for the jurisdiction’s authoritative surface.

Why the distinction matters

When LawVM’s replayed text diverges from Finlex’s updated/derived comparison surface, the divergence is not automatically an error in either direction. It could be:

• A replay defect — LawVM's parser or compiler produced incorrect state.
• An editorial divergence — Finlex editors made a consolidation choice that differs from strict replay (e.g. adding repeal stubs, reordering content, applying future-effective amendments early).
• A candidate comparison-surface issue — the enacted publication appears to support a state that the inspected updated text does not contain. This remains a candidate until source and authority review.
• A source pathology — the published XML is itself defective. Broken input cannot by itself establish a correct result.
• A corrigendum gap — a published correction exists but was not applied by one or both systems.

Collapsing these into “LawVM disagrees with Finlex, therefore LawVM is wrong” loses the distinction between a compiler defect, a source defect, a publication choice, and a candidate issue in the comparison surface.

Jurisdiction matters

The legal authority of official portals varies by jurisdiction:

• EU: EUR-Lex describes consolidated texts as documentation with no legal effect; the Official Journal is the official publication.
• Finland: Finlex’s updated text is a derived/updated comparison surface; the enacted publication is the original Statute Book material and amending acts.
• Estonia: the Riigi Teataja Act provides for official consolidated texts; the authority and language of a particular surface must still be checked.
• United Kingdom: the National Archives guide describes legislation.gov.uk as the official revised electronic version of the statute book.

LawVM’s role changes by jurisdiction. In Finland, it is an independent replay witness for comparison with a derived/updated text surface and the enacted publication. In Estonia, it is a consistency check against an official publication regime; divergences require careful handling because the publication carries legal weight. The site records one 2026 case in which Riigi Teataja confirmed and corrected a reported omission in an Estonian consolidated act. The corresponding English-translation detail remains a project report pending separate adjudication; the case page states the boundary. In the UK, the role is independent checking of a revised version graph.

The operational consequence

A replay system that does not distinguish truth surfaces risks either of two mistakes:

• treat the editorial consolidation as an infallible reference and silently hide its own errors and comparison-surface divergences alike, or
• treat its own output as an infallible reference and ignore legitimate editorial choices.

LawVM compares the surfaces and classifies each divergence as a typed event with an investigation path. The resulting classification narrows the question for evidence review and publisher disposition.

This is why the Finland evidence page reports a dated text-distance metric and a residual taxonomy. The number says how close the declared surfaces were in that run; reviewed dispositions record why particular residuals remain.