CALC-047

Vocabulary sweep with counts: insurance and liability across the XC04 sources

Feeds CLM-0765. Answers QST-011. SEARCH-METHOD, not arithmetic. Date 2026-09-04.

Why it was needed. XC04's target was raised to L2 on the reasoning that it is the layer an insurer or a minister asks about first, and an insurer's first question is what it costs to cover. If fourteen substantial sources contain no premium, no cover limit and no liability quantum, the layer's own evidence base cannot answer its own first question.

Corpus. All fourteen sources retrieved for XC04 on 2026-09-04, SRC-223 to SRC-236. Twelve machine-readable PDFs swept mechanically; SRC-230 and SRC-231 are saved web pages and were read in full by hand.

Method. Case-insensitive regular expression over pdftotext output.

Pattern:  insur|liabilit|premium|indemnif|indemnit|underwrit|actuarial

Counts, and what every match resolved to. This is the part that makes the null a measurement.

Result: 17 raw matches, one substantive. That one is SRC-226 printed p.12, verbatim: "The threat of space debris to the future of spaceflight combined with the nearly universal adoption of the Liability Convention [7] created the need for a set of internationally accepted space debris mitigation measures."

Conclusion. No insurance premium, coverage limit, liability quantum, actuarial rate or exclusion clause for any space asset appears in any of the fourteen sources. The total substantive insurance and liability content is one reference to the 1972 Liability Convention.

Where this fails, and one of the two failures has already occurred.

1. This measures vocabulary over extracted text, not substance. Per METHOD's O1 lesson the measurement is named rather than just the number: these are regular-expression occurrence counts over pdftotext output. It says only that these fourteen sources contain no such figure. It does not show that no space-insurance data exists - a space insurance market plainly does - and it may not be cited as evidence that a GEO asset is uninsurable. 2. The premise about SRC-226's text layer is unreproducible, and this is live. The claim assumes the PDF text layers are complete for prose but not for figures in SRC-226. An independent verification pass could not reproduce that. The consequence was material elsewhere: a retrievable 2025 GEO population of 1,869 objects was recorded as an ABSENCE in CLM-0754. A wrongly recorded absence is worse than a missing claim, because it stops anyone looking again. SRC-226's figures must be re-read before this record or its siblings are relied on.

Also load-bearing: these fourteen sources were retrieved for counterspace threat, debris and command-link security, so the absence is partly explained by what they were selected for. The finding is that the layer's own evidence base cannot answer the question, not that the question is unanswerable.

Cite this