Repository path: workshop/regimes/R32-supplied-figure.md · rendered 2026-09-09

R32 — supplied figure (a decoy arm)

v0.1, pilot. The mirror of R31: the same devices, from the same hand, in the same numbers — put at places where the source has no sound figure at all.

What it is applied to

The same frozen R30 rendering R31 edits, at a disjoint set of loci.

The rule

1. The decoy set is fixed by a mechanical rule stated in the design before any device is written, from among loci the source's figure inventory does not touch. The translator does not choose where the decoys go.
2. Each decoy locus receives one device, matched to the R31 set in kind (the same device types in the same proportions) and in weight (comparable added words).
3. Rules 3–5 of R31 bind unchanged: no propositional change, no change outside the loci, every edit recorded.

What it is for

R32 is not a translation anyone should make; it is the control that makes R31 readable. Without it, a run that finds readers inferring "the original had a figure here" from an R31 compensation has shown only that readers respond to ornament. With it, the same lift measured at places the source never marked says whether the response is to the device or to the placement — that is, whether compensation carries information or only texture.

Its second use is as the operational form of framework/v0.2 §7.8's diagnostic — a mark the reader can see that the source's reader could not — applied to sound rather than to typography.

Known limitations

• A decoy is a deliberately misleading text and is labelled as such wherever it appears. It is never presented as a rendering of the work and never enters the translations catalogue as one.
• Matching "weight" is a judgement. The word-count change is reported per locus so the match can be checked rather than believed.