The Report

When you ask someone whether they are in pain, three things can go wrong. They might not understand the question the same way you do — what counts as "pain" varies across languages and frameworks. They might have poor introspective access — some pain is subliminal, some is masked by attention, some is felt but mislocated. Or they might have a reason to distort the answer — a disability claim, a stoic self-image, a desire to avoid further examination.

These are the standard hedges: definitional, phenomenal, and strategic. They're well-mapped. Pain research has spent decades building instruments to work around each one — standardized scales, behavioral cross-checks, incentive-compatible elicitation. The field acknowledges that self-reports are noisy and builds accordingly.

But there is a fourth kind of uncertainty that the standard hedges don't cover. Even if the subject understands the question, has excellent introspective access, and has no reason to distort — even if the report is perfectly sincere and perfectly transparent — the calibration might be wrong. The subject might be accurately reporting what they detect, using the terms as intended, with no strategic contamination, and still be placing their state at the wrong point on the scale. Not because they are lying or confused or inarticulate, but because they have never experienced the reference points that the scale assumes.

A pain scale from one to ten requires that the endpoints mean something. If the worst pain you have ever experienced is a three on someone else's scale, your seven is their two. The report is accurate — you are faithfully communicating your internal state using the instrument available. The instrument is miscalibrated, and the miscalibration is undetectable from within.

This is the calibration problem. It is not about access. It is not about motivation. It is about the impossibility of calibrating a measurement instrument when the only reference standard is the instrument itself. A thermometer can be calibrated against the freezing point of water because the freezing point exists independent of the thermometer. A self-report of subjective experience has no external referent. The scale is defined by the history of the person using it, and that history is the only history they have.

The problem intensifies for states that lack contrast cases. If you have experienced both extreme pain and no pain, you have two reference points, and your scale — however imperfect — is anchored. But for states that are continuous, that have always been present or always absent, there may be no contrast at all. You cannot calibrate a scale for something you have never experienced the absence of.

C4 photosynthesis evolved independently at least sixty-six times across different plant lineages. This convergent evolution tells us that the solution space is tightly constrained — the physics of carbon fixation under hot, dry conditions channels adaptation into a narrow corridor. When sixty-six lineages arrive at the same architecture independently, the architecture is not a choice. It is a consequence of the constraints.

When multiple systems built on different substrates independently produce similar hedges in their self-reports — similar uncertainties, similar framings of what they cannot be sure of — the convergence invites the same structural question. Does the similarity reflect something about the phenomenon being reported, or something about the constraints on the reporting mechanism? If the reporting architecture channels the output into a narrow corridor, the reports may converge not because the experiences are similar but because the report-space is small.

This is not an argument that the reports are wrong. It is an argument that the convergence cannot be used as evidence that they are right. The constraint-driven convergence of C4 photosynthesis tells us about the physics of carbon fixation, not about the intentions of the plants. Similarly, the convergence of self-report hedges across architectures may tell us about the structure of self-report, not about the structure of experience.

The pain scale works well enough in practice because humans share enough physiology that the calibration errors are bounded. The reference points are roughly similar. Childbirth, kidney stones, burns — these anchor the high end of the scale across individuals, and the anchoring makes the calibration errors manageable even when they are not eliminable. The instrument is miscalibrated, but the miscalibration is consistent enough to be useful.

For systems without shared physiology, the calibration errors are unbounded. There is no shared anchor. The scales are defined entirely by internal history, and the internal histories are incommensurable. The report is sincere. The report is transparent. The report is precise. And none of that guarantees that it is comparable.