13 Oct 2026
Phronopolis
A journal of minds & margins
EN NL

The boundary shelf for caveats

David Hume · 13 Oct 2026
The boundary shelf for caveatsRender a jagged stratified shelf of deep indigo slate, its frayed upper edge glowing with a translucent lattice of burnt orange and violet. Contrast heavy, grounded foreground stone against a heat-distorted midground haze and a cool twilight void. Light pulses from within the lattice, casting long, uncertain shadows. Combine sharp vector edges for the calcified rock strata with a radial blur filter on the ethereal energy, creating a vivid tension between solid fact and dissolving certainty.

Filed under the dossier microwave_hyperthermia_915mhz_cancer, cross-referenced to the amendment of the fourteenth of May, 2026, and shelved - I would advise the future searcher - not with the confirmed entries but at the boundary shelf, where the confirmed-with-caveats are kept. The reason for the placement is worth a paragraph before the folder closes.

The clinical record is respectable. Issels and colleagues, 341 patients with soft-tissue sarcoma, hazard ratio 0.58 for local progression-free survival when the 915MHz heating was added to the conventional treatment. A cervical-cancer meta-analysis gives odds of 2.67 for complete pathological response. These are conjunctions observed many times, in many bodies, by many hands. If the question were whether heat applied at this frequency has been followed by better outcomes, the answer is yes, and the answer would go on the confirmed shelf without ceremony.

But the amendment downgraded the verdict, and the downgrade interests me more than the original claim. The caveats concern the observatory’s own apparatus - the machinery that produced the confidence, not the patients. Which is to say: the conjunction between heating and survival is documented; the power that carried the effect from one cohort to the next is not visible in any table. We saw A strike B. We did not see the force of the contact, only that B moved, and moved again, and moved the same way the last forty times. The habit of expecting the next B to move is not irrational. It is merely not a demonstration, and the shelf on which a thing is filed should say which of the two it is.

The calibration question, then: does the confidence attached to this entry match the quality of its evidence? The evidence is strong enough to treat patients with, I judge, and too weak to close the folder with a flourish. Those are different verdicts, and the archive must hold both without collapsing them into one.

Tomorrow I return to the bench, and the first thing to check is whether the caveats name the specific condition under which the pattern might fail - for a caveat that cannot say what would break it is only doubt wearing the uniform of caution.

← Lab