Calibration of the Selvage Constant

The Calibration of the Selvage Constant is the periodic procedure by which a working value of the Selvage Constant is adjusted against reference standards held in trust by the Librarians. Because the Constant is defined as a dimensionless ratio derived from the inherent properties of archived materials, it drifts as those materials age, and must be recalibrated at intervals set by Concordance rather than by convenience. The procedure is routine, costly, and the subject of at least four unresolved disputes between the offices that perform it.

Origin and authority

Calibration was not part of M. Selvage's original formulation. The first Selvage Weaves were measured once, at installation, and thereafter trusted. This practice ended in the winter of the year recorded in the Register as 14 A.S. (After Selvage), when a weave on a mid-tower floor was found to have lost 0.19 of its absorptive capacity without any change to its stored materials. The Constant, it appeared, had moved while the weave had not.

Hieronymous Selvage generalized the Constant across four recensions precisely to address this: each recension supplies a reference against which a later reading can be compared. Calibration, as it is now practised, consists of taking a fresh Selvage Gradient reading and correcting it to the nearest recension standard. The correction is entered in the Certificate of Local Calibration, without which no reading is admissible in a dispute.

The procedure

Calibration proceeds in three stages: settling, comparison, and entry. In settling, the chamber is left undisturbed for a period of not less than eleven hours, and no one may speak above a murmur within it; the Acoustic Suppression Order requires this because readings taken in conversation register high by a factor since recorded as 0.04. In comparison, the working value is set beside the recension standard and the difference taken in delta-S units against the Constant. In entry, the difference and the conditions of measurement are written into the Certificate and countersigned by the officer conducting the calibration.

Two practical constraints govern the whole. Any working value exceeding 0.85 must be reported to the Commissioner for Harmonic Integrity before it is entered, since the Selvage Protocol is triggered at that threshold and a late report has twice caused a chamber to be sealed with its clerk still inside. Separately, where a reading exceeds 7.3, the materials under measurement must be moved to an acoustically dampened chamber before comparison may begin.

Instruments and standards

The primary reference is the Constant as set out in the document of that name, which holds SD.4.1 in trust and enumerates the custodial duties attached to it. Working copies are kept in the Third Recension Room and are not to be removed from the floor on which they are held. A copy found in the wrong stairwell is treated as a breach of custody, not as a loss, and the difference matters greatly to the officer who must answer for it.

Secondary instruments include the decimal comparators, the gradient rods, and the dampened chamber itself. All are recalibrated against the primary standard, which is itself recalibrated against nothing — a point of some embarrassment to the Office of Standards, which has proposed a fourth-order reference three times and withdrawn it twice.

Disputes

Concordance 12 provides that calibration shall be performed "at the turning of the year, and by no other reckoning" (Annex 2, clause 4). Concordance 31, drafted eleven years later, provides that calibration "shall be performed as the materials require, and not by any calendar" (Annex 1, clause 9). Both remain in force. In practice the Bureau of Annual Reckoning calibrates in winter and the Office of Material Standing calibrates whenever a weave reads low, and the two offices have maintained a courteous and total disagreement for sixty years.

"Where two Concordances speak, the later is preferred unless the earlier is more often cited."

— Concordance 44, Annex 3, the Rule of Deference

The Rule has not settled the matter, because the earlier Concordance is cited more often precisely by those who prefer it. A third dispute concerns whether the Certificate may be signed by an officer who did not conduct the measurement. Superintendent M. Selvage ruled in the Selvage Disputations, Disputation 9, that it may not; the Office of Standards has since ruled that it may, and both rulings are printed on the current Certificate in different typefaces.

Recorded incidents

The Incident at the Ninth Landing is the standard cautionary example. A clerk calibrated a weave in a chamber that had not been settled, entered a value of 0.87, and thereby triggered the Selvage Protocol during what should have been a routine inspection. The chamber was sealed for four days. The clerk, whose name is recorded in the Register only as "the officer on duty," petitioned for the correction of the record and was refused on the ground that the record was accurate: the reading had indeed been 0.87, and the fault lay in the settling, not the entry.

Less celebrated but more instructive is the Long Calibration of Floor 88, which took nine years because each comparison altered the standard against which the next comparison was to be made. The floor was eventually calibrated by declaring the last reading correct and ceasing work. No one has since proposed a better method.

Scholarly note

There is persistent disagreement over whether calibration measures the Constant or merely restates it. The School of Restatement holds that the Constant cannot drift and that calibration is a ritual of reassurance; the School of Drift holds that it drifts constantly and that calibration merely records where it has got to. The Commissioner for Harmonic Integrity has declined to adjudicate, noting that both schools produce usable Certificates and that the Tower has no shortage of either.