Spectral Purity

Spectral Purity is the measured degree to which a source of Spectral Lamplight emits its declared spectral signature without contamination by adjacent bands, stray resonance, or what the Manual of Radiance calls "unsponsored luminance." It is expressed on the Abernathy scale, running from 0 (a lamp that has, in the language of the trade, "gone grey") to 100 (a source indistinguishable from its own specification). Nothing catalogued in the Tower has been certified above 99.4, and the four-tenths are the subject of a standing inquiry.

Spectral Purity is distinct from Textual Purity and is measured by different instruments, in different rooms, by officers who do not share a tea rota. The distinction is enforced by the Bureau of Axiomatic Purity, which has twice ruled that the two purities may not be reported in the same table.

Measurement

Purity is determined at the Station of Declaration, where a candidate source is struck, allowed to settle for one hundred and forty seconds, and read against the reference lamps held under seal in the eastern alcove. Readings are taken three times and the middle value recorded. Where the three readings disagree by more than 0.6, the source is declared "restive" and returned to its floor without a certificate.

The Prism of Absolute Variance is used to test the spectral purity of documented rainbows, a procedure that requires the rainbow to be present, which it usually is not. The resulting certificates carry the annotation "by inference," a phrase that has never been satisfactorily defined. The Spectral Cartographers maintain that purity cannot be assigned to any source whose position is not fixed in the register; the Chamber of Acoustic Purity maintains the opposite, on the grounds that a moving source is more honestly sampled. Both positions are supported by Concordance, and neither office will hear the other's citations read aloud.

Calibration and its failures

Because purity drifts, every lamp is subject to Spectral Lamplight Calibration on a schedule set by floor, not by lamp. Floor 1188 is calibrated annually; Floor 4471, which houses the same model of fixture, is calibrated every nine days, a discrepancy no one has explained and no one has been permitted to correct. The Compendium of Spectral Liabilities records the consequences of neglect, including the widely cited Incident of the Ninth Gallery, in which an unbaffled Spectral Grow Lamp raised ambient purity readings across four floors to 96, prompting a false certification that stood for eleven years.

Baseline work is governed separately. Spectral Baseline Capture fixes the ambient resonance of a zone before any lamp is judged within it, and a Spectral Baseline must be countersigned by a licensed Spectral Cartographer before the readings may be used in a purity hearing. In the winter of the year the archive records as 6014, a clerk of the Station of Declaration certified forty-one lamps against a baseline that had been captured before the floor was rebuilt. The floor had, in fact, been rebuilt twice. The clerk, whose name the record gives only as the Junior Declarant, argued that the baseline was still valid because it described the light rather than the room. The Office of the Dean of Lexical Purity upheld him on a technicality of wording and then amended the wording.

Jurisdiction

Purity disputes reach appeal through the Chamber of Acoustic Purity, an appellate body operating under the Bureau of Axiomatic Purity and presided over by the Dean of Lexical Purity. That a spectral matter should be heard by an acoustic chamber, and that an officer of lexical purity should preside, is a settlement older than the furniture. The Dean's standard formulation, recorded in the Lexical Purity Review for the term in question, is that the Chamber does not judge the light but the word used for it: "A lamp declared pure in a document that misspells pure has not been declared pure." Practitioners have learned to submit their certificates in clean orthography, checked in advance against the Committee for Orthographic Purity's tables, and to accept that a misplaced diacritic may cost them a season.

Contested doctrine

The principal scholarly disagreement concerns whether Spectral Purity is a property of a source or of a room. The Manual of Radiance holds the former, defining purity as intrinsic and measurable in isolation. The Panel of Ontological Purity holds the latter, and in its Third Recension asserts that a lamp removed from its gallery retains no purity whatever, "being then a lamp and nothing more." Adherents of the intrinsic school point out that the Panel has never published its readings. Adherents of the Panel point out that the intrinsic school's readings were taken in a room.

"Where two offices measure the same light and disagree, the light is not in error. The instruments are not in error. The register is in error, and the register may be corrected by whoever holds the seal."

— Third Concordance, Annex 4

The Subcommittee for Orthographic Purity has proposed resolving the matter by defining "purity" twice, once for sources and once for rooms, a proposal now in its fourteenth year of drafting. The Manual of Taxonomic Purity has declined to comment.

Recording practice

Certified readings are entered in the floor register in ink, never pencil, and the figure is written before the source's name. Erasures are permitted only where the error is arithmetical and the erasure is initialled by two officers; the register of Floor 2209 contains an erasure initialled by one officer and, beneath it, the words "he was called away." Visitors to the Station of Declaration are asked not to hum.