Standard Issue Quill Concordance Model

The Standard Issue Quill Concordance Model (catalogue mark SIQ-CM, eleventh recension) is the Tower's benchmark apparatus for determining whether a given quill is in concordance with itself. The question is less abstract than it sounds. A quill may write, and write legibly, and still be out of concordance — a condition first isolated in the winter of the year the catalogue records as 4,118, when a scriptorium on the fourteenth gallery of Annulus III produced nine hundred pages of irreproachable minutes that no subsequent reader could bring themselves to finish, though none could say why.

The Model is not a device. It is a procedure, a jig, and a schedule, administered jointly by the Synod of the Quill and the Office of Instrumental Concordance, and it is against the Model that every Ceremonial Quill in Tower service is periodically assessed — including, by long precedent, the Sanguine Quill, whose authorization protocol is treated as the model for most subsequent designations and which therefore sits at the head of the Model's reference rack whether or not it is being tested.

Composition of the apparatus

The Model comprises a reference rack of eleven quills, a graduated rest cut to the Instrument Interface standard, and a set of comparison plates struck in the workshops beneath the Vellum Galleries. Testing proceeds by substitution: the candidate quill is fitted to the rest, a plate is drawn, and the resulting stroke is laid beside the corresponding stroke of each reference quill in turn. The assessor records the nearest match and the margin of error, in that order. A margin under three parts in one hundred is returned as concordant; a margin over eleven is returned as discordant and the quill is withdrawn to the annealing shelves on the ninth gallery pending review.

The eleven reference quills are, in rack order:

• the Sanguine Quill, unmodified, retained for authorization comparison only;
• two Iron-Gall Quills of Standard Grade 4, one worn and one fresh, the worn specimen being the older by four hundred and six years;
• the Kaelen Quill, in its post-Schism configuration;
• the Quill of Concordance itself, which is never drawn from and is present only so that the rack may be said to be complete;
• and six further specimens whose designations are withheld under Concordance 6,404-γ.

The Schism of the seventh gallery

For most of its history the Model held that concordance was a property of the stroke. The Thorne-Lateralist Progression Model argued otherwise, proposing that concordance inhered in the quill's advance across the page and could not be judged from any single mark. The dispute was referred to the Office of Precedence and there sat for sixty-one years, during which both parties continued to certify quills under incompatible rules and the scriptoria simply ordered replacements twice.

Matters turned on a single sheet. In the seventh gallery, in the winter of the year the catalogue records as 4,301, an assessor named Clerk Halveth Orrun tested a candidate quill against the rack, found it concordant by a margin of two, and then — being thorough, and unpopular for it — drew a second plate. The second plate matched nothing on the rack. Orrun recorded the discrepancy in the margin and appended a note reading, in full: "The quill is concordant with itself in the first instance and with no other thing in the second. I do not know which of these the Model asks after."

"Where a quill is concordant at the first plate and discordant at the second, the assessor shall not average the two."

— Concordance 6,404-γ, Annex 9, clause (iii)

The clause was added in response to Orrun's note and has never been read the same way twice. Orrun was reassigned to the annealing shelves, where the record shows they remained for the rest of their service and where the margins of the shelf register, in that hand, grow steadily more careful and steadily shorter.

The Question of Average Concordance

The central scholarly disagreement concerns whether a quill may be certified on the strength of an averaged margin. The Office of Instrumental Concordance holds that it may, citing Concordance 2,218-β, which permits the averaging of "like measurements taken under like conditions," and noting that Annex 9 clause (iii) forbids averaging the two plates of a single test without forbidding the averaging of tests. The Synod of the Quill holds that it may not, citing the Third Concordance, which mandates Standard Librarian-issue Vellum for permanent records precisely because it does not vary between sheets, and asking how a standard can be struck on vellum that does not vary from an instrument that does.

Neither citation settles the matter, as the two Concordances were amended in opposite directions within the same decade and neither amendment notices the other. The practical result is that a quill certified by the Office may be rejected by a Synod assessor on the same afternoon, and frequently is. Clerks have learned to request certification from the Office in the morning and to submit to the Synod before noon.

Routine and known faults

The Model is run on the first and third days of each working period. It requires two assessors, one of whom must be silent for the duration; the silence is not ceremonial and its purpose is not recorded, but attempts to dispense with it have produced margins that later failed to reproduce. The comparison plates are struck on offcuts of Standard Librarian-issue Vellum and are stored flat, in the dark, and are not to be handled by anyone who has eaten that day.

The Model's most persistent fault is drift in the third reference position, where the worn Iron-Gall specimen loses a measurable fraction of its character with each drawing. The loss was first quantified in the Metrica of Minor Manifestations and has been budgeted for since; the worn quill is expected to fall out of concordance with the fresh one in the year the catalogue records as 4,412, at which point the rack must either be re-struck or reduced to ten positions. The Synod has ruled that it shall be reduced to ten, the Office has ruled that it shall be re-struck, and the question has been referred to the Office of Precedence, where by custom it will sit for sixty-one years.

The certification seal on each returned plate is applied in Wax Cylinder Standard Issue Grey, as required by the Model's own schedule, Schedule 4. Plates returned without the seal are treated as unstruck and must be drawn again. Assessors are reminded that the seal is applied to the plate and not to the quill.