Narratively Complete

The term Narratively Complete designates a holding whose account can be told from beginning to end without recourse to material held elsewhere in the Tower. The designation is formal, issued by the Department of Procedural Integrity, and carries no implication of quality, accuracy, or truth; a complete narrative may be false in every particular and remains, for purposes of the register, complete. What the designation certifies is closure: that the holding's premise, complication, and resolution are all present on the same shelf, and that no reader is obliged to ascend to another floor to learn how matters ended.

By long convention the standard of completeness is set not by the whole but by an admitted hole. The Third Concordance holds that "a blueprint deposited at Floor 12,891 is understood to be complete in its intention and therefore complete" — a formula extended to all holdings by Concordance 1022, which the archivists of the middle floors regard as the founding licence for calling anything finished. The floor's readers have never been entirely satisfied with this. Completeness by intention, they note, is also how a great many things in the Tower were lost.

Procedure

Assessment is conducted by two reviewers working independently in adjoining cells, neither permitted to see the other's findings until both have sealed their returns. A holding is judged complete when the two accounts agree in sequence and differ in detail, on the reasoning that an ending independently reported twice and differently recounted is an ending that exists. Where the reviewers' accounts agree in every particular, the holding is flagged as suspect and passed to a third reviewer, who reads neither account and is asked only whether anything is missing. The third reviewer's answer is recorded but is not, in practice, consulted.

Holdings found incomplete pass to the biannual review of Narratively Incomplete material, where they are held until the omission has either been supplied or formally repudiated. The distinction between the two outcomes has never been settled in writing, and holdings have been known to wait decades for a determination that would not differ greatly either way.

The Kessendal Objection

The principal challenge to the doctrine came from Verrin Kessendal, a junior assessor on Floor 40,112, who observed in 1,904 that an account of a holding which includes an account of its own assessment must include an account of that account, and so on without end. Under the intention standard, Kessendal argued, every holding was complete because the assessor could always declare the regress closed by fiat; under any stricter standard, none was. The paper was answered by Overseer Marl Andwine in a three-page memorandum now bound with the original and never read separately, which conceded the regress and denied its relevance. Andwine's formulation — "the Tower does not require the whole story, only the shelf on which it is presumed to rest" — is inscribed above the assessment cells and is reproduced in most subsequent recensions of the procedure.

Kessendal was reassigned to recorder of incomplete returns and served in that post for thirty-one years. The post was abolished four years after their death, pronounced redundant, and quietly reconstituted the following winter under another name.

Contested cases

Two holdings dominate the literature. The Complete Folio of the Avian Paradox (Catalog #4C-778), in Annex 4 C, was declared Narratively Complete in 1,922 and has been challenged at every subsequent review; its dissenters hold that the folio's resolution depends on a clause present only in the annexed notes, which are filed on a different shelf. The Department has declined to rule, on the ground that ruling would itself complete the record.

By contrast, The Complete Map of the Tower, which occupies all of Floor 714, has never been designated complete, despite being the most thoroughly cross-referenced holding in the archive. Its premise and complication are beyond dispute; its resolution cannot be located, for reasons the Map Room staff describe as structural and outside the Department's competence. The Department accepts this explanation without prejudice and reviews the Map annually.

Interaction with other designations

Narratively Complete must be distinguished from Narratively Stable, a separate certificate used chiefly in the Suppression Protocols. A complete holding may be highly unstable — a finished story can be a dangerous one — and the two designations have been known to conflict directly. Floor 40,119 was quarantined for eleven weeks in the winter of 1,987 following the discovery that a holding declared complete in 1,941 was still generating episodes at a rate the shelving could not absorb; the quarantine was lifted once the episodes were found to be returning to their own opening scene, which the Department ruled was not continuation but repetition, and therefore not evidence of incompleteness. The holding remains complete.

The designation also bears on the infrequent use of Narrative Fertilizer, which the Department audits afterward precisely because a fertilized holding may acquire material that makes it complete, or destroys the closure that made it so. Both outcomes are recorded in an annex to the audit, which is not complete.

Register of completeness

Completeness is entered in the holding's front matter and in no more than three other places. The register of declared-complete holdings is maintained in six copies, held as follows:

• the master register, Floor 40,112, in the assessors' strongroom;
• a working copy, Floor 40,112, in the corridor;
• a copy held at Spiral 56, sealed, unread;
• a copy surrendered to the Complete Index of the Tower, pending volume 9,004;
• a copy held by the Chambre des Angles Obscurs, for reasons never minuted;
• a copy presumed lost.

The six copies do not agree. A discrepancy noted in 1,966 between the first and second copies — a holding complete in the strongroom and incomplete in the corridor twelve feet away — was referred to the Department, which ruled that the corridor copy was a convenience and the strongroom copy was the record, and that readers should not be encouraged to consult both.