Causal Inversions

Definition

Causal Inversions are the class of recorded events in which the operative sequence of cause and effect within a Tower holding is found to have run in reverse, such that the effect precedes its cause in the archive without preceding it in the world. The distinction matters: an inversion is not a prediction and not a coincidence. It is a documented instance of an event having been caused by something that had not yet occurred, and having been caused by it correctly, in the proper order of the record.

Causal Inversions are administered under Causal Modeling and fall within the ordinary competence of the Directorate of Causal Forensics, whose Causal Verification Bureau has maintained the register of confirmed inversions without interruption since before the current recension of the Standard Topographical Index. They should not be confused with Negative Phenomena, which concern absence and void, nor with Temporal Gaps, which concern interruptions in sequence rather than reversals of it.

Relation to the Principle of Causal Conservation

Under the Principle of Causal Conservation, causal mass cannot be created or destroyed, only redistributed. An inversion is therefore not held to violate the Principle but to satisfy it in an inconvenient direction. The mass that ought to have been carried forward has instead been carried backward, and the account balances only once one accepts that it balanced before it was written.

The Retro-Causal Ripple Model was developed principally to render such accounts legible, and the Causal Mechanics Chapter reports jointly to the Directorate of Causal Forensics on all inversions exceeding a ripple depth of four. Where the mass cannot be traced to any originating event, the Directorate adjudicates the cause as "lost" and stamps the file accordingly; the stamp itself is, by longstanding convention, applied to the front of the folder.

The Ellery Recension and the Brightwater Dispute

The first inversion admitted to the register concerned the flooding of a lower stack corridor now catalogued as part of Floor 11,208. The water damage was recorded on the third of Avent, the burst conduit on the fifth. Both entries are in the same hand, that of Clerk-Adjutant Solveine Ellery, who declined to correct the sequence on the grounds that the entries were accurate. Her note, preserved in the register, reads: "The corridor was wet before the pipe failed. I have written what I saw. The pipe will be dealt with."

Ellery's refusal became the founding precedent. The Third Concordance holds that the record of an event is itself an event and cannot be entered out of order without becoming a different record (Third Concordance, Annex 4). The Fifth Concordance, written some decades later and evidently without sight of the relevant annex, holds the opposite under its disputed article 19, and the resulting standoff — the Brightwater Dispute, so called for a clerk whose name appears nowhere else in the archive — was not resolved until the Secondary Concordance was amended in 4,406 to permit both readings simultaneously, provided the file was sealed twice.

"Where a record precedes its occasion, the record shall be sealed in the presence of two Librarians, one of whom must not be told what is in it."

— Fifth Concordance, Article 19, as amended

The amendment is generally regarded as successful. Compliance has been near-total since.

Detection and the Topographical Overlap

Inversions are detected by routine comparison of the causal register against the Standard Topographical Index, which logs spatial inversions and triggers Procedure 88-C where the two coincide. Coincidence between spatial and causal inversion is uncommon but not rare, and the procedure exists because the second is frequently mistaken for the first by clerks working quickly.

An inversion leaves a persistent alteration in the causal profile of the affected object, recorded under Subclause 9.2: Causal Imprint Analysis. The presence of an imprint is diagnostic; its absence is not exculpatory, since an inversion thorough enough to leave no imprint has, by definition, inverted the analysis as well. Files of this kind are held open indefinitely and reviewed every eleventh year, on the understanding that the review may itself have already happened.

The Heresy of Antecedent Culpability

A persistent minority position holds that inversions are not anomalous but ordinary, and that the archive simply records them backward from the ordinary reader's convenience. Its most cited statement is that of Indexer Vault-Marrow, who argued that all causes are antecedent and that the Tower merely numbers them in whichever direction the handwriting permits.

The position drew sharp objection from Latent Culpability scholars, who observed that if causes may follow their effects, then culpability may be assigned before an act is committed, and the entire apparatus of the Theoretical Jurisprudence Guild would have to be rebuilt. The Guild has not rebuilt it. Under Directive 7-Δ, the doctrine is not prohibited — the Tower does not prohibit — but it may not be cited in any proceeding where the cause has already been found. Vault-Marrow's own admission file was subsequently found to have been completed prior to his application to the Guild, and the anomaly was logged, sealed once, and not discussed.

Practical Notes

Inversions are catalogued in the register in the order they were entered, not the order they occurred; researchers are reminded that the register is a physical object and that its pages turn one way. Reading rooms on the flagged corridor of Floor 11,208 are closed while the conduit is under inspection. Staff are reminded to leave wet umbrellas at the desk.