Preregistration of the Lateral-Difference Identity Trial

Abstract

Conditions fixed in advance for an identity trial on a determination of the corpus's own working doctrine. The candidate distinguishes this own obtaining from another own obtaining within the same who-region where both obtain the same repeatable manner. It is compared against three published articulations of the same corpus. The candidate is worded so as to fit none of the possible answers, and the trial may return identity, division, failed identity, or suspension; novelty is established only by failed identity against all three references. Where no common determination-space can be stated, the finding is suspension and not failed identity. Deposited before any classification is performed, and before the volume that has received the candidate begins drafting. Version 1.1 enters printed amendment 1 under §9: the three reference articulations and the candidate are fixed to version identifiers rather than concept identifiers, since the references are reconstructed at page level and a concept identifier designates whatever version is current at the time of reading. No reference was replaced or added. The candidate, determination-space, controls, and verdict conditions are unchanged. Entered before any classification was performed. Version 1.2 enters printed amendment 2 at §8-ter under §9: an office-identity test, required before any identity output may be returned where the candidate and a reference share substantial content and differ in office or operative role. The test has two parts, both printed whatever they return — the question each determination settles, stated in its own source’s terms, and the conditions under which each would be false — with suspension returned where the two parts disagree. The amendment was occasioned by a defect in the instrument as first drafted, which offered four identity outputs and supplied no test able to distinguish two determinations built from shared materials from one determination appearing under two offices. The candidate, the references, and the determination-space are unchanged; what is added is a test and not a preferred answer, and the cost of each possible outcome is fixed in advance. This version is deposited before the comparison is re-run. The first classification is not withdrawn and remains public as returned. Version 1.3 deposits both classifications performed under this instrument. The first is published as returned and is not withdrawn; it was a correct answer to a question the instrument had put defectively, offering four identity outputs and supplying no test able to distinguish two determinations built from shared materials from one determination appearing under two offices. The second was conducted blind in a fresh session under the amended instrument, with the three cold reconstructions reused unedited, and it governs. Both returned failed identity against all three fixed references. The governing classification locates the novelty in the outward co-disclosure of other HOWs rather than in the inward reciprocal difference, and the preregistered candidate is narrower than what was tested; the verdict is accordingly bounded to the conjunction of the lateral difference with its outward articulation. The preregistration file in this version is a re-rendering of the file deposited at version 1.2. The markdown is byte-identical; the PDF differs in bytes and not in content, its extracted text, page count and word count having been verified identical before publication. Version 1.2 carries the original rendering and remains available for byte-level comparison. Version 1.4 enters printed amendment 3 (§8-quater), extending the reference set from three references to seven by a screening of every open deposit created after Ethical Ontology, Volume I and before the candidate. The screening rule, every exclusion with its reason, and the fixed DOIs of the added references are printed in the preregistration. It also deposits the cold-session instrument under which the added comparisons are run, so that the conditions claimed for those sessions can be checked rather than taken on assurance. No comparison under this amendment has been performed and no verdict is recorded here; the candidate, the determination-space, the controls and the verdict conditions are unchanged, and the three completed comparisons stand as returned. Version 1.5 closes this trial. It withdraws nothing: every verdict stands as returned. It deposits the outputs not previously public — the classification report of the amendment-3 run, the compression finding, and all seven frozen reconstructions on which the verdicts were reached — together with a closure note printing the instrument’s own defects: the candidate as preregistered was understated, office-neutrality having been purchased by subtraction; the comparability rule converted two printed self-limitations into suspensions rather than findings; and the screening predicate did not match the candidate. Failed identity was returned against four references and three were left unadjudicated. A successor trial is preregistered separately.