Preregistration and materials for a 2x2 experiment on conceptual armament and the rated originality of generated research ideas

About this documentThis is the reading edition of the deposit: the same nineteen files, gathered into one document in the order listed below. The individual files are deposited alongside it, and the wording, the figures and the judgements here are identical to them.This English edition is a translation. The Japanese edition is authoritative. The experiment was conducted in Japanese: every instruction was administered in Japanese, every generated output is Japanese, and all scoring was performed on the Japanese text. Where a file reproduces material that was actually handed to a model or a judge, the Japanese original is given alongside the translation. For the raw generation outputs and for the scattered scoring input, the translation is given on its own, and the Japanese file named at the head of each is the record.The typographic emphasis marks used in the working files (coloured symbols marking how firmly an instruction bound the experimenter) are not carried into this edition. They carried no content.The experiment held one generator constant (Claude Fable 5 MAX) and varied whether it was armed with ten concepts published after the recorded training cutoffs of the models used, and whether it could search the live web. Two judges, blind to conditions, scored the pooled output.Two records of error are included deliberately. One aggregation used the wrong key and was rebuilt; the superseded files are retained with an invalidation notice. One generation run touched a repository status file before generating and was repeated; the deviated run is retained and is not scored.