Xenosophy: preregistrations of the next phase

Preregistration of the test of generation announced in "Xenosophy: Asking About Contact When No Common Frame Is Guaranteed" (version 0.1, DOI 10.5281/zenodo.23816129). The test asks whether the approach's procedures of variation, applied to its distinctions, lead an AI analyst to hypotheses about a published study's object that AI judges rate as testable and absent from the field more often than hypotheses given under a plain instruction to be bold, or under a generic list of heuristics. A further arm tests the practice of the author's laboratory as a whole, reported as a black box, against the plain instruction. The cases are drawn from listed pools by euro reference rates published after this record is made public. The document fixes the design, the rules of scoring and what counts against the approach before any case is drawn; it may add conditions to those of the announcement, never weaken them. Use of AI. The design and every decision in it are the author's. The document was drafted under the author's decisions by Claude Opus 5.5 (Anthropic), which also compiled the pools, compiles one of the lists of the field record and takes part in the tally; it is not independent of the test. The test uses AI systems as analysts (ChatGPT; Claude Opus 5.5), judges (GLM, Qwen; Gemini in reserve), the source of a control and a parallel tallier (DeepSeek), a compiler of the field record (Perplexity), and a tallier used only if the tallies diverge (Kimi). No AI system is an author or contributor.