SP3 THEORY TESTING PROCESS: FROM A STILL IMAGE TO A TESTABLE HISTORY

Abstract

A static image of a galaxy collision shows where luminous matter is observed now. It does not directly show where that matter was previously located, which regions were traversed and subsequently vacated, which direction material moved, or when passage occurred. Those historical quantities are nevertheless potentially recoverable from the morphology and line-of-sight kinematics of a well-observed interacting-galaxy system. The scientific difficulty is avoiding the circular procedure of first examining an interesting present-day response—such as enhanced star formation—and then choosing a collision history that conveniently passes through it. We therefore developed a response-blind reconstruction and qualification framework, RMICE-1, before application to the interacting galaxy pair NGC 4676, the Mice. The method uses Joshua Barnes's Identikit reconstruction machinery but adds a series of synthetic truth-recovery tests, information-degradation controls, coordinate-closure tests, provenance hashes, historical-state reconstruction, and explicit prereveal acceptance gates. The qualification program progressed from basic software and coordinate validation through strict-blind recovery, five synthetic information regimes, controlled removal of line-of-sight velocity information, and finally reconstruction of a time-dependent Historical State Cube. The recovered history was compared with a synthetic history that had been frozen independently. In the highest-information synthetic reconstruction, historical passage was recovered with 100% sensitivity and a 0.424% false-positive rate; passage followed by vacancy was recovered with 98.6% sensitivity and a 0.398% false-positive rate. Controlled degradation demonstrated a marked loss of reconstruction quality when the available line-of-sight velocity anchors were reduced from four to two. The final conservative calibration uses three independently degraded but still informative reconstructions—D0, D1 and D2—as an equal-vote robustness ensemble. Requiring unanimous passage agreement yielded 86.5% sensitivity, 91.7% precision and a 0.175% false-positive rate. Unanimous passage→vacancy agreement yielded 83.6% sensitivity, 89.7% precision and a 0.174% false-positive rate. For directionally interpretable regions with ensemble directional robustness ≥0.90, median angular error was 2.95° and the 90th percentile was 10.83°. Timing proved less secure: the calibrated ensemble interval contained the true weighted passage time in 69.7% of tested cells. RMICE-1 therefore permits broad historical ordering and timing intervals but does not presently license a precise physical relaxation timescale. The central result is methodological rather than theoretical: a collision reconstruction must first demonstrate that it can recover hidden synthetic history before it is permitted to make historical claims about a real galaxy collision. The real Mice response data remain sealed until the reconstructed historical predictions themselves are frozen.