In 1925, working through the mathematics that became wave mechanics, Erwin Schrödinger kept notebooks on Vedanta. He was not decorating his physics with philosophy. He had recognized that the oldest texts he knew and the newest equations he was writing described the same structure of observation, and he had no way to demonstrate it. The recognition stayed private, an accident of one man’s reading, for a lifetime.
This keeps happening. The observer problem in quantum mechanics, the dissolution of the self in contemplative phenomenology, the constructed self in neuroscience: similar structures, mutually unintelligible vocabularies. When 195 working scientists were asked to name the deepest bottleneck in science, the leading answer was coordination failure, the absence of any mechanism for doing collectively what no single lab can do alone. We build one precise slice of that mechanism, a neglected bottleneck: the ability to recognize when separated fields already hold the same structure.
Two communities need it from opposite sides. Researchers spend careers crossing vocabulary walls by hand; the landmark adversarial comparison of two consciousness theories took a consortium seven years, once. And a generation of seekers and creators, sensing real connections between traditions, is stranded between credulous overclaim and dismissive debunking, with no honest middle. Both are missing the same instrument.
The Consilience is that instrument: an engine that strips vocabulary to structure, matches across the walls, and grades every correspondence with a judge that is never a language model. It takes contemplative sources seriously enough to formalize them and science seriously enough to let it say no. The grade is the product. The rejections are published. The strongest results carry a citation and a name.
The project is founded by Anuja Khatri, a physics graduate and AI product builder, and is fiscally sponsored by the Institute for Advanced Consciousness Studies as a nonprofit. Its method encodes a division of labor she holds personally: conviction supplies the questions, and a merciless grader supplies the answers. The founder holds the vision. The grader holds the line.
Multiple typed guard rails target the known routes by which false correspondences fool both humans and models, and the grader is promoted only when it rejects a benchmark of false correspondences built to be seductive. The full guard rail set publishes with the open source release.