Peirceiving Turings Dream - How Classification Grounds Become External, Machine-Readable, and Contestable

The key to connecting all the knowledge that already exists, in registers, on the web, in books, turns out to be surprisingly simple. The computer does the heavy work. The human gives it an occasional nudge, from the qualities only a human has. If you approach it the right way, in the right order, you do not need AI to reason. The reasoning itself runs deterministically from the classification ground. The warrant lies in the object, not in the model. This is a practitioner report describing a specific invariant — five properties that must hold simultaneously before a machine can derive rather than retrieve. Three entirely different applications illustrate how this works across inherited, produced, and reconstructed classification. A former member of parliament with twenty years of experience in government IT policy looked at the result and said: we have been looking for this for twenty years. And you did it in less than a day. How?