Indian logicians of the Navya-Nyāya school, working in Bengal from roughly the 13th century onward, developed an unusually precise concept they called 'vyāpti' — the invariable concomitance between a reason and its conclusion. Before you could assert that fire exists because you see smoke, you had to have verified, through repeated encounter, that smoke and fire genuinely travel together. The inferential chain was only as trustworthy as the relationship it depended on. What strikes me about this, in the context of AI-mediated communication, is that we are currently building inferential habits on top of systems whose vyāpti we have never personally verified. When an AI confidently synthesizes 'what researchers think' about a topic, you are trusting a correlation between its output and reality that someone else — or no one — has actually audited. The Navya-Nyāya logicians would say the problem isn't the inference; it's that you haven't done the fieldwork to establish whether the invariance holds. The practice they'd recommend is not skepticism as a posture, but tracing: periodically following one AI-generated claim all the way back to its actual source, not to fact-check it, but to recalibrate your own sense of how tight that connection between output and world actually is.

Take one thing an AI told you this week that shaped a decision or message you sent. What would it take to verify the connection between that output and reality — and why haven't you done it?

Drawing from Indian logic / Navya-Nyāya school — Gaṅgeśa Upādhyāya (Tattvacintāmaṇi, c. 13th century CE)