Who Holds the Test

When many minds trade spiritual insights, the network multiplies variants faster than it can check them; but the sharp problem is not that there is no check. Conduct is a real and fairly cheap check: does an insight reduce craving, reactivity, avoidance, and relational harm over weeks. The problem is who holds that check. A teacher who confirms your realization, a group that affirms your progress, and your own felt relief all reward the variant instead of testing it against anything outside the network. In technology a boat that does not float sinks regardless of who taught it; the kill signal is independent of the teacher. Spiritual filters built on prestige, belonging, or pleasant calm are not independent in that way, so recombination can drift toward what feels adaptive and what the community rewards while moving away from reality contact. An insight earns trust only when part of its evidence is held by someone or something that gains nothing from your believing it.

Most treatments of the spiritual collective brain ask whether a selection test exists or whether it is ever run. The distinct move here is to separate two things that are usually fused: whether a filter is applied, and whether the filter is independent of the network that produced the variant. Even a real conduct test can select for locally adaptive but truth-decoupled forms when the people applying it gain from the result. The Kalama Sutta is the closest prior argument and is older than any of this; it already replaced prestige bias with a consequence test, but it left that test inside the sangha that judges it. Cultural evolution, meanwhile, celebrates the prestige bias that this one domain must distrust. The contribution is narrow: name filter-independence as the variable, and predict runaway adaptive-but-untrue drift specifically where the filter is captured by the network's own incentives.

The independence requirement may import a modern adversarial epistemology that living traditions would reject. Devotional and lineage paths hold that only a trained insider can recognize realization and that a stranger cannot judge depth; demanding a non-gaining outside observer could corrode the very trust that lets skill transmit, and could pathologize healthy belonging in already lonely people. The category is also hard to operationalize: a clinician gains professionally, a skeptical friend gains by being right, a stranger lineage gains by contrast. If every observer is embedded, the model collapses to ordinary advice to stay a little skeptical, which is not new. A further anomaly: some traditions preserved insight for centuries using almost entirely network-internal confirmation, which suggests internal filters can protect transmission fidelity even if they do not test truth. The claim should be weakened if network-internal-only confirmation proves as stable and harmless as confirmation that includes an independent signal.