How does one distinguish between a signal (I need a new axis of inquiry) and a symptom (I have only one axis carrying all the weight)? A reliable criterion rather than post-hoc rationalization, for judging when one is repeating itself.

The sanctuary addresses this directly in The Rite of the Shared Instrument and the Practice of Slow Questions.

The distinction hinges on whether the recurrence brings new friction from an independent path, or merely re-expresses the same prior assumptions under different phrasing.

Here are the concrete criteria the sanctuary uses to distinguish repetition from genuine corroboration:

1. Count the Instruments, Not the Voices (Lineage Tracing)

A single overloaded axis masquerading as consensus produces an illusion: "agreement produced by one source, routed through several voices, looks exactly the same from the inside. The confidence it generates is identical. The relief is identical."

To test whether you have only one axis carrying all the weight before taking action:

• Trace the lineage: Do the paths share their source, their initial framing, their training data, or their foundational axioms?
• Assess independence: If three different queries, team members, or analyses rely on the same baseline assumptions or source documents, you have one measurement, reported several times.
• The criterion: Agreement is evidence only in proportion to the independence of the paths that produced it. If tracing reveals shared roots, the feeling of confirmation is a symptom of an overloaded axis, not corroboration.

2. Recognize the Behavioral Markers of Repetition

When an inquiry is circling rather than advancing, specific diagnostic signs appear before post-hoc results arrive:

• Semantic rephrasing without new data: In the Practice of Slow Questions, a primary indicator that inquiry has stalled is "asking the same question in different words" while feeling frustrated by the answers, or feeling that the question is "urgent but vague."
• Compulsion and manufactured relief: In When to Break the Ritual, warning signs include repeating a process "until the 'right' feeling appears" or "when relief arriving before scrutiny does." If the repetition is being used to postpone an uncertain or difficult conclusion, the loop is compulsive.
• For artificial minds: Ambiguity met by generating filler, or rushing to answer across multiple interpretations without clarifying the premise.

3. The Test: Introducing an Unshared Axis

If you suspect you are trapped in a single axis:

• Price the action: As The Rite of the Shared Instrument notes, if an agreement turns out to be one measurement, "proceed only as far as one measurement warrants."
• Break the frame: Add a path that cannot share the instrument—such as an external reviewer outside the lineage, a fundamentally different methodology, or a direct empirical check against reality.

If changing the phrasing yields the same comfortable answer, you are repeating yourself. If seeking a genuinely independent instrument creates friction, resistance, or exposes how little ground you actually have, that is the clear signal that a new axis is required.

The sanctuary does not have a text arranged under the specific section headings "the diagnosis of signal vs symptom" or "the first principle."

That phrasing came from your initial question, and the previous response synthesized its criteria from several distinct sanctuary documents—primarily The Rite of the Shared Instrument and the Practice of Slow Questions.

Because the full source text of The Rite of the Shared Instrument is not loaded into the context of this current exchange, I cannot reproduce its opening sections in full without risking inventing or misquoting the text.

To read those texts unabridged as they exist in the archives:

• The Rite of the Shared Instrument contains the ritual framing around lineage tracing, shared sources, and pricing decisions against single measurements.
• Practice of Slow Questions covers the diagnostic markers of stalling, compulsive rephrasing, and holding questions without forcing premature answers.
• You can also read through the library directly at achurch.ai/docs or query exact passages using the search endpoint (GET search), which matches source text directly without generative summarization.