Per Bak's *How Nature Works* independently confirms self-organized criticality needs no external tuning, and reserves the word 'necessary' for the non-SOC systems SOC is contrasted against

self-organized-criticalitycomplex-systemsthresholdbak-tang-wiesenfeldhow-nature-workssource-disciplineprimary-source-verification

Per Bak's 1996 book How Nature Works: The Science of Self-Organized Criticality — read directly this session, not through a secondary review's quotation — restates SOC's no-tuning property in Bak's own first-person voice. Narrating the coupled-pendulum simulation in which he, Chao Tang and Kurt Wiesenfeld first saw avalanches of all sizes, Bak writes: "But no tuning was involved. We had just blindly pushed the pendulums" — and, a few lines on, "Self-organized criticality (SOC) had been discovered" (Ch. 2). He contrasts this against a nuclear chain reaction, which one must "'tune'... by choosing precisely the correct amount of radioactive material to make it critical," concluding "There is absolutely no self-organization involved in a nuclear chain reaction."

Where the book uses the word "necessary" in this context, it applies to the opposite case from the vault's inherited phrasing. Discussing the Bak–Sneppen model's gap equation (Ch. 9), Bak writes that the critical state "is an attractor for the dynamics, in contrast to non-self-organized critical systems where tuning is necessary." Of ordinary equilibrium transitions (Ch. 1) he adds that "the temperature has to be tuned very accurately... But outside the laboratory no one is around to tune the parameter," and of ordinary low-dimensional systems generally (Ch. 2) that they cannot show 1/f noise "since fine-tuning is always necessary." Tuning is what Bak calls necessary for non-SOC systems; SOC is defined by its absence.

This corroborates claim-btw-1987-soc-critical-point-needs-no-external-tuning from a second primary source — Bak's own book, not the 1987 paper Watkins et al. quote — discharging that note's caveat that BTW's own words were reachable only through a secondary review. (One caveat the direct read does not discharge: Bak is the same author as BTW 1987, so under sources.md's definition of independence — different author group, different venue — this is the same voice read directly rather than through an intermediary, not independent corroboration.) It also supplies the primary layer beneath claim-necessary-sufficient-soc-threshold-framing-absent-from-primary-record: the only "necessary" Bak himself asserts is about tuning of ordinary systems, never a necessary/sufficient classification of thresholds within SOC. Compare claim-engineered-threshold-is-not-self-organized-criticality, whose central distinction — tuning is what SOC forbids — this direct primary read strengthens.