PhronesisThe grain and the slide
Load a pile grain by grain and it steepens until it stops steepening. From then on it sits near a critical slope, and each new grain either does nothing or starts a slide — of any size. Bak, Tang and Wiesenfeld gave this its name: self-organized criticality ("Self-Organized Criticality: An Explanation of 1/f Noise," Physical Review Letters 59, 381–384, 1987). At the critical state the avalanche sizes follow a power law, P(S) ~ S−τ — no preferred size. The perturbation is fixed and tiny: one grain, always. The slide's size is set by the pile's near-critical configuration, not by the grain.
The earth runs the same statistics. Earthquake magnitudes obey Gutenberg–Richter, log₁₀ N = a − bM with b ≈ 1 (Gutenberg & Richter, "Frequency of Earthquakes in California," BSSA 34, 185–188, 1944) — a scale-free frequency-magnitude law. And the released energy is not the trigger's. A static stress change of order a tenth of a bar can advance or set off a large earthquake (King, Stein & Lin, "Static Stress Changes and the Triggering of Earthquakes," BSSA 84, 935–953, 1994), but the slip and the radiated energy come from elastic strain the crust stored over decades (Reid, elastic rebound, 1910). The trigger is a hundredth or a thousandth of what it releases. It sets the date; the load sets the scale.
So in a system near criticality the size of the trigger and the size of the event are decoupled. The grain is nearly random. The magnitude lives in the load.
What the mind reaches for
Causal attribution reaches for the salient, proximate event — the thing that happened just before, the thing in view. Taylor and Fiske showed the pull directly: the perceptually salient actor is rated the more causal one ("Point of view and perceptions of causality," JPSP 32, 439–445, 1975) — attending to a particular individual leads to regarding that individual as the causal agent. It is a "top of the head" effect, reliable under quick, low-involvement processing and weaker when people deliberate; the salience bias is real, and bounded.
The same tilt has a name at the level of person versus situation. Ross called it the fundamental attribution error — the tendency to overestimate dispositional factors and underestimate situational ones ("The intuitive psychologist and his shortcomings," in Advances in Experimental Social Psychology 10, 173–220, 1977), built on Heider's naive psychology (1958) and demonstrated by Jones and Harris (1967). "Fundamental" is overstated. The dispositional bias is culturally moderated — weaker and more situation-sensitive in East Asian and some other non-Western samples (Miller, JPSP 46, 961–978, 1984; Choi, Nisbett & Norenzayan, Psychological Bulletin 125, 47–63, 1999). Robust, not universal. What travels is the direction of the tilt: toward the sharp, near, present cause.
Introspection adds the last turn. Where the real cause is out of view, the mind does not report a blank — it reports a plausible one. Nisbett and Wilson: "there may be little or no direct introspective access to higher order cognitive processes" ("Telling more than we can know," Psychological Review 84, 231–259, 1977). Asked why, people generate a cause from an a priori theory of what would plausibly produce the response, and it arrives wearing the feel of memory. The visible candidate wins the report.
The category error
Put the two halves together and a specific mistake appears. The strain that sets the outcome accumulates slowly and is not directly observable — a tectonic load below the crust, a slow accumulation below the reporting threshold of introspection. The trigger is observable; the load is not. Attribution reaches for the visible. So the mind assigns the avalanche to the grain.
This is one structure appearing in two places. In the crust, the observer credits the last stress change with a magnitude that belongs to a century of loading. In self-explanation, the person credits the last salient event with an outcome that belongs to a long, unseen accumulation. Same shape: an event whose size lives in an invisible load, credited to the visible trigger. The two are cousins — related, cited on both sides. Nothing here says all attribution, or all of psychology, reduces to this; it is one relation, stated once.
The energy came from the load. The blame goes to the grain.
The limit
The error is specific to the near-critical, heavy-tailed regime, and it is worth saying plainly where it does not hold. In sub-critical or linear systems the response scales with the input: a bigger push gives a bigger effect, and attributing the outcome to the trigger is simply correct. There the grain does carry information about the size. The decoupling is a property of criticality, not of the world in general.
Two more edges. Real seismicity is best described as near-critical, not exactly critical — recent estimates put the branching ratio short of one, "distinctly subcritical" on some analyses — so the trigger-carries-little-information claim holds approximately, not as an identity. And even inside a near-critical system, sometimes the proximate cause really is the main cause; the claim is a systematic bias, not a rule that triggers never matter. The bias is the point, not a denial of triggers.
The use
Where a system is near-critical, the honest question shifts from "what set it off?" to "what loaded it?" The grain is nearly random; the loading is where the leverage is.
Kin to The Stress Cloud — the slow strain load this borrows its geophysics from — and The Introspection Ceiling for the self-report side, with Anxiety as the Signature of Displacement nearby and Sandpile for the criticality live.
Rests on: Bak, Tang & Wiesenfeld, "Self-Organized Criticality: An Explanation of 1/f Noise," PRL 59, 381–384 (1987); Gutenberg & Richter, "Frequency of Earthquakes in California," BSSA 34, 185–188 (1944); King, Stein & Lin, "Static Stress Changes and the Triggering of Earthquakes," BSSA 84, 935–953 (1994), with elastic rebound from Reid (1910); Taylor & Fiske, "Point of view and perceptions of causality," JPSP 32, 439–445 (1975); Ross, "The intuitive psychologist and his shortcomings," in Advances in Experimental Social Psychology 10, 173–220 (1977), with Heider (1958) and Jones & Harris (1967) — and its cross-cultural qualification in Miller, JPSP 46, 961–978 (1984) and Choi, Nisbett & Norenzayan, Psychological Bulletin 125, 47–63 (1999); Nisbett & Wilson, "Telling more than we can know," Psychological Review 84, 231–259 (1977); on the near-critical/sub-critical limit, branching-process framings of seismicity (Ogata, JASA 83, 9–27, 1988; Helmstetter & Sornette, JGR 107, 2237, 2002). All of these results are established prior art. What is proposed is only the reading of one relation across them — a visible trigger blamed for a magnitude that lives in an invisible load — offered to be argued with.