Nine theories · developed at length · not peer reviewed
The earlier volumes compressed. This one does the opposite: nine theories, each given a background, a mechanism, three predictions, a refutation and a way it could actually be tested. Fewer claims, longer arguments.
When a system's response time exceeds the timescale of its driver, the accumulated lag behaves as a debt: it must eventually be repaid at a rate higher than the one that created it, and the repayment is visible as an abrupt correction rather than a gradual adjustment.
Timescale lag · Systems across scalesThe failure rate of a maintained system is set not by its condition but by the ratio of its degradation rate to its repair capacity, so two systems with identical maintenance budgets can fail at very different rates if their degradation rates differ.
Repair capacity · Biology and engineeringAny system that measures or manages a process at a resolution coarser than the process's own scale pays a tax proportional to the ratio of the two, and the tax appears as either missed events or false alarms — never as a neutral loss of detail.
Granularity · Measurement and governanceThe frontier between efficiency and redundancy is convex rather than linear, which means systems cannot buy partial redundancy cheaply — and that collapse thresholds are set by the curvature rather than by the mean position on the frontier.
Trade-off geometry · Complex systemsSelection at one level of organisation generates variance that becomes the raw material for selection at the next, and the coupling is strong enough that the rate of adaptation at the higher level is set by the variance produced below rather than by selection pressure above.
Selection across levels · EvolutionIn any complex system, the fraction of outcome variance attributable to a specific cause has a ceiling set by the ratio of intervention rate to natural variability, and no amount of statistical sophistication raises it.
Causal inference · MethodThe expected effect size recovered by a replication is a predictable fraction of the originally published effect, set by the ratio of original sample size to replication sample size and by the original study's selection threshold — so shrinkage is forecastable rather than surprising.
Effect sizes · MethodThe energy cost of biological regulation has a floor set by the ratio of signalling molecules to noise, so tighter control is always more expensive and organisms should be found operating near, not far below, that floor.
Biological control · PhysiologyAn explanation is only valid within a band of scales set by the ratio of its mechanism's characteristic length or time to the phenomenon's, and applying it outside that band produces confident errors rather than obvious ones.
Theory choice · MethodAt one claim per paragraph, a theory is a slogan. These nine are given room to be wrong in specific ways — each has three predictions that could come out differently, a refutation that would end it, and a test that some existing body of data could already settle.
Entry page: garth.html. There is deliberately no index.html.
Publication records: source hashes · original README · original manifest · additive publication note