Contribution
A standard of evidence. It separates what would be seen if a system had alternative attractors from what would also be seen without them, and says which observations can discriminate.
Key insight
A sudden change, a bimodal distribution, or a state that persists are all what alternative stable states predict, and all are also predicted by simpler causes. What discriminates is reversibility. The system returns when conditions are restored past the original point, or a temporary push leaves a permanent change, and the return time is long enough to exclude mere slowness.
Assumptions
- The control factor is external to the system, or much slower than it.
- Recovery is not blocked by isolation (for example, no source of recolonists).
Key results
- Field indicators (Figure 3): jumps in time series, multimodal frequency distributions, dual relationships to a control factor. Each has alternative explanations.
- Experimental indicators (Figure 4): initial-state dependence (Elodea and Lemna), lasting shifts after temporary fish removal in shallow lakes, hysteresis in charophytes in Lake Veluwe.
- Traps: Lake Zwemlust stayed clear for six years after biomanipulation and then slipped back. Hysteresis-like loops appear whenever the response is slow relative to the driver.
- Positive feedbacks “only lead to alternative attractors if they are sufficiently strong”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Field patterns alone cannot establish alternative stable states | strong | each indicator shown to have alternative explanations |
| C2 | Experiments can come close to proof | moderate | worked examples, each with stated caveats |
| C3 | Slow erosion of resilience is neglected in favour of obvious triggers | moderate | argued, with examples |
| C4 | Positive feedback is necessary but not sufficient | strong | model results |
Concepts
- regime shift — an abrupt change to a contrasting regime, a word preferred to “stable state” because regimes fluctuate.
- hysteresis — different paths for forward and backward change in a control factor.
- resilience — here, the size of the basin of attraction.
Connections
Restates the theory of LIT-720 and works in Holling’s resilience framework (LIT-741).
Bearing on the record
Source of THEORY-136. It explains the reef case the record already holds (LIT-711; THEORY-128): an algal state that vanishes within 30 days of reopening the fish-exclusion cages fails the test, and the review’s own Caribbean narrative is about slow erosion of resilience. It also bears on THEORY-058, which says mental disorders are self-reinforcing networks that outlast their triggers. Outlasting a trigger is one of the hints this review calls inconclusive. No relation is declared.
For the owner’s essay (§5.4), it is the test protective lock-in must pass. The emergency regime would have to persist after the precipitating incident is resolved, or show hysteresis, and not simply be slow to unwind.
Limitations
Ecological examples only; spatial scale and environmental noise are noted as poorly explored.
Open questions
How to apply the experimental tests to systems too large or slow to manipulate, which is the problem for organizations as for oceans.