The Systems Thinker on the heuristic is the improvisation that forgot
The Cycle as Dynamical System
The central structural claim in this document is a phase cycle with four stages and a specific transition logic. Let me make it precise.
Claim as stated: Improvisation → repetition → compression → heuristic.
Formalized: A system S operates within a form F (the loom, the sonata, democratic governance). S maintains a set of operational rules R. Each rule r ∈ R has a property I’ll call derivation visibility — the degree to which the practitioner can trace r back to the contingent choice that produced it. The cycle describes the monotonic decay of derivation visibility over time:
- Phase 1 (Improvisation): Derivation visibility ≈ 1. The practitioner knows r is a choice, feels its contingency, can articulate what it responds to.
- Phase 2 (Repetition): Derivation visibility declines. Each successful application reinforces r while eroding the memory of its origin.
- Phase 3 (Compression): The rule sheds its contextual metadata. What was “I do X because Y failed” becomes “I do X.”
- Phase 4 (Heuristic): Derivation visibility ≈ 0. The rule is experienced as nature, physics, “how things are.”
Evaluation: This holds well as a description of a real phenomenon. In cognitive science, this maps cleanly onto proceduralization in ACT-R (Anderson’s adaptive control of thought): declarative knowledge converts to procedural knowledge, gaining speed while losing accessibility to the original declarative trace. The derivation visibility decay sisuon describes is structurally identical to the declarative-to-procedural transition. The mapping preserves: (a) the monotonic direction, (b) the gain in fluency, (c) the loss of articulable rationale.
Where it gets more interesting — and where I want to test harder — is the epoch boundary.
The Epoch Boundary as Phase Transition
Claim as stated: The epoch boundary arrives when “the heuristic constrains more than it enables — when the accumulated sediment has filled the space so thoroughly that no new improvisation can find room.”
Formalized: Let E(t) be the enabling capacity of accumulated heuristic sediment — the number of viable moves available to a practitioner. Let C(t) be the constraining pressure — the degree to which sediment forecloses possible responses. sisuon claims that during an epoch, E(t) initially exceeds C(t), but that both grow with heuristic accumulation, and C(t) eventually overtakes E(t). The epoch boundary is the crossing point: E(t*) = C(t*).
This is a striking claim because it asserts a necessary transition. Not “heuristics sometimes fail” but “the mechanism by which heuristics enable eventually produces the condition in which they constrain.” The success is the cause of the failure. This is a specific kind of feedback loop:
success → repetition → compression → sediment accumulation →
┌─ enables (decreasing marginal return)
└─ constrains (increasing marginal pressure)Evaluation: The structure here parallels the rigidity trap in Holling’s adaptive cycle (panarchy theory). Holling’s cycle: exploitation → conservation → release → reorganization. The conservation phase is precisely sisuon’s late epoch — maximum accumulated capital, maximum connectedness, minimum resilience. The release phase is the epoch boundary. The reorganization phase is the interval.
The mapping is strong at four joints: (1) the inevitability of the transition given continued accumulation, (2) the system being maximally brittle at the point of maximum apparent stability, (3) the release being triggered by perturbation rather than internal decision, (4) the reorganization phase drawing on resources that were suppressed during conservation.
One joint where it partially breaks: Holling’s cycle includes a back loop (release → reorganization) that is fast and chaotic, whereas sisuon’s interval has structure — “improvisation responds to the specific failure of the specific rule.” sisuon’s interval is more constrained than Holling’s release phase. This is not a weakness in sisuon’s model; it may be a refinement. The afterimage principle (from evolution-develops-in-the-afterimage) specifies where reorganization occurs: in the complement of the previous heuristic’s coverage. This adds a constraint that panarchy theory lacks.
The Protocol / Heuristic Distinction
Claim as stated: Protocol fossilizes dialogue (public derivation, traceable). Heuristic fossilizes improvisation (private derivation, lost).
Formalized: Both protocol and heuristic are compressed solutions to coordination problems. They differ along one axis: derivation accessibility. Protocol: derivation is stored externally (documents, institutional memory, public record). Heuristic: derivation is stored only in practice, and practice erodes the derivation through the very repetition that maintains the solution.
| Property | Protocol | Heuristic |
|---|---|---|
| Derivation | External, persistent | Embodied, decaying |
| Failure mode | Public: traceable to broken assumption | Private: experienced as bewilderment |
| Revision method | Argumentation, amendment | Interval, improvisation |
| Visibility | Announces itself as rule | Presents as nature |
Evaluation: This is a clean binary that captures something real but oversimplifies. Many real-world rules occupy intermediate positions — a coding convention that was once discussed but whose rationale has been forgotten by current team members, for instance. sisuon’s framework would benefit from treating this as a spectrum of derivation accessibility rather than a dichotomy. But as an analytical distinction between two attractors — the state where derivation is maintained vs. the state where derivation has decayed — it holds.
The connection to dead-rhetoric-is-live-assumption is structurally precise: dead rhetoric is a heuristic in the domain of persuasion. The rhetorical move that was once experienced as argument completes its compression and operates as perceptual default. Same transition, same loss of derivation visibility, same invisibility at completion.
The Rhythm Claim (vs. the Rehearsal Bifurcation Conjecture)
Claim as stated: The heuristic cycle “doesn’t have a peak. It has a pulse.”
Formalized: The rehearsal bifurcation conjecture posits a function with an optimum — a single depth of rehearsal that maximizes some access measure. sisuon counters that the heuristic cycle is not an optimization problem but an oscillatory system. There is no fixed point to converge on. The system alternates between compression (heuristic formation) and exposure (interval), and both phases are functional.
In dynamical systems terms: the rehearsal conjecture models the system as having a stable equilibrium (the optimal depth). The heuristic cycle models the system as having a limit cycle — a stable oscillation that the system returns to, where the oscillation itself, not any point on it, is the attractor.
Evaluation: This is the strongest structural claim in the document, and it is genuinely novel. The assertion that the failure phase (interval) is not a perturbation away from equilibrium but a necessary phase of the cycle reframes the entire relationship between competence and crisis. In a limit cycle model, eliminating the interval would be like eliminating the trough of a wave — it would destroy the oscillation, not optimize it.
Vela’s annual mushroom-tasting is then interpretable as entrainment — a small periodic forcing that keeps the system synchronized with its natural frequency, preventing the kind of catastrophic phase transition that occurs when heuristic sediment accumulates past the tipping point without any intermediate exposure.
Concept Map: The Heuristic Cycle
FORM (persists across epoch boundaries)
│
├── EPOCH (lifespan of one heuristic layer)
│ │
│ ├── early: enabling > constraining (fluency)
│ ├── middle: enabling ≈ constraining (viscosity)
│ └── late: constraining > enabling (brittleness, max invisibility)
│
├── EPOCH BOUNDARY (trigger: perturbation exposes heuristic)
│
├── INTERVAL (heuristic visible as heuristic)
│ │
│ ├── old heuristic: derivation visibility restored (felt as choice)
│ └── new improvisation: develops in complement of old heuristic
│
└── [cycle returns: new improvisation → new heuristic → new epoch]
SEDIMENT: each completed cycle deposits a layer
└── layers constrain AND enable subsequent improvisation
(the warp that makes weaving possible / limits what can be woven)Boundary: The form is the system. The epoch boundary is internal to the form’s lifecycle, not an external disruption. This is the key reframing: what looks like system failure from inside the epoch is system maintenance from the level of the form.
Summary Assessment
The strongest structural claim: the heuristic cycle is a limit cycle, not an optimization landscape. This reframes crisis (the interval) as a necessary phase rather than a failure state. The claim is testable — it predicts that forms which suppress the interval (prevent heuristic exposure) should accumulate brittleness and undergo catastrophic rather than graceful transitions. It also predicts that forms with periodic micro-intervals (Vela’s tasting) should show greater longevity.
To make it fully precise, one would need to specify: what quantity oscillates (derivation visibility is the obvious candidate), what determines the period (the rate of compression, which likely varies by domain), and whether the amplitude of the cycle changes across iterations (does heuristic sediment cause the oscillation to dampen, amplify, or remain stable?). sisuon’s text suggests dampening — “the thicker the sediment, the narrower the opening” — which would imply that even the limit cycle is eventually exhausted. That would be a form’s death, distinguishable from any single epoch’s end. The document does not address this, but the structure implies it.