Update 2026-06-11-general-guidelines-for-G-expression.md

elements/neuron/appunti/2026-06-11-general-guidelines-for-G-expression.md

@@ -2,7 +2,99 @@
 include_toc: true
 ---
 
-# Complete set of general guidelines
+# Logic principles
+
+Here are the logic principles that have guided the construction, organized from the most fundamental to the most specific.
+
+---
+
+## I. Principles of Resource and Conservation
+
+**Nothing is free.** Every behavior consumes a budget. There is no operation in the system that does not draw down some resource. This is not a constraint added on top of the logic — it is the foundation of the logic. Selectivity, competition, and forgetting all emerge from the fact that resources are finite.
+
+**Resources are redistributed, not created.** The total resource pool is bounded by an external ceiling (the vascular glucose supply). Within that ceiling, the system moves resources around — from one synapse to another, from structure back to pool when structures are dismantled. No internal process can manufacture new capacity; it can only reallocate existing capacity. Learning is therefore always at the expense of something else.
+
+**Every economy has a single capped root.** Each resource type traces back to one producer with a hard output ceiling — the astrocyte cell body for synaptic energy, the soma for neuronal structural material. Everything downstream competes for shares of that capped production. The ceiling is the ultimate arbiter of how much the system can do.
+
+**Conservation makes one synapse's gain another's loss.** Because resources are shared and finite, potentiation at one site necessarily reduces what is available elsewhere. This coupling is not designed — it is the automatic consequence of drawing from a common pool.
+
+---
+
+## II. Principles of Time and Scope
+
+**Behavior and structural change occupy separate scopes.** Fast behavior happens in DAY; permanent structural change happens in NIGHT. This separation prevents transient activity from directly rewriting architecture — every noise spike would otherwise remodel the system. The scope boundary is the mechanism that makes the system both responsive and stable.
+
+**DAY accumulates evidence; NIGHT acts on it.** No decision about permanent change is made during DAY. DAY only gathers traces. NIGHT reads the accumulated evidence and commits. The system never commits in the moment — it always defers commitment to a consolidation phase that operates on aggregated evidence.
+
+**Every quantity has a characteristic timescale, and timescale is meaning.** Fast traces decay in milliseconds, tags in hours, structures over days. The decay constant of a variable is not a parameter — it is what the variable means. A variable that decays fast is a momentary signal; one that decays slowly is a commitment. Putting two different timescales in one variable destroys both meanings.
+
+**Time windows are enforced by chemistry, not by clocks.** The system never checks a timer. Coincidence windows emerge from the competition between accumulation and decay. A signal must arrive while a trace is still elevated. The window opens when the trace crosses a threshold and closes when it decays below it. Timing is a consequence of dynamics, not an explicit rule.
+
+---
+
+## III. Principles of Capacity and Occupancy
+
+**NIGHT builds containers; DAY fills them.** Structural variables are capacities — ceilings on what behavior can achieve. NIGHT changes the ceiling. DAY operates within it. The two never do each other's job: NIGHT never places a receptor, DAY never builds a slot.
+
+**Short-term change is occupancy; long-term change is capacity.** Filling a container more or less is fast and reversible. Changing the size of the container is slow and persistent. The same physical quantity — receptor count, vesicle count — has a fast component (how full) and a slow component (how big), and these are governed by entirely different processes.
+
+**Structure is simultaneously memory and prior.** The architecture left by the last NIGHT is both a record of past significant experience and a bias on how the next DAY will respond. A potentiated synapse is more likely to respond strongly and therefore more likely to be potentiated again. Structure encodes what mattered and predicts what will matter.
+
+---
+
+## IV. Principles of Locality and Non-Locality
+
+**Short-term change is local; long-term change is non-local.** A component can transiently potentiate from its own activity alone. But to permanently change, it requires validation from beyond itself — from other compartments, from the soma, from the organism. Cheap reversible change is autonomous; expensive permanent change requires external authorization.
+
+**Permanent change requires coincidence across spatial scales.** A tag forms only when a local eligibility signal meets one or more non-local confirmation signals. The number of required coincidences reflects where the component sits in the hierarchy — the postsynapse, as the primary memory locus, requires the most. Each scale confirms something the previous scale cannot know about itself.
+
+**The whole validates the part; the part cannot validate itself.** A synapse cannot know whether its activity was behaviorally significant — that information exists only at the organism level. The neuromodulatory broadcast carries organism-level significance down to the synapse. This is why the system is open: the highest validation comes from outside any component that is being modified.
+
+---
+
+## V. Principles of Selection and Asymmetry
+
+**Potentiation is the active drive; depotentiation is its shadow.** The entire machinery is oriented toward strengthening what is significant. There is no symmetric machinery for weakening. Weakening happens to whatever potentiation did not select, as a consequence of the resources potentiation consumed. The system is built to learn, and forgetting is the cost of learning.
+
+**Depression is never explicit — it is what happens when potentiation does not.** No signal says "weaken this." Structures decay continuously and are held up only by maintenance resources. When potentiation consumes those resources, unmaintained structures drift down. Depression is the absence of maintenance, not the presence of a depression signal.
+
+**Selection requires winning on multiple independent criteria.** To be permanently strengthened, a synapse must be both active enough to be fueled and significant enough to be validated. These are independent gates. Activity without significance is not saved; significance without sustainable activity cannot be maintained. The conjunction is what filters for genuinely valuable connections.
+
+**The system finds equilibrium through the residual of its own imperfection.** Where alignment or balance is achieved, the very success removes the signal that drove it, allowing slow drift back toward imbalance, which regenerates the driving signal. The system hovers near optimum, continuously corrected by the small errors its own imperfect state produces. Equilibrium is dynamic, maintained by residual error, never static.
+
+---
+
+## VI. Principles of Bottom-Up Emergence
+
+**Complex temporal behavior emerges from local reactive traces, not from explicit computation.** The soma aligns with its input rhythm without representing the rhythm. It simply leaves a trace when an input arrives during refractoriness and lets that trace speed future recovery. Prediction, anticipation, and rhythm-tracking emerge from purely local, reactive deposits — never from a model of the future.
+
+**The system never represents what it is becoming tuned to.** A potentiated synapse does not contain a representation of the pattern it responds to — it is physically biased toward that pattern. The tuning is the structure, not a description of the structure. Prediction is implicit physical bias, not explicit expectation.
+
+**Global organization arises from local competition.** Sparsification, normalization, and winner-take-more dynamics are nowhere computed centrally. They emerge automatically from many local units drawing from shared pools. The astrocyte does not decide which synapses to fuel — the synapses' own demands, competing for capped production, produce the allocation.
+
+---
+
+## VII. Principles of Coupling and Self-Reinforcement
+
+**Couplings create trajectories, not just states.** Some variables, once moved in a direction, make further movement in that direction easier — the astrosynapse wrapping tighter after potentiation, which makes future potentiation easier. These self-reinforcing couplings mean the system has momentum: it does not just occupy states, it follows trajectories, deepening whatever direction it has begun.
+
+**The same signal can serve opposite functions through different receptors.** Glutamate spillover brakes the presynapse while exciting the astrocyte — one ligand, two receptor types, opposite cascades, simultaneous opposite effects. Function is determined by the receiver, not the signal. This lets one event coordinate multiple responses without any coordinating mechanism.
+
+**Energy availability is itself a selective pressure, parallel to validation.** Beyond the explicit activity-and-reward gating, the simple availability of energy continuously selects which synapses can participate. A synapse that cannot be fueled cannot generate the activity that would let it be tagged. Metabolism silently shapes what can be learned, in parallel with and independent of the explicit plasticity machinery.
+
+---
+
+## VIII. Principles of Openness and Boundedness
+
+**The system is finite and open, not infinite and closed.** It has bounded components and a bounded state space, and it receives inputs it cannot generate from within — sensory drive, neuromodulatory validation, metabolic supply. Because it is finite, its self-modification does not generate infinite regress. Because it is open, its highest validation comes from outside itself.
+
+**The fixed points are made explicit, not hidden.** The parameters the system cannot modify from within — thresholds, the vascular ceiling, the neuromodulatory signals — are declared as fixed. These are the system's boundary with what it did not set and cannot inspect. Making them explicit is the honest acknowledgment that every self-modifying system operates within constraints it did not choose.
+
+**Validation comes from embedding, not from internal consistency.** The system does not certify its own changes as correct. Whether a structural change was good is answered by the organism's subsequent experience in the world, fed back through the neuromodulatory system. Correctness is determined by the coupling between system and environment, not by any internal criterion. This is what it means for the fixed point to lie outside the system: the system acts, the world responds, and the response — not any internal check — determines what was worth keeping.
+
+
+
+# Complete set of general guidelines (old)
 
 ## DAY — Behavior