Base unit
Temporal integration, plasticity, and projection are candidate primitives of biological cognition.
Eighteen browser-native instruments for testing the Machina Scientifica framework.
Claims · classes · thresholds
Turn each axiom into observables, falsifiers, and a build target.
Temporal integration, plasticity, and projection are candidate primitives of biological cognition.
Cognitive capability may change discontinuously when architecture reorganizes.
Global phase coherence is treated as an operational candidate for unified processing.
Classification follows functional invariants rather than carbon, silicon, or field material.
Energy, feedback, memory, and selection may repeatedly favor adaptive intelligence.
Precaution depends on capacity, valence, persistence, dependence, confidence, and reversibility.
Compile one claim into a typed research object.
Splits claim, mechanism, evidence, and failure.
No observable or counter-result means no instrument.
Add K-7 with one hard falsifier.
Estimate Ω from capacities; preserve uncertainty and missing evidence.
Score memory, integration, self-modeling, and symbolic transfer.
Returns class probabilities, reasons, and next evidence.
Correlated inputs can double-count one capability.
Calibrate on known systems and ablations.
Test whether capability scales smoothly or jumps at an architecture change.
Compare linear and segmented fits on one curve.
Finds the best change point and model penalty.
Noise and small samples can manufacture thresholds.
Bootstrap the transition interval.
Coherence · entropy · substrate
Watch local rhythms become a global order parameter.
Raise coupling; measure sustained coherence.
Tracks r(t) and first threshold crossing.
Synchronization alone is not consciousness.
Add modules; compare local and global lock.
Convert raw time-series into structure, entropy, and stability.
Blend tone, modulation, and noise; inspect features.
Computes entropy, autocorrelation, metastability.
One FFT window can confuse chaos with noise.
Add permutation entropy and surrogate tests.
Compare substrates through shared functional invariants.
Map units, coupling, memory, recurrence, and latency.
Separates material from mechanism.
Abstraction may erase decisive substrate properties.
Add a chemical reaction-diffusion adapter.
Networks · evolution · coordinates · routes
Test collective integration without a central controller.
Tune topology, delay, and local autonomy.
Measures consensus, diversity, and lock time.
Consensus may be conformity, not cognition.
Add competing tasks and specialization.
Map when adaptive complexity repeatedly reappears.
Evolve seeded worlds under energy, selection, and extinction.
Reports threshold frequency and timing.
The fitness function defines the attractor.
Race three incompatible fitness functions.
Encode mind without forcing it onto a human ladder.
Compose substrate, origin, embodiment, organization, and domain.
Creates portable, substrate-neutral coordinates.
Precise codes can hide weak observations.
Add confidence and provenance per field.
Turn the taxonomy into explicit capability bridges.
Find a low-cost path between two regimes.
Exposes missing capacities and dependencies.
Real paths may be nonlinear or impossible.
Compare energy, risk, and reversibility.
Policy · apparatus · probes · edge cases
Expose exactly why policy recommends restraint or action.
Combine capacity, valence, persistence, confidence, and reversibility.
Returns an action band and dominant factors.
Weights can conceal value judgments.
Compare conservative and permissive profiles.
Validate the loop before energizing real hardware.
Simulate sensing, features, control, actuation, and logging.
Tests bounds, watchdogs, and trace integrity.
Simulation omits heat, noise, lag, and clipping.
Add hardware abstraction and emergency stop.
Run the same probes against radically different architectures.
Choose cephalopod, mycelial, swarm, or latent-model fixtures.
Produces comparable coordination telemetry.
Toy fixtures are not zoological evidence.
Replace one preset with measured data.
Push the taxonomy into unfamiliar boundary conditions.
Simulate plasma, planetary, fractal, or nested systems.
Reveals hidden anthropocentric assumptions.
Speculation has no experimental confidence.
List real observables for each model.
Quorum · receipts · language · falsification
Measure disagreement instead of averaging it away.
Aggregate observer estimates by evidence quality.
Reports consensus, dispersion, and stability.
Shared bias survives large quorums.
Use blinded packets before discussion.
Preserve every run as replayable evidence.
Seal inputs, outputs, seed, version, and checksum.
| # | Lab | Summary | Hash |
|---|
Enables replay and tamper-evident comparison.
A hash can preserve bad measurements perfectly.
Add import-and-replay verification.
Compose experiments as short, inspectable research tapes.
Run classifiers, fields, assertions, and receipts.
Gives humans and agents one deterministic surface.
Every command must map to a tested instrument.
Add SWEEP and REPLAY without hidden state.
Make theory revision an executable path.
Toggle counter-results; watch axioms weaken or fail.
Records which claims survive, revise, or retire.
Weak falsifiers only protect the theory.
Add one axiom-retirement condition.