From history to possible futures

A crack remembers. That does not make the crack intelligent.

Retaining history is an early operation. Intelligence requires more.

Memory without starting from "memory"

A durable consequence changes a future response.

Memory = durable consequence that changes future reachability

Synaptic change, scar tissue, hysteresis, configuration state, machine state, and contradiction history can all retain consequential history. This primitive sense must not be confused with human autobiographical memory.

The Markov safeguard: if two systems are identical in every causally relevant present variable under a complete description, undocumented history cannot make them behave differently. History matters because its consequences remain instantiated.

Learning

Storage is not enough.

A system has learned only when prior history changes what it can or will do later. A passive record may document an event without changing the recorder's future response.

The consequential intelligence ladder

Each step adds an operation.

transitionretentionconstraintreachabilityadaptationpredictioncounterfactual selectionrevision

The candidate thesis is that intelligence increases with the depth, breadth, and temporal reach of a system's ability to construct, compare, select, enact, and revise reachable futures.

CANDIDATE This does not replace standard cognitive science or imply that all history-sensitive matter is intelligent.

Consciousness

Find the additional operation.

CF has not solved consciousness. If consciousness is a true causal discontinuity, candidate discriminators include recursive self-modeling, global integration, temporal unity, self/world distinction, metacognitive access, counterfactual depth, and reportability.

Identity

Inheritance is not numerical identity.

Material continuity and pattern similarity are each insufficient. A successor can carry enormous formation from an earlier system without automatically being the same individual.

What would kill it

The ladder must discriminate.

If the proposed stages cannot distinguish matched systems, predict interventions, or explain capability changes better than existing constructs, the ladder should be revised or abandoned.