ThinkingTrace | BraInference

Mechanism

Thinking is treated as movement through a space of concepts: either mapping a region or finding a route from a starting state to a goal. When the concepts are words from a SelfOrganisingVocabulary whose identity is fixed by ContentAddressedMinting, each step of the route can be written down as the word it passed through, in the order taken, with the revisions kept rather than overwritten. The record is then a walk on the graph of words: a ModelAsCompression of the thinking that keeps its route and drops the rest. Because the words are shared, two agents' walks are comparable, recurring walks become visible, and a walk left in the shared graph cues later agents the way a trail does in Stigmergy. The trace adds nothing to a single agent's thinking; its value is in what others, or the same agent later, can do with the route.

Invariants

Steps are named with shared, content-addressed words, not private labels
Steps are kept in the order taken, and a step may use only what came before it
A revision keeps the earlier step and the reason it was left, rather than deleting it
The trace is left where other agents can read it

Failure modes

Steps written in private words cannot be compared or reused
The record is reconstructed afterwards, so it shows the shortest path rather than the route actually taken
Uniform tracing of every step buries the few steps where the route changed direction
The trace stands in for the thinking: later agents repeat the route without re-checking that it fits
Near-duplicate words minted along the way fragment the shared vocabulary

Builds on

Used by

FuzzyAlgorithmReferences

Identity

The name is the hash of the definition. Change a word and it becomes a different pattern.

Sema ID: sema:ThinkingTrace#mh:SHA-256:dabd0b9391acbafa02d17431051643b6ca599bf3950de182a3c64d1e8e9a71c4
Reference: ThinkingTrace#dabd
Ring: 2
Tier: 2

View in graph JSON