AIM Framework glossary
Habit
Canonical definition
Habit-timescale action-selection at the Decision Hub, driven by the cached-value (habit) store intrinsic to the Decision Hub itself, without a fresh Ground-Resolution cycle. The Decision Hub acts on this cached store directly; the store is built up and updated in place over repeated cycles rather than read live from the Belief Register through the Observer. A Habit is therefore not motivationally sourceless: the cached values it acts on are laid down by repeated cycles of fresh integration in a stable context, so the store is written by what the Individual has done and not by what the Thinker currently believes, and holds a value with no record of what produced it. The automaticity of Habit is a feature of its bypassing fresh Ground-Resolution, not of any bypassing of stored value. Habit runs automatically only where two conditions hold: the Decision Hub's cached-value store holds a value for the Object, and the current context matches closely enough that the Observer's live Salience flag does not diverge from the cached expectation. Where either condition fails, the automatic path is not available and a Ground-Resolution cycle is recruited. Where the cached store holds no value for the Object (an unfamiliar Object with no cached attribution), there is nothing for the Decision Hub to act on, so the Decision Hub cannot act by Habit and the Thinker must resolve the Object from scratch. Where a posterior exists but the context is novel or unusual (the Observer flags a mismatch between what the cached value predicted and what is actually present), the Habit still fires first on the current pass, because the automatic route is faster than deliberation, and the novelty then registers as the trigger that recruits the Thinker to run a Ground-Resolution cycle after the fact. That cycle revises the Belief Register through a Belief-Register Update, which the Observer reads on its next cycle; the Ground-Resolution cycle does not retune the cached store, which changes only as repeated cycles in the new context lay down new values. Novelty-driven recruitment is source-neutral: it gates the Observer channel as such and does not depend on whether the routed posterior is A-dominant, I-dominant, or M-dominant. Habits can be decoded by altering conditions and observing which motivational source reasserts itself: a Habit formed from A-dominant posteriors will seek functional substitutes when the habitual Object is removed; a Habit formed from M-dominant posteriors (confabulated or otherwise) will seek status-equivalent substitutes. Habit is not a fourth motivational source but sub-deliberative selection on the Decision Hub's own cached-value store, gated by whether the context is familiar enough to spare the Thinker a fresh Ground-Resolution cycle.
Explanation
A Habit is what happens when the system acts without stopping to think the choice through afresh. Normally the Thinker runs a Ground-Resolution cycle before a decision, working out what an incoming signal means and what it is worth right now. A Habit skips that step. Instead the Decision Hub acts straight off its own cached-value store, a running record of what past cycles settled the Object was worth, which the Decision Hub holds itself and can act on without waiting on anything else. That is why a Habit is fast: it is running on a stored answer rather than working out a new one. This also means a Habit is not motivationless, and not a fourth kind of drive alongside Appetites, Intrinsic Motivation, and Mimetic Desire. It is running on old conclusions. Whatever drove the past cycles that laid the cached value down, a genuine bodily need (A), a genuine engagement (I), or a status pull (M), is what the cached value now carries forward and what is driving the behaviour, though the value keeps no record of which it was, and it keeps driving the behaviour until something makes the Thinker look again. Because the value is pre-stored, it does not correct itself just because circumstances have changed; correction requires a fresh Ground-Resolution cycle, whose revised posterior reaches the Observer through a Belief-Register Update and the Decision Hub's store through repeated cycles. A Habit runs automatically only while two things hold: the Decision Hub's store holds a value for the Object, and the situation is familiar enough that the Observer's live Salience flag still fits what that value predicted. If there is no stored value (an unfamiliar Object), there is nothing for the Decision Hub to act on, so there is no Habit to run and the Thinker has to work the Object out from scratch. If there is a stored value but something about the situation is novel or unexpected, the Habit still fires first, because the automatic route is faster than deliberation, and the novelty then acts as the trigger that calls the Thinker in to re-resolve the Object; what the Decision Hub's store holds changes only as the new situation repeats. This is why you can catch yourself doing the habitual thing and only afterwards realise the situation had changed. You can read what a Habit is really running on by changing the situation and watching what the system reaches for. Take away the habitual Object, and a Habit built on A-conclusions will look for a functional substitute (something that meets the same underlying need), while a Habit built on M-conclusions will look for a status-equivalent substitute (something that restores the same standing). The automatic part is only the skipping of fresh thinking, not the skipping of stored value; the cached value is exactly what the Habit is acting on.
Epistemic status
Framework-Specific Construct.
Derivation
Chapter 9, Section 9.8.1; Chapter 9, Section 9.8.5; Chapter 4, Section 4.4.8; Chapter 9, Section 9.8.2.