AIM Framework glossary
Thinker
Canonical definition
The reader-facing name for Intrinsic Motivation (I) used in the Minimal Architecture Specification and throughout the treatise where the architectural role of the component (the two architecturally inseparable functions of continuous prior-maintenance and Bayesian ground-resolution, plus the locus of deliberative inference) is under discussion rather than the Axiom-1 source-taxonomy role. The Thinker is not a distinct component from I; it is the same Disaggregation Primitive specified under its reading-surface function rather than under its source-taxonomy label. It is introduced in Chapter 4 Section 4.4 in the course of specifying the minimal architecture, and is used in preference to "Intrinsic Motivation" wherever the architectural function (prior-maintenance across the four prior-classes, outward template-supply to the Observer, Bayesian resolution of the Decision Hub's Scalar into ground-attributions, write-access to the Belief Register, being the sole locus of deliberative inference in the architecture, being the sole architectural target of the Knowledge Patch) is the point being made. Use "Intrinsic Motivation" or "I" where the Axiom-1 taxonomic role is the point being made, and "Thinker" where the architectural function is the point. The two labels denote the same component. The Thinker is the thinking. What a reader would recognise as conscious cognition (inference, Hypothesis-formation, vocabulary acquisition, revision of Belief, deliberate sustained engagement with a process) is what the Thinker does. The Thinker's ground-resolution function is not invoked on every cycle: where the Decision Hub's cached-value store already holds a value adequate to the context, action-selection proceeds at habit timescales on that store without recruiting the Thinker (see Habit). The Thinker is recruited only by an Object the Belief Register cannot already resolve, so novelty and prediction-mismatch are the Thinker's activation condition rather than incidental inputs; Curiosity is the felt signature of that recruitment and Boredom the felt registration of its exhaustion. Ordinary sensory input, including information received from other Individuals, reaches the Thinker at its deliberative timescale and is not routed through the Decision Hub. The Decision Hub's Scalar is the additional channel on which ground-resolution of integrated valuation operates.
Explanation
The Thinker is just another name for I, used when we are talking about what the component does in the architecture rather than where it sits in the three-source taxonomy. The Thinker thinks. It holds the mental library the rest of the system needs (what hunger feels like, what food looks like, who counts as a Model, how the world is carved up into categories), and it figures out why something has caught your attention (whether it matters because the body needs it, because engaging with it is worthwhile, or because other people are paying attention to it). The Thinker name makes that job easy to remember: it is the part that thinks. The previous technical name ("prior-maintainer-and-Bayesian-resolver") was accurate but read as jargon, and once Chapter 4 has done the work of specifying what I does, the short name "Thinker" can carry the rest of the treatise. We call it "Intrinsic Motivation" when we are talking about the three sources of motivation and where I fits among them, and we call it the "Thinker" when we are talking about how the system is wired and what each component is responsible for. Same component, two labels for two different kinds of discussion. Affective posteriority specification. Because the Thinker is the slow, serial, deliberative faculty, it operates by its very function downstream of the fast automatic integration that emits the Common-Currency Scalar and Phenomenal Affect. It therefore cannot act on either before both are already formed. This lateness is entailed by the Thinker's functional definition and does not depend on any particular empirical timing result; the two-stage structure of emotional processing observed empirically is a corroboration of it, not its ground. The Thinker's available operations with respect to an affective episode are, first, to process the already-delivered Phenomenal Affect after the fact, and second, to revise the prior loaded into the loop so as to condition the next episode of the same type, which can both attenuate the magnitude and shorten the abatement of subsequent affect but can never eliminate it (see Affective Latency of the Knowledge Patch, Confabulation Race).
Epistemic status
Canonical.
Derivation
Chapter 4, Section 4.4.3; Chapter 4, Section 4.4.6; Appendix A, Section A.5.6.
Related terms
- Intrinsic Motivation (I)
- Disaggregation Primitive
- Dashboard Gauge Principle
- Minimal Architecture Specification
- Decision Hub
- Body Manager
- Observer
- Conscious Cognition
- Bayesian Updating
- Hypothesis Space
- Belief Register
- Confabulation
- Knowledge Patch
- AIM Vocabulary Acquisition
- Epistemic Suspension
- I-Override
- Confabulation Rationality Lemma
- Signed Valuation Principle
- Phenomenal Affect
- Affective Latency of the Knowledge Patch
- Confabulation Race
- Curiosity
- Boredom
- Memory Store