Glossary

Agency

• The capacity to maintain internal order (homeostasis) by distinguishing “Me” from “Not‑Me,” updating a model (Context), and acting in a closed loop (sense→update→intervene). See 🧭Agency & Delegation.

Agent surprisal (subjective information)

• How unexpected an outcome $i$ is for a specific agent with beliefs $Q$: $surprisal = −log Q(i)$. Distinct from source surprisal $−log P(i)$. See 📊Information: Objective vs Subjective.

Autocatalytic set

• A chemical network where members catalyze each other’s formation, enabling self‑maintenance. Used here as a testbed for proto‑agency. See 🧪Autocatalytic Sets & Proto‑Agency.

Closed loop (control)

• A cycle where sensing updates a model which selects an action that changes future sensed data. Minimal operational marker of agency. See ✅Operational Criteria.

Complex amplitude (quantum)

• $ψ$ with modulus ($|ψ|^2$: outcome statistics) and phase (context‑sensitive relations). Choice of basis/interferometry can expose phase as observable structure. See 🧿Quantum Foundations.

Context

• An agent’s model: priors, expectations, values, goals, and policies used to interpret Texts and act. Often denoted $Q$ when probabilistic. See 📝Text–Context–Interpretation.

Delegation of agency

• The moment an event enters an agent’s closed loop so one alternative becomes the enacted outcome (e.g., choosing a measurement basis and registering a result). See 🧭Agency & Delegation.

Discreteness (events)

• Agents register changes via thresholds and sensitivities, turning continuous processes into sequences of significant events. See 🧩Core Concepts.

Dual Reality

• The claim that every phenomenon has two complementary sides: Objective (Energy) and Subjective (Information/Context). Both should be made explicit in explanations. See 🌓Dual Reality.

Dual second law

• Pairing: (1) In closed systems, entropy does not decrease (objective). (2) Open systems/agents can export entropy and build local order by using free‑energy flows (subjective/process). See 🔥Open‑System Thermodynamics.

Energy (in this framework)

• The conserved physical substrate of change (objective). What sensors ultimately register. See 🌓Dual Reality.

Free energy (predictive)

• A scalar measuring model mismatch/cost; reduced when agents improve predictions under constraints. Related to negative log evidence and control costs. See ✅Operational Criteria.

Glossary (this page)

• A compact, linkable dictionary of core terms used across the site.

Hierarchies (levels)

• Organization across scales (molecules → cells → organisms → groups). Each level has cognition limits: some Texts from higher levels cannot be correctly interpreted below. See 🧬Hierarchies & Quasi‑Organisms.

Info‑cloning

• Formation of similar Contexts across many agents via shared priors/instincts/protocols rather than individual learning. Tends to increase uniformity. See 🧬Hierarchies & Quasi‑Organisms.

Information (objective vs subjective)

• Objective: source properties (Shannon), entropy $H(P)$, surprisal $−log P(i)$.
• Subjective: agent surprisal $−log Q(i)$ given the agent’s Context $Q$.
• Learning reduces the gap (e.g., $KL(P \|\| Q)$). See 📊Information: Objective vs Subjective.

Interpretation

• The mapping from Text to meaning and action using an agent’s Context. See 📝Text–Context–Interpretation.

KL divergence (Kullback–Leibler)

• A measure of mismatch between two distributions; used here to track how far an agent’s Context $Q$ is from a source $P$. See 📊Information: Objective vs Subjective.

Model (of the Universe)

• A structured representation (often probabilistic) an agent uses to predict, explain, and control. Here, synonymous with Context when owned by an agent. See 🧭Agency & Delegation.

Operational criteria (for agency)

• Minimal, testable markers: closed loop; counterfactual sensitivity (different actions → detectably different future Texts); model improvement over time under resource budgets. See ✅Operational Criteria.

Phase (quantum)

• The argument of a complex amplitude encoding relations among alternatives. Becomes observable through appropriate setups (e.g., interferometry). See 🧿Quantum Foundations.

Quasi‑organism

• A stable collective that shares and reproduces a common Context (species‑typical repertoires/instincts), behaving organism‑like at its scale. See 🧬Hierarchies & Quasi‑Organisms.

Research directions

• Compact, falsifiable project ideas aligned with existing toolchains across fields. See 🔬Research Directions.

Shannon information

• Objective, distribution‑based measures (entropy $H(P)$, surprisal $−log P(i)$) describing sources, not specific agents. See 📊Information: Objective vs Subjective.

Text

• Any objective carrier: signals, traces, data, records — what sensors can register, independent of who reads them. See 📝Text–Context–Interpretation.

Thermodynamics (open systems)

• Study of energy, work, and entropy in systems exchanging matter/energy with the environment; here paired with an explicit agent layer. See 🔥Open‑System Thermodynamics.

Work from coherence (quantum thermo)

• Protocols converting quantum coherence (off‑diagonals) into extractable work or information under control operations. See 🎛️Quantum Control & Feedback.