What is Intelligence? A Cycle Closure Perspective

TL;DR

This work proposes a cycle-closure perspective on intelligence rooted in the topological law $\partial^2=0$, where transient fragments cancel and only closed cycles persist as invariants, forming memory that enables prediction. It defines three core principles: Prediction Requires Invariance, Structure-before-Specificity (SbS), and Context-Content Uncertainty (CCUP), and introduces Memory-Amortized Inference (MAI) as a computational implementation that combines temporal bootstrapping and spatial bootstrapping to collapse specificity into structure. Semantics emerge from persistent cycles, which then scaffold syntax, with SBs and CCUP guiding dynamic alignment to minimize joint uncertainty. The framework is argued to unify cognitive science, machine learning, and evolutionary biology, illustrating how memory and prediction arise from invariant topological structures and how bootstrapping across time, space, and social interaction contributes to the evolution of intelligence. The paper also outlines concrete neural mechanisms (e.g., oscillatory phase coding) and proposes a spiking-network blueprint for testing SbS, CCUP, and MAI, highlighting potential advances in neuromorphic design and in-memory computation. Overall, the work offers a principled, mathematically grounded path to understanding and engineering intelligent systems that exploit cycle invariants and amortized closure.

Abstract

What is intelligence? We argue for a structural-dynamical account rooted in a topological closure law: \emph{the boundary of a boundary vanishes} ($\partial^2=0$). This principle forces transient fragments to cancel while closed cycles persist as invariants, yielding the cascade $\partial^2\!=\!0 \Rightarrow \text{cycles (invariants)} \Rightarrow \text{memory} \Rightarrow \text{prediction (intelligence)}$. Prediction requires invariance: only order-invariant cycles can stabilize the predictive substrate. This motivates the \textbf{Structure-before-Specificity (SbS)} principle, where persistent structures ($Φ$) must stabilize before contextual specificities ($Ψ$) can be meaningfully interpreted, and is formalized by the \textbf{Context-Content Uncertainty Principle (CCUP)}, which casts cognition as dynamic alignment that minimizes the joint uncertainty $H(Φ,Ψ)$. We show that \textbf{Memory-Amortized Inference (MAI)} is the computational mechanism that implements SbS\,$\rightarrow$\,CCUP through dual bootstrapping: \emph{temporal} bootstrapping consolidates episodic specifics into reusable latent trajectories, while \emph{spatial} bootstrapping reuses these invariants across latent manifolds. This framework explains why \emph{semantics precedes syntax}: stable cycles anchor meaning, and symbolic syntax emerges only after semantic invariants are in place. In an evolutionary perspective, the same closure law unifies the trajectory of natural intelligence: from primitive memory traces in microbes, to cyclic sensorimotor patterns in bilaterians, to semantic generalization in mammals, culminating in human symbolic abstraction by natural language. In sum, intelligence arises from the progressive collapse of specificity into structure, grounded in the closure-induced emergence of invariants.

Figures (6)

• Figure 1: $\partial^2=0$ enforces the dot-cycle dichotomy.Left: An open chain $\sigma$ has a nonzero boundary $\partial\sigma$ and collapses to a dot (class in $H_0$), carrying no relational content. Middle: The boundary operator squares to zero: $\partial(\partial\sigma)=0$. Right: A closed chain $\gamma$ with $\partial\gamma=0$ persists as a homology class $[\gamma]\in H_1$, i.e., a cycle that encodes order-invariant structure.
• Figure 2: Trivial, nontrivial, and order-invariant cycles.Left: A boundary of a filled region is trivial in $H_1$. Middle: A loop around a hole cannot bound any 2-chain, so it represents a nontrivial homology class. Right: Once a trajectory closes into a cycle, its homology class depends only on the multiset of moves, not their order: order permutations yield the same $H_1$ class.
• Figure 3: SbS as a corollary of the dot-cycle dichotomy. Open fragments cancel as boundaries, whereas closed loops persist as invariants. Closure (via $\partial^2=0$) and persistence funnel contextual variability $\Psi$ into structural invariants $\Phi$, which anchor memory and prediction; specificity is layered afterward.
• Figure 4: Dynamic alignment under CCUP as a feedback loop. Left: context $\Psi$ contains open, order-dependent fragments. A forward inference stage projects $\Psi$ toward content $\Phi$. A mismatch detector compares projected context to persistent cycles and drives a controller that adapts $\Psi$ so that joint uncertainty $H(\Phi,\Psi)$ decreases. Topological closure ($\partial^2=0$) cancels boundary terms in the loop, ensuring that only closed structures persist as carriers of prediction.
• Figure 5: Cycle of MAI. Instead of recomputing $\Phi^* = \arg\min \mathcal{L}(\Psi, \Phi)$, the system reuses prior trajectories: $\Phi_{t+1}$ and $\Psi_t$ guide memory-based retrieval via $\mathcal{R}$, and bootstrapping $\mathcal{F}$ updates the latent state $\Phi_t$. The process forms a self-consistent loop grounded in structured memory.

Theorems & Definitions (46)

• Lemma 1: $\partial^2=0$ Enforces the Dot-Cycle Dichotomy
• Corollary 1: Topological closure yields cycles
• Remark 1: Topological vs. cycle closure
• Proposition 1: Nontrivial homology classes as substrates of memory
• Proposition 2: Memory enables prediction
• Example 1: Invariance in chemotaxis and phototaxis
• Corollary 2: Forecasting as Entropy Reduction
• Remark 2: Invariance cycles as semantic backbone
• Remark 3: Cycle closure leads to order invariance
• Theorem 1: Homological Equivalence of Permuted Cycles