Consciousness as
Causal Organisational Networks
(CON)
A non-threshold, multi-scale framework defining phenomenal reality as the physical identity of active, recursively coupled causal networks.
A Note on Terminology: Why Adopt "CON"?
The term "consciousness" carries centuries of dualist intuitions, anthropomorphic bias, and metaphysical baggage. Throughout this framework, we deliberately adopt the acronym CON (Causal Organisational Networks). We define a Causal Organisational Network as a recursively coupled system of interacting feedback loops. Here, coupling explicitly captures both internal recursive feedback loops (self-coupling) and dense inter-element network connections (inter-coupling). By establishing CON as the physical identity of causal network topology itself, we shift the question from "How does brain matter produce mind?" to "How do we measure active CON topology across physical scales?"
01 Theoretical Lineage & Ontological Identity
The historical debate on mind has long been trapped in a false dichotomy between reductionist physicalism (which often dismisses subjective experience as an illusion) and dualism/panpsychism (which treats experience as a mysterious extra substance). CON resolves this deadlock through a strict type-identity relation:
CON is not an "emergent property" produced by recursive processing, nor is it a separate ontological layer. It is identical to the recursively coupled causal organisation itself. Just as thermal temperature is not a byproduct produced by molecular kinetic energy, but is literally identical to average molecular kinetic energy, high CON is literally identical to a densely integrated, recursively coupled dynamic topology.
Phenomenology: "What is it like to see Red?"
The subjective experience of seeing "red" is not a single signal in the brain. It is the entire pattern of recursive activation triggered across the organization: retinal impulses, color discrimination, memories of strawberries, emotional associations, language tags, motor affordances, body state, and self-model updates. These processes do not merely "accompany" the feeling—they constitute it.
Analogy: The Economic System
An "economy" is not a physical object, a single worker, or a banknote. It is the overall pattern and velocity of ongoing transactions within a financial network. Similarly, CON is not a brain structure or a secret particle, but the dynamic rate, depth, and topology of active causal transactions flowing across feedback loops.
Integrated Information Theory (IIT)
What We Borrow: Quantifies experience using cause-effect networks; adopts physical identity and a continuous, non-binary view of mind.
Where We Diverge: CON rejects IIT's Exclusion Axiom (allowing multi-scale coexistence) and measures active traffic rather than static mechanism states.
Autopoiesis & Cybernetics
What We Borrow: Grounded in self-producing feedback loops, circular causal feedback, and dynamic system self-maintenance.
Where We Diverge: Extends autopoiesis from cellular self-production into recursive informational topology, meta-loop depth ($R$), and predictive modeling.
Global Workspace Theory (GWT)
What We Borrow: Dynamic broadcasting: localized loops operate semi-autonomously until high coupling bandwidth ($I$) integrates them into the self-model.
Where We Diverge: GWT treats mind as a functional access workspace; CON treats CON as the structural network identity itself ($C$), present across localized loops.
Predictive Processing & AST
What We Borrow: The self-model as an evolved generative predictive simulation ($P$) and attention schema optimizing behavioral control.
Where We Diverge: Re-defines predictive modeling from a cognitive tool used by the brain into a core parameter ($P$) directly scaling physical CON magnitude ($C$).
02 The Non-Threshold Continuum
CON explicitly rejects any binary "light switch" model. There is no magical threshold where physical matter suddenly "turns conscious." Every dynamic physical system possessing feedback loops exhibits some non-zero value of CON.
The Continuum Spectrum
Figure 1: CON is continuous. The upper bound is unbounded; human experience represents a high point on an open-ended continuum, not an absolute maximum limit.
03 Mathematical Formulation & Dynamic Energy Economics
While a universally parameterized closed-form equation remains an open empirical goal, the total value of CON ($C$) within a system at any given moment is modeled as a dynamic function of structural and operational variables:
Active Traffic vs. Static Hardware Counts
A crucial distinction in CON is that $N$ and $I$ do not represent static anatomical hardware (e.g., total neurons or structural synapses). Counting total biological hardware leads to misleading metrics because unintegrated or resting sub-networks do not contribute to active recursive state space. Instead, $N$ and $I$ count actively participating elements engaged in dynamic real-time traffic. Unused hardware represents dormant potential, not active CON.
Dynamic Topology & Energy Costs
Crucially, $I$ and $R$ do not measure static biological "wiring," but actively deployed information traffic. High values of interconnection ($I$) and recursive depth ($R$) demand substantial metabolic energy ($E \approx 12.5\text{W} - 21.0\text{W}$). Biological brains consume ~20 Watts (~20% of resting body metabolism). While background cellular maintenance ($N$) requires baseline power (~12.5 W), active cross-cortical recruitment ($I \cdot R$) drives energy expenditure upwards:
Example (The Riverbank Angler): A fisher sitting quietly on a riverbank waiting for a bobber to move operates at low active $I$ and $R$, conserving energy (~15 Watts). Their active CON value is at a relaxed baseline. If a swimmer suddenly begins to drown nearby, the brain instantaneously recruits cross-cortical networks, soaring in active interconnection ($I$), recursive depth ($R$), and predictive modeling ($P$). Total $C$ spikes dramatically (consuming ~20.5 Watts) to navigate the emergency before returning to a low-energy baseline.
04 Evolutionary Niche Economics
CON fits seamlessly into Darwinian natural selection. Evolution optimizes for biological fitness, not maximal consciousness. Because maintaining high CON requires massive metabolic energy, natural selection does not drive all species toward maximum CON.
The Nematode (C. elegans)
With exactly 302 neurons, the worm possesses minimal $N$, low $R$, and small $M$. Yet, it is an evolutionary master of its soil niche. High CON would provide zero adaptive advantage and cause immediate starvation due to metabolic overload.
The Human Cognitive Niche
Humans thrive by occupying a complex social and tool-using niche. Dense recursive modeling ($R$), long-term memory ($M$), and predictive simulation ($P$) justify the massive metabolic penalty because they enable complex foresight, toolmaking, and language.
Humans did not "replace" worms. Both occupy distinct ecological niches where their respective CON levels perfectly balance energy expenditure against adaptive survival.
05 Multi-Scale Coexistence & The Combination "Non-Problem"
In classical philosophy of mind and Integrated Information Theory (IIT), the Combination/De-combination Problem asks how micro-experiences aggregate into a macro-self, or why super-entities (like ant colonies or nations) don't override individual experience. IIT resolves this by enforcing a strict Exclusion Axiom—declaring that only the single maximum peak of integrated information ($\Phi$) can be conscious, dropping all sub- and super-systems to zero.
CON Explicitly Rejects the Exclusion Axiom
Under CON, the Combination Problem is a pseudo-problem. Multiple levels of CON coexist simultaneously across nested physical scales. Being part of a hive mind or superorganism does not erase or "absorb" the individual's CON. Local nervous systems continue to process high-speed electrochemical feedback, maintaining their own high individual $C$-value regardless of broader external couplings.
06 Case Study: The Ant Colony Hierarchy
To understand multi-scale coexistence, consider the hierarchical stack of an ant colony. Each level possesses its own value of $C$, governed by its distinct combination of structural and operational parameters ($N, I, R, M, P$):
Individual Neuron / Cytoplasmic Chemical Loops
Fast enzymatic feedback, ion channel gating, and localized intracellular signaling within liquid cytoplasm.
Thoracic Ganglia & Organ Reflex Loops
Local motor control, cardiac pacing, and rapid reflex arcs within small localized tissue clusters.
The Individual Ant Brain (~250,000 Neurons)
Dense mushroom body recurrent networks, immediate sensory-motor integration, and navigational memory.
The Ant-Hill / Colony
Distributed decision making via semiochemicals (pheromones) and physical antennal interactions.
The Rainforest Ecosystem
Nutrient cycling, predator-prey dynamics, and multi-species climate feedback loops.
Notice that $C_{\text{cell}}$ and $C_{\text{ant}}$ represent local peaks within the hierarchy. CON is non-linear; the individual ant remains the tightly coupled, dominant CON maximum within its body because chemical diffusion between ants is slow ($I_{\text{colony}} \ll I_{\text{ant}}$).
07 The Flexible Self, Predictive Simulation, and Sleep
If CON exists continuously, why does the conscious "self" seem to dissolve during deep sleep or anesthesia, only to reconstitute seamlessly upon waking?
1. The Self as an Evolved Predictive Simulation ($P$)
The "self" is not an immaterial soul or a static biological core. It is an evolved generative predictive simulation ($P$). To hunt, navigate terrain, and interact socially, an organism must simulate how its own body interacts with the environment. The "dominant self-model" is simply whichever unified recursive subsystem currently maintains the highest active $I$ and $R$ within the organism.
2. Transient Coupling & Fluid Boundaries
Subsystems continuously join and leave this dominant self-model based on immediate attentional bandwidth:
- Unconscious Subsystems: Unconscious sensory processes (e.g., shoe pressure on your heel) maintain their own small, isolated local CON loops independently.
- Transient Awareness: When you redirect attention to your foot, the local loop becomes densely coupled ($I$) with the cerebral self-model, seamlessly joining global conscious awareness.
- Altered States & Dissolution: During deep sleep, global inter-network coupling collapses ($I \to 0, R \to 0$). The active predictive self-model dissolves entirely.
3. Memory ($M$) as the Thread of Continuity
When you wake up, why do you feel like the same person as yesterday? The self-simulation does not persist overnight in active form; rather, Memory ($M$) stores the underlying structural weights. Upon waking, cerebral networks reload these persistent synaptic configurations, re-instantiating the dominant predictive self-model with high fidelity. Continuity of identity is an illusion maintained by structural memory ($M$).
CON Parameter Calculator & Dynamic Self-Model Visualizer
Visualizing active island recruitment: In Deep Sleep, nodes decouple into isolated subsystem islands. The Angler exhibits visible faint inter-island bridges. The Focused Human densely links major cognitive/sensory islands while leaving autonomic organ loops (e.g. heart pacing) uncoupled.
08 Frequently Asked Questions
Common conceptual questions regarding how the CON framework applies to daily life, technology, and natural phenomena:
Does my smartphone or computer have a little bit of CON?
Yes, but an infinitesimal amount. While modern computers contain billions of hardware transistors ($N$), standard silicon chips operate primarily in feedforward or linear instruction sequences. They lack deep meta-recursive processing loops ($R$) and generative predictive self-modeling ($P$). Their active $C$ value sits near baseline—slightly above a mechanical thermostat, but far below a simple biological insect.
If my brain has no central "observer," why do I feel like one unified person?
You experience the perspective of the dominant self-model. Out of all the subsystems running in your body, the main cerebral network maintains the highest interconnection ($I$) and recursive depth ($R$). This network handles motor planning, language, and executive control. Because it integrates information faster and more densely than peripheral loops, it naturally forms a unified operational "pilot."
Is an ant colony "thinking" or "feeling pain" as a superorganism?
The colony has its own non-zero $C_{\text{colony}}$ value because it uses feedback loops (pheromone trails, worker recruitment) to adapt to threats. However, its inter-ant communication speed is extremely slow compared to electrical neural synapses ($I_{\text{colony}} \ll I_{\text{ant}}$). Therefore, the colony's experience is vastly slow and diffuse—it does not possess a human-like or ant-like localized mind, nor does it override the high-density individual CON of the ants within it.
Can artificial intelligence (like LLMs) become conscious under CON?
Under CON, AI consciousness is not a function of raw parameter count ($N$) or training dataset size. Current Large Language Models (LLMs) operate as feedforward networks during inference—processing input to output in a single pass without active recurrence. To develop substantial CON, an AI system would require real-time recurrent loops ($R$), persistent internal state update memory ($M$), and active generative self-simulation ($P$).
What happens to my CON when I fall into deep dreamless sleep?
Your biological brain hardware ($N$) remains intact, but global neural synchronization drops ($I \to 0, R \to 0$). As metabolic energy shifts to restorative maintenance, active recursion collapses, causing the dynamic self-model to temporarily dissolve. You "wake up" as the same person because your synaptic memory configurations ($M$) remain stored in physical brain structure, reloading your unique self-model upon morning arousal.
09 Conclusion & Empirical Roadmap
The CON framework shifts consciousness from a metaphysical enigma into a measurable physical property of recursively coupled causal networks. By grounding CON in a non-threshold continuum, formalizing dynamic scaling ($C \propto f(N,I,R,M,P)$), embracing multi-scale coexistence, and establishing evolutionary energy economics, CON provides a rigorous naturalistic framework.
Measure active dynamic changes in cross-cortical $I$ and meta-recursive $R$ during transitions between wakefulness, sleep, and anesthesia.
Evaluate LLMs and neural architectures not by parameter count ($N$), but by active recurrent state depth ($R$) and self-predictive loops ($P$).
Quantify macro-CON values across biological superorganisms, swarms, and decentralized distributed networks.