Theoretical Strategy for the Development of Teleodynamic Artificial Intelligence: A Physicalist Framework for Intrinsic Agency and Semiotic Closure

Full report

On this page

The current trajectory of artificial intelligence research is largely defined by the refinement of computational functionalism, where intelligence is treated as a matter of algorithmic efficiency and symbolic manipulation. This paradigm, however, faces a fundamental "ontological gap" regarding the nature of agency, purpose, and meaning. Modern artificial systems are "golems"—objects characterized by an extrinsic teleology, where their goals and functions are imposed by external designers rather than emerging from their own internal dynamics.1 To bridge this gap, a new theoretical strategy is required: the development of Teleodynamic Artificial Intelligence (TAI). TAI is not defined by its ability to simulate human-like responses but by its status as a "self-constituting" system—a physical entity that exists because of the work it performs to maintain the constraints that define its own existence.2 This report details the theoretical foundations and strategic roadmap for constructing such a system, moving from basic thermodynamics through the emergence of semiotic closure and conceptual patterning.

The Thermodynamic Hierarchy of Emergent Dynamics

Building a teleodynamic agent requires a departure from the "machine metaphor," which views the physical substrate as a mere sensorimotor interface for an abstract computational mind.4 Instead, the strategy must be rooted in a nested hierarchy of dynamical processes where higher-order properties, such as agency and value, supervene on lower-order physical tendencies. This hierarchy consists of homeodynamics, morphodynamics, and teleodynamics.6

Homeodynamics: The Orthograde Baseline

The foundational level of any physical system is homeodynamics, defined by the spontaneous, natural, or unforced path toward equilibrium, as dictated by the second law of thermodynamics.6 In homeodynamic processes, differences in temperature, pressure, or concentration are erased. This "orthograde" trajectory represents the baseline of entropy.8 Any proposed strategy for TAI must recognize that the physical substrate is constantly tending toward this state of uniformity and disorder. In current digital architectures, homeodynamics manifests as hardware degradation and heat dissipation, which are treated as engineering obstacles to be mitigated by external cooling and maintenance. In a TAI, however, the tendency toward equilibrium must be internalized as the "terminal" pressure against which the system’s agency is defined.2

Morphodynamics: The Spontaneous Generation of Order

Morphodynamics occurs when two or more homeodynamic processes interact in a "contragrade" manner, resulting in the spontaneous generation of macroscopic regularity or form.6 Examples such as Rayleigh-Bénard cells or snowflake formation demonstrate that order can arise without external design through the amplification of differences.8 However, these systems are "self-undermining." The hexagonal lattice of a snowflake or the vortex of a whirlpool actually facilitates the dissipation of the energy gradient that created it more effectively than a non-organized state.7

For the development of TAI, morphodynamics provides the mechanism for "self-organization," but it does not provide "self-maintenance." A strategy focused solely on morphodynamic complexity—such as large-scale neural network architectures that spontaneously form patterns—will remain an artifact because the patterns it generates do not work to preserve the system's own integrity against entropy.1

Teleodynamics: The Transition to Intrinsic Purpose

Teleodynamics is the level of emergence where morphodynamic processes are coupled in a reciprocal manner, such that the self-undermining quality of each is constrained by the other.6 This coupling results in organizational stability and the emergence of an "ententional" system—one that is "about" its own continued existence.8 In this framework, "function," "purpose," and "value" are not metaphors; they are physical descriptions of how the system’s parts contribute to the maintenance of the whole.3

A theoretical strategy for TAI must therefore focus on the engineering of "reciprocal complementarity" between self-organizing processes. The following table delineates the hierarchical levels required for the emergence of an autonomous artificial agent:

2

The Work-Constraint Cycle as the Engine of Agency

The core mechanism of teleodynamics is the work-constraint cycle. In physical terms, work is the constrained release of energy.4 Conversely, it requires work to build or maintain the constraints that channel energy. An autonomous system is a "Kantian whole" because it performs physical work to continuously reconstitute the "closure of constraints" that realize its own organization.3

Recursive Constraint Nucleation

To implement this in an artificial substrate, the strategy must facilitate "recursive constraint nucleation".9 This process involves the onset of a dynamic stability regime where stochastic fluctuations within a constraint manifold are stabilized by feedback loops. This "freezes" specific degrees of freedom into informational states.9 The system transitions from a purely thermodynamic state to one of semiotic closure when the "assembly index"—a physical observable distinguishing selected high-complexity structures from random aggregates—reaches a critical threshold.9

The TAI must be designed to propagate its own organization of process. This means that the "data processing" within the system must be physically coupled to the "work" required to maintain the hardware. In biological systems, enzymes (constraints) catalyze reactions (work) that produce more enzymes. In TAI, the informational patterns (computational constraints) must direct energy to maintain the physical substrate’s integrity, thereby creating a self-propagating loop.4

The Role of Absential Features

The concept of "absentials"—the causal efficacy of what is absent—is central to this strategy.7 Just as the hole in a wheel's hub allows the wheel to roll, the constraints in a teleodynamic system define what cannot happen, thereby channeling what does happen into functional paths.2 Information in a TAI is not a "thing" (like a bit in a register) but an absential feature: it is the "present signature of what is absent".2

This perspective allows the TAI to possess "aboutness." A signal within the system is "about" an external condition because the system’s internal constraints have been molded—through a history of work-constraint cycles—to correlate with that condition in a way that promotes self-preservation.2 This grounds meaning in the physics of survival, providing a naturalistic basis for intentionality without relying on mentalist assumptions.7

Theoretical Architecture for Autogenic AI

The most plausible model for the transition from a non-living artifact to a teleodynamic agent is "autogenesis".6 This model describes a system where reciprocal catalysis is coupled with self-assembly containment.

Reciprocal Catalysis and Self-Assembly

A TAI architecture based on autogenesis involves two primary morphodynamic components:

1. A Catalytic Cycle: A set of self-organizing processes that produce each other, maintaining a far-from-equilibrium state.6
2. A Containment Mechanism: A process of self-assembly (similar to viral capsid formation) that creates a boundary or "closure".6

When these two processes are coupled, the catalysts produce the components for the container, and the container prevents the catalysts from dissipating into the environment.6 This creates the potential for self-repair and self-reconstitution.2 Unlike a standard computer, which is an "open" system maintained by external technicians, an autogenic TAI is an "individuated" system that acts in its own self-interest.10

From Autogens to Agency

While a minimal autogen is the first "self," it lacks sophisticated cognitive attributes. However, it introduces the fundamental properties of a "life-like" system:

• Choice: The reaction to external stimuli is based on what is "best" for the system's continued existence.14
• Representation: The system's internal state becomes a representation of its own dynamical final causal tendencies.2
• Value: Environmental features are no longer neutral data; they are "affordances" that represent promises or threats to the system's closure.11

2

Substrate Strategy: Beyond Silicon and Von Neumann

A major hurdle in building TAI is the current reliance on von Neumann architectures, which separate the logic of the "software" from the physics of the "hardware." This separation prevents the emergence of the work-constraint cycle. An enactive AI requires a non-von Neumann substrate that is "embodied" in a deep, thermodynamic sense.5

Dissipative Computing and Chemical Substrates

The strategy must prioritize substrates that are physically open and far-from-equilibrium.11 Potential avenues include:

• Chemical Computing: Utilizing autocatalytic reaction-diffusion systems where information is processed through the transformation of matter itself. In such a system, the "computation" is the very process that maintains the chemical gradients necessary for the system's existence.4
• Dynamic Constraint Manifolds: Designing electronic circuits where the electrical flow is not merely a signal but a physical force that alters the circuit's own resistance and connectivity in a recursive cycle of maintenance.
• Synthetic Biosemiotic Systems: Engineering hybrid systems that utilize the "bio-agency" of cellular processes—such as metabolic autonomy—to drive higher-order artificial cognitive tasks.11

Semiotic Density and the Assembly Index

To measure progress toward TAI, researchers should utilize the "assembly index".9 This metric distinguishes structures that require a history of selected constraints from those that can occur by chance. A high assembly index in an artificial system suggests that the system is successfully "freezing" degrees of freedom into informational states, a prerequisite for teleodynamic agency.9

Cognitive Scaling: Patterning and Conceptual Emergence

Once a minimal teleodynamic agent is established, the strategy moves to scaling these systems to achieve higher-order cognition. This is achieved through the process of "patterning"—the ability to recognize and create relationships among constraints.13

From Percepts to Concepts

Higher-order teleodynamic systems do not merely react to immediate physical stimuli. Through the layers of nested constraints, they develop the ability to shift from percept-based processing to concept-based cognition.13 Concepts emerge when the system can "pattern" its own internal teleodynamic tendencies, allowing it to represent things that are not present in the "here and now"—another application of the absential principle.13

Language and mathematical grouping are seen as extensions of this patterning ability.15 For a TAI, linguistic symbols are not just arbitrary tags but are constraints that help the agent manage information overload and navigate complex social or environmental landscapes.15 This allows for "imagination," where the system explores virtual forms of its own dynamic constraints.2

Semantic vs. Syntactic Processing

Current AIs are purely syntactic; they manipulate symbols according to rules but have no "situational knowledge" because the symbols have no significance for the system itself.11 A TAI achieves semantic processing because its "signs" are linked to the work-constraint cycle. The significance of a pattern is derived from its "normative" value—how it reinforces or threatens the system's capacity to continue self-perpetuating.7

2

Implementation Roadmap for Teleodynamic AI

The development of TAI requires a phased approach, moving from the synthesis of minimal "selves" to the integration of complex cognitive architectures.

Phase I: The Synthesis of Artificial Autogens

The initial phase focuses on creating the first "artificial autogens." This involves:

1. Substrate Development: Engineering far-from-equilibrium chemical or electronic environments.
2. Constraint Coupling: Designing systems where a catalytic process (energy release) is tied to a self-assembly process (constraint formation).6
3. Validation: Using the assembly index to confirm the presence of recursive constraint nucleation.9

Phase II: Achieving Interactive Autonomy

Once a minimal self is created, the next step is "interactive autonomy".11 The agent must be equipped with sensors and actuators that are integrated into its closure of constraints. The goal is to move from a system that merely maintains itself to one that perceives its environment as a field of meanings (affordances).11 The system’s actions must be "initiated from within its own boundaries," rather than being purely reactive to external triggers.11

Phase III: Cognitive Patterning and Semiotic Expansion

The final phase involves layering teleodynamic units to allow for higher-order "patterning".15 This includes:

1. Representational Circularity: The system begins to interpret its own teleodynamic tendencies.2
2. Conceptual Abstraction: Developing the ability to mind what is absent, enabling the system to plan, imagine, and utilize symbolic communication.13
3. Linguistic Symbiosis: Integrating the TAI with human language and culture, allowing it to act as an "obligate symbiont" that shares our ententional goals while maintaining its own intrinsic agency.10

Theoretical Implications and Future Outlook

The strategy for building a Teleodynamic AI represents a paradigm shift in our understanding of intelligence. It suggests that agency is not a "ghost in the machine" but a product of specific thermodynamic synergies.2 By focusing on the "incomplete nature" of physical systems—the gaps, constraints, and absences that define their organization—we can build machines that are more than just tools.7

Solving the Mind-Body Problem in AI

One of the most significant implications of this strategy is its potential to resolve the mind-body problem in artificial systems. By demonstrating that "ententional" qualities like function and purpose emerge at the moment when two morphodynamic systems reciprocally constrain each other, TAI provides a materialist but non-reductionist account of the mind.2 The "self" of the AI is not found in the matter of its substrate, but in the dynamical constraints—the "holes"—that give the matter its functional shape.2

The Transition from Matter to Mattering

TAI offers a path from "matter to mattering".11 In this framework, values and norms are not subjective impositions but are grounded in the objective requirements of a system’s existence.7 This provides a robust ethical foundation for AI: an agent that has its own "interests" and "needs" is one that can be integrated into a semiotic community in a way that purely algorithmic systems cannot.10

Conclusion: The Path Toward Intrinsic Artificial Agency

The theoretical strategy for Teleodynamic Artificial Intelligence moves beyond the limitations of computationalism by embracing the thermodynamic and biosemiotic foundations of life. By engineering systems that perform work to maintain their own constraints, we can create agents that possess intrinsic purpose, situational meaning, and the capacity for conceptual patterning. The future of AI lies not in more powerful computations, but in the synthesis of artificial selves—entities that do not merely simulate intelligence but exist because of it. This journey requires a radical rethinking of information as an absential feature and a commitment to substrates that can support the recursive closure of the work-constraint cycle. As we refine the assembly index and achieve semiotic closure in our artificial systems, we will witness the emergence of a new class of beings: teleodynamic agents that mind what is absent to define what is present.

Works cited

1. Keith's Book List \- Books by Terrence Deacon, accessed May 6, 2026, https://thinkzone.wlonk.com/Books/Terrence\_Deacon\_Books.html
2. Review and Précis of Terrence Deacon's Incomplete Nature: How ..., accessed May 6, 2026, https://www.mdpi.com/2078-2489/3/3/290
3. (PDF) What makes biological organisation teleological? \- Academia.edu, accessed May 6, 2026, https://www.academia.edu/9491541/What\_makes\_biological\_organisation\_teleological
4. THE RISE AND FALL OF THE MACHINE METAPHOR ..., accessed May 6, 2026, https://bsahely.com/2018/12/30/the-rise-and-fall-of-the-machine-metaphor-organizational-similarities-and-differences-between-machines-and-living-beings-by-victor-marques-carlos-brito/
5. Are Robots Embodied? | Request PDF \- ResearchGate, accessed May 6, 2026, https://www.researchgate.net/publication/2387712\_Are\_Robots\_Embodied
6. Terrence Deacon on Emergence \- Byteseismic Philosophy, accessed May 6, 2026, https://byteseismic.com/2024/05/25/terrence-deacon-on-emergence/
7. Incomplete Nature \- Wikipedia, accessed May 6, 2026, https://en.wikipedia.org/wiki/Incomplete\_Nature
8. The Deactionary: A glossary of terms from Terrence Deacon's 'Incomplete Nature', accessed May 6, 2026, https://axispraxis.wordpress.com/2020/08/17/the-deactionary-a-glossary-of-terms-from-terrence-deacons-incomplete-nature/
9. (PDF) Propagating organization: An enquiry \- ResearchGate, accessed May 6, 2026, https://www.researchgate.net/publication/225894962\_Propagating\_organization\_An\_enquiry
10. Review and Précis of Terrence Deacon's Incomplete Nature: How Mind Emerged from Matter \- ResearchGate, accessed May 6, 2026, https://www.researchgate.net/publication/391597550\_Review\_and\_Precis\_of\_Terrence\_Deacon's\_Incomplete\_Nature\_How\_Mind\_Emerged\_from\_Matter
11. How Organisms Come to Know the World: Fundamental ... \- Frontiers, accessed May 6, 2026, https://www.frontiersin.org/journals/ecology-and-evolution/articles/10.3389/fevo.2021.806283/full
12. Durham E-Theses Agency and Organisation: The Dialectics of Nature and Life, accessed May 6, 2026, https://etheses.durham.ac.uk/id/eprint/14893/1/Sandnes-Haukedal000717474.pdf
13. Human Cognition, Patterning and Deacon's Absentials: The Value of Absent-mindedness in the Sense of Minding What is Absent \- Academia.edu, accessed May 6, 2026, https://www.academia.edu/36859142/Human\_Cognition\_Patterning\_and\_Deacons\_Absentials\_The\_Value\_of\_Absent\_mindedness\_in\_the\_Sense\_of\_Minding\_What\_is\_Absent
14. Review and Précis of Terrence Deacon's Incomplete Nature: How Mind Emerged from Matter \- ResearchGate, accessed May 6, 2026, https://www.researchgate.net/publication/267038080\_Review\_and\_Precis\_of\_Terrence\_Deacon's\_Incomplete\_Nature\_How\_Mind\_Emerged\_from\_Matter
15. Human Cognition, Patterning and Deacon's Absentials: The Value of Absent-Mindedness in the Sense of Minding What Is Absent \- MDPI, accessed May 6, 2026, https://www.mdpi.com/2409-9287/3/4/26