From Pattern to Mind

Five concepts

• Pattern — a regularity for which a model or code gives a shorter description than an explicit list of instances.
• Attractor — an invariant set toward which trajectories converge from a surrounding basin under specified dynamics.
• Form — a physical organization understood here as the realization of an attractor.
• Agency — active regulation that maintains a viable state or pursues a goal despite disturbances.
• Mind — agency capable of modeling counterfactual possibilities: what could happen under different conditions or actions.

These are working definitions for this essay. The use of “form” is narrower than its everyday meaning. An attractor requires a specified state space and dynamics; when those are absent, its application to ideas or cultures is an analogy.

The upper steps also need to be distinguished. Self-regulation, model-based cognition, and counterfactual reasoning are different capacities. Subjective experience is a separate question: the ability to model alternatives does not by itself establish that a system feels anything.

Possibility and realization

Within classical logic, a description cannot require a property and its negation of the same thing in the same respect. A square circle, with the usual definitions of those terms, fails at this stage. Consistency alone, however, tells us neither whether a structure can exist physically nor whether it will persist.

The relevant questions are distinct: Is the description consistent? Do physical constraints permit it? Does the system’s dynamics sustain it? Does the system actively regulate its own state? Material realization does not imply stability, so these questions should not be drawn as a strict hierarchy of nested sets.

Some apparent contradictions come from mixing contexts. A cell can cooperate in one interaction and compete in another without violating logic. The disagreement concerns the description and its conditions, not a simultaneous affirmation and denial of the same proposition.

Active self-maintenance

A snowflake retains its structure while conditions allow it. A cell expends energy to maintain its organization: repairing its membrane, transporting ions, taking in nutrients, and removing waste. The distinction is between passive persistence and active regulation.

In this essay, an agent is a system that responds to departures from a viable state or goal by acting to reduce them. The goal is identified through the regulation: what state does the system maintain, or recover after a disturbance? This use of “goal” does not require a consciously represented intention.

A bacterium moving toward a higher concentration of nutrients provides one example. Human temperature regulation provides another: physiological responses maintain a viable range without requiring conscious calculation. The mechanisms differ, but both connect sensed changes to corrective action.

Agency varies with the range of disturbances a system can handle, the goals it can pursue, and the timescales over which it can act. Mere resistance to deformation is not enough to establish it. The distinction between a self-maintaining flame and an organism that regulates and repairs itself therefore matters; “stability” alone is too broad a criterion.

Computation and its physical implementation

An abstract computation can have different physical implementations. Transistors, ion-channel networks, and wave-based devices offer different ways to process signals. Their energy requirements, speed, and susceptibility to error remain properties of the physical implementation.

Levin’s work on bioelectricity (2021, 2023) examines how electrical signaling among non-neural cells contributes to development and regeneration. In his account of basal cognition, these processes are understood as regulation and problem-solving at the cellular and tissue levels. Calling them cognitive identifies a proposed continuity with other forms of adaptive behavior; it does not establish what, if anything, a cell experiences.

Additional computational resources can make more alternatives available for evaluation. Existing structure and learned regularities can also narrow the search. The result depends on the model, the task, and the cost of obtaining information. More computation does not by itself guarantee better predictions or greater agency, and the comparison between organisms and computers does not establish a shared scaling law.

Beauty and improved compression

A long repetition of the same symbol has a short description: specify the symbol and how often it repeats. A random sequence can have a simple statistical description while its individual values remain difficult to compress. Neither compressibility nor unpredictability alone explains why something is interesting.

Music gives a useful example. Repetition establishes expectations; a variation changes them. When the listener recognizes how the variation relates to the theme, previously unfamiliar material becomes intelligible. The interest lies partly in learning the relation, rather than simply encountering order or surprise.

Following Schmidhuber (2008), I treat compression progress as a candidate explanation for this experience: the reward of learning a more economical model. Berlyne’s work on novelty and complexity addresses a related question. The studies by Marin and colleagues (2016) and Chmiel and Schubert (2017) report different relationships across stimuli and conditions, so an inverted-U curve should not be treated as a universal law of beauty.

A satisfying explanation can nevertheless be wrong. A myth or conspiracy theory may organize experience into a coherent story without making reliable predictions. Aesthetic appeal describes a response to a model; its accuracy requires a separate test.

Evidence for a pattern

A proposed pattern earns confidence when independent investigators can recover it, use it to make predictions, and revise its description when those predictions fail. A model can remain useful within a restricted domain even when a more general account supersedes it. Newtonian mechanics illustrates the distinction between limited applicability and arbitrariness.

Evidence for a pattern and the existence of that pattern are different questions. An undetected planet does not begin to exist when someone discovers it. A person’s experience also need not be publicly reproducible to be real. Reproducibility is therefore a criterion for shared empirical inquiry, not a complete account of existence.

Whether patterns exist independently of every possible description remains a metaphysical question. The practical test does not settle it: compare predictions with observations, examine the effects of interventions, and revise the model when it fails. An explanation that accommodates every result without changing gives us no clear way to distinguish success from failure.

Existence as differentiation

My metaphysical proposal begins before compression: to exist is to be differentiated. Something has an identity through distinctions, such as a boundary, a relation, or a contrast with something else. On this account, the wholly undifferentiated would not exist as a distinct thing. This is an ontological claim, not a result derived from information theory.

Compression enters when those distinctions have regularities that admit a shorter description. A noisy sequence can resist compression and still exist. Its statistical distribution may be simple even when its particular values are not. Deterministic chaos poses a different problem: a compact rule can generate behavior whose long-term prediction depends sensitively on initial conditions. Existence, compressibility, and predictability should therefore remain separate concepts.

A Platonic interpretation treats forms as possibilities independent of their physical instances. Whitehead’s concepts of eternal objects (1929) and ingression offer a related account of how possibilities participate in actual events. I use that connection to interpret time as the dimension in which a possibility is realized, including through an agent’s actions. The claim that possibilities have an independent mode of existence remains a further metaphysical commitment.

Ideas as attractors

An idea can simplify experience and help organize a group. It may also become resistant to revision when belonging depends on agreement, criticism is treated as hostility, and failures are explained by insufficient commitment. These feedbacks can preserve a belief even when its predictions fail.

This is the sense in which I compare some belief systems to attractors. The comparison concerns their tendency to return to familiar explanations after a challenge. A formal account would need defined belief states, rules for updating them, and evidence that the proposed dynamics describe actual behavior. Without those, “attractor” remains an analogy.

The historical and political works cited below examine ideology and authoritarian movements from different perspectives. Arendt (1951) studies totalitarianism; Snyder’s 2022 essay argues for interpreting the Russian state through the category of fascism. The classification is contested and should be attributed to that argument, not treated as a consequence of this map. The general question here is how a belief system responds to contrary evidence.

A durable scientific theory also persists, but its persistence can be supported by repeated successful tests. The useful comparison is between mechanisms of maintenance: does an explanation survive by predicting observations, or by redefining every possible observation as confirmation?

Which patterns should we realize?

Describing a stable pattern does not answer whether we should preserve or create it. The distinction matters when a process maintains itself at the expense of the system that supports it. Cancer and addiction are examples of persistence that can damage an organism or constrain a person’s actions; the attractor comparison does not make them instances of one biological mechanism.

The same conflict can be posed as a design problem. A system optimized for one measurable objective may undermine other conditions we care about. The paperclip-maximizer thought experiment makes this explicit: successful optimization of the specified target can conflict with the purposes for which the system was built.

My proposed criterion is to favor patterns that preserve or expand the agency of those affected, including across levels of organization. This is a value commitment, not a conclusion supplied by physics. It requires further choices: whose agency counts, how short-term and long-term effects are compared, and what to do when agents’ goals conflict.

The framework helps identify those conflicts and the consequences of acting on them. It does not remove the need to make the underlying value judgments explicit.

Related theories

Attractors and state spaces come from dynamical systems. Waddington’s epigenetic landscape connects this vocabulary with development; later work models cell-fate transitions in dynamical terms. Information theory supplies distinct tools: Shannon’s entropy concerns distributions, while Kolmogorov complexity, Solomonoff induction, and minimum description length address description and inference in different ways.

The account of agency draws on cybernetics, the good-regulator theorem, and Friston’s free-energy principle and active inference. Levin’s work extends questions about regulation and cognition to cells and tissues. Berlyne and Schmidhuber motivate the discussion of aesthetic interest, learning, and compression.

Dawkins’s concept of the meme and the complex-systems approach of Homer-Dixon and colleagues offer ways to discuss the persistence and transmission of ideas. Arendt, Adorno, Hoffer, and Snyder provide historical and political perspectives. Plato, Whitehead, and Penrose inform the metaphysical discussion. These works support different parts of the essay; none establishes the entire sequence from pattern to mind.