Attractor Models of Behavioural Stability

Someone begins exercising through a structured programme.

The sessions are scheduled. Other people expect them to attend. The location, timing and exercises have already been decided. For several weeks, exercise becomes part of their routine.

Then the programme ends.

Without the same schedule and accountability, sessions become less frequent. Competing demands take their place. Before long, the person has returned to something resembling their earlier pattern.

If they already changed their behaviour, why did the change not remain?

We can observe that supportive structure was present, some of that structure was later removed and behaviour changed across those conditions. An attractor model offers one possible interpretation: the programme may have temporarily supported different behaviour without fully changing everything that made the earlier pattern likely.

Return is not inevitable, and the example does not prove that the person has a formal behavioural attractor. It introduces a way of understanding how behaviour can change temporarily while retaining a tendency to return.

Stability does not mean standing still

We often imagine stability as the absence of movement. A stable thing stays where it is.

But many systems remain stable through movement.

Standing upright requires continual adjustment. The body does not hold one perfectly fixed position. It shifts, detects imbalance and responds. The exact position changes from moment to moment, yet the broader pattern—remaining upright—persists.

Behaviour can show a similar kind of organisation.

A stable behavioural pattern does not require someone to perform exactly the same action every day. Behaviour may fluctuate with energy, context, attention and opportunity while remaining within—or returning toward—a recognisable pattern.

One person may exercise more during one week and less during another while remaining generally active. Another may make several attempts to establish a routine but repeatedly return to long periods without exercise.

In both cases, stability concerns the pattern across time rather than any isolated moment.

Stability is about the organisation of movement, persistence and return—not the absence of movement.

What an attractor represents

A dynamical system is represented by variables that change over time. Its state is the current configuration of those variables. Its trajectory is the path it follows as that configuration changes.

An attractor is a state, cycle or set toward which a system tends to move under relevant conditions. In a formal model, trajectories from a surrounding range of starting states converge toward—or remain organised around—that attractor over time.

A stable resting point can be an attractor, but an attractor does not need to be motionless. A recurring cycle can also represent stable organisation. What matters is the pattern formed by trajectories, not whether the system stops moving.

The starting states that tend toward an attractor form its basin of attraction. In simpler language, the basin represents the range of modelled conditions from which the system tends toward the same pattern.

An attractor is not a hidden object inside the system. It is not one cause controlling everything else. It represents the organised result of interacting processes.

Researchers can study attractor dynamics in bounded forms of human behaviour. In sensorimotor coordination, for example, people may settle into stable movement patterns. Researchers can perturb the movement and measure whether it returns, changes or reorganises. Existing coordination tendencies can also influence how easily a new pattern is learned.

That is a stronger test than simply observing repeated behaviour. Recurrence alone does not demonstrate a formal attractor. Establishing one requires decisions about the relevant variables, repeated observations across time and a model of how the system evolves.

The landscape is a model, not a literal map

One simplified way to visualise attractors is as valleys in a landscape.

Imagine a ball resting in a valley. If it is pushed slightly uphill and released, the shape of the landscape guides it back down. The valley represents an attractor, while the surrounding slopes represent its basin of attraction.

Within this simplified picture, a deeper or steeper valley can represent a stronger tendency to return. A shallow valley can represent a pattern that is easier to leave. Several valleys represent several possible stable patterns.

But a person is not merely a passive ball.

Deliberate action can alter feedback, environments and future possibilities. A person can seek support, practise a response, restructure a situation or create consequences that influence what becomes likely next. Those actions occur within existing conditions, but they can also become part of what changes those conditions.

There are no literal valleys inside a person. The landscape depends on which variables a model includes and how their relationships are represented. It may change as the person, environment and relevant feedback change. Not every dynamical system can be described exactly as downhill movement across a fixed surface.

The model also does not capture the whole person. It represents selected processes relevant to the pattern being studied.

The landscape is therefore a visual representation of tendencies—not a complete psychological map or a picture of destiny.

Why behaviour may return

Return to the exercise programme.

During the programme, several conditions may have supported the routine simultaneously:

• predetermined sessions;
• a prepared environment;
• social accountability;
• fewer decisions about what to do;
• immediate signs of progress;
• protected time.

When the programme ended, several of those conditions changed together. Exercise had to compete with other demands without the same structure.

The person may still have developed skills, confidence or positive experiences. Temporary support can contribute to lasting change. But performing a behaviour under one set of conditions does not guarantee that it will remain equally likely when those conditions change.

This creates a useful distinction:

Temporary change alters what the system is doing. Reorganisation alters—or begins to alter—what the system tends to do under relevant conditions.

The two can overlap. What the system does now can also contribute to what it becomes more likely to do later.

One action can change future expectations. Repetition can alter skill, familiarity and feedback. A temporary structure can produce experiences that continue after the structure disappears. Reorganisation can begin gradually rather than arriving as one complete shift.

Returning to an earlier pattern does not mean the attempt achieved nothing. Nor does return identify the mechanism by itself. It may provide a reason to investigate which conditions continue to make the earlier pattern likely.

More than one pattern can be stable

A formal system can contain more than one attractor. This is called multistability.

Conceptually, this can help us interpret why someone displays one recurring pattern in a structured environment and another when structure is absent. They may respond one way when rested and another when depleted, or communicate openly in one relationship and defensively in another.

These examples illustrate context-dependent organisation. They do not, by themselves, prove that the person has multiple formal psychological attractors.

The model nevertheless helps correct the assumption that everyone has one permanent behavioural destination. Different conditions can make different patterns more accessible or stable.

A person may therefore appear inconsistent across contexts while behaving quite consistently within each one. Behaviour emerges through interaction between the individual and the conditions in which action occurs.

Change means altering tendencies, not controlling every action

If stability does not require identical behaviour, change does not require perfect consistency.

A behavioural system may be reorganising even when someone occasionally returns to an earlier response. What matters is whether relationships among possible actions and states are changing.

Over time:

• an earlier pattern may become easier to leave;
• another response may become more accessible;
• disruption may be absorbed differently;
• the system may return less often or less quickly;
• the same conditions may begin producing a different response.

These are ways the model can represent change, not instructions for how a person must intervene.

Individual actions still matter. They occur within an organisation that influences what follows, and they can contribute to changing that organisation. Learned expectations, environmental cues, social responses, current capacity and earlier consequences may all be relevant without any one factor controlling the pattern.

This does not mean that every part of life must be redesigned before change can last. It means that durable change may involve more than temporarily moving behaviour away from its familiar path.

People act within conditions that shape what is likely. Their actions can also become part of what changes those conditions.

What the model explains—and what it does not

Attractor models help explain how behaviour can remain stable while continuing to vary.

They can represent:

• several processes jointly supporting a recurring pattern;
• temporary support changing behaviour without guaranteeing maintenance;
• return after a disturbance;
• different contexts supporting different patterns;
• change altering tendencies without eliminating variation.

The model also offers a less moralising interpretation of behavioural return. Returning to an earlier pattern need not mean that someone lacked commitment or secretly wanted to fail. It may be consistent with conditions that continued making that pattern more likely.

But attractor language has limits.

The model cannot, by itself:

• prove that a recurring routine is a formal attractor;
• identify every process influencing it;
• predict someone’s next action;
• determine the correct intervention;
• replace psychological, biological or social explanations;
• remove agency from human behaviour.

An attractor is a model of patterned tendency. It is not a hidden force, fixed identity or predetermined destination.

Behaviour can remain likely to repeat without being destined to repeat. The conditions that support a pattern can also change.