A Relational Frame Theory Account of Parental Alienation, Part 2: A Computational Model

Introduction: From Theory to Testable Model

In a previous theoretical article, we proposed that Relational Frame Theory (RFT) offers a powerful behavior-analytic framework for understanding parental alienation. The core thesis was that alienation can be understood as a pathological excess of rule-governed behavior, where a child's actions and beliefs about a parent become decoupled from their direct, contingency-shaped experience. Conversely, genuine estrangement was framed as contingency-shaped behavior: a rational response to genuinely harmful or aversive experiences.

While this theoretical distinction is useful, it remains abstract. How can we test it? How can we explore the variables that make a child more or less susceptible to alienating rules?

This article proposes a direct answer: we build a machine to think about the problem. We will specify a concrete agent-based computational model that translates the principles of RFT into a virtual laboratory. Critically, every parameter in this model is grounded in published empirical research rather than arbitrary assumption. This model will allow us to simulate the complex interactions between parents and a child, operationalize the specific behaviors that define a parent's behavioral profile, and watch the mechanics of alienation unfold in a controlled, measurable environment.

Background: The RFT Engine of Alienation

As established by Hayes, Barnes-Holmes, and Roche (2001), Relational Frame Theory is a theory of human language and cognition. Its central tenets provide the engine for our model:

Derived Relational Responding: Humans learn to arbitrarily relate stimuli and derive relations that were never directly taught. A child learns that "Dad" is related to "good" through experience, but can then derive that things associated with Dad are also "good" without direct training.
Transformation of Stimulus Functions: When the function of one stimulus changes, the functions of all related stimuli can transform. If an alienating parent successfully frames "Dad" (a stimulus) as "dangerous" (a new function), the child's bedroom in Dad's house, his favorite activities with Dad, and even his memories can take on this new aversive function.
Rule-Governed Behavior: Behavior can be controlled more powerfully by verbal rules ("Don't trust your father, he is manipulative") than by the direct contingencies of experience (Dad is consistently kind and loving). In healthy development, rules and experience are in constant dialogue; in alienation, this dialogue is severed.

Our model is designed to simulate exactly this severance.

The Model: An Agent-Based Simulation

An agent-based model is a computational simulation that allows us to create virtual "agents" (in our case, a Child, a Target_Parent, and an Alienating_Parent) and define the rules for their interaction. By running the simulation thousands of times, we can observe the emergent properties of the system.

Why a Three-Agent System?

We deliberately constrain the model to the parent-child triad. While teachers, therapists, grandparents, and the court system can provide competing verbal rules or alternative safe-base contingencies, in forensic practice courts overwhelmingly focus on the interactions between the two parents and the child. Modeling only these three agents reflects the clinical and legal reality of how alienation is assessed. Future iterations may introduce external agents, but the core dynamics we seek to model operate within this triad.

The Child Agent

The Child_Agent is the central agent, and it is not passive. It is an active participant whose internal state and behavior create feedback loops that influence both parents.

Properties:

belief_in_target_parent: A float from -1.0 (total rejection) to +1.0 (total acceptance). Starts at +0.5.
rule_susceptibility_state: A float from 0.0 to 1.0 representing the child's current level of cognitive fusion and reduced psychological flexibility.

What rule_susceptibility_state means, precisely:

This variable is not a vague proxy for "emotional dysregulation." It is a specific, operationally defined measure of the degree to which the child treats verbal rules as equivalent to direct experience.

At 0.0 (Low Susceptibility): The child is in a state of high psychological flexibility. They can experience the alienating parent's verbal rules as words or thoughts, separate from their own direct experience. The impact of alienating behaviors on belief is minimal.
At 1.0 (High Susceptibility): The child is in a state of total cognitive fusion. They are unable to differentiate their own experience from the parent's verbal rules. The alienating rules are treated as direct threats or absolute truths, and their impact on belief is maximized.

How rule_susceptibility_state is measured and changes:

The Pervasive Dysregulation Behaviors (such as EmotionalLability and IdealizationDevaluation) from the alienating parent are the primary drivers that increase the child's rule_susceptibility_state score. This score then acts as a multiplier on the child's base_sensitivity_to_rules, making them more vulnerable to alienating verbal rules.

The score also has a natural decay rate per time step, representing the child's baseline capacity for self-regulation when not under active dysregulating influence.

How the child's age changes the model:

Research by Kelly and Johnston (2001) identified that developmental stage is a key contributing factor to alienation. Baker and Darnall (2006) found empirically that older children and girls were more likely to be reported as severely alienated. We incorporate this as follows:

For a young_child agent (ages 4-9):
- Higher base_sensitivity_to_rules (0.20): Reflects developmental reliance on caregiver instruction and limited capacity for independent perspective-taking.
- Slower decay rate (0.02 per step): Fewer self-regulation tools; the effects of dysregulation linger longer.
- Lower impact from identity-based attacks: Young children are less susceptible to IdealizationDevaluation because their identity is less peer-referenced.
For an adolescent agent (ages 10-16):
- Lower base_sensitivity_to_rules (0.12): Greater capacity for abstract thought and perspective-taking provides some initial buffer.
- Higher impact from identity-based attacks: IdealizationDevaluation and ForcingAThreateningChoice have amplified effects, as they target adolescent vulnerabilities around identity, social standing, and autonomy.
- Faster but more volatile decay rate (0.05 per step): Emotional states spike higher but can decrease faster, modeling the more volatile but transient emotional regulation of adolescence.
- Potential for rebellion behavior: At high rule_susceptibility_state combined with high belief_in_target_parent, the adolescent agent may exhibit overt_rebellion against the alienating parent, a dynamic not available to the younger child agent.

The Child as an Active Agent (Feedback Loops):

The child's current belief_in_target_parent score generates behavioral outputs that influence both parents:

When belief drops below -0.3: The child begins to exhibit avoidance behaviors toward the Target Parent (refusing calls, reluctance at transitions). This feeds back into the Target Parent's stress level.
When belief drops below -0.6: The child exhibits overt_rejection (hostile statements, refusal to visit). This significantly increases Target Parent stress and may reinforce the Alienating Parent's behavior (as the campaign appears to be "working").
When belief rises above +0.3 despite alienating pressure: The child may exhibit approach behaviors toward the Target Parent, which can trigger an escalation in the Alienating Parent's behavior (perceived loss of control).

The Alienating Parent Agent

The Alienating_Parent_Agent is defined by a behavioral profile grounded in the empirical taxonomy established by Baker (2005) and Baker and Darnall (2006). Their survey of targeted parents (n=97) found that alienating parents use a mean of 8 or more strategies simultaneously (range 1-21), and that the tactics closely parallel those used by cult leaders (Baker, 2005): requiring excessive devotion, emotional manipulation, denigration, creating the impression of danger, deceiving the child about the other parent's feelings, withdrawal of love as punishment, and erasing memory of the target parent.

Behavioral Lexicon (Alienating Parent):

The behaviors are organized into three categories. Impact weights are derived from the relative severity implied by Baker and Darnall's (2006) taxonomy and the dose-response findings of Forsstrom et al. (2023), where behaviors involving fabricated accusations and visitation sabotage showed the strongest association with severe alienation outcomes.

| Category | Behavior | Impact on Child's belief | Impact on Child's rule_susceptibility_state | Empirical Basis | | :--- | :--- | :--- | :--- | :--- | | Level 1: Undermining | Badmouthing | -0.08 | +0.05 | Baker & Darnall (2006): most common strategy | | | LimitingContact | -0.06 | +0.08 | Forsstrom et al. (2023): visitation sabotage increases gradually with PA degree | | Level 2: Distortion | ReframingMemory | -0.15 | +0.10 | Baker (2005): "erasing memory" cult tactic | | | TellingChildTargetDoesNotLove | -0.18 | +0.12 | Baker (2005): "deceiving child about other parent's feelings" | | | CreatingFalseAccusation | -0.25 | +0.15 | Forsstrom et al. (2023): false accusations show strongest dose-response with PA severity | | Level 3: Coercion | ForcingAThreateningChoice | -0.30 | +0.20 | Baker (2005): "requiring excessive devotion" / loyalty bind | | | WithdrawingLoveAsPunishment | -0.20 | +0.25 | Baker (2005): "withdrawal of love as punishment" cult tactic |

Pervasive Dysregulation Behaviors (BPD-like patterns):

Research consistently identifies alienating parents as presenting with "paranoid, histrionic, or narcissistic personality traits" along with "affective disorders, suicidal ideation, and lack of resilience around separation and loss" (Haines et al., 2020, as cited in Matthewson & Haines, 2022). These behaviors do not directly target the other parent but create an unstable environment that increases the child's vulnerability.

Dynamic Profile (Not Static):

The alienating parent's behavioral profile is not a fixed dice roll. Drawing on Kelly and Johnston's (2001) finding that protracted litigation and humiliating separations are key escalation factors, we introduce external "event triggers" that shift the behavioral distribution:

Baseline period: 40% Dysregulation + 50% Alienating (Levels 1-2 predominant) + 10% Positive
During litigation/court events: 25% Dysregulation + 65% Alienating (Levels 2-3 escalate) + 10% Positive
When child shows approach to Target Parent: 30% Dysregulation + 60% Alienating (Level 3 spikes) + 10% Positive
When child shows overt rejection of Target Parent ("victory" state): 50% Dysregulation + 30% Alienating (reduced, as campaign appears successful) + 20% Positive (idealization phase)

The Target Parent Agent

Previous versions of this model assumed the target parent was a consistent, healthy parent throughout the simulation. This is clinically unrealistic. Research demonstrates that targeted parents undergo significant behavioral degradation under the stress of alienation.

Empirical grounding for the Target Parent's stress response:

Johnston (2003) and Kelly and Johnston (2001) describe targeted parents as presenting with "a history of passivity, emotional constriction, and over-accommodation."
Haines et al. (2020) found that targeted parents "may withdraw from their child after realising they cannot meet the alienating parent's demands."
The Nordic study (Forsstrom et al., 2023, n=1,212) found that 54.8% of targeted fathers reported depressive symptoms (mild to pronounced), and 40.1% reported "bad" well-being.
Harman et al. (2019, n=610) found targeted parents experience significantly elevated depression, PTSD symptoms, and suicidal ideation compared to non-targeted parents.
Four separate studies reported high levels of suicidality among alienated parents (as reviewed in Matthewson & Haines, 2022).

Target Parent Behavioral Profile:

The Target Parent has two profiles that shift based on the child's behavior toward them:

Baseline (Child's belief > -0.3):
- 85% Positive Contingencies (AttentiveListening, SharedEnjoyableActivity, ConsistentAffection, ReliablePresence)
- 15% Negative Contingencies (minor: InconsistentBehavior, occasional frustration)
Under Stress (Child's belief < -0.3, reflecting avoidance/rejection):
- 55% Positive Contingencies (still trying, but with reduced effectiveness)
- 25% Stress Responses (EmotionalWithdrawal, OverAccommodation, AnxiousPursuit)
- 20% Negative Contingencies (ReactiveAnger, InconsistentBehavior, Desperation)

This creates a tragic feedback loop: the child's rejection increases the target parent's stress, which degrades their parenting, which provides the alienating parent with real (if stress-induced) evidence to support their narrative.

The Experiment: Simulating Alienation vs. Estrangement

With the model fully specified, we run the core experiment. The simulation runs for 100 "days" under three conditions:

Condition A (Control): A child interacts with a Target Parent using the Baseline profile. The Alienating Parent agent is inactive. This provides our baseline for healthy development.
Condition B (Estrangement): A child interacts with an Estrangement_Inducing_Parent (80% Negative Contingencies, 20% Positive). The Alienating Parent agent is inactive. We predict the child's belief score will trend negative, directly and rationally tracking the consistently negative experiences.
Condition C (Alienation): The child interacts with a Target Parent (initially Baseline profile) and an Alienating Parent (BPD-like profile). The Target Parent provides positive experiences, but the Alienating Parent provides a steady stream of alienating and dysregulating behaviors. The Target Parent's profile degrades to "Under Stress" as the child's belief drops.

Each condition is run with both a young_child (age 7) and adolescent (age 13) agent parameterization.

Predicted Results

We predict the following outcomes, visualized as graphs produced by the simulation:

Graph 1: Belief Over Time

In the Alienation condition, we predict the child's belief_in_target_parent will decline into negative territory despite the majority of direct interactions being positive. The decline will be steeper for the young_child agent (higher base rule sensitivity) but ultimately reach a lower floor for the adolescent agent (due to identity-based attack vulnerability).

Graph 2: Experience-Belief Correlation

We calculate the Pearson's r correlation between the Target Parent's actual behavior valence and the child's belief score.

Estrangement condition: r > 0.7 (strong positive correlation; belief tracks experience)
Alienation condition: r < 0.2 (near-zero or negative; belief is decoupled from experience)

This decoupling is the quantitative signature of pathological rule-governed behavior.

Graph 3: Target Parent Behavioral Degradation

A new graph showing the Target Parent's behavioral profile shifting over time from "Baseline" to "Under Stress" as the child's rejection increases. This visualizes the tragic feedback loop and demonstrates that the target parent's behavioral changes are a consequence, not a cause, of the alienation process.

Graph 4: Child's Rule Susceptibility State Over Time

Shows how the child's rule_susceptibility_state increases over time in the Alienation condition due to the alienating parent's dysregulating behaviors, and how this increased susceptibility amplifies the effect of subsequent alienating verbal rules.

Limitations and Future Directions

This model has important limitations that must be acknowledged:

Impact weights are derived from relative severity rankings in the literature, not from direct causal measurement. No study has yet quantified the exact numerical impact of a single instance of badmouthing versus a single instance of a forced loyalty conflict. The weights in this model reflect the relative ordering established by dose-response studies (Forsstrom et al., 2023) and clinical severity taxonomies (Baker & Darnall, 2006; Gardner, 1998). A primary purpose of this simulation is to conduct sensitivity analyses to determine how robust the model's predictions are to changes in these values.
The model simulates behavioral patterns, not internal mental states. When we say the alienating parent exhibits "BPD-like" patterns, we are modeling the observable behavioral outputs described in the literature, not diagnosing the agent with a personality disorder.
The three-agent system is deliberately simplified. Future versions may introduce external agents (therapists, courts, extended family) to model protective and exacerbating factors.
Feedback loops are approximations. The specific thresholds at which the child shifts from avoidance to overt rejection, or the target parent shifts from baseline to stressed, will require calibration against longitudinal case data.

Conclusion: A New Tool for Research

This computational model represents a significant step toward empirical, replicable, and falsifiable science in the study of parental alienation. By grounding every parameter in published research and translating the mechanisms of Relational Frame Theory into a simulation, we can generate precise hypotheses about which behavioral patterns are most corrosive, which children are most vulnerable, and which interventions are most likely to re-establish the link between a child's experience and their beliefs.

The next step is to run this experiment.

References

Baker, A. J. L. (2005). The long-term effects of parental alienation on adult children: A qualitative research study. The American Journal of Family Therapy, 33(4), 289-302.

Baker, A. J. L., & Darnall, D. (2006). Behaviors and strategies employed in parental alienation: A survey of parental experiences. Journal of Divorce & Remarriage, 45(1-2), 97-124.

Bentley, C., & Matthewson, M. (2020). The not-forgotten child: Alienated adult children's experience of parental alienation. American Journal of Family Therapy, 48(5), 509-529.

Forsstrom, D., Maehle, E., & Skjorten, K. (2023). Parental alienation: A valid experience? Scandinavian Journal of Public Health, 51(7), 1065-1073.

Gardner, R. A. (1998). The parental alienation syndrome: A guide for mental health and legal professionals (2nd ed.). Creative Therapeutics.

Haines, J., Matthewson, M., & Turnbull, M. (2020). Understanding and managing parental alienation: A guide to assessment and intervention. Routledge.

Harman, J. J., Leder-Elder, S., & Biringen, Z. (2019). Prevalence of adults who are the targets of parental alienating behaviors and their impact. Children and Youth Services Review, 106, 104471.

Harman, J. J., & Biringen, Z. (2016). Prevalence of parental alienation drawn from a representative poll. Children and Youth Services Review, 66, 62-66.

Hayes, S. C., Barnes-Holmes, D., & Roche, B. (2001). Relational frame theory: A post-Skinnerian account of human language and cognition. Kluwer Academic/Plenum.

Johnston, J. R. (2003). Parental alignments and rejection: An empirical study of alienation in children of divorce. Journal of the American Academy of Psychiatry and the Law, 31(2), 158-170.

Kelly, J. B., & Johnston, J. R. (2001). The alienated child: A reformulation of parental alienation syndrome. Family Court Review, 39(3), 249-266.

Matthewson, M., & Haines, J. (2022). The impact of parental alienating behaviours on the mental health of adults alienated in childhood. Children, 9(4), 475.

Verrocchio, M. C., Baker, A. J. L., & Marchetti, D. (2018). Adult report of childhood exposure to parental alienation at different developmental time periods. Journal of Family Therapy, 40(4), 602-618.