Parental Alienation and the Allegation of Pseudoscience

1. Introduction

The breakdown of a couple relationship is one of the most complex contexts for psychological, social and legal intervention. Although many separations develop into relatively stable forms of co-parenting, some are characterised by persistent conflict, prolonged litigation and substantial difficulties in maintaining the child’s relationship with one or both parents. In this context, the literature describes loyalty conflicts, triangulation, control of the other parent’s involvement, resistance or refusal of contact, justified estrangement and different forms of interference in the relationship between the child and a parent (Fidler & Bala, 2020; Kelly & Johnston, 2001).

Among these concepts, parental alienation has become one of the most disputed. Since Richard Gardner’s work in the 1980s, the subject has undergone significant conceptual and empirical development. Contemporary literature is not confined to Parental Alienation Syndrome (PAS). It includes the study of observable behaviours, multifactorial models of resistance or refusal of contact, measurement instruments, expert practices and distinct interventions, each of which must be assessed separately in scientific terms (Doughty et al., 2020; Fidler & Bala, 2010, 2020; Harman et al., 2022; Kelly & Johnston, 2001).

At the same time, legal, political and institutional debate about the use of the concept in court has intensified. A number of authors and organisations describe parental alienation as pseudoscience or, in Alsalem’s (2023) report, as a “pseudo-concept”. These criticisms draw attention to genuine risks, including the dismissal of allegations of violence, the use of unrecognised labels and the application of insufficiently evaluated interventions. The legitimacy of those concerns does not, however, remove the need for an epistemological examination of the field or for a distinction between its different levels.

Classification as pseudoscience does not follow merely from controversy, lack of consensus or empirical limitations. It is a category from the philosophy of science whose application requires analysis of testability, falsifiability, operationalisation, replicability, the cumulative development of knowledge and responsiveness to criticism (Lakatos, 1970; Mahner, 2013; Pigliucci & Boudry, 2013; Popper, 1959; Thagard, 1978).

1.1. Six levels of analysis

To avoid reaching a single conclusion about different objects, this review distinguishes six levels. This distinction is central: a valid methodological objection at one level cannot automatically be transferred to all the others.

The central thesis is deliberately limited. The allegation of pseudoscience has partial merit when directed at the historical formulation of PAS and may be relevant to expert practices or interventions that rely on circular reasoning, unvalidated instruments or measures without sufficient empirical support. It is not sustainable when applied indiscriminately to the study of alienating behaviours, multifactorial models of resistance or refusal of contact, or psychometric research that advances testable hypotheses and remains open to criticism and revision.

The central question is therefore this: at which levels, and under what conditions, is it scientifically defensible to use the classification of pseudoscience? The purpose is not to validate automatically allegations made in court, defend a diagnosis or recommend a particular intervention. It is to distinguish allegation, hypothesis, empirical evidence, consensus and institutional position, while keeping the best interests and rights of the child at the centre.

Back to contents

2. Method and nature of the review

This paper takes the form of a critical narrative review of the literature, supplemented by documentary analysis. It is not presented as a systematic review or meta-analysis, since its principal purpose is not to estimate prevalence, calculate an effect size or quantitatively evaluate a specific intervention. The question is simultaneously empirical, conceptual and epistemological.

The choice of a critical narrative review reflects the heterogeneity of the object under examination. The expression parental alienation may refer to specific behaviours by a father, a mother or third parties, processes of psychological influence, changes in family relationships, a relational condition characterised by unjustified rejection, a diagnostic proposal, an explanatory hypothesis or a category used in legal proceedings. These meanings cannot be assessed through a single type of study or combined into an undifferentiated conclusion.

This review examined quantitative and qualitative empirical studies, psychometric research, systematic and narrative reviews, academic books and chapters in psychology and family law, literature in the philosophy of science and measurement theory, documents issued by professional associations and public bodies, international reports, judicial decisions and relevant advocacy texts needed to understand the controversy. The sources consulted cover literature available up to July 2026 and are not treated as epistemically equivalent.

• An empirical study may test associations or hypotheses, but it does not automatically determine the legal characterisation of conduct.
• A judicial decision may document the use of a concept in proceedings, but it does not validate a psychological construct.
• A professional document may guide practice, but its value depends on the methodology and sources it uses.
• A communication from an advocacy organisation may identify problems and experiences, but it does not replace a representative sample or a systematic review.

Priority was given to primary sources, studies with described methods, publications presenting divergent positions and work capable of clarifying the development of research and debate. Literature supporting the recognition of parental alienation was examined alongside critical literature, including studies concerning allegations of violence, gender bias, instrument validity and the risks of interventions.

A systemic and multi-method forensic psychological assessment should be based on multiple sources of information. Portuguese literature emphasises the need to distinguish data, inferences and conclusions and to integrate interviews, instruments, observation, court records and information from different participants (Agulhas & Anciães, 2017; Matos et al., 2011; Paulino & Alho, 2018; Sani, 2017; Simões et al., 2017; Weiner & Otto, 2013).

A conclusion concerning resistance or refusal of contact cannot be based exclusively on the account of a father or mother, a statement by the child, an isolated instrument or the presence of conflict between the adults. It requires differential assessment of hypotheses, including abuse or neglect, direct domestic violence or violence through an intermediary, high-intensity parental conflict, prior failures in the relationship, child-related factors, alienating behaviours and combinations of several causes.

The review adopts an approach of symmetrical critical scrutiny. The risk of discounting a genuine report of violence must be treated seriously. That risk does not, however, justify excluding in advance the hypothesis of psychological influence, instrumentalisation or obstruction of the relationship. Equally, the possibility of alienating behaviours does not permit the presumption that every rejection is unjustified.

The 2026 version of the Portuguese Psychologists’ Association document Intervenção Psicológica em Situações de Separação/Divórcio [Psychological Intervention in Situations of Separation/Divorce] was also consulted exclusively to identify bibliographical references and analyse technical formulations. Its current status is discussed in section 14.6. The document is not treated as current professional guidance or as autonomous evidence for any scientific conclusion.

The scope of the conclusions must remain limited. This review does not determine whether parental alienation is present in a specific case, validate every allegation made in court or automatically recommend changes of residence, restrictions on contact or coercive interventions. The question is more specific: is there a basis for classifying as globally pseudoscientific the study of alienating behaviours, unjustified rejection and the relational processes grouped under the expression parental alienation?

Back to contents

3. What does it mean to classify something as pseudoscience?

3.1. The demarcation problem

The distinction between science and pseudoscience is traditionally known as the demarcation problem. Although Popper’s criterion of falsifiability is often presented as a definitive solution, philosophy of science offers no single, consensual test capable of automatically classifying every theory, practice and research programme (Laudan, 1983; Mahner, 2013; Pigliucci & Boudry, 2013).

The difficulty does not make the distinction useless. Presenting unsupported claims as scientific knowledge may cause serious harm in health care, education and justice. Yet the indiscriminate use of the label pseudoscience may prematurely exclude legitimate hypotheses and replace methodological examination with an appeal to authority.

An event or behaviour is not, in itself, science or pseudoscience. The classification applies more appropriately to theories, methods, practices, research programmes or systems of claims that assert scientific status. The statement “parental alienation is pseudoscience” may therefore mean different things: that Gardner’s syndrome is invalid; that later models are invalid; that certain instruments do not measure what they claim to measure; that particular interventions lack evidence; or that interfering behaviours themselves do not exist. These propositions are not equivalent.

3.2. Falsifiability and testability

For Popper (1959, 1962), a scientific theory must make claims capable of coming into conflict with experience. A hypothesis that can explain every possible outcome, accommodates all facts retrospectively and never identifies conditions that could weaken it takes no genuine empirical risk.

An assessment of alienation would be epistemically problematic if every behaviour by the child were interpreted as confirmation: rejecting the parent would confirm alienation; accepting contact would confirm concealed alienation; and denying influence would confirm that the child had been effectively conditioned. Such a model would be immune to refutation.

The relevant question is not merely whether an abstract hypothesis can in principle be falsified, but whether researchers specify expected findings, observations inconsistent with the hypothesis, alternative explanations, measurement methods and the circumstances in which a conclusion should be revised.

3.3. Research programmes and the development of knowledge

Lakatos (1970) proposed assessing research programmes over time. A progressive programme generates new predictions, responds to anomalies and refines its concepts. A degenerating programme protects its hard core through ad hoc adjustments, meaning explanations added solely to avoid refutation without generating empirically fruitful new knowledge.

This perspective is particularly appropriate here. Assessment cannot remain confined to Gardner’s writings. It must consider whether later research remained attached to his original formulation or developed different models, exclusion criteria, instruments, validity studies, multifactorial approaches and differential assessment of hypotheses.

Disagreement among Gardner, Kelly and Johnston, Warshak, Baker, Bernet, Fidler, Bala, Harman, Saini and critical authors does not establish the validity of any position. It does show, however, that the field is not a single immutable doctrine protected from debate.

3.4. Progress, critical community and responsiveness to evidence

Thagard (1978) added the ideas of comparative progress and a critical community. A theory becomes particularly problematic when it progresses less than competing explanations and when its proponents show little commitment to resolving empirical problems or comparing hypotheses.

Signs commonly associated with pseudoscience include immunisation against negative findings, excessive dependence on testimony, selective use of favourable evidence, failure to compare alternative hypotheses, vague concepts, rejection of independent scrutiny and treating criticism as evidence of persecution. These signs must nevertheless be distinguished from poor science. Error, inadequate samples and failures to replicate also occur in recognised scientific disciplines.

Science is not defined by the absence of error but by the presence of public, criticisable and corrigible procedures. The Reproducibility Project: Psychology demonstrated substantial replication difficulties in studies published by established journals, without implying that psychology as a whole was pseudoscientific (Open Science Collaboration, 2015).

3.5. A multi-criterion application to the six levels

The boundary between science and pseudoscience should not depend on a single test. The following table applies ten dimensions to the six levels identified above. The assessments “low”, “partial”, “substantial” and “variable” are analytical judgements, not quantitative scores or meta-analytic findings.

Level 1 displays the greatest weaknesses. Historical PAS relied chiefly on clinical observations, contested definitions and an insufficiently tested causal structure, and did not obtain recognition as an autonomous diagnostic category. The criticism of pseudoscience may therefore legitimately reach some of the stronger claims associated with the historical syndrome.

At levels 2, 3 and 4 there is empirical research, peer-reviewed publication, critical debate and conceptual reformulation. The evidence remains uneven and incomplete, but these characteristics are more compatible with a developing scientific research programme than with a doctrine closed to refutation (Doughty et al., 2020; Harman et al., 2022; Marques et al., 2020).

At levels 5 and 6, the assessment depends on the actual practice. A systemic, multi-method and transparent expert assessment does not have the same status as a conclusion based on a label or checklist. Similarly, the existence of a relational problem does not automatically validate a particular reintegration programme. Each intervention requires its own evidence of effectiveness, safety and adverse effects (Mercer, 2022a, 2022b; Warshak, 2020).

Back to contents

4. Constructs and inference in forensic psychology

4.1. Observable phenomena and latent constructs

Many of the objects studied in psychology are not directly observable. Intelligence, anxiety, attachment, hostility, coercive control, trauma, emotional abuse and loyalty conflict are constructs inferred from indicators. The expert observes responses, patterns of behaviour, accounts, relationships and consequences which, together with theoretical knowledge, support reasoned inferences.

The inferential nature of a construct does not make it pseudoscientific. The relevant question is whether the inference rests on clear definitions, theoretically coherent relationships, replicable observations and the ability to distinguish the construct from neighbouring phenomena.

Cronbach and Meehl (1955) developed the concept of construct validity precisely for attributes that cannot be observed directly. Validation requires the construct to be located within a network of theoretical and empirical relationships and the formulation of predictions about how it should relate to other variables.

The question “does parental alienation exist?” must therefore be broken down. What phenomenon is being designated? Which indicators are used? How is it distinguished from conflict, abuse or justified rejection? Which findings would be inconsistent with the model? How consistently do independent assessors reach the same conclusion?

4.2. Validity is not a binary property

In contemporary psychometrics, validity is not a permanent seal placed on a test. It concerns the degree to which evidence and theory support interpretations made from scores and the uses to which those interpretations are put (American Educational Research Association et al., 2014; Messick, 1995). An instrument may be suitable for exploratory research yet unsuitable as the basis for changing a child’s residence.

Cronbach’s alpha is a statistical coefficient indicating the extent to which items in a scale yield internally consistent responses. On its own, it does not establish that the scale measures the intended phenomenon, is unidimensional, or accurately distinguishes between different situations.

4.3. Specific requirements of forensic assessment

Forensic psychological assessment differs from clinical assessment in its audience, context and consequences. The person being assessed may not be seeking help, may have incentives to omit or exaggerate information and knows that the report may affect a judicial decision. The expert must distinguish data, reported information, inference, hypothesis and conclusion.

In proceedings concerning the exercise of parental responsibility, a child’s refusal may arise from justified fear, exposure to violence, harmful parental conduct, loyalty conflict, intentional or unintentional influence, developmental factors, difficulties in moving between homes, or combinations of these elements.

Agulhas and Anciães (2017) advocate a systemic assessment that integrates both parents, the child, family interactions, court documents and external sources. Inability to assess the whole system is a limitation that must be stated and reflected in the degree of confidence attached to the opinion.

This approach is incompatible with automatically inferring alienation from a checklist. A single indicator is rarely specific. The use of adult language, for example, may reflect influence, but it may also arise from real experiences, repetition of terminology or the child’s attempt to organise a complex situation.

4.4. Construct, diagnosis and relational problem

Another recurring error is to presume that a phenomenon can be studied scientifically only if it is recognised as a medical diagnosis. A psychological construct, behavioural pattern or relational problem does not need to constitute a disease in order to be investigated.

Parental alienation may be examined as a set of interfering behaviours, a process of influence, a relational pattern, an outcome expressed through rejection, an explanatory hypothesis or a diagnostic proposal. Evidence supporting one of these levels does not automatically validate the others.

Back to contents

5. Historical development of the concept

5.1. The phenomenon did not begin with Richard Gardner

Descriptions of children becoming intensely aligned with one parent and rejecting the other pre-date the expression Parental Alienation Syndrome. Research on the effects of divorce and on family systems had already described alliances, triangulation, loyalty conflicts and hostility towards a parent following separation (Johnston et al., 1985; Minuchin, 1974; Wallerstein & Kelly, 1976; Westman et al., 1970).

These earlier descriptions do not show that those authors used Gardner’s criteria or causal explanation. They do show that the relational patterns later associated with parental alienation were not invented by one author.

5.2. The formulation of Parental Alienation Syndrome

Gardner introduced the expression Parental Alienation Syndrome in 1985, drawing on his experience in disputes concerning residence and contact. He described a pattern in which the child participated in a campaign of denigration against a parent without sufficient justification, arising from a combination of influence by the other parent and the child’s own contributions (Gardner, 1985, 1987).

The formulation included manifestations such as a campaign of denigration, weak reasons for the rejection, lack of ambivalence, automatic support for the parent with whom the child was aligned, absence of guilt and extension of the rejection to the wider family. Gardner maintained that the label should not be applied where abuse, neglect or conduct by the rejected parent was sufficiently serious to explain the refusal (Gardner, 1998, 2002).

The limitations are important. The proposal relied mainly on clinical observations, applied the medical concept of a syndrome to a complex family dynamic and lacked validation studies equivalent to those now required. Some of Gardner’s positions on gender, sexuality and child sexual abuse are also the subject of justified criticism.

5.3. Kelly and Johnston’s reformulation

Kelly and Johnston (2001) moved away from a syndrome presumed to be caused by an alienating parent and shifted the analysis to the child’s situation and the family system. They proposed a continuum distinguishing positive attachment to both parents, preference for one, intense alignment, justified estrangement and alienation characterised by persistent rejection without sufficient justification or clearly disproportionate to the child’s actual experiences.

The model incorporated characteristics of the child, the conduct of the different parents, family vulnerabilities, conflict, new relationships, siblings, the extended family and the judicial context. Deliberate intention to manipulate ceased to be a necessary condition.

Fidler and Bala (2010, 2020) developed this approach and emphasised the need to distinguish affinity, alignment, justified estrangement, alienation and hybrid cases. They recognised that some allegations of alienation are unfounded and that, in other cases, the conduct of parents contributes to the situation.

5.4. Dynamics of resistance and refusal of contact

Later work increasingly used more descriptive and less diagnostic language. Garber (2007) situated resistance or refusal of contact within the wider context of separation. Friedlander and Walters (2010) argued that interventions should be tailored to the specific causes of the problem.

This development consolidated the notion of hybrid cases, in which influencing conduct coexists with genuine reasons for the child’s dissatisfaction. The rejected parent may have weaknesses or may have contributed to the conflict without those behaviours explaining a total and enduring rejection.

Systemic literature also describes relational fusion (enmeshment), characterised by diffuse psychological boundaries between the child and one parent. The child may adopt the adult’s emotions and interpretations as their own without any need to establish a conscious strategy of manipulation (Friedlander & Walters, 2010).

5.5. Alienating behaviours as an empirical object

Another line of research moved away from seeking a global syndrome and began to study specific behaviours: conveying to the child a persistently negative image of the other parent, restricting communication without sufficient justification, making the child feel that the other parent does not love them, asking the child to choose, using the child as a messenger or confidant and obstructing access to educational or clinical information (Baker, 2005; Baker & Darnall, 2007).

The study of observable behaviour has a methodological advantage: it does not require prior acceptance of a syndrome. It is possible to investigate whether a particular adult interferes with the child’s relationship with the other parent without deciding in advance what effect occurred or what caused the rejection.

However, the existence of potentially alienating behaviour does not, by itself, demonstrate that parental alienation occurred. Frequency, duration, context, intention, impact and alternative explanations must be considered.

5.6. The five-factor model

The so-called five-factor model was systematised to differentiate parental alienation from other forms of rejection (Bernet, 2020; Bernet & Greenhill, 2022). It considers: resistance or refusal of contact; a previous positive or sufficiently functional relationship; the absence of abuse, neglect or conduct by the rejected parent sufficiently serious to explain the refusal; the presence of multiple alienating behaviours; and associated manifestations in the child.

Its intended primary function is to prevent rejection alone from being treated as proof of alienation. The word factors does not mean that the model necessarily results from factor analysis, the statistical technique used to determine whether different items cluster into common dimensions. It is primarily a clinical and forensic decision-making framework.

Studies such as Morrison and Ring (2023) have reported favourable results regarding consistent application of the criteria in brief descriptions of hypothetical cases. Garber and Simon (2023), however, criticised the model for potentially encouraging a binary division, producing anchoring bias and prematurely narrowing the inquiry.

Back to contents

6. Essential conceptual distinctions

6.1. Parental Alienation Syndrome

Parental Alienation Syndrome corresponds to Gardner’s specific formulation. It includes a set of manifestations attributed to the child and a causal hypothesis relating to the influence of one parent. Criticising the existence of an autonomous syndrome, its medical label or its original criteria does not automatically resolve the question of alienating behaviours or unjustified rejection.

6.2. Alienating behaviours

Alienating behaviours are acts or omissions capable of harming the child’s relationship with another parental figure. They may be conscious or unconscious, intentional or reactive, persistent or episodic. A behaviour does not necessarily determine an outcome.

6.3. Parental alienation as a relational outcome

At this level, the expression describes a serious deterioration in the relationship between the child and one parent, marked by persistent rejection without sufficient justification or clearly disproportionate to the child’s actual experiences. The outcome may be influenced by multiple factors and does not automatically imply that a single parent deliberately caused the situation.

6.4. Resistance or refusal of contact

Resistance or refusal of contact is a behavioural description. It does not, by itself, contain a causal explanation. Refusal may arise from abuse, fear, harmful parenting behaviour, anxiety, loyalty conflict, influence, logistical difficulties, developmental factors or combinations of these elements.

6.5. Justified estrangement

Justified estrangement occurs where resistance is proportionate to significant negative experiences, such as violence, abuse, neglect or conduct by a parent sufficiently serious to justify rejection. The difficulty is that proportionality is not always self-evident.

6.6. Hybrid cases

Hybrid cases combine legitimate reasons for dissatisfaction with behaviours that amplify the rejection. Their existence prevents a simplistic opposition between alienation and abuse. Hypotheses may coexist, and the contributions of different people are not necessarily of equal seriousness.

Back to contents

7. Terminology adopted

The expression unjustified rejection is used as a shortened category. Where genuine reasons for dissatisfaction exist but are insufficient to explain the intensity, duration or extent of the refusal, the formulation rejection clearly disproportionate to the child’s actual experiences is used.

Back to contents

8. Current state of empirical research

8.1. Size and nature of the body of research

The claim that there is no empirical research on parental alienation does not correspond to the current state of the literature. Marques et al. (2020) identified 43 empirical studies published in peer-reviewed scientific journals between 2000 and 2018. The existence of dozens of studies does not demonstrate that all are robust or that the construct has been definitively validated. It demonstrates that the debate can no longer be described as an opposition between a theory without research and criticism grounded in science.

Most of the studies analysed used quantitative, retrospective and cross-sectional designs. A retrospective study asks participants to recall earlier experiences; a cross-sectional study collects data during a single period. These designs can identify associations but have important limitations in determining cause and effect.

8.2. Main areas investigated

The literature covers instrument development, identification of behaviours, family patterns, differentiation between alienation and justified estrangement, psychological consequences, links with emotional abuse, professional perceptions and use in judicial proceedings (Marques et al., 2020).

The research is fragmented and methodologically uneven. Many studies depend on the account of a single father or mother, or on adults reconstructing their childhood. Others use selected samples drawn from associations, services or forensic settings. Conceptual heterogeneity makes comparison difficult because different studies use the same expression to measure different realities.

8.3. Alienating behaviours as an observable object

A significant part of the research examines specific behaviours: persistent denigration of the other mother or father, unjustified restriction of communication, creation of loyalty conflicts, use of the child as a messenger or source of information and interference with access to school or clinical information.

Identifying these behaviours does not resolve the causal question. The existence of the behaviour, the existence of rejection and evidence that the former contributed significantly to the latter must be distinguished.

Rowen and Emery (2019) found that denigration of the other mother or father did not necessarily produce greater closeness to the person engaging in it. In some circumstances, it could produce a boomerang effect and damage the child’s relationship with the father or mother who conveyed the negative image. The finding contradicts automatic causal accounts and demonstrates that effects depend on context.

8.4. Psychological and emotional consequences

Miralles et al. (2023) conducted a systematic review of long-term emotional consequences in adults who reported childhood exposure to parental interfering behaviours. The studies found associations with depressive and anxious symptoms, lower self-esteem, attachment difficulties, feelings of loss, abandonment and guilt, and problems in family relationships.

The conclusions require caution. The studies were predominantly retrospective and did not follow the same individuals from childhood. The review supports associations, but not a single causal relationship. It also remains difficult to separate the specific effects of alienating behaviours from the general effects of severe parental conflict, violence and family instability.

8.5. Prevalence

Prevalence estimates must be treated with particular caution. Some studies ask fathers and mothers whether they consider themselves targets of alienation; others measure behaviours; others analyse court proceedings or children who refuse contact. These measures are not interchangeable.

Until a broadly shared operational definition is applied to representative samples and accompanied by procedures for differential assessment of hypotheses, it is not possible to present a credible universal rate. Uncertainty does not mean zero prevalence, but it precludes categorical figures.

8.6. Studies of judicial proceedings

Studies of judicial decisions may provide information about who makes allegations, how courts respond, the relationship with reports of violence and possible gender differences. A judicial decision does not, however, constitute an independent clinical observation, and published proceedings represent only a proportion of cases.

Research by Meier and colleagues highlighted the possibility that allegations of alienation may adversely influence the assessment of reports of violence. Harman and Lorandos subsequently challenged the construction of the datasets, the coding and the inferences. The controversy is examined below.

8.7. Overall assessment

The research allows two extreme positions to be rejected. It is incorrect to claim that contemporary research is limited to Gardner’s writings and contains no studies. It is also incorrect to claim that the number of publications demonstrates definitive validation. There is a real and growing empirical body of research, marked by convenience samples, retrospective designs, conceptual heterogeneity, few longitudinal studies and difficulties in separating alienation, conflict, violence and justified estrangement.

Back to contents

9. Validity, reliability and limits of measurement

9.1. Reliability is not validity

Reliability concerns whether a procedure produces sufficiently consistent results. Validity concerns whether it correctly identifies what it claims to measure. Several assessors may agree and still all be wrong.

A thermometer that consistently reads two degrees above the actual temperature may be reliable because its results are stable, but it is not valid because the readings do not correspond to the true temperature.

9.2. The Morrison and Ring study

Morrison and Ring (2023) examined the reliability of the five-factor model using six brief descriptions of hypothetical cases, often termed vignettes in psychological research. They reported high consistency in the application of the criteria, including high statistical coefficients.

In this context, the intraclass correlation coefficient measures the similarity of different ratings. Cronbach’s alpha measures the consistency of responses across elements of the structure. The results provide evidence of reliability under the specific conditions of the study.

They do not establish that the model correctly identifies real cases. Hypothetical descriptions are simpler and more controlled than actual families, where accounts conflict, information is incomplete and violence, harmful parental conduct and third-party influence may coexist.

Nor do the findings establish sensitivity, specificity, positive predictive value or negative predictive value. Sensitivity is the ability to identify correctly cases in which the phenomenon is present; specificity is the ability to exclude correctly cases in which it is absent. Positive predictive value is the proportion of positive classifications that truly correspond to the phenomenon, while negative predictive value is the proportion of negative classifications that are correctly excluded. These values also depend on the frequency of the phenomenon in the population assessed.

9.3. The five-factor model as an investigative framework

The model may be useful as an organising framework because it requires consideration of the previous relationship, the reasons for rejection, interfering behaviours and the child’s manifestations (Bernet & Greenhill, 2022). It should not operate as an automatic test in which five affirmative answers produce a definitive conclusion.

Garber and Simon (2023) warn of confirmation bias, the tendency to search primarily for information supporting an initial hypothesis. This criticism does not make the framework useless, but it shows that it should generate questions rather than replace a systemic assessment.

9.4. Questionnaires and self-report

Self-report instruments allow information to be gathered about experiences that may not have been recorded by professionals or courts. They nevertheless have limitations associated with memory, retrospective interpretation, participant selection and the contextual meaning of conduct.

A scale may show good internal consistency while failing to distinguish adequately among manipulation, parental conflict, protection from risk and harmful conduct by a parent. Such instruments are generally better suited to research on groups than to determining, in isolation, what occurred in a particular family.

9.5. Need for multiple sources

Forensic assessment should not depend on a single instrument, interview or account. It should integrate interviews with parents, hearing and assessment of the child, observation of interactions, school and clinical information, court documents, the history of the relationship, allegations of violence and information from independent professionals.

Comparison of data obtained through different methods and people is often called triangulation of information. A conclusion becomes stronger when independent sources converge and more uncertain when the data conflict or come from only one person.

Back to contents

10. Areas of convergence and divergence

10.1. Areas of convergence

Despite the controversy, there is meaningful convergence around several propositions:

• Contact difficulties are multifactorial and may arise from violence, abuse, neglect, harmful parenting behaviour, anxiety, loyalty conflict, influence or combinations of several causes.
• Not every rejection is alienation. There are justified rejections, unjustified rejections and hybrid cases.
• Adult behaviour can influence the child’s relationships, although the frequency, causation and magnitude of those effects remain disputed.
• Violence and abuse must be investigated and cannot be excluded because alienation has been alleged.
• False positives and false negatives are both possible and may have serious consequences.
• Current research has limitations and requires better instruments, longitudinal studies, replication and evaluation of interventions.

10.2. Areas of divergence

Disagreement remains regarding terminology, diagnostic status, the validity of criteria, prevalence, gender distribution, interventions and the relationship with domestic violence.

Some authors consider parental alienation a useful expression. Others argue that the term introduces a premature causal conclusion and prefer contact problems between the child and a parent or dynamics of resistance and refusal. The controversy is particularly intense in relation to changes of residence, suspension of contact and intensive reunification programmes.

Back to contents

11. False positives and false negatives

11.1. Two classification errors

Assessment of resistance or refusal of contact takes place under conditions of uncertainty. The hypothesis of alienating behaviour and the hypotheses of abuse, neglect or violence must neither be presumed nor excluded at the outset. The following matrix sets out the four possible outcomes.

11.2. Asymmetric costs of error

The costs of the two errors are not identical in every case. A false positive may minimise violence or abuse, discredit the child or the parent seeking to protect them and lead to the imposition of unsafe contact. Where there are credible indications of immediate danger, protection and risk assessment must take priority.

A false negative may entrench unjustified rejection, prolong the instrumental use of the child, intensify loyalty conflicts and make recovery of a significant relationship progressively more difficult. In this quadrant, procedural delay may turn a potentially reversible situation into a prolonged breakdown.

11.3. Risk management and proportionality

Reducing both errors simultaneously requires systemic and multi-method forensic psychological assessment, specific investigation of violence, explicit comparison of hypotheses, multiple sources of information, language proportionate to the degree of certainty and revision of conclusions when new evidence emerges (Deutsch et al., 2020; Drozd & Olesen, 2004; Fidler & Bala, 2020).

The decision should consider the nature and seriousness of the risk, the likelihood of occurrence, the reversibility of consequences, the child’s age and development, the history of the relationship, the possibility of safe contact and the quality of the available information. Methodological symmetry does not mean that the risks are materially equal in every case; it means that no hypothesis is ruled out as a matter of principle.

Back to contents

12. Differential assessment of competing hypotheses

12.1. From description to explanation

The most cautious wording at the outset of an assessment is a contact problem between the child and a parent or resistance or refusal of contact. These formulations describe the situation without anticipating its cause.

The initial question should be: why is this child resisting or refusing contact with this parent, in this family context and at this point in time? Garber and Simon (2023) argue that binary questions may prematurely narrow the assessment and propose examination of the child’s relational ecology.

12.2. A guiding sequence for the differential assessment of competing hypotheses

The following sequence, inspired by the decision trees of Drozd and Olesen (2004), the differential approach of Fidler and Bala (2020) and the literature on parenting-plan evaluation (Saini et al., 2016), organises the analysis according to the priority of safety and explanatory parsimony. It is neither a rigid order nor an automatic procedure: hypotheses may coexist, be examined in parallel and be revised throughout the assessment.

1. Investigate allegations and indicators of abuse, neglect or exposure to danger.
   Are there accounts, indicators, documents or observations consistent with physical, sexual or psychological abuse, neglect or exposure of the child to danger? The absence of a criminal conviction is not, by itself, equivalent to the absence of risk.
2. Investigate direct domestic violence or violence through an intermediary.
   Are there patterns of intimidation, surveillance, threats, instrumental use of the child or coercive control that may explain fear, refusal or protective action by a parent?
3. Examine high-intensity parental conflict without demonstrated instrumentalisation.
   Is the child exposed to hostility, destructive communication, difficult transitions or prolonged litigation that may explain the resistance without a specific pattern of influence?
4. Reconstruct the history of the relationship.
   Were there attachment difficulties, prolonged absence, harmful parenting practices, emotional unavailability or other problems before separation that wholly or partly explain the estrangement?
5. Assess factors relating to the child.
   May age, temperament, development, mental health, separation anxiety, special needs, sibling relationships, school or routines contribute to the resistance or refusal?
6. Gather information from multiple sources and reconstruct the chronology.
   Do the accounts of the parents, child and third parties, clinical, educational, social and court records, and observations of interaction converge? When did the resistance and relevant behaviours begin, and how did they develop?
7. Investigate possible alienating behaviours.
   Once priority has been given to assessing safety risks, persistent patterns of denigration, induction of fear without sufficient justification, emotional conditioning, obstruction of information or loyalty pressure should also be investigated, without presuming causation and without treating the remaining hypotheses as excluded.
8. Consider hybrid cases.
   May genuine reasons for dissatisfaction coexist with behaviours that amplify, rigidify or prolong rejection beyond what the child’s experiences would explain?
9. Distinguish data, impact and causal inference.
   Which behaviours were observed or corroborated, what effects do they have on the child, and to what extent does the available material permit causation to be attributed to one person, a dynamic or the interaction of several factors?
10. State the degree of support for each hypothesis.
    Where the information does not allow hypotheses to be distinguished safely, the report should identify the uncertainty, missing material, competing explanations and the evidence that could alter the conclusion.

12.3. Competing hypotheses and continuous revision

The assessment should specify which elements support and which weaken each hypothesis. Seeking contradictory information reduces confirmation bias. Whenever new material emerges, the formulation should be revised, including the possibility that violence, conflict, relational failures and alienating behaviours coexist (Deutsch et al., 2020; Garber & Simon, 2023).

12.4. Assessment of the whole family system

Assessing the child alone, or only one parent, rarely permits an understanding of family dynamics. The assessment should examine the history of the relationship, the onset of resistance, behaviour before and after separation, accounts given in different settings, educational, clinical and social information, interactions, each parent’s capacity to support the child’s relationship with the other parent, violence, fear, the extended family and the impact of judicial proceedings.

No source is infallible. Parents may omit information or interpret it selectively; children may combine experience, influence and a need for protection; professionals are also subject to bias.

12.5. Hearing the child

The child’s views are essential but should not be confused with an automatic decision. Hearing means understanding what the child feels, has experienced and fears; how the child represents each parent; what pressures the child identifies; how beliefs were formed; and whether ambivalent feelings can be expressed.

The child does not need developed verbal language in order to be heard. Young children communicate through behaviour, play, drawing, bodily reactions and the quality of interactions. Assessment should be adapted to age and forms of communication.

Hearing the child should not require the child to choose between parents. Its purpose is to understand the child’s experience, not to transfer responsibility for resolving adult conflict to the child.

12.6. Indicators are not standalone proof

Adult expressions, lack of ambivalence, rejection of relatives, intense refusal, repeated accusations or absence of guilt may be relevant, but none of these elements, in isolation, permits a conclusion that alienation exists. They must be interpreted in context, over time and against alternative hypotheses.

12.7. Separate assessment and intervention

Even where alienating behaviours are identified, no particular intervention follows automatically. Decisions about contact, residence, support or protective measures should consider safety, age, the previous relationship, current resistance, capacity for change, available resources and evidence of effectiveness. Identifying the problem and selecting the response are distinct scientific and legal stages.

Back to contents

13. Parental alienation, emotional abuse, coercive control and violence through an intermediary

13.1. Three levels that should not be confused

The relationship between alienating behaviours and violence should be analysed in degrees. Not every interference, negative comment or conflict constitutes domestic violence. Persistent and serious patterns that place the child in loyalty conflicts, unjustifiably destroy a significant relationship or convey false representations of abandonment and danger may amount to psychological maltreatment. Where the child is used to control, intimidate or punish the other parent, the conduct may form part of violence through an intermediary or coercive control.

13.2. Emotional abuse of the child

Emotional abuse may include rejection, humiliation, threats, isolation, instrumentalisation or distortion of the child’s emotional reality. Some behaviours studied in the alienation literature overlap with these categories: falsely leading the child to believe that they were abandoned, making affection conditional on rejecting the other parent, creating fear without sufficient justification, punishing displays of affection and involving the child in legal conflict.

Childress (2015, 2016) conceptualised these dynamics as forms of psychological abuse of the child and domestic violence exercised through the parental relationship. This is a theoretical model published mainly in books and professional essays by the author, without robust independent empirical support for the full set of propositions. It should therefore be retained as a historical and theoretical reference, not as a source of authority equivalent to peer-reviewed research.

Harman et al. (2018) systematised the hypothesis in a review article published in Psychological Bulletin, comparing alienating behaviours with categories from research on psychological abuse and family violence. Mercer (2022a, 2022b) and Milchman et al. (2020) challenged the strength of the analogy, causal attribution and the sufficiency of evidence concerning specific harms. The possibility that particular patterns amount to emotional abuse must therefore be distinguished from the broader proposition that all alienating behaviours are, by definition, family violence.

13.3. Violence through an intermediary and the scope of the expression “vicarious violence”

The expression vicarious violence has acquired particular prominence in Spain and in the Ibero-American sphere. In several legal and institutional policy definitions, it refers to violence by a man against children or other people in order to cause suffering or harm to a woman. This delimitation corresponds to the scope of policies addressing gender-based violence.

A gender-delimited legal definition does not, however, demonstrate that the psychological mechanism of using a child to harm another person can occur only in that direction. Where the purpose is to describe the mechanism generally, the more precise expressions are violence through an intermediary or instrumental use of the child to harm the other parent.

13.4. Violence through an intermediary and coercive control

Violence through an intermediary and coercive control are not alternative concepts. Coercive control describes a continuing pattern of domination, intimidation, surveillance, isolation and restriction of autonomy. Use of the child may be one strategy within that pattern.

For example, a parent may threaten to prevent contact, use the child for surveillance, conceal information, make contact conditional on demands or induce fear. The child is the instrument; coercive control is the wider pattern of power and domination.

The overlap with parental alienation is only partial. Instrumentalisation may occur without producing rejection, unjustified rejection may develop without a clear intention to punish, coercive control may occur without use of the child, and emotional abuse may occur without violence directed at the other parent.

13.5. The risk of circularity

A methodological risk arises where every alienating behaviour is defined in advance as abuse and the presence of that abuse is then used as proof of alienation. To avoid circularity, the analysis should begin with observable behaviour, frequency, context, impact, alternative explanations and any pattern of control. Only then should psychological or legal characterisation be considered.

13.6. Violence and alienation may coexist

Public presentation of the debate tends to create an opposition between domestic violence and parental alienation. Several configurations are nevertheless possible: justified rejection of a parent who behaves violently; use of an allegation of alienation by an abuser; instrumental use of the child to control the other parent; or coexistence of violence, harmful parenting behaviour and relational influence.

The notion of a hybrid case should not be used to distribute responsibility automatically between both parents. The existence of conduct by both does not imply equal seriousness, origin or impact.

13.7. Psychological and legal characterisation

The expressions emotional abuse, family violence and coercive control used in psychology do not automatically correspond to the elements of a criminal offence. Any criminal-law characterisation depends on the applicable legislation, the conduct established, seriousness, repetition, the relationship between the people concerned and the evidence produced.

Back to contents

14. The contribution of Portuguese literature

14.1. A non-monolithic national debate

Portuguese literature does not present a single position on parental alienation. The volume published by the Centre for Judicial Studies brings together divergent legal, psychological, psychiatric and sociological perspectives and expressly states that the texts represent their respective authors rather than a uniform scientific position of the institution (Centro de Estudos Judiciários [CEJ], 2018).

For analytical purposes, three principal strands may be distinguished, without exhausting the debate: a legal-critical strand, centred on the risks of PAS and its use in proceedings involving allegations of violence; a forensic-psychological strand, centred on systemic assessment of family dynamics, the formulation of hypotheses and forensic caution; and an ecological-systemic intervention strand, oriented towards constructive transformation of conflict, co-parenting and coordination among professionals.

14.2. The legal-critical strand

Sottomayor (2011, 2018) questions the scientific status of Gardner’s PAS and warns of the risk that its use may minimise the child’s voice, allegations of domestic violence or sexual abuse, and protective behaviour. These criticisms are relevant to level 1 and to forensic practices that turn a label into a causal presumption or automatic measure.

The scope of those works is predominantly legal and critical of the judicial application of PAS. They are not empirical studies designed to test the existence of interfering behaviours, loyalty conflicts or rejection lacking sufficient justification. Criticism of the syndrome and its misuse should therefore not be confused with proof that all the relational dynamics concerned are non-existent.

14.3. The forensic-psychological strand

Agulhas and Anciães (2014, 2017, 2023) place resistance or refusal of contact on a relational continuum and consider multiple explanations, including genuine negative experiences, changes in attachment, triangulation, loyalty conflicts and highly litigious separations. Their approach rejects automatic conclusions and requires assessment of the family system, interactions, documentation and different sources of information.

Sani (2017) emphasises that psychological expert assessments in interparental conflict provide technical assistance to courts and require a comprehensive analysis of family dynamics, relationships and attachments, combining theoretical knowledge, technical competence and the best interests of the child. More generally, Simões et al. (2017) and Paulino and Alho (2018) reinforce the requirements of appropriate instrument selection, reasoned inference and clear definition of the scope of the expert report.

Marques et al. (2020) add an empirical contribution by identifying and describing published research while simultaneously recognising the existence of a body of research and its methodological limitations. This intermediate position is particularly relevant: it does not turn the quantity of studies into definitive validation, but neither does it reduce contemporary literature to Gardner’s original formulation.

14.4. Ecological-systemic intervention in parental conflict

In the field of psychological intervention, Delgado-Martins (2017) proposed an ecological-constructivist model for cases of high-conflict divorce and parental alienation. She advocates a systemic and contextualised intervention based on systematic, naturalistic observation of family interactions in everyday settings, the involvement of parents, interdisciplinary coordination, continued follow-up and the promotion of more functional co-parenting.

This contribution is relevant for two reasons: it shifts attention away from an exclusively individual or diagnostic reading towards relational dynamics and the contexts in which behaviour occurs; and it supports a constructive and transformative approach in which parents are involved, whenever the safety of the child and those concerned permits, without starting from a prior and definitive allocation of the roles of alienating parent and alienated parent, and are made jointly responsible for seeking solutions to the conflict.

It is an intervention method informed by the literature and developed from the author’s professional experience. Its specific effectiveness was not evaluated in that article through a controlled empirical study.

Some of its formulations, in particular the general classification of parental alienation as domestic violence and the accumulation by the same professional of therapeutic, mediating and potentially forensic roles, also require critical analysis in the light of differential assessment, professional impartiality and role separation.

14.5. National convergences

Despite the disagreements, there is growing convergence on the need for systemic and multi-method forensic psychological assessment, appropriate hearing of the child, investigation of allegations of violence, rejection of automatic labels, declaration of limitations and separation between psychological conclusion and legal decision (Agulhas & Anciães, 2017; Sani, 2017; Simões et al., 2017).

14.6. Document of the Portuguese Psychologists’ Association

In May 2026, the Portuguese Psychologists’ Association (Ordem dos Psicólogos Portugueses, OPP) published the professional-practice support document Intervenção Psicológica em Situações de Separação/Divórcio [Psychological Intervention in Separation/Divorce Situations]. Its contents included section 1.4, entitled “The case of parental alienation: What happens when the children and young people in the family are involved in marital conflict?”

The version then available was consulted during preparation of this article, particularly to identify bibliographical references and technical formulations. At the date of publication of the present work, the file was no longer available on the website of the Portuguese Psychologists’ Association. To the best of our knowledge, the document is currently under review.

For that reason, the document is not presented as professional guidance currently in force or as independent proof of any scientific conclusion. The sources it cites are considered individually, according to the nature, methodology and quality of each publication.

Back to contents

15. Critical analysis of the report of the United Nations Special Rapporteur

15.1. Material legitimacy of the concerns

Alsalem’s report (2023) examines decisions concerning children’s residence and contact in contexts involving violence against women and children. It identifies risks that must be recognised: the use of alienation allegations to minimise violence, continuation of control through judicial proceedings, insufficient professional training, inadequate hearing of children and the imposition of coercive interventions without sufficient evidence of safety and effectiveness.

These concerns are materially legitimate. An allegation of alienation should not suspend investigation of violence, convert protective action into automatic evidence of manipulation or lower the evidential burden required for measures that profoundly affect the child. GREVIO’s guidance for Portugal reiterates the need for screening and risk assessment in proceedings concerning residence and contact and rejects the use of so-called PAS in proceedings with a history of domestic violence (GREVIO, 2025).

15.2. Political and institutional value and epistemological value

The report is an official document of a United Nations special procedure and has political and institutional relevance to the protection of human rights. It is not a systematic review, a meta-analysis, clinical guidance or a consensus statement by scientific organisations. The document’s institutional authority does not remove the need to justify, according to epistemological criteria, classification of a domain as pseudoscience.

Testimony and submissions from organisations are appropriate for identifying patterns of concern, possible violations and victims’ experiences. Without additional methodology, they do not permit estimates of prevalence, establishment of causation or exhaustive assessment of the scientific literature.

15.3. Methodological limitations

The report does not present a structured literature-search strategy, the databases consulted, inclusion and exclusion criteria, assessment of study quality, analysis of risk of bias or a procedure for resolving contradictory findings. Nor does it formally apply the criteria of the demarcation problem before Alsalem (2023) uses the expression “pseudo-concept”.

The six levels of analysis are also compressed. The original formulation of PAS, alienating behaviours, multifactorial models, instruments, expert practices and interventions are frequently treated as a single block. This fusion allows legitimate limitations at level 1 and in some practices at levels 5 and 6 to be transferred to the whole of levels 2 to 4.

The selection of sources is heterogeneous and does not always distinguish the evidential value of empirical articles, books, legal documents, organisational submissions and testimony. The report also does not sufficiently address the body of research after Gardner or the methodological controversy between Meier and colleagues and Harman and Lorandos (Harman & Lorandos, 2021; Meier et al., 2022).

The World Health Organization’s decision not to include parental alienation as an autonomous category in ICD-11 confirms that it is not a recognised health diagnosis. It does not, however, demonstrate that all behaviours, relational processes or contact problems are non-existent (World Health Organization [WHO], n.d.). Likewise, the Family Justice Council (2024) rejects PAS while simultaneously recognising the possibility of operationally defined alienating behaviours.

The Parental Alienation Study Group disputes the report and brings together literature that it considers ignored or insufficiently addressed. That group adopts an advocacy position in support of the concept; its documents should therefore be identified as advocacy and used to locate arguments and sources, not as an independent arbiter of the debate (Parental Alienation Study Group [PASG], 2025).

15.4. Critical synthesis

Criticism of misuse of the concept is well founded and should inform training, investigation of violence, procedural safeguards and evaluation of interventions. The additional inference in Alsalem’s report (2023), according to which parental alienation is globally a “pseudo-concept”, is epistemologically weak because it does not distinguish levels of analysis or apply a formal method of demarcation.

Back to contents

16. The allegation of pseudoscience: analysis of the main arguments

16.1. Reducing the debate to Gardner and the risk of the genetic fallacy

A recurring rhetorical strategy consists in reducing all contemporary research and practice to Richard Gardner, the circumstances in which he formulated PAS and positions of his that are now widely contested. For analytical convenience, we refer to this strategy here as the ad Gardner argument: instead of separately assessing contemporary propositions, data and models, the entire debate is traced back to Gardner’s biography and original formulation. These elements are relevant to assessing that formulation, the quality of its sources and the legitimacy of its original diagnostic claims.

The origin of a hypothesis, the characteristics of its proponent and the historical context in which it was formulated may be relevant to critical assessment. Moving directly from that problematic origin to the falsity of all subsequent propositions may, however, involve a genetic fallacy: a claim is rejected because of its origin without sufficient examination of the arguments, data and reformulations that currently support or contradict it. This caution neither validates Gardner’s original model nor makes its history irrelevant.

It is therefore necessary to separate the validity of Gardner’s PAS, the existence of behaviours capable of interfering with a parental relationship, the usefulness of multifactorial models, the quality of instruments and the evidence concerning practices and interventions. Criticism of the first question does not automatically answer the others (Kelly & Johnston, 2001; Warshak, 2015).

Criticising Gardner does not remove the need to assess research that preceded his work, reformulated it or departed from his propositions.

16.2. “It is not included in DSM-5-TR or ICD-11”

Parental alienation is not an autonomous mental disorder recognised in DSM-5-TR or ICD-11. It should not be presented as a medical diagnosis. Diagnostic systems do not, however, catalogue every human behaviour, family problem or legally relevant phenomenon.

16.3. “There is no scientific consensus”

There is no consensus on terminology, diagnosis, criteria, prevalence, instruments, causality and interventions. This must be acknowledged. The absence of complete consensus is not, however, a sufficient criterion for pseudoscience. The relevant questions are which propositions attract convergence and which remain disputed.

Even critical publications recognise that some fathers and mothers may use harmful strategies to turn a child against the other mother or father (Milchman et al., 2020). Criticism of exaggerated claims of consensus is justified; the non-existence of all such behaviours does not follow from that criticism.

16.4. “The construct lacks scientific validity”

The most substantive criticism concerns validity. There is no single, widely recognised and sufficiently validated test that makes it possible to conclude from a score that a child has been alienated. Nor is there a specific marker that automatically distinguishes alienation, abuse and conflict.

There are, however, different levels of evidence: scales concerning behaviours, studies of representations, reliability research, group comparisons and analyses of consequences. A study may provide supporting evidence for one component without fully validating a diagnostic category.

16.5. “The models are circular”

Circularity occurs when rejection is treated as proof of alienation and every explanation offered by the child is then interpreted as a further consequence of alienation. A model of this kind is protected against refutation and should be rejected.

Differential models make the hypothesis testable when they require a prior relationship, an identifiable change, absence of sufficient reasons for rejection, interfering behaviours, convergence between sources and rejection of the hypothesis where there is evidence of violence or sufficiently serious conduct.

16.6. “The indicators are vague or non-specific”

Intense rejection, adult language, absence of ambivalence, hostility towards relatives and fear may occur in several situations. They have low specificity and do not identify a single cause. The criticism is correct when lists are used as diagnostic proof. The scientific response is differential assessment of hypotheses, not the elimination of every construct that shares indicators.

16.7. “There are no reliable prevalence data”

The criticism is largely justified. Studies measure different behaviours, perceptions, refusals of contact and judicial decisions. There is no sufficiently established universal rate. The uncertainty precludes categorical quantitative claims, but it does not demonstrate that the phenomenon is non-existent.

16.8. “Research relies on retrospective reports and biased samples”

Many studies use adults recalling childhood, fathers and mothers who consider themselves targets of alienation, or participants recruited from specific groups. These samples carry risks of selection, reconstructed memory, lack of independent confirmation and limited representativeness.

The criticism does not apply uniformly to every study. The literature includes research using multiple sources of information, analyses of judicial decisions, experimental studies of parental denigration, psychometric work and group comparisons. The available research remains methodologically incomplete, but it is not wholly incapable of testing and progress.

16.9. “Contrary research is ignored”

Selecting only favourable studies is a genuine problem. Findings such as those of Rowen and Emery (2019), according to which denigration may produce a boomerang effect, must be integrated and require automatic causal models to be abandoned. The same standard applies to critical literature: selecting Gardner, Meier or Mercer while omitting empirical studies and differential models produces an equally partial representation.

16.10. “The construct discredits reports of violence”

The risk is serious and documents a problem of forensic use. It does not demonstrate the non-existence of the behaviours. Construct validity, assessment accuracy, admissibility of evidence, proportionality of the decision and safety of the intervention must be distinguished.

The appropriate response requires prior investigation of violence, differential assessment, multiple sources, adversarial scrutiny, professionals trained in both areas and language proportionate to the degree of certainty.

16.11. “It is a gendered theory against mothers”

Part of the history of the construct reproduced stereotypes of vindictive or manipulative women. It is necessary to investigate whether similar allegations produce different outcomes according to gender. The capacity to engage in the behaviour, its frequency and the institutional response must nevertheless be distinguished.

Evidence that men and women may engage in alienating behaviours answers the first question, but does not demonstrate equal frequency or equality of power. Possible asymmetry in judicial responses also does not demonstrate that the psychological mechanism can occur in only one direction.

16.12. “Changing the label merely rescues the theory”

Replacing the word syndrome without changing the assumptions does not constitute progress. However, the reformulations by Kelly and Johnston, resistance-refusal models and multifactorial approaches changed the object of analysis, the role of the child, the centrality of intention, the consideration of multiple causes and the inclusion of hybrid cases.

16.13. “There are no validated interventions”

Evidence concerning interventions is one of the weakest areas. Many studies have small samples, no control groups, non-random selection, limited follow-up and links between authors and programmes. Recognising a problem does not automatically validate treatments or judicial measures applied in its name.

16.14. “The number of publications proves that it is science”

The number of publications does not demonstrate quality. An extensive bibliography may include empirical articles, commentaries, decisions, chapters and advocacy. Quantity shows only that it is incorrect to claim that the field is limited to Gardner’s writings.

16.15. “Institutions reject or recognise the construct”

Institutional positions must be cited precisely: institution, document, date, version, object and method. Generic claims that the scientific community rejects or recognises parental alienation are symmetrical forms of appeal to authority.

16.16. Application of the criteria

Rigid and circular versions of PAS may be difficult to refute. Contemporary differential models formulate conditions of inclusion and exclusion. Terminology remains heterogeneous; evidence of validity and reliability is partial; there are comparisons of hypotheses, debate, publication, criticism and reformulation.

The analysis does not support classifying the entire body examined as consolidated science. Nor does it support classifying it globally as pseudoscience. Its status varies according to the specific hypothesis, method, instrument and intervention.

Back to contents

17. Limitations of the current state of knowledge

Research is sufficiently extensive to prevent the claim that no science exists, but it remains far from a fully stabilised domain of inquiry. The principal limitations include heterogeneous definitions, the absence of widely shared criteria, convenience samples, few longitudinal studies, reliance on retrospective accounts, difficulty establishing causation, insufficient replication and limited evaluation of interventions and adverse effects.

17.1. Limitations of the present review

This work is a critical narrative review rather than a systematic review. No PRISMA protocol was implemented, no meta-analysis was conducted and no formal, reproducible assessment of risk of bias was undertaken for each study. The selection sought to represent divergent scientific positions and relevant institutional sources, but it remains subject to the interpretative choices of the review authors.

Terminological and methodological heterogeneity prevents aggregation of many findings. The literature considered was limited to sources identified up to July 2026, and later publications may alter specific conclusions. Legal, professional and advocacy documents were included in order to analyse the controversy, not to attribute to them evidential value equivalent to peer-reviewed empirical research.

These limitations call for caution: the review permits assessment of the reasonableness of a global classification as pseudoscience, but it does not provide a quantitative estimate of the average methodological quality of the available literature, the prevalence of the phenomenon or the effectiveness of interventions.

17.2. Conceptual heterogeneity

The same expression may refer to negative comments, refusal of contact, serious deterioration of a relationship, intentional interference, unintentional influence, a condition of the child, a pattern of violence or an allegation in legal proceedings. Future research should separate behaviour, relational process, outcome, causal attribution and clinical or legal characterisation.

17.3. Sampling problems

Much of the research uses convenience samples comprising parents who identify themselves as targets of alienation, association members, service users, families in litigation or specialist professionals. These samples are useful for studying specific experiences but do not permit generalisation to the population.

17.4. Retrospective studies

Retrospective accounts have value but are subject to forgetting, reconstruction of memory, influence from later experiences and difficulty establishing temporal sequence. Conclusions should be proportionate to the method used.

17.5. Scarcity of longitudinal studies

Studies are needed that follow families over time and permit observation of the onset of resistance, preceding events, development of relationships, factors predicting recovery and later consequences. Without follow-up, it is difficult to determine whether behaviours preceded rejection or arose in reaction to it.

17.6. Causation

Much of the literature identifies associations. Causation requires temporal sequence, consistency, consideration of alternatives and replication. Family systems are reciprocal: rejection may increase a parent’s anxiety, leading to intensified controlling behaviour, which in turn aggravates the rejection.

17.7. Absence of a universal reference standard

There is no independent and widely accepted procedure that determines with certainty whether alienation is present or absent. This difficulty limits assessment of sensitivity, specificity and accuracy. If an instrument is compared only with the opinion of experts who use the same assumptions, validation may become circular.

17.8. Independence and conflicts of interest

Some programmes are evaluated by the professionals who developed them. Conflicts of interest may include work in programmes, associations, expert assessment, legislative campaigns, victim support or instrument development. They do not automatically invalidate research, but they should be declared symmetrically.

17.9. Adverse effects of interventions

Studies should not measure only whether contact was restored. They should assess anxiety, fear, trust, relationship quality, intensification of conflict, retraumatisation, interruption of significant relationships and the effects of abrupt changes of residence.

17.10. Participation of children

Many studies focus on parents or judicial decisions. Children’s direct experience should be studied in a manner adapted to age and forms of communication, without treating children as infallible sources or passive objects.

17.11. Cultural and legal differences

Much of the research was produced in the United States, Canada, Australia and some European countries. Results cannot be transferred automatically to Portugal because of differences in legislation, courts, expert assessment, services and residence arrangements.

17.12. Publication bias and polarisation

Positive findings are more likely to be published. In a polarised debate, there may also be selective choice of journals, repeated publication by the same groups, hostile readings of opposing positions, exaggeration of findings and confusion between research and campaigning. Pre-registration, accessible data, independent replication, null findings and conflict-of-interest declarations are priorities.

17.13. Forensic application

A statistical finding about groups does not automatically determine what occurred in a particular family. Research can inform individual assessment but cannot replace it.

17.14. The risk of turning caution into inaction

An impossible demand for absolute certainty may prolong harmful situations. The appropriate response is to state uncertainty, adopt proportionate and reversible measures where possible, monitor outcomes and review decisions when new information emerges. Scientific caution is not paralysis.

Back to contents

18. Priorities for future research

• Develop shared operational definitions distinguishing alienating behaviour, conflict, resistance, unjustified rejection, justified estrangement, hybrid cases, emotional abuse and violence through an intermediary.
• Conduct longitudinal and prospective studies following families from the early stages of separation.
• Use representative samples to estimate the frequency of behaviours, refusals of contact and distribution by gender and socioeconomic context.
• Promote independent validation of instruments, including convergent and discriminant validity, inter-rater reliability, sensitivity and specificity.
• Directly compare alienation, justified estrangement, domestic violence, intense conflict, harmful parental behaviours and hybrid cases.
• Evaluate interventions using transparent criteria, comparison groups, independent evaluators, extended follow-up and recording of adverse effects.
• Develop Portuguese research on expert assessments, judicial decisions, duration of proceedings, hearing children, professional training and the outcomes of measures.
• Study gender, power and institutional response without assuming symmetry or asymmetry in advance.
• Separate description of behaviour, causal explanation and the intervention decision, recognising that each step requires its own evidence.

Back to contents

19. Discussion

19.1. The error of treating parental alienation as a single entity

The controversy arises in part because the same expression is used to refer to a syndrome, behaviours, unjustified rejection, resistance to contact, instrumentalisation, expert practices, interventions and allegations in legal proceedings. The validity of each level must be assessed separately.

19.2. Phenomenon, construct and theory

The phenomenon consists of observed behaviours and relational patterns: the child refuses contact, a parent repeatedly conveys a negative image of the other parent, communication is obstructed or the child is pressured to choose.

The construct is a theoretical category used to group and interpret behaviours, relationships and indicators considered to be associated, such as alienating behaviours, resistance to contact or parental alienation.

The theory seeks to explain how those behaviours and relationships develop, which mechanisms are involved and under what conditions they may contribute to rejection. There may be strong evidence for some behaviours without the causal theory being fully validated.

19.3. It should not be treated as a presumed diagnosis

The literature does not provide a basis for treating parental alienation as an autonomous mental disorder. Expert language should describe behaviours, relationships, hypotheses, the material supporting or weakening each explanation and the degree of certainty.

A methodologically appropriate formulation would be: “Persistent denigrating behaviour and obstruction of contact were identified, associated with a deterioration in the child’s relationship with the other parent which, on the available material, is not sufficiently explained by violence, abuse, neglect or conduct by the rejected parent sufficiently serious to justify the refusal.”

19.4. Global classification as pseudoscience

Analysis of the demarcation criteria does not permit the whole body of research and practice to be classified globally as pseudoscientific. There are empirical studies, assessed instruments, contradictory findings, published criticisms, responses, conceptual changes and competing models. This does not mean that every study is scientific or that every practice has sufficient evidence.

Some versions of PAS, professional practices and claims may display pseudoscientific characteristics. It does not follow that all research on alienating behaviours, unjustified rejection or instrumental use of the child is pseudoscience.

19.5. Developing science is not consolidated science

Rejecting a global classification as pseudoscience does not authorise the available research to be presented as fully validated. The most appropriate formulation is a heterogeneous, controversial and still developing body of research: there are testable questions and empirical inquiry, but important uncertainties persist and practical applications sometimes exceed the evidence.

19.6. Two symmetrical forms of denial

One position denies the possibility of alienating behaviours and treats every allegation as a strategy used by abusers. Another denies the possibility of justified rejection and interprets every form of resistance as influence. Both are incompatible with differential assessment.

19.7. False positives and false negatives

Child protection requires simultaneous attention to both. A false positive may minimise violence and impose unsafe contact. A false negative may permit instrumentalisation, entrench unjustified rejection and destroy a significant relationship. Methodological symmetry does not mean that risks are equal in every case; where there is immediate danger, safety takes priority.

19.8. Violence through an intermediary, “vicarious violence” and gender

In several legal contexts, the notion of vicarious violence describes a man’s use of children to cause suffering or harm to their mother. That definition may be appropriate within gender-based violence policy. Where the purpose is to describe the mechanism irrespective of gender, the expressions violence through an intermediary and instrumental use of the child should be used.

A conceptually neutral approach does not require inequalities to be ignored. It is possible to recognise that women are disproportionately affected by particular forms of violence while also recognising that different people may use children to control or punish. Research should distinguish normative definition, prevalence, gender distribution and institutional response.

19.9. The position of the child

The child should be treated neither as an infallible source nor as a person whose account may be discarded because influence is alleged. Hearing should respect age, maturity, communication, context, consistency, development of the account and the right not to choose between parents. Young children also communicate through behaviour, play, drawing and interaction.

19.10. Time as a forensic variable

Delay may be necessary to avoid premature decisions, but it may also entrench rejection, weaken bonds and make recovery more difficult. Caution should combine timely assessment, protection where risks exist, proportionate interim measures, monitoring and review. A late decision may be formally correct yet materially incapable of repairing the relationship.

19.11. The role of courts

Courts should not decide on the basis of a label. They should examine established behaviour, its effect on the child, reasons for resistance, current risk, capacity to support significant relationships, proportionality of the measure and arrangements for monitoring. The expression parental alienation must neither lower the evidential burden nor operate as a prohibition on investigating interference.

19.12. The role of forensic psychology

Forensic psychology should avoid conclusions that exceed the data and also avoid refusing every conclusion because absolute certainty is unavailable. A sound opinion distinguishes observed, reported and corroborated information, hypotheses, inferences, limitations and level of confidence. It should also state what could alter the conclusion.

19.13. Interventions and proportionality

Evidence concerning interventions is weaker than evidence concerning the existence of interfering behaviours. Measures should be graduated and evaluated for safety, relationship quality, wellbeing, maintenance of outcomes and the child’s experience. Restoration of contact cannot be the sole indicator of success.

19.14. Advocacy, science and public policy

Organisations supporting recognition of the concept and organisations seeking to restrict it participate legitimately in the debate. Their documents should be identified as advocacy where that is their nature. The quantity of testimony or declarations does not replace research. Evidence also does not determine public policy automatically; legal and normative judgment is required.

19.15. Application to the question of pseudoscience

Classifying a heterogeneous body of research and practice globally as pseudoscience requires more than demonstrating controversy, lack of diagnostic recognition, weak studies, judicial misuse, a problematic origin or insufficiently evaluated interventions. It would be necessary to demonstrate lack of testability, immunisation against criticism, absence of empirical research, rejection of adverse findings and incapacity to evolve. The literature does not support that global conclusion.

Back to contents

20. Conclusion

This review did not seek to determine whether every allegation of parental alienation is true, whether Parental Alienation Syndrome should be recognised as a diagnosis or whether particular interventions should be applied.

The question examined was different: is it scientifically defensible to classify globally as pseudoscience six different levels of knowledge and practice brought together under the same label?

Gardner’s proposed Parental Alienation Syndrome has important methodological and historical limitations and is not an autonomous diagnostic category recognised in the DSM-5-TR or ICD-11. The allegation of pseudoscience may legitimately apply to strong claims associated with the original formulation of PAS.

The absence of diagnostic recognition does not, however, entail the non-existence of all the associated behaviours, relational processes or family problems. Professionals should describe behaviour and relationships, state hypotheses and avoid unrecognised nosological labels.

Empirical research exists on alienating behaviours, multifactorial models, contact problems, associated consequences and measurement instruments.

The evidence is heterogeneous and incomplete, but displays sufficient testability, peer review, critical debate and cumulative development to prevent global rejection of the entire body of research as pseudoscience.

Research does not provide a single test capable of identifying alienation with high accuracy in every context. Assessment requires a systemic and multi-method forensic psychological approach, differential assessment of competing hypotheses, investigation of violence, multiple sources and explicit acknowledgement of uncertainty.

There are serious risks of false positives and false negatives. A justified rejection may be classified as alienation, thereby minimising violence. Instrumental use of the child and persistent obstruction may also go unrecognised, allowing emotional and relational harm to become entrenched.

Particular patterns of alienating behaviour may amount to psychological maltreatment, violence through an intermediary or strategies embedded in coercive control. This characterisation depends on the behaviour, context, duration, intention, power structure and impact, and does not follow automatically from a label.

The report of the United Nations Special Rapporteur identifies legitimate risks concerning domestic violence, gender bias and improper use of the concept. Its methodology is not, however, sufficient to demonstrate the global pseudoscientific status of the domain.

Critical literature is right to reject claims of absolute consensus, known prevalence, definitive validity or proven effectiveness.

Supportive literature is right to point out that criticism of the syndrome does not eliminate interfering behaviours or rejection lacking sufficient justification.

The appropriate response is not to accept or prohibit a word. It is to investigate specific behaviour, hear the child, assess relationships, test alternative hypotheses, protect against violence and recognise the possibility of parental interference. In proceedings concerning the exercise of parental responsibility, the child’s interests are protected through rigorous assessment, timely procedure, adversarial scrutiny, proportionality and review of decisions when new information emerges.

Back to contents