When Violence Is Learned: Social Learning, Role Models, Development and the Neurobiology of Relational Behaviour

By Deborah Casey

23rd August 2026

The question of how violence can become embedded within families requires us to move beyond the idea of a simple “cycle of abuse”. Violence is neither transmitted mechanically from one generation to the next nor reproduced inevitably by every child exposed to it. Human development is considerably more complex. Children are active participants in their own development, interpreting what they observe through relationships, culture, temperament, cognition and experience. Nevertheless, the social environment in which a child develops provides an important source of information about how people behave, how relationships operate, how conflict is managed and how power is exercised.

The developing child cannot therefore be considered independently from the relational environment surrounding them. Family relationships constitute one of the earliest and most influential social learning environments, through which children encounter models of affection, communication, authority, conflict, aggression, reconciliation, cooperation and control. What children observe does not automatically become what they later do, but repeated observations can contribute to the repertoire of behaviours and relational expectations from which they subsequently construct their own responses. This makes the study of violence inherently developmental, because the question is not simply what behaviour occurred, but what meaning the developing child may have attached to what they repeatedly witnessed.

Albert Bandura's social learning theory provides one of the most influential frameworks for understanding this process. Social learning theory proposes that behaviour can be acquired through observation, modelling and imitation rather than requiring the individual to experience every consequence personally. Importantly, the observer also attends to what happens to the person being observed. Behaviour that appears to produce reward, status, compliance or control may become more salient than behaviour that consistently produces negative consequences. Bandura subsequently developed this approach through social cognitive theory, placing greater emphasis on cognition, interpretation, self-regulation and human agency. The significance of this distinction is considerable because it means that children are not simply passive recipients of behavioural information; they are interpreting what they observe and developing beliefs about the circumstances in which particular behaviours appear effective.

The significance of modelling therefore extends well beyond the copying of individual physical behaviours. A child who observes an adult shouting, threatening, humiliating or using physical aggression is also observing what happens to the recipient of that behaviour. The child may see another person becoming frightened, withdrawing, complying or attempting to appease the aggressor. They may observe whether other adults intervene, whether the perpetrator experiences consequences, whether the behaviour is subsequently minimised, or whether the relationship continues largely unchanged. Consequently, the model may communicate something about relational power as well as about aggression. The developing child may begin to encounter the implicit possibility that aggression can be used to alter another person's behaviour.

This is why power dynamics are central to understanding the development of abusive behaviour. Violence within relationships is rarely simply an isolated physical act; domestic abuse can involve patterns of coercion, intimidation, domination and control through which one person restricts another person's autonomy. A child who repeatedly observes such interactions is therefore potentially witnessing a model of power as much as a model of violence. They may be observing how authority is established, how resistance is treated, how boundaries are challenged, how compliance is obtained and how control can be maintained within intimate relationships.

Social learning theory consequently provides an important framework for understanding intergenerational violence, but it should not be treated as a complete explanation. Contemporary research identifies multiple interacting pathways, including observational learning, social-information processing, emotional regulation, attachment, stress responses, individual characteristics and wider social and cultural influences. Intergenerational transmission is therefore better understood as a complex developmental process in which several mechanisms may operate simultaneously rather than as a straightforward process in which one generation simply teaches the next to behave violently.

Social-information-processing perspectives add another important dimension because children do not merely observe behaviour; they interpret social information. They learn to recognise social cues, attribute intentions, generate possible responses and evaluate the consequences of different behaviours. A child repeatedly exposed to aggression may consequently develop particular expectations about conflict, including beliefs about whether another person's behaviour is threatening, whether aggression is an effective response, whether retaliation is necessary, and whether violence is likely to achieve a desired outcome. These processes can influence later behaviour, but they remain influenced by subsequent experience and individual interpretation.

This helps explain why children exposed to similar environments can develop very different trajectories. One child may reproduce behaviour that they have observed, another may become intensely fearful of conflict, another may consciously reject violence, while another may become highly controlling or vigilant. Exposure therefore represents a potential developmental influence rather than a predetermined outcome. The same experience can be interpreted differently by different children, and later relationships, protective adults, education and personal agency can alter the trajectory.

The concept of the role model is particularly important within this framework. Children encounter many adults, but not every adult has the same developmental significance. Parents, grandparents, older siblings, teachers, coaches and other socially important individuals can become reference points against which children evaluate behaviour and social expectations. The perceived status of the model can therefore matter. A person who is admired, feared, charismatic, confident or socially influential may command greater attention than someone who carries little significance within the child's social world.

This becomes especially complicated when an admired individual displays behaviour that is harmful to others. A child may encounter an adult who appears powerful, confident or impressive while simultaneously behaving abusively. The developmental message is therefore potentially contradictory. The child may recognise the harmful behaviour while continuing to admire the person demonstrating it, creating a complex process of social interpretation in which positive and negative evaluations coexist.

The relevant question is consequently not simply whether the behaviour was observed, but how the observer understood the person who demonstrated it and the consequences that followed. If aggression appeared to produce compliance, if intimidation appeared to prevent challenge, or if manipulation appeared to secure loyalty, the child may potentially learn something about the instrumental value of those behaviours. This does not mean the child consciously decides that such behaviour is morally acceptable. Rather, the behaviour may become represented as one possible method of exercising influence or managing interpersonal situations.

This distinction between observing an action and learning the principle underlying that action is particularly important. Modelling is broader than simple imitation because an observer can extract general rules from what they see and subsequently apply those rules in circumstances that are quite different from the original situation. A child does not necessarily reproduce the exact words, gestures or actions of an abusive adult. Instead, they may potentially internalise a more abstract relational expectation, such as the belief that when another person resists, pressure, intimidation or manipulation can be used to restore control.

The family can consequently function as a social environment in which relational rules are repeatedly demonstrated. A child may observe adults resolving disagreements through communication, negotiation and mutual respect, or may repeatedly witness conflict being managed through shouting, intimidation, humiliation, withdrawal, threats or violence. Over time, these experiences can contribute to the child's developing understanding of what relationships are and how people behave when their interests conflict.

The distinction between modelling and imitation is therefore essential when considering domestic abuse. A person does not need to reproduce the exact behaviour they witnessed in childhood for developmental learning to have occurred. What may be reproduced is an underlying relational script concerning power, entitlement, conflict or emotional regulation. Similarly, exposure to manipulation can potentially influence later expectations about loyalty, alliances and interpersonal control without producing an identical repetition of the original behaviour.

This is particularly relevant to family relationships in which children are deliberately or repeatedly placed against one another. When an adult creates divisions between siblings, encourages competing loyalties or uses one family member to influence another, the child is encountering a model of relational power that operates through triangulation rather than necessarily through direct physical aggression. Such experiences may influence developing understandings of trust, loyalty, alliance and betrayal, although again they do not determine how the child will subsequently behave.

The developing brain provides the biological context within which these processes occur. Childhood and adolescence involve substantial neurodevelopment, including continuing maturation of systems associated with emotional regulation, social cognition, reward processing, decision-making and behavioural control. Social experience therefore occurs alongside an evolving neurological capacity to interpret and regulate behaviour. The brain is continually responding to experience, integrating information from the social environment and adapting as the individual encounters new circumstances.

This is where the discussion of mirror neurons requires particular care. Mirror-neuron research has contributed to understanding how observing another person's actions can engage neural systems involved in action representation. However, the popular suggestion that mirror neurons provide a single neurological explanation for imitation, empathy or the acquisition of complex social behaviour considerably oversimplifies the evidence. Contemporary neuroscience understands observational learning as involving distributed networks that integrate action observation with attention, social cognition, emotional processing, reward and executive control.

Mirror-like mechanisms may therefore form one component of observational learning, but they cannot adequately explain the development of complex relational behaviour by themselves. When a child observes an act of aggression, the brain is not simply registering the physical movement. The child is simultaneously processing who is acting, who is receiving the behaviour, the emotional significance of the interaction, the apparent intention of the perpetrator, the response of the recipient and the consequences that follow. The meaning attached to the behaviour is therefore fundamental.

The concept of imprinting also requires careful qualification. Classical imprinting is an ethological concept referring to particular forms of relatively rapid learning during sensitive developmental periods, famously demonstrated in animals. Human development should not simply be described as though children become permanently “imprinted” by parents or caregivers in the same biological sense. Human attachment and social development are more complex, and contemporary developmental science distinguishes classical imprinting from the formation of attachment relationships and internal relational expectations.

Nevertheless, the broader concept of sensitive developmental periods remains important. Early experiences can have particular significance because the developing child is constructing expectations about the social world at a time of profound dependence upon caregivers. It is therefore more scientifically defensible to describe early relational experiences as contributing to developing expectations and patterns of emotional and social response rather than suggesting that children are permanently imprinted by individual experiences.

Attachment theory provides another complementary perspective because children develop expectations concerning whether significant others are available, responsive, predictable and safe. These expectations have often been conceptualised as internal working models of relationships. Importantly, these models are not fixed neurological programmes. Human development remains plastic, and later relationships, experiences and reflective processes can modify earlier expectations.

This capacity for change is particularly important when considering exposure to domestic abuse. Research has identified associations between childhood exposure to violence and later behavioural, emotional and relational difficulties, including increased risks of both victimisation and perpetration in some populations. However, the existence of an increased statistical risk does not establish an inevitable developmental pathway. Protective relationships, alternative role models, social support, emotional regulation, education and later experiences can all influence the trajectory.

The existence of positive role models is therefore as important to understanding prevention as negative role models are to understanding risk. If children can learn behaviours through observation, they can also observe adults resolving conflict without violence, respecting boundaries, accepting responsibility, communicating openly, sharing power and treating others with dignity. Social learning theory therefore has implications not only for understanding how harmful behaviours can be reproduced, but also for understanding how healthier relational behaviours can be demonstrated, reinforced and learned.

The consequences surrounding behaviour are particularly important. A child does not only observe what a perpetrator does; they also observe what happens because the perpetrator did it. When aggression results in compliance and is followed by little accountability, the social information available to the child differs considerably from a situation in which aggression produces meaningful consequences and is clearly rejected by significant adults. Similarly, when respectful communication produces cooperation and positive relationships, those behaviours provide an alternative model of how interpersonal problems can be managed.

The family response therefore becomes part of the learning environment. Silence, intervention, accountability, minimisation, forgiveness, confrontation and boundary-setting can all communicate information about the social acceptability of particular behaviours. A child is observing the entire relational system rather than simply the perpetrator and victim as isolated individuals.

This is one reason domestic abuse should not be regarded solely as a private interaction between two adults when children are present. Children can become observers of the relational system in which violence occurs, and the patterns of power, fear, conflict and resolution surrounding them can become part of their developmental environment. The effects of exposure will vary considerably, but the possibility of developmental influence warrants recognition within safeguarding, prevention and therapeutic practice.

At the same time, there is an important ethical danger in discussing intergenerational transmission. If developmental explanations are used carelessly, responsibility can gradually migrate backwards through generations until the person who perpetrated the abuse becomes almost invisible. One generation is explained by another, which is explained by another, until explanation begins to resemble excuse.

Developmental science should never be used in that way. A person may have experienced significant adversity and nevertheless remain responsible for how they treat others. A person may have witnessed violence and consciously reject it. Equally, a person who has not experienced significant childhood adversity may still become abusive. Developmental history can help identify pathways of risk and resilience, but it does not determine moral or behavioural outcomes.

The most useful framework is therefore one of interaction. Biology interacts with experience, experience interacts with interpretation, interpretation interacts with relationships, relationships interact with social and cultural expectations, and observed behaviour interacts with reinforcement and consequences. Early experiences may influence later expectations, but subsequent experiences can challenge those expectations, and individuals remain capable of reflection, adaptation and change.

This provides a more sophisticated understanding of what is sometimes called the “cycle of abuse”. Rather than imagining violence as something automatically transmitted from parent to child, it is more accurate to understand intergenerational violence as a probabilistic developmental process involving multiple pathways. Observational learning may contribute to behavioural scripts; exposure to threat may influence emotional and physiological responses; attachment experiences may influence relational expectations; social-information-processing patterns may influence interpretations of conflict; and wider cultural beliefs may shape understandings of gender, authority, entitlement and acceptable behaviour.

The child is therefore neither a blank slate nor a predetermined product of their environment. Development occurs through an ongoing interaction between the individual and the world around them. What children observe matters, but so too does what they make of what they observe.

This is particularly important when considering role models. Children learn from what significant adults demonstrate, but they also learn from what those adults tolerate, reward, condemn and permit. They observe who possesses power, how that power is exercised, how victims are treated, whether boundaries are respected and what happens when boundaries are crossed. They also observe adults who demonstrate alternative possibilities: that strength can coexist with restraint, that disagreement does not require domination, that authority does not require humiliation, and that intimate relationships can accommodate difference without coercion.

The implications for prevention are therefore substantial. If harmful relational behaviours can be learned through observation, then preventing violence requires more than telling children that violence is wrong. Children need opportunities to observe alternative models of communication, emotional regulation, accountability, equality and respectful conflict resolution. They need significant adults who demonstrate rather than merely describe healthy relationships. They also need environments in which abusive behaviour is recognised, challenged and associated with meaningful accountability rather than normalised or concealed.

Ultimately, the developing brain is not simply a biological mechanism for reproducing whatever it sees. It is part of a dynamic developmental system in which observation, emotion, cognition, relationships, social meaning and experience continually interact. Mirror-like neural mechanisms may contribute to the representation of observed actions, but the development of complex relational behaviour depends upon much more: the identity of the model, the meaning attached to the behaviour, the consequences that follow, the power dynamics within the relationship, the child's own interpretation and the subsequent experiences that either reinforce or challenge what was learned.

Understanding these processes allows us to hold two propositions together without contradiction. People can be profoundly influenced by what they experience and observe during development, while also remaining responsible for the behaviour they subsequently choose. Developmental influence can help explain vulnerability, learning and the emergence of behavioural patterns, but it does not remove agency and it does not excuse abuse.

Perhaps this is why the language of a “cycle” can ultimately be less useful than the language of developmental pathways. A cycle suggests inevitability; a pathway allows for divergence, interruption and change. Children can inherit experiences without inheriting destiny. They can encounter destructive models without reproducing them, and they can encounter healthy models that provide an alternative understanding of power, relationships and human dignity.

The central question, therefore, is not simply whether violence is learned. The more important questions are how it is learned, what relational meanings accompany it, which behaviours are reinforced, how power is modelled, how the developing brain interprets repeated social experiences, and what enables those pathways to be interrupted. When these questions are considered together, social learning theory, role modelling, observational learning, attachment, developmental neurobiology and the study of power dynamics provide a much more nuanced account of how relational behaviour can travel across generations without assuming that it must.