Murray Bowen, one of the founders of family systems therapy, spent years watching families in clinical settings. What he noticed was that anxiety inside a family rarely stayed put. It moved. It travelled from a parent to a specific child, skipped a sibling, and reappeared a generation later in a grandchild who had never met the original source. Bowen came to describe this drift as the family projection process, and he folded it into a wider framework he called multigenerational transmission. No one had to shout. No one had to hit. The signal moved on its own.
That claim — that a family can hand down its anxiety across three generations without a single overt argument — remains one of the more provocative ideas in twentieth-century psychiatry. It also happens to be one that modern epigenetics, developmental psychology, and trauma research have quietly started to corroborate, from different angles, using tools Bowen never had.
The clinician who watched families instead of patients
Bowen trained as a psychiatrist before conducting a project that admitted entire families — mother, father, and an adult child diagnosed with schizophrenia — to live together on the ward. The idea was radical for its time. Most psychiatry in the 1950s treated the patient as an isolated organism. Bowen treated the family as the organism and the patient as its most visible symptom.
What he saw during those live-in observations reshaped the field. Families produced a kind of emotional weather. One member absorbed most of the storm. Another stayed calm at the cost of staying distant. The parents’ unspoken worries seemed to land, with unnerving precision, on one particular child — usually the one most attuned to them.
Bowen synthesised these observations into eight interlocking concepts. The family projection process was one of them. Multigenerational transmission was another. Together they described a mechanism: chronic anxiety in one generation gets funnelled onto a child, that child grows up with a slightly lower capacity to tolerate uncertainty, and when they become a parent they funnel their own anxiety onto one of their children in turn. Three generations. No raised voices required.
What “projection” actually looks like in a living room
The word projection is doing a lot of work here, and it does not mean what a Freudian would mean by it. In Bowen’s usage, projection is a process — a slow, repetitive, often loving pattern in which a parent’s own unresolved anxiety gets fixed onto a child as if the child were the problem.
It looks ordinary. A mother who worries constantly about her son’s health monitors him more closely than his siblings. He grows up sensing that his body is fragile, that vigilance is love, that separation is dangerous. A father who never resolved his own fear of failure watches his daughter’s schoolwork with an intensity he does not apply to his other children. She learns that her worth is contingent and that her father’s mood is her responsibility. Nothing about these interactions requires an argument. The transmission happens in the tone of a question, the length of a pause, the choice of which child to look at first when the phone rings with bad news.
Recent meta-analytic work summarised by Nature’s research intelligence index on intergenerational dynamics of childhood anxiety lines up neatly with what Bowen described decades ago. Parental modelling of fear — the way a caregiver reacts to a stranger, a dog, a doctor’s waiting room — shapes infant responses to novel stimuli, with the effect being largest in temperamentally inhibited infants. Highly sensitive children absorb the parent’s fearful appraisal most readily. The child in Bowen’s framework who becomes the focus of parental anxiety is, in modern terms, the temperamentally inhibited infant paired with an anxious caregiver.
Why the quietest parents transmit the most
The families Bowen studied were rarely dramatic. He wrote that the most intense projection processes he observed happened in families that appeared, from the outside, to be functioning well. Everyone was polite. No one shouted. The anxiety was displaced into worry about a child rather than expressed as conflict between the adults.
This is where the modern research on covert parental patterns intersects with Bowen’s framework. The Nature summary of intergenerational parental depression research notes that the everyday emotional climate of caregiving — not discrete events — is where much of the transmission happens. It is the pattern, not the incident, that lands.
A parent who never raises their voice can still communicate: your emotions are inconvenient, your needs are dangerous, my mood depends on your behaviour. The child absorbs this the same way they absorb grammar — without being taught, without being able to name it, and without the ability to compare it to any other family until adulthood.
The three-generation frame
Bowen’s most contested claim was that this process compounded across generations. He described what he called the multigenerational transmission process: over successive generations, the child most caught in the projection tends to marry someone with a similar level of what Bowen called differentiation of self — the capacity to hold onto one’s own thinking under emotional pressure. Their most susceptible child inherits slightly less differentiation than they had. And so on.
He argued that after roughly three generations of this compounding, families could produce a child whose baseline anxiety was severe enough to manifest as major psychiatric illness. This was Bowen’s controversial explanation for why schizophrenia sometimes appeared to “run in families” without a clear genetic pattern.
The specifics of that claim have not held up in modern psychiatric genetics. But the three-generation frame has aged surprisingly well in other domains. Rachel Yehuda’s group at Mount Sinai has spent decades studying Holocaust survivors and their adult children, and in a 2016 study in Biological Psychiatry the team reported altered methylation on the stress-response gene FKBP5 in both the survivors and their offspring — the first demonstration in humans of a link between a parent’s trauma and a measurable change in the next generation’s biology. That work reached the children of survivors, the second generation; whether such marks travel further, into grandchildren with no direct exposure, is a harder question the human evidence has only recently begun to address.
More recent work reaches that third generation directly. A 2025 study in Scientific Reports followed three generations of Syrian refugee families in Jordan and found distinct DNA-methylation signatures in the grandchildren of women who had been pregnant during wartime violence — fourteen sites tied to that germline exposure, and twenty-one more in people who had lived through the violence themselves. The mechanism proposed is not identical to Bowen’s — it runs through methylation on stress-response genes rather than family dynamics, and the authors are careful to note how limited the human evidence still is — but the empirical shape rhymes with what he described. Something moves through the family line that is neither a memory nor a shouted lesson.
The mechanism, in plain terms
Contemporary trauma researchers now describe the transmission through multiple overlapping channels. An overview in Psychology Today on intergenerational trauma and resilience describes the pathways of transmission.
First, direct behavioural modelling — the child watches how the parent handles a stranger, a bill, a hospital visit. Second, communication style — what gets named, what gets whispered, what is forbidden to discuss. Third, physiological co-regulation — a chronically anxious parent produces a chronically anxious autonomic environment, and the infant’s nervous system calibrates to it. Fourth, epigenetic modification — methylation patterns on genes involved in stress reactivity can be altered by prolonged parental stress and, in some cases, appear to persist in the next generation.
Bowen only had access to the first two channels. He inferred the third from clinical observation. The fourth was invisible to him. But the shape of the transmission he described — quiet, cumulative, targeted at one child in particular, invisible to the family itself — turns out to describe something that operates on multiple biological layers at once.
The child who gets chosen
One of the more unsettling parts of Bowen’s model is the claim that the projection process rarely lands evenly on all siblings. There is usually one child — the one most temperamentally attuned, most emotionally available, most willing to absorb the parent’s mood — who takes the largest share. Siblings from the same household can grow up with wildly different levels of baseline anxiety not because they were treated with obvious unfairness, but because one of them was the emotional lightning rod.
This matches what clinicians who work with adult children of covertly controlling parents report. In a family with a parent who never raised their voice but ran the household through guilt and silence, one sibling often emerges as the vigilant one — the one who reads the room, manages the parent’s feelings, cannot say no. The others may look more troubled or more successful on paper, but the vigilant child carries the anxiety.
A Medical News Today overview of intergenerational trauma notes that people carrying it often describe the same cluster decades later: guilt, chronic anxiety, and a hypervigilance that never fully switches off — a background sense that something is wrong. None of them can point to a single argument that produced it.
Where the research is heading now
The mechanistic side of Bowen’s hypothesis is being taken up by molecular biologists who would not necessarily call themselves family theorists. A 2018 review in World Psychiatry by Rachel Yehuda and Amy Lehrner lays out the putative epigenetic pathways — methylation on glucocorticoid-receptor and stress-response genes — through which the effects of parental trauma might reach a child’s biology. The findings are not settled. The mechanisms by which such marks would survive gametogenesis remain contested. But the empirical claim — that a parent’s, and possibly a grandparent’s, experience can show up in a descendant’s stress biology — is no longer fringe.
On the intervention side, work summarised by Open Access Government on children of parents with severe mental illness emphasises what Bowen would have recognised immediately: the point of intervention is not the diagnosed adult and not the affected child in isolation, but the pattern between them. Support the parent’s capacity to regulate their own anxiety and the child’s exposure drops. Increase the child’s access to a differentiated adult — someone who does not need them to manage feelings — and the transmission attenuates.
What breaks the chain
Bowen was not fatalistic. His clinical work was built around the idea that a single member of a family, if they could raise their own level of differentiation, could shift the entire system. He called this working on self. It did not require confronting the parent. It did not require an argument. It required noticing the pattern and refusing, quietly, to participate in it in the same way.
Modern resilience research points in a similar direction. Beliefs about the self and about ageing turn out to have measurable biological consequences. A Yale study led by Becca Levy, which tracked people for more than two decades, found that those who held more positive beliefs about their own ageing lived about seven and a half years longer — a gap Yale’s own summary noted was larger than the years attributed to regular exercise or to not smoking. Cognition, in other words, is not decoration on top of biology. It is a lever on it. The child who eventually names the pattern in their family — who sees that the anxiety they have carried since age four was never really theirs — has already begun to interrupt the transmission.
That interruption is what Bowen was after when he sat in the ward in the late 1950s and watched families eat dinner together. He was not looking for the argument. He had already worked out that the argument was not the point. The point was the silence around the argument, the way it settled onto one child’s shoulders, and the way that child would carry it, gently and without complaint, into a marriage and a nursery of their own. Sixty years on, in cortisol assays and methylation maps and meta-analyses of paternal anxiety, the outline of what he saw is still legible.
