The Myth of Codependency

The word codependency has become so familiar within the alcohol and drug field that we rarely stop to ask where it came from. There is also an interesting difference in how the word tends to be used. In the UK, codependency is often used fairly loosely to describe an unhealthy alliance between two people in which one person's behaviour protects, facilitates or accommodates the behaviour of another. A partner pays the debts, rings the employer, conceals the drinking or repeatedly rescues somebody from the consequences of their drug use. In this sense, codependency has become almost interchangeable with enabling.

However, this is not the original meaning of the idea.

The history is considerably stranger.

The alcoholic wife

The intellectual roots of codependency can be traced to theories about the wives of alcoholic men that emerged particularly strongly within American psychodynamic thinking during the middle of the twentieth century. The assumption was not simply that living with a problem drinker affected the wife. It was that there might have been something psychologically unusual about the woman before she entered the relationship.

One particularly influential example was Thelma Whalen's 1953 paper Wives of Alcoholics: Four Types Observed in a Family Service Agency. Whalen classified wives into four personality types and attempted to explain their relationships with alcoholic men through their own psychological needs. (PubMed)

The reasoning went something like this.

A woman with a strong need for control might unconsciously select a man she could dominate. A woman with powerful needs to care for or rescue somebody might select a dependent man. Another might obtain psychological satisfaction from suffering, martyrdom or occupying the morally superior position in a relationship.

The alcoholism therefore served a psychological function for both partners.

This eventually became known as the disturbed personality hypothesis.

It generated an extraordinary reversal in thinking.

Instead of asking what prolonged exposure to somebody's problematic drinking might do to their partner, clinicians began asking what might be wrong with the partner that caused her to enter — and remain in — such a relationship.

Some formulations went further still. If the husband stopped drinking, the wife's psychological equilibrium might supposedly be threatened because she had lost the dysfunctional relationship through which her own needs were being met. The woman was no longer simply living with alcoholism. She had become part of its explanation.

Then researchers began testing the idea

The problem was that increasingly the evidence did not behave as the theory predicted.

In 1973 Patricia Edwards, Cheryl Harvey and Paul Whitehead published a major critical review of research on wives of alcoholic men. They specifically compared the disturbed personality theory with alternative explanations, including the idea that apparent psychological problems in wives might be consequences of living under severe and prolonged stress. (PubMed)

The evidence for a distinctive pathological personality was poor.

There was a much simpler explanation for many of the behaviours being observed.

Living with somebody whose drinking is unpredictable and disruptive is stressful.

A partner may become anxious because something anxiety-provoking is happening.

She may monitor alcohol because alcohol has repeatedly predicted trouble.

She may take control of money because household money keeps disappearing.

She may take over responsibilities because somebody else is no longer reliably carrying them out.

She may become hypervigilant because experience has taught her that small changes in another person's behaviour can precede a crisis. What looks like pathology when removed from its context may make considerably more sense when the context is restored. The behaviours did not necessarily reveal the personality of the woman. They might reveal the circumstances in which she was living. For a while, the pathological-wife explanation appeared to be disappearing. Then, remarkably, much of it returned.

The resurrection of codependency

During the late 1970s and particularly the 1980s, a new language began developing within the American treatment and recovery movement. The alcoholic increasingly became the chemically dependent person. The person living alongside them became the co-dependent. But the concept did not remain confined to the partner of an alcoholic.

It expanded.

Writers including Sharon Wegscheider-Cruse, Timmen Cermak, Anne Wilson Schaef, Melody Beattie and John Bradshaw progressively developed codependency into a much larger psychological theory. It became fused with family-systems ideas, dysfunctional-family theory, self-esteem, compulsive behaviour, shame, emotional abandonment and the emerging language of the inner child.

John Bradshaw's enormously successful Bradshaw On: The Family, published in 1988 and based on his television series, presented addiction and codependency within a wider theory in which dysfunctional rules and patterns were transmitted through the family system from one generation to another. (Google Books)

The direction of causation had begun to change again. Addiction could now be interpreted not merely as the problem of an individual but as a symptom of a dysfunctional family system. The partner or child was not simply affected by addiction. Their own codependency could become part of the process through which dysfunction — including addiction — was generated or perpetuated. And unlike the relatively narrow theory of the alcoholic's wife, almost anybody could potentially be included.

You could become codependent through living with an alcoholic.

Or through having an alcoholic parent.

Or an alcoholic grandparent.

Or another compulsive person.

Or through growing up within a dysfunctional or emotionally restrictive family.

The boundaries of the concept were expanding rapidly.

From idea to movement

On 22 October 1986 the first meeting of Co-Dependents Anonymous — CoDA — was held in Phoenix, Arizona. Its founders deliberately constructed a Twelve-Step fellowship modelled structurally on Alcoholics Anonymous. CoDA's own historical material records the origins of the movement and the first meeting in October 1986. (codependents.org)

This was an important moment.

Codependency was no longer simply an idea circulating among therapists and writers.

It had acquired its own recovery movement. At the first National Conference on Co-Dependency in 1989, a group of leading advocates attempted to agree a definition. A contemporary account reported the resulting definition as:

“a pattern of painful dependence on compulsive behaviors and on approval from others in an attempt to find safety, self-worth and identity.”

Recovery, they stated, was possible. (Los Angeles Times) The emphasis increasingly moved towards recovering a lost authentic self: recognising one's own needs, establishing boundaries, rejecting oppressive family rules and separating personal identity from the needs and expectations of others. For some people this was undoubtedly liberating.

But there was a rather significant problem. What exactly was the disease from which they were recovering?

So, what is codependency?

This should have been a straightforward question. It wasn't.

As the literature expanded, so did the characteristics attributed to codependency.

Lists eventually contained around 140 different characteristics in some formulations, while other influential checklists ran still further. To put that in perspective, this was no longer a tightly specified psychological construct. Almost every aspect of relationships, self-esteem, emotion, responsibility, sexuality, control, anger, trust and decision-making could potentially become part of the condition.

More troublingly, some characteristics were opposites. Melody Beattie's enormously influential Codependent No More illustrates the problem particularly well. Her list of characteristics describes codependent people as potentially extremely responsible and also extremely irresponsible. They may respond passively, aggressively, or combine passive and aggressive responses. (goodtimesalways.com)

Think about what that does to a psychological theory. If being highly responsible indicates codependency and being highly irresponsible indicates codependency, what level of responsibility demonstrates that someone is not codependent? The same problem occurred throughout the concept.

A person might trust too easily.

Or be unable to trust.

They might desperately seek relationships.

Or avoid relationships because they fear abandonment.

They might become controlling.

Or feel completely powerless.

They might suppress anger.

Or express too much anger.

They might be obsessively attached to another person.

Or emotionally withdraw from them.

The problem is not that human beings cannot display apparently contradictory behaviours. Of course they can. The problem is that a theory claiming to identify a particular disorder has to specify what observations would not support it. Otherwise almost any behaviour can be retrospectively interpreted as evidence for the theory.

And what if I say I'm not codependent?

Here the problem becomes even more serious. Denial was itself incorporated into the language of codependency. A person who accepted that they were codependent could therefore be regarded as demonstrating insight. But a person who rejected the diagnosis could be regarded as being in denial.

This creates what philosophers of science sometimes call a self-sealing proposition.

The theory explains both the evidence supporting it and the evidence apparently contradicting it. Accept the diagnosis: evidence of codependency.

Reject the diagnosis: evidence that your codependency prevents you recognising it.

Apparently independent and self-sufficient? That too could be interpreted as a defensive manifestation concealing the underlying problem.

An extraordinary illustration of this culture was later described by psychologist William Miller. A young woman attended a family session at the residential service treating her father's alcohol dependence. She was subsequently told that she herself suffered from potentially fatal codependency. When she protested that she remembered a happy childhood and considered herself psychologically normal, this was interpreted as evidence that she was so disconnected from reality that she could no longer recognise normality.

The conclusion had effectively become impossible to escape.

That is not how a scientific hypothesis should operate.

Psychology has to allow itself to be wrong

Science does not require an idea to begin life with perfect evidence. Most important theories begin with observation. Somebody notices a pattern. They construct an explanation. That is exactly what psychology should do.

But generating the theory is only the beginning. The next stage is much less glamorous.

We have to try to prove ourselves wrong. Can the concept be clearly defined?

Can different researchers reliably identify it? Can it be distinguished from existing concepts such as anxiety, depression, trauma, low self-esteem or insecure attachment?Does it predict something that those established constructs do not? Do people supposedly suffering from it actually display the proposed characteristics more frequently?

And what observation would lead us to conclude that the theory itself is incorrect?

By the early 1990s even researchers sympathetic to investigating codependency were acknowledging the definitional problem. A 1991 review in the Journal of Clinical Psychology specifically called for operational diagnostic criteria because of confusion surrounding what codependency actually meant. (PubMed) A subsequent critical review found something revealing: although there was some agreement about broad themes, there was virtually no agreement on a workable definition, and remarkably little research had examined many of the assumptions of the model. Much of the literature rested upon clinical observation and personal experience. (PubMed)

This doesn't mean researchers found absolutely nothing. They did. And this distinction is important.

Did research ever find codependency?

There is empirical evidence for some individual behaviours associated with the idea.

For example, a 1991 experimental study found that women with an alcoholic parent behaved differently from controls when given an opportunity to help an apparently exploitative person. The authors interpreted this as evidence for a particular codependent behaviour. (PubMed)

Researchers have also constructed questionnaires capable of measuring clusters of characteristics labelled codependency. So it would be inaccurate to claim that every observation ever made by codependency advocates was imaginary. But this is not the same as demonstrating that codependency exists as a single distinct psychological disease.

If I construct a questionnaire containing measures of self-sacrifice, anxiety, low self-esteem, difficulty asserting needs and excessive concern for others, I may well find people who score highly on it. That demonstrates that those characteristics exist. It does not necessarily demonstrate that they collectively constitute a separate disease called codependency.

This is an important distinction. Research can demonstrate the existence of the bricks without demonstrating the existence of the building. And once researchers studied families affected by alcohol and drugs without beginning with the assumption of codependency, a very different picture emerged.

What the research found instead

Jim Orford, Richard Velleman, Alex Copello, Lorna Templeton and colleagues developed what became known as the Stress-Strain-Coping-Support model. Its starting position is almost the reverse of classic codependency theory. Having a close relative with a serious alcohol, drug or gambling problem represents a substantial and often prolonged stressor. That stress can produce psychological and physical strain. Family members attempt to cope.

And the effects of that stress and their ability to cope are strongly influenced by the quality of support available to them.

Most importantly, the model begins by regarding affected family members as ordinary people dealing with extraordinarily difficult circumstances, rather than people displaying evidence of an underlying pathological personality. (University of Birmingham)

Research across different populations has been able to measure these processes with considerably more precision. Measures of stress, psychological and physical symptoms and different coping responses have demonstrated acceptable reliability and validity across studies involving family members affected by alcohol, drugs and gambling. (Wiley Online Library)

This produces a very different interpretation of the same behaviour.

A mother repeatedly giving money to her drug-dependent son may indeed be doing something that ultimately makes the situation worse.

A wife continually ringing her husband's employer to conceal his drinking may be preventing him experiencing consequences.

A partner may spend enormous amounts of time monitoring, checking, pleading, threatening or rescuing.

None of those behaviours automatically requires us to conclude that the person has a disease.

They may be coping strategies.

Some coping strategies work.

Some work temporarily.

Some cease to work when circumstances change.

Some inadvertently make the situation worse.

And desperate people sometimes keep repeating things that have stopped working because they have no idea what else to do. That is very different from saying their behaviour reveals an underlying pathological need for the addiction to continue. Codependency turns coping behaviours into personality traits. The research suggests we should often understand them as attempts to manage extremely difficult circumstances.

The harm experienced by families is real

One thing the codependency movement unquestionably recognised was that alcohol and drug problems do not affect only the person consuming the substance. The effects on families can be profound.

A 2020 review of 56 studies found affected family members experienced increased stress and burden, poorer mental health, reduced quality of life and social adjustment, with qualitative studies also describing stigma, self-blame and social isolation. (PubMed)

These are not trivial effects.

Concerned others can become anxious, depressed, frightened, socially isolated and physically exhausted.

Their finances may collapse.

Relationships deteriorate.

Children may be affected.

Violence and aggression may occur.

Their lives can gradually become organised around predicting and responding to somebody else's behaviour.

But recognising these consequences does not require us to diagnose the family member with codependency. Indeed, doing so may distract us from what is actually happening.

If a woman experiencing depression, anxiety, trauma or chronic stress is instead told that these experiences are symptoms of her codependent personality, an explanatory label can obscure conditions for which established treatments already exist.

Children are affected — but they are not programmed

The same caution applies to children raised around problematic alcohol or drug use. Parental substance problems can profoundly affect childhood. There is strong evidence that family environments influence both risk and resilience and that family processes can affect the development of substance problems, particularly among young people. (PubMed)

But risk is not destiny.

This became particularly important because the adult-children movement sometimes presented remarkably deterministic accounts of what happened to children raised in alcoholic or dysfunctional families. Velleman and Orford's research provides an important corrective. Their study comparing young adults who had grown up with parental drinking problems with a comparison group found substantial differences in reported childhood adversity and family disharmony.

Yet overall adult adjustment was remarkably similar between the groups. (the University of Bath's research portal)

Childhood experience matters enormously.

But people also develop.

They adapt.

They encounter new relationships and environments.

They acquire new skills.

They construct lives that are not simply repetitions of their family of origin.

Psychological theories should describe risk without converting it into destiny.

Families can influence addiction without causing it

There is another important distinction. Rejecting codependency does not mean pretending families are irrelevant. Families influence behaviour. All families do.

Family relationships can increase or reduce risks associated with substance use. Family members can inadvertently reinforce behaviours. They can also become enormously important resources for change.

Modern family interventions explicitly use these relationships therapeutically. Reviews identify interventions designed to help relatives encourage treatment engagement, participate constructively in treatment and improve their own wellbeing independently of whether the substance-using relative changes. (PubMed)

What the evidence does not require is the enormous additional leap from:

family behaviour can influence substance use

to:

the codependent personality of the family member causes the addiction.

Influence is not causation.

And coping is not pathology.

Help the concerned other

Perhaps the most important development in modern research has been the recognition that concerned others deserve help in their own right. Not merely because changing their behaviour might make somebody else stop drinking. Not because they need to recognise an underlying disorder. And not because their help-seeking demonstrates that the entire family is sick. They deserve help because living alongside serious substance problems can be extraordinarily difficult.

A 2023 systematic review and meta-analysis examined 19 studies of psychosocial interventions for family members affected by somebody else's substance use. Individually delivered interventions reduced depression and distress, while studies also found improvements in coping and family functioning. The authors cautioned that many studies were small or methodologically limited, but the overall direction was favourable. (DOI)

Again, this requires no assumption of codependency.

It requires us to listen to what the person is experiencing, examine what they have tried, identify what is and isn't helping, strengthen support and help them rebuild areas of life that have been progressively consumed by another person's substance use.

Why did codependency become so successful?

It would be easy simply to ridicule the codependency movement. That would miss something important. Codependency emerged because there was a need.

As alcohol and drug problems became increasingly visible, professional treatment remained overwhelmingly focused on the person using substances.

Partners, parents, children and other family members were frequently offered very little.

They might be told to detach.

They might be advised not to interfere.

They might be told the drinker had to “hit bottom”.

But there was often remarkably little practical help for the person going home that evening to deal with the consequences.

The codependency movement stepped into that vacuum. It said:

You matter too.

It acknowledged that people could lose themselves while trying to save somebody they loved. It gave people a language through which to talk about relationships, boundaries and self-neglect. It created groups where experiences that had previously been hidden could be discussed openly. For many people that was enormously valuable. And that helps explain why the concept became so powerful.

It was addressing a problem that professional services had largely neglected. The mistake was not noticing the suffering. The mistake was turning one explanation for that suffering into a disease.

The myth of codependency

Psychology needs theories. Human behaviour is extraordinarily complicated. We make progress by looking for patterns and developing ideas capable of explaining them.

Psychology is often exceptionally good at generating those ideas.

Where the discipline has sometimes been less successful is ruling out the ideas that don't survive testing.

That second process is every bit as important. A psychological theory should be capable of being wrong. We should be able to define what evidence would cause us to abandon it.

When a theory can explain one behaviour and its opposite, acceptance of the theory and rejection of the theory, dependence and independence, responsibility and irresponsibility, it has stopped taking much scientific risk. Almost anything can be made to fit.

The history of codependency illustrates this beautifully. Researchers did eventually confirm something extremely important. Living alongside serious alcohol or drug problems can have profound effects upon families.

People experience stress and strain. They develop ways of coping. Some are effective.

Some are ineffective. Some inadvertently perpetuate difficulties. Social support matters.

Professional support can help. Families can influence recovery. Children can be profoundly affected while also showing remarkable capacity for later adjustment. All of that is supported by credible research. What we do not need in order to understand any of it is a hidden disease called codependency.

Codependency emerged in response to a genuine need and, for many people, the movement provided support when professional services provided almost none. That contribution should be recognised. But compassion for the people helped by an idea cannot become evidence that the idea itself is true.

Psychology must do both parts of its job. It must generate explanations.

And it must be willing to let them go. Concerned others do not need a theory that tells them what they secretly are. They need us to understand what has happened to them, what they have done to survive it, what is helping, what is no longer helping, and how they can begin rebuilding their lives.

We do not need to diagnose them.

We need to listen to them.