Sense of coherence: the overlooked ingredient in sustainable change

Aaron Antonovsky was working through survey data on an ordinary subject: how Israeli women of different backgrounds were adapting to menopause. Among the women who answered were some who had survived Nazi concentration camps.

The expected finding was there, and it was stark. As a group, the survivors were in worse health than the women they were compared with, on both physical and emotional measures. What Antonovsky couldn't leave alone was the rest of the result. A substantial minority of the survivors were in good health. Not merely undiagnosed. Adapted: working, raising children, taking part in a life1.

Under the model the field was running on, that group was noise. Stress causes disease; extreme stress causes extreme disease. So the survivors who were well must have been in denial, or lucky, or wrongly counted.

Antonovsky went the other way. He treated the odd result as the finding, and spent his life on the question it raised. Not what makes people ill. What makes people well, even under conditions that should have destroyed them.

His answer is called sense of coherence. It names a lasting sense that life makes sense, that you can meet what it asks, and that doing so is worth the effort. Few measures of how a person feels have been tested as widely, or predict health and coping as steadily.

Yet it's almost absent from the way behaviour change is designed, sold and taught. That absence is the subject here. The claim is one a reasonable person can argue with. Whether a change holds depends less on the quality of the plan than on whether the person's life currently makes sense to them. Our advice perfects the plan and stays blind to what has to be there first.

The question that got inverted

Medicine is built around one approach: find what has gone wrong, name it, remove it. In an emergency this is the only thing that works, and a framework that forgets it isn't worth reading.

But it carries an assumption: that health is the absence of disease. Antonovsky's data wouldn't sit inside it. If stress causes illness, exposure should predict outcome, and among his survivors it didn't. Something was sitting between what people endured and what it did to them.

Health and disease, he argued, aren't two states but one line, and everyone is somewhere on it, moving. So what pushes a person toward the well end, given that nobody is ever at either end? He was asking why, given how hard the world is, anyone is well at all.

Ordinary medicine pulls people out of a river downstream, and preventive medicine asks who pushed them in. Antonovsky said there's no bank to stand on. Everyone is in the river, and what matters is who swims better2.

Three questions the body asks before it agrees to anything

Sense of coherence has three parts2. Antonovsky built a survey to capture them. The framing is more useful than the tool.

Comprehensibility — does life make sense? Do events feel ordered, or do they arrive as noise? This doesn't ask for certainty, only for enough structure. Think of a week of conflicting orders, or a month of symptoms nobody explains. The tiredness that follows isn't weakness. Being unable to predict is costly to the body, whether or not anything bad happens.

Manageability — can I meet it? Are the resources enough for the demands? Hard things can feel manageable when the support is there, and small things can feel impossible when it isn't. What matters is the ratio, not the difficulty. It's why the same request lands as interesting on one Tuesday and unbearable on the next.

Meaningfulness — is it worth doing? Antonovsky held this the most important of the three, and his reason was structural rather than sentimental. The first two describe whether a thing can be done. The third decides whether the system will bother. Without it, effort is a cost with no return, and living systems don't fund those for long.

The three interact, and they fail in ways worth telling apart. You can see a problem in full detail and have nothing to meet it with. You can have all you need and no reason to spend it. You can care about a situation you can't read.

Read those three questions against the way health advice arrives. Almost all of it answers a fourth one: what should I do. That's not what the person is asking.

What the evidence will carry, and what it won't

The record is substantial. Monica Eriksson and Bengt Lindström pulled together well over a hundred studies. Higher coherence went with better health, as people reported it, and the link was strongest for mental health4. A second review found the same pattern for quality of life5. It has been tested for decades by people with no stake in the answer.

The steadiness across cultures is the strongest part of that record. A scale built by one researcher, with one population, then carried into dozens of languages, largely held4.

Now the limits. They matter as much as the result.

Most of that evidence is cross-sectional, which means each study looked at people at a single moment and can't tell us which thing came first. It's also mostly self-reported: the person filled in both surveys about themselves, and no one checked from outside.

One survey asks how manageable life feels. The other asks how healthy you feel. When the two line up, part of what has been shown is that people who are doing badly say so twice. The link to physical disease that a doctor can measure is thinner and less steady than the link to mental health and wellbeing.

The tool itself is contested. Antonovsky wrote a paper on its structure. The question he examined there has never been settled: does the survey measure one thing or three3? The statistical tests haven't agreed.

Scores also track closely with scores for anxiety and depression, closely enough that a serious reader has to ask how much separate ground they cover. Antonovsky's claim that this sense settles around age thirty looks wrong, too. Reviews suggest scores tend to rise across the lifespan4, which turns a fixed trait into a thing that moves.

One gap matters most. Whether coherence can be raised on purpose in adults is almost untested. Nearly all of the evidence describes people who already have more or less of it.

So: a strong link, unusually steady, found in country after country. Not a proven lever. Anyone who tells you that raising your sense of coherence will improve your health is stating a hypothesis in the grammar of a finding.

The arrow points both ways, and the data can't separate them

This is the honest centre of the problem.

The natural reading is that coherence protects health. The equally plausible reading is that health produces coherence. Someone who sleeps well, has money, isn't in pain and can see next year clearly has excellent grounds for finding life clear and manageable. Their high score may be an accurate report of good circumstances rather than a resource that produced them. Work on wellbeing and ageing finds the link running both ways12.

Both are probably true, and they probably feed each other. Keeping the body's stress systems switched on has a real cost, and that cost builds up instead of clearing between episodes10. A life that never makes sense is a life spent braced. But a plausible mechanism isn't a proven direction, and these studies mostly took both readings on the same afternoon.

What survives that doubt is worth keeping. Even if coherence is mostly an effect, it's still an unusually good sign of the conditions a person is living in. It works as a reading whichever way the arrow points.

Coherence is made of resources, not resolve

Antonovsky's second book gave the idea its foundations, the part people usually skip2. Coherence, he argued, is built out of what he called generalized resistance resources. These are the supplies a life runs on: money, close ties, knowledge, a stable culture, bodily strength, ritual. Coherence is what having enough of them, over years, feels like from the inside.

So a person doesn't decide to have it. It builds up when a life holds enough of what a human being runs on. Social connection isn't a nice extra here. People with weak social ties die at higher rates than people with strong ones. The margin is like that of risk factors which carry public health campaigns11.

It also explains a pattern in the research that's usually read backwards. People who reach their goals report less effortful self-restraint than people who fail, not more13. That finding is small in scale and shouldn't be oversold. But it points the same way as work on self-determination theory9. What a person genuinely wants lasts far longer than what they do under pressure. Change that lasts looks less like sustained effort than like a person doing something that fits.

Prescribing change to someone whose resources can't fund it isn't a neutral act. It's prescribing an experience of failure, with a moral gloss attached.

Our reading: coherence isn't evenly distributed

Here the publication departs from Antonovsky, and the departure should be labelled. He measured sense of coherence globally, as one stance toward life as a whole, and the evidence rests on that global measure.

We don't think it works that way in practice. This next part is our reading, not a finding. Coherence appears to be specific to areas of life. A person can find their working life entirely readable: the rules clear, the resources enough, the work worth doing. The same person can find their own body close to unreadable, sending signals that mean nothing and energy that comes and goes without explanation.

If that's right, it predicts a useful thing. Change will hold in the areas a person already finds coherent, and slide off the ones they don't. That holds no matter how much they want it, or how good the plan is. Where they have no footing, they start four times and stop four times, then conclude they're the problem.

We state this as a reading of patterns, not a tested result. It hasn't been shown in the form set out here, and should be held that loosely. It's at least specific enough to be wrong. It predicts that scores taken area by area would beat one global score at forecasting where a change sticks. Nobody, as far as we can establish, has run that test. Until someone does, this is a lens and not a result.

Why the advice was fine and the change failed anyway

The standard account of failed health behaviour puts the shortfall inside the person. Not enough discipline. Not enough commitment. Not enough want. The coherence frame offers a less flattering reading.

An instruction arrives. Eat this way. Train like this. Sleep by then. It enters an area the person finds unreadable, thin on resources, or not obviously worth it. The instruction may be excellent. It lands in a system with no way to make sense of it, no capacity to fund it, no reason to put it first. It doesn't fail because it was wrong. It fails because nothing was in place to receive it.

This also explains what the field files under differences between people and leaves alone. The same advice works very well for some and not at all for others. The ones it works for tend to be the ones who already had coherence in that area. The plan takes credit for an outcome the groundwork was carrying.

There's a second pattern here, routinely mistaken for success. When one area is coming apart, a person leans harder on the one where their footing is firm, because effort there still brings a return. The training gets more careful while the relationships thin out. The load widens while the base narrows.

Meaning is the load-bearing wall

Of the three parts, meaning most resists being engineered, and it's the one Antonovsky weighted heaviest.

Viktor Frankl wrote from inside the same catastrophe Antonovsky was studying from outside. He argued that finding meaning in one's circumstances worked as a resource for survival, not as comfort6. That's testimony and philosophy, not data, and it should be read that way.

But the research has moved some distance toward it. A sense of purpose predicts lower death rates across adulthood. Purpose was measured first and the deaths counted later, with other factors held constant7. Carol Ryff's work on psychological wellbeing made purpose one of its core parts8. That work treats meaning as a feature of a working life rather than a mood.

None of that tells you how to acquire meaning if you don't have it. The honest answer is that we largely don't know. What the evidence supports is narrower. Meaning isn't decoration. When it's missing, the system's answer to "is this worth the energy" is no. No amount of tactical improvement to the plan changes that answer.

The implication runs further than it first appears. Say coherence has to be there before change, not after it. Then much of what we record as adherence is measuring the wrong thing. Not whether the instruction was good, but whether it reached someone in a position to receive one. The person who couldn't is filed under weak will and sent away with a better plan. It'll arrive in the same area, meet the same absence, and fail the same way. Nobody in that exchange is looking at the only thing that was ever going to decide it.