Cognitive Dissonance and Anorexia: Optimizing for Tolerable

Part 4: The protracted dangers of partial dissonance reduction within illness.

Posted Nov 23, 2020

“The theory of cognitive dissonance obviously has many implications for everyday life. In addition to throwing light on one's own behavior, it would seem to carry useful lessons for everyone con­cerned with understanding human be­havior in a world where everything is not black and white.” — Festinger, 1962

Having set out the basics of cognitive dissonance and their relevance to eating disorders in the first three parts of this series (see 1, 2, and 3), in the last two parts, I’ll go into more depth on cognitive dissonance as a pro or anti recovery principle in anorexia by putting cognitive dissonance in dialogue with the concept of optimization. I’ll suggest that the automatic drive to dissonance reduction is itself in a meta-conflict with the more reflective desires we may have for our lives. The concept of optimization may allow for a more-than-partial, more-than-temporary resolution.

The most important implication of the theory is that dissonance generates ways of thinking and acting that are optimized for dissonance reduction. That is, whatever their other merits or drawbacks, if a strategy is successful at reducing dissonance, it has a high chance of being selected over other strategies that might reduce the dissonance less readily, even if they have profound advantages in other respects. Thus reducing dissonance may hijack, say, achieving health, by offering more immediately satisfying ways of improving how your life feels. 

In this sense, cognitive dissonance theory offers a way of reframing perhaps the most profound of the common contributors to eating disorder survival: the fact that attitude change (to make attitudes consistent amongst themselves) is a lot easier than behaviour change (making behaviours consistent with attitudes). This helps explain why it’s so common to get stuck for years in the insight-and-inaction stage: You’re not doing nothing in that phase — you’re practising all kinds of clever little tricks to try to reconcile the brutal opposition of (something more or less like) wanting to be happy and being addicted to something that prevents it. This feels like real cognitive work, because it is, but it’s also work that is optimized for perpetuating itself by sustaining the conditions in which it’s needed.

Because attitude change can only ever be partially successful if the physiology is crumbling around your ears, the asymmetry between the cultivated set of attitudes (e.g. any number of cognitive accommodations to the idea that this is as good as life gets) and the unchanged behaviours (which are very obviously not as good as life gets) constitutes an unresolved dissonance that can last for years or decades. The insight into illness grows and the impossibility of action does likewise, because growing insight into intolerability is repeatedly counterbalanced by tools for marginally increasing tolerability. 

The dissonance is not permanently soluble by anything but significant behaviour change, but there will always be another little variant to try that avoids this distasteful fact: another therapist to go and talk to about your childhood, another body image workbook to complete, another eagerly pocketed confirmation that people without eating disorders are unhappy too, another cultivated reinforcement of how much more sense it makes to work all evening than spend it with other people.

This activity is, crucially, achieving something for you: It is making the state of your life more tolerable to you. And that is precisely the problem: The more this state of your life is made tolerable, the more your motivation to seek something more than merely tolerable dwindles. By making you “not sick enough to get better,” your success in this phase of your life is your failure for the rest of your life.

The ease of eliminating dissonance by denying responsibility seems instructive here too, in understanding the persistence of the pre- or pseudo-recovery phase. We always do our best to avoid/minimize personal responsibility for unwanted action outcomes, and this is easily and rightly done with something called an illness. With any illness, regardless of where it straddles the physical/psychological non-divide, acknowledging that “it’s not my fault” is often an important precursor to ditching the shame and moving on. But anorexia has the uneasy status of an illness many of whose component parts (skipping the meals, never skipping the treadmill) may feel like choices made for personal and/or normative reasons. And the acknowledgement that I never chose this can expand into a convenient helplessness when it comes to recovery: I didn’t choose to get ill and there’s nothing I can do about still being ill now. I didn’t really choose to stop eating breakfast, and I can’t now start eating it again. In this case, the crucial distinction between fault (past) and responsibility (future) is being ignored. The liberation of ditching the fiction of free will for a more real, more acceptant, yet not morbidly fatalistic model of human agency is also being precluded. If you keep saying it’s not your fault that you continue not to get better (whether because of biological fate, or thanks to other people modelling such messed-up behaviours, or for any other reason), what is going on is, above all, comforting dissonance reduction that also has the side-effect of keeping you ill. 

Optimizing for “tolerable” is also happening when people engage over years or decades with therapeutic strategies that aren’t working for them. The effort justification effect is the finding I mentioned in part 1: that people come to like a goal they’re trying to reach as a function of the amount of effort they expend to reach it (as long as that effort is voluntarily expended). Talk therapy might involve unearthing painful memories, talking about things that make you feel shame, paying lots of money, etc., and might precisely thereby heighten the appeal of the goal you’re trying to reach: the fully recovered state.

This would be a splendid effect if the methods were viable, since the more you go through in the attempt, the more you’ll care about succeeding. But if the methods are not effective, effort can dangerously become mistaken for efficacy, and failure to recover be misinterpreted as a failure of effort, not method. Again, dissonance reduction efficacy isn’t the same as recovery efficacy, and by putting a lot of effort and resources into something that is convincingly enough like recovery (especially when it’s professionally packaged that way), you can reduce dissonance without the more frightening effort of changing behaviour. Thus you get hooked on the feeling rather than the results of hard work, and nothing much changes.

This fiddling-around-with-dissonance-reduction stage is likely exacerbated by the fact that dissonance, as an arousal state, may be dampened down by malnutrition and the associated depression, meaning more costly reduction strategies like behaviour change are less likely to be sought where they’re most needed. Nowhere in the cognitive dissonance literature have I found any mention of physiological change as a dissonance reduction option, mostly because physiological change tends to be driven by behavioural change. But it’s relevant here in the sense that changed behaviours (e.g. eating more, exercising less) have physical effects that can be powerfully and even implausibly dissonance-reducing: e.g. gaining weight from underweight generally makes you less obsessive about avoiding weight gain than when you were thinner. Your dissonant starting point might be, say, doing all you can to avoid gaining weight even while theoretically despising the idea of wasting a life (any life, let alone yours!) trying to be small. Getting bigger (as a central part of all the repair and regeneration involved in recovery from malnutrition) is what makes it possible to smooth away the agonising opposition into congruence. At long last, it’s possible for you not only to subscribe to all the reasons a life shouldn’t be reduced to weight control, but also to genuinely not care—because you’re living, every day, your ability to cope, and more than cope, see your life blossom. But if you’re too weakened and depressed to do more than attitude manipulation, you’re more likely never to get the chance to find this out. 

Finally, cognitive dissonance theory also supports the hypothesis that, as found in the Body Project intervention research, a higher degree of initial severity may be helpful in heightening the dissonance and change the reduction strategies at play. I’ve long been sceptical of the strident claims made for the “early intervention is key” principle, and in this framework, we can expect that the dissonance may need to reach a certain strength to be worth expending serious effort to reduce. In the honeymoon phase of weight loss, compliments, hunger highs, and a sense of heady specialness, the odd niggling little doubt about whether this is really OK can easily be resolved with a quick bit of attitude manipulation or selective evidence-seeking. If you’re alone, obsessed, depressed, frightened, bored, and yet can’t stop doing any of the things that make these facts remain true, then the fragility is far greater and the chance of choosing something to demolish not massage it far higher. In this, I think, there is more than a glimmer of hope that true intolerability will be the agent of its own demise—if dissonance reduction optimizing for tolerable doesn’t get there first.

Read on for the final part in this miniseries, on how to get to better than tolerable.


Festinger, L. (1962). Cognitive dissonance. Scientific American207(4), 93-106. Paywall-protected journal record here.