Why We Don’t Change—Even When We Know What’s Wrong |
Insight alone rarely leads to change, even when patterns are clearly understood.
The brain prioritises predictability, even when those predictions can cause pain.
Beliefs formed early in life are reinforced and resist contradiction over time.
Change requires experiences that feel real and salient enough to challenge expectations.
People often assume that insight leads to change. If you can see the pattern clearly enough—why you push people away, why you shut down, why relationships feel unsafe—then things should start to shift.
In practice, it rarely works like that.
A therapist had been working for years with a young man whose childhood had left him certain of one thing: that asking for care was pointless, and probably dangerous. He had learned to protect himself by not really looking at people, not in the moments that mattered. One day in session, mid-disclosure, he commented that the therapist looked flat, disengaged.
She noted, gently, that he hadn’t actually looked at her in quite some time. She asked him to look now.
He did. What he saw stopped him—her face was warm, concerned, unmistakably present. He said it felt like his mind had played a trick on him.
It had. And in some ways, that trick is the whole problem.
Why harmful patterns are so hard to shift
People often know, in the abstract, that their patterns aren’t helping them. They know they push people away, or collapse in relationships, or can’t ask for help without feeling exposed. Insight isn’t necessarily the issue. So why doesn’t knowing seem to change anything?
The neuroscience framework of predictive processing provides an explanation here. The basic idea is that the brain doesn’t just take in information from the world—it’s constantly generating predictions about what it’s about to encounter, based on prior experience, and checking those predictions against what actually happens. Where there’s a mismatch—a prediction error—the model updates. That’s learning in its simplest form.
The problem is that beliefs formed in adverse environments don’t simply fade over time. They are repeatedly confirmed by the kinds of situations in which they first developed. If you grow up expecting people to be critical, unavailable, or unsafe, you are likely to encounter, or interpret, interactions in ways that reinforce that expectation.
Over time, the system becomes less open to contradiction. When something doesn’t fit the belief, when someone is actually kind or available, it may be dismissed, minimised, or explained away. The new information doesn’t fully register.
So what can look like someone being stubborn or unwilling to see what’s in front of them is, in fact, something more automatic: The mind is relying on what it already “knows”, rather than updating itself on the basis of a single, inconsistent experience.
Here’s the part that takes a moment to sit with: Such beliefs don’t persist only because people can’t change them. They also persist because they serve a function—they offer something the brain values: predictability. A world in which people are reliably cold or unavailable is at least a world you can anticipate. You know how to move in it. Stability, in that sense, can matter more for survival than accuracy.
Letting go of that expectation isn’t just about “updating a belief.” It means stepping into something less certain. You no longer know how the other person will respond or how you should respond in turn. That uncertainty can feel more threatening than staying with something painful but familiar.
What this framework misses
Predictive processing offers a strong account of how people get stuck, and also suggests what therapy needs to do. For beliefs to shift, something has to happen that doesn’t fit what the person expects, and it has to feel real and salient enough to matter. For example, if someone expects others to withdraw when they open up, and the therapist instead stays engaged, attentive, and steady, that creates a mismatch—in predictive processing terms, that’s a prediction error—and those moments are what allow beliefs to start changing.
But this only takes us so far.
Predictive processing explains how beliefs change when something doesn’t match what we expect. It gives us a clear mechanism for why new experiences sometimes lead to learning. But it leaves a crucial question unanswered.
If these beliefs provide stability—if they help us anticipate the world, even in painful ways—why would anyone move toward situations that might challenge them? Why risk stepping into something uncertain, where the old rules may no longer apply?
In other words, predictive processing explains how change happens. But it says much less about why people would move toward that change in the first place.
What fills this gap—and why do people not just passively update their beliefs but actively test whether it is safe to change? More, in my next post.
Li, E., McCollum, J., Krieger, J., Winter, S. E., Duane, D., & Silberschatz, G. (2025). Predict to control, test to master: Integrating predictive processing and control–mastery theory in understanding how psychotherapy works. Journal of Psychotherapy Integration. Advance online publication. https://dx.doi.org/ 10.1037/int0000386
There was a problem adding your email address. Please try again.
By submitting your information you agree to the Psychology Today Terms & Conditions and Privacy Policy