Imagine your interactions with someone as colored marbles in a jar. Each encounter adds a new marble: red for moments of connection, support, and trust, and blue for disappointments, conflicts, or betrayals. Under normal circumstances, drawing an occasional blue marble doesn’t force you to discard the many red ones; instead, you update your overall picture, acknowledging that people can be both kind and hurtful in different situations.

When operating optimally, the Bayesian method lets you refine your understanding gradually. For example, if your trust in someone is at 95% and they become temporarily unavailable during a crisis, you might adjust that trust slightly—to 90%. However, when emotions run high, the weight assigned to new information can be dramatically altered. In those moments, minor evidence may receive disproportionate emphasis, overwhelming all previous data.

Consider how this plays out with our schemas: imagine you’ve always believed that hard work guarantees success. Then you encounter someone who worked diligently for years but faced setbacks due to economic downturns or structural barriers. This creates cognitive dissonance. You now have two options: Schema assimilation (forcing the new information into your existing framework by thinking, “They must not have worked hard enough”) and schema accommodation (modifying your belief to “Hard work is important but doesn’t guarantee success; other factors matter too”).

Most of us can eventually accommodate this contradiction because, while it challenges our worldview, it doesn’t threaten our fundamental sense of safety. This is healthy Bayesian updating—incorporating new evidence to refine our models of reality without wholesale rejection of previous understandings.

At the core of this phenomenon is a computational mechanism known as precision weighting. Think of it as an amplification system that determines how much certainty we assign to our beliefs. This mechanism evolved to help us stay safe by prioritizing signals that were once crucial for survival.

Under normal conditions, drawing a blue marble here and there doesn’t upend our overall positive view. Precision weighting remains flexible, allowing our jar to contain a balanced mix of experiences. However, when early experiences with caregivers involved trauma or chronic shame, people develop a fundamentally different way of processing social information. Instead of maintaining one continuously updated model, they develop two separate schemas (all-good or all-bad), with the ability to suddenly transform between these extremes, and assign extremely high certainty to whichever model is currently active.

Under intense stress, things start feeling irreconcilable; the evidence, no matter how contradictory, is seen as incompatible with the new model. In this state, the mind constructs a representation with such unwavering certainty that it completely overwrites the previous version, dismissing any conflicting signals. It’s as if, upon encountering a significant blue marble, the jar suddenly empties of red ones, leaving behind a single, unchallengeable reality.

This is exactly what happens when a close friend fails to support in the way you may have desired during an important life event, like a graduation or a wedding. Despite years of reliable care, a single instance of mismatched support can trigger a complete model swap. Instead of gradually adjusting your belief, your trust plummets to zero, and you become absolutely certain: “They don’t care about me. They never did.”

This reaction feels entirely justified in the moment – after all, what could be more important than being there for a once-in-a-lifetime event? The thought process is absolute: ‘Someone who truly cared about me would never let me down this way. Since they did let me down, they must have never actually cared.’ There’s no middle ground in this moment. The very definition of caring becomes incompatible with their behavior, making it impossible to maintain the previous view of them as someone who values you. Your mind must resolve this contradiction, and it does so by completely rejecting the ‘caring friend’ model. This isn’t a choice – it’s your brain’s attempt to make sense of an experience that otherwise feels unbearably confusing.

What makes this perceptual shift so dramatic is the certainty with which the individual dismisses all previous evidence of the friend’s care and commitment. That time they congratulated you on your promotion becomes seen as manipulation (“they just wanted to stay close to someone successful”), deep conversations are reinterpreted as mere data collection for future psychological warfare, and consistent support is dismissed as a calculated façade to create reciprocal debt. The birthday card where they wrote “you mean the world to me”? Classic love-bombing technique straight from the Manipulator’s Handbook, chapter 3, paragraph 2.

This isn’t just selective interpretation; it’s your brain shifting between two entirely separate models for understanding the same person. Instead of maintaining a complex, integrated representation where your friend is generally caring despite occasional shortcomings, your mind activates a different schema—one perhaps rooted in prior experiences with a caregiver characterized by neglect or abandonment. It’s like your brain suddenly decides your friend isn’t just having a bad day—they’re actually a secret agent whose decade-long mission was to infiltrate your life just to disappoint you at this precise moment.

In that heated moment, what’s missing is context—the deeper story behind your friend’s disappointing behavior. Maybe they’re quietly struggling with exhaustion or personal issues you know nothing about. Perhaps they genuinely thought they were supporting you, unaware their efforts didn’t align with what you needed. Or maybe you never clearly communicated your expectations in the first place, unintentionally leaving them guessing. It’s even possible that you’re feeling more overwhelmed than you realize and unknowingly projecting some of that internal chaos onto your friend. And sure, maybe they’re actually a bit of a shitbag and you’re just now noticing—that happens too! The point is, you simply don’t know yet. Problems arise when absolute certainty takes hold, leaving no space for nuance, reflection, or empathy. Certainty locks you into rigid, black-and-white conclusions about who your friend “really is,” whereas embracing uncertainty gives you room to consider alternative explanations and genuinely understand what happened.

This intense drive toward certainty often reflects something deeper: it may have roots in early relational experiences, especially those involving caregivers. When we leap to rigid conclusions about someone’s intentions or character, we might actually be replaying a fundamental confusion learned long ago.

This breakdown often originates in a profound early double bind: learning to find safety in fundamentally unsafe people. Picture a world where the person meant to be your safe haven is also the source of your deepest hurt. For a developing mind, the core challenge isn’t just reconciling contradictory perceptions—it’s finding safety and security when your only available attachment figure is sometimes dangerous.

The child faces an impossible dilemma: they need the relationship for survival but must also protect themselves within it. To resolve this, our mind splits the experience into two divergent narratives that remain entirely separate. One narrative constructs an idealized “good parent,” a comforting illusion that provides hope when love is most needed. The other preserves the memory of the “bad parent,” capturing the raw, painful reality of harm. Because maintaining both views simultaneously would make attachment impossible, these experiences become compartmentalized. The child might even see themselves as “all bad” to preserve the caregiver’s goodness, believing “if I’m the problem, they are safe.” When we need to preserve the illusion of a “good parent” despite evidence of harm, we often resolve the contradiction by blaming ourselves. If mom can’t control her rage, then perhaps you provoked it; if dad couldn’t protect you, maybe you weren’t worthy. This early maneuver sets the stage for later relationships, where we instinctively assume fault to maintain attachment.

A parallel pattern emerges when children sense their authentic selves are unacceptable to the caregivers they depend on. To cope, these children develop an adaptive self—a constructed version that meets others’ expectations at the expense of authentic needs and feelings. Children who grow up with caregivers who demand they meet adult emotional needs develop adaptive strategies to earn conditional love—striving to be the perfect, trouble-free child. The specific form this adaptive self takes varies depending on what each caregiver required—perhaps becoming the emotional caretaker, the academic achiever, or the child who never causes problems. This represents another kind of split, but now in our self-concept: an idealized, acceptable self versus a hidden, authentic self. The lesson becomes: you can be loved, but only for being someone you’re not.

And here’s a truth we need to shout from the rooftops: the good parent doesn’t exist, and neither does the perfect version of ourselves or anyone else we construct in our minds. These idealized versions are cognitive constructions—safety mechanisms that once protected us but now limit our ability to embrace the beautiful messiness of real human connection.

These patterns maintain themselves through precision weighting. The mind assigns extremely high certainty to whichever schema preserves emotional safety, distorting evidence to fit rather than updating the schema. This isn’t irrational—it’s a logical response to overwhelming childhood threats.

When the “dangerous other” representation activates, even kindness becomes suspect as the mind rewrites history, recasting support as manipulation. Because abandonment feels life-threatening, people often assign high precision to their own “badness” to maintain the relationship. Being “bad” feels safer than being abandoned. It’s like thinking, “If I’m the problem, at least I can still have people in my life—even if they’re just sticking around to remind me how terrible I am.”

The adaptive self requires constant vigilance because the underlying shame is tied to attachment. Revealing one’s true self risks losing vital love. Anyone threatening to expose inauthenticity becomes an existential threat. Devaluing them safeguards the idealized self and fragile attachment. Tragically, this undermines genuine connection, as real love requires the authentic self. It’s like wearing a full-body costume to every interaction and then wondering why no one gives you a proper hug.

What’s truly chilling is being on the receiving end of someone’s devaluation. Research demonstrates that when someone persistently views you through a devalued lens, it can fundamentally alter your self-understanding—even if you once had a robust self-concept. It’s not merely unpleasant; it’s computationally destabilizing to your sense of who you are. When someone operating from a negative split model encounters positive evidence about you, they don’t update their model—they reinterpret the evidence to fit their existing narrative. This creates an impenetrable interpretive wall where every action you take becomes further proof of their negative view.

Consider how this works with even something as simple as a thoughtful birthday gift. When splitting is active, the gift’s interpretation flips entirely: “This isn’t thoughtfulness—it’s a calculated manipulation strategy with a gift receipt. They chose a hardcover book to literally weigh down our relationship. They signed the card with just ‘Love’ instead of ‘Lots of Love’—that’s a 33% reduction in affection from last year! And they wrapped it in blue paper when everyone knows blue is the official color of emotional sabotage and passive-aggressive hydration. They bought me a plant? Obvious metaphor for how they expect this relationship to wither and die. Chocolate? They’re clearly implying I need to sweeten my personality. What’s fascinating isn’t just the reinterpretation, but how it feels like you’ve cracked a Da Vinci-level code so obvious that only a fool would miss it.

The language of devaluation often contains extreme terms like “pathetic,” “disgusting,” or “worthless”—words that signal not just disagreement but a fundamental rejection of the other person’s humanity or value. These aren’t casual descriptors; they’re markers of a mind that has shifted into seeing someone as irredeemably bad—as if they’ve been caught sacrificing puppies to summon demons, not just forgetting to text back.

Split representations create distinctive relational dynamics. When two people with these patterns interact, they activate each other’s defenses. One person, accustomed to viewing themselves as “bad,” may readily accept devaluation, while someone prone to devaluing others might seek partners who reinforce that negativity. A sudden shift to negativity in one partner can trigger self-blame in the other, leading to a rapid escalation of mutual fears and a toxic equilibrium—like a chaotic dance where both are locked into self-reinforcing maps of reality for years. Picture a relationship where one person says, “You’re 15 minutes late; you clearly don’t respect my time,” and the other responds, “You’re right, I’m terrible, I don’t deserve you anyway.” Instead of a simple conversation about traffic, we’ve rapidly escalated to existential relationship doom. It’s like watching a tennis match where each player is determined to hit the ball into their own court.

It’s important to distinguish between the distortion of splitting and legitimate recognition of disappointing behaviors. Feeling let down by someone’s actions is completely natural, especially in two types of situations that deeply activate our attachment systems: First, during meaningful milestone events where we expect celebration and acknowledgment. Second, during personal crises—times of illness, loss, or significant transitions—when we’re at our most vulnerable and support feels essential. When someone we trust doesn’t show up in the ways we hoped during these moments, the disappointment is particularly valid.

The challenge lies in how we process this disappointment. Sometimes patterns of behavior do warrant reconsidering a relationship. The key difference between healthy disappointment and splitting isn’t about whether we feel hurt—it’s whether we can still see the whole person, rather than reducing them to their best or worst moment. Healthy responses allow us to say, “I wish they had acted differently, and we may need to talk about it,” while still acknowledging, “This person has been caring in many other instances.” It’s about maintaining a balanced perspective while still honoring our authentic feelings.

So, what can be done about all this?