Why We Never Forget Our Starter

There is something unreasonable about how strongly people feel about a fictional creature they chose at age nine. Decades later, adults who have not touched a Game Boy since middle school can still name their starter without hesitation. Bulbasaur. Charmander. Cyndaquil. The memory is precise in a way that most childhood memories are not. Something about that first choice lodged itself into long-term storage and stayed there. This is not nostalgia in the ordinary sense. Nostalgia is usually diffuse — a warm feeling about a period rather than a specific object. Pokemon starter attachment is different. It is targeted. It has a face.

The Mechanics of the Bond

Pokemon was designed around a single constraint that turns out to be psychologically significant: you cannot have everything at once. You pick one. That scarcity creates investment in a way that abundance cannot. If the game gave you all three starters, none of them would matter as much. The choice forces you to identify with one creature and implicitly reject the others. That moment of identification is where the attachment begins. What follows reinforces it. Your starter is present for the entire early game — the tutorials, the first gym, the moments when you are still learning the systems. It accumulates shared history with you faster than any other Pokemon in your party. Psychologists call this the mere exposure effect, but it goes deeper here because the exposure is tied to challenge and reward. You struggled together. That is the substrate of bonding.

Parasocial Attachment Has Real Cognitive Signatures

Research from the University of Amsterdam found that parasocial relationships — one-sided emotional bonds with fictional characters or media figures — activate many of the same neural regions as real social relationships. The brain does not cleanly separate "real companion" from "represented companion" when the interaction is sufficiently rich and repeated. A separate study from Stanford's Human-Computer Interaction Group looked at how players narrated their relationships with non-player characters in long-form games. Participants consistently used relational language — words like "loyal," "trustworthy," "let me down once" — even when describing entities with no actual behavior beyond scripted responses. The attribution of personality and history was generated almost entirely by the player, not the game. This matters for understanding Pokemon specifically. Starters in the original games had almost no individual personality. They were sprites that followed commands. The personality players remember is the one they built themselves through projection and narrative.

The First Companion AI Question

What happens when the companion is actually responsive? AI companions today can remember previous conversations, adjust tone based on emotional cues, and develop what feels like a consistent personality over time. The question researchers are beginning to ask is whether starter-style attachment — that early, formative bond that persists across decades — is possible with AI, and what conditions produce it. The evidence suggests the conditions are similar: early exposure, repeated interaction during a learning or challenging period, perceived reciprocity, and scarcity of alternatives. When someone begins using an AI companion during a difficult transition — a new city, a job change, a period of social isolation — the attachment formation mirrors what happens with Pokemon starters more than it resembles typical software adoption.

The Tangent Worth Taking

There is a phenomenon in animal behavior called filial imprinting, famously documented in Konrad Lorenz's work with greylag geese. Young animals attach to the first sufficiently stimulating social object they encounter during a critical developmental window. The attachment is fast, deep, and resistant to revision. Humans do not imprint in the technical sense. But there may be softer versions of this — sensitive periods in childhood or during major life transitions when attachment systems are more open to new bonds. The Pokemon starter experience may be effective partly because it happens during exactly such a period for most players: early childhood or early adolescence, when identity is still forming and the emotional system is calibrated toward novelty. If that is true, then AI companions that enter someone's life during a transition period may form stronger and more durable bonds than those adopted during stable periods. The timing matters, not just the quality of the interaction.

What Stays

A study from the University of Tokyo on long-term game memory found that emotionally charged early-game decisions — including starter selection — were recalled with significantly higher accuracy than other game content, including plot events and boss battles. The choice was the memory anchor. This has implications for how we think about AI companion design. The first interaction is not just onboarding. It is potentially the moment from which everything else is organized in the user's memory. Getting it right — making it feel like a real choice, with real stakes — may matter more than any subsequent feature. People do not remember their starter because it was the most powerful or the most useful. They remember it because it was first, and they chose it, and it was there when everything was new.