Possible Selves

In the AI Story

Before AI, the distance between a possible self and a testable provisional identity was measured in years. The backend engineer could imagine building interfaces, but getting there required months of study, practice, and the accumulation of a second body of technical knowledge layered on top of the first. The imagination-to-artifact ratio was large enough that most possible selves remained precisely that: possible but untested. The gravitational pull existed, but the orbital distance was too great for contact.

AI collapsed the orbit. The backend engineer who, with Claude Code, builds a complete user-facing feature in two days has made physical contact with a possible self that was, until that moment, theoretical. She has worn the identity of "someone who builds interfaces." The possible self has become, however briefly, a provisional self — an identity inhabited rather than merely imagined. This is new not in concept but in speed. The acceleration is genuinely generative, because Ibarra's research consistently shows that the primary obstacle to successful career transition is not the absence of good options but the inability to test them.

The population of possible selves available to any given professional has expanded dramatically. The developer in Lagos, whose possible selves were previously gated by access to capital, training, and institutional infrastructure, can now test identities that were structurally unavailable a decade ago. This is the democratization of capability read as identity theory — what is being democratized is not merely the ability to build things but the ability to explore who you might become.

The risk that accompanies this expansion is what Ibarra's framework would describe as possible-self inflation: the accumulation of possible selves at a rate that outpaces the reflective capacity to evaluate them. When experiments can be run back-to-back, the person may accumulate experiences at a speed the identity system cannot process, producing a condition of identity diffusion — rich in data, poor in integration.

Origin

Markus and Nurius's 1986 American Psychologist paper introduced possible selves as a bridge between self-concept research and motivation theory. The authors argued that self-concept contains not only representations of the current self but also of selves the person could become, wanted to become, or feared becoming — and that these future representations exerted real behavioral effects. Ibarra adopted the construct in her 1999 Administrative Science Quarterly paper on provisional selves and made it central to her subsequent work on career transitions.

Key Ideas

Possible selves are testable hypotheses. They are not fantasies but working propositions about future identity that can be empirically evaluated through identity experiments.

Gravitational pull on present behavior. Possible selves shape which risks feel worth taking and which efforts feel worth sustaining, regardless of whether they are ever realized.

AI collapses the testing distance. What previously required years of skill acquisition before a possible self could be tested now requires a conversation. The orbital mechanics of identity exploration have been fundamentally altered.

Democratization of exploration. The population of people who can test possible selves has expanded beyond the historically privileged, though access remains unequal along lines of connectivity, language, and infrastructure.

Speed without integration produces diffusion. The abundance of testable possible selves becomes pathological when the rate of testing exceeds the rate of reflective integration.

Debates & Critiques

Critics have questioned whether the concept of possible selves, developed in a slower era, remains coherent when possible selves can be sampled at industrial speed. Does a possible self briefly inhabited through an AI-assisted weekend project carry the same developmental weight as one pursued through months of deliberate effort? Ibarra's framework suggests not — the possible self only becomes identity-forming when tested repeatedly across varied conditions. A second debate concerns whether AI-mediated testing produces authentic contact with a possible self or merely simulates contact. The backend engineer who builds an interface with Claude's help has demonstrated capability but may not have experienced the identity-forming struggle that unmediated attempts would have required.