Why behavioural confirmation requires a power asymmetry—and how a target's self-concept breaks the loop.
The Myth of Automatic Confirmation
In both cognitive science and algorithmic design, the self-fulfilling prophecy is often treated as a frictionless loop. A perceiver holds an expectation about a target, behaves in accordance with that expectation, and subtly coerces the target into confirming the initial belief. In automated recommendation and user-profiling systems, this is the standard critique of predictive personalization: the model assumes a user prefers a narrow band of content, serves only that band, and points to the resulting clicks as proof of the user's underlying disposition.
Yet empirical social psychology demonstrates that behavioural confirmation is neither automatic nor symmetrical. As research on social perception makes clear, expectations do not effortlessly rewrite the behaviour of the person on the other side of the interaction. Whether a prophecy fulfils itself depends on two strict boundary conditions: the distribution of power in the encounter, and the structural certainty of the target's own self-concept.
Copeland’s Asymmetry: The Role of Power
In 1994, John Copeland tested whether a perceiver's expectations alone are sufficient to elicit confirming behaviour from a target. Perceivers were led to believe that the person they were about to interact with was either introverted or extraverted. Crucially, Copeland manipulated the relative power between the two participants: in half of the pairs, the perceiver held the power to accept or reject the target as a teammate for a cash-prize game; in the other half, the target held that power over the perceiver.
Independent observers who rated the recorded interactions found what Copeland termed prophecies of power. High-power perceivers reliably triggered the self-fulfilling prophecy, steering targets into behaving as introverts or extraverts according to the initial prime. Low-power perceivers did not. Stripped of structural leverage, low-power perceivers spent less time probing the target and more time attempting to ingratiate themselves and be liked. An expectation only reshapes a target's behaviour when the perceiver controls the payoff matrix of the interaction.
Self-Verification and the Resistant Target
Even when a perceiver holds an expectation, the second link in the chain—the target's behavioural compliance—is routinely broken by the target's internal priors. Most classic demonstrations of the self-fulfilling prophecy, such as Rosenthal and Jacobson’s (1968) classroom studies or Snyder and Swann’s (1978) interview paradigms, relied on targets remaining ignorant of the false impressions held by their evaluators.
When targets become aware that they are being miscast, the mechanism reverses. Hilton and Darley (1985) demonstrated that participants who knew an perceiver held a false expectation actively behaved in ways that forced the perceiver to abandon the preconception. Furthermore, as Swann and Ely (1984) showed in their work on self-verification, when a perceiver's expectation clashes with a target's confident self-concept—such as an interviewer expecting introversion from a target who strongly identifies as extraverted—it is the perceiver's belief that gives way, not the target's behaviour.
Implications for Algorithmic Profiling
When we build or evaluate systems that model human agents—whether in interactive media, behavioural targeting, or AI alignment—treating the human user as a passive surface for behavioural confirmation is a category error. A system only shapes user behaviour when it holds gatekeeping power over an outcome the user cannot get elsewhere. Where the user retains agency and a high-certainty self-concept, heavy-handed predictive framing does not produce compliance; it triggers active behavioral correction.
In our work with the NeuraScene engine at Arcanamatic, this distinction governs how we structure narrative and commercial persuasion. You cannot brute-force an audience into a behavioural role by projecting an expectation onto them. Social perception is a two-way street: the people we judge have their own prophecies to fulfil.
Method note. This piece was built in NeuraScene's Advertising Mode using an explicit psychological frame:
- Persuasion trigger: Anchoring
- Target affect: Flow State
- Gestalt principle: Common Region
Anchoring frames the core theme of how initial expectations set a baseline in social interactions before power and self-verification intervene. Common Region visually binds the high-power inspector and low-power promoter inside the single circular boundary of the manhole rim, guiding the viewer's eye directly to the contested interaction while maintaining the cognitive absorption of a Flow State.
