A trader predicts a bank will fail, and it fails — the next week, on schedule. A guru tells a room of followers they will "manifest" a specific outcome, and a suspicious number of them get it. A teacher is told a student is gifted, and the student's grades climb. Every one of these looks, from the outside, like proof that belief bends reality: that some people have a talent for making futures happen just by naming them. That reading is wrong, and the actual mechanism is more interesting than magic, because it is fully mainstream science — sociology, game theory, and economics that already won two Nobel Prizes for describing exactly this.
The theorem that started it
In 1928 the sociologist W. I. Thomas wrote the sentence that names the whole field: "If men define situations as real, they are real in their consequences." Robert K. Merton picked up the idea in 1948 and gave it its modern name, the self-fulfilling prophecy, illustrated with the example that is still the cleanest one available: a perfectly solvent bank, hit by a false rumor of insolvency, actually becomes insolvent — because the rumor changes behavior (depositors rush to withdraw), and the changed behavior produces the exact outcome the rumor predicted. Notice what did no work in that story: no one read the future, no one bent physical law, nothing paranormal occurred. A false belief altered real actions, and the real actions altered a real, checkable financial fact. That is the entire mechanism, stated in one sentence, in 1948.
The proof it isn't hand-waving
Merton's bank-run story stayed an illustration until 1983, when economists Douglas Diamond and Philip Dybvig built it into a formal mathematical model of banking — work that won them the Nobel Memorial Prize in Economic Sciences in 2022. Their result is exact and uncomfortable: an ordinary bank, taking in short-term deposits and lending them out long-term, has two mathematically legitimate equilibria given the identical same set of underlying facts. In one equilibrium, everyone expects everyone else to leave their money in, so no one runs, and the bank is fine. In the other, everyone expects everyone else to withdraw, so everyone withdraws — and the bank fails, even though nothing about its loans or its solvency has changed. Which equilibrium actually happens is not determined by the bank's fundamentals at all. It is selected by which future depositors, as a group, expect. Diamond and Dybvig didn't discover that beliefs are magic; they proved that some ordinary systems have more than one mathematically valid future, and shared expectation is the mechanism that picks which one gets realized.
Why the original, quotable version of a prediction matters
If shared expectation is doing the selecting, then the practical question becomes: how do thousands of independent people end up expecting the same thing, without a meeting, a vote, or a central coordinator? Thomas Schelling answered this in 1960 (Nobel Prize, 2005) with the concept of the focal point: when people need to coordinate but can't communicate, they gravitate toward whichever option is most salient, most obvious, most easy to name — not the option that's objectively best, just the one that's easiest for independent minds to land on simultaneously. This is the real, unglamorous skill behind a prophecy that "comes true": not clairvoyance, but producing a version of the prediction vivid and simple enough that a stranger who hears it once can repeat it accurately to someone else. A quotable frame is a coordination device. It is the same skill a slogan or a brand name requires, not a supernatural one.
Two more results explain how that initial coordination device snowballs into an actual outcome. Mark Granovetter's 1978 threshold model of collective behavior shows that populations with nearly identical individuals can produce wildly different collective outcomes — a riot versus no riot, a run versus no run — depending only on how many people are needed before the next person joins, a number that varies slightly from person to person. Once enough early adopters cross their own threshold, everyone with a slightly higher threshold follows, and the cascade completes itself; a prediction doesn't need everyone to believe it at once, just enough people to trigger the next ring of believers. Sushil Bikhchandani, David Hirshleifer, and Ivo Welch formalized the adjacent mechanism in 1992: rational people, watching the visible choices of the people before them, will rationally start copying those choices even when their own private information disagrees — an information cascade. A handful of early, visible followers of a prediction can make a much larger number of later observers fall in line, not because the later observers were fooled, but because copying visible predecessors is, individually, a perfectly rational strategy. Put these two results together and you get an accurate anatomy of how a single quotable claim, repeated by a first small ring of listeners, turns into mass behavior: not one giant leap of collective faith, but a threshold cascade running on ordinary, individually rational imitation.
The mechanism also runs inside one person
None of this requires a crowd. Albert Bandura's self-efficacy research, starting in 1977, is one of the most replicated findings in psychology: a person's belief in their own capability measurably changes how much effort they invest, how long they persist after setbacks, and — through that mediated path — their actual rate of success. The belief doesn't teleport the outcome into existence; it changes behavior, and the changed behavior changes the outcome, exactly like the bank run, just inside a single nervous system instead of across a crowd. Placebo research adds a physiological layer to the same point: expectancy-driven placebo analgesia is reversible by the opioid-blocker naloxone, which means the pain relief runs through the body's own real endogenous opioid system — a belief triggering a genuine biochemical cascade, not an illusion and not magic, just an ordinary neurochemical pathway that expectation happens to be able to switch on.
The honest caveat
The single most famous claimed case of expectancy effects, Robert Rosenthal and Lenore Jacobson's 1968 "Pygmalion in the classroom" study — teachers told at random which students were "about to bloom" saw those students' IQ scores rise — deserves the same scrutiny this post has applied to everything else. Later, larger reviews, notably Lee Jussim and Kent Harber's 2005 synthesis of decades of follow-up research, found that teacher-expectation effects on student performance are real but considerably smaller and more conditional than the original study's popularized version suggests, concentrated mainly among younger students and in the early days of a new teacher-student relationship. The honest state of the evidence is: expectancy effects on other people's performance exist, are measurable, and are much weaker than the folklore built on top of one 1968 study.
What ODTOE adds: naming the same mechanism formally
ODTOE's starting claim is that realized reality is a function of an observer applying an operator to a field of open configurations, R = Ô(Ψ); an individual observer's coherence toward a specific configuration C is written B(O, C). The corpus's own postulate for combining many individual observers (collective observation, postulate P5) formalizes exactly the asymmetry that makes bank runs and cascades work: collective observation probability is P_coll(E) = 1 − ∏ᵢ(1 − Bᵢᵏ), a disjunctive combination across observers rather than a conjunctive one. That single design choice is the whole story mathematically: a conjunctive requirement (every single person must believe, multiplied together) would make mass belief almost impossible to reach and fragile to any one holdout, exactly the opposite of what bank runs and information cascades actually look like. A disjunctive combination means the collective probability climbs toward certainty as more individual Bᵢ rise, even while plenty of individuals still have low B — precisely Granovetter's threshold cascade, restated as an inequality.
A later synthesis in the corpus goes one step further with what it explicitly labels a candidate lemma, not a settled theorem: a shared configuration acts as a collective attractor A once the group's pairwise coherence S(A) clears a threshold, and a system's trajectory reaches the fixed point associated with that configuration if the coherence gradient at the starting point is oriented toward A. The corpus's own worked examples for this exact pattern are not exotic: a passionary cluster of a shared cause, a scientific community converging on a shared paradigm, a creative partnership, a stable family. A following that forms around a quotable prediction is a fifth instance of the identical pattern — a group whose individual Bᵢ toward one specific future configuration has climbed, collectively, above the threshold that makes that configuration reachable rather than one of the many equally possible configurations sitting unrealized in the wider space Fix(Φ). Reaching that specific fixed point doesn't require the field of possibility to contain only one option; systems with real, mathematically legitimate alternate futures (Diamond and Dybvig's two banking equilibria are the cleanest instance) are exactly the systems where this kind of selection is visible at all.
Every era gets the oracle its coordination technology allows
None of the mechanisms above are historical curiosities. They run, unmodified, on any platform that lets a claim reach many people at once, which is why the internet has not produced a new kind of prophet — it has produced a cheaper distribution channel for the same five mechanisms above, and correspondingly more visible cases of prophecies fulfilling themselves in public, in real time.
The clearest modern instance is the GameStop short squeeze of January 2021. A Reddit community converged on a single, quotable claim — the stock was undervalued and heavily shorted, so coordinated buying could force short-sellers to cover at a loss — and then acted on it as a bloc. The claim did not describe an inevitability; it was one of several mathematically available price paths, exactly in the Diamond–Dybvig sense, and which path occurred was selected by how many people traded as though it were already true. Cryptocurrency markets supply an even more measurable case: a 2021 academic study by Ante found that individual tweets from Elon Musk produced statistically detectable, short-term price movements in Dogecoin and Bitcoin — a single, highly quotable, easily repeated statement (or emoji) from one account moving a market, the Schelling-focal-point mechanism operating at the speed of a retweet.
What the platform changes is not the mechanism but its cost. Duncan Watts and Peter Dodds's 2007 study of opinion formation, and Matthew Salganik, Dodds, and Watts's 2006 "Music Lab" experiment published in Science, both point at the same uncomfortable finding: cascades are driven less by a few uniquely gifted "influentials" and more by a critical mass of ordinary people who are easily influenced by what they see others doing. In the Music Lab experiment, identical songs became hits or flops in different simulated "worlds" depending only on which early downloads happened to be visible first — the same content, wildly different outcomes, decided by early social proof rather than by quality or by any special property of whoever posted first. That is the rigorous version of the point this whole post has been making: an "oracle" does not need clairvoyance or charisma beyond the ordinary; a platform that makes an initial signal visible to enough people, at the moment enough of them are primed to copy each other, is sufficient on its own.
This is also why the honest caveat from Rosenthal and Jacobson's classroom study generalizes cleanly to modern influencers: fame and follower count are not the same variable as causal power. A prediction from an account with ten million followers is not more likely to be true; it is more likely to reach the threshold density of initial believers that Granovetter's model and Bikhchandani, Hirshleifer, and Welch's cascade both require to tip. Robert Cialdini's social-proof principle names the same lever from the psychology side: people use the visible behavior of others as evidence about what they themselves should do, which is precisely the raw material a large following supplies in bulk. None of this requires the account behind the prediction to be unusually wise, only unusually reachable — the same P_coll(E) = 1 − ∏ᵢ(1 − Bᵢᵏ) climbs toward certainty whether the initial Bᵢ increases came from a 1907 bank teller's rumor or a 2021 verified account's post, because the formula never asked where the individual coherence came from, only how many observers it reached.
So: no magic, and no extraordinary ability
Put the pieces back together and the "gift" a successful prophet needs shrinks to something almost mundane: producing an original, memorable, easy-to-repeat frame — a Schelling focal point — vivid enough to become quotable and specific enough that early listeners can act on it identically. Everything after that first spark runs on mechanisms with decades of citations behind them and, in two cases, a Nobel medal: individual belief changing individual effort and physiology (Bandura, placebo neurobiology), a threshold cascade turning a small ring of early believers into a large one (Granovetter, Bikhchandani–Hirshleifer–Welch), and a system with more than one legitimate equilibrium letting shared expectation pick which future actually happens (Diamond–Dybvig). Through the ODTOE lens, that whole chain is one disjunctive collective-observation postulate and one candidate reachability condition, doing in formal notation exactly what Merton's bank run did in a single sentence in 1948: a belief, once it is shared by enough people to cross a coherence threshold, is not describing a future — it is one of the ordinary, well-documented ways a future actually gets selected out of the several that were genuinely available.