PsychologicalDUAL-USE
Apophenia
What it is
The tendency to perceive meaningful patterns, connections, and intentions in random or unrelated data — faces in clouds, messages in noise, plots in coincidence.
How it works
Real-world examples
- •The “face on Mars” in a 1976 Viking orbiter photograph became evidence of an ancient civilization for a generation of enthusiasts; higher-resolution images from 1998 and 2001 showed an ordinary eroded mesa.
- •Michael Drosnin's 1997 book The Bible Code found assassinations and disasters encoded in the Hebrew text; McKay, Bar-Natan, Bar-Hillel and Kalai (1999) found comparable “predictions” in Moby-Dick and showed that the method finds messages in any long text.
- •Connect-the-dots boards on both sides — the QAnon “decodes” from 2017 onward and the elaborate Trump-Russia link charts circulated in 2017-2019 — assembled real names, dates, and photographs into designs the evidence did not support; the special counsel substantiated Russian interference and numerous contacts but not the coordinated conspiracy the charts depicted, and no QAnon prediction has come true.
- •Technical-analysis chart patterns — head and shoulders, cup and handle — appear just as readily in randomly generated price series, as Malkiel demonstrated with coin-flip charts in 1973; the sequences that “worked” are the ones that get shared.
- •Gambling systems from roulette “due” numbers to slot-machine timing rest on perceived streaks in independent events; casinos display the recent history at the roulette table precisely because players see patterns in it.
- •In a jealous or anxious relationship, a partner's ordinary lateness, a new password, and a laugh at a text message are assembled into a story; the story then filters what is noticed next.
Ethical guidelines
Where the line is
Pointing out a real pattern with its data and its denominator is analysis; the line is crossed when a persuader assembles selected dots, supplies the frame, and withholds the misses and the base rate, so that the audience experiences a pattern the persuader knows the evidence does not contain.
- ●Before presenting a pattern as a finding, state the prediction in advance and test it forward; a pattern found after the fact in searched data is a hypothesis, not a result.
- ●Do not present coincidence chains as evidence without the denominator — how many connections could have been drawn, and how many were.
- ●Teaching people to notice pareidolia and coincidence is legitimate; exploiting their pattern-hunger by feeding them dots and a frame is not.
- ●“There are no coincidences” is a claim about the world that is false; do not use it to make your audience stop asking for evidence.
How to defend against it
- ►Ask what you would expect to see if the data were random, and check whether that is what you are seeing; most coincidence chains are exactly what randomness produces with enough dots.
- ►Apply the law of truly large numbers (Diaconis and Mosteller): with millions of people and events, one-in-a-million coincidences happen daily. Ask how many chances there were for this one.
- ►Pre-register your pattern: write the prediction down and see whether it holds in new data. Patterns that exist only in hindsight are apophenia by definition.
- ►Count the misses. For any system, decoder, or prophet, ask to see the predictions that did not come true, and weight accordingly.
- ►When a narrative explains everything and is confirmed by everything, including the absence of evidence, treat that as a property of the narrative rather than of the world.
References
- Conrad, K. (1958). Die beginnende Schizophrenie: Versuch einer Gestaltanalyse des Wahns. ThiemeOrigin of the term Apophänie in the description of early psychosis.
- Foster, K. R., & Kokko, H. (2009). The evolution of superstitious and superstition-like behaviour. Proceedings of the Royal Society B, 276(1654), 31-37The evolutionary model showing when over-detection of patterns is adaptive.
- McKay, B., Bar-Natan, D., Bar-Hillel, M., & Kalai, G. (1999). Solving the Bible Code puzzle. Statistical Science, 14(2), 150-173The refutation of the Bible Code by finding equivalent codes in other texts.
- Diaconis, P., & Mosteller, F. (1989). Methods for studying coincidences. Journal of the American Statistical Association, 84(408), 853-861The law of truly large numbers recommended in the defenses.
- Miller, J. B., & Sanjurjo, A. (2018). Surprised by the hot hand fallacy? A truth in the law of small numbers. Econometrica, 86(6), 2019-2047The statistical correction suggesting the hot hand is not purely illusory.
Last reviewed
Suggest a correction