Precognition

1 · What it is

Knowing before

Precognition is the claim that information can come from the future: a dream that matches next week’s news, or a sense of dread before an accident.

Presentiment is the bodily version: a change in heart rate, sweat or brain activity a few seconds before something random happens, such as a startling picture appearing on a screen.

Because the target does not exist yet when the guess is made, nobody can leak it by accident, which makes precognition one of the cleaner things to test.

2 · How people practise it

Write it down first

Practice is mostly record-keeping. People keep dream journals and check them against later events, and keep prediction logs with every entry dated before the outcome is known.

The date is the point. A hit only counts if it was written down beforehand, in words specific enough to be wrong. Without a record, the hits are remembered and the misses fade.

3 · What the labs tested

Targets chosen after the guess

From the 1930s, card tests were run in which the target was picked at random only after the person had made their guess. Charles Honorton and Diane Ferrari pooled these forced-choice studies from 1935 to 1987 and reported a small effect above chance.

In 2011 the psychologist Daryl Bem published nine experiments with more than a thousand students. They took standard psychology tests and ran them backwards in time: for example, people seemed to remember words better if they practised them after the memory test. All but one came out above chance.

Presentiment studies measure the body instead. A 2012 meta-analysis reported that physiology shifted slightly before emotional pictures, compared with calm ones.

4 · Where it’s contested

A turning point for psychology

Bem’s paper used the standard methods of the day, and that was the problem it exposed. Eric-Jan Wagenmakers and colleagues argued in the same issue that those methods had made weak evidence look strong.

Stuart Ritchie, Richard Wiseman and Chris French repeated the memory experiment three times and found nothing. Jeff Galak and colleagues ran seven experiments with 3,289 people, and also found nothing. Bem and colleagues replied with a meta-analysis of 90 experiments that reported a small positive effect.

On presentiment, Sam Schwarzkopf argued that expectation and analysis choices could produce the pattern without anything coming from the future.

The cleanest test so far came in 2023. The Transparent Psi Project was designed by supporters and sceptics together, registered every detail in advance, and published its data as they came in. 2,115 people took part; the hit rate was 49.89% against 50% chance. It did not support precognition.

Supporters point to the meta-analyses; critics to the preregistered null. Both now agree on how the next test should be run.