The Global Consciousness Project: Did the Numbers Know Before We Did?

Machines around the world were generating random numbers. Researchers wondered whether those numbers might change when millions of people shared a moment of shock, grief, or celebration.

Then came September 11, 2001.

Later examinations of the data raised a disturbing possibility: had something changed before the attacks began?

That question helped turn an unusual consciousness experiment into a story about machines that could see the future. The research supports a much more cautious account.

What Happened

The Global Consciousness Project began in 1998 under Roger Nelson. It developed from research associated with the Princeton Engineering Anomalies Research laboratory, commonly called PEAR.

Instead of one mysterious machine, the project used a worldwide network of physical random-number generators. Researchers collected their output and looked for statistical departures from expected randomness during major events.

The proposed explanation was that shared human attention or emotion might somehow influence physical systems. That was the hypothesis being tested, not an established property of consciousness.

The Princeton connection was genuine, but it needs qualification. The project’s own description identifies it as an independent collaboration and explicitly states that Princeton University did not sponsor it.

On September 11, thirty-seven devices were reporting data. The project’s published analysis described unusual statistical behavior surrounding the attacks. One measure appeared to change several hours before the first plane struck.

That apparent early change became especially intriguing. If the proposed effect came from people reacting to an event, why would it begin before they knew anything had happened?

But the project’s analysis also included a later caution: the September 11 results alone should not be treated as proof of global consciousness.

Did It Predict the Attacks?

The cited records do not establish an advance warning identifying the attacks, their location, or their timing.

A pattern discovered in recorded data after a disaster is different from a forecast issued before it.

The word “prediction” can also cause confusion here. In an experiment, researchers may predict how a statistical test will behave. That does not mean they predicted the historical event being studied.

The project’s own media page rejects claims that its data are predictive. It describes its main hypothesis as changes occurring during major events, while separately acknowledging research into possible anticipatory effects.

Why It Matters

Independent researchers Edwin May and James Spottiswoode examined the September 11 data. They acknowledged that the project’s formal test reached its stated significance threshold, but found that the result depended strongly on the selected time interval. A slightly shorter window, or one extended by thirty minutes, would not have reached that threshold.

Their analysis did not support the claimed response to the attacks. They also praised the project for making its data available for independent examination.

This is the central problem: random data contain fluctuations. To establish a warning system, researchers would need a rule fixed beforehand, then test it against future events while counting false alarms and missed events.

An unusual graph can suggest a question. Establishing that consciousness affected the machines—or that the machines detected the future—requires additional evidence.

The lasting mystery is how to distinguish a meaningful signal from a pattern that becomes compelling once we know what happened.

T Funk Take

The part that gets my attention is the claim that the numbers changed before anyone knew what was coming.

If that holds up, it raises a lot of questions. What were the machines responding to? People who already knew something? Something about the event itself? That would all be speculation.

First, I want to know how often the numbers did the same thing and nothing happened.

If these machines were running all day, every day, there must be plenty of strange-looking stretches to examine. How many did anybody notice before there was a disaster to connect them to?

I give the researchers credit for making the data available. Let other people check it. That is how you find out whether you have something.

But if someone says a machine predicted an event, I want to see what it told us beforehand. What should we expect? When? Where?

Finding a pattern afterward gives us something to investigate. An accurate warning beforehand would give us something very different.

Until the numbers can tell us something before the headlines do, the Global Consciousness Project stays open in the archive.