Project Stargate: the government's psychic spies

There was no single night in this file. There were many afternoons, and most of them looked the same.

Scene · reconstructed from the paperworkPicture it. A room, a desk, a blank sheet of paper. Somewhere else, a target has been chosen, and the person at the desk has not been told what it is. That is the point of the exercise. According to the records the CIA has released[cia-stargate], US Army and intelligence units spent years asking people to describe distant targets by mind alone, and to write and sketch what came to them. The file doesn't say what any one of them felt in that chair. It says that the chair was paid for.

The practice had a name: remote viewing. The claim behind it is simple to state and very hard to test: that a person can describe a place, a building or an object they cannot see, using nothing but attention. From the 1970s until 1995, according to the reading room[cia-stargate], the United States government ran programmes to study and use it. The names changed over the years. The one the public remembers is Stargate.

I want to be careful about what kind of file this is. It is not a story of a strange light or a lost object. It is a story of a bureaucracy taking an extraordinary claim seriously enough to give it offices, staff and memos. The CIA's own collection[cia-stargate] holds thousands of pages of it. That is what makes it unusual: most cases in this archive are short of paper. This one has more than any reader can finish.

One of the men who sat in those rooms has since spoken publicly. Paul H. Smith, a former Army remote viewer[smith-stargate], describes his time in the programme in a long interview. It is a participant's account, not an evaluation, and this file treats it that way. But it reminds us that these sessions were done by real people, chosen by the government, on government time.

So the scene, kept inside the papers, is this:

• soldiers and civilians, chosen and paid (the record[cia-stargate]);
• targets they were not told (the record[cia-stargate]);
• descriptions and sketches, written up and filed (the record[cia-stargate]);
• a programme that ran, under one name or another, for about two decades (the record[air-1995]).

What I keep coming back to is the ordinariness of it. No ceremony, no glowing screen. A pencil and a sealed target. If remote viewing was ever going to show itself, it was going to show itself in a filing cabinet. And the filing cabinet is now open.

The paper trail is the strongest part of this file, so I will walk it slowly.

The collection. The CIA's FOIA Electronic Reading Room holds a Stargate collection[cia-stargate]: declassified documents on Stargate and the remote-viewing projects that came before it. The release and accession dates for the collection are still on our list to verify, but the documents themselves are public and readable.

Grill Flame, 1979. One of the predecessor names was Grill Flame. In the collection there is an Army message dated 12 April 1979[cia-stargate] about changes to the Grill Flame programme. It mentions a review committee of senior scientists. It reads like any other administrative message about any other project, which is exactly what makes it strange. The page shown in this file is page 1 of that message.

Star Gate, the briefing. Later the effort took the name by which it is remembered. The cover of this file is the title slide of a Defense Intelligence Agency briefing[cia-stargate] headed Project Star Gate. It was marked SECRET NOFORN, meaning not for release to foreign nationals. According to the reading-room record, it was released in 2000.

The evaluation, 1995. Then the document that closes the file. On 29 September 1995 the American Institutes for Research delivered An Evaluation of Remote Viewing: Research and Applications[air-1995], commissioned by the CIA. Its authors were Michael D. Mumford, Andrew M. Rose and David A. Goslin. It looked at both the laboratory research and the operational use.

Two reviewers. The evaluation package included outside reviews. One came from the statistician Jessica Utts of UC Irvine[utts-air-review], whose assessment argued that the statistical effect in the laboratory work was established. The other came from Ray Hyman of the University of Oregon, who disagreed[hyman-1995]. Two qualified readers, the same data, opposite conclusions about the laboratory side.

What the trail adds up to is this: the programme existed, and its budget lines and its end[air-1995] are documented. The Finding in this file is confirmed, and the grade is documented, because the government's own papers say so.

What the trail does not contain is equally important. It does not contain any repeatable effect accepted by the wider scientific community. The dossier's analysis[air-1995] states that plainly, and I will not soften it.

There is one more thing the paper shows, quietly. Programmes like this do not run on one memo. They need renewals, reviews and sign-offs, year after year. The reading room[cia-stargate] is full of them. Someone kept saying yes. In 1995, someone finally asked whether the yes had ever paid off.

The turn has a date: 29 September 1995.

That is the date on the AIR evaluation[air-1995]. The CIA had commissioned an independent look at two decades of work, and the answer came back in two parts.

On the laboratory side, the reviewers split. Jessica Utts[utts-air-review] argued that the statistics showed an effect. Ray Hyman, the other reviewer[hyman-1995], did not accept that conclusion. I find that split honest and worth keeping in view: this was not a case of everyone laughing a programme out of the room.

On the operational side, the evaluation was blunt. According to the report's conclusion[air-1995], remote viewing had not provided intelligence that was useful in operations. That is the sentence that mattered to the people paying for it. An intelligence tool is judged on whether analysts can act on what it produces, and the evaluation found they could not.

Then the programme was ended[air-1995]. The timeline[air-1995] records the evaluation and the closure together. As far as the government was concerned, the file was closed, and this archive keeps its status as closed too.

Here is the fair other side, in the file's own terms. Former participants, among them Paul H. Smith[smith-stargate], argue that some sessions produced striking hits, and that the 1995 review looked at too little of the data to judge the whole programme fairly. That is a participant's view. It deserves a hearing, because these were not people off the street; the government chose and paid them (the record[cia-stargate]). It does not, on its own, overturn an evaluation.

And here is the mainstream reading, stated just as plainly. What the documents show is that the United States ran remote-viewing programmes for about two decades (the record[cia-stargate]). What they do not show is that remote viewing works. No repeatable effect has been accepted by scientists (the analysis[air-1995]).

So where exactly does the story bend? Not at the moment someone proved remote viewing false. That never happened, and the file does not claim it. The bend is quieter. It is the moment the question changed from is something there? to has it ever helped? The first question can be argued about forever, as Utts[utts-air-review] and Hyman[hyman-1995] showed. The second had an answer, and the answer was no.

That leaves a question the evaluation was not asked to settle, and it may be the most interesting one in the file.

The programme is closed. The file is not quite finished.

The question this archive keeps open is a human one: why did it last so long? If the results were as weak as the 1995 evaluation[air-1995] found, why did officials, year after year, approve another review, another briefing, another room with a blank sheet of paper? The reading room[cia-stargate] shows that they did. It does not, as far as I have read, explain why.

Here is what the documents leave unsettled:

• The laboratory dispute. Utts[utts-air-review] and Hyman[hyman-1995] read the same experiments and disagreed about whether there was a statistical effect. The evaluation did not end that argument; it set it aside in favour of the operational question.
• The participants' case. Former viewers such as Paul H. Smith[smith-stargate] say the review saw too little data. The file records that claim as the other side, not as a finding.
• The full scope of the collection. The release dates of the Stargate collection[cia-stargate] are still to be verified, and thousands of pages remain for anyone willing to read them.
• The people. The file's own notes list reviewers and researchers whose official pages are still to be checked. I won't add names here that the dossier has not verified.

And here is what is not open. The programme existed, ran for about two decades, and is documented in released government files (the record[cia-stargate]). The AIR evaluation[air-1995] found no operational value, and the programme closed. Nothing in this file shows that remote viewing works, and the file's Finding does not pretend otherwise.

I keep coming back to the dry Grill Flame message of April 1979[cia-stargate], with its committee of senior scientists. Somebody wanted the claim reviewed properly, even then. It took sixteen more years for the review that ended it. Perhaps it was fear that a rival would get there first. Perhaps it was a handful of sessions that felt too good to ignore. The file does not say, and I won't pretend it does.

The documents, the evaluation and the timeline are below. The files tell us what was done. They do not tell us everything about why.