Edwin C. May - Star Gate research director and nuclear physicist

Type: named-figure source-of-record
Subject: Edwin Charles May, nuclear physicist and parapsychology researcher
Credibility: ~51 composite, with conventional scientific background and program role ~90, archive and publication contribution ~75, experimental psi evidence ~35, operational remote-viewing claims ~25, and claim handling ~55
Principal roles: SRI remote-viewing researcher from 1975, project director from 1985, and director of the later SAIC research program through 1995

Short assessment

Edwin C. May was a conventionally trained experimental nuclear physicist who became the longest-serving researcher and later director of the government-funded remote-viewing program usually grouped under Star Gate. His education, physics publications, security-cleared SRI work, program leadership, and later archival contribution are well documented. He is a strong historical source for how the research program operated and what its advocates believed they had found.

The evidence for May's central scientific and operational claims is much weaker. The 1995 government-commissioned review accepted that the selected laboratory experiments produced statistically significant results, but it did not establish a paranormal cause and found no actionable intelligence value in the operational sample. Jessica Utts reached a more favorable scientific conclusion, while Ray Hyman and the final evaluators emphasized methodological alternatives and the absence of useful operations. Later reconstruction of a prominent SAIC experiment found inconsistent accounts of who handled and judged the records and could not rule out information leakage.

May's later theories are serious attempts to make testable distinctions, but they remain speculative. Decision Augmentation Theory recast apparent mind-over-matter results as precognitive selection rather than a physical force, yet researchers using a larger Princeton database disputed its headline statistical result. His operational success stories are specific and sometimes traceable to program records, but the public versions are selected examples rather than complete case series with independent scoring.

May's claim handling is mixed but includes an unusually clear public correction. In 2024 he said that "everything I said" in his 1996 criticism of the Star Gate closure explanation was wrong after later CIA releases changed his view. He remains much more confident about target entropy and operational utility than the public evidence supports, but openly withdrawing a long-standing, reputation-protecting argument materially improves the assessment.

May is also a strong new candidate for Jacques Vallée's pseudonymous "Gilbert". A public identity hint can be read directly as EDWIN MAY, and Vallée's journals establish a long personal and professional relationship with May. His documented role was researcher, director, and judge rather than elite operational viewer, and no public record connects him to Gilbert's Central America travel or alleged physical encounter. The hint raises May to the leading identification hypothesis without verifying the encounter story.

Nuclear-physics training and Star Gate leadership are documented

An official 1983 SRI biographical sketch records May's 1962 physics degree from the University of Rochester, 1968 doctorate in nuclear physics from the University of Pittsburgh, postdoctoral work at the University of California, Davis, and conventional publications in nuclear-reaction research. The Rice University finding aid says he joined SRI as a consultant in 1975, became a full-time senior research physicist with a Secret clearance in 1976, and succeeded Hal Puthoff as project director in 1985.

That record supports May's expertise in experimental physics and his access to the classified research program. It does not make parapsychology an extension of his established nuclear-physics work. His conventional publications concerned particle and nuclear reactions, while the remote-viewing program required separate experimental and statistical support.

May's later organizational biography says he controlled about 70% of the program's roughly $20 million in funding and 85% of its data. Those figures are self-reported and difficult to audit as exact percentages, but the broader claim that he directed most of the later research is consistent with the declassified record. When the SRI effort ended, the program moved to Science Applications International Corporation. Rice's historical note incorrectly says May founded SAIC; the company already existed, and May established his research effort within it.

One bibliographic caution is worth preserving. May's later biography highlights Physical Review Letters volume 18, page 962 as his report of the first measurement of a singlet deuteron state. The American Physical Society's metadata for that article lists B. L. Cohen, E. C. Cohen, and T. M. O'Keefe, not May. Other conventional publications do list him, including a Physical Review paper on deuteron disintegration. The mismatch is a CV or metadata problem around one highlighted citation, not evidence that his physics career was invented.

The 1995 review found a laboratory anomaly but no operational utility

The American Institutes for Research evaluation was commissioned as the government decided whether to continue Star Gate. It reviewed ten SAIC laboratory experiments and a sample of later operational work. The evaluators found statistically significant laboratory results, but said the cause was unresolved and that methodological alternatives had not been eliminated. In the operational material, they found occasional correct general descriptions mixed with wrong and inconsistent details, no case in which remote viewing guided an operation, and no actionable intelligence value. They recommended ending operational use.

Statistician Jessica Utts's separate assessment for the review argued that anomalous cognition had been established and that the SRI and SAIC results formed a consistent pattern with conceptual replications elsewhere. Her analysis matters because it prevents the official review from being summarized as a simple null result. Utts had also collaborated and published with May, including on Decision Augmentation Theory. AIR deliberately selected her as the favorable expert and Hyman as the skeptical expert, so her report is expert support for the program rather than an independent replication of it.

May's 1996 response raises concrete scope objections. He says AIR examined only the program's weak final operational period, excluded earlier positive cases, did not interview important users and participants, and gave too much weight to an older National Research Council review. These are legitimate reasons not to treat the reviewed operational sample as a complete audit of the program's history. May goes further and argues that the outcome had effectively been set in advance. He acknowledges that he cannot prove that conclusion, and the procedural complaints alone do not establish deliberate suppression.

In a 2024 Shawn Ryan interview, May withdrew that explanation directly. At 3:12:27 he said, "everything I said in that article was wrong." He preserved the factual complaint that AIR reviewed one year rather than the program's full history, but said later CIA releases convinced him that post-Cold-War budget cuts, rather than a predetermined review, put Star Gate on the chopping block. This is a substantial self-correction because it abandons an argument that protected the program and his own leadership from an unfavorable evaluation.

A prominent SAIC experiment could not be reconstructed securely

Richard Wiseman and Julie Milton's critical reconstruction of SAIC Experiment One examined a study that Utts treated as particularly successful. The published report did not specify enough procedural detail to determine who handled the viewers' original response sheets, whether May judged originals or pre-feedback faxes, how much contact he had with non-blind experimenters, or whether the trials were judged in an order that could reveal repeated target cues.

Their follow-up exchange records several inconsistent recollections from May and other team members. May initially described judging original responses, later said he used faxes, and the team's final reconstruction did not have surviving records sufficient to settle the sequence. Wiseman and Milton accepted that one proposed contact path was probably unimportant, but found that other possible leakage paths remained. This does not explain every SRI or SAIC result, but it materially weakens the evidentiary value of an experiment presented as a tightly controlled centerpiece.

May's role as the study's judge also concentrated too much authority in the program director. A blinded judge can be methodologically legitimate, but the experiment needed a contemporaneous procedural record and clean separation from people who knew target information. The later inability to establish those conditions is the important failure.

Decision Augmentation Theory made a testable claim that did not hold up cleanly

The Decision Augmentation Theory papers by May, Utts, S. James P. Spottiswoode, and Christine L. James proposed that apparent psychokinesis in random-number-generator studies could instead arise when participants unconsciously selected favorable moments through precognition. The model was phenomenological: it attempted to distinguish patterns expected from selection from patterns expected if consciousness physically pushed a random system. It did not identify a physical mechanism, and the authors explicitly said that a quantum explanation would be premature.

This was a useful scientific move because it generated a statistical comparison rather than treating every anomaly as evidence of a force. May later described the result in a 1996 interview as an incontrovertible rejection of force-like effects in the main random-generator database and said most researchers had accepted it. York Dobyns and Roger Nelson subsequently reported that the 8.6-sigma result depended on an erroneous analysis and that the full Princeton Engineering Anomalies Research database rejected Decision Augmentation Theory by more than five standard deviations. Their objection came from within the same pro-psi research community. It does not decide whether the underlying anomalies were real, but it shows that May's strong claim of a settled interpretation did not survive technical dispute.

May's later physical theories follow the same pattern. He and collaborators explored whether target entropy correlated with remote-viewing performance, whether performance varied with local sidereal time, and whether information might travel backward in time before being processed by the brain. Reported effects were marginal, declined within studies, or shifted in later data, and no carrier or stable neural mechanism was identified. The work is hypothesis generation, not established physics.

In 2024 May nevertheless said there was "no doubt" about the entropy correlation, that nine studies agreed, and that the relationship was well established. He was more restrained about his proposed information carrier and brain mechanism, calling the model testable, saying it might be wrong, and acknowledging that he had no idea what the carrier or a psychic equivalent of the retina might be. The contrast captures his public style: appropriately tentative about mechanisms, but too certain about mixed experimental findings.

Operational success stories lack a complete public denominator

In a 2012 first-person history, May describes remote-viewing contributions to the search for kidnapped General James Dozier, the recovery of a politically connected woman's missing daughter, a 12-for-12 test against the proposed MX missile basing system, and a directed-energy series he says achieved 80% accuracy and 69% reliability. These are concrete claims with names, dates, and proposed mechanisms for use. They are more useful than vague assertions that the program worked.

The article presents selected successes without the complete task set, fixed scoring rules, contemporary customer records, or unsuccessful comparison cases needed to measure operational value. The missing-daughter story depends on May's later account of what the FBI did with three viewers' combined material. The Dozier account says a drawing later matched a park and cathedral in Padua, but it does not show that the information led investigators to him. The MX and directed-energy numbers depend on program-defined scoring and are not reconciled with AIR's finding that the operational record did not guide decisions.

May is therefore a valuable witness to what the program attempted and what its leadership counted as success. His retrospective selection and interpretation cannot substitute for a complete independent operational audit.

In the 2024 interview, May gave two broader customer-use figures: he said the CIA returned with 41 missions and that 17 of 19 end users requested more work. Repeat requests show institutional interest, but the figures come from May, do not measure whether the information changed an operation, and are not reconciled with AIR's evaluation of actual utility.

Archive and publication work make the program more auditable

The Edwin C. May papers at Rice University comprise 54 boxes and about 35 linear feet of reports, correspondence, experimental records, and administrative files from 1974 through 1995. May also helped publish the multivolume Star Gate Archives after declassification. These are substantial contributions because they expose details that can support, weaken, or correct the program's later stories.

The archive does not resolve every claim. Important procedural records and complete operational scoring remain missing, and the finding aid contains at least one historical error. Still, making a large research record available distinguishes May from claimants whose work remains inaccessible or exists only in interviews.

The public Gilbert hint points to May but the biography remains incomplete

In September 2026, Australian researcher Rhys Dalton-Morgan said publicly that he knew Gilbert's identity and that he had answered it cryptically in the post. The end of "engaged," start of "with," and end of "an" spell ED-WI-N, immediately followed by the word "may." That is a more direct name extraction than the initials of "lab in a mountain," which spell LIAM and match Bill Ray's documented pseudonym. Dalton-Morgan did not confirm either reading or explain how he learned the identity, so the extraction identifies the likely intended answer rather than its factual basis.

May fits several of Vallée's clues unusually well. Vallée's published journals document visits to May's home, dinners and private discussions, May's invitation for Vallée to return as an SRI consultant, and repeated classified project meetings after Vallée's compartmented clearance was restored. May had official access, an old friendship with Vallée, a highly technical style, and control over experimental records.

The strongest conflict is role. Public records consistently identify May as a researcher, director, task designer, and judge, not as an exceptionally capable operational remote viewer. In the 2024 interview, he even described himself during a personal anecdote as "not psychic at all." No public record found places May and Vallée together in Central America, says May was taken into a laboratory inside a mountain, or attributes a physical entity encounter to him. His extensive publications also contain no clear version of Gilbert's promised personal account. Those gaps keep the identification below even confidence despite the apparent name clue.

Incentives and public posture

May spent much of his later career defending the scientific and operational value of the program he directed. He has led research organizations, sought grants, published books and papers, and appeared in media about Star Gate. That creates professional and reputational incentives to emphasize positive results. It does not by itself invalidate his work, and his archival releases expose his claims to more scrutiny than withholding the record would.

His public posture mixes methodological restraint, overstatement, and correction. He rejects macroscopic psychokinesis, criticizes loose quantum explanations, and describes a failed nine-month search for extraordinary abilities in India. He has characterized disputed statistical conclusions as incontrovertible and presents mixed entropy findings as settled. His explicit 2024 withdrawal of his earlier closure explanation shows that he can revise a central public claim when later records change his view.

Credibility assessment

What raises May's credibility

His physics education, conventional publications, security clearance, SRI employment, and later program leadership are supported by institutional and declassified records.
He generated explicit protocols, quantitative models, and testable distinctions rather than relying only on anecdotes.
AIR and Utts both recognized statistically significant laboratory results, even though they disagreed about what the results established.
His archive and publication work made a large part of the program available for independent historical and methodological review.
He rejects some popular paranormal explanations, including macroscopic psychokinesis and premature quantum claims.
He publicly withdrew his 1996 explanation of Star Gate's closure after later CIA releases contradicted it.

What lowers May's credibility

The government review found no actionable operational value, while May's contrary examples remain selected cases without a complete public denominator.
A central SAIC experiment could not be reconstructed consistently enough to rule out information leakage, and May gave conflicting procedural recollections.
Decision Augmentation Theory's headline statistical result was directly challenged using a larger database after May described the conclusion as settled.
His later entropy, sidereal-time, retrocausal, and neural-processing ideas remain unreplicated or speculative rather than established physics.
He still presents the target-entropy relationship and operational repeat-customer figures more confidently than the available independent record supports.

Net assessment

May rates ~51 overall. His conventional scientific background and documented program role rate ~90. His archive and publication contribution rates ~75. Experimental evidence for psi rates ~35 because statistically unusual results exist, but their cause and robustness remain disputed and at least one important experiment had unresolved control problems. Operational claims rate ~25 because selected success stories do not overcome the official finding of no actionable intelligence value. Claim handling rates ~55 because he is technically specific, admits uncertainty about proposed mechanisms, and made a substantial public correction, while continuing to describe contested experimental conclusions as settled.

Position relative to other UAP and remote-viewing figures

May rates near Gary O. Langford and Keith Harary overall. He has a stronger documented leadership and publication record, but greater responsibility for disputed program claims. He rates above Hal Puthoff in this framework because his later claims are narrower, his theoretical papers make clearer testable distinctions, and his archival record is unusually substantial. He rates below Robert Jahn because May's public case remains more tightly tied to a program he directed and defended, and its claimed operational benefits did not survive independent review.

Role-category placement

May belongs among credentialed program insiders whose institutional role is firmly established but whose claims about anomalous performance and practical value remain disputed.

Followup items

Obtain May's published response to Wiseman and Milton's Experiment One critique and compare each procedural assertion with the surviving experiment records.
Retrieve the full Dobyns and Nelson paper and independently reproduce its Decision Augmentation Theory comparison against the earlier May analysis.
Audit the MX missile, Project Rose, Dozier, and missing-daughter cases from contemporaneous tasking, scoring, and customer records rather than retrospective summaries.
Find the source records and denominator behind May's 41 CIA missions and 17-of-19 repeat-customer claims.
Resolve the authorship mismatch for the Physical Review Letters citation highlighted in May's later biography.
Ask Dalton-Morgan whether the EDWIN MAY reading is his intended identity clue and what evidence beyond the wordplay supports it.
Search Vallée and May's correspondence, travel records, and unpublished archive folders for Central America, a mountain laboratory, an entity encounter, and the protocol tables Vallée says he reviewed.