Epstein Data Wiki

Charles L. Harper Jr.

Properties2
tagsperson, epstein-science-philanthropy, john-templeton-foundation, fqxi, cryptography-in-nature, research-meeting-proposal
aliasesHarper, Charles L., Charles Harper, Chuck Harper, Charles L. Harper, Jr.

60 min read

How he appears in these documentsAs a correspondent of Jeffrey Epstein in his own voice, under the sender name “Charles L. Harper Jr.” and the signature “Charles Harper” or “Chuck Harper”, and as the author of two documents that carry no e-mail headers at all: a proposal on his own printed letterhead and a paper on grantmaking method12
WhenOverwhelmingly 2010. Taking for each document the earliest full calendar date printed anywhere in it, 148 of the 203 fall in 2010 and 31 in 2009; the remaining 24 are accounted for in the note3
How he was introducedThrough Barnaby Marsh, over three weeks in May 2009. Epstein opened on 5 May with “charles harper-eastern europe ideas?”, asked again on 28 May with the two words “charles harper?”, and Marsh answered “He left the foundation… still, a very good guy for you to talk to…”45
His position at the moment of the introductionAlready departed. Every document read for this page that shows him acting is later than that message, and Epstein himself described him eight months afterwards as “former head of templeton”56
What he proposedA “very-high-dynamism two-day research meeting” on the theme “Cryptography in nature”, at a budget summary total of “$118,000 + Meeting T&L & special events”, in an eighteen-sheet proposal dated 4 November 2010178
A refusal he reportedThat Max Tegmark and colleagues “offered support but did not think that FQX should do it”, and that “the real reason was prudence to avoid FQX, already quite ‘controversial’ via Templeton support to become even more so via Epstein-related sponsorship of prizes”9
Not to be confused withCaleb Harper, a different man of almost identical corpus size: 192 documents against this page’s 198-document name union, and zero documents in common10
Documents read at the image for this pageThirty sheets across twenty-two documents, at 400 dpi and two at 600, each named in its footnote

Charles L. Harper Jr. is a science-philanthropy executive who corresponded directly with Jeffrey Epstein, chiefly through 2010, and who wrote and priced a research project for him. He had been Executive Director and Senior Vice-President of the John Templeton Foundation, and the release holds a thirty-five-sheet paper of his over that title; but the correspondence with Epstein begins after he left the Foundation and not before, and no document read for this page shows him acting for the Foundation in any of it.

The sequence matters enough to state at the top, because the obvious reading of the file is the wrong one. Over three weeks in May 2009 Epstein asked Barnaby Marsh, then an officer of the Foundation, about Harper twice; on 29 May Marsh answered “He left the foundation… still, a very good guy for you to talk to…”. Epstein asked for a contact, used it eight months later, and what follows in the release is a year of direct correspondence with a man who, on both of the documents that speak to it, had already gone.

Two men named Harper, and why the surname alone is useless here

This release holds a second, unrelated Harper of almost exactly the same size. Caleb Harper, of the MIT Media Lab, appears in 192 documents. The population used for this page, the union of the phrases "Charles L. Harper" (140 documents), "Charles Harper" (100) and "Chuck Harper" (47), comes to 198 documents, and no document is in both sets.310 The two records do not even overlap in time: this page’s traffic is 2010, and Caleb Harper’s is later.

A search on the bare surname is worse than useless. "Harper" alone returns 834 documents, of which 444 belong to neither man.3 The four families that account for most of those 444, found by reading the out-of-union returns rather than by guessing at them, are a Florida state legislator named in a newspaper clipping, two unrelated Harpers on Palm Beach County property rolls, the publisher HarperCollins, and the artist Charley Harper.11 Where a document below says only “Harper”, this page says how the identification was made.

How he enters the record: May 2009

The earliest document in this page’s population is dated Tuesday 5 May 2009, and it is Epstein asking Barnaby Marsh, then an officer of the John Templeton Foundation, about Harper for the first time. The message has an empty subject field, and its whole body is one line:4

“charles harper-eastern europe ideas?”

Marsh answered at 20:05:22 the same day: “He is travelling a lot- I’m trying to find time to discuss with him…more soon…”.4 Two days later he reported that he had “finally heard from Chuck Harper, who is visiting anthropological sites in South Africa at the moment and has been a little hard to reach”, a message already carried on Barnaby Marsh’s page and not restated here.

Three weeks after that, Epstein asked again. On Thursday 28 May 2009 at 5:57 PM he sent Marsh a second message, again with an empty subject field, whose whole body was two words:5

“charles harper?”

Marsh answered at 8:48 the next morning:

“He left the foundation… still, a very good guy for you to talk to…”

Epstein replied that Friday afternoon at 4:32:39 PM:

“i’d love to,, he has great ideas and experieince.. can i have a contact.. or feel free to give him mine.. — will he be replaced or will you move up , (my hope)”

Four things on that sheet are worth separating. The subject line of Epstein’s original message is a printed field with nothing in it, and “charles harper?” sits in the body; the reply’s subject is “Re:”, which is what an empty subject produces.5 The doubled comma in “i’d love to,,” and the misspelling “experieince” are on the image and are not repaired here; the other doublings on that line are doubled full stops. Marsh’s e-mail address is barred in the addressee field and again in the attribution line. And Marsh’s answer is in the past tense, left, which is the only statement read for this page, from anyone, about Harper’s position at the moment Epstein was put in touch with him.

Eight months later Epstein described him the same way, unprompted, to a third party whose name the release withholds. On 28 January 2010 he wrote:6

“novak, and krause coming for weekend, spent day with charles harper former head of templeton.”

Those two documents are the whole of the departure evidence, and they agree. No document read for this page shows Harper acting for the John Templeton Foundation at any date after May 2009, and this page does not describe him as doing so. The thirty-five-sheet grantmaking paper that carries his Foundation title is undated on its face and is discussed below only for what it says about him.2

January 2010: the contact is used

Seven and a half months after Marsh’s answer, Epstein used it. On Saturday 16 January 2010 at 6:40 PM he wrote:12

“chuck,, got your address from Barnaby,, may i call you at your convenience?”

Harper answered the same evening, “Sure. I welcome your call. I am in a theatre just now (in intermission)”, and offered times. The extraction of that sheet renders the name in Epstein’s message as “Bamaby”; the image prints Barnaby, and the difference is the commonest optical fault in this release.123

The next day Epstein telephoned and invited him to Palm Beach. Harper wrote to Epstein, copying one person whose name is barred, under the subject “Possibly visiting Palm Beach next weekend at the invitation of Mr. Epstein”, and addressed the body to that barred person:13

”( Thank you Jeff for your call today, and for your kind invitation to join Martin Nowak and his team coming down to visit you in Palm Beach next weekend! )”

He added that he knew Martin Nowak well and would be in Cambridge that week, and he pasted in two JetBlue fares he had looked up himself on Expedia. Epstein’s message of that afternoon, quoted beneath his own and running onto the second sheet, is three words: “please coordinate a trip”. The same day Epstein wrote to Nowak directly: “i spoke to charles harper„ he would like to come with you to palm beach”.14

Nowak then postponed, and on 18 January at 6:46 PM Epstein wrote to Harper: “charles„ martin delayed his trip„ you would still be welcome to come and spend a day. will organize everything for you , we have lots to discuss or if you prefer, wait until martin can come.” Harper replied at 9:31 PM with three dated travel options and the line “Looking forward to meeting you and to fun discussions.” Epstein chose the first, and named the person who was to make it happen. On two copies of his reply that name is barred in the middle of the sentence, and the extraction of one of them closes the gap and prints the sentence as though it were continuous:1516

“option 1 sounds great// ▮ please organize”

Two of the other copies print the same sentence whole, and the word the bar covers is a forename:17

“option 1 sounds great// sarah please organize”

Both are in this page’s own population and both were reached by the same search as the barred pair. Each prints the name three times: in Epstein’s instruction; in Harper’s quoted message, where the barred copies read “Please l=t me and ▮know which might be” and these read “Please let me and Sarah know which might be”; and in the salutation “Dear Sarah,”.17 The producing party barred the name on two copies of one message and printed it on two others, and one of those two is the same document whose first sheet is cited above for the travel arrangements. That is a redaction inconsistency on the producer’s side, verified at the image on both sides of it and not a fault in the text layer; the release names her and this page reports the forename it prints. No document read for this page gives more of her name, and none is supplied here.

Eleven documents in the release carry that sentence, four extracted with the digit and seven with a capital I in its place, and all eleven were read for this page: nine close the gap in the extraction and two print the name in both layers.18

Harper wrote to her that night. His arrangements message of the small hours of 19 January opens “Dear Sarah,”, with its sole To: field barred and its Cc: naming Epstein. He picked the later Friday flight, JetBlue 427 out of Boston at 8:20 pm on 22 January, asked for “Any reasonable overnight lodgings” near the airport, proposed a 7:20 pm return to Philadelphia on the Saturday at a fare he had priced himself, and asked her plainly: “Would you be so kind please to purchase these flights for me as e-ticketings and send me the confirmation details electronically.”19

Two things follow, and they are separate. The first is who paid. Later that morning she put the question to Epstein in one line, on the sheet that also prints her name:2017

“Do I purchase his tickets from Couq or personal?”

Epstein’s reply is “arrive fri night,, put in hotel”. He did not answer the question. No document read for this page records which account the tickets were bought from, or that they were bought at all, and this page does not assert that the Couq Foundation paid for anything. What the exchange establishes is that buying a visiting academic’s air tickets out of Couq was a live option in the mind of the person doing Epstein’s arrangements, and that she had to ask which it was to be.

The second is the clock, which matters because it is used again below. Of the copies read, one prints the “option 1” message’s own sending time as “Tuesday, January 19, 2010 3:06 AM” in a header field with no zone; another quotes the same message as “On Jan 18, 2010, at 10:06 PM”. The difference is exactly five hours, which in January is Eastern Standard Time against UTC. Her own question runs the same way: its header field reads 3:17:12 PM and the copy that quotes it reads 10:17 AM.151617

The visit happened. On 24 January Epstein wrote “Thank you for yesterday. there turns out to be a black physicist society, that was an eye opener, einstein letter to besso is not clear…” and on 25 January Harper replied under the subject “Super to meet you!”:21

“Super to meet you! Excellent day. Thanks very much for inviting me. We seem to have many vision overlaps and a good resonance on strategies. I’m interested to explore ways to assist you.”

The rest of that message is a list of links to templeton.org, to a symposium series Harper describes in his own words as one “I created” (the Humble Approach Initiative), and to the Cambridge University Press volume Universe or Multiverse? that came out of it. He was no longer at the Foundation; what he was showing Epstein was the format.21

March 2010: “would it be appropriate to use your name”

Through February and March the traffic is short and practical: calls arranged from airports and taxiways, “Good to hear from you”, “I can be free this Saturday”.22 Then on 30 March 2010 Epstein put a project to him by telephone, and Harper answered the same evening under the subject “Quick question on confidentiality and strategy”. His first question is the one that matters:23

“1. For initial preliminary research, would it be appropriate to use your name or alternately refer simply to a “serious philanthropic donor who generously supports advance research” ? (My initial go-to person would be Anton Zeilinger of the University of Vienna, —a good friend who is precisely the kind of broad-minded visionary with super worldwide connections for getting started.)”

The other three questions ask whether it is a two-day event, how far the meeting may be described as serious philanthropy leading to grants and possibly “a formal RFP program”, and whether an anniversary might be used as a hook, “(say) Alan Turing’s seminar cryptographic work in 1940 at Bletchley Park 70 years ago ?”.23 The typeface is a monospaced one and the sheet carries no redaction.

The record answers the first question a month later, in Harper’s own words. On 30 April he wrote: “I’ve acted upon it immediately by sending a letter to Max Tegmark. I have asked him for initial confidentiality, but (as we agreed in our initial discussion) have explained that the identity of the donor would not be confidential in this project”.24 No document read here records Epstein’s side of that agreement in his own words; the sentence is Harper reporting a decision already taken between them.

April and May 2010: the essay contest, and the refusal

The plan at the end of April was an essay competition run through FQXi, the Foundational Questions Institute, whose winners would then be invited to the meeting. Harper set out an “off-the-cuff” budget of three prizes totalling $40,000 and “FQXi administration and advertizing: $35,000”, total $75,000, and offered to do the work himself for nothing, “presuming a modest effort situation”, because he was “in a new job situation”. Epstein replied the same day:25

“didn;t get the note.. however i agree to all.. with adjustment , first prize 15k send 10 third 5, fqxi. 25 for overhead. re your time , i agree that this time its a no cost. nut would like to explore a more formal realtionship”

A month later it had failed. Early on Saturday 29 May 2010 Epstein wrote one line:26

“I have been swamped. how do we move forward with the fxq”

The sheet prints “fxq”. Its own extraction does not: the text layer of that sheet reads “ficq”, as does one further copy, while four other documents quoting the message extract it as “fxq”. The image is the authority and the string is “fxq”. Two hours and eleven minutes later Harper answered, and the answer is the most consequential passage in his file:9

“I did contact Max Tegmark about a month ago to propse the essay contest approach we discussed. He and his colleagues offered support but did not think that FQX should do it. Reasons they gave were that they saw the topic as too narrow and too technical compared to the essay contests they have been doing. It is possible that the real reason was prudence to avoid FQX, already quite “controversial” via Templeton support to become even more so via Epstein-related sponsorship of prizes. If so, I think they their perspective is not unreasonable in view of their agenda to build relationships with other donors, though ironically what you and I are working on would develop a nice topically-rich and deep program linkage with another major donor. My general advice in this matter is that we should build a relationship slowly and build trust.”

Read as a documented fact, that paragraph establishes three things and no more: that Harper approached Max Tegmark on Epstein’s behalf in about April 2010; that Tegmark and colleagues declined to run the contest and gave a reason of scope; and that Harper’s own reading, offered to Epstein as speculation and marked as such in the words “It is possible that the real reason was”, was that the refusal was about being associated with Epstein’s money. Nothing read for this page is Tegmark’s account of his own reasons, and this page does not assert what they were.

Two paragraphs on, he draws the operational consequence: “Since we would not do an essay contest via FQX, the organizing cycle can be faster to do the meeting.”9

The ordering of that Saturday is settled and not assumed. Harper’s message is printed with an explicitly zoned attribution, “May 29, 2010 8:12:19 AM EDT”, on one copy; the header field of another copy of the same message reads “Sat 5/29/2010 12:12:19 PM” with no zone. The same copy prints Epstein’s message as “On May 29, 2010, at 6:00 AM”, and the header field of Epstein’s own copy reads “Sat 5/29/2010 10:00:53 AM”. Both pairs differ by exactly four hours.92627 On six of the seven pairs read for this page the zoneless field at the head of a message runs four or five hours ahead of the quoted attribution of the same message, the UTC-to-Eastern interval, which is what fixes the field as UTC and the attribution as local: five hours on the January pair, EST, and four again on 11 July and 30 October.151628 The claim is bounded to those pairs and does not generalise, because a seventh pair on this page’s own sheets runs to seven hours. The attribution line is rendered in the zone of whichever client quoted the message, so it is not a fixed offset from the field above it; the exception is set out in the note.27

June to October 2010: keeping it alive

From June the correspondence is Harper trying to convert enthusiasm into a decision. On 21 June he wrote under a subject line the sheet prints as “Do you want to move forward on the cryptograpjy-in-nature project ?”, quoting his own earlier note: “Question: Do you want to move forward on the cryptography-in-nature project ? You had contacted me for a discussion a few weeks ago… However, we have not been in communication since.”29 The same morning he wrote to Roger Schank, copying Epstein, about a book title.30 On 23 June he offered Friday 2 July for a meeting at MIT with Scott Aaronson and Seth Lloyd.31 On 11 July he sent a numbered ten-point plan, including a suggestion that the meeting be made “more strategic than that of ‘just a good interesting meeting’” by using it to shape “a small initial grants competition to offer, say, 5 grants at the $10K-to—$25K level with a total grants budget of, say, $75K to $100K”, and a list of possible venues:32

“MIT, Nantucket your home in Palm Beach, your home in the Virgin Islands, your home in NYC”

On 15 October he wrote: “Call me if you’d like to move forward on the ‘Cryptography in Nature’ project. I have a few weeks free and can focus on the task and make-it-happen.” Epstein answered that he was travelling in Europe.33 On 30 October Epstein asked “did you send something”, Harper answered “No. (were you expecting something from me ?…)”, and Epstein replied:28

“I think you should send a proposal that includes your time.”

4 November 2010: the proposal

Five days later Harper sent it. The covering message of Thursday 4 November 2010 at 10:30 PM is four lines, “Please find attached a project discussion and proposal document in reply to your request”, and the attachment survives in the release in its own right, as an eighteen-sheet document under a printed letterhead reading “FROM THE DESK OF / CHARLES L. HARPER JR.” with a redaction bar at each side of it, a third at the foot centre, and the web address WWW.VISION-FIVE.COM printed at the lower left.341 The release holds three copies of it. Its opening sentence:1

“Thank you for a request for a proposal to support my activity to develop the “Cryptography in Nature” project you have envisioned.”

The extraction renders that one printed letterhead in four different ways across the three copies and their sheets: “FROM 1* DESK OF”, “FROM 14E DESK OF”, “FROM “HE DESK OF” and “FROM DE DESK CF”. The project title is rendered “Coptograrn in Nature” and “Cgpograrn is Nature”. None of those readings is what the sheet prints; the letterhead is “FROM THE DESK OF” and the title, in italic inside quotation marks, is “Cryptography in Nature”.135

The body is a talent-scouting document. He names eight researchers he had “discovered” himself (Yaakov Benenson, Alan Aspuru-Guzik, Mohan Sarovar, Peng Yin, Radhika Nagpal, David Soloveichik, Erik Winfree and Gregory D. Scholes), explaining that he had deliberately gone outside the lists Aaronson and Lloyd had already given him because “One of my concerns has been to avoid gathering together a number of already closely-connected research colleagues linked with MIT”.36 Three sheets later he puts a number on those lists, in a bold heading: his eight are “(ADDITIONAL TO THE ~20 NAMES I HAVE FROM MEETINGS WITH SCOTT AARONSON AND SETH LLOYD)”.37 He flatters Epstein as “an outstanding talent scout yourself”, and on the following sheets proposes five scientific anniversaries falling in 2011 that could be used as hooks, from Turing’s 1936 paper to Feynman’s 1981 lecture.3638

The proposal proper opens on printed page 14, under a rule, with a single bold line, “The following proposal provides specifics on tasks, time and budgeting.”, and states the object:7

“To develop a very-high-dynamism two-day research meeting involving about 8-10 researchers (plus perhaps up to about 4 observers) focused on the theme:

“Cryptography in nature”

with a goal to explore, frame and inspire new deep innovation agendas. The project includes developing practical options for developing post-meeting strategic research sponsoring opportunities.”

The work tasks run from there onto printed page 15, and include “to play the role of meeting manager representing the donor at the event” and “To be available to vector attendees with any what-to-do issues regarding reimbursement via Jeffrey Epstein’s operations if/as needed”. Then the money:39

“Time estimate: 6 weeks FTE preparation effort + direct meeting activity time (~5 days) + follow-up strategy development (~5 days). Total: 8 weeks Fee: $90,000”

Expenses are on the same sheet, “Limited to $10,000” and payable on initiation.39 Printed page 16 then carries the summary (planning management and running $100,000, contingent honoraria $10,000, pre-meeting exchange $6,000, extra for co-chairs $2,000, meeting travel and lodging “TBD”) and the total:8

“Total: $118,000 + Meeting T&L & special events”

Direct meeting expenses, the same sheet says, “would be covered directly by Jeffrey Epstein’s people.”8 An appendix revives the essay contest as a post-conference option, “via the FQX network (if Max Tegmark agrees. Max would be invited to, and warmly welcomed to the meeting as an expert-observer)”, priced at three prizes of $15,000, $10,000 and $5,000 with “Overhead to the FQX program: $25,000”. Those figures are printed beside a stated “Total expense: $45,000”, which is $10,000 less than the parts he prints above it, and the discrepancy is on the sheet and is not corrected here.40 The same appendix offers as an alternative to “invest substantially in one to a few researchers in the same modality as with Martin Nowak (which seems to have been a huge success)”.40

What Epstein did with it

Five hours later, at 7:29 by the zoneless field at the head of his own message, Epstein forwarded the whole proposal to a third party, barred in the addressee field but named in the body of the message and again by the reply it drew, and offered to fund him instead:41

“I d rather do a conference on hacking in bioligy.. v=ruses hosts. self deception, encryption, code theory, Pablos, if =ou would like to organize a hacking in nature conference i will support.”

Pablos Holman answered that afternoon, and his reply is the only outside assessment of Harper’s proposal in the file. He declined the role Epstein had offered him and then, in its last substantive line, told Epstein to proceed with Harper’s event anyway:42

“Thanks for the offer, these are exciting concepts to be working with. Honestly, as much as I love the challenge, I’m out of my league in these discussions. I’m sorely undereducated about biology, and would just add to the noise floor.

This proposal describes a pretty cool event, but it does feel like getting together a pretty obvious group of folks who probably know each other’s work already. What I could imagine contributing would be some of our experience in hacking computer systems…

Just go ahead with this event and let me throw a couple hackers into the mix.”

Harper, who was copied on nothing of that, wrote to Epstein the same morning: “Thanks for quick feedback! I can begin to work on issues you’ve raised if you decide to go forward contracting my services.”43

He went on sending material for another week: a Cambridge University Press book of his own “(w 16 Nobel Laureates!)” on 5 November, a note on Darwin and deception in orchids on 12 November at 1:57 AM, and eleven and a half minutes after that “one more interesting fact…”, on the 550th anniversary of Alberti’s De componendis Cifris falling in 2011.444546

That is the last document in this page’s population in which he speaks. No document read for this page records a contract, a payment, a meeting, or a reply to those last three messages, and this page does not assert that any of them happened. The search made for one is set out below.

On the circulation list, and after

Separately from the project, Harper was on Epstein’s science distribution list. On 22 June 2010 at 9:24 AM Lesley Groff sent a one-question circular from Epstein, “What are your summer plans? Can you bring me up to speed on what you are w=rking on?”, to a copy list on which the names printed between the bars are Mark Jude Tramo, Victoria C Stodden, Lawrence Krauss, Lee Smolin, Nathan Wolfe, Lisa Randall, “=an Dubno”, Pablos, Danny Hillis and “Ch=rles L. Harper Jr.”.47

Those = characters are printed on the sheet. They are not an extraction fault: the image itself carries them, in the copy list and throughout the message body (“cours=”, “inf=nts”, “w=rking”), which is what a failure to decode a quoted-printable transfer encoding looks like when it is then printed to paper. The practical consequence is that three copies of that circular are invisible to every correct-spelling search for this man’s name, and they were reached for this page only by searching the damaged form.473

After November 2010 the file goes quiet. Every document in this page’s population dated 2011 or later, one in 2013, one in 2014 and eight in 2015, is an automated LinkedIn network digest addressed to Epstein, in which Harper appears as one of Epstein’s connections alongside names including Robert Trivers and Dan Ariely. All thirteen were classified from the sender name on their own top-level header, which in every case is a LinkedIn notification address; one was read in full. None of the thirteen is correspondence between people, and none is an act by anybody.48

The Templeton paper

The release holds two copies of a thirty-five-sheet paper whose title block reads:2

“Optimizing Program Impact and Cost-Effectiveness Philanthropic Venture Partnership Opportunities for the Non-profit Community Charles L. Harper, Jr., D.Phil. Executive Director/Senior Vice-President John Templeton Foundation Radnor, Pa”

It carries no date on its face, and nothing read for this page shows how or when it reached Epstein’s files. It is cited here for one thing only: of everything read for this page it is the only document he wrote himself that states his Foundation title, and its subject is the method he was later selling: introducing “a variety of entrepreneurial concepts, practices, and standards of planning and evaluation into non-profit activity”.2 Everything else about that paper, and about the Foundation, belongs to the page for the John Templeton Foundation.

How the population was measured

Search forms and returns, run against the full-text index on 2026-09-11, counting distinct document identifiers. The three phrase forms of the name return "Charles L. Harper" 140 documents, "Charles Harper" 100 and "Chuck Harper" 47; their union is 198 documents, and that is the population the desk measured.3

Five further documents were reached only through forms the correct spelling cannot see, and they are added here, making 203. Two are Harvard publication lists citing a volume he edited, reachable only as "CL Harper". Three are the copies of the June 2010 circular, whose extractions carry "Ch=rles L. Harper Jr.", "Ch=rles L Harper J" and "Ch=rles I.. Harper Jr.".3 Forms tried that returned nothing of his: "Charies Harper", "Charles Harpe", "Chas Harper", "C. L. Harper", "Dr. Harper", "Charles Harpar", "Charles Harpur", "Charles Hatper", "Charles Haiper", "Chuck Harpe". The desk’s damaged-spelling generator produced two mechanical variants of "Charles L. Harper" and both returned zero.3 The 203 is therefore a floor, not a ceiling; a name carried in a form nobody has guessed is not reachable by any of these queries.

Act against mention, measured over all 203 and not sampled. A document counts as an act if it carries something he wrote. On that test, 62 documents of 203, 31 per cent, are acts: 52 in which the message at the head of the document is his, 5 in which he speaks only inside a message quoted or forwarded by someone else, and 5 non-e-mail documents he authored, being the three copies of the November 2010 proposal and the two of the Templeton paper. All sixty-two are named in the note.49

A further 43 documents carry him in an addressee or copy field with no word of his in them, and 98 are mentions by others, thirteen of those being the LinkedIn digests.49

This differs from the desk’s own sample, and the reason is mechanical. A sample of fifteen classified by the top-level From: field gave three acts. Two faults in that test pull the figure down. The extraction of this production frequently renders the label From: as Fran: or Front, so a document whose header is his is read as having no sender at all; and a document with no e-mail headers, such as the proposal and the paper, falls into “named by others” however plainly he wrote it. Correcting for both, and classifying authored documents as acts, roughly a third of his file is him acting rather than being discussed.49

What these documents do not establish

Whether anything was paid, signed or held. Searches of the full-text index for "Cryptography in Nature" and "cryptography-in-nature" return 17 documents, every one of them inside this population and none later than 12 November 2010. "hacking in nature" returns 2, both of them the November 2010 forward and its reply. "Harper" with "retainer" and "Harper" with "wire transfer" each return nothing. "Harper" with "invoice" returns 16 documents, none of them this man: fourteen carry no legible instance of the surname at all and two are an interior-design business trading under it. "Chuck Harper" with 2011 returns nothing.50 No document recording a contract, a payment or a meeting was located, and this page does not assert that none exists; only that these searches did not find one.

Who paid for the January flights. The question was asked on the sheet and not answered on it. See above.

What Max Tegmark’s reasons were. The release holds Harper’s report of them and his speculation about them, and nothing from Tegmark.

Any employment, role or title after the Foundation beyond what he wrote himself. He describes himself in April 2010 as being “in a new job situation” with “a serious MIT-focused project in process with the Philanthropy Roundtable”, and his signature block from January 2010 onward, and the foot of his own printed letterhead in November, carry a web address at the vision-five.com domain. No document read for this page says what that is, whether he owned it, or what role he held there, and none of those is supplied here.25201

Anything about the person Epstein told to organise the trip beyond her forename. The release names her, on two of the eleven copies of one message, as Sarah; no document read for this page gives a surname, a role or an employer for her, and none is supplied here.

Privacy and citation notes

No personal telephone or fax number, private e-mail or postal address, identity or account number or date of birth is quoted, paraphrased or located anywhere on this page, and nothing here says which sheet carries which. Each of those seven classes was searched for by hand in the finished text: telephone and fax by the digit groupings and by the words tel, fax, cell and phone; e-mail by the address shape; account, routing and card numbers by runs of eight or more digits and by the words account, routing, swift and iban; identity numbers by the three-two-four grouping; dates of birth by DOB, born, birth date and the infobox Born row; addresses by a house number followed by a thoroughfare word, and by a two-letter state code before a five-digit code. The only hits were the word account in its ordinary sense, the word telephone in a quoted subject line and in descriptions of redactions, and the document identifiers themselves, which are eight digits long. Redaction bars are reported only where the fact of the bar is itself the finding: the name inside Epstein’s “please organize” instruction, the barred To: field of one January travel message and the barred Cc: field of another, and the bars flanking the proposal letterhead.

Of the 97 identifiers cited below, registers.py returns “not in any register” for every one, so all are citable and quotable. That verdict is also what the tool returns for an identifier that does not exist, so it was checked against a known SUPPRESSED entry as a positive control, which returns a different verdict; every identifier here was separately resolved by opening its rows.

What is left to other pages

Barnaby Marsh’s side of the May 2009 exchange is his page’s, not this one’s. In particular, Epstein’s closing clause to him, “will he be replaced or will you move up , (my hope)”, is about Marsh’s advancement inside the Foundation and not about Harper, and it is quoted here only as part of the message in which it stands.5 Marsh’s report of 7 May 2009 that he had “finally heard from Chuck Harper, who is visiting anthropological sites in South Africa at the moment” is already on that page and is not restated.

Everything about the John Templeton Foundation itself, its officers, its application process, its grant rules and the money entered against its name, belongs to the page for the Foundation. The thirty-five-sheet grantmaking paper is that page’s anchor; this page takes from it only the author block and the subject.2

The Couq Foundation page carries what is established about that entity. What is added here is one unanswered question on one sheet, and it is not evidence that Couq paid for anything.20

Martin Nowak, Lawrence Krauss and the MIT Media Lab have pages of their own; this page cites their names only where a document read here prints them.

Footnotes

  1. Charles L. Harper Jr., proposal to Jeffrey Epstein, 4 November 2010, 18 sheets. EFTA01735494 p.1. Read at the image at 400 dpi on 2026-09-11. The letterhead is set in letter-spaced grey small capitals reading “FROM THE DESK OF” over “CHARLES L. HARPER JR.” with a rule beneath; a redaction bar stands at each side of it and a third at the foot centre of the sheet; “WWW.VISION-FIVE.COM” is printed at the lower left and the folio “1” at the lower right. The addressee block reads “Jeffrey Epstein” with “(18 pages)” set against the right margin. The project title is set in italic inside double quotation marks. His own errors — “fascinatingly deep and difficult which engages”, and “Adelman” for the cryptographer’s name — are on the sheet. ↩ ↩2 ↩3 ↩4 ↩5 ↩6

  2. Charles L. Harper, Jr., Optimizing Program Impact and Cost-Effectiveness: Philanthropic Venture Partnership Opportunities for the Non-profit Community, John Templeton Foundation, Radnor, Pa, undated, 35 sheets. EFTA00605502 p.1; a second copy at EFTA02444770 p.1. EFTA00605502 p.1 read at the image at 400 dpi on 2026-09-11. The title is bold italic; the subtitle, the author block, the affiliation and “Radnor, Pa” are regular type; “Overview” is underlined and “Contents” bold. No date appears on the sheet, and no redaction. The second copy was not rendered. ↩ ↩2 ↩3 ↩4 ↩5

  3. Full-index FTS5 phrase searches over text_content, run on 2026-09-11, counting distinct document identifiers and never scoped by dataset. "Charles L. Harper" 140 documents (200 pages) · "Charles Harper" 100 (111) · "Chuck Harper" 47 (47) · union 198 · "Caleb Harper" 192, intersection with the union nil · bare "Harper" 834, of which 444 lie outside both men. Punctuation is a token separator in this index, so "Charles L Harper", "Charles L. Harper Jr." and "Charles L. Harper, Jr." are not separate queries and return the same or a subset population. Additional forms that returned documents of his outside the union: "CL Harper" 2 documents, both outside · "Harper Jr" 140, three outside, of which one is a homograph · "Ch=rles L. Harper" 2, both outside; "Ch=rles I.. Harper Jr" 1, also outside. The bare damaged forename "Ch=rles" returns 6 documents, the three above and three homographs: an airport name in an itinerary and a different man’s sender line, twice. Forms returning nothing of his: "Charies Harper", "Charles Harpe", "Chas Harper", "C. L. Harper", "Dr. Harper", "Charles Harpar", "Charles Harpur", "Charles Hatper", "Charles Haiper", "Chuck Harpe", "Harpe r", "Ha rper". scripts/ocr_variants.py "Charles L. Harper" --count generated two measured-glyph variants and both returned zero; "Chuck Harper" generated none. The five out-of-union documents added to make 203 are EFTA00760692, EFTA02443859 (Harvard publication lists citing “In Spiritual Information, ed. CL Harper”), and EFTA02403481, EFTA02530296, EFTA02684558 (the June 2010 circular).

    YEARS. The figure quoted in the infobox takes, for each of the 203 documents, the earliest full calendar date (day, month and year together, in any of the three printed forms) appearing anywhere in its text; bare four-digit years are ignored, because the proposal copies print 1936, 1945, 1981 and other bibliographic years that are not dates in the correspondence. On that method: 2003: 2 · 2009: 31 · 2010: 148 · 2013: 1 · 2014: 1 · 2015: 8 · no full calendar date at all: 12. Total 203. The two 2003 documents and one of the twelve undated are LinkedIn digests; the rest of the twelve are contact sheets, the Templeton paper, the forensic contacts export and the two Harvard publication lists. A different rule gives different years: counting bare years instead moves thirteen documents out of 2010, so the rule is stated rather than the number alone. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8

  4. Barnaby Marsh to Jeffrey Epstein, “RE:”, Tue, 05 May 2009 20:05:22 +0000, quoting Epstein’s message of “Tuesday, May 05, 2009 11:52 AM”. EFTA00774603 p.1, a one-sheet document. Read at the image at 400 dpi on 2026-09-11. Marsh’s reply is printed in blue; his own address is barred in the sender field. On the image the “Subject:” of Epstein’s quoted message is a printed label with nothing after it and “charles harper-eastern europe ideas?” stands well below it as the body; the extraction of this sheet joins the two and prints “Subject: Charles harper-eastern europe ideas?”, which is a subject line the sheet does not carry. Two standalone copies of Epstein’s own message, EFTA00881996 p.1 and EFTA01828991 p.1, carry it with an empty subject and the line as body, and were read in the text layer only. The 11:52 AM stamp is Eastern against the 15:51:39 +0000 on the standalone copies, four hours, agreeing with the convention derived below. ↩ ↩2 ↩3

  5. Jeffrey Epstein to Barnaby Marsh, 29 May 2009, subject “Re:”, quoting Marsh of the same morning and Epstein’s own message of Thursday 28 May 2009 5:57 PM. EFTA01829457 p.1, a one-sheet document. Read at the image at 400 dpi on 2026-09-11. The sheet prints two commas, not a low-double-quote, in “i’d love to,,”, and prints “experieince”; neither is repaired here. The subject field of the 28 May message is a printed field with nothing after it, and “charles harper?” is the body, set well below the header block; the reply’s subject is “Re:”. Marsh’s address is barred in the addressee field and again inside the attribution line. ↩ ↩2 ↩3 ↩4 ↩5

  6. Jeffrey Epstein to a recipient whose name and address are barred, “Re:”, Thu, 28 Jan 2010 13:08:35 +0000. EFTA00888203 p.1, a one-sheet document. Read at the image at 400 dpi on 2026-09-11. The sentence carries a terminal full stop on the image. “novak” and “krause” are Epstein’s spellings; that “novak” is Martin Nowak rests on the message of eleven days earlier addressed to him by name (see the next-but-one note), and that “krause” is Lawrence Krauss rests on Harper’s message of 4 February 2010 referring to “our last brief call from last Saturday when you were visiting with Lawrence K.” ↩ ↩2

  7. The same document, EFTA01735494 p.14. Read at the image at 400 dpi on 2026-09-11. The line “The following proposal provides specifics on tasks, time and budgeting.” is bold and set between two dashed rules; “PROPOSAL” is bold and centred; the theme is bold, centred, in double quotation marks, and printed with a lower-case n — “Cryptography in nature” — which differs from the italic, capitalised “Cryptography in Nature” on printed page 1 of the same document. The folio reads “PAGE” centred and “14” right. ↩ ↩2

  8. The same document, EFTA01735494 p.16. Read at the image at 400 dpi on 2026-09-11. “BUDGET SUMMARY:” is bold and centred; the five line items and the bold total are set as a right-aligned column. The four stated figures sum to the stated total. ↩ ↩2 ↩3

  9. Charles L. Harper Jr. to Jeffrey Epstein, “Re:”, header field “Sat 5/29/2010 12:12:19 PM”. EFTA02417288 p.1 of 2. Read at the image at 400 dpi on 2026-09-11. A second copy, forwarded by him three minutes and forty-three seconds later and dated “Sat, 29 May 2010 12:16:02 +0000” on its own header, carries the same message with an explicitly zoned attribution, “Date: May 29, 2010 8:12:19 AM EDT”: EFTA00753487 p.1 of 2, also read at the image at 400 dpi on 2026-09-11. Eight documents in the corpus carry this passage. Harper’s own errors — “propse”, “I think they their perspective”, “very string affirmation” — are on both images and are not repaired. The quotation marks around “controversial” are his. ↩ ↩2 ↩3 ↩4

  10. "Caleb Harper" returns 192 documents against the 198-document union used here, and the intersection of the two sets is empty, computed as a set difference over the resolved identifiers rather than inferred. The identification of him as a distinct person rests on the documents themselves, which place him at the MIT Media Lab and in traffic later than this page’s. No document of his is cited anywhere on this page. He is a separate subject and has no page at the time of writing. ↩ ↩2

  11. The 444 documents outside both men were not read exhaustively; the families named here were found by reading every return of the additional forms listed in the population note that fell outside the union, and by reading the returns of "Harper" conjoined with "invoice". They are: a Florida state legislator named in a newspaper clipping at EFTA00259898, on a pages row of 22,844 characters which is a container and not a sheet, so no page pin is given; two unrelated Harpers on Palm Beach County property rolls, in HOUSE_OVERSIGHT documents 016552, 016695, 016696 and 016698; the publisher HarperCollins and the imprint Harper Perennial, in book bibliographies; the artist Charley Harper, at HOUSE_OVERSIGHT_017526, whose single pages row runs to 221,293 characters and is likewise a container, not a sheet; and an interior-design business trading under the surname, at EFTA01925822 and EFTA02714298. UNVERIFIED, from the text layer. This is a list of families found, not a decomposition of the 444, and no total is claimed for them. ↩

  12. Charles L. Harper Jr. to Jeffrey Epstein, “Re: telephone call”, Sat, 16 Jan 2010 23:59:57 +0000, quoting Epstein of 6:40 PM the same day. EFTA00764989 p.1, a one-sheet document. Read at the image at 400 dpi on 2026-09-11. The image prints “Barnaby”; the extraction renders it “Bamaby”, the rn-to-m fault the desk has measured across this corpus. ↩ ↩2

  13. Charles L. Harper Jr., “Possibly visiting Palm Beach next weekend at the invitation of Mr. Epstein”, Sun, 17 Jan 2010 19:24:12 +0000, and quoting Epstein’s “please coordinate a trip” of 17 January 2:00 PM. EFTA00765012 pp.1-2 of 2; a second copy at EFTA02431324 pp.1-2 of 2. EFTA00765012 p.1 read at the image at 400 dpi on 2026-09-11. The To: field names Jeffrey Epstein; the Cc: field is barred and so is the name in the body salutation, which reads “Dear ▮,”. Epstein’s quoted line “please coordinate a trip” falls on the SECOND sheet of each copy, beneath the attribution that ends the first, which is why both pins are given as a range; the second sheet and the second copy were not rendered. The sheet prints the parenthesis of the blockquote above with a space, “next weekend! )”. ↩

  14. Jeffrey Epstein to Martin Nowak, no subject, Sun, 17 Jan 2010 19:32:22 +0000. EFTA00888401 p.1; a second copy at EFTA01821452 p.1. UNVERIFIED, from the text layer. The extraction renders Epstein’s doubled comma as a low-double-quote here as it does elsewhere; the form on the image has not been checked for this document. ↩

  15. Jeffrey Epstein to Charles L. Harper Jr., copy barred, “Re: Chuck Harper visit to Palm Beach / planning details”, header field “Tuesday, January 19, 2010 3:06 AM” with no zone stated. EFTA02528279 p.1 of 3. Read at the image at 400 dpi on 2026-09-11. A redaction bar sits between “great//” and “please organize”. The quoted message beneath is Harper’s of “Mon, Jan 18, 2010 at 9:31 PM” with three travel options, and a further bar stands where he writes “Please l=t me and ▮know which might be” — the sheet prints the damaged “l=t” and it is not repaired here. The = characters through this sheet are printed on the image. ↩ ↩2 ↩3

  16. Jeffrey Epstein, “Re: Chuck Harper visit to Palm Beach / planning details”, header field “Tue 1/19/2010 4:09:22 PM”, to two barred recipients. EFTA01822442 p.1 of 4. Read at the image at 400 dpi on 2026-09-11. This copy quotes the message of the preceding note as “On Jan 18, 2010, at 10:06 PM” — five hours earlier than that copy’s own header field, which is the January interval between UTC and Eastern Standard Time. The same redaction bar appears after “option 1 sounds great//” on this copy too, scanned separately; the extraction of it closes the gap and prints the sentence continuous. ↩ ↩2 ↩3

  17. An unnamed sender, From: field barred, to Jeffrey Epstein, “Re: Chuck Harper visit to Palm Beach / planning details”, header field “Tue 1/19/2010 3:17:12 PM”. EFTA02431251 p.1 of 2, and the same three strings again at EFTA02431284 p.2 of 2. Both sheets read at the image at 600 dpi on 2026-09-11. The sheet is the same message as the Couq question in the note above: its first body line is “Do I purchase his tickets from Couq or personal?”, and beneath it Epstein’s quoted instruction prints option 1 sounds great// sarah please organize, with two spaces after the double stroke and the forename in lower case. Harper’s quoted message below prints “Please let me and Sarah know which might be” where EFTA02528279 p.1 prints a bar, and the salutation “Dear Sarah,”. Barred on this sheet: the sender field, the address in Harper’s attribution line, his signature block, and the name after “FOR”. registers.py returns “not in any register” for this identifier. ↩ ↩2 ↩3 ↩4

  18. Full-index phrase searches, 2026-09-11: "option 1 sounds great" returns 4 documents and "option I sounds great" returns 7, with no document in both sets, giving 11. All eleven were read, and the eleven extractions were enumerated one by one. Four were rendered: EFTA02528279 p.1 and EFTA01822442 p.1 at 400 dpi, both carrying the bar, and EFTA02431251 p.1 and EFTA02431284 p.2 at 600 dpi, both carrying the name. Of the eleven extractions, nine close the gap after “great//” and two print “sarah” — the two rendered at 600 dpi, and on those two the image agrees with the extraction. The other seven were read in the text layer only. So the bar is visible on no extraction at all: an extraction-only reading of this sentence either shows the name or shows nothing. An earlier draft of this page reported the count of returns without having opened them, and said so; that is what the reading corrects. ↩

  19. Charles L. Harper Jr., “Re: Chuck Harper visit to Palm Beach / planning details”, header field “Tue 1/19/2010 3:27:54 AM”; the salutation reads “Dear Sarah,”. EFTA02431284 p.1 of 2. Read at the image at 400 dpi on 2026-09-11, and its second sheet at 600 dpi. The sole To: field is barred; the Cc: field names Jeffrey Epstein. Its second sheet is one of the two that print the forename, and is cited for that below. He selects “ex BOS at 8:20 pm on JetBlue #427” for Friday 22 January, asks for “Any reasonable overnight lodgings” at or near the airport, proposes a 7:20 pm return to Philadelphia on the Saturday at a quoted $244.70, and asks: “Would you be so kind please to purchase these flights for me as e-ticketings and send me the confirmation details electronically.” A further block beneath the line “(Data: Dr. Charles L. Harper, Jr.” is barred. ↩

  20. The question “Do I purchase his tickets from Couq or personal?” is printed on the same sheet as the preceding note, inside the quoted message of “Tue, Jan 19, 2010 at 10:17 AM”, whose sender’s name and address are both barred on that copy. Read at the image at 400 dpi on 2026-09-11. Epstein’s reply at the head of the sheet is “arrive fri night,, put in hotel” and does not refer to the question. Harper’s signature block on the same sheet prints the web address www.Vision-Five.com, with the adjacent telephone field barred. The same message survives at EFTA02431251 p.1 as its own top-level message, header field “Tue 1/19/2010 3:17:12 PM”, where the sender field is again barred but the body prints the sender’s forename; that is the sheet in the note below. ↩ ↩2 ↩3

  21. Charles L. Harper Jr. to Jeffrey Epstein, “Super to meet you!”, Mon, 25 Jan 2010 14:49:03 +0000, quoting Epstein of 24 January 10:59 AM. EFTA00765503 pp.1-2 of 2; a second copy at EFTA02432034 pp.1-2 of 3. Epstein’s quoted message of 24 January falls on the SECOND sheet of each copy, which is why both pins are ranges. UNVERIFIED, from the text layer. The links he sends are to templeton.org pages for the Humble Approach Initiative and to the Cambridge University Press catalogue entry for Universe or Multiverse?; the words “research symposia series I created” are his. ↩ ↩2

  22. Charles L. Harper Jr. to Jeffrey Epstein, “Re:”, Wed, 17 Feb 2010 14:38:10 +0000 (EFTA00763796 p.1) and Sat, 17 Apr 2010 16:07:03 +0000 (EFTA00760097 p.1). UNVERIFIED, from the text layer; neither sheet was rendered. ↩

  23. Charles L. Harper Jr. to Jeffrey Epstein, “Quick question on confidentiality and strategy”, header field “Tue 3/30/2010 5:36:20 PM”. EFTA02428217 p.1 of 2. Read at the image at 400 dpi on 2026-09-11. The sheet is set in a monospaced typeface, carries no redaction, and runs to four numbered questions; the extraction of it stops in the middle of question 3(d) and omits 3(e) and question 4 entirely, so the quotations here are taken from the image. The indentation above follows the sheet. ↩ ↩2

  24. Charles L. Harper Jr. to Jeffrey Epstein, “Re: Re-send: quick update”, Fri, 30 Apr 2010 14:18:11 +0000, one hour and seventeen minutes after the message in the preceding note. EFTA00759439 p.1. Read at the image at 400 dpi on 2026-09-11. The same sheet proposes a six-month timescale and names “Artur Eket, Anton Zeilinger, Vletko Vedral and Scott Aaronson” as examples of the people to be invited; those spellings are as printed. ↩

  25. Jeffrey Epstein to Charles L. Harper Jr., “Re: Re-send: quick update”, Fri, 30 Apr 2010 13:01:07 +0000, quoting Harper of 8:41 AM the same morning and, beneath that, Harper’s note of the preceding Friday. EFTA00733490 pp.1–2 of 3. Both sheets read at the image at 400 dpi on 2026-09-11. Page 2 prints the budget as “*Budget (conceptual, not negotiated. Off-the-cuff:): Prizes (3): $40,000. Top: $25,000, 2nd: $10,000, 3rd: $5,000 / FQXi administration and advertizing: $35,000. / Total: $75,000”, and Harper’s “I’m in a new job situation and have already a serious MIT-focused project in process with the Philanthropy Roundtable” and “I am happy to do this on a no-cost basis presuming a modest effort situation.” The telephone field in his signature block is barred. Epstein’s misspellings — “didn;t”, “nut would like”, “realtionship” — are on the image. ↩ ↩2

  26. Jeffrey Epstein to Charles L. Harper Jr., header field “Sat 5/29/2010 10:00:53 AM”, no subject field printed at all. EFTA01812622 p.1, a one-sheet document. Read at the image at 400 dpi on 2026-09-11, and the string is the reason the sheet was rendered. The image prints “fxq”. The extraction of this very sheet reads “ficq”, as does one further document; four other documents quoting the message extract it as “fxq”. A phrase search over the whole index for “how do we move forward with the fxq” returns 4 documents and for “…ficq” returns 2, and the two sets do not overlap. The image settles it. The header block prints To, From and Sent and no Subject line — an absent field, not an empty one. The word is not expanded here: the corpus’s own vocabulary in the same correspondence is “FQX”, “FQXi” and “fqxi”, and Harper’s reply two hours later is about FQX, but no document read for this page states what Epstein meant by the four letters he typed. ↩ ↩2

  27. The zone convention used on this page is derived, not assumed, and is checked on six separate pairs. (1) 29 May 2010: the message at EFTA02417288 p.1 has the zoneless header field “Sat 5/29/2010 12:12:19 PM” and is printed at EFTA00753487 p.1 as “May 29, 2010 8:12:19 AM EDT” — four hours. (2) The same sheet prints Epstein’s earlier message as “On May 29, 2010, at 6:00 AM”, against the zoneless “Sat 5/29/2010 10:00:53 AM” at EFTA01812622 p.1 — four hours. (3) 18–19 January 2010: “Tuesday, January 19, 2010 3:06 AM” against “On Jan 18, 2010, at 10:06 PM” — five hours, which is the same offset in Eastern Standard Time. (4) 5 May 2009: “Tuesday, May 05, 2009 11:52 AM” at EFTA00774603 p.1 against “Tue, 05 May 2009 15:51:39 +0000” on the standalone copies, four hours. Five of those six sheets were read at the image at 400 dpi on 2026-09-11; the standalone copies of the 5 May message were read in the text layer only. Two further pairs, 11 July and 30 October 2010, agree at four hours and were read in the text layer only.

    AND THE EXCEPTION, WHICH IS THE MORE USEFUL HALF. EFTA02684807 p.1 prints the zoneless header “Friday, November 5, 2010 7:29 AM”; EFTA00752490 p.1 quotes the same message as “On Nov 5, 2010, at 12:29 AM, Jeffrey Epstein wrote:”. That is seven hours, not four, and both sheets were read at the image at 400 dpi on 2026-09-11. Seven hours behind UTC is Pacific Daylight Time, and the client that produced this quotation belongs to the correspondent who answered it, not to Epstein. So the interval is a property of the QUOTING CLIENT, not of the production: the field at the head is UTC in all six pairs, but the attribution beneath it carries whatever local zone the quoter’s machine was set to. Every other pair read for this page happens to have been quoted by an Eastern client, which is why they agree at four and five hours. Any conversion on this page that rests on the four-or-five-hour interval rests on that coincidence and is scoped accordingly. ↩ ↩2

  28. Jeffrey Epstein to Charles L. Harper Jr., “Re:”, header field “Sat 10/30/2010 3:47:54 PM”, quoting Harper at 10:45 AM and Epstein at 9:33 AM: EFTA01982566 p.1 of 2. Read at the image at 400 dpi on 2026-09-11; all three quotations are exact on the sheet, the To: field names Charles L. Harper Jr. with the address barred, and the subject is “Re:”. Harper’s “Ok” of Sat, 30 Oct 2010 16:05:14 +0000, quoting the same instruction as sent “Oct 30, 2010, at 11:47 AM”: EFTA00753799 p.1, UNVERIFIED and not rendered. The four-hour difference between the two renderings of the same message agrees with the interval derived above. ↩ ↩2

  29. Charles L. Harper Jr. to Jeffrey Epstein, header field “Mon 6/21/2010 12:23:18 PM”. EFTA02411719 p.1, a one-sheet document. Read at the image at 400 dpi on 2026-09-11. Both spellings are printed on the one sheet: the Subject: field prints “cryptograpjy-in-nature” and the quoted note below it, set in a monospaced face with > markers, prints “cryptography-in-nature”. The difference is one character and it is the document’s own. His telephone number is barred in both the signature and the quoted signature. ↩

  30. Charles L. Harper Jr. to Roger Schank, copy to Jeffrey Epstein, “Re: Re:”, header field “Mon 6/21/2010 12:52:47 PM”. EFTA02411665 p.1 of 2. UNVERIFIED, from the text layer. ↩

  31. Charles L. Harper Jr. to Jeffrey Epstein, “Quick check on a possible date for an MIT meeting”, header field “Wed 6/23/2010 12:20:38 PM”. EFTA02411951 p.1. UNVERIFIED, from the text layer. ↩

  32. Charles L. Harper Jr. to Jeffrey Epstein, “Re:”, header field “Sun 7/11/2010 1:48:32 PM”. EFTA02410375 pp.1–2 of 4. UNVERIFIED, from the text layer. A later forward of the same message, EFTA00754357 p.1, prints its attribution as “July 11, 2010 9:48:32 AM EDT”, four hours apart, agreeing with the convention derived above. The venue list is quoted here as the sheet sets it out, one line each. ↩

  33. Charles L. Harper Jr. to Jeffrey Epstein, “Re: “Cryptography in Nature” project”, Wed, 20 Oct 2010 18:38:36 +0000, quoting his own message of Fri, Oct 15, 2010 at 7:58 PM and Epstein’s reply of “Oct 15, 2010, at 1:59 PM”. EFTA00754357 p.1 of 4. UNVERIFIED, from the text layer. ↩

  34. Charles L. Harper Jr. to Jeffrey Epstein, “Project discussion & proposal document attached”, header field “Fri 11/5/2010 2:30:00 AM”, the attachment named on the sheet in a damaged form beginning “Eoste” and ending “11-4-10”: EFTA02417368 p.1. A second rendering of the same message, dated “Thu, Nov 4. 2010 at 10:30 PM”, at EFTA02022606 p.1. UNVERIFIED, from the text layer; neither sheet was rendered, and the attachment filename is therefore reported only as a damaged string and not reconstructed. ↩

  35. Three copies of the proposal are in the corpus: EFTA01735494, EFTA00726413 and EFTA02417369, 18 sheets each. Their extractions render the same printed letterhead as “FROM 1* DESK OF”, “FROM DE DESK CF” and, on later sheets of one copy, “FROM 14E DESK OF” and “FROM “HE DESK OF”; the sheet count in the addressee block as “(18 pages)”, “(IS pages)” and “(I13 pages)”; and the project title as “Coptograrn in Nature” and “Cgpograrn is Nature”. Only EFTA01735494 p.1 was rendered, and it is the authority for every reading of that letterhead and title given on this page. None of the damaged forms is repaired anywhere on this page, and none is quoted as though it were what the sheet prints. One consequence is worth stating for anyone re-checking the pins: because the extraction of printed page 1 of every copy carries the title in damaged form, a text-layer search for “Cryptography in Nature” does not find it on printed page 1 and finds only the lower-case “Cryptography in nature” on printed page 14. The printed-page-1 reading given on this page is from the image and not from the text layer. ↩

  36. The same document, EFTA01735494 p.2, carrying the eight names and the sentence about avoiding “already closely-connected research colleagues linked with MIT”. Read at the image at 400 dpi on 2026-09-11. The eight names stand as a centred list with each affiliation in parentheses; the folio reads “PAGE” and “2”. The extraction of this sheet renders the sixth name “David Solovcichik”; the sheet prints “David Soloveichik”, and the page follows the sheet. The other seven are correct in both. No number appears on this sheet — it says only “the lists of names I’ve obtained from them” — so the “~20” is pinned to printed page 5 and not here. ↩ ↩2

  37. The same document, EFTA01735494 p.5. Read at the image at 400 dpi on 2026-09-11. The heading is bold and runs over five lines; the parenthesis quoted above is its second sentence. The same sheet carries the last two of the 2011 anniversaries, headed “Birthday connections:” in italic, for Leonard M. Adelman and Scott Aaronson, and the note that 2011 is “95 years since the birth and 10 years since the death of Claude Shannon”. “Adelman” is the sheet’s spelling. ↩

  38. The same document, EFTA01735494 p.4, the list of 2011 anniversaries. UNVERIFIED, from the text layer; not rendered. ↩

  39. The same document, EFTA01735494 p.15. Read at the image at 400 dpi on 2026-09-11. “PROJECT MANAGEMENT DIRECT BUDGET” is bold; “Fee:” and “$90,000” are bold, as is “Limited to $10,000”; the words “very high quality” in the parenthesis above them are italic. The folio reads “PAGE” and “15”. ↩ ↩2

  40. The same document, EFTA01735494 p.17, the appendix of post-meeting options. Read at the image at 400 dpi on 2026-09-11. The heading “POSSIBLE POST-MEETING PROJECT OPTIONS & BUDGETS” and the two option headings are bold; the three figures sit in a right-hand column. The sheet prints “Prizes (3): 1st: $15,000; 2nd: $10,000; 3rd: $5,000 Total: $30,000” and “Overhead to the FQX program: $25,000” and then “Total expense: $45,000”. Thirty thousand and twenty-five thousand is fifty-five thousand. The error is the document’s own and is reproduced, not repaired. The sheet also carries “(This can be done with or without management oversight by Charles Harper)”. ↩ ↩2

  41. Jeffrey Epstein, “Fwd: Project discussion & proposal document attached”, Friday, November 5, 2010 7:29 AM, addressee field barred. EFTA02684807 p.1 of 2. Read at the image at 400 dpi on 2026-09-11. The = characters in “v=ruses” and “if =ou” are printed on the image, as they are throughout this sheet; they are not an extraction artefact. “bioligy” is Epstein’s spelling. That the recipient is Pablos Holman rests on the name in the body of the message and on the reply in the next note, not on the barred field. ↩

  42. Pablos to Jeffrey Epstein, “Re: Project discussion & proposal document attached”, Fri, 05 Nov 2010 18:01:53 +0000. EFTA00752490 p.1 of 2. Read at the image at 400 dpi on 2026-09-11. The sender field prints the single word “Pablos” with the address barred, and the message is signed “Pablos.”; the identification as Pablos Holman rests on that and on Epstein’s own “Pablos, if =ou would like to organize” in the message it answers, not on anything further. The ellipsis in the blockquote above stands for two further sentences on the same sheet in which he offers what he could contribute (“describing lots of the attacks we use on computers to see if they can come up with biological analogues”) and adds “I’ve probably got it backwards and I should be looking to these bio guys for ideas on how to break computers more effectively!”. The “Just go ahead with this event” line is the last substantive line of the message and is the reason this sheet was rendered. This copy also quotes Epstein’s forward with no = substitutions at all, where EFTA02684807 p.1 prints them throughout: the decode failure is a property of that copy, not of the message. ↩

  43. Charles L. Harper Jr. to Jeffrey Epstein, “Re: Project discussion & proposal document attached”, Fri, 05 Nov 2010 11:15:58 +0000. EFTA00752492 p.1. UNVERIFIED, from the text layer. Nothing read for this page shows Harper being copied on the forward of the preceding two notes. ↩

  44. Charles L. Harper Jr. to Jeffrey Epstein, header field “Fri 11/5/2010 6:52:49 PM”, subject beginning “FYI: Just out, my latest edited book: (w 16 Nobel Laureates!): Visions of Discovery”. EFTA02414652 p.1 of 7; the remaining sheets are the publisher’s web page. UNVERIFIED, from the text layer. ↩

  45. Charles L. Harper Jr. to Jeffrey Epstein, “Very interesting research area on the evolution of deception”, header field “Fri 11/12/2010 1:57:43 AM”. EFTA02415652 p.1. UNVERIFIED, from the text layer. ↩

  46. Charles L. Harper Jr. to Jeffrey Epstein, “one more interesting fact…”, header field “Fri 11/12/2010 2:09:20 AM”. EFTA02415704 p.1. UNVERIFIED, from the text layer. This is the latest document in the 203-document population in which he speaks, measured by taking the latest date appearing anywhere in each of the 62 act documents. ↩

  47. Mark Jude Tramo to Lesley Groff and Jeffrey Epstein, “RE: Jeffrey Epstein”, Thursday, June 24, 2010 11:27 PM, quoting Groff’s circular of Tuesday, June 22, 2010 9:24 AM. EFTA02530296 p.1 of 2. Read at the image at 400 dpi on 2026-09-11. The copy list is broken by redaction bars; the names printed between them are as given in the body above, and “Ch=rles L. Harper Jr.” is printed with the equals sign, as are “Victoria C Sto=den” and “=an Dubno”. Two further copies carry the same list: EFTA02684558 p.1, whose extraction reads “Ch=rles I.. Harper Jr.”, and EFTA02403481 p.1, “Ch=rles L Harper J”; neither of those two was rendered. ↩ ↩2

  48. Thirteen documents in the population are automated LinkedIn network digests addressed to Jeffrey Epstein, in which “Charles L Harper Jr” appears among his connections: EFTA00638515, EFTA00674041, EFTA01152883, EFTA01196168, EFTA01201383, EFTA01204296, EFTA01206755, EFTA01744449, EFTA01747695, EFTA02366918, EFTA02398354, EFTA02400513 and EFTA02703440. UNVERIFIED, from the text layer; none was rendered. Ten documents in the population have an earliest full calendar date of 2011 or later, and all ten are on this list: one 2013, one 2014, eight 2015. Of the remaining three, two carry no calendar date earlier than a 2003 one printed inside the digest’s own markup, and one carries no full calendar date at all, so the year of those three is not established here. A different method gives different years for this set — counting bare four-digit years rather than full dates moves one document out of 2013 — which is why the method is stated. Robert Trivers and Dan Ariely are named in the 29 January 2015 digest. ↩

  49. Classification run over all 203 documents on 2026-09-11 and not sampled, then checked by eye. For each document the full text of every page was read by script for sender labels of the forms From:, Fran:, Front, Frcm and Fmm — the last four being renderings the extraction of this production produces for From: in the stacked To: / From: / Sent / Subject print layout, and the name following each was tested.

    ACTS, 62 of 203.

    (a) 52 documents whose own top-level message is his: EFTA00745785 · EFTA00749187 · EFTA00752492 · EFTA00753253 · EFTA00753487 · EFTA00753799 · EFTA00754357 · EFTA00754361 · EFTA00754883 · EFTA00759439 · EFTA00760097 · EFTA00760380 · EFTA00763796 · EFTA00763986 · EFTA00764222 · EFTA00764223 · EFTA00764287 · EFTA00764356 · EFTA00764546 · EFTA00764989 · EFTA00765012 · EFTA00765501 · EFTA00765503 · EFTA02022606 · EFTA02410375 · EFTA02411665 · EFTA02411719 · EFTA02411951 · EFTA02414652 · EFTA02415446 · EFTA02415652 · EFTA02415704 · EFTA02416498 · EFTA02417288 · EFTA02417368 · EFTA02417997 · EFTA02418004 · EFTA02422737 · EFTA02425284 · EFTA02428217 · EFTA02430298 · EFTA02430309 · EFTA02430343 · EFTA02430639 · EFTA02430819 · EFTA02430831 · EFTA02430860 · EFTA02430923 · EFTA02431015 · EFTA02431284 · EFTA02431324 · EFTA02432034.

    (b) 5 documents in which he speaks only inside a message quoted or forwarded by someone else: EFTA00727201 and EFTA02427794, both headed by a message of Lesley Groff’s carrying his reply beneath; EFTA00752490 and EFTA02684807, the November forward and its reply; and EFTA00893585.

    (c) 5 documents with no e-mail headers that he wrote: the proposal at EFTA01735494, EFTA00726413 and EFTA02417369, and the paper at EFTA00605502 and EFTA02444770.

    43 documents carry him only in a To: or Cc: field. The remaining 98 are mentions, of which 13 are the LinkedIn digests and 2 are the Harvard publication lists citing a volume he edited.

    Three false positives were removed by eye before the figure was taken: one LinkedIn digest and the two publication lists, each of which a script counting any Harper-shaped sender string would have called an act. An independent classifier run by the gate on this page reached 61 rather than 62, differing on the top-level set alone; the two documents in (b) headed by Lesley Groff are the class that separates the two runs, and they are acts on the test used here because they carry a message he wrote, while not being top-level. The desk’s own pre-brief sample of 15 gave 3 acts by classifying on the top-level From: alone; the two corrections that matter are the damaged sender label and the header-less authored documents.

    The label list above is complete for this page and incomplete for the corpus. The gate measured, over pages_fts, that 10,134 of 799,692 pages carrying an e-mail header shape hold no From token at all, that the five forms above reach 5,238 of them, and that twelve further renderings (rom, Fnxn, Fromm, Fm, Frm, Frorn, Fron, Fram, Erom, Frcrn, Ftom, Fiom) reach about 485 more. Re-running this page’s classification with all seventeen forms reclassifies none of the 203. ↩ ↩2 ↩3

  50. Absence searches run on 2026-09-11, with the forms and the returns stated so they can be repeated. "Cryptography in Nature" → 17 documents, all inside this population, none dated later than 12 November 2010. "cryptography-in-nature" → the same 17. "hacking in nature" → 2, both the November 2010 forward and its reply. "Harper" with "retainer" → 0. "Harper" with "wire transfer" → 0. "Harper" with "invoice" → 16 documents, all sixteen read: fourteen carry no legible instance of the surname anywhere on the returned page and two are the interior-design business named in the homograph note. "Chuck Harper" with 2011 → 0. "Vision-Five" → 15 documents, all inside this population. What a disconfirming document would say, written before the searches were run: it would carry a dollar figure beside his name, or the words engagement, agreement, invoice, retainer or wire, or a 2011 date beside the project title. None of those strings appears on this page except as search terms, so the searches are not circular. No such document was found; this page does not assert that none exists. ↩

Backlinks