Frederick Bligh Bond: The Architect Who Dug Glastonbury by Automatic Writing

Between 1908 and 1921 an English church architect directed the excavation of Glastonbury Abbey. Ten years in, he published a book explaining that a good deal of his digging had been guided by automatic writing — messages from monks dead for four centuries. He was fired. What almost nobody does is read the book he actually wrote, which is freely available and says something quite different from what both his admirers and his critics report.

Who Frederick Bligh Bond was

Frederick Bligh Bond was born in Marlborough, Wiltshire, on 30 June 1864, the son of a clergyman who was also his teacher. From 1888 he practised as an architect in Bristol. The buildings are real and some still stand: board schools across the city, Cossham Memorial Hospital, the schools of medicine and engineering at Bristol University, the Music School at Clifton College.

His reputation, though, rested on church fabric. In 1909 he published Roodscreens and Roodlofts in two volumes with Dom Bede Camm, and it made him an acknowledged authority on the interior architecture of the English parish church. This matters for everything that follows: when Bond looked at a broken wall footing, he was reading it with more background than almost anyone else in the country.

He was also, from early on, a joiner of societies: Freemasons in 1889, the Theosophical Society in 1895, the Society for Psychical Research in 1902, the Societas Rosicruciana in Anglia in 1909. As early as 1899 he had argued in print that the dimensions of the Glastonbury buildings encoded gematria — number symbolism drawn from scripture.

In 1908 the Church of England appointed him director of excavations at Glastonbury Abbey. The post was unpaid. He held it for thirteen years.

How the sittings worked

Bond did not work alone. His collaborator was John Allan Bartlett, a retired naval captain and, like Bond, a clergyman's son, who wrote under the name "John Alleyne". Bartlett held the pencil. Bond sat beside him, kept his hand on Bartlett's for contact, and asked the questions. The script came out in a mixture of Latin, Middle English and modern English, often fragmentary, sometimes illegible.

The first sitting took place on 7 November 1907, before Bond's official appointment. The principal voice that answered called itself Johannes — a monk named Johannes Bryant — and others followed. Bond eventually gave the whole group a name, the Company of Avalon, and published a further volume about them in 1924.

The scripts were not tidy revelations. They gave directions, dimensions, building histories — and complaints about the difficulty of communicating. One early sitting records a voice cutting in to say: "I think I am wrong in some things. Other influences cross my own."

What Bond said his method was — and what he denied

Here the primary text does something unexpected. The Gate of Remembrance runs to roughly 45,000 words, and it is freely readable. Search it for the vocabulary everyone assumes is there and it is mostly missing. The word "ghost" appears zero times. "Occult" appears once — and that once is inside a Latin phrase in the script itself, not in Bond's own prose. "Supernatural" appears once, in a sentence arguing against invoking it.

What is there instead is the vocabulary of psychology. "Subconscious" appears fifteen times, "automatism" nine, "telepathy" eight. And Bond states his position without hedging. Writing of himself and Bartlett in the third person, he records that "neither F.B.B. nor J.A. favoured the ordinary spiritualistic hypothesis which would see in these phenomena the action of discarnate intelligences from the outside upon the physical or nervous organisation of the sitters." They regarded that view, he adds, as something like the truth turned inside out.

His own proposal was that every individual consciousness is a fragment of a larger whole, and that the script reached into what he called "a cosmic Memory, conscious or unconscious, active or latent". Whatever one makes of that, it is not a claim about talking to dead people. He even names the failure mode himself, asking why a story produced automatically, if it turns out to be fiction, should be thought worse in origin than the fiction writers produce on purpose.

And then, in the same passage, the warning that is the most quotable line in the book and the least quoted: "Above all, let us not be superstitious. There is no need to invoke the action of supernatural agencies of a malevolent sort to explain the outcome of our own fallibility."

The Edgar Chapel: 580 feet and a dotted line

The centrepiece of the legend is the Edgar Chapel, built at the east end of the great church around 1500 and completely lost. The script gave its position, its length, and — crucially — the shape of its east end: not the flat wall an English chapel of that date would normally have, but a polygonal apse, the corners cut off at an angle. "The ending of the chappel was at an angle, the sides makyng as it were a baye in the east wall there."

Bond's own account of how that apse reached print is the most important paragraph in the book, and he wrote it himself. The excavation had cleared the rectangular end of the chapel and hit a sheer face of clay with nothing beyond it; nothing resembling "walls at an angle" had been found. Meanwhile a report was due for the Proceedings of the Somerset Archaeological Society, and two seventeenth-century sources, Hollar and Hearne, both gave the abbey church a total length of 580 feet — a figure the excavated rectangle fell short of.

So, in Bond's words: he "was fully persuaded that Hearne and Hollar are giving us fact", and therefore "boldly drew a plan of the excavated chapel with an addition in dotted lines, making out the requisite dimension and showing a polygonal annexe or apse". He published it in the 1908 Proceedings, and a facsimile appeared in The Treasury at Christmas 1908.

Only afterwards, over the Christmas holidays, did he spot "a small squarish patch of pale brown discoloration in the clay" and trace a wall stump running away at an angle. In his own summary table the result is recorded in capitals as proved by discovery in January 1909, after the plan had gone to press. Read charitably, that is a prediction confirmed. Read plainly, it is an architect filling a documented gap with the shape his reading suggested, publishing it, and then finding something he could fit to it.

The contradiction inside his own book

In the back of The Gate of Remembrance Bond sets out a scorecard: each claim from the script, then a heading called "EXISTING DATA" recording what was already known, then the result. Under item VII, "As to the polygonal east end", the existing-data line reads in full: "No existing data from which such might be inferred." That is the evidential core of the whole book.

Two items later, under item IX, Bond writes something that cannot stand beside it. Discussing an older chapel on the same axis, he notes that the antiquarian Phelps "has preserved a reminiscence of this chapel, and he suggests a semicircular end, which might mean a polygon." And in his opening historical chapter he had already described Phelps's published plan of the church — a plan showing, in dotted lines, a short projection at that very point "given a semicircular or apsidal end".

So a printed plan showing a rounded or angled east end at that location existed, Bond had read it, and he discusses it twice in the same volume in which he certifies that no such data existed. This is not a critic's reconstruction. It is an internal contradiction, in his own book, discoverable by anyone who reads all of it.

The same pattern shows up in the dimensions. The script gave the chapel's length as "30 virgas" — ninety feet — and Bond admits in his preface that it also said fifty at one point, calling this one of "some very obvious misstatements". Ninety feet was already in the documentary record he and Bartlett had been reading.

What the archive showed a century later

Between 1904 and 1979, Glastonbury Abbey saw thirty-six seasons of excavation under eight directors, and almost none of them published properly. From 2007 a project led by Professor Roberta Gilchrist at the University of Reading went through the whole surviving archive, added a new geophysical survey, and published the result through the Society of Antiquaries, with the full data openly available.

On the Edgar Chapel apse, the finding is blunt. The apsed termination suggested by the automatic writing was not confirmed by Bond's excavations. And Gilchrist records what he did anyway: despite the absence of archaeological evidence, Bond showed the apsed chapel on his published plans, and reconstructed it in the layout of the ruins on the site.

That last clause deserves a moment. Visitors walking the abbey grounds today, at the east end, are walking a ground plan that contains a feature which came out of a sitting with a pencil and not out of a trench. It was laid into the turf and it stayed there.

None of which makes the thirteen years worthless. Bond's excavations genuinely established the position and dimensions of lost buildings, including the Loretto Chapel. The reassessment's complaint about his years is the one it makes about most of the site's directors: the work was not analysed and not published, so the evidence had to be rebuilt from paperwork decades later.

Four claims from the script, and where each one stands

What the script said	Bond's verdict in 1918	Where it stands now
The chapel ends "at an angle" — a polygonal apse	Proved by discovery, January 1909	Not confirmed by the excavations; drawn on the plans and built into the site layout regardless (Gilchrist 2015)
Length "30 virgas" (90 feet)	Correct, after a contradictory "50" was discarded	Ninety feet was already in documentary sources both men had read (Wilkins 1923)
"Thin walls and poore foundations in the new work"	Proved by excavation	Consistent with what was dug; also what an architect would expect of a late, hurried addition
The Loretto Chapel, with nine sheets of structural detail	Confirmed in outline	The building was genuinely located; the decorative detail in the script was never independently testable

His contemporaries: McCabe and Wilkins

The criticism started immediately and it was specific. The rationalist writer Joseph McCabe, in Is Spiritualism Based On Fraud? (1920), pointed at the preparation rather than the results: Bond and Bartlett had, he wrote, "steeped themselves, all through the year 1907, in the literature of the subject. They read all that was known about Glastonbury, and lived for months in the medieval atmosphere."

What gives that line its force is that it is not really an accusation. Bond had said as much himself. The first sittings fall in November 1907, at the end of exactly the period of immersion McCabe describes, and Bond presents his own deep reading in the abbey literature as a qualification. The critic and the author agree on the facts and disagree only about what they mean.

The Reverend H. J. Wilkins went further, publishing a book-length examination under the title False Psychical Claims in 'The Gate of Remembrance' Concerning Glastonbury Abbey (Bristol, 1923), with an earlier summary in The Spectator in June 1922. His conclusion: "there is absolutely nothing supermundane in the whole of the script… All that is true in the script could be gathered from historical data or reasonably conjectured by intelligent observation of existing facts and conditions." Bond replied in the SPR's journal in 1924; Wilkins replied to the reply a month later.

Modern assessments have added mechanism rather than new outrage. The archaeologist Kenneth Feder notes the absence of any scientific controls, and that the towers the script located were recorded in a document Bond had almost certainly seen. The psychologist Chris French has written at length on the ideomotor effect — the way small, genuinely unconscious muscle movements produce writing the writer does not experience as their own.

After Glastonbury

Bond's dismissal came in 1921, from Joseph Armitage Robinson, Dean of Wells, who objected to mediumship being used in archaeology. The Edgar Chapel controversy had by then grown large enough to be raised in Parliament.

He kept working. From 1921 to 1926 he edited Psychic Science, the quarterly of what became the College of Psychic Science. In 1926 he emigrated to New York as education secretary of the American Society for Psychical Research and editor of its journal. He was ordained in 1932 in the Old Roman Catholic Church of America and raised the following year to a rank equivalent to bishop.

Two episodes from those years complicate the picture of a credulous man. In the long dispute over the Boston medium Mina Crandon — "Margery" — Bond began as a strong supporter and ended as a sceptic, backing the accusation that the wax thumbprints attributed to her dead brother were fraudulent. Taking that side cost him his ASPR position. He returned to England in 1936 and died at Dolgellau in Wales on 8 March 1945, aged eighty.

The other episode is a footnote in legal history. Working with the medium Geraldine Cummins on automatic scripts attributed to a first-century writer called Cleophas, Bond claimed the copyright, on the grounds that the material had been addressed to him and that he had typed it up. The case reached the High Court as Cummins v Bond [1927] 1 Ch 167. Mr Justice Eve declined to rule on whether a first-century author existed, held that any such author was in no position to claim, and awarded the copyright to the woman who had physically written the words. It is still cited in copyright teaching today.

What actually holds up

Strip away a century of retelling and three things survive. Bond really did direct thirteen years of excavation and really did locate lost buildings. He really did use automatic writing while doing it, and said so in print, which is why he lost the job. And the single most famous confirmation — the polygonal apse — was published as a dotted line before it was dug, was never confirmed archaeologically, and was nonetheless laid out in the ruins where it remains.

What does not survive is the version in which a sceptical scientist was handed information he could not possibly have had. Bond was not sceptical; he had been in the SPR for five years and the Theosophical Society for twelve before the first sitting. He was not short of information; he was one of the best-read men in England on exactly this kind of building. And his own scorecard's claim that no existing data could have suggested an apse is contradicted two pages later by Bond himself.

The real figure is more interesting than the legend. Bond rejected the spiritualist account of his own phenomena, proposed a psychological one, named the risk that automatic writing simply produces fiction, and warned his readers against superstition. Then he published a building he had not found. Both halves are in the same book.

Frequently asked questions

Did Bligh Bond really find the Edgar Chapel through automatic writing?

He found the chapel, and the script did give directions. But the most-cited detail — the polygonal east end — was drawn in dotted lines and published in 1908 before anything was found there, on the basis that two seventeenth-century sources gave the abbey a total length the excavated rectangle did not reach. Bond describes doing this himself in The Gate of Remembrance. The apse was never confirmed by excavation.

Did Bond claim he was talking to ghosts?

No, and he says so explicitly. He records that neither he nor his collaborator favoured "the ordinary spiritualistic hypothesis" of discarnate intelligences acting on the sitters, and proposed instead a shared "cosmic Memory" reached through the subconscious. In roughly 45,000 words the book uses "ghost" not once, and "supernatural" once — in a sentence arguing against invoking it.

Why was he dismissed?

Because he published his methods. The Gate of Remembrance appeared in 1918 and made the use of automatic writing public; the Edgar Chapel dispute reached Parliament; and in 1921 Joseph Armitage Robinson, Dean of Wells, removed him. The objection was to mediumship as an archaeological method, not to his competence as an architect.

Can I read the original book?

Yes, and it is the best correction to most accounts of it. The Gate of Remembrance (1918) is out of copyright and available in full at Project Gutenberg as ebook 48568, and scanned at the Internet Archive. The modern archaeological reassessment of all thirty-six excavation seasons is published openly through the Archaeology Data Service.

What is the ideomotor effect?

It is the well-documented production of movement by expectation rather than intention: small muscle actions that the person performing them genuinely does not experience as voluntary. It is the standard explanation for planchette writing, dowsing rod movement and Ouija boards, and psychologists including Chris French have applied it to the Glastonbury scripts. It does not require anyone to be lying.

Sources

• Frederick Bligh Bond, The Gate of Remembrance (Oxford: Blackwell, 1918) — full text, Project Gutenberg ebook 48568. All quotations from the script, the preface and the summary tables are taken from this text.
• Roberta Gilchrist, "Glastonbury Abbey: the archaeological story", Current Archaeology 320 (2016); and the Glastonbury Abbey Archaeological Archive Project, archive and reports at the Archaeology Data Service.
• Joseph McCabe, Is Spiritualism Based On Fraud? (London: Watts & Co., 1920), p. 141.
• H. J. Wilkins, False Psychical Claims in 'The Gate of Remembrance' Concerning Glastonbury Abbey (Bristol: Arrowsmith, 1923); summarised in The Spectator, 10 June 1922.
• Richard Coates, Frederick Bligh Bond (1864–1945): A Bibliography of his Writings and a List of his Buildings (University of the West of England, 2015).
• Melvyn Willin, "Frederick Bligh Bond", Psi Encyclopedia (Society for Psychical Research, 2023).
• Kenneth Feder, Encyclopedia of Dubious Archaeology (Greenwood, 2010), pp. 43–44.
• Cummins v Bond [1927] 1 Ch 167.

If this kind of case interests you, we have written about what the archaeology at Glastonbury Tor actually found, about Alfred Watkins and the ley lines he never called energy lines, and about geomancy as a divination system with its own rules.