storiesThe charm in the wall

Britain keeps a tally, and it is the largest number in this subject.

Northampton Museum maintains a Concealed Shoe Index. By 1998 it held more than 1,100 reports of shoes found in the fabric of buildings, mostly British but some from as far away as Canada. By 2012 it held 1,900, of which almost half date from the nineteenth century.

The earliest reported find is behind the choir stalls in Winchester Cathedral, which were installed in 1308. The buildings are otherwise ordinary and then suddenly are not: country cottages, town houses, manor houses, hospitals, workhouses, factories, public houses, two Oxford colleges — St John's and Queen's — a Benedictine monastery in Germany and a Baptist church in Cheshire. In the United States the finds concentrate in New England and the north-east, the regions first colonised from East Anglia. They also turn up in Australia.

June Swann's analysis of the index, published in 1996, gives the hiding places by share. Chimney, fireplace or hearth: 26.2 per cent. Under the floor or above the ceiling: 22.9 per cent. Almost as many again in the roof. The rest come from around doors and windows, under stairs, and among the foundations — and of the thousands reported, only thirteen in Swann's analysis were in foundations proper.

The shoes themselves are specific. Almost 98 per cent have been worn and many have been repaired. Most finds are single shoes rather than pairs. All sizes are represented, from babies to adults, with adult male and female shoes at 21.5 and 26.5 per cent — and about half of all the shoes found belonged to children. Archaeologists call a cache of this kind a spiritual midden.

Nobody who hid one wrote down why, so every account here belongs to somebody reading the deposits afterwards, and this record names them rather than choosing between them.

Brian Hoggard, an independent researcher, observes that the places the shoes turn up — chimneys, floors, ceilings, doors, windows, roofs — suggest that at least some were concealed as charms protecting the occupants against evil influences such as demons, ghosts and witches.

The explanation most scholars favour, on the account this project read, runs on scent: a witch is drawn by the human smell of a worn shoe, enters it, and is trapped, being unable to reverse. That explains the 98 per cent.

The archaeologist Ralph Merrifield suggested that an unofficial fourteenth-century English saint, John Schorne — said to have trapped the Devil in a boot — may be the source of the belief, a legend possibly converting an older folk practice into an approved Christian rite. Merrifield also recorded the Lancashire custom called smickling, in which a woman wanting to conceive tried on the shoes of a woman who had just given birth, which is the evidence behind the competing fertility reading.

M. Chris Manning, an archaeologist and architectural historian, proposes instead a household spirit: the belief in a helpful domestic being runs from Ireland to western Russia, the English and Scottish brownie and hob could be driven off by a gift of clothing, and the Russian domovoi could be drawn to a house by an old boot or bast shoe under the stove or hung in the yard.

Swann raised one further possibility, carefully, about the thirteen foundation shoes: a connection with the Carthaginian practice of putting infants into building foundations, with the shoe standing in for the person.

The other common deposit is a stoppered vessel, and the first description of one is a ghost story.

In 1681 Joseph Glanvill printed it in Saducismus Triumphatus. A Suffolk man's wife has been long in a languishing condition and is haunted by something in the shape of a bird. A travelling old man tells him to cork her urine in a bottle with pins, needles and nails and set it to the fire; the cork blows out with a report like a pistol and nothing is gained. The old man returns and changes the instruction — bottle it the same way and bury it — and she recovers.

The earliest surviving examples are salt-glazed stoneware Bartmann jugs, called Bellarmines or Greybeards, brown or grey with a bearded face embossed on them; later ones are plain glass. They go under the hearth, inside a wall, or into the farthest corner of the property, and the whole class works on the premise that it stays hidden and unbroken — which is why bottles buried under fireplaces have turned up only after the rest of the building was pulled down.

One of them has been analysed. A glass bottle dug from the threshold of a Cornish house appeared on Antiques Roadshow in 2016, filmed at Trelissick; the glass specialist Andy McConnell tasted a little of the contents and guessed port or wine, noting a rusty flavour and the nails. A 2019 episode gave the laboratory answer: Loughborough University found "urine, a tiny bit of alcohol, and one human hair", with brass pins dating from the late 1840s, and an ostracod.

The counts elsewhere are small and precise. In the United States fewer than a dozen possible witch bottles have been identified. The archaeologist Marshall Becker was the first to identify one in an archaeological context — the Essington witch bottle, recovered on Great Tinicum Island in Delaware County, Pennsylvania. A mid-nineteenth to early twentieth-century enslaved or tenant site in Dorchester County, Maryland, yielded one whose cork was bristling with straight pins. In 2016 a bottle of nails came out of the hearth of a Civil War site in Virginia. And in 2020, during the renovation of the De Zwarte Ruiter pub in Turnhout, a sixteenth-century jug used against witchcraft was found — the first such find in Europe outside England.

The European deposits go into a house already standing. The older practice puts the objects in before the walls exist, and it was formal, state-scale work.

Egyptian foundation deposits are mudbrick-lined pits or holes cut at set points beneath a temple or a tomb and filled with ceremonial objects — amulets, scarabs, food, miniature tools — meant to keep the building from falling into ruin. Deposits are documented at Hatshepsut's tomb, KV20.

Mesopotamia buried figures with names on them, from the Early Dynastic period onward, to mark the boundary of sacred ground. Early pegs are limestone or clay; the buried figures are cast copper alloy, running from plain inscribed cones to fully modelled human forms. A peg of Ur-Namma, c. 2112–2094 BCE, carries his bust and an inscription to Inanna, lady of Eanna. Figurines of Shulgi were found under the Temple of Nimintabba at Ur. Deposits are recorded at Ur, Uruk and Nippur, and the Temple of Ningirsu at Girsu produced pegs dated to about 2130 BCE; the practice runs to at least the reign of Rim-Suen of Larsa, 1822–1763 BCE.

Curses went under buildings too. Egyptian execration texts naming enemies were written on figurines or pottery, smashed, and buried beneath a building under construction. In Mesopotamia, incantation bowls were buried face down under thresholds and in the corners of houses.

Inference — the two habits are not the same practice and this atlas does not claim descent between them. What they share is the mechanism: a thing put where it cannot be checked, by people who expected it to keep working unseen. The Ur-Namma peg and the Cornish bottle are four thousand years and four thousand miles apart, and both were buried on the understanding that burial is what makes them work.

The sources for both classes above name the same short list of things hidden in buildings since at least the early modern period: written charms — what archaeologists call apotropaic marks — dried cats, horse skulls, concealed shoes, and witch bottles. Shoes are by far the most common items discovered.

For dried cats and horse skulls this atlas has a name and no number. Neither account read here publishes a find count, a site list or a date range for them, and this record will not estimate one. That is a gap on this disk, not a gap in the ground — the material exists, and a reader who wants the figures should go to the specialist literature on concealed animal remains rather than take a guess from here.

Inference — the whole subject is shaped by that asymmetry. The custom of hiding shoes appears to have more or less died out during the twentieth century, with only about fifty recorded instances after 1900 — among them a pair of Norvic women's high-leg boots put into the foundations of a new factory in 1964, and an estate worker's shoe at Knebworth House in 1991, replacing a court shoe found behind panelling. Those two are on the record because somebody wrote them down at the time. Everything else in this record arrived when a wall came down, which means the curve of finds over time is a curve of demolition, renovation and building survey — and any claim about how common the practice was, as against how often it is now discovered, is not available from these numbers.