Our Deep History Our Deep History A museum of unexplained ancient sites and relics
United States rooms

Kensington, Minnesota, and other reported finds, United States

Runestones in America

Reported finds from 1898 to 1971 (one stone dates itself to 1362)

The Kensington stone in Alexandria, Minnesota. The plaque under it states one reading. The pale patch on the lower left is on the inscribed face, where the 1910 geological report describes a skin of calcite. Mauricio Valle, CC BY-SA 4.0. Shown unedited.

A Minnesota farmer lifts a rune-cut slab from under an aspen in 1898, and the carving gives its own year as 1362. Oklahoma, Maine, and Rhode Island have stones of their own. The one Norse camp that has a calendar year, in Newfoundland, has no runes among its published finds.

Why this is extraordinary

In the fall of 1898 a Swedish immigrant farmer, Olof Ohman, was clearing land on a knoll in Solem Township, Douglas County, Minnesota, about three miles northeast of the station at Kensington. The Minnesota Historical Society's report of 1910 puts the day at 8 November. The stone lay face down under about six inches of soil, with a poplar, or aspen, rooted over it. Ohman was there, his sons Olof Emil, twelve, and Edward, ten, and the neighbor Nils Olof Flaaten. Visitors who saw the stump did not agree on the tree. Some said at least ten years, others twenty or thirty, and one guess was about forty. Sam Olson of Kensington remembered a trunk about ten inches through just above the stone, and roots flattened where they had pressed the rock. That flattening is the part that matters. A root takes the shape of whatever it grows against, so the stone had been under the tree for the life of the tree.

The slab is greywacke, a hard grey sandstone. The 1910 report measures it at 36 inches long, 15 inches across the face, and 5½ inches thick, about 230 pounds. Later accounts from the same society, and the runologist Henrik Williams, say about 200 pounds. Nine lines of runes run down the face and three more run along the edge. In the translation the society's encyclopedia publishes, eight Swedes and twenty-two Norwegians are on a journey of exploration from Vinland through the west. They had camped by a lake with two skerries, a day's travel north of the stone. They went fishing. They came back and found ten of their men red with blood and dead. The Latin letters AVM, read as Ave Maria, ask to be saved from evil. Ten more men are with the ships, fourteen days' journey from "this island." The year is 1362.

If that year is the year of the cutting, Norse travelers were in western Minnesota a century and more before Columbus, and more than three centuries after the only Norse camp in the Americas that archaeologists have confirmed. If the year was cut in Ohman's own century, the slab is a different kind of document. It is a piece of immigrant Minnesota, made with a runic alphabet that turns up in nineteenth-century Sweden, and it has kept a town, a museum, and a century of argument. Both readings are published. The stone does not sign either of them.

The same question, with different stones, shows up in Oklahoma, on the Maine coast, and in the mud of Narragansett Bay. One tablet said to have been found on the northern plains in the 1730s is lost, so there is nothing left to read. None of these is the Newfoundland site. Parks Canada describes L'Anse aux Meadows, found in 1960, as eight sod buildings and a handful of Norse objects, a bronze pin, a spindle whorl, a lamp, a needle, iron nails and slag. A runic inscription is not in that list. In 2021 a tree-ring study in Nature put the felling of three trees there in the year 1021, by counting out from a spike of radiocarbon that cosmic rays left in the air in 993. The dated Norse site has buildings and iron. It does not, in the finds Parks Canada lists, have a rune.

Sandstone shelters in eastern Paraguay have been read as Norse writing too. The published case is a set of books from the 1970s, and the excavation report catalogues those walls as Indigenous petroglyphs. Both readings are in the Amambay rock art room.

Why it matters

If any of these cuttings is medieval, Norse travel in North America is not only a short stay on a Newfoundland shore around 1021. It reaches the Great Lakes country, and on some readings the Arkansas, with writing in tow. If the cuttings are later, they are still a record of how runes, immigrant memory, and a good stone made a past people could argue over. You can weigh each stone on its own. They do not all stand or fall together.

A rune is a letter from the alphabets used in northern Europe from late antiquity through the Middle Ages, and again, in some districts, long after. The older row has twenty-four letters. A shorter row replaces it in the Viking Age. Greywacke is a hard sandstone. A skerry is a small rocky island. Vinland is the name in the Norse sagas for a land west of Greenland. AVM on the Minnesota stone is three Latin capitals, read as a short form of Ave Maria. The Newfoundland date is not a rune. It is a count of tree rings outward from the year 993, when a burst of cosmic rays raised the radiocarbon in the air all over the world.

Photo tour

5 photographs, unedited. On a phone, swipe the row. Choose one to look closely.

The Kensington stone in Alexandria, Minnesota. The plaque under it states one reading. The pale patch on the lower left is on the inscribed face, where the 1910 geological report describes a skin of calcite.Mauricio Valle, CC BY-SA 4.0
Both carved surfaces, published by George T. Flom in 1910. The edge, on the left, carries the lines that include the year.George T. Flom, Public domain
The face in the Alexandria case. Sharp grooves are what both geological readings start from. Winchell called them fresh. A 2003 study argued the weathering was older than about 200 years.Lorie Shaull, CC BY 2.0
The park on Poteau Mountain. This frame is the layered sandstone of the ravine. The eight marks are not readable in it.MARELBU, CC BY 3.0
The Shawnee stone as displayed. Five marks are cut across the face. The label is the exhibit's, and the reading MEDOK is Farley's transliteration, not a caption this photograph finishes.Heironymous Rowe, CC BY-SA 3.0

See the evidence at the source

These pictures are not openly licensed, so they are not copied onto this page. Each link opens in a new tab.

What is still unexplained

These sit beside the consensus account, with the same visual weight. None of them is marked solved.

Both readings are published The words, and the runes, on the Kensington stone The encyclopedia translation is the one visitors meet. Eight Goths, glossed there as Swedes, and twenty-two Norwegians, a journey from Vinland, two skerries, ten men dead, ten men by the sea, fourteen days, and the year 1362. Williams, publishing in 2012, prints a close reading that does not match it in every word. He has Götalanders, people of southern Sweden, not the island of Gotland, whose name in this spelling would be guter. He has "2 shelters(?)" where the encyclopedia has skerries. He allows "acquisition journey" beside "exploratory journey" for the debated compound, and "sea/lake" and "island/peninsula" where the other version chooses. The story is the same shape. The dictionary is not. The word at the center of the fight is opdagelsefärd, or however the runes are typed. It looks like modern Danish and Norwegian for a journey of discovery. Williams notes that the medieval record has no such word. He also sets out the other reading. If the rune that looks like d can be a t, the word can be rebuilt as an Old Swedish compound for a journey of taking land. That compound is not written down anywhere. It is a possible build from real pieces. He asks which is the probable one, and he is plain that no single word proves the text cannot be medieval. A text is the whole of it. In his judgment it does not look like any medieval Scandinavian text he has seen. He also names the words that pull the other way. Of, rise (twice), äptir (twice), illu, and mans are awkward if you want a nineteenth-century carver with only modern Swedish and Norwegian. From, in this shape, he says, never occurs in an Old Scandinavian text. The nearest form he cites, fruam, is from an old dialect on Runö, an island off Estonia. Richard Nielsen's search through charters of 1355 to 1375 found parallels that Williams, having checked them, says mostly do not hold. The ones that do hold are enough, Williams writes, that a medieval origin is not as easy to dismiss as specialists had said. A few words may be judged impossible for the Middle Ages. A few are just as hard to explain in the 1890s. His balance still falls on the nineteenth century, and he says that case is not finished either. One finished piece of it is a book. Williams writes that Ohman owned the 1891 edition of an encyclopedia called The Well-Informed Schoolmaster. On page 62 are the old forms aptir and manss. On page 64 a Lord's Prayer dated to 1300 ends frælse os af illu, save us from evil. Those are three of the medieval-looking words, a few pages apart, in a book in the finder's house. A carver who had that page did not need a university. A carver who never opened it can still have cut the same words. The book is a path, not a signature. The letters are the other half. Breda at the University of Minnesota and Curme at Northwestern, in 1899, concluded the inscription was not medieval. The 1910 report's word for their verdict is fraudulent, and the stone went back to Ohman that March. Williams's census, in 2012, is that every academically trained specialist in Scandinavian, English, or German studies who had published on the stone agreed, with two exceptions he names. Robert Hall and William Thalbitzer published in support of a fourteenth-century origin. A geologist, Scott Wolter, is the further name on his list. Williams's own line is that the evidence he has reviewed supports a nineteenth-century inscription, and that he will not call it a fake. A fake, he writes, needs an intention to deceive, and the intention is not known. Plenty of late runestones in Scandinavia are simply late. Where the odd runes themselves come from is a newer argument, and it points at Sweden in Ohman's century. In work published in 2003 and described by Williams the next year, sheets dated 1883 and 1885 turned up at the dialect archive in Umeå. A tailor, Edward Larsson, had copied a row his brother Carl Emil wrote down, a "secret style" of journeymen's runes. Williams says the row is closely related to the Kensington runes, with differences, and that where they differ the Minnesota forms are the older-looking ones. Those older shapes were not unknown. They are on the Larsson sheet too. The Swedish National Heritage Board's runologist Magnus Källström goes further with objects, not paper. A measuring stick at Haverö in Medelpad carries the same runes. It is laid out to the Stockholm ell, 59.38 centimetres, a unit that could not be used after 1888, so the stick is very hard to date after that year. In Hassela, about thirty kilometres away, red-chalk writing in a farmhand's room and a neighboring barn uses the same alphabet. Two of the barn notes are dated 1870 and 1877. The Board's account says these are the first such runes dated before 1879, the year Ohman came to Douglas County, and that the district is just over a hundred kilometres from the place he left. Källström's published comment is that the writing mixes Swedish, Norwegian, and English features, uses modern word forms, and matches no runic system of the Middle Ages. A 2017 note by Staffan Fridell and Mats G. Larsson starts from that late date and asks what dialect the language is. They treat rise, with a weakened end vowel, fiske, a worn past tense, and fro, for "from," as the sort of small habits a speaker does not invent on purpose. A fourth word, which they read skelar and translate as sheds, is the one Williams marks with a question and the encyclopedia reads as skerries. The note is a narrowing, not a signed confession. It is the dialect parallel on this side of the argument. The other side is Nielsen's charters, Hall, Thalbitzer, and the fact, which Williams states, that no witness has ever proved Ohman cut the stone. Who held the chisel is a separate quarrel, and it is not settled by the language. In 1976 Minnesota History published an interview in which Walter Gran said his father John had told him, in 1927, that Ohman made the inscription. Williams, in 2012, writes that this deathbed story does not hold up, and that no evidence has surfaced that Ohman was the author. The Heritage Board's account leaves three possibilities open. Ohman, someone close to him, or someone who wanted to make fun of him. The Board's sentence is that we may never know. The runes can be nineteenth-century even if the hand is never named.
The geologists do not agree The weathering of the Kensington stone Newton H. Winchell examined the slab for the 1910 committee. His first impression, printed in the report, is that the cutting looks recent, not 548 years old. The edges of the chisel work are sharp. Crushed powder of the stone is nearly white. That powder is gone from the runes on the face, which he says means some years have passed, and it is still sitting in most of the runes on the edge. He did not think the powder, by itself, could give an age. Calcite is the sharper part of his argument. A layer of it covers part of the inscribed face, deposited in a crack before the slab split from its neighbor. Calcite in the weather turns to powder. Winchell writes that he can see no natural way for that skin to survive 548 years in the open. If the slab was split then, it had to lie face down for most of the time. If it was not buried, and the calcite is still intact, the inscription was made less than a hundred years before he saw it, and probably less than thirty. He also notes that the stone spent eight years, from its return in March 1899, as a stepping-stone by Ohman's granary, and he attributes a fresh polish on the calcite edges to that traffic. The face was down when the tree was grubbed. Burial protects a soft mineral. So does a recent date. His report says both. The society's encyclopedia says the 1910 committee led by Winchell concluded that the inscription is genuine. The same encyclopedia says a 2003 analysis by the geologist Scott F. Wolter concluded that the inscriptions are more than 200 years old, and that this remains highly contested. Williams, in 2012, describes the recent geological work as a restatement of Winchell that had not been published as a scientific paper and had no scientific support. The 2003 report itself is in the society's manuscripts. A relative age, older than about 200 years, would sit badly with a carving of the 1880s or 1890s. It would sit comfortably with 1362. It would also sit comfortably with any decade before about 1700. Weathering of this kind is a comparison, not a calendar. Winchell already said a decisive age could not be drawn from the powder. The committee's verdict and the sharpness of the cuts are both in the published record. They are not the same sentence.
A name, a date, or a later hand Heavener On Poteau Mountain above Heavener, in Le Flore County, Oklahoma, a great slab stands in a ravine. Gloria Farley, who grew up in the town and spent decades on the stone, describes it as 12 feet high, 10 feet wide, and 16 inches thick, with eight marks pecked across the west face, growing from about 6½ inches to about 9½. Local people called it Indian Rock. Farley writes that Carl F. Kemmerer found it while hunting in 1913, and that a copy he sent reached the Smithsonian by 1923. The reply she records is that the characters are plainly runic, and that the guess was a carver with a Scandinavian grammar as a guide. She gives signed memories from neighbors. Wilson King and two other hunters, according to his son's statement, saw the carving before the son's birth in 1874. Luther Capps said he saw it in 1898. A sawmill man's daughter remembered being lifted to the letters in 1904. A Choctaw family tradition, as a local historian passed it to her, put a hunting party at the stone in the 1830s. These are testimonies she collected. They are not a dated trench. The readings of the eight marks diverge on purpose. Farley reports Richard Nielsen's transliteration GLOMEDAL, an old name Glome plus dal, valley, so "Glome's Valley," a land claim. He placed that language between about 600 and 900, probably around 750, and he read the Poteau stone as a memorial to the same man under the nickname Gloi. Alf Monge, a Norwegian-born cryptographer, published a different use of the same marks. Farley recounts it. The runes are not letters. They are numbers, taken from each sign's place in two alphabets, the older twenty-four-letter row and the younger sixteen-letter row. Second and eighth from the younger row, the other six from the older, the numbers make 11 November 1012, St. Martin's Day. He said the date could be checked eight ways. Nielsen's word reading and Monge's number reading cannot both be what the carver did. Each one is a full use of the line. A later date has published defenders too. In the Chronicles of Oklahoma, Leslie A. McRill's notes, discussed in a 1966 follow-up, suggest a French boundary post from the land grants along the Arkansas beginning about 1720. The same pages report a 1963 manuscript by Monrad, held at the Oklahoma Historical Society. Monrad kept most of the marks in the older rune row, in use to about 750, and found one mark in a form used among Swedes in the sixteenth to eighteenth centuries, when runes had a learned revival. That one late shape was his reason for a recent carving, and the Chronicles writer points it at a Swedish captain Bossu names among German colonists on the Arkansas in the 1720s. Farley also records the Smithsonian's 1923 guess of a grammar book. She records a check of the 1897 survey. The nearest line ran off the cliff half a mile away, so those surveyors did not cut this slab and walk on. A grammar, a revival alphabet, and a French post are three different laters. They share a refusal of the year 1012 and of the year 750. The rock itself is part of the argument. Farley reports geologists, among them the state geologist W. E. Ham, rating the stone about 7 on the Mohs scale, hard to break with a hammer, and saying pecked lines in that sheltered, vertical face could last more than a thousand years. Ham's account of the slab, as she gives it, is that it fell from the cliff long ago into the north-south position it still has. Kemmerer had checked that face against the north star and thought people had set the stone up. A fall and a raising are both on the page. The park around the slab is public. The inscription's century is not.
Small stones, several alphabets Poteau and Shawnee Farley's chapter is the detailed record of the two smaller Oklahoma stones. On 20 September 1967, after a hard rain, two twelve-year-old boys, Mike Griffeth and Henry McBride, found marks on a ledge at Terry Hill on the edge of Poteau, filled with red clay. They pried the piece off with a crowbar, broke it through the second sign, glued it, scrubbed it, and took it to school. The stone was about 15 inches, with seven characters an inch and a half to two inches high. In December they brought an eighth piece, an X, that fitted the left end. A geologist, Paul Richardson, told Farley the rock is Savanna Formation sandstone of Pennsylvanian age, about 7 on the Mohs scale, and that an inscription in it could last a thousand years. Tool marks showed in the grooves because the cut reached a lighter layer. Monge read the line as another hidden date, 11 November 1017, five years after his Heavener date, and he took one sign as a numeral from an Easter table. Nielsen read words, a memorial "magic to Gloi," and tied it to the Heavener name. The boys' scrubbing and the fresh break are in the same account as the thousand-year hardness. A hard stone keeps a mark. A scrubbed, newly split mark is also what a finder in 1967 hands you. The stone, Farley writes, went to the Kerr Museum. The Shawnee stone is smaller. In August 1969 Jim Estep, with his seven-year-old brother-in-law Mikel Lindsay and another boy, David Sersen, turned over an oval rock in weeds in the northeast of Shawnee, about a mile from the North Canadian River, a tributary of the Arkansas. Farley measures it at 14½ by 11½ by 2 inches, and 14 pounds. Richardson identified red Permian sandstone of that district. Five runes, 1½ to 3 inches high, run about 8 inches across. Farley and the circle around Landsverk transliterated them, all from the older alphabet, as MEDOK. Someone compared that to Madoc, a Welsh prince placed in America in 1170 by a later tradition. Farley herself calls that very improbable. Welsh writers did not use Norse runes of the third century, and Medok is not Madoc. Monge instead read a third calendar puzzle, 24 November 1024. His three Oklahoma dates, 1012, 1017, and 1024, fall within twelve years, which is either a pattern or what a single method returns when you give it similar marks. The top edges of the Shawnee stone, Farley notes, are more worn than the bottom, which would fit a stone that once stood in the ground. The findspot was bulldozed for houses before it could be tied to a grave or to anything else. No metal turned up in the strip they could still search.
A map, a date, and a runic verdict Spirit Pond In 1971 Walter J. Elliott brought three stones to Harold E. Brown at the Bath Marine Museum. The New Yorker, writing in February 1972, places that visit in the previous June, and the stones on the banks of the Morse River at Popham Beach, in Phippsburg, the water people also call Spirit Pond. Two were covered with symbols. The third carried a drawing Brown's circle took for a map of Popham Beach. A local reader, Winthrop Stowell, read a date on the stones as 1011 and a word, "hoop," as an attempt at a Norse word for a bay. Elliott sent the stones to O. G. Landsverk, who wrote back that they were undoubtedly authentic and that a translation would follow. Einar Haugen published the examination the Maine State Museum asked for. The 1974 paper in Visible Language, as its abstract states it, did four things. The runes are not those used in the eleventh century or later. The spelling, and the use of runic numerals, do not match runic and scribal practice. The grammar is not recognizable as Old Norse. Peculiarities in the cutting relate directly to the Kensington inscription. His conclusion is that medieval Scandinavians could not have carved them, and that they probably date from after 1932. A coast, a map-like drawing, and a date of 1011 are the case for a medieval visit. A runologist's finding that the letters, the numbers, and the grammar belong with the Minnesota stone, and after 1932, is the case against. Landsverk's letter is the other published yes. The stones are at the Maine State Museum. This room does not have an open photograph of them.
Undeciphered The Narragansett stone The Rhode Island Attorney General and the Department of Environmental Management announced in 2013 that the Quidnessett Rock, also called the Narragansett Rune Stone, was back in state custody. The Coastal Resources Management Council had reported it taken from the tidal water off Pojac Point, in North Kingstown, between July and August 2012. It was returned on 16 April 2013. The state release calls it a metasandstone of a Rhode Island formation, 7 feet long, 5 feet high, and 2½ feet in the third dimension, inscribed with two rows of symbols, "which some have indicated resemble ancient Runic characters." The state asked the University of Rhode Island to examine it and to try to learn the origin of the marks. No charges were brought. The same release is careful about the earlier history. The Historical Preservation and Heritage Commission was first told of the rock in the 1980s. The commission could not find it in older inventories of the bay. One reason it offers is that as early as 1939 the rock was up on land and may have been buried. Erosion at Pojac Point later left it 20 feet from the extreme low-tide line, so the marks showed only in a short window of the tide. The commission's line, printed by the state, is that the rock's significance as a cultural resource has yet to be resolved, and that it is still worth protecting. The commission also notes other marked rocks on this bay, studied since colonial times, of which Dighton Rock is the famous one. So the case for a Norse text is that the marks come in two rows, that some observers have said they look like runes, and that a 1939 upland position, if it holds, means the rock was not invented for the 1980s. The case against is that nobody has published a reading, the state says the origin is undetermined, and a resemblance is not yet a language. An undeciphered stone can be old, young, runic, or a row of marks that only remind a modern eye of runes. The 2013 papers do not choose.
Lost The Vérendrye tablet This one is a story without a stone. Pierre Gaultier de Varennes, sieur de La Vérendrye, does not mention it in the journal and letters of his western expeditions. A few months before his death in Montreal in December 1749, at a dinner, he told the Swedish naturalist Pehr Kalm. A 1962 study of that conversation, from Kalm's travel writings, is the account. In a country of woods and prairie that Europeans had not lately entered, the party found large stones set in the ground, some of them piled, none of them carved. In one oblong block a small rectangular tablet was set, about a French foot by five inches, cut with characters the Frenchmen did not know. They took it to Montreal. Jesuits there later told Kalm they had seen it and had recognized Tatar writing. Hjalmar Holand, the Minnesota stone's long advocate, argued that a Tatar label could be a misreading of runes, because the scripts share shapes a reader who is not a runologist can confuse. The 1962 article reports that argument and finds the setting, a tablet socketed into a larger block in a rough wall of stones, a poor fit for a Norse monument. Both of those are judgments about a text no one can reopen. No copy of the characters is part of the published story. A lost inscription can be runic, Tatar, or something else, and the only evidence is who said what in 1749. That is enough to keep the tablet in the room. It is not enough to translate.
Dates in the marks, not a measured sunrise Sky readings No azimuth, no horizon height, and no solstice or star alignment of any of these find-spots is published in the sources used here. The sky material that does exist is of two other kinds. At Heavener, Kemmerer checked the slab's west face against the north star and concluded the stone had been set up on a true north-south line. You would observe that by sighting the face at night and seeing whether it runs with the star that does not set. A north-south slab, if people raised it, could be a marker of direction, the way a boundary or a memorial sometimes faces a cardinal line. Farley's report of Ham, the state geologist, gives the other observation. The slab fell from the cliff and landed in that orientation. A fallen block can point north without anyone aiming it. There is no degree in either account, and no claim about a rite tied to Polaris beyond Kemmerer's inference that a true line means a builder. The other sky reading is a calendar hidden in the letters. Monge treated Oklahoma runes as numbers. Each sign's place in the rune row is a numeral. Mix the older and younger rows, he argued, and the line yields a day, a month, and a year. Heavener came out 11 November 1012, St. Martin's Day, a feast already on the medieval church calendar, checkable eight ways by his count. Poteau came out 11 November 1017, using a sign he said was a numeral from the table that finds Easter. Shawnee came out 24 November 1024. You would "observe" this by counting, not by standing on the mountain at dawn. What it would have meant, on his account, is a learned puzzle of the kind he attributed to medieval clerics, a date concealed so that another clerk could recover it. Nielsen's reading says the same marks are words, Glome's valley and a memorial, which uses up the signs and leaves no numbers. A word and a date cannot occupy one string unless the carver meant both, and neither published method has a second cryptographer's independent repeat printed in the sources here. Kensington's 1362 and Stowell's 1011 at Spirit Pond are dates written as dates, in the text, not a code and not a sightline. The Narragansett marks have no published calendar reading. The Vérendrye tablet has no surviving characters to test. The sky in the Newfoundland study is different again. The year 993 is a spike of radiocarbon from cosmic rays, found in a tree ring, and 1021 is a count of rings from that spike to the bark edge. It dates a felling. It does not aim a building at a star.

Location and dating

Location
The pin marks the reported find on the Ohman farm, the southeast quarter of section 14, Solem Township, Douglas County, about three miles northeast of Kensington station, as the 1910 report places it. It is not the museum. The slab itself is on display at the Runestone Museum in Alexandria, Minnesota. Heavener Runestone State Park is on Poteau Mountain in Le Flore County, Oklahoma. The Poteau stone came from Terry Hill at the foot of Cavanal Mountain, and Farley places it at the Kerr Museum with the Shawnee stone, found about a mile from the North Canadian River in Shawnee. Spirit Pond is the Morse River shore at Popham Beach, Phippsburg, Maine, and the stones are at the Maine State Museum. The Narragansett stone was 20 feet from the low-tide line off Pojac Point, North Kingstown, Rhode Island, when it was removed in 2012. The Vérendrye tablet's find-spot is not a modern town in Kalm's account. L'Anse aux Meadows, at the northern tip of Newfoundland, has its own room. The rune question and the camp question are weighed there as well as here. Grooves in the Amambay hills that a 1970s reading called runes are weighed in Amambay rock art.
Period label
Current consensus — open to revision The Newfoundland trees were felled in 1021, on the 2021 count. Monge's Oklahoma dates are 1012, 1017, and 1024. Nielsen's window for those stones is about 600 to 900. The Kensington text says 1362. Monrad's one late rune, and McRill's French post, sit in the sixteenth to eighteenth centuries and around 1720. The Swedish sheets and chalk are dated 1870, 1877, 1883, and 1885. Ohman found the Minnesota stone in 1898. Heavener was reported to the Smithsonian by 1923. Poteau was found in 1967, Shawnee in 1969, Spirit Pond in 1971. The Rhode Island rock was reported to the state in the 1980s, with a possible upland position in 1939, and it was removed in 2012 and returned in 2013.
How this dating was done
For the American stones there is no radiocarbon of the cutting and no dendrochronology of a tree sealed under a rune. The clock is the script and the language, a kind of typology, set beside weathering described by geologists and beside dated runes in Sweden. The Newfoundland year is the other method. A cosmic-ray spike in 993 is found in a ring, and the rings are counted to the bark. That is radiocarbon used as a marker inside a tree-ring count. Monge's dates are a numerical reading of letters, not an alignment. No archaeoastronomy of a sunrise is in the sources used here. Kemmerer's check was the north star, without a published degree.
What could change it
A word, or a rune, that exists in only one century and is present here as a system, not as a single rare pick. Or a geological comparison published in a form another laboratory can run, on this rock, against a scale that does not depend on a gravestone in another climate. For Narragansett, a reading that other epigraphers can repeat. For the Vérendrye tablet, the tablet.
Published ranges, drawn as bars. A bar is a claim about a span, not a point.
Nielsen's window for the Oklahoma stonesContested
Monge's Oklahoma datesContested
Newfoundland felling
Year on the Kensington stoneContested
Haugen's limit for Spirit PondContested
600 CE1250 CE1971 CE
  • Nielsen's window for the Oklahoma stones. Farley's report of his dating, probably around 750, from the older rune row and the names he reads.
  • Monge's Oklahoma dates. 11 November 1012 at Heavener, 11 November 1017 at Poteau, 24 November 1024 at Shawnee, read as numeral codes.
  • Newfoundland felling. Three trees in the 2021 Nature paper, counted from the 993 radiocarbon spike. A different site, and a different kind of evidence.
  • Year on the Kensington stone. The inscription's own date. The argument is whether the cutting is that year.
  • Haugen's limit for Spirit Pond. His 1974 conclusion that the cuttings probably date from after 1932, and before Elliott showed them.

What is actually there

Open questions

Question Status
Is the Kensington year, 1362, the year of the cutting? Open. The inscription says so. The language, the nineteenth-century rune row, and the weathering have all been published on more than one side.
Do the eight Heavener marks spell a name, encode 11 November 1012, or belong to a later hand? Open. Nielsen, Monge, and Monrad each use the line differently.
What do the Narragansett rows say? Open. The state asked for a study in 2013 and did not print a reading.
Is the Spirit Pond drawing a medieval chart or a picture of Popham Beach? Open as a picture. Haugen's 1974 paper places the writing after 1932. Landsverk had said the stones were authentic.

The current narrative

Where publication stands today. It is the best-supported reading, and it is open to revision. It is not a verdict.

The current narrative — open to revision

Cut in the nineteenth and twentieth centuries

The marks were cut by people in North America who already had runes from books, from immigrant memory, or from a nineteenth-century Swedish alphabet. The Kensington text dates itself to 1362. The carving, on this reading, does not. The Oklahoma, Maine, and Rhode Island stones are separate jobs, some of them modern on their own evidence, and none of them is required by the Newfoundland camp.

What supports this

  • Breda and Curme in 1899, and Williams's 2012 balance, place the Minnesota language and runes outside the medieval corpus.
  • The 1883–1885 Larsson sheets, the 1870 and 1877 Hassela notes, and the Haverö ell-stick match the odd Kensington runes and predate the find.
  • Ohman owned an 1891 book that prints aptir, manss, and a 1300 prayer ending in illu.
  • Haugen in 1974 ties Spirit Pond to Kensington and to a date after 1932.
  • Monrad's manuscript finds one Heavener sign in a sixteenth-to-eighteenth-century Swedish form. The Smithsonian reply Farley records guessed a grammar.
  • The Narragansett rows have no published translation. The state's 2013 line is that their origin is unresolved.
  • Parks Canada's artifact list for the 1021 camp does not include a rune.

What pushes back

  • Winchell's committee, as the society's encyclopedia states it, called the Kensington inscription genuine. Wolter's 2003 report argues the cuts are more than 200 years old.
  • Williams himself says a few Kensington words are as hard to explain in the 1890s as others are in the 1360s, and that Ohman's authorship is unproved.
  • Farley's collected memories put eyes on the Heavener marks in the nineteenth century and, by one family tradition, the 1830s.
  • Nielsen reads Heavener and Poteau as names in the older rune row, about 600 to 900.
  • Landsverk called the Maine stones authentic. Stowell read 1011.
  • A 1939 note in the Rhode Island file allows the Narragansett rock an upland, possibly buried, position long before the 1980s.

What would settle it

This is the reading most specialists who publish on the Minnesota language have set down, and Williams states it as a weight of evidence, not a closed file. It does not by itself date Heavener, and it does not identify a carver. The Newfoundland camp stays dated either way.

Emerging narratives

Other published readings. Each may be right. None is adopted here unless the evidence carries it.

An emerging narrative

Medieval travelers, with writing

One or more of the stones was cut by Norse people in North America, in 1362 on the Minnesota text, and earlier in Oklahoma if Nielsen's names or a plain runic reading holds. The Newfoundland camp shows that Norse ships did cross. These stones would be the writing that camp does not have.

What supports this

  • The Kensington text tells a coherent story and dates it.
  • Hall and Thalbitzer published for a fourteenth-century origin. Nielsen's valid charter parallels, Williams allows, make a medieval text harder to rule out than it once looked.
  • The 1910 committee's conclusion, in the encyclopedia's summary, is that the inscription is genuine. Burial under the aspen explains fresh calcite.
  • Heavener's eight marks have a published name reading in the older alphabet, and local testimony of the carving before 1874.
  • A real Norse site at L'Anse aux Meadows, with iron and sod houses, shows ocean crossing was possible centuries before 1362.

What pushes back

  • The Heritage Board and Williams describe the Kensington runes as a nineteenth-century type, now found on dated Swedish objects from the 1870s and 1880s.
  • The discovery word, the missing case endings, and the grammar are the usual grounds for a late date, even after Nielsen's parallels are given their due.
  • No rune is in the published Newfoundland finds, so a camp there does not confirm a slab in Minnesota.
  • Haugen's Spirit Pond paper finds the writing dependent on Kensington and probably post-1932.
  • The Vérendrye characters are lost, and the Jesuits' word for them was Tatar.

What would settle it

This reading uses the one date the Minnesota stone states, and the one Norse site that is independently dated, and it asks the language to be a dialect at the edge of the map. Williams's test is whether that dialect is probable as a whole text. He does not find it so. The Oklahoma name reading is a different century and does not need 1362.

An emerging narrative

A hidden calendar in the Oklahoma marks

The Heavener, Poteau, and Shawnee lines are not words. They are Monge's numeral puzzles for 11 November 1012, 11 November 1017, and 24 November 1024, built by counting each rune's place in two alphabets.

What supports this

  • Farley records Monge's three dates, his eight checks for Heavener, and his use of an Easter-table numeral on the Poteau stone.
  • The three results cluster inside twelve years, which is what he offered as a pattern rather than three accidents.
  • Those dates sit within a few years of the Newfoundland felling in 1021, a little before it and a little after.

What pushes back

  • Nielsen's word readings use the same signs as names and a memorial, so they are not left over to be numbers.
  • Monrad and the Smithsonian reply Farley quotes place at least part of the Heavener line in a much later alphabet.
  • The method is Monge's. A second, independent count that recovers the same three days is not in the sources used here.
  • Kensington already prints 1362 as a year. It does not need this code, and Haugen did not read Spirit Pond as one.

What would settle it

A calendar code is a precise claim, which is why it can be checked, and why a single worker's checks are not yet a second worker's. It stands or falls on the Oklahoma three, not on the Minnesota slab.

Should the current view be reevaluated?

Open to revision

The stones, the find stories, the 1362 text, the 1021 fellings, and the Swedish rune notes of the 1870s and 1880s are all published. So are a committee verdict of genuine, a geologist's argument for more than 200 years of weathering, a runologist's balance toward the nineteenth century, a name reading, a calendar reading, and an undeciphered boulder in Rhode Island. None of those cancels the others by being the louder one. The Newfoundland camp is the dated Norse presence. It is also the site whose published finds do not include a rune. The room stays open because the evidence a reader needs is on both sides of every stone, and because several stones are not the same problem.

What drives this

  • A Minnesota slab found in 1898, dated 1362 in the carving, with sharp cuts, a calcite skin, and an aspen on top of it.
  • Nineteenth-century Swedish runes, on dated notes and an ell-stick, compared to the Kensington alphabet by the Swedish National Heritage Board and by Williams.
  • Heavener readings as Glome's Valley, as 11 November 1012, and as a later rune form, plus two smaller Oklahoma stones.
  • Spirit Pond in 1971, with Landsverk on one side and Haugen's 1974 paper on the other.
  • An undeciphered Narragansett stone, removed in 2012 and returned in 2013, and a lost tablet known from Kalm.
  • Three trees felled at L'Anse aux Meadows in 1021, and no runic inscription in Parks Canada's find list.

This is the museum’s own weighing of the evidence. It is not a visitor vote, and it is separate from the slider.

Weigh it yourself

Anonymous. Nothing here is stored except on this browser. Your lean does not change the assessment beside it.

Cut in the nineteenth and twentieth centuries Unsure Medieval travelers, with writing

If the contested reading is right

What would follow if a reading that is not the consensus turned out to be true, and what could settle it. This explains the stakes. It does not adopt the reading.

The Kensington year is the year of the cutting

Norse travel inland to western Minnesota in 1362 would be a written record, not only a saga and a Newfoundland camp. The language of the slab would be a dialect the handbooks had under-sampled. The Swedish barn runes would still need an explanation, as cousins of the alphabet rather than its source.

Why it is still worth a test

Charter forms and rune rows published side by side with this text, word by word, and a weathering study another laboratory can repeat on this greywacke.

The Oklahoma marks are Monge's November dates, or Nielsen's names

A runic presence in the Arkansas drainage, either around 1012 to 1024 or around 600 to 900, would be a journey the Newfoundland houses do not record. The two readings exclude each other, so only one of them can be the carver's.

Why it is still worth a test

A fresh transliteration, published with photographs, that states which alphabet each sign belongs to before any date or name is built.

The cuttings are all later than the voyages

The stones become evidence for nineteenth- and twentieth-century use of runes in America, including a real folk alphabet from Hälsingland and Medelpad that crossed the Atlantic with emigrants. The Newfoundland date stands, and it stands without these slabs.

Why it is still worth a test

For Narragansett, the study the state requested. For Kensington, the carver is a separate question Williams and the Heritage Board both leave open.

Sources

  1. Kensington Runestone Minnesota Historical Society, MNopedia
  2. The Kensington Rune Stone, report of 1910 Collections of the Minnesota Historical Society
  3. The Kensington Runestone: Fact and Fiction Swedish-American Historical Quarterly, 2012
  4. How the runes went from Hassela to Minnesota Swedish National Heritage Board, 2020
  5. The Dialect of the Kensington Stone Futhark 8, 2017
  6. The Oklahoma Runestones Gloria Farley
  7. Recent data on the Heavener Runestone The Chronicles of Oklahoma
  8. The Rune Stones of Spirit Pond, Maine Visible Language, 1974
  9. U.S. Journal: Maine — Runes The New Yorker, 5 February 1972
  10. Return of the Narragansett Rune Stone Rhode Island Attorney General and Department of Environmental Management, 2013
  11. Autour de la pierre de Kensington Annales de Normandie, 1962
  12. Evidence for European presence in the Americas in AD 1021 Nature, 2021
  13. Culture and history, L'Anse aux Meadows Parks Canada

Image credits

← United States World map