This guest post was kindly contributed by Richard Myers Shelton when a conversation was developing over a previous article. His contribution:
The complement of the classical 7-course labyrinth is highlighted in Erwin’s recent post “The Complementary Classical 7 Circuit Labyrinth” (20 September 2020). The term “complement” is due to Andreas; see his post of 2 July 2017. The complement of a labyrinth visits the courses in reverse order: the complement of the classical labyrinth, for example, traces the courses in the order (5 6 7 4 1 2 3), just the reverse of the standard classical order (3 2 1 4 7 6 5). Photos of a modern example of this design on the banks of the Rhine near Duisburg are featured in Erwin’s post.
My comment on the post pointed out that “complementary classicals” are not unknown historically. The complement of the 15-course classical labyrinth was reported near Borgo (modern Porvoo, Finland, some 50 km east of Helsinki) by Johan Reinhold Aspelin in a letter to the Berlin Society for Anthropology, Ethnology and Prehistory. (The letter is included in the report of the Society for 17 Nov 1877, in the Zeitschrift für Ethnologie, vol. 9, 1877, pp. 439–442.)
This is one of two stone labyrinths near Borgo mentioned in Aspelin’s letter; the other is a straight-forward Baltic-style labyrinth with a central spiral and separate exit path. Both are illustrated in the letter, and are reproduced in Figs 2 and 3 of Nigel Pennick’s “European Troytowns” (http://www.cantab.net/users/michael.behrend/repubs/et/pages/pennick.html). The complementary classical at Borgo is also illustrated in Fig. 125, page 148, of Matthews’s Mazes and Labyrinths.
Figure 1: Aspelin’s drawing of Borgo
But the Borgo complement has a peculiarity, easily visible in Aspelin’s drawing: while it does trace the two meanders in reverse order (inner first, then outer), the path does not lead to the center. Instead it dead-ends in the outer meander: the outermost course 1 is connected to course 7 instead of course 6; course 4 therefore has no escape, forcing it to dead-end. In fact, the labyrinth incorporates a thick wall of rocks, completely isolating the left side from any access to the center. The difference can be seen by comparing the two level charts below: one for the complement of Classical-15, one for Borgo.
Figure 2: Level chart for complementary Classical-15
Note: The entrance is at the bottom left, the center is at the top right (unlike as in the diagrams by Andreas or Erwin). Figure 2 is self-dual.
Figure 3: Level chart for the Borgo labyrinth
In my comments, I characterized the dead-end as a mistake; and Erwin and I exchanged a few emails about whether the mistake was in the labyrinth itself or in Aspelin’s drawing of it. But in fact, several early drawings of labyrinths from Scandinavia show features we commonly think of as mistakes – paths that don’t lead to the center, or branching paths that force the walker to make a choice, or even portions of the path that are completely isolated from the exterior (and sometimes from the center as well).
Another good example of this is Karl Ernst von Baer’s well-known diagram from 1844 of the stone labyrinth on Wier Island in the Gulf of Finland (known today as South Virgin Island). This has a fork in the path, and while one choice leads to the center, the other dead-ends near the perimeter.
Figure 4: von Baer’s drawing of Wier
I doubt that these early researchers were being careless in their diagrams. On the contrary, they were earnestly trying to preserve a rapidly vanishing past from oblivion, carefully recording these objects for posterity. I conclude, therefore, that these anomalies were present in the labyrinths themselves. But we ought not to conclude that what we see as anomalies are mistakes. Instead, it is we who make the “mistake”: namely, of assuming that the people who built these labyrinths intended them to be walked as we walk them today; of assuming that any labyrinth that is not walkable that way must be mistaken.
Some of the odd labyrinths probably were mistakes. A pattern that shows up with some frequency is what you get by drawing a classical labyrinth from a seed pattern that forgets to include the four dots inside the four angles.
And it is clear from various accounts that labyrinths in Germany and England were indeed often meant to be walked or run ceremonially or in contests or games, particularly in association with spring-time celebrations of May Day or Easter or Whitsun – and that this custom appeared later in Scandinavia as well.
(In this regard, it is curious that English and German are the only languages whose word for Easter recalls the pagan goddess Ēostre, as recorded in Bede; other languages refer to the Christian nature of the holiday, or to Passover or the end of Lent.)
But the evidence and the stories from Scandinavia (and further east into Estonia and Russia) hint at a darker purpose: many of these devices were probably intended as traps, perhaps inheriting the idea that led the Romans to place labyrinths near entry-ways to ward away evil.
Christer Westerdahl’s article “The Stone Labyrinths of the North” (Caerdroia 43, 2014) lists several contexts where labyrinths would have been intended this way: near graveyards (to keep the dead in their graves), near ancient burial mounds (to hold back their ancient and possibly non-human inhabitants), near gallows (against the vengeful spirits of executed criminals), along coastlines (against trolls or other bad luck seeking to follow the fishing boats, or even to hold ill winds and currents at bay).
I am particularly struck by examples from Iceland. In “The Labyrinth in Iceland” (Caerdroia 29, 1998) Jeff and Deb Saward tell of their quest to locate all the recorded Icelandic stone labyrinths. They found that only one still survives, at Dritvík on Snæfellsnes. When the Sawards saw it, this labyrinth was heavily overgrown, but someone has since restored it as a typical Baltic-style labyrinth. A drawing by Brynjúlfs Jónssonar from around 1900, however, shows a different plan, with four separate paths, some ending in dead-ends and one completely isolated.
Figure 5: Jónssonar’s diagram for Dritvík, ca. 1900
The Sawards also found three labyrinths carved on old wooden bed-boards preserved at the National Museum of Iceland in Reykjavík. One of these has the 7-course classical design, but two others (NMI 3135 and NMI 5628) share identical plans with isolated paths. In fact, while this shared plan is not the same as the old Dritvík plan (because it has two more courses), it shows the same general arrangement, with a large meander on one side opposite two smaller ones on the other. And in all three cases, the path that dead-ends in the large meander also dead-ends in the center. These features are clearly not haphazard; the same general design principle (the “Icelandic way”?) was at work.
Figure 6: Diagram for NMI 3135
Figure 7: Diagram for NMI 5628
Why include dead-ends or isolated paths at all? The stories seem to indicate that the ordinary classical or Baltic designs were considered effective at slowing trolls down long enough so that a boat could get safely away across the water.
But if the design contained in its very construction the magic of “unwalkable-ness”, it could be even more effective! The design itself becomes imbued with the property of entrapment or imprisonment. In this way, might it not become all the more powerful at holding evil things at bay? It would not just slow them down; it could hold them fast!
To us today this doesn’t seem entirely logical. But sympathetic magic isn’t built strictly on logical analogy alone; our irrational hopes and fears get mixed in as well. Consider the wall in the Borgo labyrinth: The dead-end by itself should have slowed the trolls and ghosts down. Why add that massive wall along the side of the meander?? Logically, this seems entirely superfluous, as the trolls and ghosts can just turn around and retrace their path to get out. But somehow that wall must make the trap seem that much more secure.
— Richard Myers Shelton, 17 December 2020