Home > Essays and Notes > Imagined Spaces

Imagined Spaces

Blog post for MORROIS

by Sean Winslow

When John was in an early stage of this project, we were discussing names. Being a project that arose out of his study of romance literature, he was inclined to look through a dictionary at 'R' words. My personal interest is in the mapping of imaginary space, so I wondered if there might be an appropriate romance location that would be friendly to being an acronym including the words "Mapping," "Places," and "Romance." My favourite set of romances are the Tristan and Isolde stories, and I wondered whether the name of the forest to which they escape might be appropriate. John confirmed that it was the forest of Morrois; "Places" became "Spaces," and the name of the project practically wrote itself: Mapping of Romance Realms and Other Imagined Spaces. The "Imagined" aspect of the name is partially accidental, but spoke to a significant issue in the mapping of romance, which is that we are taking locations referenced in fiction and reifying them cartographically. This presents some challenges of periodization and the ambiguity of historical boundaries, subjects which John has given a good deal of thought to, which I am sure he will treat in the future, here and elsewhere. I have always been interested in fictional and imaginary spaces, and from the time John first described MORROIS to me, I have been intrigued by the challenge that it presents to a cartographic representation of romance, and it is this subject that I am going to begin to explore here.

I am of course not alone in this concern: Umberto Eco and Alberto Manguel have both written books on the subject, the latter filled with maps and illustrations of the places discussed. As I was preparing to write something for this blog, the Library of Congress did a multi-part blog post series (Introduction here) about mapping imaginary space, presenting imaginary or otherworldly places like Paradise and Hell, Tolkien's Middle Earth, Middle Earth and Westeros/Essos, the fantastic realms of children's stories, and maps of places that are "Half-Real, Half-Imaginary." Elsewhere, Ta-Nahesi Coates has written about mapmaking as world-building for his work on the new Black Panther comics, relating his early exposure to maps of imaginary spaces to his childhood playing Dungeons & Dragons and reading its supplements. Indeed, the worlds of Dungeons & Dragons and other role-playing games (more on King Arthur Pendragon, below) have spawned multiple atlases of their fantasy spaces (see, for example, this high-resolution map of Faerün, the default setting for DnD as well as the world of the immensely-popular Baldur's Gate series of computer games). Dwarf Fortress became the first video game exhibited at the Museum of Modern Art, largely on the strength of its visually-interesting ASCII maps of randomly-generated worlds. There are a number of thriving internet communities devoted to mapping fantasy game spaces, such as Reddit's r/mapmaking/, and Google+'s Map-Making in Games. I may have even taken a turn at making maps, myself:

Torchbearer

The author's map, for the Torchbearer game

All of these maps of fantastic and imagined places have one signal advantage that we lack in the mapping of romance realms: the ability of the map-maker to make things up. When making a map of Middle Earth, Tolkien was working with a tabula rasa, but also unlimited authority as the creator (or, to use his term, "sub-creator") of Middle Earth, free to exercise his will upon the shape of the world. That is distinctly less the case for MORROIS, as we are working with a map of a real Earth; investigation is our remit, not invention.

Thoughts about the Taxonomy of Mappableness in Romance Literature

Thinking about coding in the truth of a place by making a decision that locates it on a map requires that we divide the knowable and the mappable from the unknowable or the unmappable, and consider how we might represent the latter categories, and how we might create problems by omitting them. A couple of preliminary divisions suggest themselves: Otherworldly versus Worldly, and Fictional verus Real. Both categories will have ambiguous cases, of course.

  Worldly Ambiguous Otherworldly
Real Tintagel    
Ambiguous Camelot Avalon Heaven / Hell*
Fictional Vortigern's Cave   Faerie

* Obviously, highly contentious whether this is Real or Fictional.

Real Worldly places

Uther and Ygerne, 274–280, from the Prose Merlin
And than the duke counseiled
with his peple and seide, "I have but two castelles that agein the kynge may holde."
But the tweyne wolde he holde as longe as he hadde lyf. And he devised to leve
his wif at Tintagel, and with hir ten knyghtes, for he knewe that castell hadde no
doute of no man; and hymself wente to another castell that was of lesse strengthe,
and it stuffed in the beste wise that he myght, and seide that he myght not his
other londes agein hym diffende.

Tintagel, the seat of the Duke of Tintagel (elsewhere the Duke of Cornwall) and legendary birthplace of King Arthur is, regardless of the historicity of the legend, a real place. It even has recently-excavated Dark Age ruins, confirming the general consensus that it existed as a place in the ambiguous period of history when the Arthurian legends are set. We can put a dot on the map at Tintagel and unproblematically associate references to Tintagel to that spot. A real, known place, it represents the easiest kind of element for us to map.

Real Worldly places with shifting boundaries

Alliterative Morte Arthure, 26–33:
   When that the king Arthur   by conquest had wonnen
Casteles and kingdomes   and countrees many,
And he had covered the crown   of that kith riche
Of all that Uter in erthe   ought in his time:
Argayle and Orkney   and all these oute-iles,
Ireland utterly,   as Ocean runnes,
Scathel Scotland by skill   he skiftes as him likes, 1
And Wales of war   he won at his will

The next level of difficulty are similarly real places, like Scotland and Wales that are real geographic places that can be identified, but which might have unknown boundaries due to the ahistoricity of the Arthurian stories. We can do our best to put shapes into our map representing these places in the period when the text was written, to show what the author was familiar with when using those terms, and the period when the story is set, to show what a historical memory of the period might represent. Given the ahistoricity of legend this is problematic, but we can make decisions about how to represent these places and draw them on a map.

Ambiguous sets of Real Worldly places

Ambiguous sets of real places: the Islands in the Ocean / Sea could refer to Britain's principle islands (Orkney or Anglesey, Man, and Wight), a larger set of islands including Iceland, Gotland, Ireland, Norway, Denmark, and the Orkneys, or literally every island in the Atlantic (and even Mediterranean).
See Luttrell, "Arthurian Geography: The Islands of the Sea." Neophilologus 83:2, 187-196.

Ambiguously-Real Worldly Places

Things start to get really problematic when we try to map Ambiguously-Real Worldly places. Is Camelot a legendary analogue for a real place? The Camelot page of the Robbins Library Camelot Project sensibly offers that "since Camelot is a legendary place, it is perhaps futile to speak of its location," but nevertheless points out that it is identified variously with Winchester and Cadbury. In Malory, Camelot is Winchester, but other cities have claimed this title, even those so far out of the Arthurian heartlands as Colchester (which would have been in Saxon lands). The multiplicity of identification perhaps strengthens the case that Camelot is not, in fact, another name for a real place, but wholly fictional. How to map this? If Camelot is wholly fictional, then it is off the map entirely, which is unsatisfying insofar as it is the capital of realms which are mappable, reachable to them by roads and no special conveyance, and its absence is conspicuous. If one believes that Camelot can be identified with a real city, this is doubly unsatisfying because the identification has no representation on the map. As a way of acknowledging all of the various claims to Camelot, the potential sites can be mapped as a set, perhaps with weights associated to each location, but this is itself unsatisfying because, firstly, Camelot is one, not various, and secondly, because we cannot represent the potential un-realness of Camelot through this set of points on a map.

Thinking about how to map Camelot led me to Gregg Stafford's (the author of the King Arthur Pendragon role-playing game) Pendragon Page, specifically, his page of maps for the game. Pendragon is a table-top roleplaying game taking place in what might be described as a 'pocket timeline,' placed firmly in Dark Age Britain at the start, but becoming increasingly magical and like the late-medieval context of Malory and Chretien (as viewed through the lens of Pre-Raphaelite art and later artistic innovations like Ivanhoe and "The Lady of Shallott"), but then receding in magic and to join our timeline after the death of Arthur. It is a fascinating work of what is popularly called "neo-medievalism" in conference circuits, and I am particularly fond of it for the way that it illustrates the books with the fictional arms of the knights and others who figure in the setting. After all, as I said above, when making a map of a fictional place for a fictional setting, one has the advantage of making things up. In fact, the first edition of Pendragon includes a beautiful map of Britain (warning: 50MB file) within the game's Arthurian setting, where the author has placed many unknown fictional locations neatly on the map. In this first edition map, Camelot is somewhere near Yeovil, presumably Cadbury Castle, one of the two locations the Camelot Project mentions. In an interesting (at least to me) turn of events, this identification changes in later editions of the game, where Camelot, presumably following Malory, is changed to Winchester. Even in a fictional game setting that has great freedom to make arbitrary decisions about locations, Camelot is hard to place.

Pendragon 1stPendragon 4th

Camelot, from the 1st (top) and 4th (bottom) editions of King Arthur Pendragon.

While perhaps the best solution, linking all mentions of Camelot to a description page describing the place and all the places that it is associated with feels like retreating, from the perspective of a mapping project. Insofar as the point of a mapping project is to map, not mapping one of the central locations in romance has a disappointing quality to it.

Fictional Worldly Places

This is a somewhat difficult one, but for this distinction, perhaps we might identify Vortigern's Cave. While arguably Otherworldly, the cave is mappable, insofar as it is part of the Margate Caves, though it is presumably hidden from mundane view. In this sense, it is similar to the more contemporary Platform 9¾, a fictional place which can nevertheless be mapped very precisely. Compare this to the somewhat more ambiguous case of 221B Baker Street, the famous address of Sherlock Holmes: in Arthur Conan Doyle's time, Baker Street did not have numbers as high as 221, so the address was explicitly fictional and un-mappable, but Baker Street has since been extended past 221; though 221B is not an address there, there is a blue plaque, and for many years the occupants took mail for Sherlock Holmes sent to that address.

Fictional Ambiguously-Worldly Places

Is Avalon an analogue for a real island? As a somewhat glib test, I entered Avalon into Google Maps; it gave a great list of businesses, the Avalon Peninsula, and nothing particularly useful if I wanted to find Avalon, the island. Not to be deterred, I tried some other searches. Avalon Island is similarly ambiguous, but entering Island of Avalon, Google was absolutely sure that I was looking for ... an art gallery in Orlando, Florida. My (predictable) problems with Google Maps underscore a basic problem with the interaction of mapping as representation and the fuller range of space and place that exist in literature. Whereas I did in some way hope that some Google engineer had pointed the search for Avalon at Glastonbury Tor (a fraud, but one with the respectability of hundreds of years of history) or inserted an easter egg, an imaginary place has and is no place in a map of the real world.

Fictional Otherworldly Places

Sir Orfeo, 347–356:
In at a roche the levedis rideth,
And he after, and nought abideth.
   When he was in the roche y-go,
Wele thre mile other mo,
He com into a fair cuntray   
As bright so sonne on somers day,
Smothe and plain and al grene -
Hille no dale nas ther non y-sene.
Amidde the lond a castel he sighe,
Riche and real and wonder heighe.

Some places are so explicitly Otherworldly that there is little hope to map them in any way that makes sense. While we might assume that the rock that the ladies ride into is in the vicinity of Winchester, Sir Orfeo's capital (though he has left it for ten years at this point, mind), the "fair country" that he arrives at is explicitly not a place that maps to the county of Hampshire. Faerie, as a realm, links to our world in romance, but it does so in ways that are unknowable and unpredictable to men, and is often experience by those wandering lost in trackless locations (further complicating any idea of mapping them). Even if we could identify these locations where the adventurer came upon faerie or faerie came to them, the realm that these places link to does not follow any rules of mapping or have any known relational geography that we could use to map it.

What to do About Otherworlds

I said above that the problem with otherworlds is that they are not representable on our map of the real world, but perhaps that is not entirely true. I also said that we are bound by the real and do not have the luxury of drawing our own map of fantastic places, but that might also be untrue. Though otherworlds do not correspond in any mappable way with the real world, there are categories of otherworlds that might be useful to sort references to, and we have plenty of empty space in the oceans to draw our own maps, even as simple as blank circles representing Faerie, Heaven, and Hell.

Special thanks to Christopher Berard, an excellent Arthurian specialist, who suggested several sources for me to read while preparing this post. He has not seen an advance copy, and the inevitable errors are mine alone.