Return to Features Page


Ruminations on the Borderlands of Cartography

© 2007, J. B. Post

The question is often asked “What is a map?” Out here on the Edge, the question might just as well be asked “What is not a map?” Before getting involved in that, let me propose the metaphor of the Map Family Picnic. A metaphor is not an analog and some poetic license has to be granted.

For those who are not familiar with family picnics, they are generally gatherings of individuals who are related by “blood” or marriage, often claiming a common descent from some historic individual. Even here, some problems arise. Let's say the descendants of Harvey Schwartzmier have gathered together and rented a tent or pavillion. Along comes someone with the last name of “Schwartzmier” but who is a direct descendant of Billy Schwartzmier, Harvey's brother. Does he get in? No. A kinship is recognized, but the descendants of Billy are not invited. Now if the family picnic expands back a generation to Jacob Schwartzmier, the father of both Harvey and Billy, then he gets in. But as of now, he is out. Probably never go back that generation because Billy was a ne'er do well whose marriages were socially unacceptable to the family at the time. And there is the problem of Aunt Hilda, a woman universally loathed, but who has to be allowed in because she married Uncle Waldo and is the mother of Schwartzmiers.

With this in mind, let's become gatekeepers at the Map Family Picnic and decide who gets in and who gets left out. There is still the problem of seating inside the tent, but that's a problem for those inside the tent to work out once the attendees start to arrive.

All the Rand McNally, Hammond, General Drafting, Bartholemew, etc. maps get ushered right in. Of course, the great historic cartographic works get in (and probably a special table) without a problem—except here comes a map of Schlaraffenlande arm-in-arm with Seutter's Mappa Geographiae Naturalis. Do maps of imaginary places get in? I think so because they certainly look like maps and are called such. We might want to put them at their own table, but they get in.

Now come a gaggle of cartifacts. These are things with maps as designs so the map part gets in, but how about the whole object? I'd have to say “yes” because the map part is bonded to an object, you might say “married” to it. Damned annoying letting that UPS truck in though. But wait, we can keep the truck out. Wine labels with maps on them are cartifacts and it is the label, not the bottle, which is the cartifact (unless a map is etched into the glass) even if it is convenient to store and display the labels still on the bottles. So, only the panel from the UPS truck is the cartifact with its rendering of a partial globe.

Something more vexing is the cartogram. We want to include the cartogram, but, if we move to Edward Tufte's meta-level of envisioned information,* the pie chart, the bar graph, and the cartogram can be informationally equivalent and are merely three different ways of presenting the same data. We don't want to allow the bar graph into the map family picnic. But if it shows the same stuff, if it is informationally equivalent, then why doesn't it get in? It doesn't look like a map, while the cartogram not only looks like one, it is one. Which brings us back to “What is a map?” and “What isn't a map?” However distorted the cartogram may appear, it still retains a rudimentary sense of spatial relationship among the data presented. The bar graph can arrange the list of, say athlete's foot cases, alphabetically by country, by geographic region, from highest incidence to lowest (or the reverse), the choice being dictated by the emotion one wants to generate in the data presentation. Canada will always be north of the United States in a cartogram, however skewed the areas may be in size or orientation.

* Edward Tufte: Envisioning Information, 1990: Cheshire, CT: Graphics Press

Here we may have the beginnings of the family resemblence. Spatial relations between the parts. In the more traditional maps, this reflects the spatial relationship between parts of the world as represented in the best manner possible considering the times. Maps of imaginary places only reflect the relations among the imagined parts, but spatial relations they are nonetheless. Or can't we speak of spatial relations among imaginary things? I think we can if we recognize the imagined nature. While Slobbovia may be a null class, maps of Slobbovia are not.

Let's say the bar graph really wants to get in and comes back in a different guise. Remember those plates in 19th century geography books and atlases which show the relative heights of mountains and the lengths of rivers? These are essentially bar graphs using pictographic elements for the mountains and cartographic elements for the rivers. Maybe we can dismiss the mountains, but those river presentations sure look like maps. The only relation among the rivers shown is comparative length, something rather abstract, but some of those plates had lakes, tributaries, and such also shown. All drawn out along a line and not showing the twists in the actual river, but they sure look like maps. I'm not sure about these guys: maybe we can sneak them in as cartograms. I'm just not sure.

Now let's say the bar graph has a little outline of the country at the top of each bar. No, this is not a map—but we could think of it as a cartifact.

Here come the globes. Sure, they get in and we don't even have to call them maps on a ball. This includes celestial spheres as well, though no way they represent a surface in the two dimensional sense, they still represent a spatial relationship. Trotting along are the orreries. Often these are called “models” and not considered cartographic items in inventories. Well, is a map a model on some meta-level? Here we play with words and another element of maps is their symbolic nature as well as just scale reduction so we can recognize a relationship without letting all models in. The orreries might end up at a table pushed up against the globes, but that's a problem for those inside the tent.

Another troublesome relative is the topographic or landscape print. These, by and large, are not maps. We can recognize a kinship, so close in some cases we want them in. The medieval city views are welcome as are many panoramic (“bird's-eye”) views of more modern times. We can sneak them in when they have something on them which isn't out in the “real” world such as a name in the middle of the street or a letter floating in the air next to a steeple keyed to text in the margin. It does, however, seem strange to say that the first hundred copies of a panoramic view without any street names are not maps but the second hundred pulled after street names were added to the stone are. This is a strange world.

There are prints and then there are prints. While not exactly sure where they will be seated inside the tent, I would allow the Harvey Hutter (perhaps “Steinberg/Hutter” is a better term) type views inside. However distorted, they do show space and a kind of relationship. Similarly, those strange hybrids—which in the foreground are landscape prints, but in the background fade off not into a distance but into a map—are passed through to bedevile those concerned with seating. Shaker maps, however pictographic, are maps.

While constructed mazes can have maps of their labrynthian ways, there are more drawn mazes than constructed ones. Do we say these are self-referential? I don't know if we should let them in.

In Philadelphia at one time, selected streets and sections had their own banners and symbols. As an instance, the banner for JFK Blvd. had a square with a line running to another square and wavy lines crossing the line between the squares. The squares represented City Hall and the 30th Street Station, the wavy lines represented the Schuylkill River which JFK Blvd. crosses. The line, of course was the street itself. This is so abstracted, I don't think this is a map. Oh, it's certainly symbolic and spatial, but its function is design. On the other hand, if one were to draw the same pattern in the dirt with the intention of showing someone how to get from City Hall to 30th Street Station, then it would become a map.

Now we are getting into the morass of utility, intended function, and subjectivism. The proponents of science fiction and the opponents of pornography say much the same thing: “I may not be able to define or describe it, but I know what it is when I see it.” We need better. If I use something as a map, as a tool for getting around, and you do not, is it a map for me and not for you? On one level, I guess it is. The map the interior decorator frames on the wall is just a pretty print to some, but we will see it as a map. Without sounding mystical and speaking of “essences,” some defining features of a map need to be commonly recognized. Even if we know one when we see it.

“Symbolic representation of spatial data” looks pretty good. But the human mind is constructed to always push the envelope and seek exceptions. That map from The Hunting of the Snark with nothing inside the neat line, has no data, yet it calls itself a map. It has parts in itself (top, bottom, left, right, corners), but it shows nothing other than itself.

Other symbolic representations are out there and we really don't want to have too many of them swell the tent. Diagrams of where to put the feet might be called “dance maps.” Musical notation might be said to be mapping sound. Fingering diagrams for beginning musicians are certainly spacial. Genome mapping shows relationships. Modern molecular models are more sophisticated than the old benzene rings, but those old formulae might be construed as mapping since some spacial relationships were represented symbolically. Diagrams used by accupuncturists, phrenologists, and reflexologists represent bodily space even if ultimately the meaning given to those parts turns out to be wrong.

Let's concede a point and agree that maps and models are related. Maybe maps can validly be viewed as a special case of models, the special case falling within specific parameters. Just as some languages distinguish between “carte” and “plan” with both being clearly maps to us English monolinguists, we can call maps models with specific characteristics and hope that keeps the model people quiet. Musical notation isn't so much mapping sound as depicting sound over time, the axes being frequency/pitch and time. This would be similar to a proposed mapping of football (either sort) where the ball is represented by a line sig-zagging between the top and bottom of a streamer representing the movement between goal posts over time. Music is out, football is in because an object is represented as moving in space.

With molecular mapping, I would have to accept the old benzene rings, but the newer computer graphics can have their own picnic.

Are they all in the tent? Let's see, now, maps are symbolic representations of spacial data and are spacial objects themselves. But, wait: here come the cybermaps. The first question we ask of them is “Is cyberspace spacial?” Certainly not by any standard definition, but we seem always to have trouble with standard definitions out here on the Edge. The cybermap itself is certainly spacial even if it is only a pattern of dots on a screen. What it represents can vary from a simple listing of contents for some “site maps” to a representation of complex relationships using space (often distance as well as size) to indicate them. A colleague in Europe has pointed out that he is closer to me via e-mail, than he is to a friend in the Netherlands with whom he must communicate via the national mails. This can be represented graphically.

This use of space to represent something other than space is what the cartogram does, so the cybermap isn't new in that sense. Does this mean that the cybermap representing relationships is to be admitted into the tent? The first reaction is to say “yes,” but look at the color wheel lurking over there: the color wheel and related representations certainly indicate relationships, but I don't think any of us would call them “maps.” Keeping in mind maps of imaginary places, can we say that the relationships have to be somehow related to space in the physical world? Tempting. But some bar graphs and pie charts can indicate geographic data in their own way.

And what about timelines? Certainly a temporal relationship among events is depicted, but not always a spacial one. We can certainly indicate temporal changes on maps by using color to indicate the rise/decline of political entities or line thickness to indicate the manpower of advancing and retreating armies, but the old traditional timeline of who was whose contemporary or which empires existed a world away from each other at the same time wouldn't get in the tent. But if they were modified to also represent spacial relationships in some manner, we would at least have to think about the matter and examine the manner of that spacial representation.

Oh, No! Here come the animals with blotches on them which sort of look like maps. These always amuse and are great fun, but they don't get in. At least not now. Another element has to be conscious creation and accidental blotches, however amusing, are not conscious creations. But what if there comes a time when we can so control genetic inheritance that we can decide what blotches are on the sides of our pigs and cows? I suspect advertising messages will come first, but map-like things will surely follow. Let's leave that question for the future and look at another animal question. Is the dance of the bees which seems to communicate the direction and distance of flowers to other hive members a map? Personally, I think the dance is better considered as comparable to sailing directions and driving directions. Yes, they all have navigational implications and are geographical in a way, but I don't think we can extend “map” to mean the dance of the bees. We have enough trouble with human minds that we can't begin to speculate on what sort of “mental map” is created in what passes for the mind of a bee by the dance.

And as far as animals with map-like blotches on them, they don't get in the tent as family, but we might consider letting them in as entertainers.

I must confess now to a prejudice against letting the bar graph and pie chart in. Maybe I've been backwards and I should say “I know what isn't a map when I see it.” I want to be as inclusive as possible when dealing with maps without letting in the bar graph et al. There may not be any way to do this other than to put maps with science fiction and pornography in that category of “we know … .” This is a matter for keener minds than mine to ponder. And, once having finally figured just who gets in the tent, the matter of seating inside the tent becomes an issue, the matter of the taxonomy of cartography and how different maps relate to each other, will have to be confronted.