Mapping the Unmappable:
Visual Representations of the Internet as Social Constructions
Adam B. King
SLIS-Strategic Leadership in Information Science
Center For Social Informatics
Bloomington, IN 47405
Recently I saw a poster entitled "The San Franciscan's View of The World." This poster showed a fascinating image of the world. San Francisco neighborhoods, especially the most famous, charming and notable ones, swelled to fill three-fourths of this map of the “world.” Nearby cities like Los Angeles and Las Vegas were depicted as lone outposts on the edge of a great abandoned frontier. On the horizon, the East Coast of the United States and the rest of the world were presented as a hair-thin trace on the map, New York being no closer or more relevant than Calcutta.
This poster was based on a famous magazine cover depicting the “New Yorker's view of the world,” and which has been replicated hundreds of times for various cities, states, and regions. The obvious joke and subtle truth of these posters is that the residents of cities see their homes as all-encompassing and overwhelmingly important despite the tiny fraction of the earth's surface they occupy. Such biased representations of the earth's surface might not be a bad thing. If you spend ninety percent of your life in one city, or as is more often the case, in one borough of a city, such a slanted map would be much more practical than the traditional map of the world showing great stretches of uninteresting ocean and inaccessible deserts. Whether a map is useful or accurate depends on your interests: Are you looking for a streetcar stop, or are you sailing around Cape Horn in a thirty-foot sloop?
Deciding what sort of visual representation is necessary and adequate to usefully depict a terrain is the major challenge of cartography. This challenge grows when the terrain being mapped is less stable and less permanent than mountain ranges and rivers. Disagreeing about how to map physical reality is one thing; disagreeing about whether the objects being are mapped representing are physical at all generates a higher level of ontological confusion. This is the conundrum that one faces when one attempts to visually represent the Internet. When called upon to describe the Internet, people seem inclined to draw pictures (create visual representations) of this rather ethereal “thing.” These pictures of the Internet are rarely drawn the same by all people, but vary dramatically in form, style, and content. As such, these pictures provide a multitude of answers to the question "What is the Internet," answers that carry important implications for the future of that explosive “communications medium” (or is it a “community,” a “world,” or something else entirely?).
In this paper, I ask the following question: What are the visual representations
commonly used to portray the Internet, and what choices do different groups
make in constructing these images? To engage that question, I first
present a typology of common visual representations of the Internet. Then
I inspect images constructed for four different purposes by four different
types of organizations, with an eye toward how the choices made in constructing
the images reflect the aims and interests of these organizations. By examining
how organizations attempt to create visual representations of the inherently
abstract Internet phenomenon, we may get some insight into how groups construct
a "material culture" that advances the goals of the group.
At this moment, it may be appropriate to describe the Internet in more detail. I will refrain from doing so for two reasons: For one thing, I imagine that my audience is quite Internet-savvy, and that their time could be spent more productively focusing on other issues than by listening to boilerplate definitions of the Internet. More relevantly, for me to describe the Internet requires me to make choices about what the Internet is, what it is made of, and what metaphor best describes it. These decisions are far from being no-brainers; there is no real consensus on these issues. Proposing a “standard” definition of the Internet now only would only distract from central point of this paper: Definitions of the Internet, and the visual representations that project them, are flexible, negotiated, practical, and impermanent.
Typology of Visions
I'll begin by suggesting a preliminary typology of visual representations of the Internet. The purpose of doing so is twofold: First, given the massive variety in visual representations of the Internet, a sense of the common representation types will keep that variety from overwhelming the discussion. Second, considering the similarities and differences between these representations will begin to show the choices that people make when constructing a visual representation of the Internet. This typology should not be considered to be exhaustive; new forms of visual representations of the Internet continue to be created at a rapid pace. Several online services generate new visual representations of the Internet several times a day. Since this paper focuses more on social organization issues than on being an exhaustive compendium of visual representations of the Internet, the typology should be considered suggestive.
As in any typology (Kahn, 1998; Young, 1996), there are many artifacts that fit categories only roughly, or that fit several categories. At this point it seems more useful to consider why these examples exhibit characteristics of several categories (i.e., why the choice was made to combine types rather than choose between them) than spend time trying to shoehorn examples into one category or another. Also, this paper is most concerned with representations that attempt to portray the Internet realistically, in a way that is "accurate" or "authentic" (i.e., what Michael Lynch (1991) might call "sensible pictures"). As such, this paper doesn't consider purely artistic or emotive representations of the Internet. Still, whether a representation of the Internet is realistic or artistic (or both) is often far from clear.
The highest-level distinction between the visual representations that I have catalogued is whether the representation depicts the Internet as existing in physical reality. That is, the distinction is between pictures with a Physical versus Non-Physical referent. This distinction is key, since the decision to cast the Internet as something that exists with other physical and geographical objects in physical space implies that it essentially like those other objects. Its scaling would then reflect the limitations and constraints of physical reality, as well as its implied permanence. The decision to visualize the Internet so that it does not refer to physical reality, on the other hand, would suggest that it exists outside of or transcends physical reality. So, it would be unbound by the time and space considerations of traditional geography. Michael Curry (1998) suggests that this decision is between depicting the Internet as cyber-Space, an empty physical container modeled after (or perhaps part of) Cartesian, Newtonian physical space, or as cyberPlace, a reality that exist not on some objective coordinate system or grid, but in a non-physical context generated through cultural socialization, the creation and exchange of symbols, cognitive activity, and social interaction.
With this first basic choice, the very nature of the Internet is called into question. Thes resolution of that question a powerful influence on the perceived characteristics and possibilities of the Internet. Visual representations of the Internet that depict it as a physical reality typically draw upon the metaphor of the Internet as a communication or transportation medium (Girardin, 1998). The Internet is portrayed as existing in physical space as set of communication-channels and connection-points. Within the category of representations of the Internet as a Physical object, there are two common ways of depicting the Internet’s "substance": The Network Topology map, and the Network Traffic map.
A network topology map typically portrays the Internet as series of
links and nodes, the links representing phone or other communication lines,
and the nodes representing the separate communication points being joined
together. These network typologies can be very simple, as in this educational
representation of the Internet as a system of tin cans and strings:
More-complex representations of Network Topology include representations of actual computer-machinery as the links and nodes in the network:
The complexity of Network Topologies increases dramatically from these primarily-educational diagrams. For instance, the following is a functional diagram of an Internet-connected subnetwork operated by Nearnet Technologies:
This example demonstrates the common depiction of the Internet or a portion of it as a physical communication network made of computer nodes and physical telecommunications links between them. Implicit in this depiction is the conceptualization of the Internet as a primarily physical system similar to a highway or phone system, used for transporting information from place to place within the "real" physical world. This view stands in contrast to the more ethereal definitions of the Internet and cyberspace implicit in the representations I will discuss later.
When a primarily physical-reality based representation of the Internet is being constructed, it is easy and natural to incorporate other representations of physical reality into visual representations of the Internet. Traditional geographic maps of the world are often used as an integral component of these kinds of Internet representations. Often, these maps serve as backgrounds for Network Topology maps, upon which the network structure is laid-out in geographical space:
Up to this point, the network topology had carried only an implied physicality. By incorporating a geographical map into the visual representation, the network topology is fixed in physical space. Thus the underlying message of a network topology, that the Internet is a communications medium connecting together separate points in physical space by way of telecommunications links, is reinforced by its being overlaid on a the known geography of a region. Once traditional geographical maps are incorporated into Internet representations, the visual drama of the images tends to rise dramatically
These representations, while differing little substantively from the network topology maps that lacked geographic representations, reinforce the Internet's presence in the physical world. The Internet, by vanishing around the horizon, can be seen as the globe-spanning, continent-connecting network that it is reported to be. Interestingly, geographic backgrounds infrequently give much practical information. One cannot go to Kansas City in hope of seeing a gigantic Internet node reaching into the heavens and connecting to lustrous information-lines streaming across the sky. All the same, these representations place the network topology definition of the Internet into a physical context reinforced by a familiar underlying geography.
Geographical maps are excellent backdrops for to the second category of Physical representations of the Internet, the Network Traffic Display. In contrast to network topology maps, network traffic displays recognize the physicality of the Internet but focus on the flows of data across various portions of the network rather than on physical nodes and links. For example, the following visualization, known as the Internet Weather Report, represents the eddies of data swirling around the Internet as sets of circles and points:
Note that in this image, the physicality of the Internet is maintained without focusing on links joining computer nodes together. Instead, it treats the actual flow of data as the essence of the Internet. The physical metaphor for the Internet remains, since these data flows are "real" concentrations of electronic turbulence around the globe. The map in the background firmly places these flows within geographical space. One might say that Network Topologies and Network Traffic maps both represent the Internet as something physically "real," but the former focuses on the physical connections themselves while the second focuses on the stream of electrons flowing over the connections. As in the case of network typologies, when geographic maps are used in conjunction with Network Traffic representations, there is opportunity for considerable visual drama:
Though physical-geography based representations of the Internet are widely used, they are no more common than representations rejecting the notion that the Internet is a physical reality and giving it a geography of its own. This approach is rather imaginative, and necessarily results in a staggering variety of visual representations of the Internet. However, they are united by their tendency to treat the Internet as a non-physical thing, an ethereal terrain made up of symbolic and logical relationships in a limitless cognitive space. In this view, the Internet is not just another telephone, train, or highway network. It is a terrain unlike any that has been mapped previously, and requires different representational tools and techniques.
There are may ways to portray the Internet in these abstract terms,
but for the purposes of this typology, I focus on three major types of
non-Physical visual representations of the Internet: Conceptual, Logico-Spatial,
Conceptual representations of the Internet focus on conceptual distinctions between different parts of or perspectives on the Internet. One common type of Conceptual map portrays the different experience-areas of the Internet on something resembling a Venn-type diagram:
These images portray different portions of the Internet in relation
to one another on a visual field. They imply a certain hierarchy,
with parts of the Internet experience being inclusive of other parts. However,
despite differences in size and shape, these maps do not indicate true
physicality. The scaling is meaningless, and the spatial relationship between
parts of the representations is arbitrary. From this, we can presume
the purpose of this type of conceptual representation: They do not
measure some quantity, nor do they act as guides for navigation. They are
simply diagrams that convey conceptual relationships between different
dimensions of the online experience. Obviously, as there are many ways
of conceptualizing the Internet, Internet mapmakers could imagine and create
an infinite array of conceptual maps.
For instance, a type of conceptual representation often referred to as an "information landscape" uses the implicit scaling and dimensions of the image to represent complex conceptual distinctions:
These maps use size, color, and shape to represent the dominance of certain topics within an online discussion forum. While these images are still conceptual in nature (i.e., the different parts are differentiated only by conceptual distinctions), the underlying model is more complex than the earlier examples of conceptual representations. Looking at these maps, I find it hardly surprising that many refer to the Internet as a "frontier" upon which land may be staked out and defended. The dimensions of these images, providing both a scale and a limited border, reinforce this characterization of the Internet as a large plain of conceptual territory for pioneers to squabble over.
Another type of non-Physical visual representations of the Internet is the Logico-Spatial representation. This category includes representations that portray the Internet as a set of logical connections by placing forms in relation to one another in two or three-dimensional space. The logical conditions that relate these areas of the "information space" and set the characteristics of the nexuses they cluster around, are up to the creator of the map:
Since they are defined by conceptual distinctions, Logico-Spatial representations may seem similar to Conceptual representations. In fact, Logico-Spatial representations are a kind of Conceptual representation. However, they have two important differences from a run of the mill Conceptual map. First, Conceptual maps are generally representations that illustrate cognitive differences between aspects of the online experience. These representations are not useful for navigation. Second, since normal Conceptual maps cannot be used for navigation, their assignment of distance, depth, direction, height, and other visual elements is arbitrary. Logico-Spatial representations, on the other hand, are often used for navigation, and differences in color, distance, and size carry meaningful information to the user.
A final category of non-Physical visual representations of the Internet is the Pseudo-Physical representation. This category blends elements from several of the other categories. Pseudo-Physical representations draw upon the familiar geography of the physical world as an organizing principle. They use this image of physical geography to organize otherwise disconnected, non-spatial online information.
An excellent example of this strategy is Yahoo 3D, an alternative way of accessing Yahoo, Inc.'s organized Web-page information:
This representation depicts various kinds of information available to the user as separate buildings in a virtual town. The user can access information by walking (or perhaps "flying") down the appropriate street and into the relevant building. This approach borrows the familiarity of corporeal geography, with its complex symbolic systems and relationships, and applies it to a mass of information lacking a geography of its own. Rather than generating a new map for the new online terrain, this representation simply uses maps that people are already familiar with.
The next example makes the physical geography metaphor more directly:
This image, appearing on the Cardiff city tourist site, seems like a typical city streetmap. However, this map is “active,” meaning that the locations can be “clicked” to receive more information about the attractions. Given the low level of detail regarding streetnames, sidewalks, parking, and the like, this map is not useful for physical navigation. Instead, this pseudo-physical map uses the underlying physical geography of the city as an index for a set of online information resources. This image is in fact a representation of the Internet, though it does not actually mention or explicitly portray the Internet. In this combination, in which an implied geographical physicality renders the underlying Internet structure invisible, lies the mystery and effectiveness of pseudo-geographical maps.
In Denis Wood's The Power of Maps (1992), Wood argues that maps are fundamentally the result of representational choices about the depiction of reality. Maps and other visual representations require these kinds of choices and limitations to satisfy their purposes. Through the process of selection, a set of idiosyncratic personal knowledge or life-experience of a terrain can be communicated to a larger population of image-consumers, who then can tap into the particularized experience of the visual representation's creator. As we examine these choices, we get a glimpse into the organizations that made them.
Howard Becker (1996) once claimed that maps and other visual representations should not be treated as static objects to be examined independently of their creators. Like ships in bottles, we must not make the assumption that visual representations just "are"; these representations are the products of organizational action. By attending to the links between visual representations and the social organization that produced them, we can get a better handle on how social organizations create and manage specific material objects reflecting their material interests and particular views of reality.
Compendia of visual representations of the Internet typically do not consider issues of impression-management, choice, interest, and organization. In the next section, I will examine a set of images of the Internet to demonstrate the insight that are lost if observers consider visual representations apart from the social organizations that generate and sustain them. I focus on images representing four main interests that seem to be the dominant arenas in which visual representations of the Internet are employed: Education, Business, Politics, and Computer Science.
To the educator, clarity is key. A complex, multi-level display of the Internet may be information-rich, but it is incomprehensible to the beginner. It is often better to start easy, with a simple cognitive model that will give the novice a basic understanding of the Internet . With that simple model in place, definitions of the Internet can be extended and enriched to capture more of its multifaceted nature. A common way to portray the Internet to a novice is to use simple metaphors involving familiar objects. For instance, the tin-can network depiction of the Internet shown earlier would let the novice see the Internet as something familiar, just an extension of the concept of the tin-can telephone played with as a child:
In this image, the conceptualization of the Internet as a communications medium is visualized with commonly-known examples of connectedness and communicability. Other views of the Internet, such as those suggesting cognitive or semiotic organizations of the online landscape, are avoided in favor of something simple that novices people will understand.
While it is understandable that educator would simplify the nature of
the Internet to avoid overwhelming the novice, there are limitations to
this approach. As when any simple model is applied to real-world
phenomena, there are complexities to the online "experience" that cannot
be accurately described by simple educational models. This limitation is
evident in the following image:
In this image, common forms (i.e., the microcomputers) are used to simplify the representation of the Internet. However, as "cyberspace" is pulled into the frame of reference, the simplicity of the image is compromised. The complexity of cyberspace cannot be portrayed by this simple model, so the Internet it is simply drawn as a set of "clouds." Simple models are limited, and these limitations such representational "magic" taking place. As Internet users progress in their education, such simple models can be rejected and more sophisticated understandings can take their place.
In the next example, another limitation of simplified educational examples can be seen: Sometimes, the lure to use familiar images and representations to depict something new can be all too tempting even when such images are unnecessary or inapplicable to the new subject. The image comes from a book that is self-described as a “roadmap” of the Internet (Falk, 1994). Naturally, the temptation to draw upon the familiar image of the geographical map is considerable, and familiarity with map-use should make the task of figuring-out the Internet easier for the novice. Unfortunately, throughout the volume, there are maps of network topologies that seem unhelpful and impractical. For example, here is a picture from the book that depicts Usenet on a worldwide scale.
It is unclear how this representation is supposed to help the reader. The image depicts part of the Internet within a familiar geography, but such a depiction is not appropriate for the kind of practical navigation help that readers of this book are led to expect. Knowing the locations of nodes and links on Usenet isn’t particularly useful, for as the book explains, the entry points into the network are really matters of protocols and communication standards, not geographical locations. Moreover, the density of the Usenet network topology in some regions is so high that in this map, individual nodes and links vanish into a tangled mess of ink. A list of connection-points, a kind of phone-book, would seem to be a better way of organizing this information, and such a list is in fact provided following the map. Still, is undeniable that this map does fulfill the purpose of making the book seem like an old-fashioned highway atlas as it is described in its introduction.
Not surprisingly, computer-science and networking professionals are major producers and users of visual representations of the Internet. Images created by and for computer professionals tend to be among the most elaborate and visually dramatic computer representations of the Internet. A few examples of these images will attest to their spectacular nature:
In this example, a computer professional attempting to advertise Internet security software has created a representation of the Internet based on information gathered by a massive cycle of "ping" commands. Using the results of these commands (which seem to be a common data-gathering technique for Internet image creators), designer assembled a picture depicting the Internet as a series of neuron-type connections. The web-page on which it appears gives no instructions as to how this image may be interpreted, or even that interpretation is necessary or advisable. The image’s main function seems to be to demonstrate the technical prowess of its creator.
This image depicts the "Mbone" Internet communications backbone as a constellation of network connections floating over the surface of the United States. The image has a visual dazzle and drama similar to the previous example, and has no clearer practical purpose. Indeed, this image utilizes what appears to be a meaningless third-dimension. It seems to be primarily for visual impact, rather than for a substantive distinction (although may imply an implicit hierarchy between the Internet "backbones" and the slower, more-local Internet links). Again, the image’s purpose seems to be to indicate the programmer’s skill or hype the Internet than to help with navigation or instruction.
One category of Internet maps that clearly reflects strong organizational
influences aref political Internet maps. In the following two images,
note that state and/or national boundaries are clearly depicted, reinforcing
the embedded governing systems whose effects are being demonstrated:
In these images, the aspect of the Internet being represented is the relative frequency and type of Internet service available in the countries of the world. The divisions shown between land-areas are traditional national borders, thus giving the impression that the main differences between concentrations of Internet use are between nations. Left out in this nation-based orientation is the differential Internet exposure between different groups within countries. It may be that Internet access is concentrated among wealthy, well-educated elites in all countries. As such, wealthy elites who enjoy Internet access in various countries may be more like one another than they are like the poor in their own countries. The national-difference perspective assumed in these pictures overlook these issues, which clash with the general argument of international inequality that is being made. Again, the point is not that these maps are “wrong”; they are successful because they make their point effectively through their representational choices. The point is that if the maps were intended make a slightly different political argument (that the elites of countries have all of the Internet access, for instance), different choices would have been made in the construction of the images.
One of the most common politically-oriented representations of the Internet is the depiction of the Internet as an "information superhighway." This depiction of the Internet carries interesting metaphorical ties between the federal highway system and the Internet, including the assumption that the government can and should be able to regulate the Internet. It also brings up the historical curiosity that both the federal highway system and the Internet were originally created for the Department of Defense. Many visual representations of the Internet seem to reflect the highway metaphor, such as the following image:
This map bears a striking resemblance to an automotive map, except that roads are replaced by network links. Like other representations of the Internet, the information superhighway metaphor, and representations based upon it, influences the way people think of what the Internet is, and what it should be in the future. This metaphor has definite limitations. Harmeet Sawney (1994) argues that this depiction of telecommunications networks causes users to become obsessed with traditional transportation issues, like unit-cost-per-distance rates and the speed and total capacity of the network. The metaphor is natural; it makes sense to use familiar objects to depict the unfamiliar. Yet, clinging to a dominant metaphor blunts the search for other, more-useful visions. The transportation-loaded metaphor limits users from considering the Internet as a semiotic or cognitive construct rather than a transportation medium.
A final arena in which visual representations of the Internet are employed is in business. As the Internet has grown, so has its profit potential. Two business advertisements, both network topologies, make interesting uses of visual geography in their attempts to attract customers.
In this example, an advertisement for an American national Internet service provider, physical geography has been eliminated. However, the topology, which otherwise simply floats in space, makes extensive use of geographical place names from all over the world. The image casts the Internet as a series of connection chains upon which cities from around the world find their place. Within this web of connections, the Internet service provider itself occupies the center position. The implication seems to be that this Internet service provider is centrally-located within the worldwide network that is the Internet (a conclusion reinforced by the written claim "HostAmerica Spans the Globe!").
Strangely, the geographical place names do not seem to correspond to any normal geography in regards to position or distance. Germany is shown as being close to France, as it would be in the physical world. However, Israel, Holland, Norway, and Turkey are positioned together along one connection chain, which makes little geographical sense. It is unclear why the United States of America is wedged between Peru and Panama, or why Canada is neighbors with Thailand and Saudi Arabia. The advertisement adopts geographical references seemingly to emphasize the extent of the network that this Internet service provider can reach. The actual geographical location of the service provider is not given, just an icon denoting its centrality. The goal of the company, to advertise its “access-centrality on a vast network,” is achieved through a cautious and arbitrary use of geographical place-names on an otherwise non-geographical network topology.
This image is also an advertisement for an American Internet service provider, but it is for a local provider rather than a national one. Comparing the two advertisements will highlight some interesting differences between the two.
Except for some meaningless place references, geography is largely ignored in the first advertisement. The second advertisement, though, emphasizes physical geography in its depiction of network topology. The image of the globe fills the picture, seeming to swell beyond its boundaries. Across this picture of the globe, network links (depicted with the now-traditional curving lines) sweep around from the horizon and descend upon a central point in southern Indiana. Only links terminating in southern Indiana are depicted in the image. The implication, as in the previous image, is that this service provider occupies a central position in the Internet network topology. However, since this advertisement is a local provider rather than a national provider, the geographical location of the service provider is emphasized rather than ignored. That two different types of advertisements (local versus national) would use two different representations (a globe image versus a set place-names on a network topology) to send a similar message (that an Internet service provider is centrally located on the Internet) supports one of the main arguments of this paper. The choices made in the construction of visual representations of the Internet are neither routine nor mechanical choices. There are many ways to portray the vastness of the Internet, and to portray one’s place within that vastness. The choices between the many options reflect the purposes and interests of the specific groups that create the images.
Ultimately, the choice of how to visually-represent the Internet is a far from simple question. Implicit in any image are underlying metaphors regarding what the Internet is, and more importantly, what it could be. None of these conceptions can really be said to be "wrong." These representations and the metaphors they reflect can be seen as explorations of what the Internet is and can be "at its essence." Ultimately, the exploration phase could end, as various representations are tried and rejected and a few dominant representations and metaphors are legitimated (Sawhney, 1994). This negotiation over metaphors is a negotiation over visions of how the Internet will be financed, governed, regulated, and experienced. One must study visual representations of reality, because as those representations are adopted and legitimated, those representations become the reality.
Curry, M. (1998). Cyberspace and cyberplaces: Rethinking the identity of the individual and the place, http://dpub36.pub.sbg.ac.at/ectp/CURRY_P.HTM .
Falk, B. (1994). The Internet Roadmap. San Francisco: Sybex.Girardin,
L. (1998). Mapping the Virtual Geography of the World-Wide Web, Web -
downloaded 1998 (http://heiwww.unige/ch/girardin/cgv/www5/index.html, ).
Kahn, P. (1998). Mapping Web Sites: What is a Map? : Dynamic Diagrams.
Lynch, M. (1991). Pictures of Nothing? Visual Construals in Social Theory. Sociological Theory, 9(1), 1-19.
Sawhney, H. (1994). Information Superhighway: Metaphors as Midwives. .
Staple, G. (1995). Notes on Mapping the Net : From Tribal Space to Corporate Space. Telegeography.
Wood, D. (1992). The Power of Maps. New York: Guilford Press.
Young, P. (1996). Three Dimensional Information Visualization. Computer
Science Technical Report(12).