Sovereignty, Territory, and the Mapping of Mobility: A View from the Outside

Abstract

Theorists within and beyond the discipline of geography increasingly realize that boundaries are not simply lines that enclose and define territories. Boundaries also regulate and are reproduced by acts of movement. Movement, beyond and across, as well as within a bounded territory, serves to reproduce the territory that is being bounded. It follows that to understand the history of a territorial entity one must go beyond tracing the spatially fixed activities that occur within that territory or the discursive strategies through which the territory is made to appear natural. One must also trace the acts of movement that occur within, across, and outside the territory's boundaries and the designation of specific spaces of movement as beyond territorial control. In short, one cannot understand the construction of “inside” space as a series of territories of fixity, society, modernization, and development without simultaneously understanding the construction of “outside” space as an arena of mobility that is deemed unsuitable for territorial control. In this article, this perspective is applied to the preeminent normative territory of modernity—the sovereign state—and attention is directed specifically to the designation of the world-ocean as a space of mobility outside the boundaries of the state-society units that purportedly constitute the modern world. Through an analysis of representations of marine space on 591 world maps printed in Europe and the Americas between 1501 and 1800, this article traces the construction of the ocean as an external space of mobility, antithetical to the norm of the territorial state that also was emerging during this era.

Los teóricos dentro y fuera de la disciplina geográfica cada vez más se percatan de que los límites no son simplemente líneas que encierran y definen territorios. Los límites también regulan los actos de movimiento y son reproducidos por éste. El movimiento, más allá y al través de un territorio demarcado, lo mismo que dentro del mismo, sirve para reproducir el territorio así delimitado. De hecho, para entender la historia de una entidad territorial se debe ir más allá de la simple indagación de las actividades fijadas espacialmente que ocurren dentro de ese territorio, o de las estrategias del discurso mediante el cual se quiere dar apariencia natural al territorio. También se deben rastrear los actos de movimiento que ocurren dentro, a través y por fuera de los límites del territorio, y la designación de espacios específicos de movimiento fuera del control territorial. En pocas palabras, uno no puede entender la construcción del espacio “interno” como una serie de de territorios de fijeza, de sociedad, modernización y de desarrollo, sin comprender de manera simultánea la construcción del espacio “externo” como un escenario de movilidad que se considera inadecuado para control territorial. Esta perspectiva se aplica en este artículo al preeminente territorio normativo de la modernidad—el estado soberano—y la atención es orientada específicamente al la designación del océano mundial como un espacio de movilidad situado fuera de los límites de las unidades estado-sociedad que supuestamente constituyen el mundo moderno. Mediante un análisis de las representaciones del espacio marítimo en 591 mapas del mundo, impresos en Europa y las Américas entre 1501 y 1800, este artículo traza la construcción del océano como un espacio exterior de movilidad, antitético de la norma del estado territorial que también estaba emergiendo durante esta época.

Key Words:

• history of cartography
• mobility
• ocean
• state territoriality

关键词:

• 制图学历史
• 流动性
• 海洋
• 国属地

Palabras clave:

• historia de la cartografía
• movilidad
• océano
• territorialidad del estado

The function of a boundary is to produce and regulate a distinction between inside and outside; the movement of things across a boundary signals not its failure but its success.

—Rubenstein (2001, 289)

When borders gain a paradoxical centrality, margins, edges, and lines of communication emerge as complex maps and histories.

—Clifford (1997, 7)

Is it not odd that ours, the most nomadic and migratory of cultures, should found its polity, its psychology, its ethics and even its poetics on the antithesis of movement: on the rhetoric of foundations, continuity, genealogy, stasis?

—Carter (1996, 2–3)

We are the only people in the world who use seafaring instruments to determine our position on the ground.

—Bohannan (1964, 175)

This article is an attempt to explain an apparent paradox of modern territoriality. On the one hand, the concept of territory has a foundational role in the structure of modern society. According to conventional narratives of state territoriality, inhabitable land is divided into autonomous, sovereign units that ostensibly form the fundamental basis for one's citizenship and identity. Each of these territories, in turn, is controlled by a government that secures its boundaries, plots the land within, organizes it, and coordinates the production of value from its human and natural resources (Gottman 1973). Geographers who have studied the history of territoriality stress that this equation of social space (and individual identity) with bounded territory is a distinctly “modern” conception (Soja 1971; Sack 1986).

On the other hand, even as the territorial state was becoming the normative social institution, these territorially defined social units were turning away from spatially fixed, terrestrial homeplaces. Although scholars differ over precisely when the modern system of sovereign, territorial polities came into being, most would agree that the process was roughly concurrent with the age of exploration and the era of merchant capitalism that followed (ca. 1500–1800), an era that was characterized not just by intensified territorialization, but also by a revolution in transportation (Vance 1990; Hugill 1993). The “discovery” of new lands to bound and govern was made possible only by the “discovery” of the ocean as a space of connection and an arena of mobility (Parry 1974; Biagini and Hoyle 1999; Bender 2006). Indeed, during this era, Europe's prevailing global image changed from one in which the world was viewed as an integrated landmass of discrete places, surrounded by a practically insignificant bordering ocean, to one in which boundable space was pushed to the margins as the world came to be perceived as dominated by a world-ocean (or oceans) punctuated by island-continents (Cosgrove 2001). In other words, even as the concept of boundable space (territory) was becoming essential for political organization, maritime spaces that resisted bounding were moving to the center of modern cosmology. Similarly, especially during the first half of this era, projections of power onto space focused on claiming rights of access to routes rather than direct control over swaths of overseas territory (Brotton 1998; Mancke 1999; Steinberg 1999a, 2001; Gillis 2007).

Linked with this historical paradox regarding the formation of the state system—that the territorial state emerged concurrent with the deterritorialization of political economy and the geographical imagination—are the paradoxes about the processes of territoriality in the contemporary world that are expressed in the quotations that began this article: We live in a world in which we affirm borders by crossing them (Clifford 1997; Rubenstein 2001); we construct myths of stable societies defined by bounded territories by living in a world of movement outside those territories (Carter 1996); and we establish the locational grounding for our lives in those territories through techniques established for moving across the putatively placeless sea (which, in turn, make reference to another abstract space: the sky; Bohannan 1964). Each of these statements suggests the frailty of the binary oppositions embedded in the sociospatial logic of the sovereign, territorial state: oppositions between inside and outside; between unit and system; between land and sea; between fixity and movement; and between experienced place and relative, abstract space.

In an attempt to think through these parallel tendencies toward territorialization and deterritorialization, this article is aligned with an ever-growing body of work, across many disciplines, that attempts to delink this modern equation between the unitary citizen-subject and the sovereign space of the state-society by examining postmodern subjects who inhabit positions “between” worlds. ¹ Much of this literature, which roughly can be grouped under the rubric of critical borderland and migration studies, engages poststructuralist-inspired geographic scholarship that seeks to disrupt naturalized and historically stable conceptions of being, place, stasis, and locatedness in favor of ontologies of becoming, connection, mobility, and betweenness.

Whereas some early works in this genre tended toward what Cresswell (2006) critiques as a celebratory “nomadic metaphysics,” more sophisticated, empirically based studies generally have emphasized that individuals’ practices of mobility vary and are not necessarily liberatory; the experience of the jetsetter is very different from that of the refugee (K. Mitchell 1997; Ong 1998). Furthermore, these scholars assert that one must draw a distinction between individuals who engage in cultural mobility (e.g., postcolonial individuals who occupy identity positions “between” worlds) and those who engage in geographical mobility (e.g., refugees, migrants, members of diasporas, or border crossers who physically move between worlds or occupy borderland locations). Not all physical acts of movement involve the transgression of identity categories, and, conversely, hybrid identities can emerge in place. Indeed, it seems likely that the conflation of these two kinds of mobility facilitates short-sighted celebrations of the “nomad” as an inherently liberatory figure. ²

One reason these two forms of border crossing are so often conflated is that we see mobility against the normative portrayal of the world as timelessly and naturally divided into bordered units: sovereign, territorial states that define the scope of governance and nations. From such a perspective, any act of border crossing is inherently disruptive. To achieve distance from this normative worldview, the aforementioned scholars of critical borderland and migration studies typically focus on individuals and communities that break the mold by living in, between, or across borders. Although I am broadly sympathetic with this literature, in this article I take a different approach: I argue that to place border crossing in context we need to first rethink the assumed position of the sovereign, territorial state as the fundamental sociogeographic entity the borders of which are being crossed. To do this, I seek to trace the history by which the sovereign state, as the idealized negation of mobility, has been represented as the universal geopolitical “reality.”

Tracing the history of the territorial state is hardly an original project: The genesis of the modern, sovereign, territorial state has been well documented, as reviewed later. The emergence of the modern state, however, is almost always told as a history of constructing insides: of drawing boundary lines and governing the space within. Such a story inadvertently reproduces the modern notion of the bounded, homogenous, state-society unit that is regularly critiqued by borderland and migration scholars, even as it denaturalizes these sociospatial formations’ histories. In contrast, in this article I seek to narrate the origin of the modern state ideal by tracing the construction of its outsides. This project has been taken up by many in the critical geopolitics movement (e.g., Ó Tuathail 1996; Campbell 1998; Sharp 2001), but for these scholars the outside against which the discourse of the state as a naturally bounded entity is juxtaposed typically is the space of other peoples (whom statist discourses often represent as less civilized) or other states (which frequently are represented as naturally antagonistic). These studies go a long way toward explaining how an individual state's ideology is produced, but they fail to interrogate how the normative discourse of the generic, idealized state as an internal space with governance, unity, stasis, fixity, and society—the state idea (Abrams 1988; T. Mitchell 1999; Painter 2006)—is buttressed by its juxtaposition against an even more fundamental outside: an asocial world of disorder beyond the state system. Echoing Schmitt's (2006a) argument that the power of the individual sovereign is rooted in the sovereign's ability to exclude individuals from its protection (see also Agamben 1998), I suggest that the rise of the concept of sovereignty historically was interwoven with the designation of certain spaces as beyond the sovereign state's organizational limits. In this article, following a further elaboration of this perspective on state territoriality, sovereignty, and mobility, I sketch out this alternative history of the modern state by tracing changes in cartographic representations of the state system's paradigmatic outside—the world-ocean—during the era when the territorial state was becoming formalized as a European, and subsequently global, norm.

Territoriality and Sovereignty: Looking In and Looking Out

To analyze the role of movement across and beyond borders in the ongoing construction of the modern state system, one first must discuss the core principles of that system and its fundamental unit: the sovereign, territorial state. Although the concepts of sovereignty and territoriality frequently are used interchangeably in discussions of the modern polity, the two terms in fact reflect different organizing principles that can exist independent of each other. Territoriality (when applied to the modern state) refers to the manner in which the definition of a society's geographic limits, the organization of its processes, and the control of its people are exercised through claims of authority over bounded swaths of land. A territorial state thus can exist in isolation; it is essentially an inward-looking entity. On the outside of a territorial state might be a world of other inward-looking states; a world of nonequivalent political (territorial or nonterritorial) entities over which the ruler of the initial state might have some influence; an ungoverned “wild”; or absolutely nothing (in the event that one state manages to encompass the entire planet within its territory; Kratochwil 1986). Thus, at the abstract level, the concept of state territoriality can exist independent of a state system.

By contrast, although sovereignty has existed ever since humans have asserted absolute authority within their realm, the modern state tempers this claim to power with the recognition that it exists within a mutually exclusive system wherein other sovereigns have power within their respective, formally equivalent entities. Thus, the modern institution of sovereignty is inherently a systemic condition in that it is impossible to have a world of just one sovereign unit. Sovereignty can persist only when multiple sovereigns recognize each other's equivalency (the outward-looking aspect of sovereignty) as well as when the sovereign's subjects recognize the sovereign's authority (the inward-looking aspect of sovereignty; Taylor 1994).

Sovereignty is not necessarily defined and delimited territorially (one can think, for instance, of religions and political parties that recognize a world of equivalent, competing religions or parties, not divided territorially, but that countenance membership in only one religion or party). As Keohane states, “Rather than connoting the exercise of supremacy within a given territory, sovereignty provides the state with a legal grip on an aspect of a transnational process …Sovereignty is less a territorially defined barrier than a bargaining resource for a politics characterized by complex transnational networks (Keohane 2002, 74, emphasis in original). Keohane's assertion implies that the formation of the modern state was a strategic innovation that occurred when institutions that already had been using territoriality for political power—itself a strategy with a very long history dating to the origins of agriculture (Sack 1986) or the first premodern state (Mann 1984)—came to adopt the inward-looking and outward-looking aspects of sovereignty.

Although the association of the modern, territorial, sovereign state with the 1648 Treaties of Westphalia is now widely recognized as a historically inaccurate oversimplification (Krasner 1993; MacRae 2005), there is little consensus among political theorists and historians about when precisely the modern state did emerge. In large part, these disagreements stem from differing emphases regarding the properties that are essential components of the modern state. The earliest dates tend to be from those who stress the functional-administrative (or inward-looking) aspect of state sovereignty. Strayer, for instance, writes that in Europe as early as 1100 there were signs of “political units persisting in time and fixed in space, the development of permanent, impersonal institutions, agreement on the need for an authority which can give final judgments, and acceptance of the idea that this authority should receive the basic loyalty of its subjects” (Strayer 1970, 10). He writes that these ideas and institutions had consolidated, at least in England and France, by 1300. These relatively early dates are shared by others who similarly deemphasize the territorial nature of the modern state or its existence within a system of sovereign states and instead associate sovereignty with the functional integration of a nation, unified by a singular authority (Armstrong 1982; Hinsley 1986).

Others, who focus on the modern state less as a functional-administrative unit than as a territorial unit, date its origins somewhat later. A number of scholars who have studied national boundary lines on maps prior to the mid-sixteenth century assert that, instead of defining the geographic limits of the state, these lines merely served as imprecise indicators of zones of influence, to the extent that they signified political power at all. This meaning was exemplified by the lack of a hierarchical differentiation between the lines separating national units from those separating provinces and other subnational territories. Beginning in the mid-sixteenth century, however, borderlines began to take on a new level of meaning on maps as they strove to accurately represent national boundaries on land. By the eighteenth century, state boundary lines typically were being highlighted to differentiate them from lesser lines of division, as the domain of state authority and its very identity came to be defined by its territorial limits (Sahlins 1989; Akerman 1995; Black 1997; Brotton 1998; Biggs 1999; Kamen 2000; B. Anderson 2006), ³ although some argue that the true integration of the state as a territorial unit did not occur until the mid-nineteenth century (Maier 2000).

A third group of scholars emphasizes the systemic (or outward-looking) aspect of the modern sovereign state. Tilly (1975), for instance, asserts that the proto-state formations identified by Strayer could have developed in any of four directions: a political federation (as in the Holy Roman Empire), a theocratic federation (as in the Papacy), an intensive trading network of locality-based polities without any large-scale central organization (as in the world of medieval city-states), or a continuation of the feudal organization of weak kings and scattered fiefdoms with complex relations of fealty to the king that predominated in most of Europe in the thirteenth century. In the end, due to a series of contingent factors that prevailed in the fifteenth and sixteenth centuries, Europe went in a fifth direction, characterized by a series of formally equivalent, territorially defined, mutually exclusive sovereignties characterized by “territorial consolidation, centralization, differentiation of the instruments of government from other sorts of organization, and monopolization (plus concentration) of the means of coercion” (Tilly 1975, 27). As with scholars who stress the territorial nature of the modern state, those who equate the modern state with the rise of a system of mutually exclusive sovereignties with centralized powers tend to date the modern state's origin to somewhere between 1500 and 1800 (Ruggie 1983; Krasner 1993; Spruyt 1994; M. Anderson 1996), although there are those who contend that this system did not consolidate until the mid-nineteenth century (Thomson 1994).

Territoriality and the Organization of Space

This brief review of the literature on the origins of the state reveals that the modern state unites the inward-looking aspect of sovereignty (as occurs when a nation is unified by a single, recognized sovereign), territoriality (as occurs when the bounding and control of space is used to define the scope of a sovereign's rule and the identity of the sovereign's subjects), and the outward-looking aspect of sovereignty (as occurs when sovereigns recognize each other as rulers of mutually exclusive, functionally equivalent domains in which no nonstate actor has similar standing). When these three aspects of the modern state are combined, it becomes apparent that territorial state sovereignty is associated not merely with sovereign control of the territory within its borders but also with the overriding spatial logic that characterizes modern society and that constructs space as a patchwork of mutually exclusive, bounded entities (Ruggie 1993). In short, territoriality encompasses not just the bounding of space but also its organization, within, across, and outside state borders.

A number of scholars have developed this theme, generally building on Foucault's (2007) concept of governmentality and Lefebvre's (1991) and de Certeau's (1991) emphasis on the role that abstraction plays in the spatial practices of those who seek to decenter the places of everyday life and impart a hierarchical, functionalist socio-spatial logic. These scholars stress that the attitude toward space reflected in the process of bounding national territories is reproduced in internal territorial organization, and that this internal territorial organization constructs not just the state but the very space on which it is situated and through which it obtains its identity. The rationalization and abstraction of space embodied in the act of territorial bounding permits a discourse premised on control and management, which in turn both reflects and enables the exercise of power by those who, through control of territory, organize the social processes that transpire within. The Cartesian grid is constructed as a backdrop of abstract space on which is located physical matter with objective properties of extension (dimension, figure, weight, motion, and—most important—position). Due to this conceptual separation between the objective material world (res extensa) and the subjective sensed world (res cogitans), the difference and distance between points in res extensa can be calculated and, once space has been reduced to a mathematical abstraction, it follows that it can be bounded into territories and managed (Elden 2006).

As Sack (1986) notes, at an advanced stage of territoriality, space is viewed so abstractly that it becomes “conceptually emptiable” (and fillable) by an authority looking down from above. Combining Sack's image with Giddens's (1985) depiction of the state as a power container, the territorial state can be conceived of as a container that can be emptied and filled, and the contents of which can be rearranged across its points in space. From this perspective, places are perceived of as distinct, but relative, points on the global grid that achieve social character as groups of people transform their local nature. Eventually, these societies evolve into states that are “filled in” and “developed” through the location of production activities at fixed points within these state-areas (Bohannan 1964; Soja 1971, 1980; Burch 1994; Steinberg and McDowell 2003).

A key concept here, as stressed by Sahlins (1989), is that the rise of the territorial state is characterized not simply by the construction of its bounded territory as a homogenous administrative zone (as Sahlins charges that Alliès [1980] and Gottman [1973] emphasize), but that the territorial state constructs its space as a differentiated set of points that are amenable to being plotted (and thus manipulated and rationalized) against an abstract spatial grid (see also T. Mitchell 1991). Historically, the development of technologies and institutions for performing cadastral mapping and land surveying stand out as mechanisms through which the state has achieved a “bird's-eye” view over territory as a means toward achieving social control over people (Bohannan 1964; Kain and Baigent 1992; Vandergeest and Peluso 1995; Edney 1997; Scott 1998; Biggs 1999). Thus, Biggs (1999) locates the origins of the territorial state in national surveying efforts of the seventeenth century. As surveyors mapped royal domains, they graphically noted each village's affiliation. What had been thought of as a personal relationship came to be expressed as a territorial relationship, and, at the same time, surveyors imposed a grid of abstract space over the domain to facilitate mapping. Eventually, these two phenomena associated with surveying converged in an example of what Pickles (2003) calls “overcoding”: The abstract, geometric space of the map came to define the lived-in space of the state, and the territorial relationship between land and sovereign came to be seen as predicating the personal relationship between individual and sovereign. Other scholars have further illustrated the relationship between the way that we hierarchically map space and the way that we hierarchically organize social relationships. Knowledge of space is a crucial tool for control, and the technologies of mapping (and the underlying assumptions about society and space that enable modern mapping), joined with hierarchical systems for drawing lines and assigning names, play a crucial role in constructing instruments of sovereign domination (Akerman 1984, 1995; Carter 1987; Buisseret 1992; Ryan 1996; Edney 1997; Brotton 1998; Burnett 2000; Craib 2000; Häkli 2001; Harley 2001; Pickles 2003; Jacob 2006). ⁴

In short, these scholars emphasize the key role that the ordering of space plays in the construction of state territoriality. This is an important advance over a perspective that simply looks at the bounding of space, but, as Strandsbjerg (2008) notes, these scholars of the cartographic origins of modern state sovereignty still tend to analyze the state as an isolated unit. Given that the modern institution of sovereignty necessarily exists within a system of sovereign units, a study of the modern state (or a story of its origins) that works only from the perspective of the inward-looking aspect of sovereignty cannot be complete. As Taylor (1995) asserts, the starting point of political geography (and political history) must be a theory of the states rather than a theory of the state. In a similar vein, geographies and histories of territoriality (or the ordering of space) must examine not just the “emptiable” and “fillable” space constructed inside the territories of sovereign states but also the spaces on the outside that are designated as not being amenable to this organization of space. Otherwise, any study of state territoriality risks falling into the “territorial trap” wherein states are viewed as internally coherent units, existing ontologically prior to the overall ordering of the state system, and wherein cross-border processes can be viewed only as “international relations” among these preexisting states (Agnew 1994; see also Sparke 2005).

I am asserting here that the same “mapping” of space that permits (and is expressed by) the surveying of a national border to define an inside also defines an outside by making possible the conceptualization of a world of equivalent states existing next to each other in relative space. Indeed, for Walker (1993) and Bartelson (1995), the discursive and material division of the world into insides and outsides is perhaps the fundamental act of sovereignty. The outside is not simply the residual space remaining after declaring an inside. If that is the prevailing image of the outside, it is an image that is itself constructed in tandem with the construction of a particular image of the inside. For instance, the discourse of “containment” that prevailed during the Cold War was based on the idea of the state as a container set against a potentially threatening environment, even though this environment (or outside) was itself the territory of other sovereign states that were constructing similar discourses about themselves and their neighbors (Chilton 1996). Similarly, throughout the nineteenth and twentieth centuries, the dominant image of the ocean as an “outside” beyond the universe of state-civilizations provided a pretext for banning all social actors operating from this outside space. Yet this representation of the ocean was itself a construction of, and within, a system: The idealization of the ocean as the ultimate outside, beyond civilization, bolstered the construction of the rest of the world—the universe of territorial states—as sovereign insides (Thomson 1994; Steinberg 2001).

At the global scale, this division of the world into “insides” that matter and “outsides” that serve to facilitate development of the “insides” typically is achieved through reference to geophysical properties. In particular, the division of the world into fundamental elements—land and sea—is viewed as the originary act of the modern socio-spatial order (Schmitt 2006b). Connery thus directs attention specifically to changing representations of the sea as a means toward understanding the political division and appropriation of land: “The triumph of a particular oceanic signification is coterminous with the universalization of land concepts, a signification that, thanks to the liquid element itself, leaves no borders, furrows, or markings” (Connery 2001, 177).

Constructing Outsides Through Movement

The concentration on the construction of insides that permeates most studies of territoriality is paralleled by a concentration on processes of spatial fixity. Most studies of societal development are informed by a perspective that locates the essence of society in spatially fixed activities that occur within bounded society-state territories:

Spatially fixed acts of production and consumption that occur within a static state territory are taken as the essential acts of society and therefore the basis of “political” state power. Other economic activities—those that take place outside state territory, those that involve trade, or those that more broadly concern capital mobility—may be engaged in if they are deemed supportive of the essential polity-building activities, but neorealist and neoliberal [international relations scholars] agree that states will refrain from these subordinate activities if they are found detrimental to the state's primary political objective of maintaining authority over territory. (Steinberg and McDowell 2003, 207)

Burch (1994) similarly asserts that the modern state system rests on a manufactured distinction (which had not existed prior to 1700) between a political arena characterized by spatially fixed, tangible property and an economic arena characterized by mobile, intangible property.

If indeed the modern system of territorial sovereignty—with its binary distinction between insides and outsides—rests on a discursive construction of fixity as the domain of the political and mobility as the domain of the economic, then a good place to begin deconstructing the distinction between inside and outside might be through an investigation of the (hidden) role of mobility in the construction of the political. For Virilio (1986, 1997), the role of mobility in contemporary politics is anything but hidden. In a perhaps overly literal interpretation of Marx's aphorism about the “annihilation of space with time” (Marx 1973, 539), Virilio asserts that information, munitions, currency, and other vehicles of power now travel at such speeds that the boundaries that historically had segmented political territories—and the Cartesian grid on which this segmentation had been based—are no longer relevant. Cultural theorists similarly stress that the communities that historically have been defined by these bounded territories are being replaced by a new system of “translocality” (Appadurai 1996b) or “traveling cultures” (Clifford 1997), in which identities are formed through connections across space rather than rootedness in place.

Whereas Virilio, Appadurai, and (to a lesser extent) Clifford stress the victory of mobility as a postmodern event, others stress the role of mobility in the construction of modernity. For Latour (1986, 1987) and Law (1986), the development of standardized procedures for moving information across space permitted the combining of information and the replication of the means by which this information was gathered, and this innovation in mobility was the key that gave Europe an unprecedented advantage in the accumulation of knowledge at the center and the projection of knowledge outward as power. Cartographic theorists similarly have noted how the projection of geographic images onto an abstract grid of relative space created the illusion of a decentered—and hence objective—world of geometry, independent from that of social hierarchy and political power, and that this facilitated the cooptation, combination, and manipulation of local knowledges so as to enable the universalization of understandings (Turnbull 1989, 1996; Woodward 1990; Mignolo 1995). ⁵

Although there are significant differences among these scholars (most notably regarding whether the role of mobility in constituting identities and state power is a modern or postmodern phenomenon), they all suggest that we turn our attention toward practices and representations of mobility if we want to avoid Agnew's “territorial trap” wherein the inside of the territorial state is assumed to be a timeless fundamental sociopolitical entity. Put another way, the ordering of space and the projection of power across space is associated with the drawing of lines of movement (which can cross boundary lines) as well as with the drawing of boundary lines (that purport to restrict movement). This is a central point of Albert (1999), who, expanding on Deleuze and Guattari's (1988) concept of reterritorialization, writes:

Sovereignty is something that has to be practised through “marking” space by boundaries of various kinds—and by mapping these boundaries in an exact science …For territory to be meaningful it has to be reproduced by the enactment of challenges to it, by questionings and erasures of boundaries as markers of space, but also through the inscription of new boundaries. (Albert 1999, 61)

In other words, through the marking and crossing of boundary lines, one defines not merely the scope of what is inside the territorial unit but also the nature of the system itself, as the system is represented as being the sum of its bounded units.

Deleuze and Guattari's insights regarding the continuing function of reterritorialization through boundary drawing and boundary crossing are important because they suggest that the relationship between mobility and fixity is more complex than simply a contradiction between a political desire to capture and enclose and an economic desire to open borders. ⁶ Indeed, they make possible a wholesale reconceptualization of the relationship between state and territory, wherein the world is understood as a decentered “empire,” characterized by networks of power that continually deterritorialize and reterritorialize (Hardt and Negri 2001). At the root of this reconceptualization of power is a heightened understanding of the social significance of movement in the constitution of space: “Space is composed of intersections of mobile elements. It is in a sense actuated by the ensemble of movements deployed within it” (de Certeau 1991, 117). Hence, movement produces politics (and states) as well as challenging them.

This perspective on the hidden roles of movement, the crossing of boundaries, and the definition of outsides in the inscription of territorial states is evident in Burnett's (2000) study of the surveying of Guyana, in which he highlights the role of the traverse surveyor who works by moving across boundaries, even as his or her stated end is the defining of an inside space of fixity with supposedly inviolable boundary lines. It is also evident in Salter's (2003) genealogy of the global passport regime, wherein the construction of these “pathways” to the “outside” contributed to the construction of the boundaries across which individuals were passing and the “insides” from which they were leaving. It is evident, perhaps most clearly, in Carter's (1996) call for sociocartographic studies in which foreground and background are reversed, where the emphasis is placed on tracing the lines of movement as “drift-lanes” (as opposed to mere routes between places) on a textured ground, rather than on the boundary lines that (purportedly) divide abstract, empty space into discrete territorial units.

Mapping Movement

It is easy to issue calls for recognizing mobility as constitutive of society, but putting this recognition into practice is much more difficult, in large part because mobility elides representation: “[Mobility] is absent the moment we reflect on it. It has passed us by” (Cresswell 2006, 57). When movement is mapped onto Euclidean space, the processes of movement, wherein people engage and produce space in their everyday lives as they draw and cross borders, are replaced by static representations. As summarized by Burnett:

Carter points out that the flattening of the land into a map turned walking the land from a way of living into little more than a symbol. The peripateia of one foot following another across the contours of the terrain became merely a line on the map. Instead of being an exploration—an active engagement with place—the passage became a fixed inscription: the place became a stage. The dialogues of foot and ground, light and eye, breath and breeze all vanished, and with them the territorial claims of wanderers and nomads, those whose relationship with place was rooted in participation (methexis), not representation (mimesis). (Burnett 2000, 171)

De Certeau elaborates on this point in his critique of route maps:

It is true that the operations of walking on can be traced on city maps in such a way as to transcribe their paths (here well-trodden, there very faint) and their trajectories (going this way and not that). But these thick or thin curves only refer, like words, to the absence of what has passed by. Surveys of routes miss what was: the act itself of passing by. The operation of walking, wandering, or “window shopping,” that, the activity of passers-by, is transformed into points that draw a totalizing and reversible line on the map. They allow us to grasp only a relic set in the nowhen of a surface of projection. Itself visible, it has the effect of making invisible the operation that made it possible. These fixations constitute procedures for forgetting. The trace left behind is substituted for the practice. It exhibits the (voracious) property that the geographical system has of being able to transform action into legibility, but in doing so it causes a way of being in the world to be forgotten. (de Certeau 1991, 97)

An extreme example of the failure of the route map to convey methexis can be seen in portolan charts, which began to be produced in the late thirteenth century to accompany portolans, lists of directional headings and distances between communities along the coasts that were frequented by Mediterranean sailors. On a portolan chart, the ocean—the space of movement—is crisscrossed by rhumb lines punctuated by compass roses, which could be used by sailors to find the heading that would get them from a port on one coast to a port on the opposite coast. Although it is debatable whether portolan charts ever saw much use as navigational aids, the persistence of rhumb lines on maps through the seventeenth century (when they certainly were not being used for navigation), as well as their appearance on world maps (where, for technical reasons, they cannot aid navigation), suggests that rhumb lines were important because they symbolized the idea of the ocean as a space that could be easily crossed (Steinberg 2005; Jacob 2006). Indeed, the prevalence of rhumb lines on seventeenth-century world maps that used the Mercator projection (which, in turn, was derived from the mathematical logic of the portolan chart; Williams 1992) further suggests that rhumb lines served less to enable navigation than they did to signify the practice of navigation and, by extension, the navigability of space.

Portolan charts were maps of idealized movement. In their depiction of an ocean crossed with rhumb lines, portolan charts were prefiguratively modern. The ocean was portrayed as an idealized space of limitless mobility (one simply needed to set one's heading and skirt around the highlighted intervening islands). Yet this representation signified nothing about the experience of moving through what de Certeau calls “practiced space”: the space that time makes (Casey 2002; Steinberg 2005).

When one looks at the ocean on a portolan chart, one is struck by the fact that, for maps that were intended to be used at sea (or to resonate with the image of crossing the sea), the sea is exceptionally empty. Although Ptolemaic maps, especially in the fifteenth century (see later), featured numerous names for bays, gulfs, seas, and oceans, these toponyms almost never appear on portolan charts, even on large-scale maps in portolan atlases where there would have been ample room for their placement. Similarly, one finds few, if any, sea monsters, fish, ships, or other indicators of human or nonhuman life on the sea that might evoke the experience of the sailor moving through a marine environment. Rather, the portolan chart idealizes the act of moving along a straight route, through empty space. Although portolan charts are “route-enhancing” maps that foreground the act of (or at least the potential for) maritime movement (Woodward 1990), the movement that is represented is a disembodied process that abstracts movement from time, environment, or human experience. As frequently occurs when the act of wayfinding is depicted cartographically, the experiential aspect of movement disappears and the map loses its utility as a functional tool (Akerman 2006; Delano-Smith 2006).

The limits of the cartographic representation of movement (whether on a portolan chart or a modern route map) are given dramatic effect by geographer-artist Matthew Knutzen (2000) in his work Subway Map (Figure 1), in which he reproduces the New York City subway map but with all the route lines omitted. By its stark reduction of routes to abstracted, nonlinear symbols, Subway Map spurs the viewer to consider how all subway maps (and other route maps) are, in de Certeau's words, “procedures of forgetting” rather than representations of the experience of movement.

Figure 1 “Subway Map” by Matthew Knutzen (2000). Image courtesy of the artist.

At the root of the critique leveled by Knutzen (and by Burnett, Carter, and de Certeau) is the problem that movement involves travel through time as well as space and, as Massey (2005) and Grosz (1995) have shown, the coconstitution of space and time is incompatible with the concept of representation. So long as the world is conceived of as a set of stable points, on which and across which objects emerge and subsequently move, the distinction between a contextual, stable background of space and a dynamic, mobile foreground of time will persist, leading to incomplete depictions of mobility as a foundational social process (Elden 2005).

An alternative would be to map the experience of movement through time. In addition to various non-Western maps that depict movement through time (Harley, Woodward, and Lewis 1992–1998), this representation of movement can be seen in medieval mappaemundi that “reveal explicitly the layers of historical events, processes, and artifacts that have shaped the present landscape, as well as the objects that now exist …portray[ing] the location and distribution not only of objects and conditions but also of events and processes” (Woodward 1985, 519–20). This depiction contrasts with the modern map, where “the element of time, so important in the mappamundi, disappear[s] in favor of a world in the present tense” (Edson 2007, 14).

De Certeau praises mappaemundi for the way that they capture the experience of movement: “an outline marked out by footprints with regular gaps between them and by pictures of the successive events that took place in the course of the journey …not a ‘geographical map’ but ‘history book'” (de Certeau 1991, 120; see also Woodward 1985, 1987b). Although a mappamundi might be more “accurate” than the modern map in capturing the experience of movement through time, it is no better (and probably worse) at capturing movement through spaces that, even if they are dynamically produced, always exist in relation to other spaces. ⁷ The potentials and limits of this position are exemplified by another piece of conceptual art based on the New York City subway map, Hand-held Subway, by Nina Katchadourian (Figure 2). In an apparent invocation of Deleuze's statement (after Bergson) that “movement is no less outside me than in me” (Deleuze 1988, 75), Katchadourian (1996) has cut out the routes from the subway map and reconstituted them within her hand, leaving the viewer wondering whether, when one moves, one's body moves along routes (that are set against a fixed background of space) or whether routes move through one's body (as a collage of images, experiences, and memories that flash through one's brain as one moves through time). Thus, Katchadourian illustrates Deleuze's concept wherein “our perception contracts ‘an incalculable multitude of rememorialized elements'; at each instant, our present infinitely contracts our past” (Deleuze 1988, 74). As Massey (2005) notes, however, Deleuze's conceptualization falls short because, although the “here and now” of the self is constituted through the accumulated memories and images of passage through time, it is equally constituted through the accumulation of passages through space. Hence, although Katchadourian's artwork is to be commended for leading the viewer to consider the methexis of mobility, it fails to fully capture the practice because mobility involves movement through space as well as time.

Figure 2 “Hand-held Subway” by Nina Katchadourian (1996). Image courtesy of the artist and Sara Meltzer gallery.

Ptolemy and the “Transport Revolution”

By the sixteenth century, as Europe's “transport revolution” was commencing and the ideal of the sovereign, territorial state was beginning to congeal, both the portolan chart and the mappamundi were giving way to systems of geographic representation based on the gridded space of the Ptolemaic map. By introducing the concept of a grid for plotting points against an ideal of Euclidean space, the “discovery” of Ptolemy's Geography and its initial European publication in 1409 revolutionized European cartography. These innovations followed from Ptolemy's application of an “astronomer's vision” to a mapping of the world (Cosgrove 2001, 105). By looking up at the stars, one could locate oneself on an abstract grid. Using the grid, one could plot a world of relative locations on Earth's surface. This system of plotted points then could be used to construct a tableau of static locations for mapping natural and social features. The effect was to replace the “center-enhancing” perspective of the mappamundi with an “equipollent-coordinate” perspective, “entirely geometrical and abstract, independent of the geographical content beneath” (Woodward 1990, 119). Although the details of Ptolemy's Geography were soon surpassed by the findings of contemporary explorers, the equipollent-coordinate logic underlying his method continues to inform cartographic practice, even if it shares the stage with other spatial logics in everyday practice (Curry 2005).

Throughout this era, world maps inspired by Ptolemy made little attempt to represent movement and hence there was little depiction of the boundary lines, forged and reproduced through movement, that divided the terrestrial inside space of land, society, state, and sovereign from the marine outside beyond. This was in contrast with mappaemundi and portolan charts, which each came to produce elemental distinctions between land and sea, albeit in very different ways, as they sought to represent the process of movement. Mappaemundi sought to depict journeys through time, and these were believed to occur on land, because, as Aristotle noted, “The sea surrounds the earth like the horizon, and ships cannot work in this sea …and nobody has crossed it, and it is unknown where it ends” (cited in Johnson 1994, 38). Thus the ocean was constructed as an uncrossed, unnamed, and unmapped outside, a “phantasmic …ring which was above all a mental ring, a psychological barrier” (Biagini and Hoyle 1993, 3). Conversely, portolan charts foregrounded the ocean, the space of the mobile practice being represented on these maps. In many cases, rhumb lines ended when they reached the coast. Mirroring the ocean on mappaemundi, land areas on portolan charts (beyond the coastal points where ocean routes terminated) were external spaces, “vague areas inside the border—but beyond the contour of the map” (Jacob 2006, 111).

On early Ptolemaic charts, however, with their conception of the world as a field of equivalent points, there was no such distinction between a foregrounded “inside” and a backgrounded “outside” and no boundary lines that attempted to communicate a division between purported zones of movement and stasis. Rather, early Ptolemaic maps recognized a fundamental similarity between the earthly elements of land and sea. Regions on Ptolemaic maps simply were labels for clusters of points in space rather than designations of distinct and relevant social (or sociopolitical) units (Akerman 1984), and “in Ptolemy the lines [were] reduced to the status of markers on a surface of infinite variation” (Curry 2005, 685).

Set against the ideal of points-in-space located on a fixed grid (which could be laid across Earth's surface by looking to the sky), the distinction between land and sea, which had been so important on the mappamundi and the portolan chart, was greatly diminished. For Ptolemy (and fifteenth- and sixteenth-century cartographers working in his tradition), the ocean was a field of points that could be mapped and named just like land (Lewis 1999). Thus, whereas the ocean had few if any regional names on mappaemundi or portolan charts, world maps in early printed editions of Ptolemy's Geography were replete with names for oceans, seas, gulfs, and bays (and even more ocean names appeared on the accompanying regional maps; Table 1). ⁸

Table 1 Number of names for ocean regions on Ptolemaic world maps in printed copies of Ptolemy's Geography, 1478–1730

Ptolemy Series (based on review of 47 printed copies of The Geography in the Newberry Library's Henry Stevens Collection)	Names in outer seas (Atlantic/Indian/North)	Names in inner seas (Mediterranean/Black/Caspian)	Total ocean region names
Rome 1478 (reprinted in Rome 1490, 1507, 1508)	26	17	43
Rome 1480	29	21	50
Ulm 1482 (reprinted in Ulm 1486)	19	10	29
Venice 1511	17	14	31
Strassburg 1513 (reprinted in Strassburg 1520)	18	15	33
Strassburg 1522 (reprinted in Strassburg 1525; Lyons 1535; Vienna 1541)	12	2	14
Basle 1540 (reprinted in Basle 1542, 1545, 1552)	8	3	11
Venice 1561 (reprinted in Venice 1562, 1564a, 1564b)	9	3	12
Venice 1574 (reprinted in Venice 1598a, 1599)	9	3	12
Cologne 1578 (reprinted in Cologne 1584; Amsterdam/Frankfurt 1605; Amsterdam 1618; Frankfurt/Amsterdam 1695, 1698, 1704; Amsterdam 1730)	8	3	11
Venice 1596 (reprinted in Cologne/Arnheim 1597; Venice 1598b; Cologne 1608; Arnheim 1617; Padua 1621)	9	2	11

These maps’ turn away from the land–sea distinction that implicitly had underlain mappaemundi and portolan charts also can be seen in the way in which they used a combination of land-space and ocean-space to mark the margins of the known world (Dilke 1987). In contrast with mappaemundi, where knowable land was encircled by an unknowable ocean, or portolan charts, where a knowable (or, to be more precise, navigable) ocean was surrounded by unnavigable land, on a Ptolemaic map, all points in space were knowable (and amenable to toponyms), all were plottable, and all existed in relation to each other.

Thus, the Ptolemaic map presents a perspective wherein all points in space are equivalent and relative. Although this might seem to establish the context for a world without discontinuous categories (e.g., the binary between land and ocean), in fact, as Foucault (1973) has shown, equivalence makes calculation, and hence comparison, possible (Elden 2006; Steinberg 2006). The comparison of like units finds cartographic expression in the depiction of stable, homogenous entities demarcated by fixed boundaries, as in a map of political units or land-use management areas where the complexities of boundary crossing and boundary formation are obscured by the illusion of permanence (Harris and Hazen 2006; Harvey 2006). Such a map, which resonates with the medieval (or Platonic) depiction of the island polis (Gillis 2004; Steinberg 2005), implies that there are absolute and immutable distinctions between territorial insides (the land-space of state societies that are idealized as arenas of sedentarism, civilization, and development) and outsides (most notably the marine wild, which is idealized as a space of mobility devoid of natural or social forces that might hinder mobility).

In other words, the Ptolemaic map presented a world of points without lines, a representation that stood in contrast to many other maps of the era that implied fixed and elemental distinctions between categories of space. At the same time, though, by depicting the world as a set of relative points—a res extensa amenable to calculation—it established the basis for the modern map that depicts absolute distinctions between insides and outsides. To trace the evolution of the modern map from one that depicted a world of equivalent points to one that depicted a world of differentiated bounded spaces, the remainder of this article chronicles the cartographic construction of the ultimate outside—the asocial marine void—from 1500 to 1800.

On Analyzing What Is Not There

At this point, one is faced with a methodological dilemma. If, as Cresswell (2006) notes, mobility elides representation, then spaces of mobility are similarly difficult to represent, at least when mobility is idealized as a process that occurs beyond experience and abstracted from time, as in the subway map or the portolan chart. When mobility is conceived in this manner, the surfaces across which one moves are valued because they are (ideally) devoid of features and hence amenable to annihilation. If one's aim is to historically trace the evolution of the representation of the ocean as an empty, external space (as a counterpoint to the evolution of the representation of terrestrial states as bounded and sovereign), how does one record the development of something that is increasingly represented (or nonrepresented) as absent? There is no “there” in mobility (or its spaces) to be represented.

This dilemma is moot if one accepts the argument that maps are fundamentally nonrepresentational. Kitchin and Dodge (2007) suggest that we move beyond semiosis and view maps not as stable entities that code the world through representation but rather as processes that are enacted (or performed) through user engagement, a perspective that resonates with a broader trend in social theory toward appreciating the fluidity of objects (Law 2002). Although Kitchin and Dodge might well be correct in pointing to the instability of a map's meaning and hence in identifying the map as a nonrepresentational object, it nonetheless remains the case that maps have achieved much of their power by purporting to represent. Although the representations suggested by maps are anything but stable (the identities of consumers, producers, maps, and the referred spaces all are transformed in multiple ways in the process of “performing” a map; Brown and Knopp 2008), maps nonetheless achieve their power through representations, as map producers and users engage a web of representational conventions and preexisting conceptualizations to give meaning to the dots, lines, and shadings inscribed as cartographic text (Hanna et al. 2004; Del Casino and Hanna 2006).

It follows that the appropriate critical response might be not to turn away from semiotics but to turn toward a nuanced semiotics that recognizes the particular ways in which a map claims authority. Indeed, Jacob's (2006) study of cartographic semiotics suggests that maps are particularly powerful when representing a process that is difficult to represent (in this case, movement) or a space that similarly defies representation (in this case, the ocean). Thus a study of cartographic representations would seem to be especially fruitful for understanding the links between the construction of the ocean as an external space of mobility and the concomitant construction of a global political system of sovereign, territorial states.

Jacob notes that the standard semiotic triad of signifier (the symbol in question), signified (the idea to which the signifier refers), and referent (the material reality that the signifier represents) does not hold for maps. Whereas the authority of most signifiers is derived from their link (or at least their perceived link) with a referent, a map has a special claim to authority because it cannot be verified. Unlike, for instance, a photograph, a map does not claim to bear any resemblance to the thing that it represents: “The map's link to its referent is clearly different from the mimesis that belongs to images in general …Maps are based upon a collective consensus and an individual accord that grounds their validity, their accuracy, no doubt for lack of a capacity to measure resemblance because of an inaccessible referent” (Jacob 2006, 272). ⁹ Jacob's statement about the inaccessibility of the referent is particularly true in the case of a world map, where the referent is at a scale entirely beyond the realm of human experience.

The semiotics of the map is further complicated for Jacob because, in addition to having no direct correspondence with a referent, the signifier is also obscured: “An effective map is transparent because it is a signified without a signifier. [The signifier] vanishes in the visual and intellectual operation that unfolds its content …The map is not an object but a function” (Jacob 2006, 11). A map thus achieves its authority not when it is found to approximate reality but rather when it becomes something that one uses un-self-consciously, as a proxy for reality. In this manner, a map begins to insinuate itself as reality (see also Crampton 2003; Pickles 2003).

Just as Jacob's statement about the lack of verifiable correspondence between a map and its referent is particularly true for world maps, his statement about the lack of evident correspondence between a map and its signified is particularly true for maps that attempt to depict spaces or processes of movement. Because maps are rarely useful as way-finding tools (Akerman 2006; Delano-Smith 2006; this is especially so in the case of a world map; Cosgrove 2001), a map's depiction of a space of movement is less a signifying tool for movement and more an expression of an idea about the ease (or lack of ease) of mobility, as well as a means by which the viewer of the map idealizes herself or himself as a “located” subject who is associated with a bounded space but who maintains the ability to move across space (Conley 1996). Thus, for instance, a rhumb line should be considered as a representation of the idea that the ocean can be crossed rather than a tool that enables its crossing (and, again, this is especially so in the case of a world map, where rhumb lines are not functional; Steinberg 2005). Likewise, a line on a transit system's route map suggests the idea of friction-free movement beyond time rather than representing the actual experience of travel.

Of course, if a map is a projection of an idea then the question arises: Whose idea is being projected onto the map? Every map involves some fusion of the cartographer's personal preference and the commonly held conventions and perspectives of the intended readership. As one analyzes maps, there is always the danger, in extrapolating from individual maps to social norms and attitudes, that any particular map represents the attitude of the cartographer (or the cartographer's patron) rather than a social norm. To diminish this possibility, the analysis discussed next, in addition to using a very large data set, is restricted to printed maps, where cartographers typically feel a greater pressure to conform to contemporary norms about the world and its spaces than they do when producing manuscript maps (Jacob 2006).

In addition, the data set analyzed here is restricted to world maps, in part for the reasons noted in the preceding paragraphs and in part because, in the post-Columbian world, the cartographer creating a world map has no choice but to represent broad (but not infinite) swaths of both land-space and ocean-space. On a post-Columbian world map, the ocean cannot be ignored, but it can be presented as external to social processes. It is the evolution of this representation that is traced in the next section of this article.

To assemble the data set of printed world maps analyzed here, I first reviewed 393 maps reprinted in Shirley's (2001) The Mapping of the World: Early Printed World Maps, 1472–1700 (4th ed.), a cartobibliography of every known printed world map from that era that also includes, in most cases, legible reprints. ¹⁰ By 1700, the number of world maps being printed each year had exploded and it would have been impossible to identify, let alone analyze, the entire population. Instead, for the eighteenth century, I searched through sheet maps, reprint books, atlases, special geography, travel narratives, and other books and magazines likely to contain world maps at each of four major research libraries noted for their map collections, as well as utilizing Web-based sources. ¹¹ This resulted in my identifying and annotating 198 maps from the eighteenth century, for a total data set of 591 maps (Table 2).

Table 2 Number of maps annotated, by year of production

Range of years	Number of maps
1501–1525	21
1526–1550	26
1551–1575	34
1576–1600	63
1601–1625	55
1626–1650	48
1651–1675	54
1676–1700	92
1701–1725	35
1726–1750	19
1751–1775	45
1776–1800	99
Total	591

Externalizing the Ocean

An initial indicator of the construction of the ocean as an external space, beyond society and nature and without distinct places, can be seen in Table 1, as the number of ocean toponyms in the world maps in volumes of Ptolemy's Geography dropped dramatically during the early sixteenth century, from a high of fifty in 1480 to a maximum of fourteen in maps printed from 1522 on. Despite this trend in Ptolemaic representations, however, the sixteenth century was not an era in which the ocean simply was emptied of social, natural, and metaphysical significance. Although elements of mercantilist-era ideology constructed the ocean as a “great void” (thus prefiguring the ideal that was to predominate under industrial capitalism), the mercantilist-era ocean was also constructed as a space of social power, routes of connection, fearsome biota, scientific exploration, and romantic imagination (Steinberg 2001) and even its idealized immateriality was imbued with metaphysical significance (Connery 2006). As several scholars of empire have noted, the idea of an intervening, marine space between the core and the periphery was crucial to the discursive and material workings of the era's European empires (Richards 1993; Said 1993; Dettelbach 1996; Sidaway and Power 2005). Portuguese conquerors, for instance, claimed distant territories in the Americas not by celebrating that they had found a distant land but by celebrating that they had discovered a way to get there (Seed 1995). In short, the ocean played a crucial role in both the material life and the imagination of sixteenth- through eighteenth-century Europeans, and its materiality and social significance could not simply be wished away by the era's cartographers. Eventually, the ideal of the ocean as a “great void” beyond society, suggested by the decline in ocean toponyms on Ptolemaic maps, prevailed, and this ideal continues to serve as a counterpoint to the ideal of land-space as a universe of naturally occurring bounded “insides” (Steinberg 2001). It took several centuries, however, for cartographers to align their representation of the ocean with the emergent norm of state territoriality.

In the meantime, mercantilist-era cartographers were faced with a crisis of representation that paralleled a similar crisis in ocean regulation (Steinberg 1999b): How does one depict the ocean as a space of connection, in which case its “value” is best emphasized by portraying the ocean as a space devoid of any social influence or materiality that might interfere with the free flow of ships, while, at the same time, depicting the ocean's crucial function in mercantilist-era cosmology and political economy? This crisis echoes the aforementioned problems in representing the practices and spaces of mobility: How does one depict (and even celebrate the importance of) a process that is ideally never “there”?

To trace the changing representation of the ocean on world maps, I focus on graphic images that are placed on the ocean by cartographers. Although some might dismiss these artistic flourishes as the decorative whim of the cartographer, contrasting them with the map as an expression of scientific knowledge, this distinction between map as science and image as art is misguided (Woodward 1987a; Casey 2005; Cosgrove 2006; Wood 2006). Jacob rhetorically asks:

Do not these graphic features [and the absence of features—pes] reveal culturally determined choices, a conception of the world and a state of knowledge and thinking? Do they not constitute a fundamental level that might bring to light the purpose of the map, the intellectual, emotional, and imaginary affects that it might arouse, far exceeding a simple geographical reading? (Jacob 2006, 103–4)

Thus, for instance, the decline in ocean toponyms on Ptolemaic world maps during the early sixteenth century indicates not so much a change in scientific knowledge (it seems unlikely that knowledge of the ocean declined so precipitously during this era) but rather a changing conception of what the ocean is (or should be). This is all the more so for nontextual images which, by “reintroduc[ing] the particular into an undifferentiated space” (Jacob 2006, 159), communicate the perceived character or nature of a region.

The Sixteenth-Century Ocean: A Space of Matter

At the beginning of the sixteenth century, most cartographers, following the convention of Ptolemaic world maps, made little distinction between land and sea. They typically represented the ocean as a material, textured space through such techniques as stipling or drawing in waves (Figure 3). This was a problematic representation, though, for the emergent ideal of the ocean as an empty, external void, because a material ocean presented potential obstacles to movement. Furthermore, such an ocean presumably consisted of places with distinct natures, and these could then be enclosed by boundaries that would serve as further barriers to mobility. If capitalism is built around the fantasy of the “annihilation of space with time” and, in particular, through the designation of empty spaces of movement that are especially amenable to annihilation, then the idea of the ocean as a material, textured space of nature was incompatible with this emergent capitalist ideal. The prevalence of this representation, which might be considered a holdover from medieval perspectives of the ocean as a viscous swamp rather than a wet nothingness (Lestringant 1989; Cosgrove 2001; Gillis 2004), rapidly declined during the study period, peaking in the late sixteenth century and disappearing entirely by 1700 (Figure 4, bottom frame).

Figure 3 Detail of the Pacific coast of Central America from Ortelius (1570).

Figure 4 Median number of ships or marine biota per square meter and percentage displaying texture on the ocean, for world maps printed between 1501 and 1800, grouped by twenty-five-year intervals. On ship and biota graphs, a bold line indicates median, a box indicates interquartile range, and “whisker” indicates 1.5 times the interquartile range. Outliers have been omitted to enable presentation at a legible scale.

As cartographers began to turn away from representing the ocean as land-like, they nonetheless continued to represent it as a material space of nature and society, highlighting specifically marine characteristics by drawing in generic ships as well as marine biota such as fish and sea monsters (Figure 3). Although these new representations, which peaked between 1550 and 1650, began to establish the basis for a conceptual distinction between land and sea as practically different kinds of spaces with different features and social processes, they still implied that the ocean, like land, was filled with social processes and natural features and that therefore, like land, it could be bounded, controlled, and organized. Such an image of the ocean was incompatible with the emergent ideal of an ocean beyond society, counterposed against a world of bounded land territories that contain and define fundamental societal units. Thus, by 1700, ships and biota, like stipling and waves, had nearly disappeared from representations of the sea (Figure 4). ¹²

The Seventeenth-Century Ocean: A Space of Routes

If the sixteenth century was characterized by depictions of the ocean as a space of nature and society, the seventeenth century was dominated by depictions of the ocean as a space of routes. The first of these representations to gain ascendency was the drawing of rhumb lines that, as has been noted, served no navigational purpose on world maps but served to impart the idea that the ocean was a space on which one could navigate by constructing routes through a putatively placeless in-between space (Figure 5). Rhumb lines, which were most common in the period between 1600 and 1650 but persisted through 1700, were followed by a second depiction, most often occurring between 1650 and 1725, in which the names of basin-based ocean regions (e.g., the South Atlantic for the body of water between the east coast of South America and the west coast of Africa) were replaced by a series of regions defined by coast-hugging arcs (e.g., the Ethiopian Ocean, which arced beneath the African continent, from the South Atlantic to the Indian Ocean; Figure 6; see also Lewis 1999). Arced ocean regions, like rhumb lines, suggested that the ocean consisted of a series of routes that could be traveled. This same time period also saw the emergence of a third representation that similarly portrayed the ocean as a space of routes: generic commercial routes, such as the route to the East Indies.

Figure 5 World map from Teixeira (1630).

Figure 6 Detail of South America from Nolin (1708).

By depicting the ocean as a space of differentiated and functional routes, the ocean was cartographically represented in a manner that closely paralleled its legal designation during the same era as res extra commercium. According to this doctrine, expressed most famously by Grotius ([1608] 1916) in his treatise Mare Liberum, although the ocean provides a resource and thus can be conceived of as property of a sort, it cannot be fully possessed (with all the rights of exclusion and alienation normally associated with property) because the principal resource that it provides—an environment for navigation—is inexhaustible and, as such, is granted to all by natural law (Steinberg 2001).

This conceptualization of the ocean, both in maps of the era and in the legal doctrine of res extra commercium, was transitional. On the one hand, unlike the sixteenth-century depiction of the ocean, the ocean was no longer perceived as consisting of specific points that were rich with nature and that hypothetically could be gathered into territories, enclosed, and developed, like land. On the other hand, although the depiction of the ocean as a space of distinct routes was not the same as land-space, neither was it the antithesis of land. The ocean was still represented as a space of pathways and discrete (if not enclosable) spaces, something more material than the empty and abstract “hyperspace of pure circulation” (Castells 1996, 475) that was to prevail later as a counterpoint to the division of land into bounded territories. Although the specific ways of depicting the ocean as a space of routes waxed and waned during the seventeenth century, all of these depictions had essentially disappeared by 1725 (Figure 7).

Figure 7 Percentage of world maps printed between 1501 and 1800 displaying rhumb lines, curved oceans, or generic maritime commerce routes, grouped by twenty-five-year intervals.

The Eighteenth-Century Ocean: A Space of Mathematics and Memory

By 1700, a new set of depictions of the ocean had emerged, solidifying the ocean's position as an external void that could serve as a counterpoint to the territorial state. The first of these depictions already had begun to appear early in the seventeenth century: the ecliptic (or zodiac) line (Figure 6). Much like rhumb lines, the ecliptic line pays homage to the act of navigation without actually enabling navigation. However, unlike rhumb lines or the other navigational signifiers that were common during the previous century (arced ocean regions and generic commerce routes), however, the ecliptic line signifies that the ocean is less a space of discrete routes than an abstract, mathematical, and materially empty space of points that can be crossed at will. Whereas the earlier generation of signifiers that pointed to the ocean as a space of mobility still implied a degree of materiality, social relations, and limits, the ecliptic line suggests that the ocean is a space of pure mathematical abstraction, a dematerialized arena of potentially limitless time–space compression, or idealized annihilation.

Cartographers during the eighteenth century adopted a second means to demonstrate that the ocean was a space of mobility, by depicting the ships and routes of historic explorers such as Columbus, Magellan, and Drake (Figures 6 and 8). In this way, the ocean was presented as a space of movement that achieved its significance not through its contemporary materiality or through ongoing interactions with its nature or even its space but as an arena of history. Depictions of historic explorers’ routes celebrated the ocean—and, in some cases, specific ocean regions or specific historical interactions with the ocean—as a source of national pride and identity formation. At the same time, however, they avoided any implication that the ocean was a space for contemporary social activity or assertions of territorial power. Although explorers’ routes suggested that the ocean was historically significant and the ecliptic line suggested that it could continue to serve society through limitless crossing, these two representations also suggested that it was a space beyond the world of developable places and enclosable state-territories that was emerging on land.

The third ocean representation that was common in the eighteenth century demonstrates the extent to which producers (and, presumably, consumers) of maps had come to accept a distinction between land as a set of developable points, defined by the boundaries of the territorial state, and the ocean as an empty outside that defied territorialization, civilization, or differentiation into “places” of nature and society. The Grand Banks, the highly productive fishing ground off the coast of Newfoundland, on the one hand, had the physical properties of the ocean, but, on the other hand, had the social properties of land as it could be bounded and developed. The Grand Banks thus was a geographic outlier that defied the modern ordering of the world with its division between developable terrestrial territories and a marine nonterritory of mobility. To cope with this outlier space, cartographers of the era borrowed a series of representations that were commonly used on larger scale navigational maps to depict shallows or shoals—shading, stipling, or dotted lines—to depict the Grand Banks as an intermediate space (Figure 8).

Figure 8 Detail of the Grand Banks from Nolin (1708).

Finally, although the focus of this article has been on the construction of the ocean as an unboundable and empty space of movement—an external counterpoint to the emergent idea of the territorial state as a bounded arena of control and development-in-place—it bears noting that the emergence of these final three signifiers that suggest the ocean's status as a space beyond society coincided with the rise of territorial state boundaries on world maps (Figure 9). Indeed, a generalized linear model applied to the data summarized in Figure 9 (controlling for interyear variance by using the mean value for each year with multiple data points) reveals that the presence of each of these three signifiers is positively correlated with the presence of state boundary lines, and the correlation between the ecliptic line and state boundary lines is significant at the 0.001 level (p = 0.000274). This corresponds with the findings of others who have traced the emergence of state boundary lines from a more traditional perspective that roots the rise of the territorial state in a history of boundary drawing and the designation of “insides” (e.g., Sahlins 1989; Akerman 1995; Black 1997; Brotton 1998; Biggs 1999; Kamen 2000; B. Anderson 2006). By 1800, boundary lines that defined both insides and outsides were the norm.

Figure 9 Percentage of world maps printed between 1501 and 1800 displaying the ecliptic line, explorers’ routes, the Grand Banks, or state boundary lines, grouped by twenty-five-year intervals.

Conclusions

Since the nineteenth century, few cartographers or policymakers have dissented from the dominant view of the ocean as an external space, beyond the world of state-societies. Although Mahan (1890) and Mackinder (1904) differed greatly in their opinions regarding the geopolitical significance of the ocean—Mahan saw it as the arena on which global dominance would continue to be fought, whereas Mackinder felt that its days as a crucial surface for asserting and projecting force had passed—the perspectives of both geopolitical theorists rested on

an idealization of the sea as an unmanaged and unmanageable surface …that resonates with the spatial assumptions that permeate realist theories of international politics …As unclaimed and unclaimable “international” space, the world-ocean lends itself to being constructed as the space of anarchic competition par excellence, where ontologically pre-existent and essentially equivalent nation-states do battle in unbridled competition for global spoils. (Steinberg 2001, 17)

In other words, both sides of this keystone geopolitical debate accepted the construction of the ocean as a fundamentally external space, a counterpoint to the bounded, terrestrial state within whose borders the anarchy of global competition was contained.

Given the degree to which this externalization of the ocean has been accepted, it is not surprising that one finds few world maps today that display the ocean features that were prevalent in the sixteenth, seventeenth, or eighteenth centuries. By the twentieth century, cartographers had few qualms about making maps that implied that the ocean was fundamentally external to modern society and its foundational territorial units (Konvitz 1979). Indeed, it is noteworthy that few modern “political” world maps depict the political divisions of the ocean—territorial waters, exclusive economic zones, high seas, and so on—instead portraying an undifferentiated, featureless space. This is yet another indication of how, by the twentieth century, there was an accepted distinction between land as the space of sedentarism, civilization, and politics and the ocean as an external, asocial space across which one simply moved.

By the turn of the twenty-first century, the idea of land-space as the domain of bounded, developable insides and the ocean as an external surface for movement had become so entrenched that contemporary cartographers were faced with a representational dilemma similar to the one that had been posed by the Grand Banks: How does one depict Antarctica? Like the Grand Banks, Antarctica has the physical properties of one element (it is solid, like land) but the social properties of the other element (it resists division into sovereign, territorial space, albeit with some dissent). Thus, on the Central Intelligence Agency's (2000) political world map, land is beige (and divided by state boundaries), the ocean is blue (and undivided), and Antarctica, as an outlier space that is not quite land or territory but not quite sea or nonterritory, is designated with a unique, third color: white (Figure 10).

Figure 10 Political world map by the Central Intelligence Agency (2000).

Despite the prevalence of the ocean as external to society in today's cartographic representations, the history narrated in this article reveals that the depiction of the ocean as a space beyond territorialization, and the parallel depiction of land as a series of discrete, bounded territories, was a long time in the making. Indeed, the idea of depicting the ocean as a materially empty wilderness likely was counterintuitive to the mercantilist-era cartographer, both because one would expect a significant space to have a tangible materiality and because cartographers likely were aware that the ocean was not truly an undifferentiated mass devoid of nature. Thus, this history of ocean mapping reveals how, although maps achieve their power by attributing fixed properties to bounded and apparently natural spaces, the act of attempting to fix properties to space (by both the map's producer and its consumer) “offers productive possibilities, particularly through the emergence of constitutive spaces at/in the ‘gaps and fissures’ of the mapping processes” (Del Casino and Hanna 2006, 41). These “gaps and fissures” can occur at specific spaces that defy emergent categorizations (as, for instance, can be seen in the decision to depict spaces like the Grand Banks and Antarctica with intermediate representations that both reproduce and challenge the binary division of the world into social and extrasocial spaces) or they can be more conceptual (as, for instance, in attempts by mercantilist-era cartographers to develop an ideological framework wherein spaces of mobility were perceived as both socially significant and materially empty).

A history of how producers and users of maps attempt to fill these “gaps and fissures” so that they can be placed within a priori categories of inside and outside, fixity and mobility, and social and extrasocial, and how these “gaps and fissures” nonetheless reappear, suggests that the study and practice (or performance) of cartography can play an important role in generating new perspectives on the categories and institutions that define modern society. As Kwan (2007) notes, a range of alternative visualization practitioners, from advocates of consciousness-raising participatory geographic information systems to situationist members of the art-and-cartography movement, have made this assertion. Perhaps it is not surprising that so many of these individuals’ experiments with alternative modes of visualization focus on mobility. As the examples discussed here have shown, it is difficult to represent spaces of mobility in a manner that simultaneously portrays the experience of movement, the metric space through which one is moving, the idealized immateriality of that space, the actual materiality of that space, and the space's significance in the reproduction of social institutions that go beyond the practices or designated spaces of mobility. Thus, the depiction of mobility has emerged as an arena not just for gaining new insights on the world, but also for advancing cartographic practice.

Kwan's (2008) work is representative of these experiments in cartographic representation, as she combines the representational practices of time-geography with an attempt to depict the subjective perceptions and emotions of individuals moving through an urban environment (in this case, Muslim women in Columbus, Ohio). This work is exemplary not just because Kwan attempts to portray the experiences of individuals. It is also significant because, in representing these experiences, Kwan reconstitutes the city not as a series of points with ontologically prior characteristics but as a series of affect-laden surfaces with meanings that are realized by the individual as he or she moves across them (or, in instances when the space is perceived as too dangerous, as he or she contemplates moving across them).

In bringing mobility to the center of her mapping of both personal and urban spaces, Kwan aligns herself with a growing number of scholars who stress that mobility is a constitutive social practice (Hannam, Sheller, and Urry 2006). As this article has suggested, this new focus on mobility has far-reaching ramifications: Recentering movement (and the subjective experiences that emerge as one moves) involves rethinking a series of cartographic as well as political norms that have developed over hundreds of years and that underlie the very foundation of modern society.

Thus, this article suggests two related agendas for further research. In one sense, this article has called for inverting foreground and background on our maps. Instead of focusing attention on the bounded land territories that we take for granted as the spaces of society, what would happen if we foregrounded the spaces across which people (and commodities, information, and capital) move, so that they become conceived of as something other than the residual spaces that are left “outside” society? As this study has shown, the designation of some spaces as external spaces of mobility has played a key role in the ways in which we order our world, including the ways in which we bound and order the “inside” spaces in which mobility is purported to play only a secondary role. Although cultural studies and critical development theorists have for some time focused on the processes by which cultures are designated as “others” that are in need of development (e.g., Said 1978), much less attention has been devoted to the processes by which some of the world's spaces have been designated as being so “outside” that they lack even the potential for development. Historically, the construction of the modern system of territorial states has also involved the designation of outsides as being beyond territorial control, and as one studies the processes of drawing (and ascribing meaning to) boundary lines, one must focus on the construction of outsides as well as insides. Furthermore, this process has not ended with the formation of the juridical ideal of the modern, sovereign, territorial nation-state. There are numerous configurations by which spaces and peoples can be constructed as inside or outside the protection of the state, and this is as much a process of constructing outsides as it is one of constructing insides. ¹³

This agenda, however, also suggests a second, more radical agenda where, instead of tracing the historic externalization of a specific space (like the ocean) or social function (like mobility), we revision a world without distinct insides and outsides. Under this agenda, one would go beyond bringing externalized spaces of mobility to the foreground. Instead, one would consider that, in fact, all spaces are spaces of both mobility and fixity and all social processes are driven by ideologies of both internalization and externalization. This is an ambitious agenda. Analytical thought, after all, involves differentiating, comparing, and drawing “lines” of contrast. A first step, however, might well involve erasing the boundaries on our maps that divide inside from outside. Perhaps then we can begin working from behind the lines.

Acknowledgments

Research for this article was supported by a John S. Best Fellowship from the American Geographical Society Collection at the University of Wisconsin–Milwaukee, a Short Term Fellowship in the History of Cartography from the Newberry Library, a fellowship from the Dorothy and Lewis B. Cullman Center for Scholars and Writers at the New York Public Library, an Arthur and Janet Holzheimer Fellowship from the Institute for Research in the Humanities and the History of Cartography Project at the University of Wisconsin–Madison, and an Other Globalizations Fellowship from the Center for Cultural Studies at the University of California–Santa Cruz.

In addition, I am grateful for comments and questions from audience members at presentations of earlier versions of this article at Rutgers, Lund, Princeton, and Louisiana State Universities, the National University of Singapore, and the Universities of Washington, Wisconsin–Milwaukee, and California–Santa Cruz. Additional thanks go to Seth Bassett, Barbie Bischof, Mei-Po Kwan, and the anonymous reviewers for comments on final drafts of this article, with further thanks to Seth Bassett for assistance in preparing graphics and conducting statistical analysis.

Notes

1. Some of the classics in this literature include Anzaldua (1987), Butler (1990), Gilroy (1993), Bhabha (1994), Appadurai (1996a), Clifford (1997), Gupta and Ferguson (1997), Ong (1998), and Tsing (2004). Of particular interest here are scholars who have used this perspective to question the ontology of the sovereign, territorial state, such as Shapiro and Alker (1995), Newman (1999), Albert, Jacobson, and Lapid (2001), Sparke (2005), and Rajaram and Grundy-Warr (2007).

2. Price (2004) makes a similar point, specifically with reference to critical borderland studies. See also Cresswell (2006).

3. By 1755 the territorial demarcation of states on maps had been normalized to the point that Robert de Vaugondy felt it necessary to apologize for the omission of San Marino from his map of Europe because of its small size, even though “the republic of San Marino carries as much authority in its territory as the republic of Venice in its own” (cited in Jacob 2006, 75).

4. State formation and territorialization—and the adoption of an abstract-relational perspective regarding one's homeland—also require participation from “below,” which makes the entire process much more contingent and unpredictable. See, for instance, discussions on the potential for countermapping as a means of constructing narratives and land-tenure relations outside the state discourse that presupposes interchangeable points on an abstract grid (Peluso 1995; Sparke 1998; St. Martin 2001; Walker and Peters 2001; Fox 2002; Hodgson and Schroeder 2002) and, more generally, discussions of how mapping as a process involves participation by the map user as well as the mapmaker (Kitchin and Dodge 2007).

5. Todorov (1984) makes a similar argument with regard to the role of written language.

6. This point is elaborated on, with reference to the world of electronic communications, in McDowell, Steinberg, and Tomasello (2008).

7. This deficiency is exemplified by de Certeau's (1991, 120) remark that the mappamundi is “not a ‘geographical map’ but ‘history book'” and is in keeping with his more general tendency to set a dynamic conception of time against a static conception of space (Massey 2005).

8. The decline in ocean-region names during the sixteenth century is discussed later in this article.

9. Kitchin and Dodge (2007) make much the same point when arguing for a focus on “mappings” in which the map is continually reproduced by the user.

10. Shirley identifies 628 known printed world maps from 1501 on (1501 was used as the start date because there are no known printed world maps from the period between 1492 and 1500). Of these, 214 were excluded because the Shirley book did not contain a legible reproduction (in most cases, Shirley had chosen not to reproduce these maps because they were almost identical to maps that he did reproduce). Two additional maps were excluded because, notwithstanding their appearance in the Shirley volume, they were not, in fact, printed world maps.

In addition, because the goal of this project was to document prevailing worldviews of the time, I excluded maps that sought to reproduce some other era's cosmology, and thus I omitted thirteen maps reproduced by Shirley that had been created specifically to reproduce a classical or medieval map type or illustrate a biblical story or a metaphysical principle. Likewise, because my aim was to reproduce changes in the representation of the ocean as an “outside,” I omitted two maps reproduced by Shirley that were explicitly oceanographic in content. Finally, my goal was to trace the evolution of the “Western” worldview, a problematic perspective in that it assumes a “West” that is both unified and isolated. Nonetheless, given that the modern state system (and the idea of the territorial state with an extraterritorial outside) was a “Western” invention, even if it was one that subsequently has been modified and even influenced by non-Western societies (e.g., Thongchai 1994), I somewhat arbitrarily defined Western maps as those that used Latin lettering, leading to the exclusion of a further four maps, one each in Russian, Turkish, Armenian, and Greek.

11. The four libraries utilized were the American Geographical Society Library, the New York Public Library, the Newberry Library, and various libraries at the University of Wisconsin–Madison (including the Wisconsin Historical Society).

12. As Jacob (2006) notes, both de Certeau and Latour discuss the disappearance of ships on world maps, speculating that the ships had been depicted to pay homage to the process by which the world (and the ocean) came to be known and that their disappearance attested to geographical knowledge and its representation being taken for granted. However, the close correspondence between the rise and fall of representations of ships with the rise and fall of marine biota suggests otherwise. A correlation analysis run on the raw data (not aggregated into twenty-five-year intervals) reveals a positive correlation between the number of ships and the number of biota, with an extremely significant p value of 2.220^{- 16}. This correlation suggests that the ships, rather than representing the process by which the ocean came to be known, simply represented the ideal that the ocean was a space of social processes as well as natural features.

13. See, for instance, Gaynor's (2007) study of the Indonesian government's attitudes toward that country's outlying islands, the people who inhabit these islands, and the ocean that connects (and separates) Java from these islands. See also work that applies Agamben (1998) to the U.S. Global War on Terror and its extraterritorial interrogation and detention centers (Agamben 2004; Butler 2004; Gregory and Pred 2007).