Boundaries are mental constructs essential to human societies. We use them to make sense of the world, but boundaries both bind and blind us to our world. Boundaries discipline and divide, order and organize, include and exclude. Boundaries allow the otherwise inexplicable inequalities of gender, class, and race to appear normal. As with the broader society, so too with the modern university, for without our disciplinary boundaries of distinct epistemologies and methodologies present university structures would make little sense.[1] While this division of labour has permitted unprecedented knowledge acquisition, these disciplinary divisions also camouflage many of the inequities that characterize university life.[2]

This paper explores how a collaboration between geographers and historians to create an historical geographic information system, the Montréal l’avenir du passé, or MAP project, came to challenge disciplinary boundaries. Through our analysis of digitalized period maps and routinely generated nominal series, we began to conceive them differently. This encouraged us to rethink in a more democratic, inclusive and open manner many of the boundaries and conventions of our respective disciplines. The interdisciplinary lessons learned through this subversion of existing structures are of potential interest to progressive scholars in both the humanities and social sciences.

L’avenir du passé means the future of the past, and we chose this title because we envisaged the MAP project as a tool to allow people to see their past anew. The project is developing a multi-layered geo-referenced historical database for Montréal (Sweeny and Olson 145-54). Each layer consists of an historical map dynamically linked to a variety of period sources. Based on the availability of high quality maps for Montréal, we selected six years: 1825, 1846, 1880, 1912, 1949 and 2000. I will be drawing on my experience in working with the three nineteenth century layers in order to explore how that past might help us to see our future anew.

I will be using the significance of boundaries in nineteenth century Montréal to critique the affects of disciplinary boundaries in contemporary intellectual life. This critique consists of three parts: a brief reflection on the historical character of geographic information systems is followed by a linked discussion of the temporal character of maps and the spatial character of sources. But first, a general observation: my focus is on history and geography, two fields where respecting the fundamental boundary between manual and mental labour characterize almost all large research projects. Early on, our project decided to adopt a different strategy. All the joys and the frustrations of data entry, warping, proofing, linkage, and analysis were shared. I learned many of the most important interdisciplinary lessons through this democratic praxis.

1.0 What is and Why?

A geographic information system (GIS) is a digital representation of part of the earth’s surface, where the part is integrated into a system of global co-ordinates and the key visual elements that compose the system are objects with definable properties. Generally these objects are organized into transparent layers of vector graphics, but it is not unusual for the bottom layer to be a raster image that has been "rectified" to the co-ordinate system.[3] These virtual systems are flat, two-dimensional representations of a three-dimensional sphere, although 3-D drawings may give them the illusion of perspective. A GIS mobilizes centuries of cartographic conventions with the latest in computer technologies to ensure the illusion of accuracy. What we see is most definitely not what is there, nor is it even a reasonable facsimile. It is a radically different representation of reality, one that qualitatively reorganizes diverse social observations into a virtual world whose very detail is essential to the success of these gigantic contemporary trompe-l’œils.[4]

This inherent complexity to what is and why is further compounded when we use a GIS to explore the past. The ontology of geographic information systems is antithetical to historical understanding. Moreover, this antithesis is integral to these systems being in history. As a quintessentially advanced capitalist technology, a GIS encapsulates our particular cultural understandings of time and place through its very denial of both the significance of time and the particularity of space.

The very language we use to describe the rectification of maps to a particular projection and corresponding co-ordinate system betrays our implicit awareness of this contradiction, just as the methods we use tend to mask it. After all, the first step in constructing an historical GIS is to warp a map from the past to fit current co-ordinates, whereby we anchor a radically modified image within a new alien spatial order. All the while, we reassure ourselves by continuously checking how accurate the warped image will be. As this oxymoron of an accurate distortion suggests, present perceptions of what constitutes an appropriate spatial order always trump the historically grounded spatial perceptions of the original map. The simple positioning of north at the top of an historical GIS of Montréal illustrates how this spatial restructuring works to deny the significance of long established ways of knowing.

The cultural convention of placing north at the top of a map developed in early modern Europe. Montréal, founded in 1642, was the western most outpost of seventeenth century Europe’s invasion of North America, so maps of the region always respected this convention, but until very recently city maps applied it in a most unusual way. For three hundred-forty years, Montréal was shown as running parallel to the St. Lawrence River, which here flows north by northeast. In both popular and administrative language, however, the city was imagined as lying along a west to east axis running parallel to the river, with the area away from the river, at the top of the map, considered as North. People were not ignorant of their cardinal points, but they understood Montréal’s place in the world differently: in terms of a spatial order centred on the river. West meant upstream and the up-country of le pays d’en haut, while East meant downstream and the port of Québec. As the city grew to encompass most of the island, streets running parallel to the river were divided into their "western" and "eastern" segments at Boulevard St. Laurent, a fabled "Main" street that runs west by northwest out of the town centre. The post-war suburban boom in the Anglophone "West Island", located to the southwest of the city, and the Francophone "rive sud" or south shore, which lies to the east by northeast, shows the continued cross-cultural strength of this popular understanding.[5] Indeed, the first maps to impose a "proper" northern orientation on the city did not appear until the 1980s. Significantly, they were maps of the city’s Metro system designed for display in the subway cars, where, below ground and without any visible reference points, a newly imagined order could be more easily imposed.

As this story illustrates, human perceptions of space are culturally grounded in complex ways. Now geographic information systems assume the primacy of a knowable, "real" world, unencumbered by such idiosyncratic imaginings. In practice, however, this positivist assumption of a real world that is ontologically distinct from culturally laden understandings soon proves unworkable. On all of our nineteenth century maps of Montréal the representation of actual physical objects, be they churches, courthouses, homes, outhouses, wharves or workshops, failed to match the care that had been given to the representation of imaginary lines in space. Accurate rectifications required an almost exclusive reliance on property lines as our anchoring, reference points. Property lines were more accurate because they were more important than the built environment. This importance was both historical and practical. Property relations preceded and determined the actual shape of the city. As this primacy of a socially sanctioned imaginative ordering of space over the "real" world suggests, analytical dichotomies between anecdotal and scientific, imagined and real, speculative and verifiable are themselves more cultural than empirical.

Pragmatic empiricism has for the last sixty years been the dominant historical theory and method in the Anglo-American world (Sweeny, Staples 327-49), so the problem as posed here is not restricted to geography. However, the power to transform that geographic information systems possess acutely poses that basic concern of the human sciences: how do we know what is and why? In stripping away the historical distortions inherent in old maps, and by aligning them to a projection of an empirically verifiable physical world, a GIS offers the promise of an accurate scientific understanding of the past. Thus, an historical GIS works because it is a-historical. Outside of time the spatial co-ordinates act as absolutes. It is this a-temporal character that allows differing images of the past to be brought together, not just in close proximity, but superimposed one upon the other. Conceptually then, geographic information systems realize virtually what the capitalist system has as yet only partially achieved, most notably in its 2.3 trillion dollar a day currency market, and that is the effective annihilation of time and space.[6]

This technology, which promises to bring history and geography together in new and fruitful ways, is itself inherently destructive of the epistemological claims of both disciplines, but that does not mean we should abandon GIS. It does mean, however, that we must abandon the belief in a technological solution to our disciplinary boundary disputes. In the human sciences, to know what is and why is a conceptual problem, not a technological one. In order to develop a properly historical GIS we need to rethink not just the relationship between our disciplines, but the disciplines themselves.

2.0 Maps as Historical Sources

Each of the three nineteenth century layers of the GIS Montréal l’avenir du passé is based on a different type of map. The earliest layer uses an 1825 map by John Adams, published in New York.[7] The second layer uses an 1846 commercial map surveyed and drawn by James Cane, published in Montréal.[8] Along both sides of the original map are a series of engravings of prominent buildings, while at the bottom is a panoramic view of the city as seen from the river. Both the engravings and the view are the work of James Duncan. I am not aware of either Adams or Cane having done comparable work in another city. By contrast, the third layer uses the 42 plates of an insurance atlas for Montréal, one of literally hundreds of such publications prepared by Charles E. Goad & Company.[9]

As old artifacts, these three base maps are all historical in the popularly understood sense of the term, but they are also historical in a more profound sense. Each is in history. Each draws selectively from the past to represent its present and influence the future. This dialogue between past, present, and future is integral to the historical logic of each map. Identifying and respecting the historical logic of a source is essential to the creation of a properly historical GIS. Some aspects of these historical logics are visible to the naked eye, but others only become visible as the result of extensive use of GIS techniques to analyse the contents of each map.

When the Corps of Military Surveyors and Draftsmen was established at the Tower of London in 1805, John Adams was one of this new Ordnance Corps’ first recruits (French 8). He worked on a variety of defence projects until the end of the Napoleonic wars.[10] His half inch to the mile survey of Québec and its environs of 1822 shows a flexible mastery of the various techniques long associated with that corps’ major peace time activity: the survey of southern England and Wales.[11] It is, however, his exceptionally detailed map of Montréal that best illustrates the extension to Lower Canada "of Britain’s calm felicity and power."[12]

Drawn at a scale of 200 feet to the inch, Adams’ map appears to leave little to the imagination. Along 114 named streets and 83 lanes, Adams exhibited distinctly 459 vacant lots and 2,485 lots upon which were built 26 "Public Buildings", 2,393 "Dwelling Houses', and 2,409 "Warehouses, Stables, Barns, Sheds and every description of out Building." I say appears, for it is the very precision of Adams’ hierarchical spatial order that is imaginary.

Figure 1: St Joseph Suburb, John Adams, <em>Map of the City and Suburbs of Montreal, 1825</em>. McGill University: Rare Books Collection.

Figure 1: St. Joseph Suburb, John Adams, Map of the City and Suburbs of Montreal, 1825. McGill University: Rare Books Collection.

This false sense of order, which is only magnified by the illusionary precision of a GIS, operates on two distinct levels. First is the promise of "exhibiting distinctly every property public and private," against which we must set the reality that a survey of this type can only identify those properties that were visually distinct. Adams’ presentation of St. Joseph Suburb illustrates well the larger problem. In this popular class ward, which at two hundred-twenty square metres had the smallest median size lots in the city, 16 "properties" accounted for 45% of all land in the suburb. These properties shared a number of characteristics. On the map they appear undeveloped, they had only limited frontage, if any, onto a road, and they abutted onto one of the streams that ran through the neighbourhood. Developed properties had clearly distinguishable features, land lying fallow rarely did and so, quite understandably, the natural break of a brook or stream became a property marker.[13] By the same measure, fences, hedges, and other agricultural improvements within the gardens, orchards, and farmlands of the city’s outskirts might well be mistaken for boundaries.

The second level is more conceptual than practical and it concerns Adams’ presentation of the built environment. Not just the choice of buildings, ranging from two of the city’s 56 schools to only 8 possible outhouses, but how they are represented, tells us more about Adams than they do about the city itself. Adams’ classification system presumed that a meaningful distinction could be made between dwellings houses and other buildings. However reasonable for an English officer to believe this normative assumption that home and work constituted separate spheres may have been, it was singularly premature for a thriving artisanal centre like Montréal. Along the town’s main commercial streets of St. Paul and Notre Dame buildings with mixed commercial and residential functions would dominate for another generation. While in the majority of artisanal trades and in petty commerce, marriage and mastery were so closely intertwined that for most households Adams’ distinction would have been seriously misleading.

The historical logic of Cane’s map is also multi-faceted. Dedicated to the Mayor and Corporation of Montréal, a celebration of the city’s commercial past, vibrant present and remarkable future, is the leitmotiv of the whole piece. From a statistical portrait at the top, through the detailed map in the centre, to the panoramic sweep of the waterfront at its base, this Topographical and Pictorial Map has many stories to tell.

The statistical portrait of the city based on a recent census[14] made three succinct points. A clear majority were either "British Canadians" or "English, Irish, and Scotch." Indeed, according to Cane’s estimate of the city’s growth rate, those born outside Lower Canada would easily have outnumbered the "French Canadians" by 1846. Space was at a premium, for the city’s 44,093 residents formed only 6,252 households, an average of seven people per household. Yet, at 1,607, the "Proprietors of property" were in a distinct minority. The conclusion was clear: the city was ripe for major real estate developments.

The choice of subject matter for the eight sketches along the side borders reinforced the implicit identity politics of this singular image of a city that had, after all, become capital of the United Canadas in 1843. Five churches were featured. Entitled the "French Church – Winter Scene", the sketch of the largest church in North America, Notre Dame, foregrounded a peasant’s sleigh crossing Place d’Armes as viewed from the steps of the Bank of Montreal. The "English Church", the neo-classical Anglican Christ Church Cathedral, showed a cariole, while the "Wesleyan Church Gr St James Street" featured strolling pedestrians and a post chaise. One of the city’s five Presbyterian churches "St. Paul’s Church" and the newly completed Irish Catholic parish "St. Patrick’s" were also shown. A cart laden with hogsheads in front of the "Customs House" and a couple being driven in a carriage and two briskly by the "Odd Fellow’s Hall" represented economic and social life. People were clearly on the move, albeit at differing paces, and perhaps directions, depending on their religion and ethnicity. The final sketch showed a view of the city from a mountain estate, whose name was used as the title: "Trafalgar."[15]

Figure 2 Sulpician estate on Mount Royal, James Cane, <em>Topographic and Pictorial Map of Montreal</em>, 1846

Figure 2: Sulpician estate on Mount Royal, James Cane, Topographic and Pictorial Map of Montreal, 1846. BANQ.

Drawn to a scale of one foot to the mile, from the densely developed central core to the richly detailed slopes of the mountain, the map itself is a work of art. The attention to detail in the riding paths of the suburban estates, the evocation of formal gardens, orchards and parklands, and the elegantly named terraced housing all suggest a city of considerable affluence. Only 135 of the map’s 6,327 buildings are identified: 44 terraced homes, 25 manufacturing facilities, 16 military buildings, 15 church buildings, 8 state buildings, 7 schools, 7 emigrant sheds, 5 markets, 3 hospitals, 2 hotels, and 2 monuments. As this might suggest, in the city of James Cane the primacy of the public sphere as conceived by Adams had been eclipsed by private initiative.

Nowhere is this more in evidence than in Cane’s representation of land use. His map highlights two important processes: a more intensive use of the land and a much greater commodification of land. Table 1 shows a significant growth in the number of buildings, but of course the population had doubled.[16] What is truly remarkable is the nine-fold increase in the number of both large and very large buildings since 1825. If Cane is to be believed, then 66 buildings with a footprint as large, or larger, than the Anglican cathedral had been built in the intervening 21 years.

Table 1

The size of buildings in Montréal in 1825 and 1846

Number in 1825

Size of the building’s footprint

Number in 1846


Smaller than 60 square metres



From 60 to 100 square metres



From 100 to 250 square metres



From 250 to 500 square metres



From 500 to 1000 square metres



1000 square metres or larger





Such a remarkable growth in the size of buildings did not necessarily mean a more intensive use of land. The city could simply have spread out. As Table 2 suggests, on the whole this was not the case. In contrast to the profile of vacant lots, there was relatively little change in the size of built lots. There were two reasons for these slight differences in the size distribution of built lots: the first reflected a real change in urban land use, while the second reflected the equally real constraints on Cane as he surveyed the city. By 1846, numerous suburban villas had been constructed in St. Antoine and St. Lawrence wards on land used for orchards and market gardens in 1825. These lots were substantially larger than the norm. Indeed, lots in several of the suburban terrace developments were large enough to rank in the top quintile on Table 2.

Second, James Cane did not enjoy the felicitous power of a John Adams. He knew how important property lines were, because this is what often impeded him from doing a proper survey. Cane did not have the right to tramp across private property, so he faced a serious constraint. All too often he had to survey private land from the perspective of the public road. Smaller structures behind the principal building on the lot would have been frequently missed. In the older parts of town, where buildings were flush to the street and private cart ways were the only points of entry into the interior of a block, his ability to distinguish between properties and even to draw accurately the outlines of the buildings was greatly reduced. These constraints created many, perhaps most, of the very large new buildings[17] in the older parts of town, and they explained why the number of smaller buildings was so stable despite a doubling of the population and an increase in household size.

Table 2

The quintile distribution of lot sizes in 1825 and 1846

Size of lots with buildings

Size of lots without buildings

1825 (# = 2485)

1846 (# = 3370)

1825 (# = 459)

1846 (# = 2940)

240 sq m. or less

240 sq m. or less

400 sq m. or less

360 sq m. or less

350 sq m. or less

360 sq m. or less

810 sq m. or less

460 sq m. or less

490 sq m. or less

530 sq m. or less

1770 sq m. or less

600 sq m. or less

820 sq m. or less

970 sq m. or less

5150 sq m or less

850 sq m. or less

More than 820

More than 970

More than 5150

More than 850

By contrast, the changes in the size of vacant lots were not an illusion, or at least not the same type of illusion. They were the product of a speculative real estate boom without precedent in the city’s history. Land values in the city quadrupled between 1825 and the mid 1840s (Sweeny and Hogg 44-53). In all of the older suburbs, farmland was being subdivided into hundreds of residential lots and, as the distribution on Table 2 clearly shows, the presumed market was largely petty bourgeois.

Cane’s map did not coincide with this boom; it was an integral part of it. This complex promotional effort presented future developments as current realities. It did so with a confidence characteristic of the peak of all capitalist cycles: this time is different, the boom is real, and the bust will not come. In support of this confidence game, James Duncan used the panorama of the town to draw a history lesson.

The entire sweep of the city’s waterfront from Windmill Point in the west to the new Gaol in the east is laid before us. Pride of place is given to the new municipal market, complete with a portico that would not be built for another decade, set against a majestic Mount Royal that evokes Arthur’s Seat in Edinburgh. Stretching into the distance to the west is an almost unbroken row of four story commercial buildings. Dozens of vessels crowd the harbour, but visually the eye is drawn to four in particular. A steamboat crosses from the south shore, while four people paddle their canoe in the opposite direction. This temporal allegory is not unique. Downstream from the market is a timber raft bound for Québec and the British market. A present that will soon be past, because the largest vessel in the harbour, proudly flying an ensign the size of its water wheel, is the steam dredger that in the following year would clear a channel through the shallows of Lac St. Pierre, thereby opening the first St. Lawrence seaway. Ocean going vessels would then be able to make it up river as far as Montréal under their own steam, assuring the city’s future growth.

The Goad atlas is the most detailed of our nineteenth century maps. Drawn to a scale of one hundred feet to the inch, it is the first to provide street numbers, proprietors, construction materials, and cartways. The 42 plates of the atlas outline 12,192 distinct properties and identify 23,162 buildings as being stone or brick, 15,816 wooden frame buildings, and 1059 buildings of mixed materials. This description, while accurate in the sense that it accurately describes how the city is represented in the atlas, errs inasmuch as it suggests that the organizational principle of the atlas was actually rooted in the built environment.

Insurance atlases record information according to an established template, but towns are not built to a template,[18] and, by 1880, in significant ways Montreal did not conform to the industry’s template. The standard form of housing in both working class and petty bourgeois neighbourhoods had become the tenement, a multiplex dwelling where each floor was either a separate housing unit or one divided into two housing units that shared a common stairwell to the outside. Therefore, most buildings had more than the one civic address recorded on the atlas. Furthermore, although the cadastral numbers figure prominently, the atlas neither privileged individual lots, nor, despite appearances, did it accurately describe the built environment.

Proprietors were recorded on the map by having their names written over the properties they owned. Fire insurance was of course sold to the proprietor, not to a building or a particular city lot, so the primacy accorded the proprietor makes sense from the perspective of the insurance companies. Concretely, however, this meant that the organizational principle of the map is neither the actual physical world, nor a cadastral grid, but rather a particular understanding of the social reality controlling that world. Under the Common Law married women had no rights to real property in the nineteenth century, but such was not the case in Québec where community of property was the default marriage regime. Women, married or not, could also own property in their own right, and this was often the case for land that had been inherited. Thus, women owned substantial amounts of property in Montréal, but they only rarely appeared in the atlas (Sweeny, Property and Gender 9-34). On the other hand, husbands, who most often only shared in a community of property with their wives, appeared as sole proprietor. As this suggests, multiple interests in the same property were almost never properly recorded. Usually a single male appears as if he were sole proprietor. Thus, the information about property ownership only makes sense in light of the shifting and highly cultural gender boundaries of the late nineteenth century.

These atlases were colour-coded to identify differing types of building materials. The two major categories were brick or stone and wooden framed. Construction techniques in late nineteenth century Montréal were not, however, so simple. Stone buildings used wooden joists, beams and floors, and they generally had metal clad roofs. In 1880, almost the entire central business district would have been of stone construction. Brick was much more common in housing construction. These brick buildings were almost invariably built of deals, wooden planks 2 ½ inches thick, with wooden joists, beams, floors, and increasingly wooden mansards covering most of the top floor and tarred flat "Boston" roofs. In Montréal, "brick" buildings were merely a façade. The type of wooden frame house construction that characterized so much of the American Midwest was not unheard of in Montréal, but it was usually reserved for the sheds adjacent to or adjoining the back of the brick tenements. In Goad’s atlas of Montréal this wooden frame category was used for all buildings not made of stone or clad in brick.

Each of our base maps was a product of its time. Adams’ map reflected his imperial vision. Cane promoted a qualitatively new relationship to urban land. Goad standardized the world for his corporate clients. Their differing forms reflected these qualitatively differing visions. Adams’ map was drawn for colonial officials, to whom it offered a sweeping strategic assessment of the town and adjacent canal in a single glance. Cane’s map was drawn to promote real estate development; with its heady mixture of proposed sub-divisions, statistics, and pictorial images, Cane's map promised a radiant future of capital accumulation. Goad offered no such over arching motif, for with its 42 plates it was an administrative tool designed for clerks and insurance agents. Indeed, the conceptual parallel between the development of neo-classical economics, with its focus on incremental change, and the periodic pasting of revisions to the plates was striking; both were tools for managing mature capitalism.

Each of these three overarching logics shaped in differing ways the specific contents of the maps. These influences are not minor distortions that can be easily corrected in order to have an accurate view of the past. We cannot culturally rectify the contents of these sources, nor should we even try, for the form and contents of an historical source are intrinsically related to each other. Recognizing and respecting the diverse historical logics of past spatial representations allows us to enter into a dialogue with these maps as historical sources. Listening carefully to what each map has to say helps us to understand historically the complex interplay between time and space.

3.0 Historical Sources as Maps

The great advantage to being an historian of nineteenth century Montréal is the extraordinary variety of the available primary sources. We not only have all the "routinely-generated" nominal series of census returns, city directories, and tax rolls that are typical for an industrializing North American city, but also exceptionally rich church, manorial and court records, and, most importantly, more than a million notarial deeds. Historical databases based on these sources that I first created for use in teaching have become components of the larger MAP GIS (Sweeny, Past in the Present).

As I constructed these pedagogical databases it became increasingly clear that each of these historical sources had significant spatial components. At times, space actually determined the internal structure of the source. For example, a manuscript tax roll for the city in 1832 listed property owners of each street by ward in the order the evaluator, Jacques Viger, encountered them as he walked down one side of the street and then back up the other (Sweeny, Roll 1832). Since the roll was used to establish who met the property qualifications for voting and municipal elections were organized by ward, Viger listed a proprietor only once in each ward; all of his or her other properties in the ward were added to the first entry as Viger came to them. Even the exception to this rule had a spatial logic; when a proprietor had back-to-back properties facing on to another street, the adjacent property was listed immediately.

The first modern census of the island of Montréal, conducted in 1825 by Jacques Viger and Louis Guy, illustrates how important spatial considerations were to people at the time (Sweeny, Census 1825). On Viger’s personal copy of the census returns he literally enumerated each separate building in order to determine how many households resided in each building. The patterns constitute a social geography of the urban landscape. Now Viger was a pioneering social scientist, and his analysis of the 1825 census returns was the first major sociological study of a North American city; nevertheless, spatial dynamics drew his attention because they were key to so many different aspects of life in Montréal. Space was a defining aspect of the tensions between agency and constraint that generated these historical sources.

People were constantly on the move; indeed the question that has increasingly pre-occupied me is: how did structure emerge from this state of continuous flux? Two early city directories compiled by Thomas Doige in the fall of 1819 and the spring of 1820 illustrate the scale of change (Sweeny, Doige 1819; Doige 1820). The second edition had a 14% increase in the number of entries, from 2446 to 2798. Yet the movement was much greater than that. One in four entries from 1819 did not appear in 1820, while one in three entries in 1820 were new. Only 56% of the people and firms listed appeared in both directories. Doige cited "the unprecedented number of removals this year" as reason for the new edition, and a spatial analysis of the two directories revealed an historically significant pattern to this movement. In 1819, 1,471 people were listed with both a complete address and an occupation. One in six were not listed in 1820. For one in seven, somebody else was listed at the address. This large-scale movement, 309 people replacing 208 at 183 addresses, was not random. A third of the time, at least one of the new occupants declared the same occupation as the former resident. This was not the only scale of importance: many chose to make micro-level adjustments. Between the two directories, 104 people and firms had moved to a new address on the same street.

Space is not only physical; it can also be social, cultural, and structural. In both directories, Doige listed people and firms in alphabetical order. Such a spatial arrangement may now appear normal, but since it did not respect social rank or status it was deeply offensive to many powerful people in pre-industrial Montréal. Gentlemen and gentlewomen did not appreciate being treated in such a democratic manner and their influential opposition explains in large measure why the directory failed to find commercial success.

In a similar manner, the internal organization of monetary protests, a type of notarial deed, allows us to map credit networks. If a promissory note, draft, or bill of exchange was not honoured when due, the holder of this credit instrument could protest for non-payment. From their inception, banks in Montréal made their protests a matter of public record by having them notarized (Sweeny, Bank of Montreal & Bank of Canada). These deeds contain not only a copy of the text, but also a list of everyone who had drawn, endorsed, or discounted the note or bill. By inverting the order of these lists – that is, starting with discounter at the bank and going back up the chain of credit – it is possible to identify the informal networks that controlled access to these exceptional pools of credit. During the first worldwide capitalist business cycle, 1820-27, only a third of the 1136 people or firms identified in the 1500 notes protested by the Bank of Montréal had direct access to the bank. These networks of power evolved over both space and time.

4.0 So What Does this Mean?

As first conceived, the GIS Montréal l’avenir du passé was conceptually conservative. Our plan was simply "to people" geo-referenced maps by linking them to a variety of historical sources. Thus, the maps would provide the spatial representation for temporally specific data drawn from a variety of nominal series. This approach respected disciplinary boundaries. Indeed, it sanctioned an academic division of labour, with historians mining their databases and geographers rectifying their maps, each group secure in their disciplinary cubicles behind walls of conceptually distinct software.

As our work progressed, however, it became clear that both the historical logic of the maps and the spatial dynamics of the sources needed to be respected. This required a qualitative change in how we conceived of our system. Instead of a single fully integrated, multi-layer project, our historical GIS is evolving towards a series of distinct sources articulated to, but not limited by, their period’s geo-referenced map. Each historical source is a standalone database, with its own user interface. Each source is fully indexed with its own Boolean query engine and integrated user’s guide. It will be possible to create coherent partial datasets from each source, but it will not be possible from within our software to isolate fields from the essential context provided by the whole record. The maps themselves are treated as standalone programs, consisting of the rectified images of each period map, its associated shape files and databases, and an historical guide. Specific applications linking particular sources to the relevant map are then developed.[19] This design is modular, open-ended, and cumulative.

The epistemology this design promotes is a radically different way of knowing than the disciplinary norms in either history or geography. The ontologically distinct nature of each source is recognized, because its historical logic is respected and explained. Thus, there is no presumed superiority accorded to either an historian’s "expert" synthesis or a geographer’s scientific model. Instead, the user engages directly with diverse and contradictory digital representations of the past. This democratic design does not just accommodate difference; rather, it actively encourages subversion of existing academic boundaries by allowing people to explore the multiple, indeterminate, and contradictory nature of change through time and space.

This rethinking of boundaries is itself historically bound. Just like the sources I have been discussing, MAP’s design has an historical logic that reflects the values of our particular time and place. Although academic authority is most certainly questioned and, I hope, undermined by a more democratic ethos and practice, this process of emancipation remains fundamentally individual, rather than collective or social. This is, of course, profoundly ironic; the development of such a complex, highly malleable, and yet structured series of representations of the past can only be achieved through the collective effort of many people. Thus, while we may have overcome certain academic boundaries to understanding the past, our historical GIS does not question the central contradiction of our time: the destructive environmental impact of the individual appropriation of socially created value. Clearly, in order to see our future anew, indeed to have a future, will require much more than just greater democracy in the academy. Our project contributes nevertheless to a most laudable aim for digital studies in the humanities. By encouraging people to understand the historical specificity of differing times and places, it will stimulate some to rethink what is and to question why.

Works Cited

Adams, John. MAP of the CITY and SUBURBS of MONTREAL; Exhibiting distinctly every property public & private the course of the Water Works the River Line in front of the City and the LA CHINE CANAL from its Junction with the Port to the distance of 1½ miles above. Map. New Survey Cartographers John Adams H.P. RoylMily Surveryr and Draftsman. Engraved by Jas D. New York: Stout, 1825. Print.

Broglio, Ron. "Mapping England." The Wordsworth Circle 33.2 (2002): 70-76; "Wordsworth and Technology Mapping British Earth and Sky." Georgia Tech: School of Literature, Communication, and Culture Web. < mappingengland.html>.

Cane, James. A Topographical and Pictorial Map of the City of Montreal. Surveyed and drawn by James Cane. Lithographed by Matthews & McLee’s. The views drawn by J. Duncan. . Montréal: Robt W. S. MacKay, 1846. Print.

CAUT. Almanac of Higher Education. Ottawa: Canadian Association of University Teachers, 2008. Print.

Chesneaux, Jean. Habiter le temps. Paris: Bayard Éditions, 1996. Print.

Curtis, Bruce. The Politics of Population: State Formation, Statistics, and the Census of Canada, 1840-1875. Toronto: U of Toronto P, 2001. Print.

French, Josephine. Tooley’s Dictionary of Mapmakers, Volume I. Revised edition. London: Map Collector, 1999. Print.

Goad, Charles E. Atlas of the City of Montreal, Compiled and drawn from official plans and special surveys. Montreal: Charles E Goad, 1879. Print.

Harvey, David. "Militant Particularism and Global Ambition." Spaces of Capital: Towards a Critical Geography. New York: Routledge, 2001. 158-87. Print.

Hayward. Robert J. "Sources for Urban Historical Research: Insurance Plans and Land Use Atlases: Sources for Urban ." Urban History Review 2.1 (1973): 2-9. Print.

Maps and Plans in the Public Record Office. Volumes I & II. London: Her Majesty’s Stationer, 1967. Print.

Poovey, Mary. Genres of the Credit Economy: Mediating Value in Eighteenth- and Nineteenth-century Britain. Chicago: U of Chicago P, 2008. Print.

Seymour, W.A. et al. A History of the Ordnance Survey. Folkestone: Dawson, 1980. 57-63 and 363-365. Print.

Sweeny, Robert C.H. with the collaboration of Grace Laing Hogg. "Land and People: Property Investment in Late Pre-industrial Montréal." Urban History Review 24.1 (1995): 44.53. Print.

Sweeny, Robert C.H. Sweeny and Sherry Olson. "MAP: Montréal l’avenir du passé. Sharing Geodatabases, Yesterday Today and Tomorrow." Geomatica, 57.2 (2003): 145- 54. Print.

Sweeny, Robert C.H. "The Staples as the Significant Past: A Case Study in Historical Theory and Method." Canada: Theoretical Discourse /Discours théoriques. Ed. Jane Greenlaw, Terry Goldie, Carmen Lambert and Rowland Lorimer. Montréal: Association of Canadian Studies, 1994. 327-49. Print.

───. "The Past in the Present: Part I Epistemological Challenges of Computer-assisted Teaching, Part II, Methodological Reflections on Computer-assisted Teaching." Canadian Historical Association. 1998. Web. < cchi/Doc/CHA98_Sweeny.htm>

───. The 1825 Manuscript Census of Montréal. St John’s: MUN, 1997-2002. CD-ROM.

───. The 1819 City Directory of Thomas Doige. St John’s: MUN, 1997-2002. CD-ROM.

───. Monetary protests of the Bank of Montreal, 1821-27. St John’s: MUN, 1997-2002. CD-ROM.

───. The 1820 City Directory of Thomas Doige. St John’s: MUN, 1999-2002. CD-ROM.

───. Monetary protests of the Bank of Canada, 1821-24. St John’s: MUN, 1999-2002. CD-ROM.

───. A partial tax roll for the City and suburbs of Montréal, 1832. St John’s: MUN, 2002. CD-ROM.

───. "Property and Gender: Lessons from a 19th-century Town." London Journal of Canadian Studies 22 (2006-2007): 9-34. Web. <>

Wordsworth, William. "View from the top of Black’s Comb." The Complete Poetical Works. London: Macmillan, 1888. Print.


[1] See Poovey for a stimulating discussion of the cultural origins of the differing conceptions of value that justify the major split within the human sciences..

[2] For the most recent evidence on how far we have failed in Canada to fulfill the most basic promises of the liberal order see CAUT.

[3] Vector graphics use algebraic formulae to define the boundaries of an image and object orienting programming to specify its properties. They are compact and can be resized without loss of detail. Raster images assign a value to each pixel of a digital image, so not only are these images frequently very large and will break up if you zoom in too much, but they are not objects for the computer and so cannot be assigned particular properties.

[4] A highly engaging parallel is to be found in Bruce Curtis’s discussion of how census takers made a virtual world that was constitutive of the modern state.

[5] My former colleague Alan M. Stewart once encapsulated this complex geography with the observation that when you get off the Victoria Bridge in the South Shore suburban community of St. Lambert you are north of where you started from on Montréal island.

[6] In thinking through this insight of Karl Marx, I have found the penultimate work by the historian Jean Chesneaux and a wonderful reflection on Raymond Williams by the geographer David Harvey to be particularly useful.

[7] MAP rectified a digitalized copy in six parts prepared by the Rare Books Collection of McGill University.

[8] MAP rectified a digitalized copy in four parts prepared by the Rare Books Collection of McGill University, with additional information for damaged areas coming from the copy in the collection of the then Bibliothéque national du Québec.

[9] Charles Goad was a Civil Engineer, and the firm he created in 1875 quickly became to Canadian insurance maps what Sanborn & Co. was to insurance maps of the United States, the standard against which all else was measured.

[10] By 1807 he was drawing the plans for the Ordnance depot in Ballincollig, Cork. From 1808 to 1810 he signed numerous copies of naval defence drawings for both the Newfoundland and West Indian stations. After working on the plans for coastal defences from Wicklow to Dublin in 1810, he is known to have worked on plans for: Cork, Jersey; Falmouth and Plymouth in 1811; Plymouth and Portsmouth in 1812; Exeter Barracks and Eastern Hoe in 1813; and Chatham in 1815. (Maps).

[11] See Seymour for discussion of these methods and the instructions surveyors received..

[12] This description of the Ordnance Survey is from Wordsworth’s View from the top of Black’s Comb. See Broglio for a thought-provoking discussion of the relationship between cartography and the romantics.

[13] I think that in a number of instances Adams actually chose to leave these areas blank. In the creation of a GIS, however, these blank areas within a city block simply become irregular lots, unless one imposes one’s own discretionary reading of the map. How best to respect these choices remains a point of contention within MAP.

[14] Either Cane or his publisher Robert MacKay, who also produced the town’s trade directory, chose to present the census data of 1842 as dating from 1844, the year they were published, which conveyed the impression that the city was really booming; since their estimate of a current population of 50,000 in 1846 meant a population increase of 15% in just two years.

[15] The name of a cape off the west coast of Spain, Trafalgar was the site of the most important British naval victory in the Napoleonic Wars. It ended any threat of a French reconquest of Canada, which explains why – decades before the creation of Trafalgar Square in London – a monument to Nelson, the commanding officer who died in the battle, was raised by public subscription to lord over the city market.

[16] During August 1825, while Adams surveyed Montréal, a detailed census was conducted that indicated 22,500 people lived in town. The 1842 census, carried out at differing times in differing wards over the summer and fall of that year, indicated a population 44,093. A reasonable estimate of the population by 1846 would be 47,500.

[17] When workers reconstructed buildings in Montréal, it was a common practice to use the foundations of the previously existing buildings. As a result, it is not possible to discern whether or not a building was new simply by comparing the footprints of buildings on different maps.

[18] Which is why the companies that produced these atlases also produced detailed plans of specific properties and areas (Hayward).

[19] MAP has released three linked pedagogical applications using ESRI’s Arc Explorer 4 interface. One application mapped attendance at Protestant schools versus burial in the Protestant pauper’s grave for 1875-1885, against a backdrop of ethno-religious data from the 1881 census and rents from the municipal tax roll. Another contrasted the footprint of buildings in the year 2000 with the built environment of 1880. The third linked occupants on the tax roll to households in the 1881 census returns. We have also released Arc Explorer versions of the base maps for 1825 and 1846.