Richard Talbert, University of North Carolina, Chapel Hill
Modern Mapping Before Digitization
For most of the 150 years since APA was founded in 1869 it has been harder to make maps than many of us may imagine today in the era of satellites and digitization. I say this not to find fault or make excuses, but to provide context for the somber truth that good maps for our work in the ancient field have been the exception rather than the norm unfortunately.
One basic obstacle faced by all mapmakers for the dissemination of their work until as recently as the mid-1990s was that the product really had to be a print one, normally on paper. And, whatever the choice of production process, it was inevitably laborious and expensive; corrections and changes would be difficult to introduce, if not impossible. Copper-plate engraving was thought to give a map the sharpest definition, and so until World War I this remained the process of choice for publishers seeking top quality, if they could swallow the high cost and be patient. To engrave a large map would typically take two specialists several months, one for the linework, the other for the lettering. As likely as not, any coloring required would then be added by hand.
Post-World War I, film-based photographic processes were developed and became standard. While more streamlined, these also were laborious and expensive, and still involved manual work. A map might well comprise, say, 40 different film-layers or ‘separations’ which, captured in a single shot, made up the completed sheet. Landscape elevation was shaped by peeling and scribing, and typed place-names were gummed onto the relevant separation, even letter by letter in the case of, say, the spaced-out name of a wide region. When work began in 1989 for what became the Barrington Atlas, this photographic process was still the only option, and it was used eventually for 43 of the 102 maps in the Atlas. But in the course of the 1990s the mapmakers (Geosystems) were to abandon it completely. The Barrington maps produced in this way would be their last; their photo lab was closed as soon as final corrections were made and these 43 maps dispatched to the printer early in 2000.
The decision was taken in the mid-1990s to produce the other 59 Barrington maps digitally. I should stress what a risky, major shift this was on the part of all concerned. For representing physical landscape, it relied heavily on the first edition (just released) of the U.S. Defense Mapping Agency’s Digital Chart of the World – a product that turned out to have plenty of defects. A graver concern was that digital mapping, despite its impressive advances, was still a very fragile growth then, with nothing like the capacity or versatility that we take for granted today. Around the same time Geosystems risked other forays into digital mapping: these proved premature, and led to heavy losses. The great Historical Atlas of Canada project made four successive attempts to switch to digital production, the first three of which all failed and nearly sank the project. Some over-confident ‘experts’ were prone to say that the Barrington ‘ought’ to have been a digital atlas, not a print one, and that its entire compilation and production ‘should have been’ computerized. This is not a realistic claim, I’m afraid, unless work were to be suspended for several years. For sure, map production in the early 1990s was less taxing than it had been before World War I, say; but still, the removal of the major obstacles – and with it the transformation of mapping potential that we’ll hear about in the next three papers – is a phenomenon of only the past 20 years or so, really post-Barrington.
By now you may be feeling that I’ve put the horse behind the cart. Agreed. To produce a map brings challenges, but first obviously you must gather items of relevant data and create from them the map you have in mind. At any period before aerial photography (which began during World War I), let alone satellite imagery, achieving this goal presented challenges of its own. Shorelines could be mapped with rather less difficulty than inland regions; this was because, for much of the Greek and Roman world that we study, the British Admiralty fortunately secured authorization, funds and equipment to conduct the required surveys during the first half of the 19th century, and then published the charts that were compiled.
Mapping inland was liable to be a sterner proposition, however. By the late 19th century tools and instruments for measuring distance and height were available, certainly. But to carry out the kind of triangulation survey needed to record an extensive landscape accurately was a slow, costly operation that really only the state itself could contemplate and was prone to regard as its exclusive responsibility. Foreigners working independently were most unlikely to gain authorization to undertake it – and even a state’s own surveyors might well meet hostility from locals who shrewdly suspected that the aim was greater exploitation and control.
During the late 19th century it was common enough for states in what had once been the Greek and Roman world to be stretched for funds, and to be facing pressures far more urgent than commissioning a survey; so this work was not done, or only in a very limited way such as surveying for the construction of railroads. Even those classical scholars interested in actively mapping the Greek and Roman world did not think in terms making maps to show this ancient world exclusively as Barrington does, when there weren’t yet maps even of the contemporary landscape. Rather, they set out to contribute to the mapping of that landscape, with particular attention to marking ancient settlements, monuments and other remains. By the same token, these scholars did not share Barrington’s concern to return the landscape of their time back (where possible) to how it may have looked in antiquity. As I say, even the contemporary landscape was yet to be grasped, and in any case there was no strong sense of it being different for the most part from that of antiquity (this is a post-World War II concern).
In these circumstances, therefore, the 24 remarkable maps of Attica at 1:25,000 produced for the Deutsches Archäologisches Institut by Johann Kaupert’s team of German military surveyors near the end of the 19th century are exceptional – the work of foreigners in Greece, who had the state’s permission and external funding, and who focused on antiquity. But of course the area covered was relatively limited. Another exception was the surveys made by the (British) Palestine Exploration Fund from the 1860s. Comparison of its results with those of the short-lived American Palestine Exploration Society in the 1870s underlines how exceptional. The Americans had no theodolites and few other instruments, and so (explains a report to the British Fund in 1880): “The American maps are … valuable in giving general information concerning the roads, rivers and principal topographical features but they are not strictly accurate as to distances and smaller details”. They amounted, said the report, to a “reconnaissance” rather than a “trigonometrical survey”. [F. Cobbins, Palestine Exploration Quarterly 137.1 (2005), 15]
For long frustrating, and over a vast area rich in classical antiquities, was the situation in Asia Minor, where the Ottoman authorities would not, and really could not, make a survey. Small parties of foreigners might gain permission to explore and make observations, but this could prove tough work upcountry – with transport rudimentary, terrain punishing, and locals often unco-operative as well as speaking difficult languages. By the late 19th century the renowned German cartographer and classicist Heinrich Kiepert was at least bringing some informal co-ordination to Westerners’ piecemeal efforts in Asia Minor. Several of these travelers commissioned him to create maps from sketches and notes they made en route. Among them was the American classicist and epigrapher John Sitlington Sterrett, whose published record of journeys in 1884 and 1885 includes two maps by Kiepert for each journey. As you see, there’s no mistaking their spotty character. This is again evident in the 15 sheets of the Specialkarte which Kiepert published in 1890-91 as a synthesis of his knowledge about Western Asia Minor. The widespread emptiness suggests that a scale as large as 1:250,000 was prematurely ambitious in fact.
The decisive idea to map the entire classical world specifically, and at a uniform scale, was not proposed till 1928 when the eccentric archaeologist at the British Ordnance Survey, Osbert Crawford, put it forward. It was feasible by then in fact because, on the one hand, an international project to map the contemporary world uniformly at 1:1,000,000 scale was progressing well – the International Map of the World; and, on the other hand, its production method was the film-based photographic one. So, Crawford realized, in order to map the entire Roman Empire (as he dared to propose), which covers about 50 IMW sheets, you could cherry-pick the separations: discard all those showing modern settlement, railroads, etc; retain the ones for physical features; and then create new separations for marking Roman settlements, roads, limites, etc. This project, headed by Crawford and sponsored by the International Geographical Union, came to be called Tabula Imperii Romani.
A brilliant scheme in principle. But how feasible was it in practice ? Well, there was initial enthusiasm to draft provisional sheets – shown particularly in Italy, also in France and Spain, as well as by the British Survey in Egypt and the Ordnance Survey itself. But Crawford struggled in vain to gain agreement on just what features were to be mapped and in what styles, and he had no power of editorial control; he also had no funds to bestow or withhold. World War II halted progress, and Crawford retired; then there had to be a fresh start. When it eventually came, in the mid-1950s, sponsorship switched to the International Union of Academies. This was a boost, but the Union too offered no funds, and increasingly progress was stalled by its insistence (understandably enough) that the mapping for a country could only be done by that country. For one reason or other, some countries showed no interest and weren’t even Union members – including Turkey, with its territory straddling as many as eight IMW sheets. In several other instances, where an IMW sheet includes territory of two countries or more (as often), neighbors refused to collaborate; so either the sheet wasn’t made, or one country just went ahead leaving the other’s territory blank !
Meantime, with TIR still notionally progressing but in fact almost at a standstill, there was no comparable mapping of core areas of the classical world being commissioned otherwise. So right up till 2000 in fact, when the Barrington Atlas appeared, the ‘best’ (in inverted commas) maps of ancient Greece or Italy or Spain or France were mostly still ones (seldom large-scale) made by Heinrich Kiepert or his son Richard before World War I, either for fascicles of the Corpus Inscriptionum Latinarum or for an atlas (never completed) Formae Orbis Antiqui – all rare items of course, hard to find let alone to acquire, and very, very out of date.
APA/SCS deserves full credit for recognizing this jam and initiating steps to break it. In 1980 an APA ad hoc committee formed to review basic research tools for the ancient field, and to recommend how APA might assist in improving them, dared to describe the state of tools for cartography as “utterly disastrous”. Its report went on: “A concerted attempt to produce a uniform series of maps which show both the topography – with all the sophistication of modern cartography – and the ancient toponyms – with the accumulated knowledge of classical scholarship – would be immensely valuable.” [R. S. Bagnall (ed.), Research Tools for the Classics (Chico, 1980), 27]
Action soon followed with the formation of an APA Classical Atlas board; but do not imagine that a smooth path then leads directly to the appearance of Barrington 20 years later. Alas no, because first this board struggled for a long time – as Crawford had done – over just what should be mapped and how, by whom, to what schedule, and (as always !) where the funding would come from. At the end of 1987, with still nothing agreed or achieved and no funds raised, the board resolved to start afresh and it invited me to become project director – in part no doubt because far away in the U.K., with the help of 25 colleagues, I had produced a modest textbook classical atlas on a shoestring budget. But I had only moved across the Atlantic in 1985 (to Canada), and I was barely even aware of APA’s far more ambitious project. My chances of turning it into a success seemed slim to none, so I realized how crazy it would be to accept the invitation. Yet this was clearly an opportunity with exciting potential to be unlocked somehow, and I fell for the bait.
You can find records of how this bleak zero-point gradually led to the Barrington Atlas and its Map-by-Map Directory published by Princeton in 2000, if you dip into my new book Challenges of Mapping the Classical World (Routledge). Much in these records was written – please understand ! – with an eye to funders and to maintaining their confidence, so naturally accomplishments and progress are highlighted, with much less on details or on setbacks, failures and crises, of which there were plenty throughout, believe me. I am painfully aware that the principles I adopted, and the choices I proceeded to make about what should be mapped and how, by whom and to what schedule, were by no means the only possible options, let alone the only ones worth pursuing. But I would still maintain that they offered the most useful and viable way forward in the dire circumstances at the time, and that Barrington retains long-term value, as was always the hope.
The plan, in essence, was for a large-format bound atlas (not sheet maps) to be created in close partnership from the outset with a professional map producer and a leading publisher, for completion within a decade (energy wouldn’t last longer); its scope to be the entire classical world post-Bronze Age to Late Antiquity, but ‘only’ (ha, ha) the physical and cultural landscape (no modern features or thematic maps or city-plans); apart from overviews, at really just two scales – 1:500,000 for the core, 1:1,000,000 for the periphery, using separations from the American and British aeronautical charts that superseded the IMW; map compilation to be by 70 or more top regional experts worldwide, who would be paid and obliged to follow the editor’s instructions, as well as to document everything briefly for the Directory, and submit to peer review; the design would deliver clear, visually appealing color maps, accessible to non-specialists; and the atlas list price would be within a range that private buyers could at least consider.
In retrospect now, Barrington as produced in accordance with this plan has patent limitations, to be sure. In particular, it is not georeferenced, and no map’s coverage can be greater than a doublespread obviously; so there are constant page-breaks, and seamlessness isn’t an option. As to content, it’s also not ideal to include everything from the Archaic period to Late Antiquity on just a single map. Ideally, a region should have a suite of maps for successive periods; but then think how costs would balloon, along with Barrington’s bulk and price ! The fact is that this limitation and others derive from obstacles imposed by the production process in the pre-digital era, as already explained. Barrington’s timing, however, turned out to be uniquely fortunate, because advances in digitization quickly made it possible to start overcoming these obstacles and do much more – which leads to the next paper by Tom Elliott. Thanks in the first instance to APA’s lead, and then not least to international teamwork generously funded, the state of cartography in the ancient field has now been transformed from “utterly disastrous”.
Follow Us!