Suggested citation format:
Ciolek, T. M. 2000. Digitising Data on Eurasian Trade Routes: An Experimental Notation System. pp. 1-28 of section 5-122, in: PNC Secretariat (ed.). 2000. Proceedings of the 2000 EBTI, ECAI, SEER & PNC Joint Meeting 13-17 January 2000, University of California at Berkeley, Berkeley, USA. Taipei: Academia Sinica.
www.ciolek.com/PAPERS/pnc-berkeley-02.html


Digitising Data on Eurasian Trade Routes: an experimental notation system

Dr T. Matthew Ciolek,
Research School of Pacific and Asian Studies,
Australian National University, Canberra ACT 0200, Australia
tmciolek@coombs.anu.edu.au
http://www.ciolek.com/PEOPLE/ciolek-tm.html
To be presented at the
Pacific Neighborhood Consortium (PNC) Annual Meeting,
University of California at Berkeley, Berkeley, USA,
13 -17 January 2000
Document created: 18 Nov 1999. Last revised: 11 Jan 2000.

[Please note that the notation system originally proposed in this document has been since extensively modified and improved. For the latest version of the OWTRAD notation system see http://www.ciolek.com/OWTRAD/notation.html]

0. Abstract

The paper describes a low-tech and low-cost methodology for collecting data on Eurasian and other trade routes. It records details of their course, users, chronology, movement speeds and other variables. This newly developed methodology, the so called 'OWTRAD' data transcription system, can be applied to a wide range of primary and secondary sources. It can handle raw information in both verbal and graphic formats. It can do so regardless of variations in the sources' reliability and overall precision. It is geared to the creation of large volumes of standardised and correctable digital data sets. Subsequently, these data can be processed, either automatically or by hand, in order to establish their geographical co-ordinates and make them mappable within the context of ECAI's investigations.
I dedicate this work to my best friend and wife, Irena Goltz. May her Canberra garden, in the third millennium, always be filled with birds, butterflies and stars.

1. Introduction

If social and cultural studies are to flourish, it is essential that they take wholehearted advantage of computers and the Internet, and their unprecedented potential to store, retrieve and process vast amounts of data in real-time (Ciolek 1999). The Electronic Cultural Atlas Initiative (ECAI) (www.ecai.org) promises a major shift in the way social science research can be conducted online. However, until suitable digital data become plentiful and readily available, it is doubtful whether the viability of the project can be tested in practice.
This paper explores the possibility of producing plentiful digital data for ECAI projects. It presents a practical methodology for digital collection, storage and publication of information suitable for geo- and chrono-referencing. Naturally, such a methodology, if it is to be adopted by other researchers, needs to be compatible with existing research practice. Therefore, it needs to: The proposed methodology hopes prove simple and intuitive enough to become a part of the everyday research toolbox. It hopes to be used by anyone interested in the digitisation of data (through time and space) about communication links between various parts of the world.

2. Available information about trade routes

There is not a great deal of systematic research on trade routes. The field appears to be underdeveloped. It is commonly regarded as a minor adjunct to the wider studies of ancient history, economic history or the anthropology of cross-cultural contacts. For instance, an online search database of the Library of Congress' (lcweb.loc.gov) 12 million records reveals the existence of no more than 65 titles dealing with the topic 'Silk Road' and another 69 titles on 'trade routes' in general. Similarly, the University of California's Melvyl database (library.berkeley.edu) points to the existence of some 206 titles dealing with the Eurasia's 'Silk Road', and 264 titles on other trade thoroughfares in the other parts of the world. These figures are in stark contrast with the equivalent data on other topics. For example, in November 1998, the Library of Congress kept bibliographical details of no less than 3050 books and monographs on the rather esoteric topic of Christian and Buddhist monasticism.
The relative unimportance of the trade routes to the modern cultural studies is further confirmed by data from Table 1.
Table 1
Information about trade routes in five historical atlases
                        Maps     Maps showing data
Publication             all       on trade routes
-------------------------------------------------------
Putzger 1963             105       5
Shepherd 1976            270       9
Scarre 1988              127      34
Stone 1989               430      33
Vidal-Naquet 1992        177      25
-------------------------------------------------------
Average                  221      21
-------------------------------------------------------
The above figures indicate that less then 10% of all maps published in contemporary historical atlases show the existence and layout of the movement corridors in use for circulation of people, goods and information. However, the relative absence of institutionalised research on these matters is perfectly offset by the almost limitless wealth of raw, unprocessed information.
Practically, there is an inexhaustible amount of intelligence which casts light on the layout and attributes of the trade routes in the Old World. It can be easily found in a wide range of publications and primary sources. Such materials include histories and ethnographies of various countries and regions, as well as memoirs and travelogues of explorers and adventurers. Additionally, there are also historical studies describing the disposition and movements of various invading and defending forces. Some of these studies also mention various communication lines employed by government couriers and various postal services. Still another category of materials is formed by publications dealing with the Great Religions of the Old World. They provide useful information about temples, monasteries and other holy places, and the associated pilgrimage routes.
On the whole, materials offered by all these sources form four (see Appendix A below) major groups: The common aspect of all these publications is the bewildering variety of ways through which the data in question were obtained and documented.
Some of the information pertains to geographic areas some 3000 km wide; while other is more focused and deals with geographical locations no more than few kilometres apart. Some of the data might be based on personal experience, whereas other information can be derived from the second-hand reports or be compiled from a series scholarly investigations such as original archaeological or ethnological field-work. Finally, the available information, even the supposedly rigorously constructed maps and charts, can be of uneven quality (see Table 2).
Table 2
Problems with the maps published in five selected* historical atlases
--------------------------------------------------------------
Shortcoming                             Percentage of maps
                                      displaying the problem
--------------------------------------------------------------
No time-frame                                     9%
No legend/key to symbols                         12%
Factual errors present                           20%
No lat-long grid                                 53%
No source of information stated                  68%
Incomplete annotations of data                   69%
No scale is specified                            74%
No projection is specified                      100%
--------------------------------------------------------------
Total                                           106 cases (100%)
--------------------------------------------------------------
* atlases analysed: Putzger 1963, Shepherd 1976, Scarre 1988,
Stone 1989, Vidal-Naquet 1992
Src: Ciolek (1999).
Therefore, any attempt to bring the masses of colourful but unruly graphic and verbal information under the aegis of an ECAI study is not an easy task. A shortcut which involves digitisation of the existing map material is a tempting but impractical strategy. There is a number of reasons for exercising care: It is obvious then that a different strategy needs to be pursued. Such an alternative strategy is more cautious, and less hasty. It advocates 'deferred digitisation' of collected data and 'deferred marriage' of these materials with the GIS-style base-maps of the world.
Certainly, the raw data need to be captured and stored with equal ease and equal speed, from both textual and graphic source materials. Moreover, such simple data capture must be carried out regardless of the overall integrity and precision of the interrogated materials. Once the required information is extracted it needs to be put through a series of routine operations. It needs to be collated; brought to a common denominator; verified against the ever-growing body of 'online intelligence' dealing with the major variables; checked for any hidden problems, and, finally, made open to endless and iterative (Raymond 1998, Ditlea 1999) inspection and correction. It is only after all these steps are completed that a data set can be properly geo-referenced and formally incorporated into the framework of other GIS-style, ECAI-compatible data sets.
The crux of this alternative strategy is to embark on an intensive, large scale yet inexpensive data gathering program before more costly and more technically demanding, and time consuming data mapping and data plotting programmes are engaged in.
To use a metaphor, before we start constructing a sumptuous palace or a glorious cathedral, it is useful if we first secure a steady supply of solid, modularised building blocks. This is, however, only possible if we learn ourselves how to shape amorphous lumps of clay into the standard-shaped bricks, and also learn how to produce such bricks quickly, inexpensively and, importantly, in large amounts.

3. Trade routes: an archipelago of linked nodes

This section sets out the conceptual underpinnings of a technique for identifying and collecting data on trade routes. In this case it has been illustrated with examples taken from the 'Old World', by which I mean the combined sea and land masses of Europe, Asia and Africa. Most of the presented information tends to be self-evident, nevertheless it is necessary to make the subsequent terminology explicit. The basic notion used throughout this paper is that of 'trade routes' by which I mean spatially and temporarily persistent, land- or water-based long-distance movement corridors. These corridors connect human settlements via a series of other nodal points, and are used by groups of travellers to transport goods in a manner which minimises risks to their personal safety and security and maximises their commercial profits, while keeping the overall travel time as short as possible.
The above definition is handy. It serves as a template for conceptualising other types of routes as well. For instance 'pilgrimage routes' can be viewed as movement corridors, similar to the trade routes, but ones in which goods and the associated commercial profits are not an issue. In similar vein, 'military routes', tend not to concern themselves with matters of profit and personal safety and, instead, place emphasis on transport of people and goods in the manner which is reasonably secure from enemy's intervention and as speedily as possible.
This definition of the trade routes is influenced by the methodological work of the ECAI Silk Road team (Williams 1999), one which discerns nodes, sites, and routes. The present definition introduces three additional notions, those of links, users and commodities. Together they form the following conceptual model.

3.1. Nodes

All over the world, one can discern numerous nodes, or places of travellers' departure, transit, temporary stopover and, finally, arrival. Nodes are not pre-defined. Any distinct geographical location, as long as it is frequented by travellers, may serve as a node. There are three categories of nodes: (a) inhabited places, such as towns, monasteries and fortresses; (b) various natural features (e.g. anchorages, fords, passes, oases); and finally, (c) man-made constructions (e.g. reservoirs, harbours, ferries). Some of the nodes may function as attractants because they increase the probability that a place (e.g. a market, bridge, the site of an oracle) will be visited by travellers. By contrast, other nodes (e.g. besieged cities, robber-barons' castles, tax-collectors' outposts), may act as powerful deterrents to human traffic. Sometimes the dicision to avoid a given place is based on common sense, sometimes it is quite idiosyncratic. The nodes, of course, differ from each other in terms of their physical size and social and cultural characteristics. All of them are mappable as long as they have been assigned precise chronological and geographical coordinates, calculated from a map, or taken from the growing body of specialist registers and online databases (see also Appendix B below). Some nodes may double as sites (locations where artefacts and archaeological finds have been uncovered) (Williams 1999:585), whereas other nodes may not have any artefacts associated with them.
In terms of mapping, certain kinds of nodes such as inns or road junctions, can be characterised as points on a map. Other nodes, e.g. oases or towns, might be visualised as roundish areas enclosed by a perimeter (a wall, an irrigation ditch). These most typically have one or more central points (a temple, a forum, a well). Finally, some nodes, such as a portage area, can be best depicted as a strip with an invisible spine of minor topographical sub-nodes.
Some nodes may have unique names, since they are perceived as places, that is, distinct locations endowed with social and cultural characteristics. Other nodes, like anchorages and watering holes, may remain nameless. They have a physical existence but they are commonplace or ill-defined, and so have not established a separate social existence. Such nodes usually derive their identity from features of the nearby landscape, as well as from the role they play within topology of a particular communication/transportation network.
The spacing between nodes varies widely. It is always a function of the researcher's decision of what pair of geographical locations will serve as nodes, say, 'A' and 'B'.

3.2. Links

Often the space between two particular nodes remains unexplored and untraversed by people for long periods of time. Of course, in reality there is no such thing as a totally uncrossable stretch of space. As the WW2 memoirs of Rawicz (1997) and Harrer (1955) testify, determined individuals are able to suffer the most terrible privations for several years on end and walk through the most inhospitable types of terrain. On the other hand, such space may become a communication link when it is spanned by customary traffic.
A link is thus a stretch of space commonly used for movement between a pair of nodes. If these nodes are called 'A' and 'B' then such a link can handily be recorded the 'A=B' link. This means that our definition of a link applies not to all possible connections between pairs of nodes, but only to those which are the most prevalent.
Links can vary greatly in length. Sometimes a link is very short. For instance, distances separating two adjacent spots in a city (e.g. the warehouse and the pier) would measure only in the tens of meters. Or, it can be immense, and measured in the thousands of kilometres, like the distance between the Bosphorus and the Gibraltar Strait, at the opposite ends of the Mediterranean.
All links (apart from river and coastal links), by definition, are represented by a straight line. This is because links refer to the generalised, logical connectivity between two nodes, and not to the actual analog trajectory traced across the terrain by the travelling parties. Naturally, for short distances, a straight line is a fairly accurate representation of the real-life path taken by the travellers. In the case of longer distances, the trajectory meanders. However, if we do not have information on intermediate nodes, we must postulate that, on land and open sea, such a long-distance link is a straight line. This assumption is, of course, an expedient but temporary simplification. It needs to be replaced, eventually, with information on the intermediate nodes (as well as associated links). Finally, in the case of coastal and river communication links the mapped line can be assumed to follow fairly faithfully the contour of the shore or the river in question.

3.3. Routes

In other words, large scale distances between a pair of nodes can be defined as a route, that is an unique series of interconnected links, say, 'A=G=E=B=L=P' each of them represented by a straight line between adjacent nodes: 'A=G'; 'G=E'; 'E=B'; 'B=L' and finally, 'L=P' (see the hypothetical map above). This is, of course, different, from the one spanning the distance 'A=G=E=O=M=N=F=C'. The two routes share, of course, the common stretch between nodes 'A' and 'E', and then they go their separate ways.
Routes, by definition, tend to circumnavigate obstacles (e.g. marshes, forests, waterless wastelands, mountains, steep cliffs, fast flowing rivers, rapids). They seek to establish the least physically demanding transit, one which is traversable in all weather conditions, if possible. They also aim to be the safest and logistically most advantageous passage between pairs of attracting nodes.
Most routes are chosen so that they take advantage of the natural configuration of terrain (e.g. valleys, sheltered coastal waters, mountain passes). Other routes (and therefore, links) can be established by artificial means: the levelled or paved roads, tunnels, causeways and canals are the most commonly encountered structures. The physical width of links/routes can also vary. Sometimes, like some of the North Saharan tracks (Lewis 1997:133), they can be ten to twenty kilometres wide. Sometimes, like in the case of the ancient passage through the Cilician Gates (Golek-Boghaz) in the Taurus Mountains in Asia Minor, they can be so narrow that burden animals need to be unloaded (Fuller 1958:97), in order that they squeeze through the narrow defiles.
Generally speaking, there are four basic types of links: Naturally, there may be more than one link bringing together two adjacent nodes. For instance, there may be one major road, as well as a series of locally known trails, shortcuts and paths. As military history illustrates, the outcome of many battles, including those of the Thermopylae in 480 BCE and the Persian Gates in 330 BCE, was decided when one of the warring sides exploited a minor and generally unknown trail (Fuller 1958:228-233, Fuller 1970:55-56).
As we have already noted, a series of links strung together forms a route. Routes can have various topologies, ranging from simple chains and loops, to stars, forks, and lattices. The overall topology is, of course, very much a function of the scale of one's sampling procedure. The larger area of study, the more complex the system of routes it contains tends to be.

3.4. Users

Links and routes are created by the presence of people moving from one node to another. Such traffic of users (i.e. people, animals, and vehicles) pulsates and alternates between the routes in response to the changes in weather, time of the year, local conditions (such as famine, plague, unrest) and overall economic and political circumstances. Various individuals and groups of travellers, such as soldiers (Adams 1976), bandits, raiders, messengers and couriers, pilgrims, mendicants, merchants (Casson 1984, Curtin 1984), refugees, migrants, local tradesmen, businessmen, explorers (Newby 1974) and so forth, can move along a given route in a simple sequential fashion (i.e. one group of people after another), or in parallel with each other, or move unpredictably across the Gordian knot of links.
The users of nodes and routes display a number of attributes. These include the social, demographic and ethnic characteristics of the people traversing a given route as well as the technological, economic and logistical aspects of their activities.
Of the logistical variables, questions of the demand and supply of food, fodder, water, fuel, spare parts, and noncomestibles; the ratio of humans to animals; the movement speeds and the overall endurance; as well as the optimal and maximum carrying capacity of humans, animals and vehicles - all are of key importance. All these variables are closely interrelated and can be represented in the form of an arithmetic formula (Engels 1978). The formula balances daily food and water requirements of people, as well as those of the accompanying animals (saddle and pack) against their physical capacities to carry weights (such as supplies, trade goods, weapons, tents etc.) over extended distances. The formula shows that in the case of overland journeys, one which would involve horses, mules, oxen, donkeys (but not, apparently, bactrian camels or dromedaries), it is not possible for a body of men and animals, regardless how many of them would travel together, to proceed for "more than four days without replenishing its water in a terrain where no water or grain was obtainable" (Engels 1978:63).
Other factors also play a role. These include the ability of travellers to orient themselves and operate in the unfamiliar terrain (Lewis 1997); their ability to move surely and systematically (despite frequent adversities of weather, and presence of natural obstacles) from one node to another, and to negotiate (often without the knowledge of the local language) and secure vital supplies, accommodation, health-care, reliable guidance across a given territory, and essential politico-economical briefs about the developments on next leg of the journey.

3.5. Commodities

The final variable important to the studies of trade routes is that of commodities, the transport of and trade in which comprises the raison d'etre for countless nodes and links. These goods are evidently most diverse, and their exact nature, monetary and symbolic value, volumes traded, as well as impact on the suppliers, distributors and buyers change from one geographical area to another, and also according to the historical period. In a nutshell, the traded commodities included weapons; luxury goods (e.g. gems, precious metals, ivory, arts &craft, perfumes, incense, manuscripts and books); raw and semi-processed materials (e.g. ingots of metal, timber, tar, wax); salt, spices and medicines; foodstuffs (e.g. grain, oil, wine, vinegar, fruit) textiles, tapestries and furs; pottery and glass; animals (e.g. homing pigeons, horses, beasts of burden, circus animals); slaves; as well as a great variety of contraband, including the by now legendary silkworm.
The circulation of goods gave rise to numerous evocative names such as 'Amber-'; 'Gold-'; 'Spice-'; 'Slave-'; 'Incense-' and 'Silk-routes.'

4. The OWTRAD notation system - an overview

A tool which can possibly be used to exploit both graphic and verbal information and produce large amounts of digital data relevant to ECAI purposes is a simple notation. Such a notation, or transcription system aims to summarise any extracted data in a standard, predictable and unambiguously legible format and store them as universally legible electronic ASCII files so than any future corrections, enhancements and refinements can readily take place.
Since the experimental notation system described below was initially developed to handle data about the trade routes of the Old World (i.e. Asia, Africa and Europe), it will be henceforth referred to as the 'OWTRAD' notation. The objectives of the OWTRAD notation are two-fold. Firstly, the notation aims at extraction of standardised chunks of information from a large variety of sources. It does so by creating a predictable sequence of conceptual pigeon-holes which can be easily filled with details for each pair of communication links. If a given item of information is not available, then the notation uses the standard default value. Secondly, the notation aims at being a simple and versatile tool, one which can be used manually, or as a part of ones' use of a typewriter or computer keyboard.

4.1. Basic requirements and elements of the proposed notation

Data sets produced with the aid of the OWTRAD notation system do not aim, in the first instance, to provide detailed geo- and chrono-referenced information. They aim to create a body of information which can be subsequently easily referenced according to the specifications of the ECAI project. They aim at the production of information in a format which enables it to be soaked up, if necessary, by various data-bases and subjected to additional refinements and post-processing.

4.2. The OWTRAD notation system - basic elements

The OWTRAD notation, in its current (Jan 2000) version succinctly represents data about some 14 key variables. These are: In other words, the notation handles information about nodes, links, routes, and users. It does not deal, at this stage, with variables pertaining to the transported commodities.
For instance, a brief (in this instance 17 words) expression referring to one of the lines on a map entitled "The economic life of the Roman Empire" (Stone 1989:91) which is a part of the "'Times' Atlas of World History":
Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
identifies and describes The elements of the propoposed notation are introduced, step by step, below. Each vasic elemenent of notation is highlighted in bold letters and briefly annotated:
  1. departure node, its type, name and general geographical provenance (e.g. named or unnamed place in a given country)

  2. e.g.
    Expression: Panormus@IT (Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: there is/was a communication link "=" starting at "Panormus". The departure node is situated in the present day Italy, more specifically, on the Island of Sicily.
  3. arrival node, its type, name and provenance

  4. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: there is/was a communication link "=" terminating at "Ostia". The arrival node is situated in the present day Italy.
  5. general physical environment (e.g. open sea route, land route, river route)

  6. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: an open sea route.
  7. general context, the communicational network, the frame which the given link is a part of.

  8. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: details of the frame are not stated in the source.
  9. user details (e.g. Greeks, Romans etc)

  10. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: used by inhabitants of the Roman empire.
  11. uses of the communication link (e.g. trade route, pilgrimage route)

  12. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: used for the trade purposes.
  13. type of the communication link (e.g. general communication line, actual road)

  14. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: an unimproved movement corridor, i.e., an avenue.
  15. relative importance of the communication link (e.g. minor road, major route, etc.)

  16. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: apparently, a major route.
  17. physical length of the link (e.g. miles or kilometres)

  18. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: the length of the route between Panormus and Ostia is not stated by the source.
  19. time required to travel the distance, as stated by the source

  20. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: the travel time on the route between Panormus and Ostia is not stated by the source.
  21. chronological context

  22. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: the route was in use in the 2nd century CE.
  23. source quality and data quality

  24. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: the source is rated to be of relatively low, "C" quality, the data are to be of relatively low, "3" quality.
  25. bibliographical reference

  26. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: information comes from pages 90-91 of a book by Stone, Norman (ed.). 1989. "The Times" Atlas of World History. Third edition. London: Times Books Ltd.
  27. data set identifier

  28. e.g.
    Expression: Panormus@IT(Is.Sicily) = Ostia@IT [S frame:unspec romans trd av maj d000dd t000t m0200y QC3 (Stone 1989:90-91) XMEm0200]
    Value: the data set is called "XMEm0200", to indicate that it deals with the "X", ie. sea routes; criss-crossing the "ME", Mediterranean Sea, mainly in "m0200y", i.e. the years of the 2nd century CE.

5. The OWTRAD notation system - format

The following terminology intends to capture information originally expressed in verbal as well as graphic formats (remarks, descriptions, maps, plans, situational charts - see Appendix A) and convert it into a simple, regular written format.

5.1. & 5.2. Departure/Arrival Nodes

Place identifiers
Meaning: a node, place or other reference point in the terrain which identifies a beginning and end of a given stretch of road or other communication link.
Example: ArabiaEudaemon@YE : a place in contemporary Yemen, referred to by the source as 'Arabia Eudaemon'.
No of characters used: several chars.
Further details:
Annotations
Directional and locational annotations are enclosed in () round brackets. Note: All other variables, starting with "General environment" and finishing with "Data set identifier", are enclosed in [] square brackets.

5.3. General environment

Meaning: Overall information on the type of the communication link.
Example: "R", a route which follows the course of a river.
No of characters used: 1 char.
The range of values:

5.4. Frame (geographical context of the communication network)

Meaning: Role of the communication link within the context of other links
Example: "frame:Constantinople-BalticSea", a communication link which is a part of a route spanning the two reference points.
No of characters used: several chars.
The range of values:

5.5. Users' details

Meaning: Ethnicity of the most common users of a given set of links.
Example: viking : Vikings.
No of characters used: 6 chars.
The range of values:

5.6. Usage details

Meaning: The primary function of the link in question.
Example: plg : a route used by pilgrims.
No of characters used: 3 chars.
The range of values:

5.7. Type of the link

Meaning: Distinction between natural and artificial structures.
Example: rd : a beaten path, a road, a highway.
No of characters used: 2 chars.
The range of values:

5.8. Importance of the link

Meaning: Importance of the link, as recorded by the source for the particular time-frame.
Example: loc : local communication link
No of characters used: 3 chars.
The range of values:

5.9. Length of the link

Meaning: Information of the physical characteristics of the link.
Example: r017km : 17 kilometres, real measurement.
No of characters used: 1+3+2 chars.
The range of values:

5.10. Travel Time

Meaning: Time needed to travel a given stretch of distance.
Example: e008d : estimated measurement, 8 days.
No of characters used: 1+3+1 chars.
The range of values:

5.11. Chronological context

Meaning: Time when a given link was in use.
Example: m1410y : in 1410 CE.
No of characters used: 1+4+1 chars.
The range of values:

5.12. Data Quality Tag

Meaning: Overall trustworthiness of the information in the opinion of the collator, a way to rank credibility and precision of various sources. Better data and better sources suggest that a given set of variables should be taken more seriously than one which is a less trustworthy.
Example: QA1 : information comes from a publication which gives ample (A) detail of its methodology and its own sources; the reported data pertain to a terrain which is under 100 kms across.
No of characters used: 3 chars.
The range of values:
  • Letter Q ('q' for 'quality'), followed by a 2 digit alphanumeric code

  • Sources vary in terms of the amount of detail they provide about their data (see Table 3).
    Table 3
    Four levels of generality in publications
    -------------------------------------------------------------------------
    Level of                 Range of                      Type of
    generality              information                  publication
    -------------------------------------------------------------------------
    
    1st        data + source + context + methodology           research papers, journal articles
    2nd        data + source + context                         monographs
    3rd        data + source                                   overviews, text books
    4th        data                                            syntheses, maps, encyclopaedias
                                                                 [also, newspapers & propaganda]
    
    -------------------------------------------------------------------------
    For further discussion of issues of information quality and reliability
    
    see Ciolek 1999.
    Hence the following scheme: At the same time, data vary in terms of the amount of detail their offer, that is, in terms of their granularity. Hence the following scheme: These six values, if intersected, create a nine-fold matrix
                                    QA1 QA2 QA3
                                    QB1 QB2 QB3
                                    QC1 QC2 QC3
    In the above matrix lower values (i.e. A & 1), both for sources and their data, signal information's better quality.

    5.13. Source of data

    Meaning: The publication from which information has been extracted. A short reference to the author's name, date and page of publication is given in round brackets. The reference pertains to the bibliography associated with a given data set. This bibliography is stated in the set's meta-data section.
    Example: (Runciman 1978:184)
    No of characters used: several chars.
    Further details: all bibliographical details are provided in the meta-data section of the data-set

    5.14. Data set identifier

    Meaning: The name of the data set is the name of its file in a particular subdirectory.
    Example: XMEm0200
    No of characters used: several chars.
    Further details:

    6. The notation system in action

    Armed with the above tool we can proceed to a practical demonstration of the way OWTRAD notation system extracts information from the available sources.
    As an example I shall process information provided in a paper describing the road network in the vicinity of a prehistoric settlement of 'En Haseva in Israel. The material quoted below names 13 reference points and postulates the occurrence of a number of movement routes in an area measuring some 250 km in length and 100 km across.

    6.1. Major steps of the data-extraction process

    6.2. Transcription rules and commentaries

    The above transcription process always requires some thought on the part of the person collecting OWTRAD data. Here are some preliminary rules, which have been found useful in the work carried out so far:

    6. Putting the OWTRAD notation to daily work

    Once a data set is established other steps also need to be taken. However, they are not an integral part of the OWTRAD notation itself. These steps are four-fold:
    1. Register the data set with the ECAI Clearinghouse (Johnson 1999b) (www.ecai.org/metadata/).
    2. Obtain the geographic coordinates for as many as possible nodes listed in the data set. For the lat-long values of some of the places reported in the "ILa0600" file see Appendix B below.
    3. Maintain the newly created data set. Receive, evaluate and implement (if necessary) any corrections and emendations supplied by the users/readers of your online collection of data. Keep track of these changes in the "Version and Change History" section. Make sure that your email address can be easily located. Remember to confirm and acknowledge all input, however small.
    4. Find ways of enhancing your specialization in a particular geographical and historical field. Find online collaborators from other fields. Encourage the sharing of data, technology and methodology. Jointly discuss and implement any corrections and enhancements to your operations. Find ways of building up online catalogues of high quality georeferenced placenames. Find ways for other people, however busy or unskilled, to contribute regularly to the growth and well-being of ECAI projects.
    Between July and December 1999, in addition to creating the "ILa0600" data, the experimental notation was applied to a series of other sources as well. Altogether, the resultant data sets are as follows:
    Table 4
    Information about trade routes extracted via the OWTRAD notation
    
    DATA-SET   REGION                   DATE       INFORMATION    LINKS   SOURCE
    ----------------------------------------------------------------------------------
    ILa0600    Middle East              600 BCE    trade routes  13 links Cohen & Yisrael 1995
    XMEm0200   Mediterranean Sea        130 BCE    trade routes  46 links Stone 1989
    GRa0200    Hellenistic countries    350 BCE    trade routes  48 links Scarre 1988
    INa0100    India                    100 BCE    trade routes  85 links Stone 1989
    EGm0200    Mediterranean countries  200 CE     trade routes  44 links Stone 1989
    XBLm0800   Baltic Sea               800 CE     trade routes   9 links Humnicki & Borawska 1969a
    CZm0800    Central Europe           800 CE     trade routes  15 links Humnicki & Borawska 1969a
    UAm1000    Ukraine                  992 CE     trade routes   9 links Humnicki & Borawska 1969b
    PLm1150    Poland                  1150 CE     major roads  169 links Humnicki & Lalik 1957
    PLm1370    Poland                  1370 CE     major roads   43 links Rutkowski 1980a
    PLm1500    Poland                  1500 CE     major roads   98 links Rutkowski 1980b
    INm1550    India                   1550 CE     major roads   89 links Farooque 1977
    PLm1750    Poland                  1750 CE     major roads  200 links Rutkowski 1980c
    ----------------------------------------------------------------------------------
    On average, the OWTRAD transcription of a map or written document takes about one (1) minute per link, plus additional 10 to 30 minutes per data set to proof-read, correct, organise, annotate and place the extracted information online.
    In the case of data sets listed in Table 4 they are stored on a web site at the address of the trial OWTRAD Project (www.ciolek.com/owtrad.html). The site is intended as a kernel for a forthcoming clearing house for electronic data on the land, river and maritime trade routes of Eurasia and Africa during the period 10,000 BCE - 1815 CE. The site consists at present (Jan 2000) of five interrelated parts:
    1. an online copy of the latest version of the OWTRAD notation;
    2. an archive of the collected data sets;
    3. the beginning of a gazetteer of places mentioned in the above data sets;
    4. the beginning of an online notebook with as much of 'intelligence' as possible and other factual data pertaining to physical, logistical and social issues involved in the circulation of people, goods and information. The notebook deals with such topics as Nodes; Communication links; Technology & Resources; Users; Logistics; Speeds & Endurance; Time budgets; Staging posts; Travel times; Loading capacity; Navigation; Effects of seasons; and, finally, Commodities & quantities.
    5. a collection of research tools (ISO 3366 country codes table, ISO 8859-1 special characters table, a list of printed maps showing the course of various routes and roads)
    6. a bibliography
    All this information albeit fragmentary and incomplete, is placed online and made readily accessible to all interested users of the Net. The purpose of the site is three-fold. Firstly, it aims to provide readers with access to the current results of the methodological work-in-progress. Secondly, if freely offers collected data for use within and without the context of ECAI studies. Thirdly, it assembles together in a single spot the growing collection of tools necessary for the debugging of collected data (the intelligence notebook) as well as for their future georeferencing (the gazetteer).
    By doing so I am hoping both to receive useful suggestions and critical advice, as well as inculcate among the scholarly users of the Net a custom of programmatic, free and selfless sharing of results of one's work, however minor and unimpressive they might be. It is my sincere conviction, that social science research will greatly benefit from the attitude of openness, trust, collegial help and unimpeded access to each other's workbenches. Whatever elements of academic mystique and impression-management are lost through this process, they are splendidly compensated by the increased speed with which we learn about each others successes and blunders and the opportunity to use each other's common sense and critical gaze, to correct rapidly (and inexpensively) all sorts of shortcomings in our scholarly work.

    7. Conclusions

    For the purposes of our work on the construction of trustworthy geo- and chrono-referenced digital data about human social and cultural behaviour the described attempt at a systematic study of trade routes is very much akin to the attempts of climbing a stone affectionaly known as the Eckenstein Boulder. Oscar Eckenstein was a famous British mountaineer of the turn of the 20th century and was one of the first people research and document techniques of climbing on rocks, snow and ice.
    In summer 1956 Eric Newby and his intrepid friend, Hugh Carless were getting ready for, what they called, a "short walk in Hindu Kush". They did not have much time for proper mountaineering training. Nevertheless, they did not want to wander into the wilderness of Nuristan totally unprepared and totally unskilled. Therefore, they went to Caernarvonshire, Wales and spent quite a few anxious hours clinging upside down to the boulder learning first-hand about various moves and holds. Newby (1974:37) wrote about Eckenstein An experiment undertaken between July and December 1999 and documented in this paper suggests that the problem of how to best identify, describe and adequately map various nodes, sites, links and routes (and if possible keep track of their major users and major commodities shipped across them) is complex but not overly complicated. It can serve, therefore, as a handy training device for all budding ECAI enthusiasts.
    The OWTRAD experiment suggests that information on various historical places and transportation routes can be easily located in and extracted from a variety of sources. The data-extraction process is fairly quick and painless. It does not involve any specialised technology nor extraordinary skills. It can be accomplished with the equipment and resources we all aready have. However, it requires mindfulness and attention to detail. It also requires that we subdivide the work into a series of self-contained stages, and that by the end of each operation we place our collected data online so that our collaborators can see them and, if possible, proof-read them and debug them (Raymond 1998, Ditlea 1999). Initially, our data do not need to be complete, or perfect, or fully standardised, or fully geo- or chrono-referenced. All they require, to start with, that they start being amassed, that they are presented online in a simple, universally legible format, that they are fully referenced (with respect to the source they are derived from as well as the overall data set they belong to) and that they are made freely accessible to other people interested in the studied phenomenon.
    In sum, the present work on digitisation of information on trade routes constitutes an intriguing and promising methodological exploration. Any part of it can be borrowed by any ECAI researcher and played with and modified at will. We need to experiment freely and play wholeheartedly with our methodological apparata, our tools and technologies, and with our data. It is only through frequent and frivolous interaction with each other and the results of our work that we can learn about our limitations quickly and painlessly. We should start, however, with simple questions and even simpler materials. The study of trade routes falls into this class.
    If we manage to figure out, all within the realm of ample and straightforward historical information, how we should conduct ourselves, as individuals and as members of the scholarly community, in order to collect detailed mappable data on various points and on various connections between them, the chances are that we will be able to figure out far more difficult methodological and organisational issues as well.

    8. Appendix A - Types of information about trade routes

    Information about communication links comes in four basic formats: graphic maps, anecdotes, itineraries, and finally, verbal maps.
    i. Graphic maps
    In this category of materials the geographic information is represented predominantly nonverbally, that is, through a combination of lines, points, areas, as well as various forms of colouring and patterning. Also, words are used, to identify and annotate inhabited places, mountain ranges, rivers, and other salient features of the landscape.
    ii. Random written remarks
    Possibly the largest category of materials is formed by accidental, almost haphazard remarks on various geographical points and their possible relationships. These remarks are made within a context of narrative or commentary which is focussed on some other issue: Some of this intelligence can be, from time to time, more explicit and more deliberate: All these tiny snippets of information are, of course, of hardly any use on their own, in isolation. However, in conjunction with other materials, such as itineraries and maps (see below) can be used to construct and cross-check a rich and colourful mosaic of complementary factual data. These bits and pieces of factual data are very much like fibres of some plant material, such as hemp or cotton. If dealt with one at a time, they are extremely flimsy and not worth much attention. In larger numbers however, providing that they are tidied up, systematised and brought to a common denominator, they start forming strands of information which flesh out data derived from other, more substantial sources.
    iii. Explicit written itineraries and lists
    Another group of materials of relevance to the study of communication routes criss-crossing the world is made of various catalogues of places and itineraries. This category subsumes many primary texts, including the famous "Periplus of the Erythrean Sea" (Schoff 1912, Casson 1989) and the "Parthian Stations by Isidore of Charax" (Schoff 1914), first century CE professional guides to the maritime and land trading places linking the Mediterranean with India.
    In these sources data pertaining to inhabited places and often, distances (measured as units of space or units of time), are systematically collected, ordered and stated explicitly, as a part of a longer manual, narrative or treatise: Such detailed lists are very telling. Firstly, if plotted on a map, they tell us where, in a given area, are the customary stopping places for groups of travelling people. The presence of such stopover points is, of course, influenced by several variables (Engels 1978, Lewis 1997). Secondly, the list of recommended stopping places informs us about the distances and rates of daily march characteristic of a given category of travellers within the context of a given historical situation.
    iv. Verbal maps
    This category is made of materials where deliberate and fairly detailed accounts of two or more interlinked routes are presented in an attempt to discern and annotate all major component parts of the communication network in question.

    9. Appendix B - Gazetteer of locations listed in the data set "ILa0600"

    The materials below attempts to collection information on (i) place's preferred (i.e. 'vernacular') name (in bold characters); (ii) place's type (e.g. inhabited place, river, island etc.); (iii) the ISO 3666 code (RIPE 1997) of the country in which it is currently situated; (iv) latitude and longitude values; (v) decimal latitude and longitude values; and, finally, (vi) the list of it's variant names.
    Some of these georeferences can be acquired online through interrogation of the Getty Research Institute's (1999) Getty Thesaurus of Geographic Names (TGN) or, for central European materials, of The Jewishgen ShtetlSeeker database (Tobias 1999). Other values can be located in such gazetteers as: Seltzer (1962) and his The Columbia Lippincott Gazetteer of the World [Contents: about 159,000 entries, with about 4,500 defined via lat/long coordinates]; Lloyd's Maritime Atlas (1964) [Contents: lat/long list over 5070 ports and shipping places of the World]; "The Times" Index-Gazetteer of the World. 1965 [Contents: approx. 345,000 lat/long locations]; International Geographic Encyclopedia and Atlas (1979) [Contents: 25,000 entries + index to the Atlas lat/long coordinates for approx. 14,300 locations]; and Willett (1984) The Prentice Hall American World Atlas [Contents: two indices - USA (about 29,000 lat/long locations) and the rest of the world (about 59,000)]. For sites not identified by the above gazetters manual calculations need to be made on the basis of the large-scale maps.
    In this note latitude and longitude values which are derived from sources other than the Getty Thesaurus of Geographic Names (TGN) are noted accordingly.
    Following the TGN convention, placenames are annotated with two capital letters, or "flags", in parentheses. These flags indicate the following: C for current name; H for historical name; V for vernacular name; and O for a variant name in a language other than the vernacular.

    10. About the Author

    Dr T. Matthew Ciolek, a social scientist, heads the Internet Publications Bureau, Research School of Pacific and Asian Studies, The Australian National University, Canberra, Australia. His work and contact details can be found online at http://www.ciolek.com/PEOPLE/ciolek-tm.html

    11. Acknowledgements

    I am grateful to Monika Ciolek for useful discussions and critical comments on an earlier version of this paper.

    12. References

    [The great volatility of online information means that some of the URLs listed below may change by the time this article is printed. The date in round brackets indicates the version of the document in question. For current pointers please consult the online copy of this paper at http://www.ciolek.com/PAPERS/pnc-berkeley-02.html

    13. Version and Change History


    Site Meter
    visitors to www.ciolek.com since 08 May 1997.

    Maintainer: Dr T.Matthew Ciolek tmciolek@ciolek.com
    Copyright (c) 2000 by T.Matthew Ciolek. All rights reserved. This Web page may be freely linked to other Web pages. Contents may not be republished, altered or plagiarized.
    URL http://www.ciolek.com/PAPERS/pnc-berkeley-02.html
    [ Asian Studies WWW VL ] [ www.ciolek.com ] [ Buddhist Studies WWW VL ]