‘Fast’ and ‘Slow’: Abstract Thinking and ‘Real Experience’ in Two Mongolian Non-Pastoral Modes of Travel


This paper discusses two non-nomadic modes of transport in 1930s–60s Mongolia: the horse relay system and the goods caravan. It suggests that each of these should be seen as a ‘mobility constellation’ involving entanglements of mobility, narrative and practice, and implying different social relations and experience of the environment. It is argued that the relay system in particular involved abstract distance–speed calculation and that this enabled the conception of extensive cross-border geographies. The paper also explains why herders who took part in (fast) relay and (slow) caravan duties greatly preferred the latter.

and religious implications explored. But two important types of long-distance travel native to the Mongols, the post-relay (örtöö)1 and the goods transport caravan ( jing), have received somewhat less attention, especially in their socio-cultural aspects. The post-relay has been described by Chultemsuren (2016) and specialist historical articles on certain periods (Harris 2015;Kovalev 2016), while apart from Lattimore (1941), who described both forms in some detail, most information about the caravans comes from numerous but incidental accounts in travel writing (Cable 1943;Casella 1968;Gilmour [1883Gilmour [ ] 19701893;Grant 1862-3;Haslund 1936) and Russian trade and military expeditions (Batorskii 1889;Vasenev 1883). This paper aims to bring these two kinds of travel back into focus and to compare them. It is true that neither exists any longer, the horse-relay having been replaced by motor transport in 1948-9 in Mongolia, while the camel and ox caravans came to an end more slowly over the following decades. Nevertheless, it is my suggestion that these ways of organising movement are still interesting-first, because they reveal how Mongols mentally constructed and inhabited wide geographies, and second, because they are good cases with which to think about distance, speed and modes of travel as social constructs and as kinds of experience.
These two forms, the post-relay being considered 'fast' and the caravan 'slow' , structured two alternative and co-existing forms of organising human-animal movement and managing distance and speed. They involved distinctive sets of practices and provided different kinds experience of social and ecological environments during the journeys. They can therefore be seen as 'mobility constellations' in Tim Cresswell's expression, i.e. historically and geographically specific formations of movement that can be seen as entanglements of mobility, narrative and practice (Cresswell 2010: 17). Important for this article is that ordinary Mongol herders' reflections on the horse relay and the caravan have revealed the existence of extensive 'mental maps' (Gell 1985) that were distinctive to these two forms and yet were sometimes translatable between them. This translation was possible because the Mongols derived measures of space and time from these forms of travel. These were abstract measures that could then be used independently of any particular type of mobility and separately from the dimensions, such as the kilometre or the Chinese li, used by states to map and calculate distances and territories. Tim Ingold has argued (2012) that modular thinking in terms of abstractions is a product of modernity and is alien to 'the ground of real experience' , which is that of the embodied experience of perambulatory movement through a world (see also Legrain, this issue). Modular thinking tends to depict lines as limits within which movement is contained or as bounded perimeters, and Ingold contrasts this with wayfaring, which involves a temporal series of vistas, occlusions and transitions unfolding along myriad entwined pathways. 'It is as wayfarers that human beings inhabit the earth' , he writes (Ingold 2007: 74-85). The Mongol cases I shall describe contradict Ingold's dictums in some ways. They show that modular (countable) spatial notions existed well before 'modernity' however that is defined, and furthermore that such linear concepts were not divorced from human or animal experience manifested in bodily rhythms, tiredness, energy, and so forth. And yet, with these caveats, Ingold's distinction between 'contained' and 'wayfaring' modes is illuminating. It reflects a contrast that Mongols make themselves about forms of travel, and it helps us to see how their thinking about different kinds of mobility can be understood in a broader human dimension.

Travelling in Pre-motorised Mongolia
To see the place of the örtöö and jing in Mongol life, it is useful to consider the spectrum of kinds of mobility known to rural people at the period considered here, the mid-socialist era of the 1930s-60s. In the countryside, most forms of movement were unchanged since the late nineteenth century. Let us consider valuations of speed based on the human body. At the slow end of the spectrum, plain walking was seen as a trudge to be avoided if possible, the resort of poverty-stricken folk without a horse. Even for short distances, people preferred the velocity, comfort and elevated position of riding. Indeed, material culture was in accord with this preference: the heavy inflexible Mongol boots hindered fast walking, let alone running.2 To move even slower was a deliberate penance, observed notably in the pilgrimage practice of edging forward by a succession of full length prostrations that measured distance by covering the whole journey with the suppliant's body. While lines of such pilgrims were occasionally to be seen inching their way towards monasteries, the everyday pastoral scene presented other kinds of movement of varying speeds. James Gilmour, for example, described meeting large parties of leisurely riding nomad families; nearby a dashing young man showing off on his most splendid horse; an encampment of two or three hundred ox-carts, the oxen grazing and the drivers sitting mending the wooden wheels; and in the distance, a long train of tea-laden silent camels (Gilmour [1883(Gilmour [ ] 1970. It is in relation to the medley of these diverse movements that the post-relay was seen as 'fast' , being a set distance to be covered on horseback at a constant gallop or canter. The first exotic motor cars had arrived in Urga [Ulaanbaatar] in the 1910s. And the edges of the steppe world had already been reached in 1901 by the Russian-built Chinese Eastern Railway that ran through Hailar and Harbin to connect with Beijing. By the 1920s, Beijing was also linked by rail with Zhangjiakou [Kalgan], Hohhot and Baotou on the southern edge of Mongol lands. With the increasing dominance of the twentieth-century culture of speed, 'slowness' came to be seen not just as a hindrance to communication and trade, but as denoting developmental backwardness, and 'slowing down' was liable to be interpreted as disengagement from, or opposition to, this culture (Vannini 2014: 117). Japan more than Soviet Russia was taken with velocity. Under its occupation of Manchuria in the 1930s, the railway even boasted a 'bullet express with a top speed of 140 km per hour' (Anon. 1939: 42). Motor transport from Outer Mongolia met the railhead at Zhangjiakou. But it was scarce and unreliable and for decades could not cope with the bulk of exports and imports. As for the rest of the country, there were too few motor vehicles for general use. In the vast steppes and deserts, the earlier animal-based forms of travel dominated until the 1950s-60s. It is in the context of this spectrum of mobilities that I seek to understand the horse-relay and the camel-or ox-caravan.
This paper uses materials gathered in 1974 in the Gerelt Zam Negdel, Ar Khangai Aimak, and in Galuut Sum, Bayan-Hongor Aimak, Mongolian People's Republic. At that time, many senior herdsmen vividly remembered taking part in post-relays (örtöö) and goods caravans ( jing) as these systems had operated in the 1930s, '40s and '50s. One contribution this paper hopes to make is simply to provide some information about rural life this period. The primary (but not the only) point of view taken will be 'from below' , i.e. that of the men who worked in these systems, rather than the officials who organised them. I will argue that although the two systems were linked in certain ways and although both were important infrastructures of the same socialist Republic, they reproduced very different patterns of social relations, particularly in respect of hierarchy and equality (see also Safonova & Santha, this issue). I suggest further, and this is where Ingold is relevant, that the two forms had radically different experiential and imaginative implications as regards human ways of being in the surrounding world. Although the practices I describe have disappeared as institutional systems, I venture to think that the dispositions and sentiments they reveal may have been passed on among rural families and may still have some relevance, even to some people who have migrated to towns today. In particular, I explore one perhaps unexpected feature revealed by my respondents: the advantages and pleasures of 'slowness' .

The Horse Relay System
It is clear that Chinese and Inner Asian empires organised relay systems centrally from very early times and that they had important functions of governance. Over time, the systems expanded and contracted along with the capacity of a dynasty to achieve control over and keep up communications with its various regions (Atwood 2004;Kovalev 2016). The Mongol Empire established a particularly robust and extensive system, which in Ögedei's reign was made more efficient by confirming definite routes (zam), settled stations and trained station managers. The aim was speed-to convey persons and information as fast as possible across great distances.3 This was achieved by obligatory provision of prime horses at the stations, such that the envoy, or successive messengers, could travel almost without stopping to the destination. All of the tasks involved (service as an envoy, provision of horses, guides, forage, accommodation, food, and management of a station) were unpaid obligatory duties (alba) to the state. The norms for distance between stations, the length of journeys per rider, the weight of the letters or parcels conveyed and the metal 'pass' (paiza) licensing uses of the service were strictly regulated. Each messenger's horse wore bells to warn of its approach, and as the tinkling grew close, fresh mounts or riders were prepared, so as not to delay immediate departure (Juntunen 1991: 170). Mongol historians have estimated that there were over 1500 stations and a total network of 45,000 km covering the whole of the empire at the period of the reign of Emperor Khubilai (Chültemsüren 2016;Perenleilhündev 2011). Under the Ming, the relay system was among a variety of essentially military techniques for penetration of imperial margins; it was used along with gunpowder, cartography, bureaucratic administration and appointment of native officials to bring these areas into the orbit of the centre (Harris 2015: 72-3). Subsequently, sixteenth-and seventeenth-century Mongol law codes stipulated their own obligations with regard to relay and cartage duties (Riasanovsky 1965). Clearly this was a modular system involving abstract calculation of distances and speeds that (contra Ingold) developed long before the advent of modernity. What is significant for this article is that although post-relay systems were ancient and were used by diverse empires in global history, the Mongols thought of the örtöö as inherent to their own mode of rule, not as something foreign. It was, I suggest, part of an indigenous aristocratic practice of government that was acknowledged and reiterated throughout society (Sneath 2019; see also Legrain, this issue). This consciousness persisted into the twentieth century, even though the relay system had in the interval been operated by the Manchu Qing Empire to which the Mongols were subject.
With the setting up of border posts along the frontiers, similarly stipulated as state duty, the Qing system reached a high point in the eighteenth century. However, when the dynasty was weakened by invasion and rebellion, it lost grip on parts of the post-relay operation. By the early twentieth century, regional, private and company postal systems competed with one another and with the skeleton of the state operation. In Inner Mongolia, all of these were preyed on by bandits. However, in Outer Mongolia the state-run service, with its strictly ranked officers zalan, zangi, khuyag and ulaachin,4 seems to have been maintained relatively intact. There, it continued to serve Qing officials till 1911 and its framework was retained by the succeeding autonomous and then socialist governments.
What were the attitudes of ordinary Mongols required to serve the post relay system? Owen Lattimore, who knew Inner Mongolia well in the 1920s-30s, wrote: 'This system, much admired by Europeans since the Middle Ages, was hated by the Mongols.' (Lattimore 1962: 57). It was a particularly onerous obligation among the many duties (alba) herdsmen had to render to their lord, such as supply of milk, cheese, hides, and fuel, and labour service at the lord's headquarters. The nobles were exempt or passed the onerous aspects of örtöö duty over to their subjects.5 Horses had to be provided-and the lord would never provide a horse of his own as long as he had a subject who owned a horse. Somebody had to ride with the message, or escort the traveller from one stage to the next, and to bring back the horses-and whoever heard of a lord discharging this duty? Lattimore 1962: 57 We read further: To be a stage-rider was a tough and much hated life. Nominally, all commoners could be conscripted to ride the stages, carrying official orders or escorting travellers, but anyone who had any money at all paid a substitute to ride for him, and so only the poorest men took on the job. Lattimore 1962: 77 Lamas were exempt, and it seems that the status of shabi, subject of a Buddhist monastery rather than a secular lord, was valued at this period for the very reason that the shabinar were mostly freed from örtöö and military service. For this reason, some commoners tried to escape from princely to church domains. When lordly rule was weakened by the wars of the 1930s-40s, there were wholesale migrations-to escape from such 'feudal duties' as well as accumulated debts (Lattimore 1962: 104).
The örtöö can be seen as a 'superstructural infrastructure'-in other words, it was a transport mechanism designed more to facilitate rule than economic or social purposes. Its routes, running between governors, went through pastures but generally by-passed places where Mongols would gather, such as large monasteries or trade settlements. This can be seen from early twentiethcentury maps, some of which also indicate a bone of contention: the land that had to be set aside for the livestock, mainly horses, supplying each station. Maps show the örtöö as a sequence not just of the points where mounts would be exchanged, but of equally-sized loops of the surrounding pasture all along the way.6 This alienation of land was resented by local herders. Yet as I describe below, the idea of the relay as a linear series was not foreign to ordinary Mongols and their dislike of it was rooted in experience of the way it was actually practised in local political relations.
By the 1920s-30s, Outer Mongolia had become a socialist republic, yet the countryside had almost no modern transport or communications infrastructure such as motor vehicles, tarmac roads, telegraph, or telephone.7 Robustly built lorries and buses that could cross the steppes began to appear in rural regions only in the late 1940s-early 1950s and were not widespread till the 1960s. For decades, the country needed to maintain a horse-relay system. But how was a modernising socialist state to reconfigure this historically imperial-feudal organisation?
Here it is useful to recall the political-economic background of the time, as recounted to me by Minzhuur, the Chairman of the Gerelt Zam Negdel in 1974. He gives an interesting perspective 'from above' , from those trying to make the system work. He pointed out that in the period in question (late 1930s-early '50s) the Mongol herders were not collectivised-they had rebelled against attempts to follow the Soviet example-so the herds were still individually owned and local society contained both rich and poor as well as a certain amount of undercover private enterprise.8 Minzhuur said that the horse relay had been an essential element in getting socialist policies through to unwilling and elusive people: Poor people lived as dependents on rich families. But even their children were not interested in improvements or education; they thought they were destined to live like this. Everyone thought his way was right, so it was impossible to introduce new things.
The örtöö was seen 'from above' as a means to control this recalcitrant population. In fact, as compared with the Qing version, it was strengthened, intensified and extended further into localities. In principle a centralised state institution, the network went from Ulaanbaatar to the Aimag capitals and from there they went to two to three large 'threshold stations' (bosgo örtöö) near each regional capital. From the bosgo, routes spread out in various directions to the seven or eight Sum [districts] in the Aimag. From stations at each Sum centre, which in those days was 'settled' in the sense it had a few buildings, the lines spread out to the smallest units of all, the Bag. The Bag had a general area but no boundaries. The centre of the Bag was wherever the headman's ger happened to be, and 7 The Moscow-Beijing telegraph passed through Mongolia from 1904, but was subsequently interrupted many times by war and revolution (https://atlantic-cable.com/Books/GNT/ index.htm). The telephone reached only a few places. The railway was a single line, begun from the Russian border in 1947, reaching Ulaanbaatar in 1950 and the Chinese border in 1955. The Mongolian national airline was set up only in 1954 (according to Wikipedia, in 1956). 8 For details on household economies in the pre-and post-collectivisation periods in Ar Hangai, see Humphrey (1978).
yet the riders carried messages there. Thus, the new örtöö was more comprehensive, penetrating down to the grass-roots, than the Qing system that mainly linked Manchu officials. The hierarchy of the socialist version followed exactly the structure of the state administration of the population. In fact, since the service delivered the various edicts, resolutions and regulations to the people, it could be said that it was the state-or a significant part of it.
As before, only certain categories of people were entitled to use the örtöö: state and Party officials and couriers, army officers and soldiers returning on leave, and then only if they were furnished with an alban bichig [permit]. Local users needed a paper stamped with the seal (tamga) of the negdel chairman. Meanwhile, the herders had to provide the services at the stations. Perhaps surprisingly for a socialist country, this was not a universal state duty, but was organised according to an elaborate system of wealth calculation. Locally, it was the Sum officials who decided both who could use the service and who should serve. After the Aimag gave figures for the number of horses required from each Sum, the Sum administration gave orders to the designated households-to provide a certain number of horses and to serve a turn (usually 30 days) at the station. This was done according to another abstraction, a counting measure of the livestock units (bod mal)9 owned. Extremely poor people were exempted altogether. If someone had a large number of sheep but very few horses, they still had to supply horses to the örtöö, and this was done by 'renting' horses for the period at varying, but often exorbitant rates, from an owner with a large horse herd. The consequence was that the horse-rich households prospered, while the poorer families could take a harsh economic hit. People generally understood the system to be unfair, in that it both enhanced existing inequality and was subject to political favouritism by the Sum officials.
The örtöö stations were not settled-they had to be mobile in order to provide pasture for the horses as well as the cows and sheep that the serving families brought with them to live on. On average a minimum of 70 horses were kept at each örtöö at any one time in the 1930s.10 The conflicts with herders, known to have been frequent in earlier historical periods over pastures appropriated for state relay stations, were not mentioned to me by the Gerelt Zam elders. They were preoccupied rather by the hard, thankless work involved. Generally, the speed and distance demanded correlated directly with the importance of the communication to the state. Yet the carrier might well have 9 The bod was a counting measure for livestock. In Gerelt Zam one cow and one horse were one bod, a camel was 1.5 bod, seven sheep were one bod and 12 goats were one bod. Such conversion rates varied slightly in different places and times. 10 The bosgo stations never had fewer than 100 horses, but a small örtöö might have only 20. no idea what this message was: to them this was an alien velocity. The most urgent dispatches were carried by bosoo elchi [standing envoys], so called because they did not sit but stood in their stirrups and travelled always at a full gallop. They were expected to accomplish three to four stages per 24 hours (one honog), i.e. up to around 120 km, while the fastest and most exhausting stints could reach 200-300 km per day.11 The riding was so physically demanding that the bosoo elchi had to bind their entire bodies, including legs and arms, tightly in cloth under their clothes, to prevent jolting to their inner organs and muscles. The bosoo elchi 'knew neither night or day, neither sleeping or resting,' people told me.
The station itself was managed by up to 10 men delegated to live there for the required number of days, who were called the örtööchin [d]. They would take their families and a few animals with them, leaving their main herds with a relative. There was a headman, who was responsible for seeing that enough good horses were ready, that they were fed and watered, that appropriate food was available for the messengers12 and that a furnished ger (tai ger) was available for less hurried travellers to stay the night. The örtööchin took it in turns to organise the work of the lower-status escorts (ulaach), who were the younger men assigned to go with the traveller as far as the next station and then bring back his horse. The ulaach's job was arduous and unpleasant, since it involved acting like a servant to the traveller, and if his horse tired on the way, giving up one's own to him, trailing along later perhaps on foot, and then bringing back both horses.
As Lattimore suggested for Inner Mongolia earlier, these demands gave rise to elaborate calculations (usually payments in kind) for how to avoid personally having to serve. This was an obligatory state duty, and although there were closely bargained payments for certain expenses by agreement with the Sum,13 there was no salary as such. Those on duty were expected to provide their own food for themselves from animals they brought with them. As for the men who agreed to 'work the stages' as ulaach, often on behalf of someone else, they 11 Four örtöö stages per day was counted a medium, though not the fastest speed. All of these socialist-era speeds are slower than those reported, possibly exaggerated, from earlier historical periods: see note 3. 12 The food provided was closely regulated. It was known as 'neg khöl makh' [one leg of meat], which was a unit that could consist of various cuts of meat, depending on the species of animal. For example, the four top ribs and a shoulder of mutton counted as 'one leg' . 13 The örtööchin received two tugriks per month per horse supervised, according to one informant; another cited five tugriks per 10 horses, the usual number supervised by each örtööchin.
tended to be people from such poor households they had few other options, escapees, or drifters looking for work-like the young Sükhebaatar at the age of 14 (Lattimore 1962: 82).

Örtöö Geographies
What is interesting is that the örtöö system gave rise to its own geographies at both local and far wider scales. Everyone at this period knew that the stations were around 25-30 km from one another, a distance dictated by the ability of horses to travel fast in a given terrain. Locally, the örtööchin had to know in detail the two stages on either side of his own station, since he had to be prepared to send out his ulaach in either direction. He had to know their current location, the time it took to get to them via alternative routes, and then give the right instructions to the travellers. My respondents in Gerelt Zam who had served on the horse relay knew not only this corridor of the three örtöö: they were also aware that the nearest 'threshold station' had a broad area in its purview consisting of all the Sum districts surrounding it. They knew this because each Sum had to send horses to the threshold node as well as to their own örtöö and they had carried out this duty. But 'one örtöö' was also an abstraction that was used as a unit of measurement for distance and time. Pictured as an identical length whatever the actual irregularities of the terrain, it could be used to describe a variety of journeys that had nothing to do with the relay system itself, such as trade, migration, pilgrimage, or allocation to some distant military duty. The unit of the örtöö enabled people to maintain mental maps, sometimes consisting of vast geographies, by counting the stages as numbers, such as the 19 örtöö between Sain Shand and Kalgan [Khaalga], or the 107 between Kiakhta on the Russian border and an end point on the Amur River, and then juxtaposing these ladderlike lines. Although even a professional messenger could not possibly have made all of these journeys in person, it was evidently possible and useful to compose a mental örtöö geography of this kind. I made a sketch copy of a map in the possession of one of my respondents that detailed the numbers of örtöö over a vast Inner Asian space, from Beijing in the East to Tarbagatai in Russia in the West, from Kiakhta in the north to Ili and Urumchi in the southwest.
At first sight this sketch looks like a simple map of the late Qing imperial post-relay. But this was a personal map, and the Mongol focus is evident from the prominence given to 'Chinggis Khan's Palace' . Annotations note the character of certain routes ('waterless' , 'a hard road' , etc.). It seemed to me that this was not esoteric knowledge, although younger generations in the 1970s might have had difficulty reading the Mongol script. Older herdsmen I met in the negdel seemed perfectly au fait with the script, how to draw such a map and the mental picture of the spatial relations between the places it referred to. What is interesting, is that this image is entirely made up of notional örtöö 'distances' , like a tape measure, and that the picture entirely eliminated political borders and boundaries.
In practice, the örtöö was a measure of time as well as distance. But the length of time it indicated depended on the speed of the mode of transport. From my conversations, it seems that the maps of long distances reckoned in örtöö were actually most useful to people engaged in a very different and much slower form of travel: long-distance caravan transport. For the caravan men, 'one örtöö' was not a number of minutes, as it was for the fast envoy, but well over a 24-hour day.14 14 I was told, for example, that men reckoned the 25 örtöö from Hailar to Ulaanbaatar took 30 days and nights. If the post-relay system meant for ordinary people (even in early socialist times) enforced hardship, economic expense, carriage of alien messages, reinforcement of inequalities and subjection to orders 'from above'-all of which negated whatever pleasure there might have been in galloping on horseback-a very different picture emerges from their narratives of caravan work.

Caravan ( jing) Transport
Caravan travel in Inner Mongolia and Xinjiang in the 1930s has been vividly described by Owen Lattimore in a number of publications, especially his Mongol Journeys (1941). However, the situation in the early Mongolian People's Republic is much less well known.

Many of my respondents in Gerelt Zam
Negdel had taken part in such transports and were eager to talk about it. I had the impression that, even in the 1970s, these customs were close to their lives and to some extent still practised (although transport by then was supposedly all motorised). Taking out a caravan was known as jing teekh [caravan or load carrying].15 The classic examples described by Lattimore were camel caravans bearing loads on their backs. But there were no camels in Ar Khangai and all caravans were composed of oxen or yaks, either loaded or pulling carts (or sledges in winter). These moved at a very slow pace, such that they could easily be overtaken by a man walking. Large cart-wheels made of heavy timbers and not perfectly round accentuated the sluggishness.16 Journeys were often measured by time, days or months, rather than by distance. What is strange, at least initially, is that the horse-riding-loving Mongols highly valued the quietly ambling jing, while they disliked the dashing örtöö, and in this section I hope to explain why.
First, some economic background. In the 1930s-40s, the caravans my respondents worked on were used mainly to carry goods to and fro the relatively short distance from Ikh Tamir Sum to Altan Bulag on the Russian borderthough several of the men had previous experience of far longer journeys, 15 According to a Mongol informant, the etymology of the word jin ( jing) is related to the root morpheme 'ji-' , meaning 'pair' , appearing in words like jiguur [wings of a bird]. Here it refers to the practice of balancing loads, one on each side, as in loading a camel. 16 Mongols explained the use of such seemingly clumsy wheels as perfectly rational. They helped the wheel's grip in mounting sandy inclines, and they prevented the cart running away during steep downhill sections. For the same reasons (better grip), iron rims were not used on carts that had to cross mountains. especially to and from Hailar, which was an export/import mecca for caravans as it linked to the railroad. The Chinese and European private trading firms (puuz), which had earlier hired caravans to transport goods back and forth to China, had been squeezed out in the 1920s (Atwood 2003).17 They were replaced by Mongolian state trade organisations that also transported donations in kind from the Soviet Union. A new and different range of Russian items was carried. The caravans from the USSR imported flour, cloth, tobacco, consumer goods, tools and iron, and they took for export hides, wool and sheepskins. Most of the caravans from and to Ar-Khangai were organised by the Negdel or Sum authorities. It was a 'state duty' of the herdsmen to provide oxen and serve on these caravans. Some consumer goods arriving at the Sum/Negdel were distributed further by official 'agents' to the Bags in short caravans. As the herdsmen at this period still owned their own livestock, there were also some small private caravans taking products for sale in towns at the edge of China (fox and wolf furs, mushrooms, deer horns, leather, lamb-skins).18 Wandering traders also existed, who would buy city items from caravan men and take them to sell in the furthest camps in the Bags. This was known as ganzagyn khudlaa, incidental 'trade from the saddle-strap' . It was difficult for me to tell the extent of this informal trade from the cautious statements of the negdel members (in the 1970s such trade was forbidden). But clearly the caravan men stood to gain 'on the side' for their work in transportation. Furthermore, the respondents told me that this was truly honourable, hard manly work. It demanded physical strength, bravery and valued knowledge. In either case, whether state or private, people said that they had been pleased and proud to work as a caravan man. An average-sized state caravan would have around 30 carts with three jingchin [caravan man], each in charge of some 10 oxen (though caravans taking animal products for foreign export could be far longer, with over 100 carts). I write 'oxen' here for short, but in actuality the 'Mongol ükher' [ordinary cattle ox] was distinguished from the two different variants produced by crossing cattle with male and female yaks. The different qualities of the three types of 'oxen' were known in minute detail to these men. Some were hardier, some suited to rocky mountainous roads, others to flat plains. The animals were attached to one another in file, with the last ox carrying a bell to alert the men in case the line broke. One of the men would be stationed at the front, riding the front cart or sometimes on horseback.
The caravans had a miniature social organisation. Three or so jingchin formed what was called a 'gal' [fire], i.e. the camp fire with trivet around which they gathered to cook and eat. They took a tent (maikhan, not a Mongol ger) for shelter. Their relations, people told me, were relaxed, helpful and equal-so although when inside the tent the eldest by age would sit in the honoured spot opposite the door (khoimor), he was not in authority over the others. 'There was no darga [boss]' , people said. They readily contrasted this egalitarian ethos with the hierarchical and status-ridden organisation of the örtöö.
Everything on the jing was carefully thought out and planned. There were norms (and regulations as socialist organisation took hold) about the weight that could be carried on each cart. Papers had to be obtained documenting the number and weight of the boxes taken and stamped when handed over. An approximate route was planned in advance in accordance with the landscape and the animals' capacities, bearing in mind the many possible dangers and hardships (devastating storms, rivers in torrent, landslides, terrible frosts, well run dry, and so forth). The oxen had to be pastured en route, which was vital to maintain their strength, especially in winter, so it was not advisable to travel along a main road, where nearby pasture would already have been over-grazed. The caravans therefore often took their own, somewhat winding routes, taking account of numerous factors such as wells or streams, spots with soda (khujir) and fords over rivers, while avoiding bogs, sandy mounds, steep rocky inclines and places known to be ridden with evil spirits. Lattimore describes the complex, multi-sensorial knowledge of the land that this 'wayfaring' required. Camel men acquired it over years of experience. For example, when travelling at night, they could recognise the smell of the earth of a route or region. An old man would dismount, take up a handful of earth, sniff it and say, 'No, this is not our road, we should go in some another direction' . More than once, Lattimore saw a man dismount and feel the ground and say, 'We are all right' or 'Bear away to the north' , this being a real skill derived from knowledge of, and attention to, patches of different kinds of soil or vegetation (Lattimore 1941: 139).
Caravans had their own timing of resting and moving that responded to the energies, temperature and feeding rhythms of the animals, rather than to the human habits of herders. The former caravan men of Gerelt Zam Negdel described a routine (incidentally coinciding closely with that described by Lattimore 1941: 137-63) whereby the caravan set out in the night, well before dawn, travelled till mid-day, took a long rest for grazing, and then set out again for a second evening spell until after midnight.19 This meant very little time for the men to sleep, but it was done in order to reduce the length of time the animals would lie on the cold ground at night (which was bad for them, as they sweated a lot after their exertions pulling loads). Thus, unlike the relay messenger 'who knew neither night or day' , the caravan men paid great attention to diurnal rhythms-but those of their animals more than their own.
When travelling at night, both the timing of starts and stops and the orientation of march were regulated by the stars. For example, according to some, they stopped on winter nights when the Pleiades constellation appeared; according to others, it was Orion. They estimated directions by Venus, the Lode Star, and so forth. This astronomical referencing along the way went with much attention to the cosmological circumstances of the caravan as an event in time. The caravan must be blessed with good fortune. This involved a cosmological rationale that was entirely separate from, and incipiently at odds with, the socialist ideology. It required the coordination of the astrological status of the man initiating the caravan with its date of starting and its direction. A lama or zurkhaich [astrology specialist] had to be consulted to ascertain the auspicious day and direction. If the caravan had been ordered to go in an inauspicious direction/ date, the custom was to set out first on a short trip in the right time/direction, then return home, and set out on the official journal next day. In this and other ways, the caravan was set up to be 'blessed with fortune' (buyan khishig). There was a saying, 'We go out in the wind-horse direction. We come back in the meritorious fortune direction (buyan khishig züg)'-but what these directions actually were changed for each person every year according to the astrological coordinates. The day before setting out, the household of the caravan leader would perform the wind-horse offering (khii morin sang) and libations of milk, dried cheese and alcohol were made to heaven (tenger). The oxen or camels were purified with juniper smoke. There were certain set phrases that should 19 A man from Dund-Gov Aimak, Delgeriin Bud, aged 75, said: 'There are two kinds of jing techniques, namely Kharchin and Khalkha. The Khalkha type has two forms, 'old' (khögshin) and 'young' (ider). Ider means working at night and khögshin is resting at night and working in the day time' . The Kharchin method did not refer to the Kharchin ethnicity of the caravan men, but to a different organisation of the day, and it was used mostly by Chinese. The difference between the Khalkha and the Kharchin techniques was described by several of my respondents and is mentioned by Lattimore (1941).
be pronounced as a caravan set out and returned and at any meeting on the way. A wind-horse flag for raising fortune20 was put on one of the carts. On the journey, there were a host of customs, taboos and prohibitions to observe. First thing in the morning, tea or milk should be sprinkled in the direction of the journey. If a piece of a cart fell off, it was forbidden to burn it on the fire-it should be carried home. Bones from meals eaten must not be broken or thrown away, but burned; otherwise the oxen or men risked having their legs broken during the journey. It was not allowed to sit on one side of the cart with legs hanging down-one should sit in the middle. At mountain passes, two or three hairs from the oxen and other small offerings should be placed on the oboo. When dismantling the tent in the morning, one should take care that it fell in the direction you were going. Certain places brought a risk of dangerous disorientation, spots known to be haunted by spirits, or crossed by the running-tracks of invisible beings, or simply the cross-roads of well-worn routes where polluted/evil underground forces might surface. Detours should be taken round such spots, or if this was not possible, an offering made to the spirits of the place (Nanzatov et al. 2008: 305-6). On the way, it was forbidden to annoy noted mountains, rivers, lakes, etc. by pronouncing their names, to swear, to get openly angry, or to break things. On the other hand, songs of praise should be sung-for animals, the landscape, the file of the caravan itself. A favourite song conveyed the feeling of being far away from home, with the evocation of the distant mother waiting with a dish of milk to welcome her son's return.
When a long-distance caravan finally wends its way back to the Aimag boundary, news of its impending arrival travels fast, people told me. The children and younger brothers of a jingchin should go out to meet him, even up to 30 km, and there they give him one of their horses so he can ride home comfortably, while they come later with the oxen. The wife at home should respectfully don a hat and come out to meet him at the tethering-post, take his horse and tie it up for him. He usually brings some presents, which are put in a bag and hung up. A new felt rug (shirdeg) should be laid out in welcome, and fresh rich tea prepared in advance. The jingchin is seated in the place of honour, and soon neighbours arrive to bow and give greetings and then enquire about news. As my respondents said, these customs are very similar to those for the arrival of an honoured guest.
If all this is evidence of the aura of fortune and respect that surrounded the jing, perhaps the most important reason why men were glad to take these jobs was that it gave them a sense of freedom. The point is that the pace of the caravan was so slow that all sorts of other activities could take place at a leisurely pace alongside.21 A couple of men could go hunting and catch up later. A jingchin could stop off at the ail of a girlfriend. In the 1930s-40s, many jingchin would take things to sell at places they passed. The whole caravan could make a detour to trade tobacco in the profitable region around Hailar. All of this brought pleasures and advantages. One man described how he used to take manufactured goods (baraa) by ox caravan from here (Gerelt Zam) to distant Uliasutai in the west, bringing back raw materials. 'I was not forced to become a jingchin' , he said. 'It was all done for benefit. I would stop along the way, stay a few days with my own families, have a rest, or change tired oxen. My families were living all along the route. With them everything was done by mutual agreement, there was no documentation.' I have used the present tense occasionally in this description, because although my questions were about caravans in the past, many of the responses seemed to refer to a way of life that was not completely over. Thus, the herders also referred to present-day (1970s) transport with files of lorries, such as the seasonal transport of hay from Khangai to Gobi regions, as jing. You could buy a seat on such a lorry, they said, or room for goods you wanted to send. Most of the drivers also had a fine time doing private business along the way.

Conclusion
This article has described two different forms of non-nomadic travel that have been integral to Mongol life for centuries. Both of these involved journeys that extended far beyond a community's most distant pastures. Perhaps each of them can be seen as incipiently 'global' in their own way, since they linked Mongols to foreign countries and international trade. However, rather than attempt a survey 'from above' by taking the perspective from a metropolitan centre, I have tried to understand how these journeys were viewed by the people taking part. Some interesting points have emerged. First, that many herdsmen living in a rural collective far from any town, let alone a city, had experience of the örtöö or caravan work and consequently had available to them vast crosscontinental geographies. If the örtöö was primarily a political instrument, linking governmental nodes in a realm of sovereignty (i.e. it went no further the borders of the Qing Empire, or later of early socialist Mongolia), the jing caravan was an economic operation, a section in a network that extended literally globally. The archetypical example of this had been the camel-and oxcaravans that brought tea across Mongolia to Kyakhta on the Russian border, in which the full journey involved bringing the tea by horse, river and road from south China to the Mongolia border, then the caravan section across the steppes, then transport by sledge or cart across Siberia to Moscow, and thence to sale in Europe. A few of my herder respondents referred to this well-known historic trade. But more frequently they mentioned an Asian religious geography that also crossed national borders. The caravans in this case had transported religious texts and sacred objects originating in Tibet or Nepal via monastic towns in Inner Mongolia to Outer Mongolia and Buriat Russia. Memory of these routes had not disappeared, and one of the Gerelt Zam herders was able to sketch a map of remarkable scope linking Chita in Siberia via Hailar in Inner Mongolia to Lhasa. A second point to emerge from this material is that the post-relay, which was a rule-bound governmental practice, gave rise to the idea of units of measurement that could be counted. Not only was 'the örtöö' visualised as one in a series of identical units (Figure 1), it was associated with other such abstract entities, notably monetary payments, the livestock bod (see note 6) and the 'leg of meat' (see note 9). 'The örtöö' was used as a way to conceptualise the distances of journeys and the amount of time they would take. A description in terms of a number of örtöö would provide an approximate estimation of distance, since people knew the stations were around 30 km apart; though as for the duration of a journey, a further specification of the mode of transport would have to be given. This was nevertheless a measure that was used to visualise many other kinds of journey that took people far from their native lands. In this sense, the örtöö was a mental abstraction, but one which (contra Ingold) emerged from the exigencies of practice, such as the distance a horse could travel before exhaustion.
However, mental mapping aside, in practical life it proved almost impossible to combine the caravan with the relay system. One such attempt in the late nineteenth century was described by Alexei Pozdneyev. The supreme lama of Urga, the Jebtsun Damba Khutukhtu, regularly presented his tribute known as the 'nine whites' to the Qing Emperor in Pekin, this taking the form of a caravan accompanied by a retinue of around 20 lamas on horseback. From time immemorial, the caravan had been hosted by the princes along the way. Because they were inconvenienced by not knowing the time of arrival and were therefore not able to make proper arrangements, the princes held a meeting and decided to set up temporary relay stations in advance. They specified the number of animals to be supplied as well as the caravan's dates of setting out, arrival at the destination and return to Urga. This was done. But the problem was that each year the caravan chose a different route, the easiest, or the one where least pasture had been consumed by other travellers. A prince would set out his station with provisions, only to find the caravan had gone another way. Even though a set route with 18 stations was then designated, Pozdneyev records that the whole attempt to discipline the wandering caravan was beset with confusion (Pozdneyev [1896(Pozdneyev [ ] 1971. To return to the Gerelt Zam herdsmen's accounts, they show that in the end the measurement of distance was only one element in the agglomerations of practices, narratives and affects that constituted these two forms of travel. As social 'mobility constellations' the horse relay and the goods caravan can be distinguished from other Mongol forms, such as herding flocks, moving camp, hunting expeditions, or pilgrimage, even though some of these functions could have a caravan form on occasion. The two constellations are 'good to think with' because the contrast between them is so stark. This paper has put forward the initially somewhat counter-intuitive argument that for the rural Mongols, despite their pleasure in riding, the slowness of the caravan was infinitely preferable to the speed of the relay. Perhaps the main reason for this was that the örtöö was an artificial enforcement of speed. It exhausted men and horses. It was 'unnatural' , out of tune with the world (remember the man who said 'the courier knows neither day nor night'). The caravan was the opposite. It followed bio-ecological and animal rhythms, and it was never precise. Its meandering was adapted to the slowness of bulky animals bearing freight through a varied, sometimes perilous environment, and this practice, stubbornly pre-modern, if not exactly oppositional, allowed the caravan men to live life along the way as they wanted to.