Last week, we looked at a model for what the countryside around an ‘ideal city’ might look like. Today we’re going to introduce some complications to that model (you will recall, our ideal city existed in a perfectly flat plain of uniform fertility) and see how they change the patterns of land use which in turn determine what the land outside the city would look like. Finally, at the end, I want to talk at a bit more length about why I think the trope of ‘Lonely Cities’ distorts our view of past societies (and is a bad habit for speculative fiction).
Let’s start by bringing back our ideal city model:
The main feature of this entire model is that the intensity with which the land is utilized rises as you approach the market center, because of the importance of transport costs. Low density use (like pastoralism) is pushed to the outside while very high density use – often labor, infrastructure or cost intensive – is pulled to the center.
But the key thing to note here is that the dominating force is transportation costs. Anything that changes those costs is going to distort the model. We’ll start with human terrain:
Roads and Paths
Road systems alter our neatly circular model in exactly the way you would expect: land next to roads is brought ‘closer’ to the market center than land further away, in terms of transport costs. Close to the city, this is why the buildings outside the wall seem to ‘cling’ to the roads:
Now we should note that not all roads are the same and so there is a hierarchy of roads and paths. The most basic of these are what are now called ‘desire paths’ – areas where the shortest or easiest foot path between two destinations has seen enough traffic to tramp down the grass and flatten the dirt to create an easier walk. There is some irony in planning literature treating these sorts of paths as undesirable aberrations compared to ‘normal’ improved paths, given that ‘desire paths’ are, in fact, the original form of roads.
In an agricultural landscape, most of these basic paths are going to connect the farming village to its outlying fields. These roads and pathways can get very intricate, because they follow use patterns, as seen here in the paths today around the village of Maryeli in Messenia, Greece:
These paths – with varying degrees of artificial improvement – are going to replicate themselves outward from the city. There’s a feedback effect to this: density increases the number of pathways and pathways in turn bring land ‘closer’ to the areas they connect with. Thus, instead of looking empty, like the plains of Highgarden, the area around cities – where the population is most dense – instead tend to be a riot of improvised and planned pathways.
Against the background noise of these usage paths, we also need to set the sort of intentional infrastructure we mentioned in the last post: Roman roads, the Persian royal road, etc. It is worth noting that big, improved infrastructure projects like this are rarely conceived of for economic purposes: the Persian Royal Road was fundamentally administrative and military in nature and the Roman road system was as well. The roads aren’t built to reduce the effective distance to the market center, but they do it anyway.
The impact of these sorts of roads can actually be observed archaeologically, because they tend to come quite late historically and thus alter pre-existing settlement patterns, so we can see exactly how they induce population centers to shift around the landscape. J. B. Ward-Perkins did a fascinating study of this in Etruria (“Etruscan Towns, Roman Roads and Medieval Villages” in The Geographical Journal 128 (1962)). There, the Roman roads (built in the late third century B.C.) slowly alter the settlement patterns as they pull population towards them, causing old centers and villages away from the roads to shrink while new ones spring up.
Of course, roads are not placed at random: they tend to link key locations. This can create a reinforcing pattern: local market centers get linked to each other by roads that feed in to other, larger market centers, which in turn increases the intensity (and thus population and thus market pull) of the land use around those linked market centers. This can create an oddly literal path dependence – towns and market centers are, after all, usually not placed for economic reasons (see the last post). Yet even once their original function declines, they may persist as centers because their pre-existing roads and infrastructure makes them more desirable – one can see this in the persistence of hill-top settlements (put there for defense) in the Roman Empire, long after the security threats had been pushed many hundreds of miles away by the progress of the legions.
One more thing we need to note about roads is how they are used by animals. Remember in our last post that we have a zone of pasturage and possibly transhumance – semi-mobile raising of animals – on the outer edge of our ideal city. These shepherds are going to be doing two main kinds of movement. First, if they are transhumant, they are going to be rotating between upper and lower pastures in predictable ways, using what are likely well-worn paths to do it. So while the path network is going to thin out as we get further and further from a market center (and population drops accordingly), it won’t run to zero until we are well up into the upper summer-pastures of our transhumant shepherds.
The other motion, however, is moving herds from the periphery of our model to the market center for sale. Initially, that is going to start on the path-network, but rapidly move on to the major roads as the shepherds navigate their flocks into villages and even towns. This leads to fences and hedgerows – because the farmers whose fields are on the side of the road do not want the animals wandering off into their fields to eat all of the delicious crops there! You can actually see this pattern of improvement above in the Exeter map – the fields adjacent to roads are fenced (if you look closely), or else the homes themselves block the roads from touching them, while the fields away from roads have far less formidable field dividers:
To sum up this section then: a major city – or even a big castle town – should radiate out roads and pathways, most of which will follow paths of least resistance (because they are usage paths). Outer buildings, and often even whole villages, will tend to cling to larger roads, while also generating their own spider’s web of pathways which – if they are important enough – may in turn have structures clinging to them.
Larger roads cut through this landscape, connecting major centers, but often passing through or by smaller population centers (like villages) which may have ‘grown up’ around the road. Fields like to be next to roads, but if the road sees significant animal traffic (and it probably does!) they’ll want to build some kind of barrier to those animals (what kind will vary by region – if the soil is rocky, expect simple drystone walls or hedgerows, if not, you may see wooden fences or some combination of both).
And If There Had to be City Walls…
So those are roads designed to encourage passage. What about walls designed to forbid passage? The assumption that tends to get made here is that walls enclose a city like a bottle: no houses outside the walls, no fields inside the city. This is how the cities of Game of Thrones are shown (there are a handful of outbuildings outside King’s Landing in some scenes, but only a few), as well as places like Edoras and Minas Tirith in the Lord of the Rings (the films; Tolkien is clear in the books that the Pelennor is quite built up). But historically speaking, that’s not what we see.
First off (as you can see with the map of Lyon, above) some cities have considerable amounts of open space within the walls. Walls, after all, are frequently positioned to take advantage of terrain features which may or may not correspond to the exact size of the city. Looking back through the maps in this post and the last, it is easy to see that the horticulture zone, in particular often occupies a significant amount of space within the walls. Certainly, in both Roman and later Italian cities, small market gardens (Latin: hortus; Italian orto, but not that the word hortus may also mean other kinds of garden) were frequently located within the city walls in whatever space was available (often enclosed spaces attached to houses).
For lower-density settlements, one might find aboriculture or even agriculture located within the walls. The map of Lyon above, shows a number of wooded pastures (more on that combination in a moment) within the walls. Gallic oppida (singular: oppidum) – fortified hilltop towns – often enclosed the entire homesteads of elite families, along with smaller dwellings for less elite resident. That kind of organization speaks to both the social organization of the Gauls as well as the purpose of the oppidum itself. Reconstructs of the oppidum at Bibracte, for instance, tend to show not only considerable open space within the walls, but also less elite dwellings outside of the walls (think about the power situation when the wall protects the elite family’s cow but not the non-elite family’s…family).
Other kinds of agriculture might exist within the city wall as well: Gilgamesh boasts in the Epic of Gilgamesh that the walls he built for the city enclose more than three square miles, declaring, “One square mile is city, one square mile is orchards, one square mile is claypits, as well as the open ground of Ishtar’s temple. Three square miles and the open ground comprise Uruk” (trans. via Stephanie Dalley, Myths from Mesopotamia (2000)). While we might beware of some epic exaggeration and simplification here, the city walls of Uruk (c. 2000 B.C.) did run a 9km circuit around the city and nowhere near all of that appears to have been built up.
And, as you might have guessed by now the second assumption – no buildings outside of the wall – doesn’t really hold either. Cities tend to grow and ‘spill out’ through their walls as space inside of the walls becomes more crowded and expensive. That’s not to say walls have no effect. One thing you may have noticed with the historical maps I’ve been showing is that walls interrupt the web of paths and roads by limiting them to a handful of gates. This in turn means that population outside of the city tends to cluster around the gates themselves, rather than being evenly distributed around the wall.
But the main point to make here is that the market center circle of our ideal city diagram does not necessarily correspond to where the city walls are. In many cases, the walls cut into the forest and horticulture zones. In other cases – particularly very large cities like first century Rome – the city might completely spill out of the old walls and spread out beyond them. Needless to say, security concerns factor in here: a city that feels very secure (likely being far from any frontier) is more likely to be unwalled or to have grown beyond its walls without expanding them.
(Side note: Some cities did have rules requiring a certain clearance from the city walls; these were especially common for cities of the early modern period protected by trace italienne fortifications (star forts), which required clear firing lines in order to function properly.)
While these human interventions do have significant effects on land-use and what the outside of a city looks like, they pale in comparison to the impact of natural terrain. Let’s turn to that next:
Blue Highways: Rivers and Seas
As we’ve been discussing, transportation costs are what drives all of the land-use patterns so far discussed (though we will get to some non-transportation concerns). Now, as a car-and-bus-and-train sort of culture, we tend to think of water as a barrier to travel – rivers must be crossed at bridges, lakes must be driven around, and so on. But this is very much not the case for the movement of goods in a pre-industrial setting. More often, the water is a highway.
The usual estimates for transit costs derive from Diocletian’s Price Edict, a late Roman law setting standard prices for things, including transport. Extrapolating from the edict suggests that the cost of moving something by river is roughly five times cheaper than moving it by road. After all, even up-river, a draft animal can pull a much bigger barge (by walking a path alongside the river) than it can pull a wagon; going downriver, the animal doesn’t need to pull anything.
Consequently, the zones of dense inner settlement will extend out much further by river. Morley’s Metropolis and Hinterland (page 84, for the curious) provides some sense of this by noting where the ancient sources record fruit and vegetable production in Rome’s suburbium; the locations the sources report this activity stay closely confined to the watershed of the river Tiber which runs through Rome, although they may be as much as 20 miles distant from the city proper (though in towns of their own, some of which are fairly substantial). Equally distant sites connected only by roads, in contrast, report things like vineyards (solidly in the agricultural zone).
Moreover, rivers supply water for irrigation systems, lowering the cost of intensive agriculture. Nearly every large population center is going to built on, or very near, a river, because humans require a lot of fresh water to live. It’s a safe bet that this river basin is going to be very intensively farmed and thus densely populated compared to the surrounding countryside.
Sea travel is even faster. Moving goods by sea was around twenty times cheaper than moving by road (and thus about four times cheaper than by river). Rome provides one of the most dramatic effects of the sea on shaping the extent of a city’s hinterland. Because Rome was so massive – almost a million residents in the first century – the draw of food was immense. The river Tiber was navigable down to Ostia, Rome’s port on the coast, providing the city with fairly easy access to the sea.
In terms of movement cost, grain from Clusium (80 miles from Rome’s city center, overland with good roads) would have been several times more expensive to ship to Rome than grain from Carthage, 380 miles away by sea (check the routes via Stanford’s ORBIS project). Consequently, Rome’s agricultural layer was nothing like a circle – in Italy, the central mountain range, the Apennines, made it difficult to move grain overland from eastern Italy (less than 100 miles away), but grain from Sicily, North Africa or even Egypt could be economically shipped to Rome’s port. Morley argues – persuasively, I think – that this would have had the effect of intensifying agricultural production in a huge (but by no means even or circular) area around the city.
To sum up: Rivers distort our ideal model much the way roads do, but much more intensely. The added value of rivers for providing irrigation means that settlement along them is likely to be much denser than equally distant areas overland. For large cities, sea access complicates this picture even further, because it allows such cities to essentially relocate their agricultural hinterland (since grain keeps well and is easy to transport by ship) to distant shores.
I’ve used the phrase ‘agriculturally marginal lands’ a couple of times (or just ‘marginal lands’ for short), so now it’s time to talk about what I mean. So far we’ve been assuming a flat plain with land equally suited for farming everywhere. Rivers have introduced one wrinkle into this picture, in that by making irrigation much easier with the supply of fresh water, they make the land around them more suitable for farming.
But there are a host of other factors. Rainfall matters intensely (along with river water availability). Types of soil also matter (and matter differently for different crops; barley will thrive in soils too tough or dry for wheat and I’ve met several vintners who argue that the best wine is produced by grapes that have to struggle a bit to survive). Changing elevation also matters – farmers generally want flat fields for plowing and irrigation (and will often cut terraces into hills to turn one rolling field into many little flat ones). The land may also be still covered in trees or rocks which would need to be removed in order to use the land for farming.
When we talk about agriculture broadly, we often refer to marginal land as encompassing land which is somehow less than ideal in one or several of these factors, but not entirely infertile. If demand is low, marginal land, even fairly close to a city, might go unused, because the cost of putting it into use isn’t worth the production – if the city grows, that may change (one way we can gauge population growth in past societies without census statistics is to look for marginal land either coming under cultivation (growing population) or being abandoned (shrinking population), which can sometimes be observed archaeologically). But marginal land may also be used for some – or several! – purposes which are something a bit less productive than intensive agriculture. Speaking from the Roman context – barley seems to often have been grown on marginal land, since it was less profitable than wheat generally, but still profitable enough to grow on land too poor to support the more demanding crop.
I’ve already alluded to another of these uses – that areas of forest necessary for the city are often found on land which is rough or hilly or otherwise unsuited for traditional agriculture. Such forest land is also often put to multiple uses – silvopasture is the term for letting orchards or forests double as grazing grounds for animals (in Northern Europe, this was often done with sheep, in Italy from Roman times onward, pigs were often let into woodlands and orchards). Such multi-use spaces, taking advantage of marginal areas close to a city, could offer a mix of valuable products from land which would be underutilized otherwise.
In practice, varying land quality is going to create pockets of more intensive or less intensive use in and around our city. You can see this really clearly in the map of Wissembourg above. Orchards and fields cover the flat areas, with villages clinging to the roads, but the hilly areas are left mostly as pasture or forest, creating little pockets of more or less intensive usage.
Conclusions: Stories in the Countryside
Obviously, we have not discussed anything near the total set of factors at play here. We’ve left aside all but the most basic patterns of agriculture and have barely touched trade. I hope we’ll get to a lot of those topics later, but I want to end this two-part series by returning to our original question:
What ought the land around historical pre-modern or fantasy cities look like (at least in a European context, where we’ve focused most of our efforts)?
And I think we’ve answered this question in two ways. First with our discussion of some of the factors that govern land-use (coming at the problem from a theoretical basis) and second through primary source evidence – period visual representations of late medieval and early modern European cities.
The land outside of cities was not empty. In fact it was heavily utilized – often the product of the countryside was the main reason the city was there in the first place. Cities were surrounded by smaller settlements (often clinging to the roads and walls) like villages and even smaller towns. They were also surrounded by what we might call (to borrow John Landers’ term from The Field and the Forge (2005)) the organic economy – horticulture, arboriculture, agriculture and pastoralism, in successive rings, modified by terrain and transit costs. Between these areas were a spiderweb of roadways and pathways – often bordered by fences to control animals.
Most importantly, these places were full of people, and that’s the note I want to stress here. What is missing most of all from our Lonely Cities is not farms or fields or buildings but people. And not just a few people: most of the people. In much of medieval Europe, most than 90% of the population lived not in the city, but around it in the countryside. Urbanization rates are a bit higher in the Roman Empire, but not far above 10% at most (even in Italy, where the city of Rome was, Morley estimates urbanization was not higher than c. 20% – still far less than half; and that figure is an absolute peak for Europe pre-1750).
As I’ve said before, in a strange way, this blindness to the vast majority of the population who lived in the countryside mirrors the concerns and attitudes of medieval and ancient elites (who tend to write our sources for understanding the past). For a Roman senator or a medieval lord, the only people who “mattered” lived in cities and castles and towns; the peasantry was of little account. This disregard was often quite openly paired with elite contempt for the common folks, often phrased in moral terms (for instance, the ‘treacherous peasant’ is a commonplace of medieval (elite) literature – one assumes the peasants might well have added violent and greedy lords, but they weren’t consulted!).
The Lonely Cities of our fictional creation take this tendency and ramp it up to eleven – the peasantry isn’t just ignored but actively deleted from the countryside, leaving a world that apparently consists entirely of elites (knights, nobles, etc) and their hangers-on and no one else. We might be told the countryside is being pillaged – but we rarely, if ever, even see that countryside and its people, much less meet or get to know any of them. Take A Song of Ice and Fire – for a narrative about the destructiveness of war, we view that destruction entirely from the warrior-elite’s perspective. Not one of the point-of-view characters is actually a peasant (though a few are smallfolk turned warriors, which doesn’t quite do the job and they only get short preludes anyway). Tolkien shows us the sad evacuation of the Pelennor before the Siege of Gondor, but Peter Jackson removes this entirely (though we do get some shots of nameless panicked civilians with all of the character of Extra743).
This ties in to one of the worst tendencies I note in my own students – the tendency to unthinkingly identify with the elites of the past, to see themselves as the knight, the noble, the senator. But not the farmer, the artisan, the shepherd, the petty merchant. All too often, I see students read the class-contempt of Roman aristocrats or aristocratic medieval troubadours with horror – and then unthinkingly replicate those very patterns of thought when themselves thinking about the past (usually in the foolish assumption that nobles were smart and peasants were dumb). That attitude – adopted unintentionally and unconsciously, mind you – is poison not just to the study of history, but also to effective citizenship and leadership in the real world.
There are stories to be told in these places outside of the not-so-lonely-cities. How often do we see historical events or fantasy epics from the perspective of common folk – the farmer whose fields host the battle, the common soldier levied or hired to fight? Almost never (the one defiant recent example to the contrary I can think of is Kingdom Come: Deliverance, although the main character there turns into a professional warrior pretty damn fast; still, watching the sack of a town from the perspective of the helpless, defenseless smith’s son is a novel way to show it). Maybe we should tell more of those stories. Maybe just putting the peasants back into their countryside would be a good start.
Next Week: Something different! Probably War Elephants, but maybe something else if I can’t get war elephants ready in time.