The Ancient Paths

by Graham Robb

  A historian of the Iron Age is often left with nothing but a question mark, like a shepherd clutching his crook when the sheep have run away. Does the microscopic artistry of Celtic smiths imply an equivalent expertise on a larger scale, or was the artist’s forge a lonely light in a small, dark world? If some grand measurement and ordering of the Continent was attempted, there must have been a coordinating body or at least an efficient sharing of scientific secrets, and there are few signs of this in the unpromising debris of hill forts and farmsteads.

  Archaeologists rightly stress the muddle and diversity of the early Celtic world. The homes and possessions of the living and the dead varied as much as the climate and the vegetation. With each new excavation, it becomes ever more apparent that Celtic civilization was not suddenly imported to western Europe as a flawless operating system of beliefs, aptitudes and techniques. Yet the notion of intertribal cooperation is not as fantastic as it seems. The Gaulish tribes still had strong and separate identities four or five centuries later, when the Druids were organizing yearly pan-Gallic conferences and a national education system, and when pale-faced warriors from the Atlantic fringe fought alongside sun-bronzed oppidum dwellers from the Lower Rhone.

  Even in the sixth century BC, when Latin was an obscure Italic dialect of the lower Tiber, forms of a Celtic language were spoken throughout much of western Europe. No Gaulish warrior would have suffered the fate of the Breton-speaking French soldiers in the First World War who were mistakenly shot as foreigners. The scarcity of the evidence makes it impossible to draw a map of mutual intelligibility in early Iron Age Europe, but there is also the spectacular evidence of that other common language – La Tène art, which was not just a shared taste in décor and personal adornment but a whole world of familiar forms, stories and beliefs.16

  When the protohistoric past falls silent, the figures of Celtic legend sometimes appear along the field boundaries, speaking an unexpectedly comprehensible tongue. In Gaul, where the first signs of political cohesion emerged, there was, according to a legend recorded by Livy, a body that might have organized such a grand survey of the Celtic lands. In the sixth or fifth centuries BC, ‘supreme power among the Celts lay in the hands of the Bituriges’. The Bituriges were not simply the most powerful tribe; they were the recognized leaders of the Celts: ‘Ii regem Celtico dabant.’ The king of Keltika was always a Biturigan.

  It was on the southern edge of the Bituriges’ homeland that the Mediolanum now called Châteaumeillant existed. This exceptional place is the only Mediolanum that actually contained buildings and people in the days of the ancient Celts (here). It lies towards the west of the principal zones of early expansion, and there are few middle hills in its vicinity. Yet it stands at what appears to be the heart of Gaul, at almost equal distances from the Pyrenees and the Alps, the Mediterranean, the English Channel and the Atlantic Ocean. At the time of the Biturigan hegemony, when a federal Celtic power was first developing in Gaul, and when the old system of ‘middle places’ was giving way to a wider network, it would have been an obvious omphalos, a sacred centre which, like the Hill of Uisneach in Druidic Ireland, stood in a region where several major tribal territories intersected. Châteaumeillant was not noticeably occupied until the second century BC, but perhaps its exceptional status reflects its significance in the earlier network. Perhaps this really was the centre of Gaul when the Bituriges ruled the Celts . . .

  The map of the Heraklean diagonal in the thatched cottage now acquired two new lines: a ‘midday’ line or meridian running north to south, and an equinoctial line from west to east. The two lines formed a cross, centred on Châteaumeillant. This was the common principle of Celtic territorial organization. Celtic tribes in Switzerland, Turkey and south-eastern England divided their lands into four, just as Ireland was divided into four and centred on a fifth province. In Celtic art, the motif first appeared in the early eighth century BC, around the time of the first ‘middle places’, and then recurred in many different forms. Sometimes, as on the tiny Kermaria obelisk from Finistère, which resembles the omphalos stones of Greece, and the countless pocket-sized votive wheels associated with Herakles and the sun god, the centre of the two lines is also bisected by two diagonals. In Celtic iconography, these are thought to represent the rising and setting of the sun at the summer and winter solstices.

  When I traced the two lines on the map, something quite simple and logical appeared: the meridian that passes through Châteaumeillant is the longest straight line that can be drawn through the part of the European isthmus known as Gaul. This is the ideal line for a long-distance survey, the uninterrupted axis on which the future map will be based. The cartographer’s term is ‘line of mid-longitude’. It runs ten kilometres to the west of the Paris Meridian, which is the line of mid-longitude that was chosen in the eighteenth century for the first complete and accurate map of France. The eighteenth-century meridian bisects the centre of the Paris Observatory on the Left Bank; the Gaulish meridian bisects Nanterre, at the foot of Mont Valérien, which looks down on the Seine beyond the western edge of the Bois de Boulogne. In the days when central Paris was a swamp, Nanterre was a major river port of the Parisii tribe. Its name was Nemetoduron or ‘sanctuary of the fortified hill’.

  13. The centrality of Mediolanum Biturigum (Châteaumeillant)

  Medieval provincial boundaries partially coincide with those of Roman civitates, which were based on Celtic tribal territories. Place names indicating boundaries give an approximate idea of ancient frontier zones: Bazoches, Feins, Fin(s), Limes, Limite, Marche, Ouzouer and names including the Gaulish ‘randa’. The band of frontier places running across the centre of the map follows the much later division of the languages of oc and oïl.

  It turns out that the Gaulish meridian was well chosen for other reasons too, but this will require a grounding in Celtic science. A Druidic education could last as long as twenty years, and even the abbreviated curriculum presented in the second part of this book calls for a recreational interlude. A journey along a Druidic pathway towards the centre of Gaul and the midday sun will be a useful preparation for the course, and a chance to meet some of the gods and other ancient creatures who still inhabit Middle Earth. We shall encounter them again when the local grids that were first devised in the distant days of the Mediolana had begun to form a vast network in which the migrations and memories of a civilization would be preserved.

  5

  Down the Meridian

  The expedition set off at dawn from the cathedral city of Amiens and cycled west along the river Samara (the Somme) to join the Gaulish meridian at longitude 2.1958. Three hundred and seventy-six kilometres to the south, the Café de l’Angle on the Avenue de la République in Châteaumeillant stands on the same line of longitude and marks the middle of the Biturigan oppidum. This was in September 2009. We had been following the hypothetical line of mid-longitude towards the centre of ancient Gaul for three days – as far as this can be done without actually traversing every field and marsh – and an oddly liberating sense of disbelief had settled in. Although a few stretches of road and track adhere to the line, there is no obvious physical sign that any such meridian ever existed.

  The northernmost point of the meridian, five hundred kilometres from Châteaumeillant, lies at a place disconcertingly named Loon Plage. The ‘beach’ is a desolate zone of wind-bent poplars and container trucks queuing for the cross-Channel ferry. In the late Iron Age, when sea levels were higher than they are today, Loon was an island called Lugdunum, which means ‘fortress of Lugh’, the Celtic god of light.

  Lugdunum shared its name with several other important Celtic towns: Laon, Leiden, Loudun, Lyon and perhaps London. As an island joined to the mainland at low tide, it lay between the worlds of the living and the dead. (Some inhabitants of western shores still believe that the soul of a dying person waits for the tide to go out before leaving the body and beginning its final journey.) Like the tidal island of Ictis off the Cornish coast (probably St Michae
l’s Mount), where, according to Diodorus Siculus in the first century BC, smelted tin was carted out to merchant sailors arriving from the Mediterranean, Lugdunum may also have served as a neutral international trading post. It is better known today as a suburb of Dunkirk, the second largest French harbour on the English Channel. Perhaps the original ‘Dune Church’, first recorded in 1067, was heir to the ancient holy site, but it would take an enormous leap of faith to see the hand of Lugh in the glare of the halogen security lights that turn the sea beyond the port into a world of darkness.

  14. Gold coin of the Aedui

  This coin, which dates from the second or first century bc, was probably discovered at Tayac (Gironde). Medieval Irish legend talks of Lugh ‘the long-armed’. The epithet is thought to refer to the god’s spear-throwing prowess, but the same god can be seen here, ten centuries earlier, thrashing his sun-chariot across the sky with the long arms of daylight. The prototype of these coins was a gold stater of Philip II of Macedonia.

  The next two days were devoted to what seemed, at first, an austere, postmodern form of tourism – visiting sites where there was nothing to see, matching drab places on the ground to their colourful equivalents on the map, pursuing a journey of discovery that would discover nothing but itself. Along the meridian between Loon and the valley of the Somme (120 kilometres), Gaulish sites and places with Gaulish names occur with statistically significant regularity: three possible Mediolana, two Equorandas, two other ‘boundary’ place names and six medieval sites called ‘La Justice’ or ‘Les Gibets’ where criminals were hanged from gibbets. These places of public execution were traditionally located on tribal boundaries, especially those that lay on ancient roads. Places called ‘La Justice’ are four times as likely to occur along the Gaulish meridian as elsewhere in the region. But statistical significance is not the same as true significance. The Roman fort of Watten, which may once have been a Celtic oppidum, two empty fields called ‘les Gallois’, and a ‘Champ de Bataille’ named after a forgotten battle or a hoard of Iron Age weapons unearthed by a bemused medieval peasant were probably no more revelatory of a Celtic sun-path than the ruined starting ramps of German V1 rockets that lie on the meridian near a field called ‘le Rideau Mollien’ (one of the possible Mediolana).

  As we cycled past the soggy allotments and fishermen’s shacks along the river Somme, the mist rose from the marshes like the charcoal smudge of history book illustrations which serve to mask the areas of ignorance. A few metres from the meridian, we tethered the bicycles and walked along a gravelled path through the woods above the river. At the top of the escarpment, a wide field offered a view of the distant spire of Amiens Cathedral silhouetted against the morning sky. Amiens is usually said to have been the town of Samarobriva (‘Bridge on the Somme’), one of the most important tribal capitals in northern Gaul. A general council of all the tribes was held there in 54 BC, and it was at Samarobriva that Caesar spent the winter after his second invasion of Britain. Yet despite the bombs and building projects that have enabled archaeologists to rummage in the foundations of Amiens for the last hundred years, nothing from the pre-Roman period has been found there, and the true location of the Celtic Samarobriva remains in doubt.

  In the plains of Picardy, where a traveller’s eyes are filled with horizons, it takes some time to notice the intimate corners of the landscape. Just beneath the field, in a leafy hollow, was something resembling the back of a giant mammal slumbering in the woodland: we recognized the water-repellent roof of reed thatch and the tan-coloured daub of an Iron Age house. A path curled down into the hollow, where the tang of a turf fire hung in the air. An ancient Gaul had just completed his morning duties. Inside the house, swirls of smoke were trapped in a shaft of blue-grey sunlight slanting down from the smoke-hole. The man flung the fold of a plaid cloak over his shoulder, stooped under the lintel and the eaves, and disappeared without acknowledging our presence, which was understandable, since he would be spending most of the day trying to convince crowds of texting, tweeting schoolchildren that their Iron Age ancestors were not the unsophisticated brutes they had seen on television.

  15. Parc Samara and environs

  KEY (1–5 are field-names):

  1. Camp César or le Grand fort.

  2. Les Câtelets or Câttelets (diminutive of ‘castellum’).

  3. Camp Saint-Romain (a Christianized ‘Roman’) or Champ à Luzet (‘Coffin Field’).

  4. Fossé Sarrasin and Derrière le Fossé Sarrasin (‘Behind the Saracen Ditch’): ‘Saracen’ = ‘pagan’.

  5. Le Petit fort.

  6. Le (sic) Pierre and Chemin de Pierre: probably a paved road.

  (Two modern roads have been omitted.)

  There are probably fewer than a dozen reconstructed Iron Age houses in all of France, and the fact that one of them happens to stand on the Gaulish meridian seemed an amazing coincidence. This was supposed to be the first expedition in two thousand years – maybe the first expedition ever – to follow the north–south line from Lugdunum to Mediolanum Biturigum, and now it looked as though someone had not only stumbled on the secret but commemorated it with a fully functioning Iron Age habitat. But if the meridian had been rediscovered, there was no sign of this in the small museum of Parc Samara, ‘the nature park of prehistory’ (fig. 15). A sequence of illustrated panels guarded by a moustachioed Gaulish warrior on a rearing horse explained that this promontory above the river, which, like many other ancient fortified sites in France, had been known for centuries as ‘le Camp de César’, was the site of a Roman fort. Despite the unusual preponderance of coins from Massalia and a rampart in the local Gaulish style, the display suggested that ‘Caesar’s Camp’ had never been a Celtic oppidum. The Gaulish dwelling was based on an excavation at another site.

  Museum displays often have a misleading air of self-confidence. ‘Caesar’s Camp’ has been a mystery from the very beginning. In the spring of 1960, a schoolteacher from Amiens called Roger Agache flew over the site and noticed geometrical patterns in the fields to the east of the fort. From the open window of his biplane, he recognized the ghostly rectangles of a sanctuary complex. Agache, who died in 2011, was also a tireless sleuth on terra firma. He interviewed local peasants in the Picard dialect and collected tales of fairies’ trysts, vanished churches, and fields where the hand-plough suddenly sank into the ground as though the earth had been recently disturbed.

  The Roman fort was excavated between 1983 and 1993, and Parc Samara was opened to the public in 1988. Most of the area beyond the fort is private farmland and has remained unexcavated, but in the silence of the archaeological record, the old field-names are the captions of a treasure map. Stretching either side of the meridian and covering an area ten times the size of the Roman fort are the names that indicate the presence of a ruined Celtic oppidum.

  Parc Samara may be more important than it thinks. The medieval Peutinger Map, which shows staging posts on Roman itineraries of the fourth century AD (and perhaps also the first century AD), records a distance of ten Gaulish leagues between the previous staging post and Samarobriva, which is supposed to be the city of Amiens. But a traveller following this route – as Caesar and his legions would have done in 57 BC – reaches the future site of Amiens after only about half that distance.17 After the full ten leagues, the legions would have come to a Celtic promontory fort on the site of the future Parc Samara. This is the next place after Amiens that might have been called ‘Bridge on the Somme’, and it is far more likely to have been the tribal capital than the place now called Amiens, where nothing pre-Roman has been discovered.

  The siting of roads and bridges changed so little between the days of the Gauls and the French Revolution that the eighteenth-century maps are a useful guide even to this remote age. Here, at the site of Parc Samara, the Cassini map of 1757 shows a road heading southwest across the floodplain. The bridge, which may have been a wooden causeway like the jetties used by local anglers, crossed the Somme a few metres from the meridian, beyond
the south-east corner of the settlement. At this level of detail, modern maps which claim to depict ‘the Roman road system’ are of little use: their convincingly coherent patterns are the effect of a tidying instinct and an overestimation of Roman precision which makes all the roads point directly at the Roman town of Amiens. If the surviving sections of ancient road are plotted precisely, the map looks more like a puzzle of footprints at a crime scene. It shows not one system, but two. The road from Caesaromagus (Beauvais) heads straight for the Roman town of Amiens, but several other roads are distantly attracted, not to Amiens, but to the oppidum above the river Somme.

  These apparently chaotic lines are evidence of the great upheaval in the late first century BC. The capital that was imposed on each tribe by the Roman conquerors was rarely the same as the original chief oppidum. The Roman capitals usually lie several kilometres from the traditional tribal site and in a more convenient, less defensible location (see here and here). Shortly after the Gallic War, Amiens became the capital of the defeated Ambiani. The old oppidum was left to rot away until nothing remained of it but ditches, coins and boot-nails. Its name, Samarobriva, may then have been transferred to the new Roman town upstream.

  16. ‘Roman’ roads of northern France

  The surviving sections of road marked with arrows are oriented on the Celtic oppidum at Parc Samara (the original Samarobriva?) rather than on the Roman town of Amiens.