Your basket is empty.


There is no source that is not sacred.

Seneca, Epistles 41.3

The ruins of aqueducts and sepulchres lay about everywhere: ruins that seemed more like the forests and indigenous plants of an earth composed of the dust of the dead and the debris of empires.

Chateaubriand, Voyage en Italie


Running in a sequence of dismembered sections, the Roman aqueduct of Nîmes can be read as a kind of fragmented text. Its choked tunnels, collapsed archways, brief truncated arcades can be seen as the scattered passages, say, of some early ontological discourse. This discourse (dis– + currere: to run forth or asunder) expressed itself, no doubt, in terms of the fluent, the continuous. Even today, there’s not an isolated pier, a hydraulic cross-section that doesn’t draw one’s attention as some severed part of a dynamic ensemble: a hiatus at the running heart of a once unbroken continuum.

From the source onward, one could interpret the full length of this “linear edifice”1 as a sequence of scattered phrases, parenthetical clauses, even —at the very limit—sprays of infinitesimal punctuation. For the entire structure begs to be read. Out of so many disarticulated units, it seems to await nothing less than its own restitution through the agency of some synthesising grammar. The least element matters. Isolated plinths, conduits, panels of rose-red aggregate: there’s nothing that wouldn’t serve that restitution. Not a pebble that wouldn’t enter, finally, that inherent discourse.

Curiously enough, one hears the source before actually seeing it. It bubbles, quite audibly, from underneath the weeds of an abandoned orchard before gushing forth—fifteen, twenty metres later—as a limpid little brook. Why had the Romans chosen this particular site, these waters? The Fontaine d’Eure, located just beneath the perched medieval city of Uzès, satisfies the two basic conditions for determining an aqueduct’s point of inception. The first, of course, touches upon the quality of the water itself. It needed to be as pure, salutary as possible. Vitruvius suggests in his masterly De Architectura that one can determine the quality of water emanating from a particular source by the health of those who happen to live in its immediate vicinity. “If they are hardy, of fine complexion, subject to neither inflammations of the eye nor pains of the leg, then one can be assured of the water’s goodness.” Vitruvius also recommends sprinkling a few drops onto the finest copper plate available. If the drops leave no stain, then the water, he claims, may be considered perfectly pure.2

The second basic condition that the Fontaine d’Eure satisfies is that of altitude. It is located at a point sufficiently elevated to allow the water in its catchment basin to flow the full fifty-one kilometre length of the aqueduct—from, that is, the plateau of Uzès to the plains of Nîmes—powered by nothing more than the force of gravity. In terms of hydraulic engineering, the achievement was monumental. Not only did the waters of the Fontaine d’Eure traverse that expanse, they forded a major river (the Gardon), crossed at least eleven substantial highland valleys, and tunneled through several hundred metres of solid limestone before reaching—an estimated twenty-six hours and forty-five minutes later—the castellum or underground water reservoir of Nîmes itself. All the while it maintained a gradient that averages, throughout, no more than twenty-four centimetres a kilometre or, approximately, an inch every few hundred feet. Given the sheer immensity of the structure, this gradient can only be seen as something infinitesimal. Mass, here, was brought to serve virtually imperceptible increments of measure.

So, from the outset, the source, both for its altitude and the purity of its waters, must have attracted the libratores, the Roman land surveyors responsible for locating such waters and laying out the trajectory for an aqueduct’s passage. But for us, standing at the exact same point today, the aqueduct’s debut isn’t altogether that evident. At its inception, all we can see, running along the shadow-dappled edge of a public playground, are the first truncated sections of a derelict conduit. The first, interrupted phrases of that projected discourse. Call them, its initial stutters.

An aqueduct, however, no matter how dilapidated, draws the visitor with a distinct magnetism of its own. Unlike the oval outline, say, of an antique arena or the embedded “D” of an excavated amphitheatre, an aqueduct, sequential by its very nature, tugs at the senses, invites one to follow—in so many broken segments—its singular route across gullies and ridgelines, through orchards and scrub oak. Discourse, indeed. For we feel immediately involved, implicated in the running of this historic programme. It suggests, no doubt, other passageways: those that function at subjective levels unto themselves. So doing, an aqueduct offers us a set of analogous images for our own personal sense of rupture, fragmentation, discontinuity. Reading its vestiges, we cannot help but feel that we’re interpreting something of ourselves. That we’re attempting to reconstitute, no matter how subliminally, sections of our own, obliterated itinerary: that its text might very well underlie our own.


The Romans considered fresh running water as something precious, even sacred. Mirabilia aquareus, they repeatedly called it: the sheer wonder of the waters. Rome alone, at its peak, was serviced by no less than nine aqueducts and the city itself languished in the spray of an estimated 1,352 public fountains. Night and day, water would “overflow its basins, filling the city’s pools, brimming over the edges of both the dyer’s and the fuller’s vat, gushing up in the midst of gardens, pouring over into public baths and spacious thermae, then, purging latrines, end finally in those very gutters wherein everything, once, first originated.”3

Indeed, to appreciate the extent to which Rome invested its waters with near magical properties, one might, in passing, recall those origins. For Rome, at the beginning, was little more than a cluster of seven hills sitting in the midst of a fluvial marshland. Romulus and Remus were rescued as little babies from that very marshland, thanks to the intervention of the she-wolf. Much of Rome’s early history, as well as its founding mythology, was based, in fact, on the manner in which it came to drain those marshes and, at the same time, harness the distant spring waters of the Apennines to serve its growing needs. It might even be said that Rome (and, by extension, the entire Roman Empire) evolved out of a deep-seated dialectic between the still and the flowing, the stagnant and the animate: between an ever-inherent, regressive tendency towards death (the “death wish”) and that, far more manifest, towards life. To channel, sluice, furnish entire populations with fresh water must have represented, for the mens romana, an all-powerful affirmation of existence itself. It was a means by which Rome could express domination over its own paludal origins. Could surmount, like some lower level of consciousness, its own groundless past: the “Rome of Romulus”, as Cicero described it, “steeped in mud”.4

As perhaps its proudest achievement, the construction of aqueducts, whether for the mother city itself or any of its innumerable outlying colonies, followed the same rigorous set of operational procedures. First, as already mentioned, came the libratores. From libramentum, a noun signifying the action by which a particular surface may be determined level, the libratores employed no more than two basic instruments: the groma, an ungainly, stilt-like platform from which dangled four lead-weighted plumb lines, assuring, once adjusted, astonishingly accurate horizontal sight-of-eye readings, and the chorobate, a wooden table no less than six metres long upon which lay a traditional surveyor’s level: a glass tube with its ever-vagabond water bubble. The chorobate was particularly useful on windy days when the strict verticality of the groma’s plumb lines could no longer be assured. 

For us, making our way over the broken terrain of the Languedocian moors rife with gullies, abrupt ledges and sudden patches of impenetrable undergrowth, it’s virtually impossible to imagine how those libratores came to take such infinitesimally fine measurements. Not only did they manage to establish a gradient far too slight for our own senses to appreciate, they projected over that broken ground, in so many loops, chicanes, hill-hugging contours, a single seamless trajectory. Nothing mattered in all their calculations but the proper flow of the waters themselves. For the flow needed to be maintained at a speed slow enough to keep the waters from eroding the U-shaped channel of the conduit itself (the specus), yet fast enough to reach—in a single, uninterrupted run—the castellum of Nîmes without losing their essential freshness.

Ours, we’d claim, coming upon the barrel-vaulted entry, say, of some underground hydraulic section or a stranded pier, standing in the midst of a cherry orchard. Yes, ours, we’d repeat, taking an instinctive, proprietorial interest in something we can only claim by analogy, by the running metaphor of its disparate parts. These parts, however, suffice. Lying like clues to our own sense of a lost underlying unity, they become the objects—the cherished objects—of an intense scrutiny. Measured, correlated, treated as the semantically charged particles of that otherwise obliterated discourse, the least vestige receives extravagant attention.

A large part of the Nîmes aqueduct (of most aqueducts, for that matter) was constructed underground. Its buried sections account for over sixty per cent of its total trajectory. This was the rivus subterraneus, celebrated by Frontinus. Yet another twenty per cent runs flush against the surface of the ground (soil or bedrock), the conduit itself capped by a succession of heavy, overhanging stone slabs. These figures, however, shouldn’t surprise us. For doesn’t our own, hidden discourse—that elusive script—travel sub rosa from its very inception, concealing, as it goes, far more than it reveals? Suggesting in a subliminal grammar of its own so much more than it chooses to disclose?

Less than twenty per cent or nine kilometres of the total structure conforms to one’s traditional vision of a Roman aqueduct. Running overhead on either a solid, masonry support (opus caementicum) or, whenever its total height exceeded six metres, on a series of vaulted arches (opus arcuatum), the elevated sections of the aqueduct astonish us with their inherent unity. Even if most of those sections have undergone irreparable depredation and we find ourselves staring, for the most part, at truncated vestige, we’re struck, nonetheless, by an uninterrupted sense of the serial—the rhythmically determined—running throughout. There’s not a single, stranded archway that doesn’t exude—out of every pore of its still adherent, still articulated masonry—something of the pulsate. Caryatids, we might call them. In an unbroken processional, these water-bearers had once cradled that precious element as it travelled at a rate of no less than a hundred cubic metres a minute against the crook of their mineral necks.

We’re involved, here, with movement. Despite the apparent petrification of so many isolated elements, we’re witnesses to an initial dynamic: that drafted by the architects themselves. For no sooner had the libratores, those expert land surveyors, completed their work (laying out the aqueduct’s trajectory and clearing, at the same time, a swathe of ground thirty metres wide for its construction) than the responsibility for the entire project devolved upon the architecti. Their role, according to antique sources, was exclusively conceptual. They were charged, first and foremost, with maintaining proportio: a word we might translate as “harmony”. It’s this very harmony, this underlying sense of architectonic unity, that we continue to recognise in visiting the aqueduct’s surviving sections. Masters of number, scale, and proportion, the architecti were charged with creating—out of an endless succession of telluric accidents—the vector of a single, singular, uninterrupted, water-bearing conduit.

The final responsibility for the construction itself, however, fell upon its redemptores or building contractors. It was they who chose the materials for the entire edifice and oversaw its execution from beginning to end. We are fairly certain that these redemptores or locatur operis were native to the region in which they operated. They needed to be perfectly familiar with whatever resources happened to be immediately at hand, most especially in regard to quarries: to the quality and characteristics of the stone in each, and the competence of the quarriers from one site to the next. Furthermore, they needed to provide immediate answers to an endless series of logistical problems touching upon the acquisition, transportation, and final deployment of every conceivable material, from crushed amphorae for the cement aggregates to monolithic keystones weighing, quite often, several tons each. Their familiarity with the region needed to be total.

It’s virtually impossible to imagine the sheer magnitude of the work-sites these redemptores were called upon to oversee. We know that tens of thousands of labourers—mostly slaves—executed the bulk of the work under the immediate command of the military. Then, too, innumerable artisans, specialists in their own particular fields, assured the execution of the finer, more applied aspects of the overall construction. Along with the presence of so many teeming quarriers, stone carvers, and masons, there were the carpenters, assembling, in situ, the indispensable trusses, scaffolds, and hoists for the overhead sections; the blacksmiths, honing tools and pounding out heavy, iron holdfasts to anchor one monolithic voussoir against another; the itinerant lime-burners supplying the stonemasons with an unending quantity of fresh mortar as they travelled—at the same, relentless, yet imperceptible speed—down the full length of that gargantuan work-site. To all the above must be added the carters, cartwrights, the entire armies employed in the transportation of everything from firewood for feeding the lime kilns to the meats and fruit and wine jugs for nourishing that immense population in its immeasurably alien migration downward.

In reviewing the multiple aspects of the overall project, we mustn’t lose sight of the one structural element that ultimately mattered: the specus or conduit. For there wasn’t a bucket of mortar or a spadeful of gravel that didn’t serve, finally, as support for that sleek, lime-faced, U-shaped conveyor. Measuring no more than 1.3 metres in width and 1.8 metres in height throughout its full fifty-one kilometre throw, the conduit was the raison d’être, after all, for the entire structure. No matter how deep the aqueduct’s tunnels or colossal its vaulted, overhead arcades, nothing mattered, in the end, but that narrow, meticulously prepared corridor in which the water, precious as it was salutary, would run in a single, unbroken current. When the specus itself wasn’t carved directly in rock, its bedding consisted of poured cement plaster, its parallel uprights of joined stone masonry, and its overhead vaulting of massive tabletop limestone slabs. Throughout its entire trajectory, in fact, water would run sheltered beneath one kind of covering or another. Thus protected from the intrusions of man and animal alike as well as the injurious effects of sunlight (“ut minime sol aquam tangat”, specifies Vitruvius: the less sunlight touching the water, the better5), it would arrive perfectly fresh at its destination. Infallibly, that destination would be the city itself. As one archaeologist has phrased it, if an aqueduct was essentially rural in its trajectory, it remained ultimately urban in its ambition6. For the Romans, with their vision fixated upon the city—the idea of city as the concentrate of every civilised value—the aqueduct could only be at the service of urbanitas.

It should be noted, in passing, that the specus or corridor in which the water travelled lay lined with several coats of waterproofing mortar. Called opus signinum, this mortar consisted of a lime in which a rich aggregate of broken tile, brick, and crushed amphorae had been added. Applied in two, sometimes three layers, the mortar was then sealed by a thin, lacquer-like wash of scarlet coloured pigment. According to Pliny the Elder, this ultimate layer of waterproofing was composed of nothing less than quicklime, slaked in wine and mollified, in turn, with an admixture of fig juice and pork fat. The result, he claimed, was the most tenacious substance on earth (“resomnium tenacissima”) and would outlast even stone in sheer durability7.

It’s up to us to imagine these waters flowing at such speed, with such abundance, reflecting in their passage the unctuous reds of their immediate surroundings. To imagine the unimpaired flow of thousands of cubic metres per hour travelling—toboggan-like—down a whole series of sinuous curves only to enter, immediately after, a nearly flat, nearly imperceptible gradient for the sake of reaching—a full day later—the deep recipient basins of Nîmes itself. Up to us, finally, to imagine such a seamless construction; reconstitute, out of all its sparse, surviving segments, such a resolute view of the indivisible.


There are two ways, essentially, to read an aqueduct: ideally, as a resuscitated ensemble; or materially, as a suite of dismembered sections. We’ve chosen the latter. Beginning with the little that remains, we’ve gone from one isolated archway to another, one stout little culvert (smothered nowadays, in ilex, juniper, the iridescent pink sprays of the terebinth) to the next. Haunted, no doubt, by other conveyors, we’ve created an analogy of our own. Out of an innate sense of some fundamental continuum, running in the deepest channels of our consciousness, we’ve come to interpret vestige, here, as script, reading its fragments as ours.

Curiously enough, this analogy, this extended metaphor, seems to lose its effectiveness in the face of the aqueduct’s one major surviving monument, the Pont du Gard. For the bridge itself is far too well preserved to invoke one of those lost, mnemonic passages. It might even be said that the Pont du Gard suffers, in this context, from a certain absence of absence. Rising a full forty-eight metres over the green pellucid Gardon on three successive tiers of vaulted masonry, this “masterpiece of functional architecture” suggests nothing more than itself.8 Jean-Jacques Rousseau, visiting the bridge in the 1730s, would exult: “The overall appearance of this noble if austere work impresses me all the more for standing in the midst of a wilderness where the surrounding silence and solitude render the monument that much more striking and one’s admiration that much more entire, given that this so-called bridge is but a simple aqueduct.”9

As for ourselves, however, we need to explore that “surrounding silence and solitude”, need to rummage through the heavy undergrowth of these Languedocian moors for the sake of that fracture, that fragment, that stray archaeological signifier lying in the midst of our long vanished discourse. It’s as if only the shattered, the disjuncted, had “something to say”, as if only the estranged could speak, finally, in the name of that all too elusive integer.

So here we are, having crossed the Pont du Gard (crouching as we went within the narrow corridor of its high overhead specus) and entered, immediately after, the Bois de Remoulins. From this point onward, the aqueduct breaks, once again, into disconnected sections. Moving, now, through a wilderness of low, wind-stunted dwarf oak, we feel a sudden rush of pleasure whenever we happen upon a few metres of masoned conduit in one place, the rounded right angle of an “elbow bend” in yet another. For, so doing, we’ve identified a portion of that decimated itinerary. Encountered, once again, a piece of our own obscure puzzle. Recognised, in a brief corridor of bedrock, part of that running, self-proclaimed metaphor. For it’s not an aqueduct, finally, that we’re tracking, but ourselves: not the disparate sections of some historical monument, but the diaphanous outlines of some long-since obliterated human programme.

Monument and metaphor: the history of one illustrates the ideation of the other. Through an examination of the former, we might come to a greater understanding of the latter; taking measure of so much given, material evidence, come to appreciate all the more the drifting vocables of our own drawn analogy: that “parallelogram” we’ve created for ourselves.

We should begin, then, by distinguishing three distinct periods in the history of the aqueduct: that of its construction (essentially under the reign of Claudius in the middle of the first century A.D.); that of its maximal flow and distribution (a period covering the subsequent three centuries); then that, by far the longest and easily the most evocative, of its dismantlement and ultimate destruction. If we linger, here, on the last of these three periods, it’s only to appreciate more fully the analogies we might draw in terms of our own personal sense of the disaggregated. For comparisons clearly abound. The breakdown of such a monolithic public service, functioning for the general benefit, had already begun. We find, for example, the first cases of illicit water-tapping (the clandestine diversion of that public commodity for the sake of satisfying private, essentially agrarian needs) already manifest by the end of the fourth century. These first instances of pirated waters can be attributed to local landowners deliberately creating breaches in the aqueduct or simply canalasing its spillage for the sake of irrigating their crops in immediately adjacent areas. Piles of porous, tufaceous concretions, laying flush against the walls of the aqueduct, still attest to this practice. These piles, standing like frozen cascades, became more and more frequent as the fifth century progressed, and indicate a profound change in the operational life of the aqueduct. Colonists (essentially retired centurions, massively compensated) began replacing, at this time, the small autochthonous landowners. These new settlers would tap and thus reduce the flow of the aqueduct in ever-increasing quantities. We can only assume that they did so with the tacit approval, if not the active complicity of whatever authorities happened to be responsible for the water’s distribution.

We have then, and increasingly, the first indisputable evidence of a breakdown in the functioning of the aqueduct as originally conceived: as a uniform construct at the service of a singular entity: the res publica. More and more, the vested interests of a new landed gentry would determine the flow and distribution of that public commodity, and the aqueduct come to serve the few at the expense of the many. “Great villas and the large domains that sprang up around them with their constituent settlers (coloni) would lead to the total breakdown of the Roman aqueduct. From this new set of conditions,” writes A. Grenier, “a closed economy would emerge, replacing the rich exchanges that once existed in which cities served as metropolises.”10

Archaeologists, studying both the quantity and quality of these waters, their “palaeoflow” as evinced by tiny laminal, concretionary deposits found along the inner walls of the specus itself, have detected traces of increasing turbidness throughout the fifth century. Concretionary deposits have given them a remarkable basis for interpreting the life of the aqueduct, especially in its years of degenerescence. They’ve learned, for example, that the fresh, running waters that travelled down its conduit through the second, third and fourth centuries left a clear, near crystalline deposit called “travertine”, but that those same waters, a century later, obstructed by organic matter, tended to grow muddy and turbid, generating, as concretion, a rough, porous substance called “tufa”.

Clearly, the aqueduct no longer received the kind of continuous care that it had in the past. Circitores, maintenance workers charged with its upkeep, no longer journeyed up and down its length, scrubbing its walls clean and keeping vegetation from invading its low, lidded, ongoing chambers. Roots now dangled from the vaulted slabs overhead. Reaching into the canal itself, these roots not only inhibited the flow of the waters and raised, consequently, their level, but they also introduced the algae and bacteria which, in decomposing, would result in the aforementioned concretions. This process, called biolithogenesis, is caused by the interaction of karstic waters heavily charged with lime or calciferous matter, and organic substances introduced by the root systems themselves.

Increasingly choked by its own concretions, the aqueduct would supply less and less water to the civitas of Nîmes. By the end of the fifth century, its yield had fallen from an optimal 124,000 cubic metres per day to an estimated 14,630. At the same time, the quality of this water had gone from something limpid, crystalline, to turbid, unpotable, fit for little more than purging the city’s underground sewerage system. We may assume from these figures that the aqueduct must have suffered increasingly from the absence of a single centralised agency responsible for its upkeep. In fact, everything indicates a breakdown, an atomisation of any such authority. The unity of this eminently transmissive structure could no longer be assured.

We may also assume that the last to benefit from its waters weren’t the citizens of Nîmes, but those prosperous landowners, latifundians and the like, on their flourishing plantations. Turbid or not, the waters would continue to irrigate their orchards and vineyards, and they themselves go on benefitting individually from what had once been an exclusively public commodity. Metaphor, indeed. For how can we help but see—as if mirrored within this particular historic complex—an image of our own collective breakdown? How can we help but discover, in these remote landscapes, a paradigm for our own structural fragmentation: in its diverted waters, elements of that very same truncated discourse?

Worse, far worse, awaited the aqueduct. Having undergone de facto privatisation and fallen into the hands of so many disparate interests, it would be reduced from a rushing cataract to a mere trickle of turbid water. Even that trickle would undergo sporadic arrests as, for instance, during the Great Invasions when Nîmes was repeatedly besieged. By the sixth century, that trickle, that meagre yield, had ceased. Obliged to erect or reinforce their defence systems against the invading Franks and Visigoths, local populations fell upon the aqueduct itself as a storehouse of readily available materials. Out of its hydraulic masonry arose ramparts, watchtowers, strongholds. No doubt the invading armies, both Franks and Visigoths, did much the same themselves. Warring simultaneously against one another and the local Gallo-Roman populations, they ransacked whatever they could. Within a brief period of time, then, this singular construction, the ghost of that once gushing conduit, was converted into a public stone quarry. One could pillage at will. With its massive blocks pried loose, disassembled and carted off, the overall structure came to resemble some slain mastodon, its flanks hacked into so many hopelessly divided sections.

Throughout the Middle Ages, the aqueduct would continue to suffer the abuse (or simple indifference) of historic circumstance. Alain Malissard in his study, Les Romains et l’eau, has drawn a moving portrait of those dark times. Everywhere in the ex-Roman Empire, he tells us, aqueducts suffered the same fate. “The underground canals would slowly collapse under the weight of rubble and ruin; vestiges of an impossible level of well-being, hopelessly encrusted now, choked in silt, cracked, perforated, dismantled by anyone who happened to be in search of building materials, the overhead sections of the aqueduct would tumble, one after another, into a countryside gone thoroughly fallow.”11

This demolition wouldn’t reach its peak, however, until the Romanesque, five centuries later. Then, in what often has been called “the first renaissance”, a whole corpus of monks, lay-brothers and ecclesiastics of all rank, serving as masons, stonecarvers, and architects would set upon and pillage the aqueduct in an unprecedented manner. There was an out-and-out ravaging of the edifice during the period between 1050 and 1200. In deciding which particular sector to reclaim, three critical factors determined the choice for that befrocked workforce: first, that the sector be part of some overhead section (the stones there being more accessible, easier to dislodge); second, that the sector be as near as possible to the work-site itself, whether chapel, church or monastery; and lastly, that no other source of building material (a contemporary quarry, for example) be within immediate reach. For that reason, some of the finest surviving segments of the aqueduct run directly through areas still rich in quarries. As for the rest—the greater part of the aqueduct, that is—the Romanesque would make short work of that massive structure. Pried loose, disassembled, its stones would reappear now out of all established context. Sprouting in the form of plinths, capitals, altar pieces, they’d have undergone total dislocation. Far more than the quarried stones themselves, however, the concretions lying flush against the specus—that once waterproofed conduit—would constitute the most prized building material of all. Easy to carve and offering a single, perfectly uniform surface, these travertine encrustations could be dressed into highly decorative blocks resembling—in both granulation and lustre—marble itself. Scarcely more than fossilised water, than the petrification of that water within so many successive laminal plates, these slabs would make their way into a jambpost here, a lintel there, a particular architrave yet there again. Today, following the last remaining traces of the aqueduct itself, one is likely to find, within a swathe of land no more than two or three kilometres on either side, blocks of that lovely concretion rippling like candy stripes in the midst of so much commonplace masonry. Not only chapels, churches and monasteries became the beneficiaries of this legitimised plunder, but castles, keeps, outlying domains, even—on occasion—dovecotes, barns, the least agricultural dependency. Everywhere, individual interests, secular or sacred, came to feed on that long, winding, vectorial ruin. The res publica, that prescription in favour of the all, ended in this free-for-all pillaging by the each.

The aqueduct’s dismantling continued well past the Romanesque. Each succeeding period in history would bring its own distinctive contribution to its unmaking. One might even postulate a typology, an entire methodology, devoted exclusively to the praxis of its demolition. During the seventeenth and eighteenth centuries, for instance, peasants reclaiming fallow land in the surrounding hillsides needed a considerable quantity of stone for creating restanques: terraced retaining walls against soil erosion. The stones for these bulwarks came, often enough, from that long-since desiccated watercourse. The aqueduct’s once seamless structure was little more, now, than a narrow, ongoing strip of disposable rubble.

However, it was the construction of an irrigation canal, the Canal de Pouzin, in the 1860s that would strike a last, definitive blow to the edifice, most especially in the final fifteen kilometre leg of its trajectory. Here, the hydraulic engineers of the last century could think of nothing better than laying their own canal directly over that of the Roman: pouring, that is, their own industrial cement over the antique opus signinum, not only cloning the aqueduct but smothering it irreparably in a glut of proto-modern agglomerates.

The aqueduct today, as it approaches what was once its final destination, the castellum of Nîmes, seems to vanish altogether. It’s as if our text, our imaginary discourse, at the very moment when it might have reached some kind of conclusion, underwent a form of deletion. Suffered a kind of censorship. Granted, one can still come across the brief parenthetical clause of some barrel-vaulted conduit, here; the paraphrase of some sinuous, subterranean diversification, just there. Any hope, however, of bringing the sporadic fragments of that tenuous discourse to closure has long since vanished. In its final kilometres, the aqueduct finds itself consumed by a succession of suburbs, commercial centres, industrial parks. We’re no longer drawn, guided now, by visible evidence: by a rhythmic set of vaulted arches, or a single Piranesian archway standing in the midst of some rocky, Cezannesque landscape. There’s not even a three-metre swathe of stunted wheat running clear across a grainfield to indicate the unmistakable presence, just beneath, of that indomitable edifice. Caught now in the labyrinthian underworld of a fully serviced modern city, the aqueduct dies an obscure death. Running beneath the streets, gardens, workshops of an increasingly congested metropolis, it finds itself hopelessly broken into brief, isolated units. Some of those units have been converted into individual cisterns, cesspools, latrines, serving their particular proprietor immediately overhead. Yet others have seen their vaulted chambers turned into private storage spaces, wood-sheds, even wine cellars. Of that gargantuan, fifty-one kilometre thrust of pure human inspiration, motivated by nothing less than a single overriding public necessity, little if anything still remains within the cellulated underworld—the larders, crypts, choked passageways—of contemporary Nîmes.

Quite recently, yet another form of depredation has come to undermine the edifice. This time it’s not the structure per se that finds itself endangered, but the very idea underlying its existence. Several archaeologists have come to question the motivations that led to such an extravagant construction in the first place. Was it, in fact, really necessary, they’ve asked? Didn’t Gallo-Roman Nîmes in the time of Claudius already possess a sufficient water supply, thanks to both an abundance of natural springs and a rich network of cisterns for recuperating rainwater? Wasn’t the elaboration of such a colossal project merely an act of manifest bravura, undertaken more for the sake of prestige—sheer ostentation—than for the service it actually rendered in the form of so much pure, gushing, unarrested waters?

We, of course, can only reject such an argument. We’ve grown far too attached to the aqueduct, to its necessity both as monument and metaphor, structure and sign, to endorse such a slender hypothesis that figures alone—the simplest calculations—could readily refute. Furthermore, these ruins, we feel, have long since become an indispensable part of our culture. Even hopelessly dilapidated, they still manage to signify a sense, indicate a direction, offer (no matter how segmented) a magnanimous image of the sequential in its full, functional capacity as distributing agent. It’s not for nothing that we’ve attempted to read, and reading, interpolate those gagged channels and collapsed archways in terms of a truncated, antique text. Not for nothing that we’ve come to recognise—no matter how diaphanously—whole portions of that limpid, life-sustaining discourse as if inscribed within its still surviving sections. Ruins like this, after all, are what we’ve inherited. What, ultimately, we possess. Their lost grammar—almost certainly verdant, pellucid—is what we’ve been given, in turn, to interpret, transcribe, perpetuate.


1 Jean-Luc Fiches and Jean-Louis Paillet, “Prospections et fouilles: Archéologie d’un  aqueduc”, L’Aqueduc de Nîmes et le Pont du Gard, (Conseil Général du Gard/CNRS, 1991), p.252.

2 Vitruvius, De Architecture, Book VIII, 4.1.

3 Alain Malissard, Les Romains et l’eau (Paris, 1994), p.18.

4 Cicero, Letters to Atticus, 2.1.8.

5 Vitruvius, op.cit., Book VIII, 6.1.

6 Philippe Leveau in “L’Aqueduc de Nîmes et Le Pont du Gard, p.245. The phrase in its entirety reads: “Monument typiquement urbain par sa finalité, rural par son tracé, l’aqueduc apparaît comme un excellent symbole de la romanité.”

7 Pliny the Elder, Natural History, Book XXXVI, 58.

8 Malissard, p.177

9 Jean-Jacques Rousseau, Les Confessions, Book 6.

10 A. Grenier, Manuel d’archéologie Gallo-Romaine (Paris, 1960), p.226.

11 Malissard, p.268.

More from this issue

Requiem for a Heavyweight

My impression of Murray’s critical reception is that it is partly based on laziness: critics don’t want to deal with him. They don’t want to argue, don’t want the bad light the poems put them in as adversaries: politically correct (how unsexy) and less mobile verbally than the poet himself.
Read more

On Reticence

Would you knowingly invite a reticent person to dinner, someone like my Oxford supervisor, or my friend who believed in sealing her lips? If you had no choice, how would you deal with such people? Sit them next to Chatty Carl or Commandeering Cathy in hope that their quiet attentiveness would mop up the attention-seeking of the other. The voluble need vessels, after all, into which to pour their words.
Read more