In a postscript to his lesser-known work Histoire d’une Forteresse (1874), Eugène-Emmanuel Viollet-le-Duc (1814-1879) describes the diffusion of fortifications brought about by the introduction of ranged artillery. The morphology of fortified towns was redefined by the need to operate in increasingly large arenas of conflict.

Histoire d’une Forteresse recounts the history of a fictional town in a valley in eastern France. Positioned strategically, it undergoes a series of sieges and rebuildings illustrating the practice of fortification from Stone Age Gaulish tribes to the Napoleonic wars of the early nineteenth century. Some of the town’s seven sieges result in it’s capture, while in others the town is able to hold out against the invaders. Illustrated with the lucid metal engravings characteristic of Viollet-le-Duc’s books, the text describes military architecture as fundamentally a problem of tactically exploiting matter. Neither tactics nor constructed matter alone are able to keep the fortress secure or overrun it. Instead security consists in a strategic conjunction between matter (the physical matter of the fortress and the surrounding terrain) and military practices. The fortess is a practiced site.

The conclusion, which is the only part of the text we will be concerned with in this paper, purports to be extracted from the notes of one Captain Jean, dated 1871, a year in which French military confidence had been shaken by the Prussian seige of Paris. “Attack,” writes Jean,

“implies a shock or onset; defence is a resistence to this onset. Whether a piece of ordnance discharges a ball against a plate of iron, or a casing of masonry, or an earthwork; or an assaulting column climbs a breach, the problem is substantially the same; in either case we have to oppose to the impulsive force a resistance that will neutralise its effect.” (Viollet-le-Duc, 1876: 357)

Defensive strategy, tactics, and technology therefore, aim to oppose, resist, and neutralise the effect of shock, at whatever scale it comes. Jean’s brief essay uses a series of ten geometrical plan diagrams to describe a fundamental shift in the nature of fortification, brought about by the shock of artillery.

In the simplest case of attack, a series of blows are exchanged in one-to-one combat. Advantage can be taken by outnumbering the enemy, or alternatively, by surrounding oneself with an enclosure. Jean’s first drawing shows that a group protected in this fashion strictly limits the way attackers can engage them. No matter how the attackers configure themselves, or what numeric advantage they hold, they can only confront the defenders individually.

When ranged weapons, “projectile arms”, are taken into account, the attackers receive an advantage. They can set up a line of weapons which can converge their fire on a point on the circumference of the defenders. Because of the nature of a circular fortress, the defenders are extremely limited in their ability to converge fire on the attackers. They are constrainted to scatter their fire outwards, while the attackers can concentrate it inwards. To ameliorate this asymmetry, the fortress can be constructed with external appendages, salients, that allow the defenders to establish a broader base of fire, as well as offering better visibility of the foot of the main wall. In response, however, the attackers can reconfigure, beginning their assault by concentrating fire on the exposed, outflung salients. Defender and attacker compete for the ability to converge fire on the other while remaining protected. This is the central tension of fortification, which Jean writes,

“regulates and will always regulate attack and defense; distances alone modify its applications… The more eccentric the defence is, the more distant must be the attack, and the wider the perimeter it must occupy; but it should be observed, that the more widely the defence is extended, the more open its flanks are to attack.” (Viollet-le-Duc, 1876: 360).

The 17th century fortifications of Sebastien le Prestre de Vauban (1633-1707), based on the star-shaped forts of the Renaissance, work according to this principle [3]. By extending salients out from the main fort to form bastions, the defenders create a wider line of fire which can be concentrated at any point on a circumference around the fortress corresponding to the range of the artillery, while minimising the exposure of any individual salients to concentrated fire from attackers. In the fourth of his sequence of diagrams, the geometry of the star-shaped fort is seen to be partially material (in the form of the stone walls), and partially immaterial (in the form of sight-lines, firing ranges, and trajectories). These lines of force or agency form a coherent system with lines of solid material. The trajectories do not simply inform the walls, nor do the walls simply define the trajectories. These immaterial lines are as essentially architectural as rows of cut stone. Captain Jean notes that this worked very well for the ranges of the artillery in use in Vauban’s time, but as the range of artillery increases (he notes that late nineteenth-century artillery was effective at ranges of eight or nine kilometres), the balance swings again. In operations at these greater distances, the fortress is proportionally reduced to a point. The attacker, “taking advantage of an indefinite amount of space” is hidden and free to move, while the defender, “soon encumbered with débris of all kinds” is slowed and confused. Under these asymmetrical conditions of mobility, Jean says, the end result is never in doubt. The solution? “In proportion to the length of the trajectory, therefore, the defence must remove its defensive arrangement from the centre of the place.” (VIollet-le-Duc, 1876: 363).

In the sixth diagram in this section, Viollet-le-Duc re-plots Vauban’s geometry for these larger dimensions. Vauban’s star-forts presented fronts of about 360m, but the new situation called for fronts twelve or thirteen kilometres long. To achieve this, the salients are detached from the main fortress; instead of bastions, they are distributed as separate forts forming a defensive ring. They are positioned at the limit of their own firing range from the central fortress, so they themselves cannot be used to bombard it. The arrangement is then refined into a dual ring of forts, so that no one fort is exposed without other forts being able to provide supporting fire from the flanks. In the drawing, the solid material of the fortress appears as a small hatched mass embedded in an increasingly complex field of radii and circumferences demarcating the domains of projected firepower. Railways are necessary to connect the distant forts, and communication lines must be kept open. The territory surrounding the fortress proper becomes a zone of mobility.

In the eighth drawing, the solid material of the fortress walls is given no presence at all; the entire fortress has dissolved into arcs and radii, with certain key points marked algebraically. However, the fortress has not become obsolete, but ubiquitous. The ninth drawing reveals that each individual fort retains the articulation of salients and angles appropriate to smaller arms and closer quarters. The efficacy of the new system was no longer in the solidity of a fixed centrepoint but in the furnishing of a territory. In the tenth drawing, we are shown how this abstract geometry is deformed by the terrain itself. We see the fortress with its wall, but the arced lines of the wall no longer indicate the presence of solid material, as they did in the first drawings. The wall is the extent of the artillery’s range; it is a horizon of effect surrounding a zone of mobility rather than an enclosure.

[ This older post from David Gissen’s HTC Experiments makes an interesting connection between Viollet-le-Duc and conceptions of crowds in the nineteenth century. ]

[ Part II ]

This is the first section of a paper to be published in the forthcoming issue of Interstices. It argues for the existence of implicit theories of assemblage. I think overall I’m trying to make a connection between architectural assemblages and social or political ones – a stronger connection than just suggesting that the former represents the latter. Something like this is going to feed into my PhD research (whenever that finally gets going). I haven’t included the footnotes or anything, and it ends a bit abruptly. And its far too long for a blog post.

Laugier, Frontispiece to 2nd ed. of Essai sur l'Architecture, 1753

“Man was born free, and he is everywhere in chains.” The Social Contract begins with chains, and remains entangled in questions of binding. The chain is a figure of arbitrary constraint, and is represented as something to be thrown off. But in Rousseau’s text it is not a matter of aspiring to a state of absolute unconstraint. The very concept of society, of a social order, implies some kind or degree of attachment; and it is the proper form of this attachment which is the concern of The Social Contract.

Rousseau makes a primary distinction between the arbitrary bond of the chain and the natural bond of the family, “the oldest of all societies, and the only natural one” (1968: 50). The child is bound to the father by necessity (the maternal bond is never raised), and once the child becomes independent, this bond dissolves: the child and father are freed from this relation, and if it persists, it is by mutual consent: “If they continue to remain united, it is no longer nature, but their own choice, which unites them; and the family as such is kept in being only by agreement” (50). In this shift from dependence to agreement Rousseau locates the shift from the natural to the social. All legitimate authority, asserts Rousseau, must be based on agreement, and he sets himself the task of describing a society of this kind. Rousseau, who has occasionally been misunderstood as advocating a return to nature, actually describes the social as a second nature[1]. Natural order does not authorise social order. Social order must consist of covenants, freely entered into.

As Mark Wigley points out, Rousseau explicitly describes the constitution of social order as a building project, for which the ground must be cleared and tested, the structure carefully maintained, and collapse avoided, “as an architect who puts up a large building first surveys the ground to see if it can bear the weight”  (Rousseau, 1968: 88; see Wigley, 1993: 133). The state is a collective identity formed by very specific relationships between individual elements. By freely entering into the social contract, an aggregate is formed, a corporate body, a “public person… once called the city” (61). This agglomeration is given its internal cohesion by the social contract subscribed to by each individual. The contract is the fundamental joint, the bond or bind by which the entire social edifice takes shape and holds together. The social body acquires unity, life and will.

Although a social whole is formed, however,  the parts must remain autonomous, such that each individual has a private will distinct from the general will: “His private interest may speak with a very different voice from that of the public interest” (63). This freedom runs to the extent that the individual may at any time withdraw from the contract entirely. Society exists only so long as the social contract is freely maintained by its constituents. The freedom to renounce society is essential. The joints of Rousseau’s social structure must not be bound or fused. There cannot be forceful constraints in the social contract.

Rousseau’s social contract is in many respects a gloss on Hobbes’s Leviathan. Hobbes proposed that the state should be conceived as a collective body, of which the sovereign was the head. The famous frontispiece of Hobbes’ treatise show what he has in mind: a body comprised of individual humans as cells, all looking up towards the sovereign. Rousseau’s innovation is in shifting focus from the exterior relations to interior relations. Where Hobbes begins with the image of a human organism, and proceeds to show how society can be fitted into this authorising metaphor, Rousseau begins with individual connections, and attempts to discover what the whole body might look like. [fn useful critique of the organismal metaphor in delanda] Put simply, where Hobbes tendentially assumed the primacy of social form, Rousseau was concerned with social formation. [fn for a fuller discussion of Hobbes’ Leviathan, see McEwen]

Joseph Rykwert has suggested a correlation between Rousseau’s primitivism and that of his contemporary, Marc-Antoine Laugier. The famous frontispiece image of Laugier’s Essay on Architecture (1753) is one of the key coordinates for Rykwerts study of the idea of the primitive hut in architectural theory, On Adam’s House in Paradise. Laugier proposes that the basic elements of the classical tradition in architecture are already present in an imagined primitive scene: seeking environmental control over light, heat, dampness and air, a primitive man finds four trees arranged in a square, and constructs a raised roof; thus inventing column, entablature, and gable. Rykwert writes “Allowing for the inevitable differences between the two men, and the differing scale of their enterprises, this view of the authority of the primitive hut is not unlike that which Rousseau attributed to the family as the archetype of social organisation” (Rykwert, 1981: 44).

In his The Contribution of Art and Science to the Refinement of Manners, Rousseau describes in parallel the socialisation of human beings, and the degenerate elaboration of architecture: “Here is a calm riverbank, dressed by the hand of unaided nature, towards which the eye turns constantly, and which you leave with regret… then came the height of degradation, and vice was never carried so far as when it was seen, to speak figuratively, supported by marble columns and engraved on Corinthian capitals” (Rykwert, 1981: 46-47). How to properly house human beings is a question allied to that of proper social relations.

In his drawing for the 1755 edition of the Essay on Architecture, Laugier’s hut is conspicuous for its structural self-sufficiency. The individual elements: the still-living columns, the cross-beams and the rafters, all rest together naturally, without pins or bonds. The four tree-columns have been pruned, and the stumps of the branches become brackets to support the beams. The trees retain their leafy growth, except possibly for the front left tree, which looks as if it has been trimmed back to the trunk. The rafter branches sit up at an improbably steep angle. They rest on the beams without any evident support: under close inspection, the expected bindings are found to be absent, and the rafters do not appear to be notched onto the beams. At the ridge, the rafters rest against one another. A ridge-beam is possibly hinted at, but looks as if it is suspended under the rafters rather than providing any substantive support. Again, there is no hint that the rafters are bound or pinned together at the top; and they cannot be interwoven, because the branches are conspicuously thick and blunt.

Perhaps the gesture of Architecture personified in the foreground could be re-interpreted as a gesture of blame for the collapse of the Ionic edifice in the foreground that has attempted to follow the structural logic of the Laugier’s hut – in which case it is no wonder that the cherub appears shocked.

It is evident, of course, that Laugier did not intend his hut to be understood as an exemplar of construction practice, but as a moral “first model” (Laugier, 1756: 11). It is used to demonstrate the essential elements of architecture, and to exclude those elements which are superfluous additions, “essential defects” (12). If performs the same role (and has the same anthropological nonspecificity[3]) as Rousseau’s primitive family. But to point out the strange condition of the joints in Laugier’s image is not entirely perverse – this model does, after all, deliberately express principles of construction. And in fact, the disjointedness of Laugier’s hut is entirely consistent with his thinking about architectural attachment, and the relationship between part and whole. In the Essay on Architecture, there is little written directly concerning joints. Perhaps consideration of joints is included amongst those details with which Laugier felt disinclined “to load this little work” for fear they might “trouble and distaste the reader” (xvi). Connection and attachment are, however, important subthemes of Laugier’s text.

In the chapter of his essay which directly addresses construction, the strength of a building is said to depend on the choice of good material, disposed with consideration of loadpaths and bearing. Laugier writes, “There are three things which render a wall strong and immoveable. The foundation upon which it bears its thickness, the connection and right line of its parts” (138). It is obvious that in his text he has in mind one type of joint, stacked masonry; this is in spite of what he has asserted about the timber origins of architecture. Stones are to be laid accurately and tightly, “that there may be no void in the thickness of the wall” (141), and the use of mortar, a concession, is to be minimised. Laugier’s ideal structure would be held together by nothing other than gravity. Beams are “laid” on the columns. Columns are to “bear immediately upon the pavement, as the pillars of the rustic cabin bear immediately on the ground” (15).

For Laugier, working from the model of his primitive hut, the column was the only proper means of bearing vertical loads. Walls were to be treated as infill panels, concerned solely with sealing up a spatial envelope. Engaged columns are only permitted as a “licence authorised by necessity” (16), but they must not be lost into the mass of the wall – they should be engaged “a fourth part at most… so that even in their use they may always retain something of that air of freedom and disengagement,” (16). For Laugier, parts must remain distinct, even while they form an integrated architectural body. They must be seen to be distinct (as the columns must be seen to be distinct from the wall); and they should need a minimum of concern for attachment: there is an expected natural co-dependence of parts. The disconnection of parts which Laugier encourages could be seen as a foundational principle for later tectonic conceptions of the joint, the role and expression of which became one of the central preoccupations of modernist architecture.

Laugier and Rousseau share more than an authorising appeal to a fictional primitive scene. Both idealise connections in the same way, envisaging a kind of joint which is held together without binding. Their respective edifices, social and architectural are complete wholes comprised of individual elements, which must remain free and discrete, even as they constitute this wholeness. Both edifices are only conceivable on the basis of a very particular mereology. This joint is primitive, in the sense that it is taken to emerge from primitive social and technical conditions. Although these conditions place the joint close to nature, the joint itself is not understood as natural, except insofar as rationalism is natural. For Rousseau, there are three joints: the paternal bond, the agreement, the chain. The first is natural and primitive, the second rational and natural, the third unnatural and irrational. The social contract is of the second of these orders. Laugier fumbles the question of origin by treating it over-literally, but he too seeks to authorise architectural production by demonstrating it to be a rational and natural assembly.

See, I told you it ended abruptly. If you actually read all the way to the end, free virtual cookies for you.