Water in eighteenth-century Paris

In order to introduce these water supply projects one should first briefly describe the water resources of eighteenth-century Paris. There were a few aqueducts in Paris, mainly the aqueduc d'Arcueil, coming from the south and constructed at the beginning of the seventeenth century. There were also two important hydraulic machines taking water from the River Seine, also dating to the seventeenth century, the pumps of Notre-Dame and the Samaritaine, installed on bridges. The aqueduct and pumps were still the main water infrastructures of the city at the beginning of the eighteenth century, but had been designed mainly to supply the royal palaces of the Luxembourg and the Louvre and their respective gardens, then secondarily a few Hôtels, palaces of the nobility and religious institutions. Only a small part of this water was conducted to a few public fountains, about sixty at the beginning of the century, a number that would increase to about a hundred at the end of the century.Footnote ⁵

In any case, the fountains only provided a small part of the water used in Paris. An important proportion was drawn from numerous wells, notorious for the bad quality of their water due to the vicinity of cesspools.Footnote ⁶ These wells were also the main water resource for business activity too distant from the river. But the greater part of domestic use was provided by water carriers. These were of two kinds. Those on foot carried two buckets of about fifteen litres each, taking their water from the fountains (in fact trying to control access to these fountains) or directly from the river. Another class of water carrier used carts and were able to transport huge barrels of river water throughout the city.Footnote ⁷

Thus the main water resource was the river and the main problems were the problems of the river: winter ice hindered access to the water and the functioning of machines; summer drought reduced the river to one-fifth of its flow, turning it into a small and muddy stream. Some scientists like Jussieu were able to link the state of the river in the summer and the outbreak of epidemics such as that of 1731.Footnote ⁸ Otherwise the flow of the river was generous, but the water was extremely polluted by human activities and waste. In the 1780s Louis-Sébastien Mercier could ironically present the Seine as a ‘stream as large and black as the Styx, thick and silty’.Footnote ⁹ There had always been critics of water quality, but in mid-century strong opposition to the Seine water emerged, provoking a reaction by such people as the famous apothecary Parmentier who championed this water.Footnote ¹⁰ At the same time, often in connection with the question of water quality, the theme of water lack began to appear as a motive for new water supply projects presented as indispensable to public welfare.Footnote ¹¹ In the second half of the century the monarchy and the municipality, which had the monopoly of water infrastructures in Paris, began to encourage private initiatives but kept a watchful eye on the quality of projects. Very few of the numerous projects proposed were finally deemed satisfactory and even fewer were carried out.Footnote ¹² There were two kinds of project corresponding to two attitudes towards the Seine's water: pumping in the Seine or diverting a distant river. The main example of the pumping solution was offered by the Périer brothers' company, which installed British steam engines in the 1770s and 1780s. But the company finally failed due to financial problems and, indirectly, to the resistance of the water carriers. No further pumps were built. The best example of the alternative solution was provided in the 1760s by Antoine Deparcieux, who proposed diverting the River Yvette in the south of Paris. Project costs were prohibitive and it was not carried out. A similar authorized project finally failed at the end of the 1780s.Footnote ¹³

This paper focuses on the two projects where the notion of need was most discussed. The first is the project of Deparcieux, the most acclaimed of all projects. The approach to needs proposed by Deparcieux was taken as an example by most project authors during the rest of the century. Secondly, and at greater length, I discuss the Ourcq diversion project, finally carried out by the Bonaparte government at the beginning of the nineteenth century, having appropriated the idea from a late eighteenth-century private company. This case is especially interesting, because it has occasioned many disagreements on the question of needs.

The classical approach to need: Deparcieux and the Yvette project

What was presumably the first approach to the water-need issue was offered by Antoine Deparcieux (1703–68), a member of the Académie des sciences, who proposed a project for the diversion of the Yvette, a small river about thirty kilometres south of Paris.Footnote ¹⁴ The degradation of the hydraulic machines inside Paris, in spite of important and regular repairs, led the city of Paris to ask the Académie des sciences in 1760 to examine these machines and propose some measures. With two other members Deparcieux took part in this consultation and together they proposed urgent changes in the machines themselves. But in 1762 Deparcieux presented a memoir to the Académie which proposed the replacement of the fragile machines by a diversion of the Yvette. He would try to convince the city and the state to pay for his expensive enterprise, but the project would not be carried out. Deparcieux would get involved in a public debate fuelled by proponents of alternative pumping solutions, especially about the costs of his project and the respective qualities of Yvette and Seine water.Footnote ¹⁵ The long and vigorous controversy around the Deparcieux project drew attention to this scheme, thus giving it a role as a standard against which rival schemes were judged. It became the model for all diversion projects. Deparcieux's approach to need was thus followed, with very few changes, by most of the entrepreneurs and engineers involved in similar projects until the beginning of the nineteenth century.

In his 1762 memoir on the Yvette project Deparcieux estimated the quantity of water necessary for a big city to be approximately twenty litres per day and per person.Footnote ¹⁶ He did not say how he had arrived at this number but deemed it an average need, ‘perhaps a bit too much for the ordinary citizen and too little for the big Houses’.Footnote ¹⁷ Supposing the number of people living in Paris to be 800,000, Deparcieux concluded that an average of 16,000 cubic metres a day was needed for the private use of water in the city. Deparcieux compared this need to an estimate of the available water resources. He reckoned that aqueducts and pumps provided only 4,600 cubic metres, of which part belonged to the king and part was undrinkable. Deparcieux therefore reduced the estimate of the water resources to 3,600 cubic metres, at best, because the machines on the river were frequently out of order. He presented the important gap between the water resources (3,600) and the need (16,000) as a major argument for his enterprise. In order to solve the issue of this alleged lack of water, Deparcieux suggested diverting the river Yvette, which could provide all year round a minimum of 24,000 cubic metres. The excess eight thousand cubic metres were to be used for other purposes, mainly the cleaning of the streets and firefighting.

Deparcieux's approach to need was extremely original. The idea that a certain quantity of water was necessary for individual uses was not new, though it was seldom expressed in terms of necessity or minimum, and seldom as a subject in its own right. Jaucourt, for example, describing various cistern techniques in the Encyclopédie in 1753, calculated that a house with a given roof surface under a given precipitation (corresponding to rather dry climates) would be able to collect over a year an average of two hundred pintes a day, which made about eight pintes per person (if one took twenty-five persons living in the house). This is about 7.5 litres per person. This quantity Jaucourt deemed to be ‘more than enough for all the uses of daily life’, a common statement.Footnote ¹⁸ But such ideas remained at the individual or family level. The main originality of Deparcieux is that he estimated a global need for a whole city. This of course was related to the originality of his enterprise: provision of water for the whole city. The concept of need would remain a special feature of later projects, mainly diversion schemes, intended to supply the whole city. The concept would only marginally be used in pumping projects, since these usually proposed only limited systems at the scale of one or two neighbourhoods that could, at least in principle, be extended and revised. By contrast, after Deparcieux diversion projects were usually concerned with the whole city. Interestingly, the concept of global need and the idea of providing water for a whole city appeared when authorities began to consider water a matter of public welfare – it was expected, at least rhetorically, that water supply projects should increase public welfare.Footnote ¹⁹ The concept of global need was in this respect quite useful, since it implicitly meant that the project involved care for all and water provision for everyone.

In Deparcieux's approach need only made sense in comparison with resources: the project was supposed to address a lack of water. The way the estimate of water resources was worked out is of particular interest. Significantly, Deparcieux did not take into account the main water resource of the period, namely the water carriers. It was probably impossible for Deparcieux to make an estimate of the number of these carriers and their capacity. This water resource was not measurable and Deparcieux only took into account what could be measured – the aqueducts and pumps. But the practical difficulty of measurement exactly fitted the aim of the project, since for Deparcieux the water carriers were not to be counted as a water resource. They did not appear in his calculations because his project precisely intended to do without them. Deparcieux wanted to develop the distribution of water in the city by the creation of numerous fountains. Therefore he was only interested in the existing fountains, not in other water resources that would have to be replaced. Deparcieux's project presupposed a social redefinition. The new water supply of Paris would get rid of a great number of the water carriers. Some of those on foot with buckets would be able to continue their activity, but only from the new fountains to the houses. They would no longer be a water resource but instead a mere link in a water distribution system centred around the fountains. The project, through its definition of the water supply, decided what had to be taken into account in the estimate.

In Deparcieux's approach, lack of water thus only made sense in regard to a certain idea of distribution which did not take into account the water carriers. A similar point can be made for most of the water supply projects of the late eighteenth century: the lack of water, as conceived by the authors of these projects, entrepreneurs, scientists or engineers, was worked out in relation to an idea of the distribution, and often water carriers were not part of that idea.Footnote ²⁰ Each project was in some form a measurement of need and at the same time provided a solution to that problem. The estimate of need depended strongly on the project itself, its purposes and choices, but in the case of Deparcieux this was not yet an explicit theme.

How the practical engineer saw need – Bralle's estimate

In later projects water resources were deemed very small in comparison with need, so that need finally became the sole subject of debate. This was especially the case in the final and largest eighteenth-century project, completed in the first years of the nineteenth century, the diversion of the River Ourcq. In 1799 two entrepreneurs, Solages and Bossu, proposed a new version of their diversion project of the Ourcq. They had put forward similar projects since the mid-1780s.Footnote ²¹ Like all their fellow entrepreneurs they stressed how serious was the lack of water in Paris to show the indispensability of their enterprise. But they had not provided any figures for this alleged lack.Footnote ²² So the new government of the Consulate, very interested from the start in projects for Paris water supply, which Bonaparte deemed politically useful, decided to consult an expert.Footnote ²³ François-Jean Bralle (1750–1831) had been the water specialist for the royal administrator of the province of Picardy in the 1770s, and then for the Paris area in the 1780s.Footnote ²⁴ He kept a similar position during the Revolution as chief engineer of hydraulics in Paris and was also responsible for the famous hydraulic machine of Marly which supplied water to Versailles.Footnote ²⁵

Bralle was asked by the government to decide as a specialist about the needs of Paris in order to ‘determine the quantity to be supplied by the company’, to be set out in legislation should the project be adopted.Footnote ²⁶ His estimate was supposed to have an advisory function but of a rather prescriptive nature, since the government had announced its intention to integrate the engineer's estimate into this law. But in his report of 10 December 1799, Bralle did not consider this task to be a simple one. On the contrary, it seemed to him very difficult to achieve without further instructions:

This is an important point. Bralle stated that there could be no possible estimate without a plan of action. He acknowledged what has already been noted in Deparcieux's approach. What was to be counted as need depended on the project, on one's intention and defined purpose. But Bralle went further, stating that this first intention had to come from the government. Only the government had the legitimacy to decide what should be done. This was not a new attitude in public works, where the state had often been the initiator of projects, generally through the Ponts-et-Chaussées, the corps of state engineers specializing in public works. But this had not been the case with water supply and waterways projects, in which the state had not previously been much involved. Such projects were usually proposed by private companies with no previous state intervention.

Bralle insisted on the government's prerogatives; the estimate of the engineer and the constraints thereafter imposed on the entrepreneurs should only reflect the government's choice. He thus gave a radical answer to the typical problem of expertise: should specialized knowledge or political legitimacy decide on technical projects?Footnote ²⁸ But this conception of the respective roles of government and engineer somewhat called into question the expertise demanded from him – an estimate which could be directly integrated into a law. In Bralle's view there was no engineer who could objectively determine the needs of Paris and no government which could take necessary measures by following the expert's conclusions. The whole process had to be worked out in the opposite manner: the government should first make a general choice between ‘prodigality and parsimony’, or even more precisely by analysing the possible uses of the water, before the engineer could really estimate Parisian needs. It was not the engineer who established the facts on which the government's decision would rely. It was the government's intention that enabled the engineer to establish the facts in accordance with that decision.

The nature of the engineer's activity and knowledge was being questioned here. Bralle's statement about the necessity of an initial intention only made sense because matters handled by engineers did not exist independently of such an intention, a will to act in a certain way. Water need was not a natural object to be simply measured by a specific method. On the contrary, it had to be defined and this definition depended on the intention one had. Knowledge about the world as it is differs from knowledge about the world as it ought to be. The Ponts-et-Chaussées engineers clearly preferred the first kind of knowledge, which consisted of establishing facts about the existing world. Bralle, an engineer with a more practical culture of hydraulics, independent of the state corps, was different in this respect. He considered that, in the context of a project, knowledge about things as they ought to be was essential; need only made sense in respect to a certain idea of the future, which should first have been defined.

Bralle's approach seems quite apparent. A project is related to the future, it deals with norms and ideals, not only with things as they are. In this respect one could follow Aristotle's distinction in his Rhetoric between arguments that aim at establishing facts (the judicial mode) and arguments about human affairs with a view to action under uncertainty, where the question is to decide what is best to be done in the future (the deliberative mode).Footnote ²⁹ One argues in different ways or speaks to different people depending on whether one is concerned with the present or the future. Yet the Ponts-et-Chaussées engineers collapsed this distinction. Bralle acknowledged that needs were not numerical facts that could be established objectively; they could not be determined independently of intentions and choices, ideals and hopes. He pointed out that the government had asked him to decide about the future, about the quantity of water that should be brought to Paris. Asked for his opinion on the administration of human affairs, Bralle did not refuse to answer. Having emphasized that it was in fact a government's prerogative, he expressed his opinion on what should be done. How did Bralle conduct his estimate? He began with a reference to authorities on that matter: ‘all the scientists’, he said, ‘consider that a big city needs approximately 20 litres per day and per person’. This was the classical figure proposed by Deparcieux, usually accepted in subsequent estimates. But Bralle was very sensitive to the difference between what is and what ought to be and made it clear that this figure was not an estimate of the present consumption of Paris:

Like Deparcieux, Bralle considered that the quantity of twenty litres per day per person matched average need, the quantity that anyone would consume were there no problems in getting water. These supply problems precisely accounted for the important difference between need and present consumption. But unlike Deparcieux, for whom twenty litres were an invariable need, Bralle considered that an increase in water resources would lead to an increase of consumption and that one should therefore think big. In Bralle's scheme twenty litres was rather a projection. Bralle definitely rejected the idea that present consumption could be a relevant measurement of what had to be done, a position very different from that of the Ponts-et-Chaussées engineers. From this individual need, supposing a population of 600,000 people, Bralle deduced the need of Paris as 12,000 cubic metres. This result differed from that of Deparcieux because Bralle considered a smaller population. In the absence of a reliable census one could only speculate on such a matter. Besides the inhabitants' needs, Bralle drew up a detailed inventory of other uses of water. These included the public buildings such as hospices, jails, the Museum d'histoire naturelle, and the Palaces of the Louvre, the Tuileries and so on, to a total of four thousand cubic metres. On the other hand, he reckoned up decorative fountains, horses’ watering places,Footnote ³¹ baths, washhouses and street and sewer cleaning, all ‘objects of pleasure and healthiness’, giving at least 40,000 cubic metres. The total need of Paris was therefore of the order of 56,000 cubic metres. The approach is quite remarkable: Bralle was the only engineer whose estimate of the need was based on a calculated inventory of possible uses. Though Bralle's estimate was presented as an argument in favour of his opinion that the ‘largest possible quantity’ of water should brought to Paris, he made it possible for the government to choose what uses should be preferred.Footnote ³²

The state engineers: Bruyère's estimate and the Ponts-et-Chaussées

Bralle's report argued in favour of a very large quantity of water. The government therefore decided to have the entrepreneur's project examined by the Ponts-et-Chaussées corps.Footnote ³³ This corps of state engineers was in charge of all works of roads and bridges, on which it exercised a complete monopoly over design and implementation. But before the French Revolution these engineers managed neither water supply nor waterways. Until then they were sometimes asked to produce, examine or even construct such projects, but they had no exclusivity and were not the main experts on these questions. The technical part was usually examined by the Académie des sciences or by an architect working for a city or a province. The Revolution gave the Ponts-et-Chaussées a larger responsibility, especially over waterways.Footnote ³⁴ But since the engineers had not before been specialists in these matters, their expertise could be challenged by others. This helps explain, for example, how Bralle could have been asked in the first place. The canal projects of the Consulate period reveal that the state engineers had no fixed way of dealing with canals. In many respects their knowledge was quite uncertain.Footnote ³⁵ The consideration is significant when one seeks to explain how and why these engineers would try to present themselves as the sole possible experts on these matters.

The chief engineer, Louis Bruyère (1758–1831), worked for a year on the verification of the Ourcq project. The verification of a project was usually limited to an analysis of the documents presented by the entrepreneur or engineer. It was usually made by the inspector in charge of the district where the project was to be carried out, then presented to the assembly of the corps, which decided collectively what to do. In some cases, however, some complementary measures, mainly levels, gauges or drills, were deemed necessary and an engineer chosen for this mission, such as Bruyère in this case. Bruyère clearly took advantage of the verification to proceed to a more detailed study and to conceive a project of his own. In his April 1802 report to the assembly of the corps, Bruyère declared the Ourcq diversion project impossible, mainly because of some levelling errors, and suggested a different idea: the diversion of the Beuvronne, a smaller river much closer to Paris than the Ourcq. Bruyère estimated its daily flows as between 20,000 and 34,000 cubic metres. With Bralle's estimate of need (56,000 cubic metres) the Beuvronne would have been insufficient. Indeed, Bruyère did not accept that estimate and made a new one that would posit a much smaller need. He rejected previous estimates, emphasizing their discrepancies and the interest of the various estimators, since they usually wanted to convince the government or their shareholders that they were going to make incredible profits. Bruyère insisted that a wise government should not tolerate any kind of swindle and that therefore ‘it [was] high time not to delude oneself’ on that matter.Footnote ³⁶

Bruyère made a distinction between two approaches. On the one hand there was the strategy of interested entrepreneurs who wanted to deceive the government with self-interested and necessarily unreasonable promises; on the other hand estimates were made by the figure of the state engineer who, representing the state, should measure reality and separate truth from error. The uncertainty of grandiose, exaggerated and interested speculations without any technical foundations was to be contrasted with the certainty of the state engineer, whose skills made it possible to measure all things correctly and to decide about possibility and benefit without exaggeration. In Bruyère's conception the state engineer was the one who could measure reality, who could make all estimates without reference to value judgements and intentions of any kind, especially political. Unlike Bralle, Bruyère thought an objective estimate not only possible but precisely the job of a state engineer.Footnote ³⁷ This was his competence, his activity, and he made it in the government's name. In fact, in this approach, there was no place for the government itself, because it could only be the victim of speculators who took advantage of its incompetence and technical ignorance. Working for the good of the state, the state engineer could and should take the place of the government, because he was the only one who could make the difference between a brilliant but unfounded proposition and the bare facts. Implicitly, nobody working outside the state, however knowledgeable, could possibly propose a project for the public welfare. This was very different from Bralle's conception of a strong government, not easily influenced and the sole legitimate decision-maker. In both cases, an explicit relation was made between the respective roles of the government and the engineer and the nature of the engineer's knowledge, especially the nature of his estimates.

What did Bruyère's allegedly objective estimate look like? ‘One has got into the habit of supposing’ approximately twenty litres per day per person, the classic figure of Deparcieux once again.Footnote ³⁸ But Bruyère considered this quantity to be ‘considérable’ and supposed that the people who had first made this estimate must have included all the collective uses of water: the cleaning of the streets, bathhouses and firefighting. Bruyère did not treat Deparcieux and his followers as self-interested entrepreneurs. He introduced the classic figure with a certain respect, but only because he could give it an unfavourable interpretation: he deemed twenty litres too much for the needs of the inhabitants alone. In order to estimate the real needs of these people, Bruyère applied a very simple principle, one very common among Ponts-et-Chaussées engineers of the time. He looked at current consumption: ‘It is a fact, and everybody can check it, that a family of 10 individuals only consumes an average of … approximately 69 litres a day’.Footnote ³⁹ One had only to look around. Facts were to be found in the real world. With an average individual need of less than seven litres, Bruyère came up with a total need for the inhabitants of Paris, supposed to be no more than 720,000, of five thousand cubic metres. Bruyère supported this small figure with an estimate of the distribution at the fountains in Paris. The similarity of the two figures for individual consumption and fountain supply should have been obvious, but once again the fountain flows were a quantity that could be measured in the present world; they were facts.

Bruyère knew the needs of Paris by looking at consumption. He did not intend to increase the available quantity of water, but only to improve its distribution and quality. He took into account a possible increase of consumption because if water became easier to get, people would probably use more, but he thought that an increase to seven thousand cubic metres from the present consumption of five thousand would be extreme. So he considered seven thousand to be an absolute limit for the inhabitants' needs.Footnote ⁴⁰ This was a small quantity compared to the 12,000 of Bralle or the 16,000 of Deparcieux. It made the volume of the Beuvronne, the river he intended to divert, appear more than sufficient. Indeed, Bruyère's 20,000- to 34,000-cubic metre estimate of its flows left a significant remainder for collective uses, such as street cleaning, bathhouses and so on, uses which Bruyère did not examine in any detail.

Bruyère wanted his estimate to be ‘reasonable’. One should not be misled by great expectations and should limit oneself as much as possible to facts and certainties, to things that could be measured. That is why he simply refused to take into account the quantity of water that should be diverted for purposes of ‘public amenities’.Footnote ⁴¹ These he considered to be ‘arbitrary projects’. Since an engineer only made objective estimates, he could only consider things which could be measured in the existing world, typically the consumption of the inhabitants, and had to leave out all the uses which could not be measured, uses which were innumerable and could be imagined to be as large as one wanted. ‘I will not even try to determine such quantity’, Bruyère declared. But neither did he leave it to the government to decide about such arbitrary uses: ‘It would certainly be very beautiful to see the waters gush out and flow in every public square; but before one devotes oneself to a grand luxury, one must first provide the most necessary thing’.Footnote ⁴² This priority was not to be understood as temporary. Bruyère left no place in his project for such prodigality. The impossibility of evaluating arbitrary needs coincided with the moral imperative to keep consumption to the essentials, to the minimum. By limiting public action to the essentials – that is, to what could be measured in the real world – Bruyère made sure of the key role of the state engineer in the decision-making process, leaving no space for government interventions. What could be undertaken was limited by the competence of the engineer. This objectivist approach to the estimate of need led to a curious paradox which was in fact typical of these Ponts-et-Chaussées engineers around 1800. They often preferred the world as it is, since things can be measured, to all that could be, rejected as arbitrary, decidable without measure. In order to keep control over the decision they tried to limit discussion about the future to the measurement of the present. It was an interesting paradox for engineers whose central activity was to deal with projects, with anticipation, construction of the future. But they also had a very simple solution to this paradox.

Need or available quantity?

Bruyère's estimate and his concept of the engineer's activity would have considerable influence inside the Ponts-et-Chaussées during the Ourcq canal debates. Even before he finished his report, the government decided to divert the Ourcq. It is unclear if Bruyère's study had been forgotten by the authorities, since he had been working on it for nearly a year, or if they could wait no longer. Bonaparte had ordered the creation of large construction sites in the vicinity of Paris in order to provide work for the destitute. The government decided to construct an Ourcq canal. It did not allow the company of Bossu and Solages to execute their project. The Ponts-et-Chaussées engineers had provided a technical problem, a levelling error, to which the administrators added a legal one. The government took over the idea of the Ourcq canal and in 1802 appointed a Ponts-et-Chaussées engineer as head of the project: Pierre-Simon Girard (1765–1836). Given the prestige of the project, Girard was relatively young. He clearly owed this favour to his extreme loyalty during the final days of the Egypt expedition, of which he had been a member. His privileged status and his unusual approach to the project soon led to a major crisis in the Ponts-et-Chaussées corps, a quarrel among the engineers about what should be done about this diversion.Footnote ⁴³ Among the many issues involved, a few examples can be given. Should the Ourcq really be diverted, as the government ordered, or rather the smaller and nearer Beuvronne? What route should the diversion take, bypassing the major hills near Paris or cutting through them in monumental straight lines, as proposed by Girard? Could one use the new mathematical theory developed by Girard in order to establish the slope of the canal? Should the new canal be navigable, besides its function as a water-supply canal? This last issue was rather important in relation to the question of need, for the river Ourcq was an important waterway for bringing firewood to Paris and if the river were to be diverted entirely then the new canal would have to take over this navigation function.

In their various reports and memoirs the engineers discussed the advantages and disadvantages of the different quantities of water that could be brought to Paris.Footnote ⁴⁴ In a manner similar to that of Bralle, they considered need in relation to the possible uses of water. Girard, for example, mentioned the cleaning and the watering of the streets, embellishment by gushing fountains and the functioning of factories and mills.Footnote ⁴⁵ But unlike Bralle's evaluation, no figures were suggested for the various possible uses. This was not very surprising for some engineers like Girard, Gauthey or Bertrand, who held the radical idea that Paris could never get too much water. One should divert the largest possible quantity, because the possible uses were innumerable and it would in any case be impossible to satisfy them all. These engineers were fascinated by visions of an overflowing city, with water everywhere, described as giving life to the city like the veins and arteries in a body.Footnote ⁴⁶ While some opponents spoke of excess and evoked the difficulties of dealing with superfluous water, the enthusiasts replied with infinite needs for irrigation in the plains north-east of Paris. In a word, however much water there was, there would be use for it. For these engineers, because the uses were innumerable it was not necessary to make a precise estimate of the quantity needed for each of them for the present. Thus the estimate was made the other way round: one did not measure the quantity needed, but the available quantity. In this approach discussion about the relative importance and advantages of different uses, which should be preferred to others, was postponed for the future.

However, it is worth noting that this reasoning about available water was not specific to the engineers who were in favour of abundance. Even those engineers who claimed to know exactly the needs of Paris, and refused to divert more than was required, proposed to divert all the water available in the river they preferred. The influence of Bruyère was obvious and often explicit in these debates about the Ourcq canal. During this controversy not one engineer proposed an inventory with detailed figures of the different uses, as Bralle had done some years before. But many followed Bruyère's approach and made need coincide with the water available in the river of their choice, mainly the Ourcq or the Beuvronne. One finds only two kinds of estimate in these reports and memoirs. Sometimes protagonists offered an estimate of the consumption of the inhabitants, usually reduced to Bruyère's very small estimate of present consumption, making the bigger Ourcq seem unreasonably large. Sometimes one encounters the estimate of the water available in the river to be diverted, often with very important differences in the estimates, since the flow measurements were not standardized.Footnote ⁴⁷ The coincidence between estimated need and the water resources of the river was sometimes justified.Footnote ⁴⁸ Indeed, if one started such an important project, the costs and difficulties would not be so different if one diverted only half a river or the entire flow. If one went so far to have a supply of water, then one should take it all; it was better to take maximum advantage of the project, even if the costs were greater.

Bruyère's influence can be seen in this shared attitude. The engineers only estimated quantities that existed in the present. Their estimate of the future needs of Paris was reduced to the present consumption of the inhabitants and to the flows of the rivers. Instead of starting a discussion about the relative advantages of the different uses, they preferred to choose and measure a river. The future they constructed, the possibilities, were in the world of the present. What had to be discussed was the choice of the river that would be entirely diverted. This choice was very much related to other issues: technical advantages of diverting one river or the other, ease, cost, navigation, quality of the water and so on.Footnote ⁴⁹

Conclusion

The Ponts-et-Chaussées engineers preferred Bruyère's objective approach: establishing facts and making decisions based on facts. This approach made it possible to reject as arbitrary anything that could not be measured and therefore submitted to their control. If this approach were indeed objective, the engineers' estimates nevertheless relied on an initial intention, a preference for a given solution. An engineer who made his estimate of Parisian need coincide exactly with the river he wanted to divert thus produced an estimate that fitted a purpose. He made things as they are (the river and its flows) match things as they ought to be (need). But unlike Bralle's idea of the initial political – that is, governmental – intention, this strategy, being technical, depended completely on the engineer's competence and not on that of the government. There had to be intentions, but in the case of the Ponts-et-Chaussées engineers these intentions had to be reduced to factual technical matters to appear outside the realm of politics and to secure control over technical decisions for these engineers.

One can thus complete Alder's very general statement about engineering. What ought to be is a political problem, in the broad sense of what concerns a community as a whole. It is a problem of human affairs, in Aristotle's formulation, which are uncertain, strongly related to norms, ideals and hopes. Engineering is the use of knowledge about the world as it is, about things as they behave, and in many cases, as in the sciences, this does not exclude speculations about how things behave, in order to create a future object. Engineering is thus the articulation of knowledge about things as they are and reflection about what they should be. In this articulation lies its mixed nature, balanced between techno-scientific facts and politics. The ways engineers presented this articulation thus reveal the image of the role they wanted to impose: the reduction of technical enterprise solely to factual terms. Ponts-et-Chaussées engineers could pretend to have no political role at all, while they were indeed taking over the complete decision-making process. It would be interesting for an engineer to present himself as respectful of governmental prerogatives, to emphasize that he did not want to decide about political matters. Like Bralle, he thus appeared as a credible neutral expert, dealing with facts but not taking decisions. The Ponts-et-Chaussées attitude was exactly the opposite: they pretended that there was nothing to decide politically. They even claimed that a political decision on these matters would be useless, since one could not order nature and since the limits of nature were only known by skilled engineers.Footnote ⁵⁰

French state engineers, the Ponts-et-Chaussées especially, are well known for having had a very strong position inside the state and large control over technical decisions.Footnote ⁵¹ But this was not a naturally given situation for which these engineers did not have to fight. They were challenged from outside the state by entrepreneurs and architects who constantly claimed that they could perform the same work better and more cheaply, and who took advantage of every crisis or failure of the corps to call for reform. But they were also challenged from inside the state; the government did not always accept the control its engineers tried to keep over technical decisions. During the French Revolution the Assemblée législative almost abolished the corps, which was denounced by some as a guild, an institution appropriating part of the power that legitimately belonged to the nation's representatives. The state tried, especially through its Committee for Public Works, not to leave much to the decision of the engineers and to treat their work as mere implementation. At its start, Bonaparte's government adopted the opposite attitude, emphasizing the technical nature of the engineers’ activity and leaving them with considerable autonomy. In March 1801, when asked to pronounce on the best canal to build in order to improve navigation in northern France, the assembly of the Ponts-et-Chaussées proposed two solutions to Bonaparte, leaving it explicitly to him to decide. He returned the question to them, speaking of their ability to consider the ‘universality of circumstances’.Footnote ⁵² But this autonomy did not last long and Bonaparte soon adopted a more technical conception of the corps.Footnote ⁵³ This tension between the state and an administration sometimes perceived as dangerously autonomous continued during the nineteenth century. The position adopted by the Ponts-et-Chaussées engineers in the Ourcq case must thus be seen as a collective strategy to defend their ‘rights’, their control over decisions that the state did not always agree to let them manage.